Dummett's Game of Tarot, 1980, a few chapters

It seems to me that Dummett's 1980 book is absolutely basic to all our work, yet not easily obtainable without spending a lot of money, especially for someone new to the field who isn't yet sure if it's their "thing". To remedy that and even the playing field I want to simply post a few chapters. The most urgent are chapter 4, "When and where the tarot pack was invented", and chapter 20, "The order of the tarot trumps" . (To skip chapter 4, use the "find" function on your browser for "chapter 20" and eventually you will get there.) The original is in two columns. I am not going to take the time to change the column width, which also might introduce errors. The chapter starts on p. 65. If any copyright holder of the work protests, I will remove the post.

Addition: As of Nov. 26, 2017, I have added one more chapter, Chapter 21, "The early Italian game".

Addition: As of May 31, 2020. I have added chapter 7, "The Game of Tarot" , and chapter 8, "General features of the game".


When and Where the Tarot Pack was Invented

It was formerly believed that, in Europe, the
Tarot pack is as old as the regular pack. Indeed,
some have thought that it is older: the assertion is
still to be met with that the regular pack was
originally derived from the Tarot pack by
subtraction. It is clear from our study of the
relations between European and Oriental cards
that this latter belief must be incorrect. The
regular pack came to Europe from the Islamic
world, but the Tarot pack is a European
invention: there is no trace of the existence in
Egypt, Persia, India or China of anything in the
least resembling the Tarot pack.(1) This naturally
leads us to expect the Tarot pack to have
appeared some time later than the introduction
of playing cards into Europe: as a variant on the
ordinary type of playing-card pack, it would
hardly have been devised until the novelty of the
latter had had time to wear off. One ground that
used to be advanced for the contrary hypothesis
was the belief that the word naibi referred to
Tarot cards, while carte, cartule, etc., referred to
cards of the regular pack: but this belief was
conclusively refuted in 1900 by Robert Steele,
who showed that Italian naibi, like Spanish
naipes, was used simply to mean 'playing cards',
Tarot cards being known in fifteenth-century
1. Once again, an exception should be made for the 'Chad'
cards of Mysore. These were devised by Krishnaraj Odeyar
(1794-1868) after his deposition in 1830 by the British from
the throne of Mysore. Although several of the special forms
of pack he devised are augmented packs, in the sense of
regular packs to which additional cards, not belonging to
any suit, have been added, it is obvious that so late an
addition to the repertoire of Indian playing-card packs has
no historical significance. See Rudolf von Leyden, Chad: the
Playing Cards of My sore (India)
, privately produced, 1973.

[new column]
Italian as trionfi. (2) More recently, Mr Jan
Bauwens has claimed that a pack of playing
cards recorded in the Register of Duke
Wenceslas of Brabant as having been bought for
the Duke and Duchess was a Tarot pack, on the
ground that it contained 78 cards;(3) but a
reference to the original entry reveals that neither
it nor any of the numerous later similar entries
contains any mention of the number of cards in
the packs bought or played with, nor anything
else to suggest that these were not
straightforward regular packs. (4)

Much more frequently met with as an
argument for an early date for the invention of
the Tarot pack is that relating to a famous
fragment of a fifteenth-century hand-painted
Tarot pack in the Bibliotheque Nationale in
Paris: this comprises seventeen cards, namely the
Jack of Swords, the Fool and all the usual
triumph cards except the Bagatto, the Popess,
the Empress, the Wheel, the Devil and the Star.
The Abbe Menestrier published in 1704 an entry
from the account-book of King Charles VI of
France recording the payment in 1392 of '56 sols
parisis' to the painter Jacquemin Gringonneur
for three packs of playing cards.(5) In 1842, M.C.
2. Robert Steele, 'A notice of the Ludus Triumphorum and
some early Italian card games', Archaeologia, vol. 57, 1900,
pp. 185-200. J
3. In a booklet accompanying a reproduction of the
Mamluk pack from Istanbul published in 1973 by S.A.R.L.
Aurelia Books, of Louvain and Brussels.
4. The entry is cited in A. Pinchart, Recherches sur les cartes à
jouer et leur fabrication en Belgique
, Brussels, 1870.
5. Menestrier, 'Des Principes des sciences et des arts
disposes en forme de jeux', Bibliotheque curieuse et instructive de
divers ouvrages anciens et modernes de litterature et des arts
, vol. II,
Trevoux, 1704, p. 174.

66 Part I: History and Mystery
Leber proposed that the cards in the
Bibliotheque Nationale came from one of the
packs painted by Gringonneur; (6) and this opinion
won such wide acceptance that the cards came to
be known as the 'Tarots de Charles VI'. If this
were correct, they would be by far the oldest
surviving Tarot cards; and, not only should we
have to say that the Tarot pack came into
existence within two decades of the arrival of
playing cards in Europe, but France would
appear to have a better claim to have been the
country of their origin than Italy. In fact,
however, there is no shred of evidence to connect
the Bibliotheque Nationale pack with
Gringonneur: Chatto, (7) Merlin (8) and
D'Allemagne (9) all ascribe the cards to Italian
workmanship. W.L. Schreiber is very specific,
assigning them to Ferrara in the third quarter of
the fifteenth century.(10)

Another piece of evidence cited in a great
many books and articles on playing cards was first
presented by Count Leopoldo Cicognara in his
book of 1831:(11) a portrait in Bologna, bearing the
inscription 'Francesco Antelminelli Castracani
Fibbia, Prince of Pisa, Montegiori and Pietra
Santa, and lord of Fusecchio, son of Giovanni, a
native of Castruccio, Duke of Lucca, Pistoia,
Pisa, having fled to Bologna and presented
himself to Bentivogli, was made Generalissimo of
the Bolognese armies, and was the first of this
family, which was called in Bologna "dalle
Fibbie". He married Francesca, daughter of
Giovanni Bentivogli. Inventor of the game of
Tarocchino in Bologna, he had from the XIV
Reformatories the privilege of placing the Fibbia
arms on the Queen of Batons and those of his
wife on the Queen of Coins. Born in the year
1360, he died in the year 1419.' On the strength
of this inscription, Count Cicognara named
Castracani Fibbia as the inventor of the game of
tarocchi. Commenting on this, Carlo Lozzi cited
6. M.C. Leber, Etudes historiques sur les cartes à jouer',
Memoires de la Societé des Antiquaires de France, new series, vol.
6, 1842, pp. 256-348.
7. William Andrew Chatto, Facts and Speculations on the
Origins and History of Playing Cards
, London, 1848.
8. R. Merlin, L'Origine des cartes à jouer: Recherches nouvelles
sur les naibis, les tarots et sur les autres éspèces de cartes,
Paris, 1869.
9. Henri-Rene D'Allemagne, Les Cartes à jouer du XIVe au.
XXe siecle,
two volumes, Paris, 1906.
10. W.L. Schreiber, Die altesten Spielkarten, Strasbourg 1937,
p. 101.
11. Leopoldo Cicognara, Memorie Spettanti alla Storia della
, Prato, 1831.

[new column]
with approval an entirely just observation by L.
Zdekauer that the inscription does not attribute__
to Fibbia the invention of the game of tarocchi in
general, but only of that particular variety of it
known as tarocchino and peculiar to Bologna.(12) As
we shall see, the diminutive ending relates to the
use in this variant game of a shortened pack, in
which the 2 to 5 are omitted from every suit. Quite
evidently, such a shortened pack must be derived
from the full 78-card pack, and not the other way
around, so that, if Francescoo Fibbia really had
invented the tarocchino pack some time before his
death in 1419, the ordinary Tarot pack from
which it was derived must have been in existence
for a certain period before that: hence, if the
inscription is to be believed, the Tarot pack must
have been devised by 1400 at the very latest.

Doubt was cast upon the very existence of this
painting by Robert Steele in his article of 1900, (13)
and in this he was followed by Miss Gertrude
Moakley in her book.(14) However, in another
article written in the very next year, Steele
acknowledged its existence, speaking of 'the
famous inscription on the portrait of Castracani
Fibbia (and stating that 'the portrait is now in
the Palazzo Pallavicini in Bologna'.(15) It is not to
Steele's credit that, in this article, he did not
mention and withdraw his accusation against
Count Cicognara. The existence of the portrait
was confirmed by G.B. Cornelli in an article of
1909.(16) It is somewhat surprising that doubt
about a point so relatively easily investigated
should have been allowed to persist for so

In fact, the portrait does exist, and tallies
completely with Count Cicognara's description
of it, including the inscription.. It is, however,
far from being contemporary with its subject;
by its style, it is to be assigned to the seven-
12. Carlo Lozzi, 'Le Antiche Carte da Giuoco', La
, vol. 1, 1899-1900, pp. 37-46.
13. R. Steele, op. cit.
14. Gertrude Moakley, The Tarot Cards Painted by Bonifacio
Bembo for the Visconti-Sforza Family
, New York, 1966.
15. Robert Steele, 'Early playing cards, their design and
decoration', Journal of the Royal Society of Arts, vol. 49, 1900-
1901, pp. 317-23; see p. 319.
16. G.B. Cornelli, 'II Governo "Misto" in Bologna dal 1507
al 1797 e le Carte da Giaoco del can. Montieri', Atti e
Memorie delta Reale Deputazione di Storia Patria per le Provincie
delta Romagna
, series 3, vol. 27, 1909; see p. 3.
17 See M. Dummett, 'A note on Cicognara', Journal of the.
Playing-Card Society
, vol. II, no. 1, August 1973, pp. 14-17
(original issue), pp. 23-32 (reissue), and 'More about
Cicognara', ibid,, vol. V, no. 2; November 1976, pp. 26-34.

When and Where the Tarot Pack was Invented 67
teenth century, and probably to the second half
of that century. As recorded by Count Cicognara,
it shows Prince Fibbia holding a pack of tarocco
cards, some of which are falling to
the floor: among them can be seen the Queen
of Batons, bearing the Fibbia arms, and the
Queen of Coins, bearing the Bentivoglio arms.
The inscription is as quoted by Cicognara;
but it appears that the original inscription was
painted over and a new version painted on top,
the original one having lacked the sentence
ascribing to Prince Fibbia the invention of
tarocchino and recording the privilege granted to
him of placing his arms and those of his wife on
the two Queens. The sentence may have been
added to explain the presence of the playing
cards in the picture.(18)

A tarocco bolognese pack in the British Museum,
18. The portrait can be seen at the palazzo Fibbia, 14, via
Galliera, Bologna. I am indebted for its location to the kind
help of Signor Giancarlo Roversi, an expert on the history of
the city. The palazzo was formerly known as the palazzo
Felicini-Calzolari; in Sandro Chierichetti, Bologna, Bologna,
n.d., p. I l l , it is stated to have been built in 1497. It was
referred to by Count Cicognara as the 'casa Fibbia', and
was said by Comelli in the article cited above to have passed
from the Fibbia to the Fabbri family, and from them to the
Pallavicini; the casual remark by Steele, cited in the text,
should not mislead anyone into looking for the painting at
the palazzo Pallavicini, 45, via S. Stefano. The owner of the
palazzo Fibbia kindly allowed my friend Signor
Marco Santambrogio, a lecturer in the Philosophy
Department at the University of Bologna, not only to
examine, but also to photograph, the painting; the great
hall in which it hangs is now occupied by the Associazione
Artigiani, who were also most co-operative. I owe- my
information about the painting entirely to the assiduous
work of Signor Santambrogio. In his The Encyclopedia of
(New York, 1978), p. 33, Stuart R. Kaplan cites my
article 'More about Cicognara', saying that I there describe
the 'rediscovery' of the portrait by Signor Santambrogio.
The quotation marks are Mr Kaplan's, and suggest a direct
quotation from my article, but in fact I did not use the word
'rediscovery', and claimed nothing so portentous on Signor
Santambrogio's behalf. The painting was never lost, but,
ever since Cicognara first described it, has remained
continuously just where he said it was. Robert Steele, in his
article of 1900, and, misled by him, Miss Moakley in 1966
expressed unjust doubts whether it existed; but, since
neither of them, at the time of writing, had actually looked
for it, this hardly counts as the painting's being lost. There
is in the British Museum a complete Tarocco Bolognese
pack by the maker who used the trade-name 'al Mondo'. In
this pack, Moors replace the Papi, so it must be dated after
1725 (see Chapter 16); it exemplifies the standard pattern,
in a single-ended form and without numerals on the trumps,
and is probably to be dated to about 1750. This pack
displays the feature mentioned in the inscription on the
Fibbia portrait: the Queen of Coins holds a shield with the

[new column]
probably dating between 1725 and 1750, bears
out the statement that, in some such packs, the
Queen of Batons bore the Fibbia arms and the
Queen of Coins those of the Bentivoglio family.
The portrait testifies to the existence, in the
seventeenth century, of a local tradition. But,
because of its late date, its evidential value is
slight; in view of the lack of any other evidence
for the existence of the shortened tarocco bolognese
pack before the sixteenth century, the tradition is
unlikely to be sound. As we have seen, it was not
until the sixteenth century that the practice of
playing various card games with shortened packs
came into fashion; it is therefore probable that it
was in that century that the shortened tarocco
pack used in Bologna was first devised. The most
likely explanation is that the reason for putting the
Fibbia arms on one of the cards had been
forgotten, and that the story about Francesco
Fibbia was invented as a hypothesis to account
for it.

We have successively rejected the years 1377
(on the naibi argument), 1379, 1392 and 1419 as
bounds for the date of the invention of the Tarot
pack: one that cannot be shaken is the year 1442.
In that year there is a reference in the Registro del
for the court of Ferrara to pare uno de carte
da trionfi
, and, in the Registro di Guardaroba, one to
the purchase of quattro paia di carticelle da trionfi.^
As was remarked above, the word trionfi, or the
phrase carte da trionfi, is the ordinary fifteenth-
century Italian term for Tarot cards, while, as
in early English sources, the word 'pair' (paro or
paio) was often used to mean 'pack'. Evidently,
then, by 1442, al; least in the d'Este court at
Ferrara, Tarot cards were well known and in
some demand.

That this was also so in Milan may be inferred
from a mural painting known as 'The Tarocchi
Players' in the Casa Borromeo in that city. It
forms one of a set of three, in the International
Gothic style, on the walls of a small ground-floor
room (-now used as an office), showing young
Bentivoglio arms and the Queen of Batons one with the
Fibbia arms. The pack is 1-37 in F.M. O'Donogiiue,
Catalogue of the Playing Cards bequeathed to the Trustees of the
British Museum by the late Lady Charlotte Schreiber
, London,
19. See G. Bertoni, 'Tarocchi versificati', Poesie leggende
costumanze del medio evo
, Modena, 1917, p. 218, fn. 3, and G.
Campori, 'Le carte da gioco dipinte per gli Estensi nel sec.
XV, Atti e Memorie delle Deputazioni di Storia Patria per le
Provincie modenesi eparmensi
, vol. 7, 1874, p. 126.

68 Part I: History and Mystery
men and women of the nobility engaged in
various games. There is no agreement over which
artist painted these delightful pictures, but they
are generally dated to the early 1440s. Every
writer on art who mentions these paintings refers
to the one in which we are interested as 'The
Tarocchi Players', so that this identification of its
subject must rest on a very firm tradition. There
is nothing in the paintings as it is now to show
whether the five people depicted are playing a
game with Tarot cards or with a regular pack;
one can see the pattern on the backs of the cards,
but although the faces of two of the cards must
originally have been shown, no details of these
can any longer be seen. It is evident, however,
that the condition of the painting has greatly
deteriorated during the present century. A black-
and-white photograph of it appears in a book of
1926, (20) and shows details that have now
vanished. As far as I can see from this
photograph, the card that has just been played
by the lady in the middle of the group is the 2 of
Coins, while the man on her right is playing the
Ace of Coins; this, of course, does not help us to
decide whether they are playing with Tarot cards
or not. However, it also looks from the
photograph as though the ladies at the two ends
of the group have each put a card face up in front
of them on the table, and that these are picture
cards: if so, all trace of these cards has since
disappeared from the painting. I have not been
able to identify these cards from the photograph;
but it is possible that, when the painting was in a
better state of preservation, one or other of them
could be seen to be a triumph card, the Matto or
a Queen, thus justifying the particularisation of
the game depicted as one played with Tarot
cards; if that were not so, it is difficult to see why
the painting should have acquired its name.
Signor Vito Arienti has informed me that there is
another fifteenth-century painting of players of
tarocchi in a castle in the Val d'Aosta. He may
have been referring to a painting in the castle of
Issogne, showing people playing various games,
including three playing cards, and dating from
1470. From the illustration I have seen, in Giulio
Brochard, Valle d'Aosta, ed. Renato Willien,
Novara, 1968, p. 76 (see also pp. 91-2), it is not
evident that the cards being used are Tarot cards;
in any case, it is too late to have any bearing on the
20 Raimond van Marie, The Development of the Italian Schools
of Painting
, vol. 7, the Hague, 1926, p. 145, fig. 91.

[new column]
date of origin of the game.

A great many playing cards have come down
to us from fifteenth-century Italy. Of these, many
are sumptuous hand-painted cards made for the
nobility. The surviving cards of this kind come
from about twenty different packs: it is difficult
to give a precise figure, since some cards in
different collections may originally have
belonged to the same pack. There are nine such
packs of which more than ten cards survive: the
surviving cards of eight of these nine packs include,
in each case, at least one triumph card and
at least one suit card, so that these eight packs
were certainly Tarot packs. The three most
complete of these packs are attributed, in the
unanimous opinion of present-day art historians,
to the Cremonese painter Bonifacio Bembo,
who was born about 1420 or a little earlier and
died in about 1480. Bembo is known to have
executed several important commissions for
Francesco Sforza, who became Duke of Milan in
1450 and died in 1466, and for his successor
Galeazzo Maria Sforza, who died in 1476. On the
strength of the heraldic emblems and mottoes
appearing on many of the cards of these three
packs, it is evident that they were made for
Francesco Sforza or, in the case of the first two,
for his predecessor Filippo Maria Visconti, who
died in 1447. They are as follows.

(1) The earliest is that usually known as the
Visconti di Modrone pack, from the name of its
former owner; it is now in the Beinecke Library at
Yale University. Sixty-seven cards survive, of which
eleven are triumph cards and fifty-six are suit cards.
In the Batons suit, the numeral cards show arrows
instead of the usual staves, although the court cards
show staves, in the usual form of polished staffs. On
the numeral cards, both the Batons and the Swords
intersect, but the Swords are straight. Because the
composition both of the court cards and of the
triumphs show certain unusual features, they will be
discussed in detail below.

(2) Probably the next in date is that known as the
Brambilla pack, also called after a former owner, now
in the Brera Gallery in Milan. Forty-eight cards
survive, of which only two - the Emperor and the
Wheel of Fortune - are triumphs, the remaining fortysix
being suit cards. Here the numeral cards show
ordinary Batons, while the court cards of that suit
have arrows: Batons and Swords both intersect on the
numeral cards, but the Swords are curved in the usual
Italian manner.

When and Where the Tarot Pack was Invented 69
(3) The most complete of all the early hand-painted
packs is that usually called the Visconti-Sforza pack,
divided between the Pierpont Morgan Library in New
York, the Accademia Carrara in Bergamo and the
private collection of the Colleoni family in the same
city. Of this, as many as seventy-four cards survive
altogether, comprising the Fool, nineteen triumphs
and fifty-four suit cards. All the Batons are of the
usual type, and intersect, as do the Swords, which are,
however, straight, as in the Visconti di Modrone pack.
The subjects on the triumph cards are standard ones,
of which only the Devil and the Tower are missing.
Six of them, however - Temperance, Fortitude, the
Star, the Moon, the Sun and the World - are quite
obviously by a different artist, and are thought to have
been painted some twenty years later, by an unknown
artist of the Ferrarese school.(21) This particular pack,
or individual cards belonging to it, appears to have
served as a model for the painters of more than one
later pack.

The remaining six packs comprising more
than ten surviving cards are as follows.

(4) The most famous early Tarot pack of all is the
so-called Charles VI pack in the Bibliotheque
Nationale in Paris, already mentioned. This
comprises seventeen cards, of which only one, the
Jack of Swords, is a suit card: the rest consist of the
Fool and fifteen triumph cards, making up all the
standard subjects other than the Bagatto, the
Empress, the Popess, the Wheel of Fortune, the Devil
and the Star. The vivid, florid style differs completely
both from that of Bembo and from that' of the
unknown painter of the six later cards in the Visconti-
Sforza set; expert opinion, however, assigns the pack
to the same date and place as the latter, namely to
Ferrara in about 1470.

(5) The most complete set other than the three by
Bembo is one in the Rothschild Collection in the
Louvre, consisting of thirty-one cards. It is generally
accepted that a single card, a Cavalier of Swords, in
the Museo Civico at Bassano also belongs to this
pack, bringing the total to thirty-two. Despite a slight
divergence in the measurements cited for this card
(190 x 90 mm. as against 188 x 90 for the Rothschild
ones), this identification can scarcely be doubted: not
only the general style, but the border design, the
overrunning of the border and the arches in the top
corners all resemble the Rothschild cards, while the
trappings of the horse tally exactly with those of the
Rothschild Cavalier of Batons, and the curious
tortoise-back shield with those on the Rothschild
King and Queen of Batons. In this set, however, only
21 See Ron Decker, 'Two Tarot studies related', part III,
Journal of the Playing-Card Society, vol. IV, no. 1, August 1975,
pp. 46-52
[Transcriber's note: the measurement of the Bassano card is given as 189x90 in the Nov. 2016 Catalog Giovanni dal Ponte, Galleria Accademia, Florence, where it is part of their exhibition.]

one triumph card survives, the Emperor; the rest are
suit cards. (In his The Encyclopedia of Tarot, New York,
1978, Stuart R. Kaplan suggests that another card,
shown by him at the top right of p. 121, is also a
triumph, the Pope, the Hermit or the World; it is,
however, surely the Jack of Coins, though admittedly
a bearded Jack is a rarity. Some writers have
questioned whether the twenty-three numeral cards,
whose measurements Detlef Hoffmann gives as 186 x
93 mm., belong with the other eight Rothschild cards,
which measure 185 x 90 mm. according to Hoffmann,
and it is true that their borders do not have the wavy
lines found on the court cards and the Emperor. The
measurement criterion would be conclusive, save that
discrepancies between measurements made by
different individuals are exceedingly common.) The
general treatment, though not the individual style, is
highly similar to the Charles VI cards, and the two
packs are probably to be assigned to the same milieu.
The Swords on the numeral cards are curved.

(6) Another pack, considerably smaller in
dimensions than those so far mentioned, appears also
to have originated from Ferrara and to have been
made for the d'Este family who were Dukes of that
city; it is now also in Beinecke Library at Yale. It
consists of sixteen cards, comprising eight court cards,
the Fool and seven triumphs - the Bagatto, the Pope,
Temperance, the Star, the Moon, the Sun and the
World. The d'Este arms appear on the Queens of
Batons and Swords and the Cavalier and Jack of
Batons (the King of that suit has not survived). The
arms of the King of Naples appear on the King and
Cavalier of Swords. The style again differs from any of
the preceding packs, but has more affinity with that of
the Charles VI cards than with those by Bembo.

(7) A pack consisting of fifteen cards is in the
Museo Civico of Catania, housed in the Castello
Ursino. Eleven of them are suit cards, including the 7
and 8 of Swords with curved intersecting Swords: the
remaining four consist of the Hermit, the Chariot, the
World and one whose identity is dubious. This last
shows a naked girl reclining on a stag, wearing a coral
necklace. In her left hand she holds an object which,
since it is painted in gold on a gold background, is
difficult to decipher; in her right hand, which is
suspended above the left one, she holds another
object, also painted gold against the gold background,
which, when I saw the cards, I took to be a fan. Mr
Ronald Decker has, however, suggested to me that
she is pouring from one vase into another, *which
would identify her as Temperance: this is the only
interpretation of this otherwise mysterious figure that
I have come across. The Hermit and World cards
closely resemble those of the Charles VI pack: the
latter shows a female figure standing on a globe
holding an orb in her left hand and swinging a censer
in her right; the corresponding card in the Charles VI

70 Part I: History and Mystery

set differs principally in that the female figure holds a
sceptre in place of a censer. It thus seems reasonable
to assign this pack also to Ferrara.

(8) A set of thirteen cards described and illustrated
in full in an article published in 1954 has since largely
disappeared from public view. They were at one time
all in the possession of Mr Piero Tozzi of New York:
one (Temperance) is now in the Museum of Fine Arts
in Montreal, and another (the Jack of Cups) was in
the F. Cleveland Morgan collection in the same city,
and is stated by Stuart R. Kaplan, op. cit, p. 100, to
have passed into the ownership of Mr Cleveland
Stewart-Patterson, presumably also of Montreal.
According to Kaplan, the remaining eleven were sold
in the early 1960s to a collector in Milan. The cards
were evidently made for some member of the Sforza
family, and all but one are copied, with some
deliberate divergences, from the Visconti-Sforza pack.
Their measurements were given in the article as 170 x
70 mm., but, as pointed out by Mr Ronald Decker,
this can be seen from the full-size reproductions to be
an error: it should be 170 x 87 mm. The set consists of
one card showing only the Visconti/Sforza emblem of
a crowned serpent swallowing a woman, one numeral
card, six court cards and five triumphs - the Pope,
Temperance, the Chariot, the Wheel of Fortune and
the Judgment. The Temperance card has been copied
from the corresponding one in the Visconti-Sforza
pack executed by the later, probably Ferrarese, artist,
so that the cards must date from after the time that
those six cards were painted. On the one numeral
card, the 5 of Swords, the Swords are straight, as in
the Visconti-Sforza pack. On the Wheel card, a point of
divergence from the Visconti-Sforza card is the ladder,
heraldic emblem of the Delia Scala family of Verona,
on the clothing of the topmost figure, who wears ass's
ears, being at the height of his fortunes and about to
experience their collapse.(22)
22 Miss Moakley, in her book cited in footnote 24, draws
attention to the initials ' A. C.' on the base of the throne of
the King of Swords in the Tozzi set. She thinks that these
initials are intended as those of Antonio Cicognara, a
painter to whom many authorities have credited various
surviving fifteenth-century Italian tarocchi. The attribution is
grounded on a purported quotation from Bordigallo's
Chronicle of Cremona given in Count Leopoldo Cicognara's
book referred to in footnote 11, to the effect that in 1484
Antonio Cicognara painted a Tarot pack for Cardinal
Ascanio Sforza. As observed in more detail in Appendix 2,
the quotation is spurious; Count Cicognara was honest but
gullible. Art historians are afflicted by an avid desire to
attach artists' names to works of art, however flimsy the
evidence for it; and so, until more careful study of styles
yielded the attribution to Bembo, sets (1) to (3), and others
as well, were ascribed to Antonio Cicognara, although no
one appears to have attempted to make the elementary
check of verifying that Bordigallo's Chronicle said what it
was supposed to say; even after the attribution to Bembo

(9) The only one of these nine sets that is, almost
certainly, from a regular pack is one consisting of "
fifteen suit cards, not including any Queen, and all
badly damaged by a fire that occurred in 1904, in the
Biblioteca Nazionak in Turin. On the testimony of
W.L. Schreiber,(23) who does not, however, appear very
well informed about the matter, this set comprised
twenty-four cards before the fire. It is helpfully
reproduced in full in Kaplan's book. Unfortunately,
the composition of the set before the catastrophe does
not seem to have been recorded, save on a list kept at
the library, which Mr Kaplan reproduces and which
has itself been partly consumed by the fire. The list
starts with the Coins suit (Cavallo, Jack, Ace, 3),
followed by the suit of Cups (King, Cavallo, Jack, 3, 4,
9, 10), and then the Batons suit, of which only Ace
and 6 are legible. Of the numerals, only Ace and 3 of
Coins, 4 and 9 of Cups, 6 and 10 of Batons and Ace, 3,
6, 7 and 10 of Swords survive. There is also a Cavallo
of Swords, and three court cards whose suit-sign is
unidentifiable, a Cavallo and two Jacks. From the list,
the Cavallo cannot belong to the Batons suit, but
must be of either Cups or Coins; the Jacks likewise
cannot belong to the Batons suit. Evidently there were
no triumph cards before the fire. The Swords are
curvedl-and intersecting; on the odd-numbered cards,
other than the Ace, there is no straight vertical Sword,
but unequal numbers on the two sides. To judge by
the surviving Cavallo of Swords, the general style of
the courts somewhat resembles that of the Rothschild
had been generally accepted, the claim was made that
Cicognara had painted the six cards in the Visconti-Sforza
pack that are not by Bembo. Now Miss Moakley was
convinced that the quotation was spurious, and hence that
there was no reason to suppose that Antonio Cicognara ever
painted any Tarot cards at all. Hence she advanced two
alternative hypotheses: that the initials "A. C." had been
added some time after 1831; or that the entire set was a
modern forgery. The second hypothesis is surely unlikely: a
forger would either have made the cards more unlike the
Visconti-Sforza ones, to reduce the suspicion of forgery, or
have made them exact copies, so as to throw doubt on which
was the original, which the copy. Whether Miss Moakley's
first hypothesis is correct, or whether the initials have some
altogether different significance, I cannot say. The
hypothesis that early playing cards might be forged is not,
as such, implausible: for an example of a forged copy of a
card from the Sola-Busca tarocchi, see D. Hoffmann, Die Welt
, Leipzig, 1972, plate 23(a).
23. Die altesten Spielkarten, Strasbourg, 1937, footnote 10,
p. 102.
24. The Visconti-Sforza pack is the subject of a book by
Miss Gertrude Moakley, The Tarot Cards Painted by Bonifacio
Bembo for the Visconti-Sforza Family
, New York, 1966: all the
cards are illustrated and discussed in detail. It is also the
subject of Tarocchi: il mazzo visconteo di Bergamo e New York,
with text ty Italo Calvino and notes by S. Samek Ludovici,

When and Where the Tarot Pack was Invented 71
As already remarked, of these nine packs, eight
contained both triumph cards and suit cards,
though in one case only one triumph card has
survived and in another only one suit card. Of
any fragmentary set not containing any card
distinctive of the Tarot pack, we can never say for
sure that it was not originally part of such a pack;
but, if the pack to which the Turin cards
belonged had been a Tarot pack, the chance that
all of the fourteen surviving cards should have
been among the fifty-two that could equally well
have come from a regular pack is very low
indeed, so that we can reasonably discount this
possibility. Nevertheless, the remaining eight
packs testify to the great popularity of tarocchi
among the fifteenth-century Italian nobility,
though we should bear in mind that the greater
scope given to an artist by the triumph subjects
Parma, 1969, which also gives illustrations of all the cards.
There is also a reproduction pack issued by the Grafica
Gutenberg, Bergamo; in the United States this is
distributed by U.S. Games Systems, Inc., New York. The
Visconti-Sforza, Visconti di Modrone and Brambilla packs
are illustrated in Emiliano di Parravicino, 'Three packs of
Italian Tarocco cards', Burlington Magazine, vol. Ill, 1903,
pp. 237-52. All three of these packs painted by Bonifacio
Bembo are discussed from an art-historical standpoint in
Pietro Toesca, La pittura e la miniatura nella Lombardia, Milan,
1912 (see pp. 626-7), reprinted Turin, 1966 (see p. 218); in
R. Longhi, 'La restituzione di un trittico d'arte cremonese
circa il 1460', Pinacoteca, vol. I, 1928, pp. 55-87, reprinted in
R. Longhi, Me pinxit, Florence, 1968; Fernanda Wittgens,
'Note ed aggiunte a Bonifacio Bembo', Rivista d'Arte, vol.
XVIII, 1936; and C. Baroni and S. Samek Ludovici, La
pittura lombarda del Quattrocento
, Messina and Florence, 1952
(see pp. 91-116). The Visconti di Modrone pack is discussed
by Robert Steele, 'A notice of the Ludus Triumphorum and
some early Italian card games', Archaeologia, vol. 57, 1900,
pp. 185-200, and by Ron and Charlotte Decker, 'The
Visconti-Sforza cards in the Cary Collection', The Journal of
the Playing-Card Society,
vol. IV, no. 2, November, 1975, pp.
27-32. Eight cards from it are illustrated in Catherine Perry
Hargrave, A History of Playing Cards, Boston and New York,
1930, reprinted New York, 1966, p. 226. The Brambilla
pack was completely illustrated in a booklet called 48 tarocchi
di Bonifacio Bembo, published by the Istituto Finanziario per
l'Arte, Milan, 1971; some of the captions are incorrect.
These and several other of the hand-painted Tarot packs
discussed in the text are discussed, with several illustrations,
in an excellent article by Robert Klein, 'Les tarots
enlumines du XVe siecle', L'Oeil, no. 145, 1967, pp. 11-17,
51-2. For illustrations of the Rothschild cards, see R. Klein,
op. cit., Detlef Hoffmann, Die Welt der Spielkarte, Leipzig,
1972, plates 17(a) and 20(b), and Leopoldo Cicognara,
Memorie spettanti alia Storia delta Calcografia, Prato, 1831, plate
XI. Many works illustrate and discuss the 'Charles VI'
tarots: see R. Klein, op. cit., an anonymous picture-book,
Antiche carte da tarocchi, Rome, 1961, plates III-V; William


[new column]
must have created a strong incentive to a patron,
when ordering an expensive hand-painted set, to
specify a Tarot pack. Equally striking is the
constancy of the subjects used for the triumph
cards; despite the wide variety in their treatment,
we find always the same subjects as those known
from later packs, with the exception of three from
the Visconti di Modrone pack which will be
discussed below, and the possible exception of
the figure on the stag from the Catania pack. Of
the standard twenty-one subjects, the only one
not represented among any of the fifteenthcentury
Italian hand-painted cards surviving to us is the
Devil: but, since this figure appears on the
popular sets of tarocchi, printed by woodblock,
that have come down to us from the end of the
century, this should probably be ascribed to chance.
Andrew Chatto, Facts and Speculations on the Origin and History
of Playing Cards
, London, 1848, p. 187; R. Merlin, L'Origine
des cartes à jouer
, Paris, 1869, p. 89; H.-R. D'Allemagne, Les
Cartes à jouer du XIV au XXe siecle
, vol. I, Paris, 1906, pp. 11,
13, 15, 181-2 and opposite pp. 12, 172, 414, and vol. II,
opposite pp. 4, 18; W.L. Schreiber, Die altesten Spielkarten,
Strasbourg, 1937, p. 101; and Eberhard Pinder, 'The
history of European playing cards', Graphis, vol. 11, 1955,
pp. 246-7. For the d'Este cards, see H.-R. D'Allemagne, op.
cit., vol. II, opposite pp. 12 and 38. Some cards from the
Catania pack, including the figure on the stag, are
illustrated in D. Hoffmann, op.cit., plate 18(a); see also R.
Klein, op. cit., and Antiche carte da tarocchi, plate I, and Guido
Libertini, II Castello Ursino e le raccolte artistiche e comunali di
, Catania, 1937, pp. 112-13. The catalogue numbers
of the cards are 6425-51. One of the Turin cards is shown in
D. Hoffmann, op. cit., plate 18(b); see also R. Klein, op.
cit., Antiche carte da tarocchi, plate I, and W.LI Schreiber, op.
cit., p. 102. The Tozzi cards are all illustrated and discussed
in M. L. D'Otrange, 'Thirteen Tarot cards from the
Visconti-Sforza set', The Connoisseur, vol. CXXXIII, 1954,
pp. 54-60; see also Gertrude Moakley, op. cit., pp. 33-4, fn.
10, and Ronald Decker, 'Two Tarot studies related', part
III, Journal of the Playing-Card Society, vol. IV, no. 1, August,.
1975, pp. 46-52, particularly p. 50. R. Cavendish, The Tarot,
London, 1975, p. 140, illustrates in colour two Charles
VI cards. Kaplan, op. cit., gives illustrations of all
these set-s, as follows: (1) the Visconti di Modrone
pack, seven triumphs, pp. 88-92, and eleven suit cards,
pp. 92-5, with a colour plate of the Knight of Cups,
plate 9; (2) the Brambilla pack, both triumphs, p. 96,
and ten suit cards, pp. 97-8; (3) the Visconti-Sforza
pack, all the cards, pp. 36, 65-86, 285, with a colour plate of
the Bagatto, plate IV; (4) the Charles VI pack, all the cards,
pp. 112-16, with a colour plate of the Love card,, plate 2; (5)
the Rothschild pack, the one triumph, p. 121, and eight
court cards, including that at Bassano, pp. 120-2; (6) the
d'Este pack, all the cards, pp. 117-18; (7) the Catania pack,
two triumphs, p. 109; (8) the Tozzi pack, all the cards, pp.
100-2; and (9) the Turin pack, all the cards, p. 119.

72 Part I: History and Mystery
Besides these nine packs, there are a number of
others of which fewer cards have survived, as

(10) A set of five, consisting of four numeral cards
and one triumph, the Emperor, was acquired in 1974
from a Milanese dealer by the Fournier Playing-Card
Museum at Vitoria in Spain. Like the Tozzi cards, the
designs are based very exactly on the corresponding
cards in the Visconti-Sforza pack; the one notable
departure from the Visconti-Sforza designs is the
depiction of a three-tiered tower on the Coin in the
Ace of that suit, a heraldic emblem of the Gonzaga
family, Marquises of Mantua, according to Mr
Decker. The cards have black backs and measure 171 x
87 mm., as close as makes no difference to the
dimensions of the Tozzi cards (a discrepancy of a
millimetre or two in the measurements of different
cards from the same pack, or of the same card
measured by different people, is ndt significant). If the
backs of the Tozzi cards are also black, there is
therefore a possibility that these five cards belong to the
same pack.

(11) There are four numeral cards, one from each
suit, in the Correr Museum in Venice: the sword on
the Ace of Swords is encircled by a crown and has the
unusual feature of piercing a bleeding heart. The
cards are precisely similar in style to the numeral
cards of the Rothschild set, but, although there is no
overlap between them, they cannot actually be from
the same pack, since the dimensions do not tally (180
x 93 mm. for the Correr cards, 188 x 90 mm. - or,
according to Hoffmann, 186 x 93 mm. - for the
Rothschild ones).

(12) Another set of four cards, bought in Milan
before 1915; is in the Victoria and Albert Museum in
London: it includes two triumph cards - Death and
the Star - and two suit cards - the Jack of Coins and
the Ace of Cups. The Jack of Coins corresponds
almost exactly in design with that in the Visconti-
Sforza pack, and is in a better state of preservation: as
far as I am able to see, judging from this card alone, it
could perfectly well be by Bembo. The other cards,
however, do not in any way resemble the Visconti-
Sforza cards (though it will be recalled that the Star in
the Visconti-Sforza pack as we now have it is not by
Bembo, so that it is conceivable that the Victoria and
Albert Star resembles one by Bembo that is now lost).
Death is shown as a skeleton wielding a scythe and
wearing a cardinal's hat and robe, standing on a
black-and-white chequered floor and with a scroll
coming from his mouth saying 'Son fine'. The Ace of
Cups depicts the Cup as a fountain with a vertical
arrow between the two jets which spring from it; the
stem of the Cup bears the inscription 'nec spe nec
metu', which was the heraldic motto of Isabella
d'Este, and the Cup stands on grass; there are two
putti at its foot, one beaming a shield with the Colleoni

[new column]
arms. (25) The cards measure 167 x 85 mm. I know of no
connection between Isabella d'Este and the Colleoni
family; the cards could plausibly have been painted
for the famous condottiere Bartolomeo Colleoni (1400-
1476), who was closely associated with Francesco
Sforza at certain stages of his career; but Isabella
d'Este, marchioness of Mantua, seems a more likely

(13) Three isolated cards should probably be
grouped together. One is a Popess in the Fournier
Museo de Naipes at Vitoria. This card was bought at
the" Same "time" and from ' t ie "same~ dealer as the five
cards described under (10), but is slightly, though
visibly, smaller, measuring 170 x 85 mm.; it has a red
back, while the other five have black ones. It is a copy
of the Popess in the Visconti-Sforza pack, though not
an exact copy; the Popess's tiara, on this card,
projects further from her head. The second card is a
King of Cups in the collection of Mr N. Biedak of Los
Angeles, very closely resembling the Tozzi King of
Cups, but seen in right profile, like the Visconti-
Sforza one, not in left profile, like that of the Tozzi set;
according to Mrs Wayland, it measures 170 x 86 mm.
The third card is a Jack of Batons in the collection of
Signora C. Marzoli of Milan, measuring 170 x
85 mmj;s;and closely resembling the corresponding'
Visconti-Sforza card. I do not know the colour of the
backs of these last two cards; if it is red, it seems
probable that all three come from the same pack,
possibly one by the artist responsible for the Tozzi set.
Kaplan (op. cit., p. 103) mistakenly groups the Popess
with the other five Fournier cards.

(14) The Guildhall, London, has two pairs of handpainted
fifteenth-century cards, which are of very perceptibly
different widths, and do not come from the same pack.
The wider of these two pairs (138 x 72 mm.) consists of
the Aces of Cups and of Swords.
25. As often in heraldry, the device on these arms
represents a pun on the name of the bearer, though in this
case, one unlikely in more modern times: it consists of three
pairs of testicles (coglioni) which, by a euphemism, later
came to be called, and shown as, inverted hearts. The shield
on the Ace of Cups is parted per fess, not, as in all other
examples of these arms known to me, per pale. Kaplan, op.
cit., p. 99, remarks on the presence of a precipice at the very
bottom of the card on the Ace of Cups; it is also visible on
the Jack of Coins, though not present on the Visconti-Sforza
one. As Kaplan observes (pp. 70, 72), such a precipice is a
feature of four of the six cards not by Bembo in the Visconti-
Sforza pack, Temperance, the Star, the Moon and the Sun.
It is, moreover, to be found on three of the Tozzi cards,
Temperance, the Wheel of Fortune and the Jack of Cups.
Kaplan remarks (pp. 60, 106) that such a precipice is to be
found in the painting at the Carthusian monastery near
Pavia of Christ on the way to Calvary by Ambrogio
Bergognone (active from 1481, died 1523), but draws no
conclusion from the fact. Kaplan gives the inscription on the
Death card incorrectly as Sanfine (p. 104), which he takes to
mean 'Without end'; the first word is Son, meaning 'I am'.

When and Where the Tarot Pack was Invented 73
The former shows a strong affinity with the Victoria
and Albert card: the cup is again a fountain with
vertical arrow between two cascades of water, but
stands on a chequered floor. There is a blank scroll
behind the cup; an odd detail is a small anchor in the
top left-hand corner of the card and a small straight
sword in its top right-hand corner, looking for all the
world like suit-signs, which they obviously cannot be.
The Ace of Swords shows a short sword encircled by a
crowned serpent biting its tail; behind the sword is a
scroll with the words 'Vim vi', and above it a sun with
rays and a face, with the letters MIA above the sun.
'Vim vi' is a motto borne by various Italian families,
but I have not been able to discover one for whom
playing cards are likely to have been painted; the
motto is oddly misread by Kaplan (op. cit., p. I l l ) as

(15) The narrower Guildhall pair (141 x 66 mm.)
comprises one triumph card, the World, which is a
very close copy, laterally reversed, of that in the
Visconti-Sforza pack, and an elaborate card that may
dubiously be identified as a Jack of Batons. This
second card, which Kaplan (ibid.) mistakenly groups
with the wider pair (14), shows a crossbowman
shooting at a heron over water; the archer wears a flat
cap, there are trees behind him, and the heron is
standing by some rushes. Over the right shoulder of
the archer, not attached to anything, is a vertical
cudgel, resembling a Baton of the so-called Spanish
type. It is true that on some early Italian cards,
including the d'Este tarocchi at Yale (6), the Batons
can be rather knobbly, but, with the exception to be
mentioned below, there is nothing else at all like this;
besides, in almost all other cases, the court figure
holds his suit-sign in his hand. Moreover, the whole
design seems rather German in style than Italian. The
Guildhall catalogue records both pairs as having been
found in an old chest in Seville.

(16) A pair of cards at the Muzeum Narodowego in
Warsaw, bought in 1946 from the Potocki collection,
are both court cards, the Cavalier of Coins and the
Queen of Cups; the presence of the Queen shows that
they must have come from a Tarot pack. They show
no especial stylistic resemblance to any other of the
cards here listed.

(17) A very fine pair of Jacks, of Swords and Coins,
is at Hanover (Niedersachsisches Landesmuseum):
the style is quite unlike Bembo's, but the Coin held by
the Jack of that suit bears the Visconti-Sforza serpent.

(18) An isolated card, the Jack of Coins, is in the
collection of Signor Francesco Andreoletti of Milan,
and is a copy, though laterally reversed, of the
corresponding card in the Visconti-Sforza pack; its
measurements (140 x 66 mm,) tally closely with those
of the narrower Guildhall pair.

(19) By far the most puzzling of all is the set of nine
cards known as the Goldschmidt cards, again from

[new column]
the name of a former owner, at the Spielkarten
Museum in Leinfelden. One is a 5 of Batons; the
Batons appear in exactly the 'Spanish' form and
arrangement. Another is an Ace of Cups: as in the
wider Guildhall pair and the Victoria and Albert set,
the cup is a fountain, with two cascades of water and a
vertical arrow between them; as on the Guildhall
card, it stands on a chequered floor. The stem of the
cup is encircled by a serpent biting its tail, like that on
the Ace of Swords in the Guildhall pair, although
uncrowned and facing in the opposite direction. A
third card is surely to be identified as the Ace of
Swords, although Detlef Hoffmann has suggested that
it be equated with the Death card of the Tarot pack. It
shows a short sword, very similar to that on the
Guildhall Ace of Swords, to the blade of which is
chained a skull and the hilt of which has a pair of
crossbones superimposed. A fourth card shows a
crowned dolphin: probably this is just a heraldic
device, and the card, like the Tozzi card showing the
Visconti-Sforza serpent, was not meant to be Used in
play. The remaining five cards are a complete
mystery, (a) One shows a falconer, standing on a
chequered floor, with a little dog at his feet, a bird on
his hand and a hoop suspended from his shoulders;
floating above his shoulder is a toothed wheel,
(b) Another shows a sun, with rays and a face, very
like that on the Guildhall Ace of Swords, above a
chequered floor on which stand three metallic objects
bearing respectively, the letters a, m, c (perhaps
heraldically conventionalised mountains, or perhaps
something quite different), (c) A third shows a
bishop, again standing on a chequered floor; above
his shoulder is an anchor, exactly like that on the
Guildhall Ace of Cups, (d) A fourth shows a lady
wearing a crown, holding a model of a castle and
standing on the usual chequered floor, her gown held
by a lady in waiting; W.L. Schreiber takes her to be
an Empress, (e) The final card has no chequered
floor, and shows a lady wearing a crown and kneeling
at a prie-dieu, with a maidservant in attendance;
Schreiber identifies her as a Dogaressa, with what
right I do not know.

(20) In view of the falconer on one of the
Goldschmidt cards, it is worth mentioning also a
single, very large, card (177 x 95 mm.) showing a
falconer, also at the Spielkarten Museum at
Leinfelden. In 1955 Eberhard Pinder established that
this card was a forgery, though he did not publish this
finding. However, the card is so unlike any other
known to survive that it is probable that the forger was
imitating some original that has since disappeared; he
would hardly have gone to the trouble of producing a
forgery bearing no resemblance to any authentic
26 For colour illustrations of eight of the nine Goldschmidt
cards, see D. Hoffmann, op. cit., plate 19; for discussion
of them, see pp. 18 and 67 of the same work, the article by E.

74 Part I: History and Mystery
One of the striking facts is how frequently the
Visconti-Sforza cards were copied, sometimes
only for certain cards in a pack. It is not
especially surprising that the cards of a famous
pack should have served as a model for later
artists; but it is rather notable that it seems
always to have been the Visconti-Sforza pack
which played this role, and not, for example, the
Brambilla or the Visconti di Modrone one. There
might be suspicions of the authenticity of some of
these cards; but such suspicions could not be
founded on the mere fact that Visconti-Sforza
cards have been copied, since there is surely no
basis for suspecting the genuineness either of the
Victoria and Albert cards (12) of of the narrower
Guildhall pair (15). On the whole, I am disposed
to believe that nos. (1) to (19) are all genuine.

It is obvious that the Goldschmidt cards pose a
severe problem. It is not apparent, from the cards
Pinder in Graphis, vol. 11, 1955, p. 243, the same author's
Charta Lusoria, Biberach an der Riss, 1961, p. 89, W.L.
Schreiber, op. cit., p. 100, and R. Klein, op. cit. For a colour
illustration of the single 'Falconer' card, see E. Pinder's
Graphis article, p. 243. Pinder's later judgment that this card
was a forgery was based on a chemical analysis of the paint by
the Doerner Institut in Munich, backed by the stylistic
judgment of Dr Degenhard, of Munich, and others; I owe this
information to Frau Margot Dietrich, of the Leinfelden
Museum. For the Correr cards, see R. Merlin, op. cit., p. 66
and plates 8 and 9. For the Warsaw cards, see
Stanislaw Sawicky, 'Dwie wtoskie karty "tarocchi" w
zbiorach Muzeum Narodowego w Warszawie', Roczwik
Muzeum Narodowego w Warszawie
, vol. II, 1957,
pp. 605-24. A colour illustration of one of the wider pair of
Guildhall cards (the Ace of Swords) is in Roger Tilley,
Playing Cards, London, 1967, p. 9. For a colour illustration of
the Hanover cards, see J»A.S. Morrison, 'Gamblers' printed
art', The Penrose Annual, vol. 53, London, 1959, p 54. Three of
the Victoria & Albert cards are illustrated, two in colour, in
R. Cavendish, op. cit., pp. 126 and 140. All the cards in sets
(10) to (19), but not the Falconer card (no. 20), are
illustrated in Kaplan, op. cit., as follows: (10), p. 103; (11), p.
123; (12), p. 104; (13), pp. 103, 105; (14), p. 111; (15), pp.
104, 111; (16), p. 109; (17), p. 108; (18), p. 105; and (19), p.
110. Mr Decker cites, as a reference for the Gonzaga tower,
The Complete Paintings of Mantegna, ed. N. Garavaglia,
New York, 1967, p. 104. Of the cards in sets (l) to (20), those
I have not personally seen are the ones in Paris, namely the
Charles VI and Rothschild sets (nos. 4 and 5), those at Turin,
Warsaw and Hanover (nos 9, 16 and 17) and those in private
collections (thirteen of the Visconti-Sforza set in the Colleoni
collection in Bergamo, the Marzoli and Biedak cards in set
no. 13, the Tozzi cards, no. 8, and the Andreoletti card, no.
18). For these I have relied on photographs and on
information, including measurements, very kindly supplied
by Dr and Mrs Harold Wayland, of Pasadena, California-,
who many years ago undertook a comprehensive study of
fifteenth-century Italian hand-painted cards, but regrettably
never published the results of their findings; their help,
without which I should not have known of some of these sets,
has been invaluable to me.

[new column]
themselves, that they are Tarot cards at all: not
one of them can be identified with any assuranceas
one of the Tarot triumphs. Hoffmann equates
the falconer (a) with the Bagatto; but the single
'Falconer' card (no. 20) resembles any ordinary
Bagatto even less, and so makes this
identification doubtful. Hoffmann also equates
card (b) with the Sun of the Tarot pack; but
since the very similar sun on the Guildhall Ace of
Swords clearly does not determine the identity of
the card, it may be that, on this Goldschmidt
card, the sun is again decorative, and that the
identifying symbol is the three mysterious objects
standing on the floor. The bishop might be a
replacement for the Pope: on a sheet taken from a
woodblock, mentioned below, a female bishop
evidently substitutes for the Popess. Schreiber
might be right in saying that the lady with the
model castle is an Empress; but none of these
identifications is compelling, and the lady at the
prie-dieu remains completely enigmatic.

There is, nevertheless, a reason for regarding
the Goldschmidt cards as part of some very
unusual Tarot pack. Their iconographical links
are with the wider Guildhall pair; but there is
some reason to suppose that the narrower
Guildhall pair comes from the same pack, which
must, if so, have been a Tarot pack, since one
member of that pair is the World. The
dimensions of the narrower Guildhall pair (141 x
66 mm.) coincide as nearly as may be with
those of the Goldschmidt cards (140 x 66 mm.).
Where the wider Guildhall cards have
unpatterned gold backgrounds, the narrower
ones have gold backgrounds with patterns very
similar to those on the Goldschmidt cards. The
pattern does not seem to be exactly the same on
any two of the Goldschmidt cards, nor does the
pattern on any one of them tally precisely with
that on either of the narrower Guildhall cards;
but the pattern on one is very similar to that on
Goldschmidt card (b) (the card with the sun),
and that on the other has a clear resemblance to
those on Goldschmidt cards (a) and (e) (the
falconer and the lady at the prie-dieu). Both the
narrow Guildhall cards have black borders, just
as do the Goldschmidt cards. The only
iconographical resemblances are the Spanishstyle
Batons on the Goldschmidt 5 of that suit and
the exactly similar one on the Guildhall card
presumably to be identified as the Jack of that
suit, and the caps worn by the latter figure and the
Goldschmidt falconer. These points do not
together make the assignment of the two sets to

When and Where the Tarot Pack was Invented 75

the same original pack more than a plausible
conjecture: but since, in several of the
Goldschmidt cards and in the Guildhall Jack of
Batons, if that is what it is, we have the only
examples from these fifteenth-century handpainted
cards that present difficulties of
identification, with the sole exception of the
Catania figure on the stag, it is a tempting one.
If the Goldschmidt cards do come from a Tarot
pack, then they testify to the existence in the
fifteenth century of a type of such pack
employing the 'Spanish' form of the Latin suitsigns
and deviating greatly from the norm in the
representation of the triumph subjects, and
probably also in the selection of those subjects,
but yet having links with tarocchi of a more usual
kind, as exemplified by the Victoria and Albert
cards and by the Visconti-Sforza pack. The
implications of this possibility will be discussed
in more detail below.

The Goldschmidt cards, and their relation to
the two Guildhall pairs and to the Victoria and
Albert cards, do indeed pose a difficult problem
which is far from being solved. But if we set the
Goldschmidt cards on one side, and, with them,
the single Falconer card, almost all is plain
sailing: the smaller sets, (10) to (18), simply
confirm the impression derived from the nine
packs of which thirteen or more cards have
survived. There are a few problems about where
certain of the cards were painted or at whose
order: but their identity and the composition of
the packs from which they came are for the most
part unproblematic. Making the suggested
assumptions that the Marzoli and Biedak cards
belong with the Fournier Popess, and the
Bassano card with the Rothschild ones, we have
nine sets of from one to eight cards, of which five
come from Tarot packs and the other four could
be from regular packs. Of the five from Tarot
packs, all have some suit cards, and four have one
or more triumphs. Moreover, when it consists of
only four or fewer cards, the chance that a set
which could have come from a regular pack
actually came from a Tarot pack is significant. If
we take the denomination of a surviving card to
be random, there is of course a 2:1 chance that a
single card from a Tarot pack will be a suit card
other than a Queen. The chance that both of two
cards will be suit cards other than Queens is over
44 per cent, and, even with four cards, the chance
that none of them will be distinctive of the Tarot
pack is nearly 19 per cent. But even if we suppose
that every one of our sets from (1) to (18) that

[new column]
could have formed part of a regular pick did in
fact do so, there are, if the suggested
identifications are accepted, only five such sets
altogether as against thirteen from Tarot packs.
It is plain that the great majority of the playingcard
packs painted by hand for the Italian
nobility of the fifteenth century were tarocchi.

The Goldschmidt cards aside, the four more
fragmentary sets which include triumph cards -
the five Fournier ones (10), the four Victoria and
Albert ones (12), the Fournier-Biedak-Marzoli
trio (13) and the narrower Guildhall pair (15) -
confirm our previous impression that the
triumph subjects, though not their
representation, were standardised from an early
date. This is reinforced by the earliest detailed
reference to the Tarot pack, a sermon against
gaming from an anonymous manuscript volume
of sermons by a Dominican friar. The volume
was formerly in the possession of Robert Steele,
and is now at the Museum of Art in Cincinnati.
The bulk of the sermon was published by Steele
in his article of 1900, (27) in which he dates the
volume to between 1450 and 1470; in his
subsequent article of 1901,(28) he gives the date,
more cautiously, as between 1450 and 1480. In
this sermon, the preacher lists the twenty-one
triumph cards, together with the Matto, as if
they formed an invariable set: the subjects are
precisely the usual ones, though not in exactly
the. order most familiar to us. The same selection
of triumph subjects is confirmed by many literary
references from the sixteenth century.(29) It is
found, likewise, on certain surviving sheets of
cards, printed from wood blocks and made for
the popular market, dating from the end of the
fifteenth century. For our purpose, the sheets
showing regular packs are not of importance: I
shall list only those four which show tarocchi.
(21) Three coloured sheets for one such pack are in
the Metropolitan Museum in New York, and show, in
whole or part, twenty of the twenty-one standard
triumph cards.

(22) A sheet for another such pack, showing all
twenty-one triumphs and three Queens, is in the
Rosenwald Collection in the National Gallery of Art,
Washington, and another copy, much less well
preserved, in the Spielkarten Museum in Leinfelden.
The Rosenwald Collection has two other sheets,
probably though not quite certainly from the same
pack, showing suit cards.
27 See footnote 2.
28 See footnote 15.
29 See Chapter 20.

76 Part I: History and Mystery
(23) Yet another sheet, showing six triumph cards,
is in the Rothschild Collection at the Louvre. A
further sheet of six triumph cards, without doubt from
the same pack, is at the Bibliotheque de l'Ecole
Nationale Superieure des Beaux Arts in Paris.
Together they show the Wheel, the Chariot, the
Hermit, the Hanged Man, Death, the Devil, the
Tower, the Star, the Moon, the Sun, the World and
the Judgment or Angel.

(24) Finally, a sheet showing two numeral cards, a
fragment that is probably the Fool, and, in whole or
part, fifteen triumph cards, is at the Beinecke Library
at Yale, having been part of the Cary Collection.
Among these, there are certainly identifiable the
Bagatto, the Empress, the Emperor, Temperance,
Fortitude, the Wheel, the Chariot, the Devil, the
Tower, the Star, the Moon and the Sun: there are also
fragmentary cards that could be the Pope and Love
cards, and a female Bishop who presumably replaces
the Popess. Several cards resemble the corresponding
ones in the Tarot de Marseille pattern.(30)

A discussion of the probable places of origin of
these various popular Tarot packs will be
postponed until Chapter 20. A detailed analysis
of all the cards in the hand-painted packs (1) to
(19) and on these four sets of sheets will be found
at the end of the present chapter.

There are two late fifteenth-century exceptions
30. A fragmentary card on one of the Metropolitan
Museum sheets is probably the Moon, but might be the
Star; the other of this pair is missing. Their catalogue
numbers are 26.101.5, 26.101.4 and 31.54.159; a
composite photograph of the last two is reproduced by
Kaplan, op. cit., p. 125. The catalogue number of the
Rosenwald sheet is R 19823; the two other sheets with suit
cards are B 19821-2. See Boris Mandrovsky, 'Early Italian
playing-cards in the Rosenwald Collection, the National
Gallery of Art, Washington, D . C , Journal of the Playing-
Card Society
, vol. I, no. 2, November 1972, pp. 1 and 8.
The catalogue number of the Rothschild sheet is 3804.
The cards shown are the Chariot, Death, the Devil, the
Tower, the Star and the Moon. See W.L. Schreiber, op. cit.,
p. 104, where, however, the sheet is. incorrectly stated to
show the Sun instead of the Star. The catalogue number of
the sheet in the Cary Collection is 1-1005. The cards
definitely identifiable are the 7 and 8 of Batons, the Bagatto,
the Emperor, a female bishop presumably representing or
replacing the Popess, Temperance, Fortitude, the Chariot,
the Wheel, the Devil, the Tower, the Star, the Moon and the
Sun. There are also fragments probably to be identified as
the Fool, the Pope and Love. Half of the Rosenwald sheet
with the triumph cards is illustrated in Mandrovsky's article,
and the whole of it by Kaplan, pp. 130-1, in both cases
printed the wrong way round; the Rothschild sheet is
illustrated by Hoffmann, op. cit., plate 14(b), and it and the
Beaux Arts sheet by Kaplan, pp! 128-9. The Cary sheet has
not, so far as I know, previously been reproduced.

[new column]
to the general rule that the triumph subjects are
always the same; these both substitute individual_
classical and Biblical characters for the
generalised figures of the usual Tarot triumphs.
One is the celebrated Sola-Busca tarocchi, a
copper-engraved pack of which several examples
are extant; it was made in Venice by a Ferrarese
artist in 1491, or possibly in 1523.(31) It has the
usual number of cards in each suit, and the suit-
signs are standard; but the numeral cards are
very fancifully executed, the suit-signs not being
displayed in the usual manner, but worked into a
picture containing one or more figures. The court
cards are identified with various historical
characters, whose names are shown on the cards.
There is a Matto, but the twenty-one triumph
cards, which are numbered from I to XXI, again
depict characters of classical and Biblical history,
their names being shown on the cards; there is no
correspondence with the usual subjects.(32) The
other is a pack designed by the poet Matteo
Maria Boiardo (1441-1494). It was to have four
suits, made up of the usual fourteen cards each,
but with the non-standard suit-signs of Whips,
Eyes,""Arrows and Vases; in addition, it was to
have a Fool (Folle) and twenty-one non-standard
triumphs. Again, there was no correspondence
between their subjects, each of which
represented some quality, such as patience,
modesty, etc., and was symbolised by an
appropriate historical character, and the
standard ones.(33) Both these are evidently
31. One card bears the inscription 'Col permesso del
Senato Veneto nell'anno ab urbe condita MLXX' ('With
the permission of the Senate of Venice in the year 1070 after
the foundation of the city'). A traditional date for the
foundation of the city of Venice is 421, yielding the date
1491 for the cards; but W.L. Schreiber, op. cit., p. 105,
remarks that an alternative date is 453, yielding 1523 for the
32. D. Hoffmann, op. cit., p. 68, gives Ferrara as the place
of origin of this pack. For discussion and illustrations, see
Arthur Mayger Hind, Early Italian Engraving, London, vol. I,
1938, pp. 241-7, and vol. IV, 1938, plates 370-93. Kaplan,
pp. 126-7, illustrates twenty triumphs and three court cards.
33. Each card was to bear a descriptive tercet composed by
Boiardo; there were also to be two extra cards, bearing
sonnets by him. The resulting poems, consisting of the two
sonnets and the tercets arranged to make five capitoli, one for
each suit and one for the triumphs, were printed separately
in 1523 in a volume published in Venice and containing
poems by various authors. They were reprinted, under the
title 'I Tarocchi', together with a previously unpublished
commentary by Pier Antonio Viti da Urbino (c. 1470-1500),
by Angelo Solerti in Le Poesi Volgari e Latine di M. M. Boiardo,
Bologna 1894, pp. 313-38, with notes on pp. xxxii-xxxv,

When and Where the Tarot Pack was Invented 77
conscious departures from the norm: they in no
way call in question the existence of a norm. The
standard composition of the Tarot pack was
plainly fixed at a very early stage in its history,
despite occasional experiments such as the Sola-
Busca tarocchi and those of Boiardo. Later, as we
shall see, a number of variant forms developed;
but, in fifteenth-century Italy, the number and
identity of the cards of the Tarot pack was
completely determinate.

The important exception to this is the Visconti
di Modrone pack, which we have yet to describe.
It diverges from the norm in two ways, both in
respect of the suit cards and in respect of the
triumphs. Among the sixty-seven surviving cards
are all forty numeral cards save the 3 of Coins.
However, there are six different denominations of
court card, a male and a female one of each rank:
King and Queen, Knight and Dame (or male and
female Cavalier), and Page (or Jack) and Maid.
Although there is no suit in which all six court
cards survive, they are distributed so randomly
and again in A. Zottoli (ed.), Tutte le opere di Matteo Maria
, Milan, 1936-7, vol. 2, pp. 702-16, with notes pp. 748-9.
The title 'I Tarocchi' is not Boiardo's; neither he nor Viti uses
the word tarocchi, but, instead, trionfi (sometimes for the
twenty-one triumph cards, sometimes for the pack as a
whole). The suits represent four passions: love (Arrows),
jealousy (Eyes), fear (Whips) and hope (Vases). Each court
card depicts an appropriate Biblical or classical character.
The Fool (called by Viti macto) is called il Mondo (the
World), a reversal of the usual practice by which the World
is the highest triumph card; each of the actual triumph
cards represents some quality, such as patience, modesty,
etc., and is symbolised by an appropriate historical
character; there is no correspondence with the usual
triumph subjects. Viti's commentary is addressed to a lady
of the court of Urbino; he expresses the hope that his
patroness will have a pack made in accordance with the
designs he describes. She must have done so, since Carlo
Lozzi, 'Le Antiche Carte da Giuoco', La Bibliofilia, vol. I,
1900, pp. 37-46 and 181-6, mentions just such a pack,
though missing all the court cards and the Fool, and R.
Merlin, L'Origine des cartesà jouer, Paris, 1869, pp. 94-6 a'nd
plate 28, speaks of another copy, missing five court cards,
seven numeral cards, the Fool and all the triumph cards.
(Merlin naturally does not recognise his pack as a Tarot
pack, and Lozzi fails to connect his with Boiardo's poem.)
The pack illustrated by Merlin was very probably identical
with one sold at Christie's in 1971 to Signor Carlo Alberto
Chiesa of Milan; this was a pack printed from wood blocks,
and also missing the Fool and all the triumph cards, as well as
a few court cards and numeral cards. For more illustrations
and further details, see M. Dummett, 'Notes on a fifteenthcentury
pack of cards from Italy', Journal of the Playing-Card
vol. I, no. 2, February 1973, pp. 1-6. The pack is now
in an anonymous Swiss collection.

[new column]
as to make it impossible to suppose otherwise
than that there were originally all six in each
suit: there survive the King, Queen, Dame arid
Maid of Swords, the Queen, Dame, Page and
Maid of Batons, the King, Knight, Page and
Maid of Cups and the King, Queen, Knight,
Dame and Maid of Coins. Of the eleven surviving
triumph cards, eight represent standard subjects
- the Empress, the Emperor, Love, Fortitude, the
Chariot, Death, Judgment and the World. The
other three cards, however, represent the three
theological virtues of Faith, Hope and Charity,
subjects which do not, of course, occur in the
ordinary Tarot pack.

The Visconti di Modrone pack is the only
Tarot pack, of any kind, in which the suits include
court cards other than the usual King, Queen,
Cavalier and either Jack or Maid. There must
have been sixty-four suit cards in all: how many
triumphs there were originally, and whether a
Fool was included, it is impossible to say.
Ronald Decker has suggested that there may
originally have been only fourteen triumphs, and
no Fool, so as to make up the usual total of 78
cards;(34) but the total number of cards in the pack
34. Letter to the Journal of the Playing-Card Society, vol. Ill,
no. 1, August, 1974, pp. 23-4, 48; see also letter by M.
Dummett, ibid., vol. Ill, no. 2, November, 1974, pp. 27-31,
and Ronald Decker, 'Two Tarot studies related', part III,
ibid., vol. IV, no. 1, August, 1975, pp. 46-52 (esp. p. 50). Mr
Decker presumes that the Visconti di Modrone pack had
only 78 cards, like other Tarot packs; since it must have had
64 suit cards, that leaves only 14 triumph cards and no
Fool. There can, on this reasoning, have been no Fool, since
Mr Decker accepts my view that the three, missing Virtues
must originally have been present, and, if we add these to
the eleven surviving triumphs, we already obtain 14, and
there is no roorfl for the Fool. Mr Decker then takes the very
illogical step of arguing that, since there are only 13
(surviving) triumph cards in the Visconti-Sforza pack that
were painted by Bembo, perhaps these, together with the
Fool, were all that the pack originally contained. This is
illogical because in this pack there are only the usual 56 suit
cards, so that he is suggesting an original pack of only 70
cards, whereas the original premiss was that all Tarot packs
had 78 cards. He attempts to rescue his hypothesis by
conjecturing that the Visconti-Sforza pack had originally six
court cards in each suit; but this is obviously very special
pleading. On his hypothesis, there would, besides the suit
cards, have been seven cards in common between the two
packs: the Empress, the Emperor, Love, Justice, the
Chariot, Death and the Judgment. Seven of the triumphs
present in the Visconti di Modrone pack would then have
been removed, namely the World and the six Virtues other
than Justice, when the Visconti-Sforza pack was painted, to
make room for the Fool, the Bagatto, the Popess, the Pope,
the Wheel of Fortune, the Hermit (which originally

78 Part I: History and Mystery
is unlikely to have been seen as a significant
feature. Since four of the stock set of seven
Virtues were included among the triumphs, it
seems probable that the other three were also:
Temperance and Justice, which belong to the
standard list of triumph subjects, and Prudence,
which does not. It is just possible, on the other
hand, that what was held constant was the ratio
between the number of triumphs and the
number of cards in each suit, which, in the 78-
card Tarot pack, is 3:2; if this was also so in the
Visconti di Modrone pack, it would have had
twenty-four triumph cards, in which case it could
have contained all save one of the usual subjects,
making, if the Fool was included, a pack of 89
cards altogether; indeed, if we do not suppose
that it included Prudence, it could have had all of
the usual subjects.

However this may be, the divergence of the
Visconti di Modrone pack from the norm, both
as to the number of suit cards and as to the
subjects, if not the number, of the triumph cards,
strongly suggests that it dates from an early
period when the Tarot pack had not yet assumed
its definitive form. In fact, it is probably the
earliest of all the examples of that pack that have
survived to us. It has usually been thought to
have been made for Filippo Maria Visconti,
which would date it to 1447, the year of his death,
at the latest. All three of the Bembo packs
bear emblems and mottoes of the Visconti family,
but that does not prove that they were made
for Filippo Maria, since Francesco Sforza,
his successor, had in 1441 married his
illegitimate daughter by Agnese del Maino,
Bianca Maria Visconti, and had assumed the
name Visconti-Sforza and, with it, many of the
represented Time) and the Hanged Man. Later, when the
number of triumphs was increased by eight, this was done
by restoring, from the original set of subjects, the World and
two of the Virtues, Temperance and Fortitude, but not the
other four, and adding the Devil, the Tower, the Star, the
Moon and the Sun. All this makes so little sense, and is so
grossly implausible, that the hypothesis that demands it is
not to be entertained. What is impressive about the
fifteenth-century Tarot packs that have come down to us is
not the variation in subjects, but, on the contrary, their
invariance, given the fact that no pack has survived
complete. Certainly we must allow that, after the Visconti di
Modrone pack was made, four of the seven Virtues were
removed; the advantage of the hypothesis that that pack
contained twenty-four triumph cards (not including the
Fool as a triumph) is that it gives a reason for the removal of
at least three of them when the number was reduced to

[new column]
Visconti devices. It is indeed, virtually certain
that the Visconti-Sforza pack was made for_
Francesco Sforza. One reason given by Robert
Steele for taking the Visconti di Modrone pack
to have been made for Filippo Maria is
admittedly flimsy. He thought that the Love
card, which shows a man and woman joining
hands before a tent above which flies a winged
and blindfold Cupid, carried a reference to
Filippo Maria's second marriage. Filippo Maria
divorced his iirst wife,_ Beatrice di Tenda, in
Italian style, having her executed for adultery in
1418; in 1428, he married Maria of Savoy,
although the marriage was probably never
consummated. The tent on the Love card is hung
with shields, alternately showing the Visconti
serpent and a white cross on a red ground, which
Steele took to be the arms of Savoy. But, if the
cards were painted by Bembo, an attribution
questioned by no one, they cannot have been
made as early as 1428, and it is unlikely that
there should have been any allusion to this
unfortunate marriage at any later date; Ronald
and Charlotte Decker identify the shield with the
cross" as the arms of the Principality of Pavia, a
title held by all the-Visconti and Sforza dukes.35
The principal reason for thinking that: the cards
were painted for Filippo Maria is, however, that
the numeral cards of the Coins suit, other than
the Ace and 2, show actual coins, the gold florin
of Filippo Maria, bearing the letters 'FI MA' and
made by the imprint of an actual die; the same is
true of all the eleven surviving cards of the Coins
suit in the Brambilla pack, but not of the
Visconti-Sforza pack. The Deckers surmise,
instead, that they were made by means of 'seals
of the sort used to attach wax imprints to official
documents'; (36) this strikes me as rather unlikely,
in view of the fact that both sides of the coin are
shown: it does not seem probable that there were
two distinct seals, corresponding exactly to the
two sides of the coin. The figures on the court
cards of Swords in the Visconti di Modrone pack
bear a gold fruit on their costumes, which the
Deckers identify as a quince, a Sforza emblem;
but this need not imply that the cards were
painted after Filippo Maria's death, since
35. Ron and Charlotte Decker, 'The Visconti-Sforza cards
in the Cary Collection', Journal of the Playing-Card Society, vol.
IV, no. 2, November 1975, pp. 27-32; see p. 29.
36 Ibid., p. 31. The Deckers wish to prove that the pack
was painted for Francesco Sforza, not for Filippo Maria
[Transcriber's note: In The Visconti-Sforza Tarot Cards Dummett says that the coin-imprints on the cards are too large to be from actual coins. Hence they are from dies. On THF Marco shows that the designs on the cards do not in fact correspond to known coins. I wold add that the "rearing-horse" design remained essentially the same since the time of Gian Galeazzo, only the name of the reigning duke being changed.

When and Where the Tarot Pack was Invented 79
Francesco Sforza was in his service, as well as
being married to his daughter. The probability
seems therefore to be that both the Visconti di
Modrone and the Brambilla packs were painted
for Filippo Maria Visconti, the former being the
earlier of the two and dating from the earliest
stage of existence of the Tarot pack.

The Deckers believe that all three Bembo
packs were painted after the death of Filippo
Maria. Stuart Kaplan, on the other hand, takes
the more usual view that the Visconti di
Modrone and Brambilla packs were both
painted for him, but regards the Brambilla pack
as the earlier (op. cit., p. 107). So far as we can
tell, the composition of the suits in the Brambilla
pack was standard (or what came to be
standard); since only two of the triumphs
survive, we cannot be certain about them. If the
composition of the Brambilla pack was in fact
standard, it seems more likely that it is the later
of the two. Hankering still after an identification
of the Visconti di Modrone pack as a wedding
present, which has only tradition, not evidence,
to speak for it (and not, of course, an ancient
tradition), Kaplan makes the novel suggestion
that it was painted for the wedding of Francesco
Sforza with Bianca Maria Visconti in 1441.
Taken together with his view that the Brambilla
pack is earlier still, this yields a date rather too
soon for such a commission to have been given to
Bembo, whose earliest dateable work is from
1442. As Ronald Decker has observed, the style
of the Visconti di Modrone cards resembles
Bembo's illustrations for a History of Lancelot
dated 1446. If we assume that the Brambilla pack
was the later, we must leave time for Bembo's
receiving from Filippo Maria a second
commission to execute a set of Tarot cards; we
shall therefore probably not be far wrong if we
date the Visconti di Modrone pack to about
1445. We know from the Ferrara account-books
that the Tarot pack (carte da trionfi) was already in
existence by 1442, and was sufficiently familiar to
that court to bear a generic name. On the other
hand, I have argued that the Visconti di
Modrone cards are not likely to have been
painted many years after the first invention of the-
Tarot pack. That event may therefore be
reasonably placed at somewhere around 1440 -
the approximate date, incidentally, assigned to
the painting in the Casa Borromeo.

With the possible exception of the
Goldschmidt cards and of one or both of the

[new column]
two Guildhall pairs, all the early Tarot cards we
possess are Italian; and though, as we shall see, it
cannot be ruled out that the pack was known
elsewhere during the fifteenth century, there is
no conclusive evidence that it was. We can
therefore safely say that it was in Italy,
specifically in northern Italy, that the pack was
invented and first became popular. Furthermore,
it appears initially to have originated and have
been in use in aristocratic circles. The type of
pack of which the few sheets, printed from wood
blocks, listed above are the only remaining
representatives was no doubt, in its time, very
common. As already remarked, cheap mass-
produced playing cards are highly ephemeral,
and survive, when they do, only through some
unusual accident, whereas costly objects made
by an acclaimed artist are preserved: there are in
fact not very many more popular cards, printed
from wood blocks, surviving from fifteenth-
century Italian regular packs than there are
Tarot cards of the same type. We may therefore
safely assume that in the last quarter of the
fifteenth century the Tarot pack attained great
popularity among the lower ranks of society; this
is confirmed by the Steele sermon, the author of
which was probably not preaching to a
congregation drawn only from the nobility, and,
perhaps, by the painting at Issogne.
Nevertheless, the connection with the nobility,
and especially with the courts of Ferrara and
Milan, compels attention. We have seen that at
least two out of three, and probably more, of the
cards hand-painted for the nobility were tarocchi,
a proportion there is no reason to suppose so high
for the popular cards printed from wood blocks.
The three packs by Bonifacio Bembo were all
made for the Milanese court, the Visconti di
Modrone and Brambilla packs probably for
Filippo Maria Visconti and the Visconti-Sforza
one for Francesco Sforza. We have noted that the
Tozzi, Fournier, Biedak and Marzoli cards come
fronxat least two distinct packs, though probably
by the same painter. That painter must have had
access to the Visconti-Sforza cards in order to
make such close copies of them. The card in the
Tozzi set bearing only the Visconti-Sforza
serpent implies that that pack was intended for
the Milanese court. If Ronald Decker is right in
identifying the three-tiered tower on the Fournier
Ace of Coins as a Gonzaga emblem, that
suggests that the five cards of the Fournier set (10)
do not after all come from the same pack as the

80 Part I: History and Mystery
Tozzi cards, and that we therefore have to do
with three distinct copies of the Visconti-Sforza
pack. A possible supposition is that all three were
commissioned from the same artist by Beatrice
d'Este, who married Lodovico il Moro, the last
great Sforza duke, in 1491 and died in childbirth
in 1497: one (the Tozzi set) for her own use, one
(the five Fournier cards) as a present to her sister
Isabella, who married Francesco Gonzaga,
Marquis of Mantua, in 1490, and one (the
Fournier-Biedak-Marzoli trio) for an unknown
recipient. (The Delia Scala emblem on the figure
on the Wheel of Fortune card in the Tozzi set
remains a mystery, since that family had been in
eclipse for a century.) The Victoria and Albert
cards may also come from a pack made for
Isabella d'Este, in view of the inscription of her
motto on the Ace of Cups (though the presence
on that card of the Colleoni shield would then be
mysterious); the artist must surely also have had
access to the Visconti-Sforza cards, in view of the
exact correspondence of the two Jacks of Coins.
The painter of the narrower Guildhall pair and of
the Andreoletti Jack of Coins, whether or not
these are from the same pack, must also have
seen the Visconti-Sforza cards. In view of the
presence of the arms of the King of Naples on two
of the cards, the d'Este pack at Yale was
probably made for Ercole I, the father of Beatrice
and Isabella, who became Duke of Ferrara in
1471 and died in 1505, since he was married to
Eleanora of Aragon, the daughter of Ferdinando
I, King of Naples. We may also with reasonable
confidence assign the Charles VI, Rothschild and
Catania packs to those made for the Ferrara
court. The Ferrara account-books continue to
record orders for Tarot packs, among cards of
other kinds, for example, in 1452, in 1454 and in
1461;(37) and in 1492 Cardinal Ippolito d'Este, the
brother of Beatrice and Isabella, wrote from
Hungary, where he was staying with his aunt,
Beatrice of Aragon, Queen of Hungary, to thank
his mother Eleanora for sending a variety of
things including gilded Tarot cards {triumphi
dorati).(38) We have no way of being sure, but a
plausible guess might be that the Tarot pack
originated in the court of Ferrara, in 1440 or a
few years earlier, and was soon afterwards
adopted by the wealthier court of Milan. In any
case, it seems probable that, for the first two or
37. See the references under footnote 19.
38. See G. Bertoni, op. cit., footnote 19, p. 218.

[new column]
three decades of its existence, it was restricted to
the nobility, and only after that spread out among a
wider social circle.

Although the Tarot pack originated in the
fifteenth century, it did not originally bear that
name. The word 'Tarot' has become more or less
naturalised as an English word; it is in fact the
French adaptation of the Italian name of these
cards — tarocchi or, in the singular, tarocco. In early
sources the French word is sometimes spelled
tarau (plural taraux),tarault or simply taro. In
every other language but French and English, the
hard c sound of the Italian word has been kept -
Tarock in German (formerly often spelled Tarok
or Taroc), tarokk in Hungarian, taroky in Czech, etc.
Where the word tarocchi comes from, nobody
knows: no plausible etymology for it has ever
been suggested, and this deficiency was already
being commented on by an Italian poet, Lollio,
in 1550. (39) It is not, however, the original name of
the cards: the first use of the word tarocchi known
to me dates from 1516, once again from an
account-book of the Ferrara court.(40) Throughout
the fifteenth century, the word used was always
trionfi, or, in Latin, triumphi - 'triumphs': this
name was still in use in 1500.(41) The word trionfi,
39. 'Invettiva contra il Giuoco del Taroco': 'E quel nome
fantastico, e bizarro/Di Tarocco, senz'ethimologia,/Fa
palese a ciascun, che i ghiribizzi/Gli havesser guasto, e
zorpiato il cervello' ('And that whimsical, bizarre name
"Tarocco", without any etymology, makes plain to each that
fantasies have damaged and befuddled his brain '- 'he' being
the inventor of the game).
40. In 1516 the Registro di Guardaroba of the court of Ferrara
repeatedly records the purchase of two, or four, para de
, and similar entries occur in the following year; see
G. Bertoni, op. cit., 1917, pp. 218-19. The word tarocchi also
occurs in Francesco Berni, Capitolo del Giuoco delta Primiera,
Venice, 1526. I know no sixteenth-century use of the word
trionfi to refer to Tarot cards in general, or to the game played
with them, although it continued to be used to refer
specifically to the triumph cards. Nor do I know any
authentic occurrence of the word tarocchi before 1516. For an
almost certainly spurious one, see Appendix 2 to this chapter.
41. The word triumphi occurs in an ordinance from Reggio
nell'Emilia in 1500, forbidding games of chance, including
dice and cards, but specifically excepting 'tables' (i.e.
backgammon), chess and triumphs (hoc tamen statuto non
comprehendentur ludentes ad tabulas et scachos et triumphos cum
); see W.L. Schreiber, op. cit., p. 79, where, however,
the city is mistakenly identified as Reggio di Calabria. Such
exceptions were quite frequent, as at Brescia in 1488, Salo in
1489 and Bergamo in 1491 (see W.L. Schreiber, op. cit., pp.
78-9); in all of these cases the expression used was triumphi
or ludus triumphorum. It thus seems clear that the replacement
of the word trionfi or triumphi by the word tarocchi occurred
some time between 1500 and 1516.

Where and When the Tarot Pack was Invented 81
strictly speaking, refers only to what we have
been_calling the triumph cards, sometimes taken
as including the Fool, sometimes not. By
transference, it was used to apply also to the
game played with the Tarot pack, and sometimes
to the pack itself, including the suit cards; but the
more correct way of referring to the cards of the
Tarot pack, taken together, was as carte da trionfi.
At some time between 1500 and 1516, the new
name, tarocchi, superseded the old one, and was
thereafter invariably used as the way of referring
to these cards in Italian.

An opinion that has gained some support was
first advanced by Robert Steele, namely that the
Tarot pack was formed by uniting the regular
pack with what had previously been an
independent entity, a pack consisting solely of
the Matto and the twenty-one triumphs used on
their own, and that the early references to trionfi
should be taken as alluding, not to the composite
pack known to us as the Tarot pack, but to this
supposed 22-card pack. He based his opinion on
the text of the sermon by the anonymous
Dominican the manuscript of which was at that
time in his possession; indeed, that sermon
formed his only ground for that opinion. The
preacher inveighed, in his sermon, against three
types of game: first dice; then playing cards
{cartulae); and finally triumphs {triumphi). When
he comes to the last of these, he lists the twentyone
triumph cards and the Fool, but makes no
mention of the suit cards. Now, doubtless, if we
knew nothing of the Tarot pack save what we
learn from this sermon, we should have no reason
to think that a set of triumphi consisted of
anything but these twenty-two cards. But the fact
is that there is no other evidence whatever for the
existence of a pack consisting solely of the
triumph cards and the Matto; as we have seen, it
so happens that every fragmentary Tarot pack
that has come down to us includes at least one
suit card. The remarks of the Dominican friar
provide a very flimsy basis for contradicting the
assumption so compellingly suggested by the
actual cards that have survived, namely that the
triumph cards of the Tarot pack from the first
formed only part of a composite or augmented
pack, one containing, in addition to them, the
four suits of the regular pack. The preacher was
not, after all, trying to introduce his congregation
to vices with which they were previously
unacquainted: he was trying to wean them from

[new column]
what he regarded as vices to which they were
already addicted. He therefore did not need
carefully to inform them of the precise
composition of a trionfi pack, something they
already knew very well: he was trying, by
rhetorical devices, to convince them of his view
that all these things - dice, regular playing cards
and triumphs - were instruments of the devil; the
list of triumph cards evidently served as a
memorandum for expatiating on this topic. What
more natural than that, having left the subject of
regular playing cards, he should, when he turned
to denounce triumphs, mention only the cards
peculiar to the trionfi pack? We may agree that it
was primarily to these additional cards that the
name triumphi applied, without in the least
inferring that they ever formed by themselves an
independent pack.(42)

Miss Moakley is inclined to the same view as
Steele, but adds a further complication: she
thinks that there were also packs, consisting
solely of picture cards, but different in number
and subjects from the triumphs of the Tarot
pack, and likewise known as trionfi. On her view,
the term trionfi originally applied to cards of any
pack of a certain generic type, one consisting of
cards depicting mythological figures, personified
abstractions and the like, and only later came to
have specific application to a composite pack
formed by uniting a particular such series to the
regular four-suited pack. That there were, during
the fifteenth century, various packs answering to
this general description, Miss Moakley
undoubtedly establishes. It does not appear,
however, that they were, at any time, of
widespread use; none of them gained a hold on
general taste or remained more than an isolated
curiosity. Nor can it be shown that they were in
existence at an earlier date than the composite
Tarot pack. What is most to the point, however,
is that there is no reason to think that the word
42 Stuart Kaplan, op. cit., pp. 26, 349, offers a piece of
spurious evidence for the Steele thesis, stating that St
Anthony, Bishop of Florence, in a Treatise of Theology written
in 1457 'refers to playing cards and tarot, thus suggesting
that the trumps or trionfi were considered a separate game
from playing cards, which comprised court cards and
numeral or pip cards'. He presumably intends to refer to the
Summa Theologica of St Antoninus, Archbishop of Florence;
Pars 2 (Nuremberg, 1447), chap. 23, of this work does indeed
contain several mentions of playing cards, under the
alternative names of cartae or naibi, and their suit-signs (e.g.
§ viii, 'Unde in cartis sive naibis sunt figure non solum
baculorum, denariorum, cupparum, sed et gladiorum'). No
mention of triumphi is, however, to be found.

82 Part I: History and Mystery
trionfi was ever used for any kind of playing cards
other than Tarot cards. If Miss Moakley were
right, the references to carte da trionfi in the
account-books of the Ferrara court, from 1442
onwards, might relate, not to Tarot packs, but to
others of this more general type; but such a
generalised use of the term cannot be

The most interesting of the special packs
which Miss Moakley claims as examples of trionfi
in the alleged more general sense, and one to
which she draws particular attention, is a set of
sixteen picture cards commissioned by Filippo
Maria Visconti from the painter Michelino da
Besozzo (fl. 1394-1442) - a painter, incidentally,
to whom the murals of games players in the Casa
Borromeo have been attributed by some. This set
was sent in 1449 by a Venetian, Jacopo Antonio
Marcello, as a present to Queen Isabella, wife of
King Rene I, Duke of Lorraine. It was divided
into four groups of four, representing Virtue,
Virginity, Riches and Pleasure; each card
depicted a suitable classical divinity. The pack
has not survived, but the letter to Queen Isabella,
written in Latin, describing the pack and saying
that Michelino painted it, has.(43) The letter
applies the word ludus (game) to the set, showing
that it was really meant to be used to play some
kind of game; but there is no use of the word
triumphi in reference to the cards.
[Transcriber's note: here Dummett is mistaken. See ]

A celebrated but problematic passage in the
life of Filippo Maria Visconti, written in Latin by
Pier Candido Dezembrio (1399-1477), runs as
follows: 'He was accustomed from his youth to
play games, of various kinds ... and particularly
that type of game in which images are painted,
which delighted him to such an extent that he
paid 1500 gold pieces for a whole pack (ludum) of
them, made in the first place by Marziano da
Tortona, his secretary, who executed with the
utmost diligence images of gods, and placed
under them with wonderful skill figures of
animals and birds.' (44) There are many oddities
about this passage. In the first place, as W.L.
43. See Chapter 3, footnote 2.
44. The passage runs: Variis autem ludendi modis ab
adolescentia usus est ... plerunque eo ludi genere, qui ex imaginibus
depictis fit, in quo precipue oblectatus est adeo, ut integrum eorum
mille, et quingentis aureis emerit, auctore pel in primis Martiano
Terdonensi ejus Sectretario, qui Deorum imagines, subjectasque his
animalium figuras, et avium miro ingenio, summaque industria
Dezembrio's life is reprinted in L.A. Muratori,
Rerum italicarum scriptores, vol. XX, Milan, 1731, and the
passage will be found in col. 1013.

[new column]
Schreiber remarked,(45) playing cards were
perfectly well known when Dezembrio .. was_
writing, and it is quite obscure why he should
choose to describe them as for readers who had
never heard of them before. In the second place,
even for someone as rich as Filippo Maria
Visconti, the price for a single pack seems
staggeringly high. In the third place, as remarked
by Campori,(46) Marziano is not known to have
been a painter, and a funeral oration for him
makes no mention of his having been one.(47)
However, if the information given by Dezembrio
is at all correct, the pack described was
presumably not a Tarot pack, which does not
normally contain images of gods or pictures of
animals and birds. Hence this was probably a
pack of the kind Miss Moakley is concerned
with; but there is not in Dezembrio's text any use
of the word triumphi. The word does, indeed,
occur in what Campori cites as a contemporary
translation into Italian of Dezembrio's life of
Visconti, written by someone using the
pseudonym Polismagna, the manuscript of
which is said by Campori to be preserved in the
d'Este library; but it may quite well be that the
translator, like others' after him, was puzzled by
the passage, and assumed that it must refer to
some kind of Tarot pack.(48)

Another documentary source cited by Miss
Moakley is an inventory of the workshop of the
engraver Francesco Rosselli made in 1528.(49) This
inventory lists plates for printing a number of
remarkable games: the giuocho del trionfo del
; the giuco d'apostoli chol nostro singnore; the
giuoco di sete virtu; and the gioucho di pianeti cho loro
(the game of the triumph of Petrarch; the
game of Apostles with our Lord; the game of
45. W.L. Schreiber, op. cit., p. 100.
46. G. Campori, 'Le Carte da Giuoco dipinte per gli
Estensi nel Secolo XV, Atti e Memorie delle RR. Deputazioni di
Storia Patria per le provincie modenesi e parmensi
, vol. 3, Modena,
1874, p. 125, fn. 4.
47. The funeral oration is printed in Tiraboschi, Storia della
letteratura italiana
, vol. 6, p. 1196.
48. See G. Campori, op. cit., p. 125, fn. 3. The translation
runs: Alcuna volta zugava a le carte de triumphi. Et di questo giocho
molto si delectoe per modo che comparoe uno paro di carte da triumphi
compite mille et cinque cento ducati. Di questo maximamente auctore et
casone Martinno da Terdona suo secretario, il quale cum meraviglioso
inzegno et somma industria compite questo giocho de carte cum le
figure et imagine de li dei et cum le figure de li animali et de li ocelli
che gli sum sottoposti.

49. See. A.M. Hind, Early Italian Engraving, part I, vol. I,
London, 1938, pp. 10, 11, 305-8. The spellings are given as
in Hind.

When and Where the Tarot Pack was Invented 83
seven virtues; and the game of planets with their
borders). These must, again, have been games
with picture cards of special kinds; but they are
not labelled, generically, trionfi. The name of the
first game relates to the poem by Petrarch called
I Trionfi, and cannot, therefore, fairly be cited in
support of a general thesis.

None of the special packs so far mentioned has
survived: the only one of this kind that has come
down to us from this period is the celebrated
copper-engraved set, which exists in two
versions, known as the tarocchi di Mantegna, about
which it is invariably, and correctly, observed
that they are neither tarocchi nor by Mantegna.
They are thought to date from about 1465, and
were made by an unknown artist of the Ferrarese
school. Many have doubted that this set was used
for a game at all, on the ground that existing prints
are on paper too flimsy to be used for play; but it is
quite likely that it was originally intended for a
game of some kind. The set consists of fifty
picture cards, divided into five groups of ten
each, representing respectively social ranks,
Muses, sciences, virtues and the celestial spheres:
the cards are individually numbered, and each,
group is distinguished by a letter. Once again,
there is no evidence that they were ever referred
to as trionfi, although at a later date the term
tarocchi was attached to them by a vague

The fact is that games of this kind represent a
persistent, and natural, inclination to invent new
games to be played with packs of playing cards
having a structure entirely different from that of
the regular playing-card pack or from an
augmented form such as the Tarot pack, an
inclination already manifest in the fifteenth
century and freely indulged in by games
50. Miss Moakley also suggests that some engravings
ascribed to Nicoletto da Modena, illustrated in A.M. Hind,
op. cit., vol. VI, 1948, plates 640-7, form part of a pack of
cards; but this cannot be so, since they differ considerably in
size. The literature on the tarocchi di Mantegna is vast: for
illustrations, see A.M. Hind, op. cit., vol. IV, 1938, plates
320-69; for a survey of the literature, see D. Hoffmann, op.
cit., p. 67. For arguments in favour of regarding them as
playing cards, see Fritz Saxl, 'Verzeichniss astrologischen
und mythologischen illustrierten Handscrhiften des
Lateinischen Mittelalters in JR.6mischen Bibliotheken',
Sitzungsberichte der Heidelberger Akademieder Wissenschaften, pp.
101, 222, and Heinrich Brockhaus, 'Ein Edler Geduldspiel
"Die Lietung der Welt oder die Himmelsleiter" und die
sogenannte Taroks di Mantegna vom Jahre 1459-60', in
Miscellanea di Storia dell Arte in onore Igino Benvenuto Supino,
1933, p. 397.

[new column]
manufacturers to-day. There was a particular
vogue for such games in Italy, which lasted
through the seventeenth century, as two such
packs designed by Mitelli bear witness. In most
cases, however, the games invented to play with
packs of this kind have no strong attraction to
outweigh the disadvantage of having to buy a
special pack of cards to play them; often they
merely imitate the features of traditional games
played with the regular pack. As a result, the
packs designed for use in such games prove
ephemeral and leave no progeny. The only
exception to this is the Cuccu pack, an Italian
invention of the seventeenth century which not
only exists to this day but spread to other parts of
Europe, where it gave rise to the Hexenkarte of
Germany, no longer extant, and the Gnav pack of
Denmark and Norway and the Killekort of
Sweden, both still well known in those countries.
This was, and is, used to play a simple and
enjoyable game which, in a simplified form,
adapted to the regular pack, is known to British
children under such names as Ranter Go Round
and Catch the Ace. But, of course, this has
nothing to do with the Tarot pack, and no-one
ever thought of calling these cards trionfi or tarocchi.

There is thus no reason to suppose that the
Tarot triumphs ever formed a separate pack by
themselves; and there is still less reason to think
that they were ever regarded as just one species of
a large genus known, as a whole, as trionfi. It is
evident that the Tarot pack became immensely
popular within a short time after its invention;
but the only reasonable hypothesis is that it was
from the start a composite pack, containing the
four suits of the regular pack alongside the
additional cards to which the name trionfi
properly applies, and that, in connection with
playing cards, the word trionfi, as used in the
fifteenth century, applied only to the Tarot
triumphs or, by extension, to Tarot cards as a

There can be no doubt that it was in Italy that
the Tarot pack was invented, and there that,
throughout the fifteenth century, it was chiefly
popular; but the question when it first became
known in any other country does not admit 6f so
ready an answer. It was certainly in France that
it first became known outside its country of
origin; but it is difficult to be precise at what date
it was first known there. The earliest certain
reference to it there comes from Rabelais in 1534;
he includes it, under the spelling tarau, in his long

84 Part I: History and Mystery
list of the games played by Gargantua; tarots are
again referred to in the posthumous Fifth Book of
1564.*51) The earliest surviving Tarot pack known
to have been made outside Italy is one made by
Catelin Geoffroy in Lyons in 1557.(52) But we have
seen that the term tarocchi did not come into use
in Italy until after 1500, and we should therefore
assume the same to be true of the term tarots in
France: if there were any reference to the Tarot
pack from fifteenth-century France, we should
expect it to be by means of some such word as
triumphes. And indeed we find, once more from an
account book, that in 1496 Rene II, Duke of
Lorraine, is reported as having played at
triumphe;(51) the earliest recorded use of the word in
French as the name of a card game dates from as
early as 1482.(54) Unfortunately, we cannot be
certain that these references are to games played
with the Tarot pack. In Italy, after the adoption
of the new term tarocchi, or perhaps
simultaneously with it, the term trionfi was
transferred to a game played with the regular
pack; this new use of the word trionfi goes back at
least to 1526.(55) In France also there was a very
ancient game, played with the regular pack, and
known as Triumphe, which is also mentioned by
Rabelais. If we conceive of the Tarot pack as not
having been introduced into France until after
the adoption of the name tarocchi, that is, at the
beginning of the sixteenth century, so that the
cards were never known there otherwise than as
tarots, then we could interpret these late fifteenth-
century references to triumphe or triomphe as
alluding only to the game known from Rabelais's
time to the present day under that name. But this
supposition, although possible, is unlikely. It
implies that the use of the name 'Triumphe' tor a
card game in France is unconnected with the
51. F. Rabelais, Gargantua and Pantagruel, book I, ch. 22 and
bookV.ch. 23.
52. Seventeen cards from this pack are illustrated in Detlef
Hoffmann, op. cit., plates 15(b) and 36(a), nine of them in
colour. Nine are illustrated by Kaplan, op. cit., p. 132. The
pack is in the Museum fur Kunsthandwerk, Frankfurt am
Main, catalogue number K 1.
53. See H.-R. D'Allemagne, op. cit., vol. II, p. 212. The
references occur in the account-books of the court of
Lorraine for the year 1495-6, and run respectively:
Au Roy, le 29 avril pour jouer au triumphe a Vezelise
deux francs.
Encore audit seigneur roy le 1e mai pour jouer audit
triumphe a Vezelise deux florins d'or.
54. See F. Godefroy, Dictionnaire de I'ancienne langue francaise,
Complement, s.v. 'triomphe'. See also Chapter 9, fn. 2.
55. In Francesco Berni, Capitolo del Giuoco della Primiera.

[new column]
contemporary use in Italy of the word trionfi for
Tarot cards and the games played with them,
that it was mere coincidence that two such
similar names were used for different things. On
this theory, the transference of the term trionfi to a
game played with the regular pack might have
occurred in imitation of the name of the French
game also so played, necessitating the
introduction of a new word for Tarot cards. This
is possible; but it is not probable. For reasons
that will-not-be-set out in full until Chapter
7, it is much more likely that no coincidence
was involved: that the name trionfi was
transferred from the Tarot cards to a game
played with the regular pack precisely because
that game was in part adapted from that which
the Tarot pack was used to play, and that the
game played in France under the name
Triumphe, like other games with similar names
in other countries, originated from, the
dissemination of the same idea. If this is so, then
the game known to this day as Triumphe cannot
have come into existence until after the term
trionfi had ceased to be used specifically for Tarot
cards; and the fifteenth-century uses of the word
triumphe or triomphe must be taken as referring to a
game played with the Tarot pack, whose
introduction into France must therefore be dated
to at least about 1480. It fits well with this
hypothesis that the later reference concerns the
court of Lorraine, to which we^ have noted a
pack of playing cards made for the Milanese
court being sent as a present some forty-odd
years earlier.

Of the various hand-painted Tarot cards of the
fifteenth century, the only ones of which we
could not be certain that they came from Italy
were the Goldschmidt cards and the two
Guildhall pairs, though they had connections
both with the Victoria and Albert cards and the
Visconti-Sforza pack. Opinions about the
provenance of the Goldschmidt cards have been
very various. W.L. Schreiber assigned them to
Venice, on the strength of his identification of the
kneeling lady as a Dogaressa. Eberhard Pinder
thought they were made in the Upper Rhine
region by an Italian artist. Now the Victoria and
Albert cards are surely Italian, if only because of
the Italian inscription on the Death card; and it
is plain that the painters of the Goldschmidt
cards and of the wider Guildhall pair were
familiar with the convention used in the Victoria
and Albert pack for the representation of the Ace

When and Where the Tarot Pack was Invented 85
of Cups. Furthermore, if the narrower Guildhall
pair really is part of the same original pack as the
Goldschmidt cards, the artist must have known
the Visconti-Sforza pack, including the later
cards not by Bembo. There is therefore good
reason for thinking that an Italian artist, or at
least one acquainted with Italian cards, was
responsible for this pack. Nevertheless, Detlef
Hoffmann is surely right in fastening upon the
appearance of the Batons in their so-called
Spanish form as the most significant clue. Batons
on Italian cards of the fifteenth and sixteenth
centuries may vary somewhat in shape: but they
invariably intersect, and, like the Swords (which
do not always intersect), they invariably extend
the whole length of the card. They are never
found disposed, as in Spanish-suited packs and
as on the 5 of Batons in the Goldschmidt set,
upright and in the manner of the pips on a
French-suited card, in separate rows. It seems
unthinkable that this pack can have been made
for use in fifteenth-century Italy.

From the fact that the two Guildhall pairs
were discovered in a chest in Seville one might be
tempted to believe that the Goldschmidt cards
represent an otherwise unknown phenomenon -
Spanish Tarot cards. But this would surely be a
mistake. As has already been remarked, that
variant of the Latin suit-system which was in the
course of the sixteenth century adopted as the
national suit-system of Spain was not in origin
Spanish, but French. What little we know of late
fifteenth- and early sixteenth-century Spanish
cards suggests that at that time cards made in
Spain employed the Latin suit-system in
something very much more like what was to
become its Portuguese variant, with straight but
intersecting Swords and knobbly but intersecting
Batons, though doubtless French-made
'Spanish'-suited cards were imported in
considerable numbers. It is possible, therefore,
that the Goldschmidt cards represent a type of
Tarot pack used in some noble house of fifteenth-
century France, though there is no need to locate
them more narrowly in Provence, as Hoffmann
does. They differ too much from anything else
that has come down to us, however, for this to be
more than a conjecture. Only a single Tarot pack
survives to us from sixteenth-century France,
that by Catelin Geoffroy already referred to; and
this is no guide to the way the suit-signs
appeared on early French Tarot packs, since it
uses completely non-standard suit-signs. The

[new column]
subjects on the triumph cards are, however,
standard, and show no relationship with
the enigmatic figures on the Goldschmidt
cards. The next earliest French Tarot pack
we have is one made in the early seventeenth
century; and on this the Swords and Batons are
neither of the usual Italian shape, nor of the
Spanish one; they do, however, for the most
part intersect with one another. In all
later French Latin-suited Tarot packs, the
Italian suit-system is used. It thus appears that, if
the Goldschmidt cards really were made for use
in France, they have left no progeny and may
have been an isolated experiment; but the data
are too sparse to ground a firm opinion. It could,
indeed, be argued from the fact that the Tarot
pack was later associated so firmly with the
Italian version of the Latin suit-system that it
cannot have been introduced into France until a
time when that suit-system was no longer very
familiar, or, at least, no longer seemed quite
ordinary, on the ground that otherwise the suitsigns
would have undergone the same
modification to their 'Spanish' form that was
imposed on regular Latin-suited packs in France.
Such an argument would rest upon the
assumption that we have made that originally the
'Italian' suit-system was everywhere in use. But,
even if this assumption and the foregoing
argument are correct, this does not threaten our
conjecture that the triumphe played by Duke Rene
II and the triumphe mentioned in 1482 were games
played with the Tarot pack, or even that the
Goldschmidt cards represent the type of pack
that may have been used. The 'Spanish' variation
on the Latin suit-system was in existence by
about 1460, but it may have been invented
earlier; and the Goldschmidt cards might
represent an early phase when the Tarot pack
was known only in a few aristocratic circles.

In all Tarot packs made outside Italy, the
triumph cards bear Roman or Arabic numerals
to indicate their position in the sequence; and,
in all non-Italian Latin-suited Tarot packs after
1700, except in the Revolutionary period, and in
some seventeenth-century ones, they also
have their names inscribed in full at the bottom
of the card (save for the Death card, whose name
is usually missing). The same practice was
usually observed for the court cards as well, and
often for the Aces. Italian Tarot cards made
before the eighteenth century do not carry verbal
inscriptions (save for a few non-standard packs,

86 Part I: History and Mystery
and some occasional mottoes); and even the
practice of putting numerals on the triumph cards
seems to have come in only gradually. On the
sheet in the Rosenwald Collection in Washington,
the numbering stops at XII, the top nine cards
being left unnumbered; on the sheets at the
Metropolitan Museum, New York, the triumph
cards are numbered from I to XX, only the top
card, the World, being left unnumbered; but, on
the sheet in the Cary Collection, and on those in
the Rothschild Collection and at the Ecole des
Beaux Arts, the triumph cards bear no numerals.
There is an incomplete pack in the Bibliotheque
Municipale at Rouen from the early sixteenth
century, whose triumph cards bear numerals;
although a classicised pack, the figures can
easily be equated with the usual subjects, unlike
in the Sola-Busca tarocchi.(56) Count Leopoldo
Cicognara knew a complete example of a very
similar, though not identical, pack, and
illustrated six cards from it in his book of 1831; (57)
in his pack, there were no numerals on the
triumph cards. The triumphs of the Sola-Busca
pack itself do bear numerals. Numerals do not
seem to have been an original feature of any
of the hand-painted packs: there are numerals
on three of the triumph cards at the Castello
56. The pack is part of the Leber Collection, catalogue
number 1351-XIV. Four cards are illustrated in colour in D.
Hoffmann, op. cit., plate 23(b), and nine by Kaplan, op.
cit., p. 133. Thirty cards survive, including the Fool and
seven triumph cards. The latter are to be identified with the
usual subjects as follows: Imperator Assiriorum,
unnumbered (the numeral is presumably covered up by the
turned-over edge) - the Emperor; Pontifex Pontificum, 5 -
the Pope; Victoriae Premium. 7 - the Chariot; Omnium
Dominatrix, 10 - the Wheel of Fortune;. Rerum Edax
(Saturn), 11 - the Hermit (or Time); Perditorum Raptor
(Pluto), 14-the Devil; Inclitum Sydus, 16-the Star.
57 See L. Cicognara, op. cit., pp. 163-6 and plate XIV; the
cards are also shown in D. Hoffmann, op. cit., fig. 6. The
cards illustrated by Cicognara are the Aces of the four suits,
Cupid = Love, and Apollo = the Sun. Contrary to what is
said by D. Hoffmann, op. cit., p. 68, the pack described by
Cicognara was not the same as that at Rouen, though very
similar. The Rouen set includes the Aces of Batons, Coins
and Swords, and these differ considerably in design from
those .shown by Cicognara. Also, Cicognara describes the
Fool of his pack in detail, and it is quite different from that
at Rouen: Cicognara's Fool was a drunkard lying bn his
back, supporting, with his legs in the air, a jar marked
'Muscatello'; that at Rouen shows a man armed to the
teeth, and dressed in armour, but with genitals exposed and
urinating, and bears the inscription 'Velim fundam dari
mihi'. The Cicognara pack is ascribed by A.M. Hind, op.
cit., vol. V, London,' 1948, pp. 139-40, to Nicoletto da

[new column]
Ursino in Catania (not on the unidentified
one showing the figure on a stag), but these
have obviously been added much later; There
are also numerals on the triumph cards of the
'Charles VI' set, which are also later additions,
although, in their case, they may have
been added in the fifteenth century. Otherwise
the hand-painted triumph cards are all
unnumbered. It should not be thought, however,
that the lack of numerals in these packs is
evidence that the triumph cards did not
originally form an ordered sequence. The sermon
quoted by Steele lists them in a definite sequence,
even giving their numbers, a sequence that is
confirmed by some literary sources of the
sixteenth century. It is not that the cards did not
have an order, but just that those who used them
were expected to remember this order without
recourse to enumeration, just as they would
know the order of the court cards of any suit
without any further aid. It might seem that to
keep in mind the order of twenty-one distinct
cards is too difficult a feat for people to have been
expected to perform; but this supposition can, as
it happens, be decisively refuted. The particular
form of Tarot pack still used in Bologna, which
has changed comparatively little since the
sixteenth century, save for becoming double-
headed (it was one of the earliest standard
patterns to do so), did not, until the mid-
eighteenth century, bear numerals on any of the
triumph cards at all. Yet, in the game played
with this pack (which has also changed very
little, at least since the eighteenth century, and,
probably, since long before that), the triumph
cards have a definite ranking. Eighteenth-
century descriptions of this game list the triumph
cards by name, and never refer to them by
number, and were probably written for players
using packs without numerals on the cards. In
any case, there is a clear demonstration that the
same ranking applied before it was the practice
to put numerals on the triumph cards. Before
that time, a celebrated geographical Bolognese
Tarot pack was designed by Canon Luigi
Montieri in 1725: the main body of each triumph
card (including the Fool) carried geographical
information, and that of each suit card showed
coats of arms. (There was a great vogue in the
seventeenth and eighteenth centuries for such
instructional packs, particularly geographical
and heraldic ones; both regular packs and Tarot
packs were designed for this purpose.) In

When and Where the Tarot Pack was Invented 87
Montieri's pack, the usual symbol giving the
denomination-of each card was confined to a
small panel at the top. In each such panel on the
triumph cards is a single letter; when the
triumphs are arranged in descending order, with
the Fool at the end, these letters spell C LUIGI
MONTIERI INVENTOR, a clear indication
that, at that time, the triumphs ranked in the
same order as that which they have had from the
mid-eighteenth century until now.(58) Another
famous Bolognese Tarot pack was made (in 1664,
according to C. P. Hargrave) for the Bentivogli
family by the engraver Gioseppe Maria Mitelli
(1634-1718); the engravings were also issued in
book form, with ten cards to a page, and the
triumph cards are again arranged, in descending
order, in the usual sequence.(59) It is thus apparent
that, long before the Bolognese triumphs bore
numerals, they were arranged in a determinate
order, and that, from the early sixteenth until the
mid-eighteenth century, players were expected to
remember this order. What Bolognese players
could do up to the eighteenth century, others
could do in the fifteenth. There is therefore no
obstacle to supposing that the triumph cards
formed, from the outset, a sequence with a definite

Why, then, were these cards called 'triumphs'?
Many have tried to explain the word from the use
of the twenty-one triumph cards in play, namely
as 'triumphing' over the other cards; and we
cannot say for sure that this explanation is
incorrect. A brilliant suggestion of Miss
Moakley's is, however, more attractive. This is
that the name has nothing to do with the use of
the cards, but only with what is shown on them,
the series of triumph cards representing a sort of
triumphal procession. As documented by
58. The Montieri cards are illustrated in Playing Cards of
Various Ages and Countries Selected from the Collection of Lady
Charlotte Schreiber
, vol. Ill, London, 1895, plates 74-9, with
notes pp. 13-15. There is also a reproduction pack issued by
the Edizioni del Solleone in Lissone in 1973, edited by
Signor Vito Arienti and illustrated by Kaplan, op. cit., p.
59. Two of the Mitelli cards are illustrated in C. P.
Hargrave, op. cit., opp. p. 232; see also opp. p. 99 in 1966
edition; twenty-four are shown in Kaplan, op. cit., p. 54.
The book version was issued as Giuoco di Carte con nuova forma
di Tarocchini; Intaglio in Rome di Gioseppe Maria Mitelli
, and
was reprinted in 1970 by Huber und Herpel of Offenbach
am Main as Gioseppe Maria Mitelli, B]ologneser Tarockspiel
des 17. Jahrhunderts
. C.P. Hargrave, A History of Playing Cards,
New York, 1930, 1966, p. 232, confidently cites the date
1664, but gives no authority for doing so.

[new column]
Burckhardt and Miss Moakley, a favourite
entertainment in the courts of Renaissance Italy
was the staging of just such triumphal processions,
with floats bearing figures either derived
from classical mythology or representing
abstractions such as Love, Death, etc.: a
transformation of the utterly serious triumph of a
Roman general or Emperor into an elegant allegorical
entertainment. A frequent ingredient
in such Renaissance triumphs was the idea
underlying Petrarch's poem I Trionfi, in
which each successive personified abstraction
triumphs over, that is, vanquishes, the last;
thus, in the poem, love triumphs over gods
and men, chastity over love, death over
chastity, fame over death, time over fame
and eternity over time. The case would be
clinched if it were possible to explain the subjects
of the triumph cards of the Tarot pack as forming
a triumphal procession of this sort; but in
spite of Miss Moakley's determined efforts,
supplemented subsequently by those of Mr
Ronald Decker, such an explanation, while
plausible in principle, is difficult to make
convincing in detail. Nevertheless, in default of a
better explanation, we may accept it as likely,
though by no means certain, that it was this
association of ideas which prompted the use of
the name 'triumphs' for the additional cards of
the Tarot pack.

Appendix 1:
A Problematic Set of Tarocchi

After I had finished this book, and was engaged on
final revision, I received a copy, kindly sent me by the
author, of Stuart R. Kaplan's The Encyclopedia of Tarot
(New York, 1978), already referred to. The most
valuable feature of the book is the extensive series of
illustrations of all the sets of fifteenth-century Italian
hand-painted c'ards, and of many other Tarot packs
surviving from before the eighteenth century. I have
inserted references to Kaplan's illustrations of the
packs discussed in this chapter in the footnotes. I have
some disagreements with Kaplan's judgments; to
some of these I have drawn attention in the text or the
footnotes of this chapter. There is, however, one set of
hand-painted Tarot cards illustrated by Kaplan of
which I was quite unaware, discussion of which I
thought it best to relegate to this appendix.

The set in question comprises twenty-three cards;
Kaplan states (p. 106) that the last known owner of
the set, before the Second World War, was a British
dealer named Rosenthal, and says (p. 99) that in 1939

Part I: History and Mystery
it was offered to a leading American collector, who
refused it because he doubted its authenticity. Kaplan
supplies illustrations of all the cards (p. 99);
unfortunately, these are rather minute, so that it is
difficult to see details even with a magnifying glass. Of
the twenty-three cards, eleven closely resemble the
corresponding Visconti-Sforza ones: the Emperor,
Justice, the Cavalier, Jack, 5 and 4 of Swords, the
Queen and Jack of Batons, the King of Cups, and the
King and Jack of Coins. Four resemble the Visconti-
Sforza cards in general style, but differ in detail: the 5
of Batons, the 5 of Cups, and the 5 and 3 of Coins.
The former two differ in the arrangement of the suitsigns,
the latter two in the disposition of the scrolls
inscribed a bon droit, which is the form of the Visconti
motto consistently used in this set (the spelling is
always droyt on the Bembo cards, as on the Tozzi 5 of
Swords, though it is droit on the Fournier 2 of Coins).
Two cards, the Star and the Ace of Cups, are very
similar to the Victoria and Albert ones. The Star is
almost precisely the same, but the Ace of Cups shows
some differences: the Colleoni arms are not parted,
there is an inscription I cannot read on the upper
scroll, the cliff noted by Kaplan is missing, and,
though the stem of the 'cup' or fountain is still
inscribed nec spe nec metu, the inscription occupies two
lines instead of four. Another card in the Rosenthal
set shows only the Visconti/Sforza serpent, exactly
like the Tozzi card. The remaining five cards are: (i) a
Falconer card, very closely resembling no. (20), save
for the design on the cape; (ii) a card showing a sun
with rays and a face, as on the Goldschmidt and wider
Guildhall cards, over a castle, with a wheel and a fleur
de lys above the castle on either side, and, at the
bottom, a scroll inscribed Fortezza (a word which may
mean either 'fortress' or 'fortitude'); (iii) an Ace of
Swords, showing a dagger dripping blood, and, at the
bottom, part of a sun with an inscription I cannot
read, and two scrolls higher up on the card, marked a
bon droit and, apparently, REPUB; (iv) a Cavalier of
Batons, like the Visconti-Sforza one but laterally
reversed, and with a three-turreted castle in the top
left-hand corner, encircled by an inscription I cannot
read, and an unidentifiable object in the top righthand
corner; and (v) an Ace of Coins, showing a
cardinal in the Coin, and, according to Kaplan, an
inscription, not visible in the illustration.

It is very hard to draw conclusions about this
extraordinary set from Kaplan's diminutive
illustrations, taken from a photograph in his
possession; they deserve publication in colour and in
full size (though Mr Kaplan does not know their
measurements). The salient reason for supposing the
suspicions of the American collector who refused to
buy them to be justified is the figure of the cardinal on
the Ace of Coins; it looks very much like an attempt to
establish the set as really being, at last, from the pack
supposedly painted for Ascanio Sforza. However, the

When and Where the Tarot Pack was Invented 89
style of this Ace of Coins seems totally unlike that of
the rest of the set; it is possible, therefore, that it is a
forged addition to an otherwise genuine set. A more
subtle reason for doubt lies in the form of the Colleoni
arms on the Ace of Cups; here the device takes the
later form of three inverted hearts, not of three pairs of
coglioni (testicles), as on the Victoria and Albert card
and other contemporary presentations of these arms,
for instance in the Colleoni chapel at Bergamo (see
footnote 25). This strongly suggests that the
Rosenthal cards could not have been painted in the
fifteenth century.

If the set should nevertheless prove to be genuine
(perhaps with the exception of the Ace of Coins), it
would establish the most interesting links between
other surviving sets of fifteenth-century Tarot cards.
First, it would supply an original for the Falconer card
(no. 20), and from a Tarot pack, though whether it
represented the Bagatto, or even the Fool, or some
distinct alternative triumph subject, would remain
obscure; this would increase the probability that the
Goldschmidt cards are also genuinely from a Tarot
pack. Secondly, it would establish the sun with a face
as a device employed on various Milanese Tarot
cards, and thus would make it less likely that the
Goldschmidt cards or the wider Guildhall pair had a
non-Italian origin; the significance of this sun would
remain problematic. Thirdly, it would provide further
examples of the practice of placing small emblems in
the upper corners of cards, a practice that would still
be baffling.

Whether genuine or forged, the set poses some new
puzzles of its own. What is the significance of the
inscription REPUB on the Ace of Swords? On the
death of Filippo Maria Visconti in 1447, the citizens
of Milan, tired of their Dukes, as well they might be,
declared a republic; in 1450, however, Francesco
Sforza captured the city and proclaimed himself
Duke. Can this card be meant to contain an allusion
to the bloody suppression of the short-lived republic?
The card inscribed Fortezza cannot, in view of the
inscription, represent the Sun; if the card is genuine,
this provides corroboration of the view that the sun on
the Guildhall card we took to be the Ace of Swords
and on the unidentified Goldschmidt card is not, in
either case, the feature of the card determining its
identity. The Fortezza card, if spurious, may be meant
to represent the Tower; but, if genuine, it can hardly
do so, because that card, although it went under
various names and had many representations, is never
called la Fortezza, or even la Torre, in early Italian
sources. It is much more likely to represent Fortitude,
by a kind of visual pun, even though this subject is
normally represented by a personification; la Fortezza
is the name invariably given to this subject in the early
sources, as against the name la Forza (Strength)
usually employed in the later Tarot de Marseille-
derived packs. It is hard to avoid being impressed by

[new column]
this card. Unlike most of those in the set, it is not a
close copy of some other existing card. If we suppose it
a forgery, then to suppose it intended to represent the
Tower is to attribute a very crude mistake to the
forger. If we regard it as representing Fortitude, on
the other hand, it becomes, an ingenious and
unexpected representation of its subject, and
presupposes enough knowledge of the literary sources
on the part of any forger responsible for it for him to
be aware that the regular word used for the Fortitude
card in the fifteenth and sixteenth centuries was
Fortezza. These considerations seem to me to weigh in
favour of the authenticity of the set; and, if the
Fortezza card is genuine, the lady with the model
castle in the Goldschmidt set may be another deviant
version of Fortitude.

It thus becomes a matter of some importance for
the study of fifteenth-century hand-painted tarocchi to
determine whether any or all of the Rosenthal cards
are genuine, and, as a first step, where they are. Mr
Albi Rosenthal, of Oxford and London, who is
presumably the British dealer referred to by Kaplan,
has informed me that in the 1920s his father sold some
hand-painted Italian tarocchi to Herr von Hardt of
Switzerland, but does not know where the von Hardt
collection is now. He has also told me that at a later
date some fifteenth-century hand-painted tarocchi were
shown to him at his Curzon Street office in London,
but that these were definitely found to be forgeries.
Which of these two sets, if either, is that designated
'the Rosenthal cards' by Kaplan is unclear. It is to be
hoped that the cards themselves, or at least some
more detailed illustrations of them, become available
for examination; in the meantime, we owe a
considerable debt to Mr Kaplan for bringing the set to
public attention.

Appendix 2:
The Tarocchi of Cardinal Ascanio Sforza

In Count Leopoldo Cicognara, Memorie spettanti alia
Storia delta Caliografia
, Prato, 1831, p. 16, there is
quoted an alleged excerpt from the Chronicle of
Cremona by Domenico Bordigallo. The excerpt is in
Italian, and states that in the year 1484 the excellent
painter Antonio Cicognara (of the same family as the
Count) painted uno magnifico mazzo de carte dette de'
Tarocchi, da me veduto
(a magnificent pack of the cards
called tarocchi, seen by me) for Cardinal Ascanio
Sforza. Count Cicognara says that this passage was
communicated to him by Mgr Antonio Dragoni from
the schede (notes) of Giacomo Torresino, an
eighteenth-century Cremonese antiquarian. On the
strength of this passage, numerous art historians have
ascribed various hand-painted Tarot cards of the
fifteenth century to Antonio Cicognara, including the

[new column]
six cards of the Visconti-Sforza pack that are not by
Bembo; but the passage is almost certainly spurious.
Any historical document connected with Dragoni is
under the gravest suspicion, since he was either a
forger or the dupe of forgers, although he was
primarily concerned with documents relating to the
Dark Ages, of which he made, or manufactured, a
large collection. Torresino did indeed compose notes
on local history, using a page for each year, and
entering quotations relating to that year from various
sources; but this work, at any rate in the form in
which it survives in the Biblioteca Statale at Cremona,
stops before 1484. Bordigallo's Chronicle was written
in Latin, and has never been published; the
manuscript was located by Signor Marco
Santambrogio, of the University of Bologna, in the
Biblioteca Treccani in Milan, where, with the kind
assistance of Signora Carla Treccani degli Alfieri, he
examined it; he found that, while the entry for 1484
does contain a reference to Ascanio Sforza, namely to
record that it was in that year that he was created a
Cardinal, it mentions neither Antonio Cicognara nor
tarocchi. It is conceivable that the quotation was in
some later section of Torresino's notes that has since
been lost, but from some other source, or that it is in
Bordigallo's Chronicle, but under a later year
(Ascanio Sforza died in 1505); but the probability is
that it is quite inauthentic. In any case, the use of
the modern word mazzo for 'pack' was, so far as I am
aware, unknown in fifteenth-century Italian, which
uses paro or gioco instead; so, even if the Italian given
by Count Cicognara is a translation of some genuine
Latin original, the word tarocchi is not likely to have
occurred in that original. See M. Dummett, 'A Note
on Cicognara', Journal of the Playing-Card Society, vol. II,
no. 1, August 1973, pp. 14-17 (original issue), pp. 23-
32 (reissue), and 'More about Cicognara', ibid., vol.
V, no. 2, pp. 26-34. These two articles are cited by
Stuart R. Kaplan in his annotated bibliography (The
Encyclopedia of Tarot,
New York, 1978, p. 356), but he
mentions only their discussion of the Fibbia portrait,
not of the Bordigallo Chronicle. Mr Kaplan does,
however, state categorically (pp. 33, 351) that the
Chronicle contains no reference either to tarocchi or to
Antonio Cicognara. Though I consider this quite
probable, I cannot vouch for it, since Signor
Santambrogio had time to examine only the section
dealing with the year 1484, and I have not yet seen the
manuscript myself. From the absence of
acknowledgment to myself or to Signor
Santambrogio, the reader might naturally suppose
that Mr Kaplan was speaking on his own authority
and had examined the Chronicle in more detail than
Santambrogio had done; but this seems unlikely in
view of his mistaken assertion (p. 33) that it was in
1484 that Bordigallo wrote his Chronicle, since
anyone who had seen it would have observed that the
entries go beyond that year. From his curious

90 Part I: History and Mystery
statement (p. 374) that Torresino's notes contain
Bordigalio's Chronicle, it is equally unlikely that he
has seen them. Mr Kaplan expresses the belief (pp.
100, 107) that the initials ' A. C.' on the Tozzi King of
Swords may stand, not for 'Antonio Cicognara', but
for 'Ascanio Cardinale'; this seems somewhat
illogical, since Count Cicognara's purported
quotation from Bordigallo is the only positive
evidence either that Antonio Cicognara painted any
tarocchi or that any were painted for Ascanio Sforza,
and Mr Kaplan agrees that the quotation is spurious.
His assertion of its inauthenticity occurs during a list"
(pp. 31-3) of spurious sources, and is repeated in the
bibliography; in the section discussing the authorship

[new column]
and dates of the hand-painted packs (pp. 106-7), he
cites the purported Bordigallo quotation in full
without, indeed, endorsing it, but without repudiating
it either; only the most alert reader is likely to
remember the earlier declaration of disbelief in it. Of
course, it is perfectly plausible that Ascanio Sforza
should have had some tarocchi made for him. But, ever
since 1831, the names of Antonio Cicognara and of
Ascanio Sforza have been endlessly cited, in books,
articles and museum catalogues, in connection with
tarocchi, and it is in my view best to make no further
reference to those two individuals until some genuine
evidence of such a connection becomes available.

[There follows two tables summarizing the surviving cards in the decks 1-12, 14, 19, 22, and 24 of this chapter.]

Chapter 20, Game of Tarot

The Order of the Tarot Trumps

Those who, in the eighteenth century, when the
Italian suit-signs of the traditional Tarot pack
were replaced by French ones, also substituted
animal figures for the traditional trump subjects,
and those who later replaced these with rural
scenes, views of buildings or characters from the
drama, obviously did not think they were
depriving the pack of any essential feature. What
was essential to a trump card was its position in
the sequence, indicated by its number; for the
rest, antiquated mediaeval figures were being
replaced by subjects more to modern taste. Many
people, however, have been fascinated by thea
figures on the trump cards of the Latin-suited
Tarot pack, and have sought to uncover a hidden
symbolism lost to us. They have been convinced
that these figures must have a deeper meaning
than appears on the surface; and, in particular,
they have believed that there is a significance, not
only in the individual cards, but in the precise
order in which they are arranged. Foremost
among these have, of course, been the occultists;
but, as we saw in Chapter 6, their interpretations
have been completely arbitrary, or based on false
premisses such as the ancient Egyptian or
Hebrew origin of the cards. In so far as the
occultist interpretation has rested on anything
more than whim or demonstrably spurious
history, it has been based on the details of the
trumps in the Tarot de Marseille pattern. If we
are seeking the symbolic intentions of those who
first designed the Tarot pack, the Tarot de
Marseille is a dubious guide. We cannot feel sure
that the pattern is, as a whole, any older than the
seventeenth century; and, although the order of
the trumps which it observes goes back at least to
Catelin Geoffroy's pack of 1557, we do not know
it to hiave any exact Italian prototype.
Speculations based on false data are obviously

Not all of those who have sought to decode the
symbolism of the Tarot pack have been
occultists; some have been serious scholars, well
versed in the iconography of late mediaeval and
early Renaissance art. One W.M. Seabury wrote .
a book to prove that the symbolism of the pack
was based upon Dante; (1) Miss Gertrude
Moakley, in her fine book about the Visconti-
Sforza pack, advanced an interpretation of the
pack, supported by much evidence from Italian
art and literature; Mr Ronald Decker has
engaged in complicated speculations, linking the
pack to the astrology of the time. I am not going
to advance another such theory. I do not even
want to take a stand about the theories that have
been advanced. The question is whether a theory
is needed at all. I do not mean to deny that some
of the subjects, or some of the details of their
conventional representation; may have had a
symbolic significance obvious to fifteenth-century
Italians, or, at least, to educated ones, that
escapes us and may be revealed by patient
research; that is very likely to be the case. But the
question is whether the sequence as a sequence
has any special symbolic meaning. I am inclined
to think that it did not: to think, that is, that
those, who originally designed the Tarot pack
were doing the equivalent, for their day, of those
who later selected a sequence of animal pictures
to adorn the trump cards of the new French-
1 William Marston Seabury, The Tarot Cards and Dante's
Divine Comedy
, New York, 1951.

388 Part III: Italian Games and Italian Cards
suited pack. They wanted to design a new kind of
pack with an additional set of twenty-one picture
cards that would play a special, indeed a quite
new, role in the game; so they selected for those
cards a number of subjects, most of them entirely
familiar, that would naturally come to the mind
of someone at a fifteenth-century Italian court. It
is rather a random selection: we might have
expected all seven principal virtues, rather than
just the three we find - and, of course, we do find
all seven in the Minchiate pack, and they were
probably present also in the Visconti di Modrone
pack. With the Sun and Moon we might have
expected the other five planets, instead of just a
star; with the Pope and the Emperor, we might
have expected other ranks and degrees. But, of
course, in a pack of cards what is essential is that
each card may be instantly identified; so one
does not want a large number of rather similar
figures, especially before it occurred to anyone to
put numerals on the trump cards for ease of
identification. Certainly most of the subjects on
the Tarot trumps are completely standard ones
in mediaeval and Renaissance art; there seems
no need of any special hypothesis to explain
them. Whatever may be the truth about those
who first designed the Tarot pack, the inventors
of the Minchiate pack surely approached their
task in the spirit I have suggested: they wanted
twenty additional subjects, and they chose ones
which it was natural for men of the sixteenth
century to think of - the four elements, the
remaining virtues, the signs of the Zodiac - and
inserted them en bloc in a convenient place. I do not
think that anyone has suggested that there is any
hidden significance in the sequence of Minchiate

That is my opinion; but I do not want to insist
on it. It may be that those who first devised the
Tarot pack had a special purpose in mind in
selecting those particular subjects and in
arranging them in the order that they did:
perhaps they then spelled out, to those capable of
reading them, some satirical or symbolic
message. If so, it is apparent that, at least by the
sixteenth century, the capacity to read this
message had been lost. There are many
references to tarocchi in sixteenth-century
Italian literature, in which their symbolic
potentialities were exploited, but always in an
obvious way: no hint survives that any more
arcane meaning was associated with them.
'What else', asks Flavio Alberto Lollio in his
Invettiva contra il Giuoco del Taroco, 'do they signify;
the Popess, the Chariot, the Traitor the Wheel,
the Hunchback, Fortitude, the Star, the Moon,
Death, Hell and all the rest of this revolving
bizarrerie, save that this man [the inventor of the
game] had an empty head, full of smoke, caprices
and idle tales?' Lollio, of course, had no interest
in making much sense of the Tarot trumps; but
he could hardly have written in quite this vein if
there was generally acknowledged to be some
particular interpretation to be placed on them.
The search for a hidden meaning may be a
unicorn hunt; but, if there is a hidden meaning to
be found, only a correct basis of fact will lead us
to it.

The hidden meaning, if any, lies in the
sequential arrangement of the trump cards; and
therefore, if it is to be uncovered, we must know
what, originally, that arrangement was. In
all Italian-suited packs made outside Italy since
1700, the order is always and everywhere the
same, namely that found in the Tarot de Marseille.
There is, as we have seen, some variation
in the subjects: the Pope and Popess are replaced,
in the Tarot de Besançon, by Jupiter and Juno,
and, in the Belgian Tarot, as in de Hautot's pack,
by Bacchus and the Spanish Captain. But, from
the beginning of the eighteenth century, the
order of the trump subjects and their numbering
remain constant, in the Belgian Tarot as well as
m the Tarot de Marseille and its various
offshoots. In the seventeenth century, we have
Viévil's pack with a significantly different order:
we also have, in de Hautot's pack and the list
given in the Maison academique, the minor
variation in which the Empress was higher than
the Emperor. But we also find the exact Tarot de
Marseille order in the anonymous seventeenth century
Parisian pack; and, as remarked, we also
find it in Geoffroy's pack of 1557, so that it is
plainly of considerable antiquity in France. The
use of just the same order in all other countries of
Europe, other than Italy, simply reflects the fact
that it was to France and other French-speaking
regions, including those in Switzerland, that the
game of Tarot first spread from Italy, and from
which it spread further to yet other lands. Since
the eighteenth century, the Tarot de Marseille
order has been well known in Italy itself, being
observed in the Lombard pattern and in the
Tarocco Piemontese which remains in wide use
to this day, both patterns being descendants of
the^Tarot de Marseille. It will, however, already
h'ave been apparent to the reader, from what we

The Order of the Tarot Trumps 389
have seen of the Tarocco Bolognese, the Tarocco
Siciliano and the Minchiate packs, that the Tarot
de Marseille order was not and is not the only
order for the trump cards known in Italy. In fact,
there is no clear evidence that the Tarot de
Marseille order was ever known in Italy before
the appearance of the Lombard and Piedmontese
variants of the Tarot de Marseille in the
eighteenth century. There is, however, no one
trump order which we could set over against that
of the Tarot de Marseille as being the
predominant Italian one; rather, the evidence
yields a number of distinct orders used in
different places or by different players.

There are three types of source that we have
for the different orderings of the trumps observed
by Italian card players. First, there are the three
variant types of pack, the Tarocco Bolognese, the
Tarocco Siciliano and the Minchiate pack. The
Minchiate pack has, of course, twenty additional
trumps: but since these were inserted en bloc at a
certain point in the sequence of standard trump
subjects, we can remove them and study the
resulting order in reasonable confidence that it
represents an order observed for the trumps of
the 78-card pack at the time the Minchiate pack
was invented. Secondly, there are the early packs
that survive to us. Not all of these are any help,
since we noted that the trumps in the fifteenth-
century hand-painted packs lack numerals, and
we therefore cannot tell how they were ordered;
but we have a few packs, mostly popular ones
printed from wood blocks, in which the trumps
bear numerals. Finally, there are literary sources.
The earliest of these is the sermon against
gaming in the anonymous volume of sermons
once owned by Robert Steele; it is dated between
1450 and 1480, probably towards the end of that
period, and gives a list of all the trumps with
their numbers. A well-known list, without
numbers, is given by Tomaso Garzoni (1549-
1589) in his La Piazza Universale (Venice, 1585).
Garzoni uses a phrase which has been
understood to mean that he has taken the list
from an earlier writer, Raffaele Maffei
Volterrano (1455-1522), but no such list is
known to occur in Volterrano's writings;
Garzoni's remark may not be so intended. (2)
2 Tomaso Garzoni, La Piazza Universale di Tutte le
Profession: del Mondo, e nobili et ignobili,
Venice, 1585. The list
occurs in the chapter 'de' Giocatori in universale, et in
particolare', which is Discourse 69 of the later editions, and
is to be found on p. 574 of the Venice, 1586 edition.

remaining two sources known to me are
examples of a curious form of verse fashionable in
Garzoni's turn of phrase is curious: he says, 'Alcuni altri son
giuochi da tauerne, come la mora, le piastrelle, le chiaui, e le
carti, 6 communi, 6 Tarocchi, di nuoua inuentione, secondo
il Volteranno: oue si vedono danari, coppe, spade, bastoni,
dieci, noue, ...', and continues by listing the remaining
twelve cards of each suit, followed by the trumps in
descending order and finally the Matto; after a mention of
the French suits (as used 'con le carte fine'), he lists a
number of card games, beginning with Tarocchi and
Primiera. (In English, the quoted passage runs, 'Some
others are tavern games, such as mora, quoits, keys and
cards, either ordinary ones, or tarocchi, recently invented
according to Volterrano: in which are to be seen Coins,
Cups, Swords, Batons, the 10, the 9, ...'. Mora is a well-
known game in which each of two players simultaneously
holds up a hand, with five, two or no fingers extended: I do
not know what 'keys' are.) For some reason, Garzoni's
reference to tarocchi is much the best known of the sixteenth-
century ones, and is cited by a whole string of later writers,
including Senftleben (Andreas Senftlebius, De alvea veterum
opusculum posthumum
, Leipzig, 1667), who mistranslates
Fortezza as propugnaculum (fortress), and the notes to Saverio
Bettinelli's Il giuoco delle carte, poemetto (Cremona, 1775);
many of these attribute the list to Volterrano. Garzoni's
phrase 'secondo il Volteranno' appears, however, to relate,
not to the list of trumps, but to the apparent.observation
that tarocchi are a recent invention. Even on this
interpretation, the remark is baffling, since Tarot cards are
nowhere referred to in the Commentariorum Urbanorum
of Raffaele Maffei, called Volterrano after his
place of birth, which were first published in Rome in 1506,
nor, so far as anyone has discovered, in any other of his
writings, as was observed by Robert Steele in 1900;
moreover, the Tarot pack had existed for a hundred and
fifty years when Garzoni was writing, and for at least fifteen
when Maffei was born. The explanation appears to be that
Garzoni meant that playing cards in general were a recent
invention, and that he was alluding to the remark by Maffei
that 'Chartarum vero & sortium & divinationis ludi priscis
additi sunt ab avaris ac perditis inventi' ('To the ancient
games have been added those of cards and of lots and of
divination, invented by covetous and dissolute men'). This
remark occurs in the section 'De ludo diverso quo summi
viri quandoque occupati fuerunt' of book XXIX of the
Commentaria Urbana (p. 421 verso of the Rome, 1506, edition,
p. 313 verso of the Paris, 1511, edition, and p. 694 of the
Basle, 1559, edition; the second ampersand, present in the
1506 and 1511 editions, is missing from that of 1559). Maffei
is meaning to convey by this observation no more than that
the games he is referring to were not played in classical
times. Garzoni was not, therefore, quoting him in support of
any thesis that tarocchi were of recent invention, only as saying
that playing cards are of modern, as opposed to ancient,
origin. (I am uncertain to what Maffei was referring as
sortium & divinationis ludi, but I do not think the passage can
be treated as evidence that cards were used for fortune-
telling; we have in all three modern types of game, cards, lots
and divination, and the mention of avari suggests that Maffei
has gambling games principally in mind.)

390 Part III: Italian Games and Italian Cards

the sixteenth century and known as tarocchi
We have already come across one
example of this form, namely the Germini sopra
Quaranta Meritrice
, in which the Minchiate
trumps are used. In a poem of this kind, a set of
people is described by associating each of them
with one of the trump subjects from the Tarot
pack. In some cases, the trumps are not arranged
in any particular order, and these poems are of
no help to us; but, in two of them, the trumps are
arranged in sequence. One of these is an anonymous
poem first published by Giulio Bertoni
in an essay on 'Tarocchi versificati' in
1917; (3) it describes the ladies of the court of
Ferrara, and is dated by Bertoni to between 1520
and 1550, more probably nearer the later date. I
shall refer to this as 'the Bertoni poem'. The
other is a poem that has been attributed to
Giambattista Susio (1519-1583), an attribution
that has been contested. The poem has never
been published in full, but excerpts are given by
Ridolfo Renier in his essay 'Tarocchi di M.M.
Boiardo' of 1894; (4) it concerns the ladies of the
court of Pavia, and may have been written about

Before we look in detail at these various orders,
a word needs to be said about the names of the
cards. Tarot de Marseille trumps, and in fact all
those in Italian-suited packs made anywhere
outside Italy, bear their names as well as their
numbers, with some slight variation on these
names from one pack to another. By contrast, in
Italian Tarot packs, other than those which
derive from the Tarot de Marseille, the trumps
never have their names inscribed on them, save'
for isolated cards like the Miseria of the Tarocco
Siciliano and a few non-standard packs such as
that at Rouen. We have, therefore, to appeal to
literary sources for the names. Here, of course,
other sources, those which name the trumps
without arranging them in order, become useful.
One such is a set of five sonnets on the Tarot
trumps by Teofilo Folengo (1491-1544), a
Mantuan author of macaronic verses; these were
included in his Caos del Triperuno, a work
published in Venice in 1527 under his
pseudonym Merlin Cocai. (5) Another is Pietro
3 In Poesie, leggende, costumanze del medio evo, Modena, 1917,
pp. 215-29; see pp. 220-1.
4 In Studi su Matteo Maria Boiardo, ed. by N. Campanini,
Bologna, 1894; see pp. 256-9.
5 Le opere maccheroniche di Merlin Cocai, ed. by Attilio
Portioli, vol. Ill, Mantua, 1890, pp. 128-33.

Aretino's Le Carte Parlanti, published in Venice in
1543; and a third is Lollio’s poem, already
mentioned, published in Venice in 1550. There
are also two other tarocchi appropriati, both
concerning cardinals at a conclave: one relates to
the conclave of 1522, which elected Adrian VI
and the other to that of 1549-50, which elected
Julius III. (6) We may add to these the lists of the
trumps of the Tarocco Bolognese given in the
earliest eighteenth-century accounts of the game
ot Tarocchinp and written before numerals had
been added to any of the trump cards; Bolognese
players were evidently as conservative in their
nomenclature as in their manner of play. (7)

Most of the cards have the same name, save for
trivial variations of spelling, in all the sources.
Among these, the ones with names coinciding
with those used in modern Tarocco Piemontese
packs, and also corresponding to those of the
Tarot de Marseille (given here in the right-hand
column), are as follows.
The Fool
The Emperor
The Popess
The Devil
The Moon
The Sun
The World

il Matto
la Papessa
la Temperanza
la Giustizia
il Diavolo
la Luna
il Sole
il Mondo

le Mat
la Papesse
la Justice
le Diable
la Lune
le Soleil
le Monde

There are five more whose names always appear
in the same form in the early sources:
6 The first is no. XXXII of the Pasquinate di Pietro Aretino ed
Anonime per il Conclave a I'Elezione di Adriano VI
, ed. by
Vittorio Rossi, Palermo, 1891, also to be found in Mario
dell'Arco, Pasquino e le Pasquinate, Milan, 1957, pp. 87-8. The
second, an imitation of the first, was published by V. Cian
in his 'Gioviana', Giornale storico delta Letteratura Italiana, vol.
XVII, 1890, pp. 338-40.
7 Stuart R. Kaplan, in his The Encyclopedia of Tarot, New
York, 1978, cites yet another source; on p. 30 he states that
Antonio Francesco Grazzini wrote in Tutti i trionfi, carri,
mascherate o canti carnascialeschi andati per Firenze dal tempo del
magnifico Lorenzo de' Medicifino all'anno 1559
about the tarocchi
trumps. On p. 359, he repeats this claim, attributing the
whole book to Grazzini. Poems by Grazzini (il Lasca) are
indeed included in this collection, which he in fact edited, but
I can find no reference in them or in any other poems in the
volume to tarocchi. Perhaps Mr Kaplan was misled by the
occurrence of the term trionfi in the title: it does not there refer
to Tarot trumps or triumphs, but is used in the sense of
'triumphal processions; see the section on Festivals in J.
Burckhardt's Civilisation of the Renaissance.

Fortitude always appears in them as la Fortezza;
it is la Force (Strength) in the Tarot de Marseille,
and likewise la Forza (Strength) in the Tarocco

Love always appears in the early sources as
I'Amore; in the Tarot de Marseille it is I'Amoureux
(the Lover) and in the Tarocco Piemontese gli
(the Lovers).

Death is always called la Morte, as in the
Tarocco Piemontese; in the Tarot de Marseille, it
is almost always left unnamed.

The Star is always la Stella in the early sources;
in the Tarot de Marseille it is, likewise, l'Etoile
(the Star), but in the Tarocco Piemontese le Stelle
(the Stars).

The Angel is always l'Angelo in the early
sources, save in the Minchiate pack, in which it is
le Trombe (the Trumpets). In the Tarot de
Marseille it is le Jugement (the Judgment), though
a pack made in Strasbourg has la Trompette (the
Trumpet), and it is similarly referred to as la
in Viévil's pack. In the Tarocco
Piemontese it is sometimes labelled il Giudizio
(the Judgment), but usually as I'Angelo.

This leaves six cards whose names vary in the
early sources. For three of these, the differences
are trifling.

The Bagatto, called le Bateleur in the Tarot de
Marseille and, usually, il Bagatto in the Tarocco
Piemontese, is referred to as il Bagatella in most of
the early sources; Pietro Aretino alone uses the
name il Bagatto, and only as an alternative. In the
poem on the 1549 conclave, the form used is il
The term is il Bagatino in the Bertoni
poem, like the form il Bagattino used by Bolognese

The Chariot is usually called il Carro, as it is in
the Tarocco Piemontese, corresponding to le
Chariot in the Tarot de Marseille. In the Steele
MS. and in Aretino it appears more explicitly as il
Carro triumphale (the triumphal Chariot).

The Wheel, called la Roue de Fortune (the Wheel
of Fortune) in the Tarot de Marseille and
likewise la Ruota di Fortuna in the Tarocco
Piemontese, is usually abbreviated to la.Rota or la
(the Wheel) in the early sources; only the
two poems on the conclaves use the full term.
However, the idea is just the same, and Fortune
is almost always mentioned in connection with
this card.

The only serious variations in nomenclature are
confined to three cards:

The card known to modern players as the
Hermit - I'Ermite
in the Tarot de Marseille (le
in some related packs) and l'Eremita in the
Tarocco Piemontese - has three names in the
early sources. In the Steele MS., in the Bertoni
poem and in Lollio, it is il Gobbo (the
Hunchback). In Garzoni, in Susio and in the
poem on the 1522 conclave, as also for Bolognese
players, it is il Vecchio (the Old Man); compare
Viévil's term le Vielart. In Teofilo Folengo it is il
(Time). (Aretino and the other conclave
poem fail to mention it.)

The Hanged Man - le Pendu in the Tarot de
Marseille and il Penduto in the Tarocco
Piemontese - is l’Impiccato in the Steele MS. and
in Garzoni, and l'Appicato in Teofilo Folengo,
both meaning 'the Hanged Man'. For all the rest,
including the Bolognese players, it is il Traditore
(the Traitor).

The Tower, called la Torre in the Tarocco
and la Maison Dieu (the House of
God, or, perhaps, the Hospital) in the Tarot de
Marseille, bears a variety of names in the early
sources. In the Steele MS. it is la Sagitta, literally
'the Arrow' but more probably meaning 'the
Thunderbolt'; for Bolognese players it was la.
, also meaning 'the Thunderbolt' (compare
la Foudre (the Lightning) in the anonymous
seventeenth-century Parisian pack, those of
Viévil and de Hautot and in the Belgian Tarot).
The Bertoni poem calls it la Casa del Diavolo (the
House of the Devil), by which name it was also
known to Minchiate players; the poem on the
1549 conclave has the variant la Casa del Dannato
(the House of the Damned), while that on the
1522 conclave calls it simply la Casa (the House).
In Garzoni, Folengo and Susio it is il Fuoco (the
Fire). Lollio calls it l’lnferno (Hell) outright.
Aretino leaves it unmentioned.

In order to have a uniform terminology for
making comparisons between different orders,
without falsely implying that a particular term is
used in each of the sources, I shall in what follows
use English names (save for the Bagatto). For the
last three cards mentioned above, I shall use 'the

392 Part III: Italian Games and Italian Cards
Hermit', 'the Hanged Man' and 'the Tower'.

Among the variant packs, the Tarocco
Bolognese, considered in its older form before the
Papi were replaced by Moors, contains precisely
the standard trump subjects; it merely fails to
yield an order among themselves for the Pope,
Popess, Emperor and Empress. The Tarocco
Siciliano has several unfamiliar subjects: but the
Globe obviously corresponds to the World, we
know from Villabianca that the Ship is a
replacement for the Devil and we may reasonably^
assume that Jupiter stands in place of the Angel.
In the Minchiate pack, the 'Grand Duke' does
not correspond precisely to any card or other
packs, but clearly belongs with the two Imperial
cards. It is evident that the insertion of the
twenty extra cards has not been allowed to
disturb the order of the familiar ones; it is
especially striking that the four additional
Virtues have not been placed next to the three
that were already there, but have been grouped
together with the other new cards. We are
therefore justified in extracting from the
Minchiate pack an order for the standard Tarot
trumps. This is confirmed by a certain feature of
the design of Minchiate trumps. The trumps
cards from XVI to XXXV, that is, the twenty
additional ones, bear rosettes in the two top
corners. Of the remaining trumps, the top five
unnumbered ones and those from I to XV, only
trumps I and II have rosettes in the top corners.
A possible explanation is that the original
designer of the Minchiate pack, whose designs
were thereafter faithfully copied, used existing
designs for the trumps of an ordinary Tarocco
pack as far as he could, and invented new designs
for the additional cards. Some adjustment would
be necessary with the lowest cards, owing to the
reduction in number of the Imperial/Papal cards
from four to three; the purpose of the rosettes
may have been to indicate to the cardmaker
himself which designs he would be unable to use
for ordinary 78-card Tarocco packs.

In Chapter 4 a list was given of all known
surviving sets of fifteenth-century Italian Tarot
cards. The first twenty of these consisted of the
hand-painted cards, made for the nobility, which
have survived in considerable numbers; but, in
addition, there were listed, as nos. (21) to (24),
four popular packs, dating from the end of the
fifteenth century, from which one or more uncut
sheets have survived; the Boiardo pack and the
copper-engraved Sola-Busca tarocchi, being to a
high degree non-standard, were not included in
the list. The dating of the four-popular paekv
(21) to (24), as late fifteenth-century, is
admittedly not unshakable: any of them may be
assignable to the beginning of the sixteenth. We
may now extend the list to cover Italian Tarot
cards of the sixteenth and seventeenth centuries.
The vogue for expensive hand-painted packs died
away in the sixteenth century; accepting nos. (21)
to (24) of Chapter 4 as of the fifteenth century,
we are left with only four sixteenth-century
Italian Tarot packs of which we have any
knowledge. With the numbering continued from
Chapter 4, these are:

(25) A set of thirty cards belonging to the Leber
Collection in the Bibliotheque Municipale at Rouen.
They comprise the King, Queen, Cavalier, Jack, 9, 6
and Ace of Swords; the King, Cavalier, 9, 7, 5, 4, 2
and Ace of Batons; the 9, 8 and 7 of Cups; the King,
Jack, 9 and Ace of Coins; and the Fool and seven
trumps. The trumps are numbered with Arabic
numerals, and are the Emperor (4), the Pope (5), the
Chariot (7), the Wheel of Fortune (10), Time,
corresponding to the Hermit (11), the Devil (14) and
the Star (16). The pack is obviously non-standard,
and is a classicised one: the court figures are labelled
with inscriptions in Latin identifying them with
characters of classical history (e.g. the King of Coins
with Midas, King of the Lydians), while the trump
cards, although clearly identifiable with the usual
subjects, also have Latin inscriptions interpreting
them in terms of classical mythology (e.g. the Devil is
represented by Pluto and is labelled 'Perditorum
Raptor'). The numeral cards are very elaborate, the
Batons, in particular, being depicted as whole trees. (8)

(26) A complete pack, very closely related to the
one at Rouen, but not identical with it, was known to
Count Leopoldo Cicognara, and was described by
him in his book on playing cards of 1831. (9) He
illustrated it by all four Aces and trump cards
showing Apollo and Cupid, obviously representing
the Sun and Love cards. This pack has now
disappeared; it evidently did not have inscriptions on
8 Catalogue number 135-XIV. The cards measure 134 x 70
mm. Four are illustrated in colour by D. Hoffmann, Die Welt
der Spielkarte
, Leipzig, 1972, plate 23b, nine in black and white
by Kaplan, op. cit., p. 133, and five by H.-R. D'Allemagne,
Les Cartes a jouer, vol. I, Paris, 1906, opposite p. 186.
9 See L. Cicognara, Memorie spettanti alia Storia della
, Prato, 1831, pp. 163-6 and plate XIV; A.M.
Hind, Early Italian Engraving, vol. I, London, 1938, pp. 241
and 243, and vol. V, London, 1948, pp. 139-40; and D.
Hoffmann, op. cit., fig. 6. Hind attributes the pack to
Nicoletto da Modena.

The Order of the Tarot Trumps 393
the trump or court cards, nor, apparently, numerals
on the trumps. The Aces differ considerably from
those at Rouen; furthermore, the Fool of the Rouen
pack is quite different from that described, though not
illustrated, by Cicognara. Cicognara's Fool was a
drunkard lying on his back, supporting, with his legs
in the air, a jar marked MUSCATELLO. The Fool of
the Rouen set shows a man armed to the teeth and
dressed in armour, but with genitals exposed and
urinating; the inscription reads 'VELIM FUND AM
DARI MIHI'. Nevertheless, Cicognara's pack was a
classicised one of very much the same kind as that at

(27) A single card, showing the Devil, is in the
British Musuem. (10) The back gives the maker's name as
Agnolo Hebreo. It is unnumbered, and the design is
very similar to that on the Rothschild sheet (no.
(23)), but it is much cruder in execution.

(28) Another isolated card is in the Museo
Nazionale delle Arti e Tradizioni Popolari in Rome: it
is numbered VIII and represents Love. (11) It shows two
musicians playing a viol da gamba and a lute, and;
behind them, a pair of embracing lovers: above them,
as usual on all versions of this card, hovers Cupid
aiming his bow. The design does not correspond in
detail to any other known version, but is in no way

This may appear a meagre crop for a whole
century, but compares favourably with the single
surviving pack from sixteenth-century France. As
before, we cannot be certain, when dealing with
Italy, that we have identified all surviving cards
from Tarot packs; it is possible that some of the
suit cards that have survived in fact come from
such packs, without our having any means of
knowing this. The wealth of literary references
assures us that the game continued, in the
sixteenth century, to enjoy both popularity and
renown. It is unnecessary to list Tarocchino and
Minchiate packs made in the seventeenth
century, since they are readily identifiable as
such; when these are set aside, our list may be
extended to the seventeenth century as follows:

(29) A set of six tarocchi was found, among other
cards, at the Castello Sforzesco in Milan during
restoration work. They consist of the 6, 7 and 9 of
Swords, the 8 of Batons, the 6 of Coins and the World,
numbered XXI. They measure 138 x 68 mm., and
have backs showing a complex design identified by
Francesco Novati as depicting Ruggero and Angelica;
the design has a dotted border, folded over to form a
similar border for the faces of the cards. The backs,
and consequently the borders, have peeled off from
the 8 of Batons, 9 of Swords and the World. The suit
cards resemble those of the Tarot de Marseille very
closely, with the major exception that they bear no
numerals to indicate their rank, and the minor one
that, on the odd-numbered cards of the Swords suit,
the crosspiece of the single sword is straight, not S
shaped as in most Tarot de Marseille packs (a Tarot
de Besançon made in 1784 by Bernhard Schaer of
Mumliswil being an exception). The single trump, the
World, is similar in general design, though not in
precise detail, to the corresponding card in the Tarot
de Marseille, showing a naked female figure enclosed
in an oval wreath, with the symbols of the four
Evangelists at the corners. The card does not bear an
inscription giving its name. It is inscribed-XXI-above
the top margin, but this inscription would have been
covered by the border folded over from the back when
the card was in its original condition. Novati assigns
these cards to the early sixteenth century; but they
seem more likely to be from some date in the

(30) Two incomplete uncut sheets from a
Portuguese-suited Tarot pack are in the British
Museum. One shows the Maids of Swords and
Batons, the Cavalier of Batons and fragments of the 3
of Swords and Cavalier of Coins. The other shows
trumps bearing Arabic numerals but not names: they
consist of the Wheel (11), the Chariot (10), Love (6),
a card depicting a Sultan and numbered 5, and two
fragmentary cards numbered 20 and 21, presumably
the World and the Angel respectively. The 2 of
Swords bears the inscription 'Alia Colonna in Piazza
Nicosia'. Italian cardmakers from the sixteenth to the
eighteenth century identified themselves by signs
which, like English inn-signs, presumably hung
outside their workshops; usually the sign itself and its
name appear on the backs of the cards. The pack is
thought to have been made in Rome, there being a
Piazza Nicosia in that city. There are also in the
British Museum two fragmentary sheets from a
regular Portuguese-suited pack by the same maker,
showing the Kings, Cavaliers, Maids and Aces of all
four suits, and the 2 to 6 and fragments of the 7 and 9
of the Swords suit. A column, which was also the
heraldic emblem of the Colonna family, appears on a
shield borne by the Maid of Swords in both packs,
and on one borne by the Maid of Cups in the regular
pack; on the 2 of Swords of the regular pack appears
the date 1613. The small details of design of the cards
10 Illustrated in D. Hoffmann, op. cit., plate 14a.
11 Illustrated in Antiche. Carte da Tarocchi, Rome,
plate XII. The back shows a standing Cupid.
12 The cards are in the Raccolta delle Stampe Achille
1961, Bertarelli at the Castello Sforzesco. For Novati's articles, see
footnote 22.

394 Part III: Italian Games and Italian Cards
common to both packs are not precisely the same, but
the pattern used is identical, and corresponds closely
to that of other Portuguese-suited packs made at the
time when this had become recognised as a quite
distinct suit-system. In both packs, the Maids are
girls with long gowns, and the Swords are straight but
intersecting; further, each suit card has an index at
both top centre and bottom centre (FS, FB, FC and
FD for the Maids, RS, etc., for the Kings, CS, etc., for
the Cavaliers, AS, etc., for the Aces and S2, S3, etc.,
for the numeral cards from 2 up). In the regular pack,
the Kings are seated, there are dragons-on-t-he- Aeesand
the Batons have a shape very characteristic of
Portuguese-suited packs.(13)

(31) There are several extant examples of a
puzzling kind of pack made by a cardmaker using the
name 'Orfeo': the backs of his cards give this name
with an elaborate design of Orpheus playing his lute.
The Beinecke Library at Yale has two nearly
complete ones and another, very fragmentary, one;
two others are in the British Museum, while the
Fournier Museum in Vitoria has fragments of three
packs (and one card from a fourth).(14) These packs are
invariably catalogued as incomplete Minchiate packs.
The reason is that the trumps display exactly the
designs of the Minchiate ones: the top five trumps, left
unnumbered in the Minchiate pack, are unnumbered
in the Orfeo packs also, and those that are numbered
bear the same numbers as in the Minchiate pack.
Furthermore, as in the Minchiate pack, the Swords
are straight but intersecting. In other respects,
however, the Orfeo packs deliberately diverge from
the highly constant Minchiate designs. This is true of
the Fool and of the Aces, and in some degree of the
Kings; but the most striking divergences are in the
two lower court cards. In the Minchiate pack, the
Cavaliers, although still called Cavalli, are centaurs in
Swords and Batons, and other half-human, half-
animal creatures in Cups and Coins; but in the Orfeo
packs, they are the conventional mounted knights. In
the Minchiate pack, the Cups and Coins suits, though
not the other two, have Maids instead of Jacks as the
lowest court cards; but in the Orfeo packs, there are
distinctly male Jacks in all four suits, and their
designs in no way resemble the Minchiate ones. If
13 See Playing Cards of Various Ages and Countries selected from
the Collection of Lady Charlotte Schreiber
, vol. Ill, London, 1895,
plates 44 and 45. The Tarot pack, but not the regular one, is
illustrated in Kaplan, op. cit., p. 134.
14 The new catalogue no. of the more complete Orfeo pack
in the Cary Collection at Yale is ITA-63; the old no. was I-
11, the other Orfeo pack being 1-96. The Orfeo packs in the
Fournier Museum are grouped as no. 10 in the Italian
section of the catalogue. For the British Museum ones, see
F.M. O'Donoghue, Catalogue of the Collection of Playing Cards
bequeathed to the Trustees of the British Museum by the late Lady
Charlotte Schreiber
, London, 1901,1-59 and 1-60.

there were anything to compel us to regard the Orfeo
packs as intended for the game of Minchiate, we
should have to treat, them as exhibiting certain
deviant features; but there is not. Although the
designs of all surviving trump cards from Orfeo packs
coincide with those used in Minchiate packs, no Orfeo
pack is known which has any of those twenty trump
subjects which are peculiar to the Minchiate pack,
which, unlike every other form of the Tarot pack, had
in all forty trumps (in addition to the Fool). The
probability is, therefore, that the Orfeo packs were
originally 78 card ones.Perhaps the designs used
were not at that time regarded as the exclusive
property of the Minchiate pack, which had probably
taken them over, at the time of its invention in the
previous century, from some local standard pattern
for the normal Tarot pack; or, possibly, the
cardmaker using the sign Orfeo found it more
economical to use Minchiate blocks, so far as he
could, to produce ordinary 78-card Tarot packs. This
conclusion cannot be regarded as certain, however,
because one oddity remains. In none of the Orfeo
packs I have mentioned is there any trump below IX
(or Villi as it is written). To make a complete set of
Tarot trumps, there would have to be nine such
trumps, perhaps leaving the Bagatto unnumbered, or,
just conceivably, with a trump XVI to insert between
the extant XV (the Tower) and the five top
unnumbered trumps: in any case, they would have to
diverge in some respect from the Minchiate trumps,
which do not include the Pope and Popess and have
only the eight trumps numbered I to VIII' below the
IX. It may be mere coincidence that these nine
trumps happen to be missing from all surviving
examples of the Orfeo pack; but, until an example is
found that includes some of them, it remains a fact
that seems to call for explanation, and, until one is hit
on, we cannot feel assured that we have correctly
interpreted the pack.

Considering the rarity of pre-eighteenth-
century cards generally, we cannot regard this as
a poor haul; and we have, in addition,
seventeenth-century examples both of the
Minchiate and Tarocchino packs. Nevertheless,
the dearth of literary and textual references, save
to Minchiate, from this century suggests that the
popularity of the game of Tarocco, as played
with the 78-card pack, was distinctly on the wane
in mainland Italy,, though Minchiate and
Tarocchino continued to flourish. In the next
century, of course, games with the 78-card pack
were to enjoy a great revival in Lombardy and

Of all these packs, made in Italy between the
fifteenth and seventeenth centuries inclusive,

The Order of the Tarot Trumps 395
there are, apart from the Minchiate and
Tarocchino packs, just nine that yield an
order, complete or incomplete, for the trump
cards. Two of these are fifteenth-century ones
printed from wood blocks and preserved in the
form of uncut sheets: that in the Metropolitan
Museum, New York, no. (21) of Chapter 4; and
that in the Rosenwald Collection in Washington,
no. (22) of Chapter 4. The Metropolitan set
includes an almost complete sequence of trumps,
of which the World is definitely unnumbered,
and the rest appear to have borne Roman
numerals from I to XX. The sheets are
mutilated, so that a few of the trumps are missing
or fragmentary. Only the tops of the Pope,
Emperor, Popess and Empress survive, bearing
the numerals II to V, but only the Popess (III)
can be identified with certainty, although, given
that, the positions of the others admit of little
doubt. Another card of which only the top
remains may be either the Bagatto or the Fool:
no numeral is to be seen. There is only the right=
hand half of the Chariot, without the numeral. Of
three cards, only the left-hand halves survive: one
is easily identified as the Hanged Man, with a
numeral beginning XL. that can only be XII;
one appears to be the Star, with a numeral
beginning XV.., which would have to be XVI,
but may possibly be the Moon, in which case the
numeral must be XVII; and the third, which is
also truncated below, is unidentifiable and shows
no numeral, but is probably Fortitude. Save for
the relative positions of Fortitude and the
Chariot, it is possible to reconstruct the order
completely with virtual certainty.

In the Rosenwald set, all the trumps, without
the Fool, are printed on one sheet, together with
three Queens. The only mutilated card is the
Wheel, on which the numeral, if there was one,
can no longer be seen. Unlike on the
Metropolitan sheets, the trumps are arranged on
the sheet more or less in sequence. The bottom
line contains the three Queens and the first five
trumps from the Bagatto (I) to the Pope (V). The
middle line begins with Love (VI), Temperance
(VII) and Justice (VIII). These are followed by
Fortitude, also numbered VIII, and the Chariot,
numbered X. Evidently the VIII on Fortitude is a
mistake for Villi. There follows the Hermit,
numbered XII, the Hanged Man, which is
definitely unnumbered, and the Wheel, of which
we cannot tell whether it had a numeral. The
cards on the top line, running from Death to the
Angel, are all unnumbered, and are arranged in
a plausible order, indeed, in what, by analogy
with other packs, is the only possible order, given
the numbers assigned to the other cards. There
are two possible hypotheses about the end of the
second line. One is that the cards are arranged in
the correct sequence, but that the Hermit has
been misnumbered XII instead of XL On this
hypothesis, proposed by Sylvia Mann, the Wheel
was not numbered, and the numbering stopped at
XL An alternative hypothesis seems to me a little
more probable. This is that the Hermit is
correctly numbered XII, and that the Wheel was
numbered XI, but was located slightly out of
sequence on the block. I shall follow this second
hypothesis in the comparative table given below.

Although fifteenth-century hand-painted
packs do not usually have numerals on the
trumps, there are two exceptions to this. One is
the celebrated 'Charles VI' pack, no. (4) in
Chapter 4. A fact seldom referred to in the
extensive discussions of this pack in the literature,
is that the trumps bear lower-case Roman
numerals at the very top, in a fifteenth-century
hand. Robert Steele listed these in his 1900
Archaeologia article. The cards are in fact printed
from a wood block with the colours subsequently
painted by hand; and, on the basis of some
technical considerations concerning the process
of production, Steele asserted that the numerals
were written on the cards before they were
painted. Detlef Hoffmann has denied this,
maintaining that the numerals were added later,
and were not intended to be part of the original
designs. In this he is almost certainly right. They
do not lose their importance for that reason: they
represent an order which, at an early period,
their owner at the time thought they ought to
have. I have not seen these cards myself, and rely
on Steele for the numbering. In one particular, he
seems likely to be wrong: he gives the numeral for
the Pope as ii, and adds a question mark to show
his uncertainty about the reading. But the Pope
can hardly rank lower than the Emperor, which
is iij; moreover, in every other case, a terminal i is
written j . It is therefore probable that the
numeral on the Pope was intended to be iiij.

The other exception is the very incomplete
Catania set no. (7) in Chapter 4. On three of the
four surviving trumps Arabic numerals have at
some time been inscribed in ink; these
inscriptions cannot be contemporary with the
cards, and may be conjectured to have been

396 Part III: Italian Games and Italian Cards
made in the seventeenth century. The fourth
surviving trump is the figure on the stag, which
Ronald Decker has interpreted as an unusual
representation of Temperance; to this no
numeral has been added, presumably because
whoever added them could not identify the

From the sixteenth century we have two
remnants of packs with numbered trumps. One
is the single card at Rome depicting Love and
numbered VIII, no. (28) above. The other is the_
classicised pack at Rouen, no. (25) above, whose
trumps, readily identified with their counterparts
in ordinary packs, bear Arabic numerals and
inscriptions giving their names in Latin; I differ
from Detlef Hoffmann in equating Pluto, not, as
he does, with Death, but with the Devil.

Finally, there are three fragmentary packs
from the seventeenth century. One is the very
inomplete sheet of trump cards for the
Portuguese-suited alia Colonna pack, no. (30)
above; this has Arabic numerals, which were
presumably borne by all the trumps, since they
go up to 21. The second is the Orfeo pack, no.
(31) above, of which several examples exist; this
has Roman numerals, but the top five trumps are
unnumbered, as in the Minchiate pack. Finally,
there is the set of six Tarot cards found at the
Castello Sforzesco, of which only one is a trump,
the World, numbered XXI.

From the four literary sources and from the
three variant types of Tarot pack, we obtain
complete orders for all the trumps; from the nine
fragmentary early packs with numbered trumps,
we obtain further orders with varying degrees of
incompleteness, of which those of the
Metropolitan Museum, Rosenwald and Charles
VI packs can be reconstructed in their entirety
with very little uncertainty. Of the orders which
we know in complete detail, only two agree
exactly, those given by Garzoni and in the
Bertoni poem; the order in the Metropolitan
Museum pack almost certainly also nearly
coincides with that in these two sources, and that
of the Rouen pack may very well have done so as
well. The order in the alia Colonna pack may
have been the same as that in the Rosenwald
pack, save for carrying the numbering through to
21 and having a Sultan in place of the Pope; the
order underlying the numbering on the three
Catania cards may have been the same as in the
Orfeo pack, save that the numbering is carried at
least as far as 19. All the rest have at least minor
differences between them. Ignoring the isolated
Love card at Rome, we thus have eleven distinct
orders, all differing from the Tarot de Mlarseille

This is a very surprising fact. Games players
do not in the least mind having to master a
complicated and arbitrary sequential ordering
(for instance, the ranking, in 'civil' and 'military'
suits, of dominoes in Chinese domino games);
but they do require that any such ordering be
held constant. It is of the essence of Tarot games
that there be a determinate means of deciding
which card in any trick is the winning one, and
this necessarily requires an agreed ordering of
the trump cards; in special cases there may be
exceptions, such as the equal ranking of the Papi
or Moors in the Bolognese game, but anything of
this kind is necessarily an exception, not the
general rule, and, indeed, this is the only known
such exception. If play is to be possible, the
ranking of the trumps must be apparent to all
players and subject to no dispute.

How, then, are we to explain the variations
that we find in the order of the trumps in Italy
from the fifteenth to the seventeenth century?
One explanation, which we glanced at earlier,
can be ruled out immediately, although it has
been proposed by some, the suggestion, namely,
that the Tarot trumps did not originally have an
order at all. If what, in the early chapters of this
book, we called the triumph cards did not have
an order, they were not trumps, and the game
played with the pack cannot have been a trick-
taking game; in that case, the suit cards cannot
have had an order either, since it is only in trick-
taking games that an order is required. We should
then have to suppose that, at some time between
the first invention of the Tarot pack in the 1430s
and its spread to France and the appearance of
packs with numbered trump cards, say around
1480, a new type of game was invented, for play
with the Tarot pack, and a new employment,
within this game, found for the triumph cards,
namely as genuine trumps. We should have to
assume this, to account for the etymological
connection between triumphi and 'trump', for the
subsequent history of the game of Tarot in Italy
and in France, and for the reversed ranking, in the
two pairs of suits, of the numeral cards, already
going out in Italy by the beginning of the sixteenth
century, as the game of Trappola shows. Above
all, we should have to assume it to explain the fact
that from the author of the fifteenth-century

The Order of the Tarot Trumps 397
Steele sermon on, writers assign an order to the
trump cards. But then we are faced with the same
problem as before: why are the orders different in
different sources? The hypothesis completely fails
to explain that which it was its sole object to
explain. We have no evidence that the trump
cards ever lacked an order; we have abundant
evidence that they had an order: our problem is
that this order is not constant. It is no explanation
of the variations in the order to propose that, at the
start, there was no order at all.

A more specious explanation is that, once it
became the practice to inscribe numerals on the
trumps, it ceased to be important to maintain a
fixed order for the trump subjects. Our surprise
at the variations in order is due, on this view, to
our being accustomed to the Italian-suited packs
used outside Italy, with their unswerving
uniformity of order: but there was in fact
no need for such uniformity. In the early French-
suited packs which have animals on the trumps,
it is common to find the same selection of
animals arranged in different orders; in the same
way, the standard selection of trump subjects in
the Latin-suited packs may have varied without
causing any confusion.

The idea underlying this suggestion is that,
once the trump cards bore numerals, the players
would have identified them from those numerals
and not from the subjects depicted, as they
unquestionably did from the start with the
French-suited packs. But this idea is borne out
neither by the early Italian Tarot cards that
survive to us, nor by the literary references. It is
correct, indeed, for the Minchiate trumps, which
were too numerous for anyone to memorise the
order of any but the five top cards, the arie; and,
accordingly, in literary sources, the first thirty-
five Minchiate trumps are virtually always
referred to by number and not by name; we do
not even know what subject Minchiate trump II
was intended to represent. By contrast, the
tarocchi trumps are never referred to by number; of
the various literary texts that mention them, only
one, the Steele MS. sermon, even cites their
numbers. Of the four sets of fifteenth-century
sheets for Tarot cards, the Cary sheet and the
Rothschild/Beaux Arts sheets have no numerals
at all, while the Rosenwald sheet leaves as many
as nine cards unnumbered; and we know that at
Bologna it was not until the later eighteenth
century that any of the trumps began to be
numbered. On the trump cards of the later
French-suited Tarot packs, the numerals form
one of the most prominent features. In those used
in Germany, Switzerland, Belgium and, eventually,
France, these numerals are contained in a
separate panel extending across the card; though
this is not true of the packs made in the Austrian
Empire, the numerals are still very prominent.
This is not so with the early Italian cards. Only
on the alla Colonna trumps are the Arabic
numerals clear and regularly placed. The
numerals on the Metropolitan Museum trumps
are quite insignificantly placed; for instance, the
figures XV appear on the left of the head of the
Angel, while the remainder of the numeral, IIII, is
set on the right; on the Sun card, the figure XVIII
is set more than half-way down the card, just
above the trees on which the Sun is shining. One
can hardly suppose that these numerals were
intended as more than a last resort in identifying
the cards. Something very interesting happens in
the Minchiate pack. On the trumps peculiar to
that pack, trumps XVI to XXXV, the numerals
are placed in a scroll at the top of each card,
making them easy to pick out; but, on the fifteen
lowest trumps, the numerals are placed in a much
more random fashion in blank spaces of the
design. It is probable that these lowest trumps - or
at least those from V to XV - represent designs
originally used for trumps of a 78-card tarocchi
pack, taken over when the Minchiate pack was
first invented. The numerals on the Orfeo trumps
are, of course, placed in the same way, and those
of the Rosenwald sheet, though clearer than the
Metropolitan Museum ones, are also far from
being very prominent. As for the numeral on the
Castello Sforzesco World card, it would have
been invisible to a player, and perhaps was
intended only as a reminder for the cardmaker

From all this it is plain that Italian players
were highly conscious of the trump subjects, and
did not rely principally upon the numerals in
order to identify the cards. If the true explanation
of the variation in the order of the subjects were
that the subjects did not matter, the cards being
identified by numeral, we should expect that at
least the subjects of the really important qards
would have been held constant; yet we shall find
that the variation affects even the top three
trumps. In some orders, the Angel is the highest
trump, followed by the World and then by the
Sun; in others, the World is the highest, followed
by Justice and only then by the Angel; or, again,

398 Part III: Italian Games and Italian Cards
the World may be the highest, the Angel second
highest and then the Sun. For the variations to be
explicable on the ground that the players were
indifferent to the subjects associated with the
numbers, they woud have had to be altogether
oblivious to them for the variations at this level to
have passed unnoticed, whereas, from the way in
which the trumps are referred to in the literary
sources, it is apparent that they were not. The
most striking example of the importance of the
subjects, at least at the highest level, is the_
persistence in Piedmont of the tradition that the
Angel beats the World, in face of the contrary
numerical ordering on the cards being used. It
seems probable, rather, that the very reverse of
the present suggestion is close to the truth; that
just as, at Bologna, a player of tarocchi had, until
the later eighteenth century, to make the fixed
conventional ordering of the trumps second
nature to him, so players of other varieties of
Tarot, save for Minchiate, identified the trump
cards principally by subject, and were aware of,
and could have stated, the order =of those

If this was so, it would have been hopelessly
confusing for the players if the order of the
trumps had varied even in minor respects from
one pack of cards to another. The observable
variations in the order must therefore be due, not
to the absence of a fixed order, but to that
phenomenon evident throughout the entire
history of the game of Tarot: the extreme
localisation of specific modes of play. Again and
again we find that the players in one city or town
play only amongst themselves and do not know
those of a neighbouring town; the detailed rules,
and sometimes the whole type of game played,
diverge from locality to locality, the players in
one circle being quite unaware of the manner of
play of those in another, and, often, of their
very existence. The different orders for the
trumps that we find in Italy .must represent
different practices adopted in different cities,
presumably at a stage earlier than that at which
numerals came regularly to be inscribed on the
trump cards. Evidently, quite a short time after
the game of Tarot had first been invented,
players in various cities or regions developed
local peculiarities in their modes of play, which,
in Italy, extended to the conventional order of the
trumps; this must have happened before it
became usual anywhere to inscribe numerals on
the trump cards, and hence before the end of the
fifteenth century. Modern piayers might feel that
it would be impossible to memorise the order of
twenty-one trump subjects so accurately as to be
at once aware, without the need for reflection,
which card was superior to which; but, as we
know from the Bolognese game, this doubt is
quite misplaced. The different orders of the
trumps testify, not to a reliance on the numerals
alone, but to the existence, at an early date, of
wide local variation in the manner of play.
Whenjwe look closely at the various orders, we
find that there was far from being total chaos. A
first impression is of a good deal of regularity
which, however, is hard to specify. Now the cards
which wander most unrestrainedly within the
sequence, from one ordering to another, are the
three Virtues. If we remove these three cards, and
consider the sequence formed by the remaining
eighteen trump cards, it becomes very easy to
state those features of their arrangement which
remain constant in .all the orderings. Ignoring the
Virtues, we can say that the sequence of the
remaining trumps falls into three distinct
segments, an initial one, a middle one and a final
one, all variation in order occurring only within
these different segments.

The first segment consists of the Bagatto and
the four Papal and Imperial cards (three in the
Minchiate pack, two only in the Tarocco
Siciliano). Save in the Tarocco Siciliano, where
the Miseria or Poverta comes below it, the Bagatto
is always the lowest trump. The Pope, when
present, Is always the highest member of this
initial segment. In all known Italian trump
orders, the Emperor ranks higher than the
Empress, as one might expect; but it will be
recalled that in the pack made at Rouen by
Adam de Hautot, and also in the list given in the
Maison academique des jeux of 1659, the Empress
outranks her spouse. In Italian trump orders, on
the other hand, the only opportunity for variation
within this segment lies in the position of the
Popess. In different orders, she occupies one of
three possible positions: immediately below the
Pope and above the Emperor; below the
Emperor but above the Empress; and below even
the Empress.

The middle segment consists of five cards, of
which the typical order is, from lowest to highest:
Love, the Chariot, the Wheel, the Hermit, the
Hanged Man. That very order is found in the
Tarocco Bolognese, the Charles VI numbering
(the Wheel cannot have come anywhere but at

The Order of the Tarot Trumps 399
no. 10), the Rosenwald pack (on the above
hypothesis as to the intended order), the Steele
sermon, Susio's poem and the Viévil pack; the
alia Colonna pack must also have had that order,
unless the Hermit and the Hanged Man- were
reversed. In all other cases, the order within this
segment results from interchanging some one
pair of adjacent cards: in the Minchiate and
Orfeo packs (and presumably in the numbering
of the Catania cards), the Wheel and the Chariot
have been interchanged; in the Tarocco Siciliano
the Hermit and the Hanged Man; in Garzoni's
book, the Bertoni poem and probably the
Metropolitan Museum pack, Love and the
Chariot; and, in the Tarot de Marseille, the
Wheel and the Hermit. (If Miss Mann's
hypothesis, mentioned above, concerning the
order in the Rosenwald pack be adopted in
preference to mine, that pack forms the sole
exception to this rule.)

The final segment consists of Death, the Devil,
the Tower, the Star, the Moon,-the Sun, the
World and the Angel. These always occur
(ignoring possible intervening, Virtues) in
precisely the order just stated, with the sole but
very important exception that, sometimes, the
position of the World and the Angel are reversed, '
the World coming highest.

If, now, in the light of this analysis, we look at
the actual orders, we see that they divide into
three sharply distinct types, which I shall
arbitrarily label type A, type B and type C.
These types are to be distinguished according to
two principles: where the Virtues come; and
whether the Angel or the World is the highest
card. In type A, the Angel is the highest trump,
the World coming immediately below it. The
three Virtues, Temperance, Fortitude and
Justice, occur consecutively, usually interposed
just above the lowest card of the middle segment,
which, in orders of this type, at least whenever we
can tell, is invariably Love. Type A is not attested
by any of our four literary sources. On the other
hand, it is well supported by actual packs. All
three variant packs - the Tarocco Bolognese, the
Tarocco Siciliano and the Minchiate pack -
belong to this type; so do the Charles VI
numbering and the Rosenwald and Orfeo packs.
So also must the alia Colonna pack have done: at
least, if the three Virtues did not come between
Love as no. 6 and the Chariot as no. 10, the order
must have been very non-standard. Almost
certainly the Catania numbering also exemplifies
this type: the numbering of the Chariot as 10 and
the Hermit as 11 surely implies that all three
Virtues ranked below the Chariot, and the World
as 19 must have ranked below the Angel.

In orders of type A, the three Virtues rank
immediately above Love, except in the Tarocco
Siciliano pack, where they rank immediately
below it, and in the Tarocco Bolognese, where
they outrank the second lowest card of the
middle segment, the Chariot. In type A orders,
Temperance is always the lowest of the three
Virtues, whenever we can tell. In the Tarocco
Bolognese and the Rosenwald pack, Fortitude is
higher than Justice, but in the Tarocco Siciliano,
the Minchiate pack and the Charles VI
numbering, it is Justice which is higher.

400 Part III: Italian Games and Italian Cards
In orders of type B, something completely
different happens. In these, the World is the
highest trump, and Justice is promoted to the
second highest position in the sequence, coming
immediately below the World and above the
Angel, the third highest card. There is clearly
here an association of ideas: the Angel proclaims
the last Judgment, at which justice will be
dispensed. In orders of type B, Temperance
always comes immediately above the Pope, and
is separated from Fortitude, which comes three^
cards later, after Love and the Chariot. There is
very little variation in these orders: the Popess
does not have a stable position, and, in the Steele
MS., the positions of Love and the Chariot have
been reversed.

Type B has by far the best literary attestation,
namely in three out of our four sources, including
the earliest one, the Steele MS. The other two are
Garzoni and the Bertoni poem, which give
exactly the same type B order. The type is also
exemplified by the Metropolitan Museum pack,
and, incomplete as it is, by the Rouen pack. The
isolated card at Rome, Love, numbered 8,
probably also comes from a pack with a type B
order, although this cannot be certain, there
being one type A order, that of the Tarocco
Siciliano, in which Love bears the number 8.


We have only one certain attestation to an
order of type C as being used in Italy before the
eighteenth century, namely Susio's poem. If we
knew nothing of non-Italian Tarot cards, or of the
post-1700 Lombard and Piedmontese patterns,
we might dismiss type C as a minor curiosity;
but, as the type to which not only the Tarot de
Marseille order, but all those used outside Italy,
belong, it is of course of great importance. Susio's
order is considerably different from that of the
Tarot de Marseille; it is, in fact, almost precisely
the order found in Viévil's pack, save for the
relative order of the Empress and Popess. In an
order of type C, the World is again the highest
card in the sequence, but, this time, the Angel
comes immediately below it. Of the Virtues, it is
Temperance that is promoted to a relatively high
position, namely to just above Death and just
below the Devil; any symbolic appropriateness in
this escapes me. The remaining two Virtues are
again separated and scattered within the middle
segment, Justice in all cases coming lower. In
Viévil's and Susio's orders, Justice comes just
above the first card of the middle segment, Love,
and Fortitude just above the next one, the
Chariot. In the Tarot de Marseille, Justice comes
above the first two cards of the middle segment,
Love and the Chariot, and Fortitude above the
next two, the Hermit and the Wheel. The single
trump card, the World, numbered XXI, from the
Castello Sforzesco set, in itself of course indicates
no more than that the order was of type B or type

It will be remembered, in connection with the
following table, that, in Viévil's pack, as in those
of de Hautot and of the anonymous Parisian
cardmaker, as well as in the Belgian Tarot,
trump XVI is actually called the Lightning and
does not show a tower; nor does that by Catelin
Geoffroy, though we do not know what name it
bore. The order in the anonymous Parisian pack
coincides with that of the Tarot de Marseille; so
do those of the Tarot de Besançon and of the
Belgian Tarot, save for the difference of subjects
on trumps II and V. With the same reservation,
the order in de Hautot's pack coincides with that
given by the Maison academique. It is apparent
from the table that the order in Catelin
Geoffroy's pack, unless very eccentric, must have
been that of the Tarot de Marseille; virtually the
only alternative is that the Wheel was numbered
VIII and Justice and Fortitude X and XI, which
is quite unlikely.

While literally true, it is somewhat misleading
to say that ten or eleven distinct orders were
known in Italy before the eighteenth century.
The variations within type B are very minor ones.
Those within type A are more considerable: but

The Order of the Tarot Trumps 401

if we set aside the Tarocco Siciliano, in which we
know that some dislocation occurred, the major
deviation is seen to occur in the Bolognese order,
in which the Virtues are placed after the Chariot
instead of before it; for the rest, the variations
within this type are, so far as we know, again
comparatively minor. We may thus regard
Italian Tarot players of the fifteenth to
seventeenth centuries as having observed three,
or perhaps four, distinct basic orderings of the
trumps, with small variations from one area to

So far we have paid no attention to the
numbering of the trumps; it might be thought
that this followed automatically from their order,
but this is not so. If we study the various
numberings, we find very little in the way of any
close association of numbers with particular
subjects. Almost the only such association is that
of the number 13 with Death. Even that is not
invariable: but it occurs more frequently than the
association of a particular number with any other
card, even that of the number 1 with the Bagatto.
It is difficult to avoid the conclusion that the
cardmakers, or those for whose tastes they were
catering, regarded this association as particularly
appropriate, and strove to arrange for it. Even
today, some superstition attaches to this card
among certain card players: a student who
played French Tarot with lorry drivers in France,
using a Tarot de Marseille pack, reported that
they considered it bad luck to be dealt trump
XIII, and would not, in such an event, enter the
bidding, however strong a hand they otherwise

Now if one has the trumps arranged in an
order of type A, and begins the numbering with
the lowest trump, Death receives the number 14.
This can be seen from the Rosenwald pack,
where the numbering stops at 12: if it had been
continued, the number 14 would be assigned to
Death. It is probable that this consequence was
accepted in the alia Colonna pack; it must have
been accepted if the order in that pack was truly
of type A, and none of the cards of the middle
segment was promoted above Death. But there
was a solution which allowed Death to be
assigned the number 13 in an ordering of type A,
and this can be seen from the Tarocco Bolognese.
In the Tarocco Bolognese, only trumps 5 to 16
are numbered: but if the numbering is continued,
it will only reach 20; the Moon will become no.
17, the Sun no. 18, the World no. 19 and the
Angel no. 20. This is because Love, as no. 5, has,
not four, but five cards below it. We could not
number the Moors, because they are all equal:
but, if they were numbered, they would occupy
positions 1 to 4, leaving no number for the
Bagatto, which would thus have to be regarded
as an unnumbered card ranking below thfe
numbered ones, like the Miseria in the Tarocco

We can see that the principle of starting the
numbering with the second lowest trump,
followed in the Tarocco Siciliano and, in a
concealed fashion, in the Tarocco Bolognese, also

402 Part III: Italian Games and Italian Cards
underlies the numbering of the Charles VI
trumps. In that numbering, the Angel holds the
highest place, and has the number 20. There is
no other card which could possibly rank above it
as no. 21: the World is numbered 19, and Justice
is numbered 8. On the assumption that the
Popess and the Empress were present to occupy
between them the positions numbered 1 and 2,
the numbering must have begun after the
Bagatto, in order to bring Death out as no. 13.
The numbering of the Catania cards provides
another example.

The Orfeo pack presumably supplies yet
another instance of the practice: there can hardly
be any cards other than the five unnumbered
ones to rank above the highest-numbered one,
the 15, so that, to have a full complement of
twenty-one trumps, there must be nine cards
below the lowest-numbered one surviving in any
of the packs, the 9. If the Orfeo packs are
representative of the type of 78-card pack from
which the Minchiate pack was derived, there are
two possibilities. The sequence may have
continued down from 8 to 3 as in the Minchiate
pack, with 4 and 3 as the two Emperors, and then
have had two corresponding Empresses as nos. 2
and 1, and, below them, an unnumbered
Bagatto. Alternatively, it may have run from 8
down to 1 exactly as in the Minchiate pack, the
Bagatto thus being numbered 1, and have had
some unnumbered card ranking below the
Bagatto like the Miseria of the Tarocco Siciliano.
Of these two possibilities, the former seems a
little more likely; but yet others are thinkable,
and it would throw great light on the history of
the Italian Tarocco pack in the seventeenth
century if a complete Orfeo pack were to be

In orders of types B and C, one of the Virtues -
Justice in type B and Temperance in type C - is
promoted higher than Death, with the result:
that, when all the trumps are numbered in
sequence from the Bagatto up, Death comes out
as no. 13, without the necessity for any special
device to secure this result. In consequence, we
never find a trump sequence of either of these
types that leaves the Bagatto unnumbered and
starts the numbering with the next card. It is for
this reason that it has been possible to classify the
incomplete Rouen set as exemplifying type B. In
the Rouen pack, the Star is numbered 16. If the
pack is at all like others that are known, there are
only two possible explanations of this. Either
only four cards rank above the Star, in which
case the numbering, must start with the card
above the Bagatto; or five cards rank above the
Star, and, in that case, one of them must be
Justice. The former possibility, seems, in the
Rouen pack, to be ruled out by the fact that the
Pope and Emperor are numbered 5 and 4

We have now to enquire in which areas the
different orders were observed; and we must use
whatever, clues, we can extract from the four sets
of late fifteenth-century wood-block printed
sheets for popular Tarot packs, numbered (21) to
(24) in Chapter'4. Just as there is a temptation to
say, at first glance, that there was no fixed order
for the trumps in Italy before the eighteenth
century, so there is the parallel temptation to say
that no standard pattern was adopted for the
Tarot pack, because, among the early cards that
survive, one can scarcely find two sets exhibiting
the same type of design. This temptation is
equally to be resisted. We have seen it to be a
universal law, applying to Indian and Chinese
cards as much as to European ones, that in any
locality any specific type of playing-card pack
very rapidly assumes a stereotyped design to
which all cardmakers conform, for the simple
reason that players need to be able to recognise
each card at a glance. There is no ground
whatever to suppose that the Tarot pack was any
exception to this rule. The variations in design
that we can observe amongst surviving cards are
to be explained in the same way as those between
different orders of the trumps subjects, namely as
representing different standard patterns used in
different regions. This does not, of course, apply
to the hand-painted packs, which were luxury
items, nor to obviously non-standard packs
such as the Rouen one or the Sola-Busca tarocchi:
but, though we can never with certainty identify a
design as a standard pattern when we have only
one example of it, it is highly probable that each of
our four sets of sheets (21) to (24) exemplifies one
of the standard patterns in use towards the end of
the fifteenth century in some particular locality.
We must therefore investigate whether it is
possible to identify the areas in which those
standard patterns were used, simultaneously with
our enquiry into the regional associations of the
various trump orders.

The easiest set of sheets with which to start is
the pair divided between the Ecole des Beaux
Arts and the Rothschild Collection (no. 23).

The Order of the Tarot Trumps 403
Detlef Hoffmann describes these as Minchiate
cards, and W.L. Schreiber, presumably with the
same idea in mind, assigns them to Florence;
Stuart Kaplan characteristically hedges his bet,
describing them as 'Tarot or Minchiate cards'. (15)
Hoffmann mentions, only to reject, an
identification of them by Sylvia Mann as
Tarocchino cards; but a comparison between
them and a modern Tarocco Bolognese pack will
at once bear her out,- revealing striking
correspondences in design. If, instead, the
comparison is made with the seventeenth-
century standard single-ended Tarocco
Bolognese in the Bibliotheque Nationale, the
designs will be found to tally in almost every
detail; the one exception is the Devil, the design
of which on the Rothschild sheet is completely
different from that of the later Tarocco
Bolognese. The resemblance between the
seventeenth-century pack and set no. (23) is
overwhelmingly close in the cases of the Angel,
the World, the Sun, the Moon, the Star, the
Tower, the Hermit and the Chariot; note
particularly the rayed arcs which appear in the
upper corners of several cards in both packs.
Death in the later pack faces the opposite way,
but is otherwise similar, down to the band at the
top of the card, save for the position of the horse's
head. The direction of motion of the Wheel, and
a few other details, differ in the later card, but
they are still fairly similar. The arms of the
Traitor or Hanged Man of the later pack differ
from those of the earlier one in being bound
behind him, instead of hanging down grasping
money bags. It is only in the design of this last
card in set no. (23) that there is any similarity
with Minchiate cards, whereas, on the eight
cards singled out above, detail after detail
corresponds exactly with the Bolognese cards.
This is not, indeed, to endorse Miss Mann's
characterisation of these as Tarocchino cards,
since the use of this term presupposes that the
pack had already been shortened to 62 cards, and
this, of course, we cannot judge from twelve
surviving trumps; on the whole, it is probable
that the shortening had not yet taken place. But
we can confidently assign set (23) to Bologna,
and conclude that the standard pattern used
there from the seventeenth century to the present
15 D. Hoffmann, op. cit., p. 66; W.L. Schreiber, Die altesten
, Strasbourg, 1937, p. 104; S.R. Kaplan, op. cit.,
pp. 128-9.

day was already in existence by the end of the
fifteenth century. It will be recalled that the
single sixteenth-century unnumbered Devil card
by Agnolo Hebreo in the British Museum (no. 27
above) resembles that on the Rothschild sheet,
and is therefore presumably also to be assigned
to Bologna: the change in the design of this card
must have occurred between the mid-sixteenth
and the mid-seventeenth century.

The next in order of difficulty is the
Rosenwald set (no. 22 of Chapter 4). This is
certainly not a Minchiate pack, since it has only
twenty-one trumps. But, although the Swords
are curved, it has several Minchiate
characteristics: (i) the Cavalli, in all four suits,
are centaurs, like those in Swords and Batons in
the Minchiate pack; (ii) the lowest court figures
in Cups and Coins are Maids, whilethose in
Swords and Batons are Jacks; and (iii) the Kings
in Swords and Batons wear short tunics, those
in Cups and Coins long robes. It is true that
there is no close relation between the designs for
the trump cards and the corresponding ones of
the Minchiate pack, save for a noticeable
similarity in the case of the Hermit and the
Hanged Man; but the order of the trumps is of
type A, nearly, though not quite, corresponding
to that of the Minchiate trumps when the twenty
additional ones are removed (the positions of
Fortitude and Justice, and, apparently, those of
the Chariot and the Wheel, are reversed). It
therefore seems probable that the set represents
an early form of that standard pattern for the
Tarot pack on which the Minchiate designs were
later based, or some closely related pattern. The
Rosenwald sheets are thus very likely to have
been made in Florence (or possibly in some other
city of Tuscany such as Piistoia). (16)

The type A orders are associated with
Florence (by the Minchiate and Orfeo packs,
and, on the basis of its resemblance to the former,
by the Rosenwald one); with Rome (by the alia
Colonna pack, which was certainly made there,
and also by the Minchiate pack, which was
16 The statutes of Pistoia exempt triumphi and a game called
la diricta from the general prohibition on card games (Statuta
Pistoriensium libri septem
, Florence, 1579, Lib. V, rubrica LX,
p. 152). Whether the word triumphi refers to Tarot or to a
game with the regular pack is unclear. 1579 is very late for the
use of the word in the former sense, but the statute may be
much older than the collection; it is repeated, word for word,
in the Leges Municipales Pistoriensium nuper mandante serenissimo
Cosmo III Magno Duce, Florence, 1682, p. 210.

404 Part III: Italian Games and Italian Cards
popular there, though it originated in Florence);
with Sicily; and of course with Bologna. The
Charles VI and Catania packs were probably
made in Ferrara, but this does not help us, since
in both cases the numbering was added later,
presumably by an owner of the cards, and we do
not know their early history. The Charles VI
numbering does show, however, that type A
orders go back to the fifteenth century, and the
Rosenwald pack takes us back to the same date
(or possibly a little later, since it could be..from_
the early years of the sixteenth century). We
know by documentary evidence that Tarot cards
were in use in Bologna in 1459; of course, we do
not know what order was there observed for the
trump cards at that date, but, in view of the
intense conservatism of Bolognese players, it
seems more probable than not that it was already
of type A. Type A thus had an early origin; as
shown by the alia Colonna and Orfeo packs, it
lasted through the seventeenth century, and
survives, in the Tarocco Bolognese and the
Tarocco Siciliano, down to the present day. The
Sicilian pack provides further evidence that type
A orders were observed in Rome, since it was
probably from Rome that the 78-card pack was
introduced into Sicily in 1663, together with the
Minchiate pack. It may not have been until the
later fifteenth century that the game of Tarot
reached Florence, and it was almost certainly
from there that both Minchiate and the 78-card
game travelled to Rome, probably some time in
the course of the sixteenth century. Florence is
thus likely to have been the place of ultimate
origin of all those type A orders other than that
observed in Bologna (which, as noted, differs
from the others in placing the Chariot below the
Virtues). Although we have no direct evidence to
this effect, it is also probable that some type A
order travelled to Piedmont by at least the
seventeenth century; on no other hypothesis is it
intelligible that, after the reintroduction of the
78-card pack from France, Piedmontese players
. should have insisted on treating the Angel as
ranking higher than the World.

We have now to consider the orders of type B.
In the only other set of early popular Tarot cards
in which the trumps are numbered, the sheets at
the Metropolitan Museum (no. 21 in Chapter 4),
the order is of type B. This is the only set of early
Tarot cards that has stylistic affinities to any
surviving regular cards. There are two sheets
(1-1009 and 1-1010) in the Cary Collection at Yale
University, evidently for the same pack, no cards
being duplicated on the two sheets. In an article
about them written in 1939,iT Cary illustrated
them, together with other copies from the Magyar
Nemzeti Museum in Budapest, which I have been
told are no longer there, and of which he stated
that duplicates (which I have not seen) were sold
to the Metropolitan Museum in New York
in 1922. By comparison of Cary's sheets and
those at Budapest, it can be seen that
together they made up a complete regular pack of
48 cards (that is to say, without the 10s). The
three sheets of the Metropolitan Museum Tarot
pack include no numeral cards of any of the suits,
but they have three Kings, three Queens, all four
Cavalli and two Jacks. The Kings from the two
packs are highly similar, though not identical: in
both they wear short tunics and sit beneath
arches. The Cavallo of Batons (not quite complete
in the Budapest sheet) appears to be identical in
the two packs, a rather curious design in which
the mounted knight holds the open-mouthed
head of some animal. The Cavallo of Coins is
quite different: on the Budapest sheet, the knight
appears to be riding an ostrich, whereas, in the
Metropolitan Museum Tarot pack, he is
mounted more conventionally on a horse, and
holds one coin while another is at his horse's feet.
Not enough of the other two Cavalli on the Cary/
Budapest sheets can be seen to be sure how far
they resemble the Metropolitan Museum ones.
The Jack of Cups on the Cary sheet is identical
with that in the Metropolitan Museum Tarot
pack: he is drinking from a cup held in his left
hand, and carries a pipe in his right hand. The
Jacks of Swords in the two packs are, however,
quite unlike: that on the Cary sheet is in the act
of sheathing his sword, while the Metropolitan
Museum one bears his upright.

Other early Italian cards survive having close
affinities with these two packs. In the Benaki
Museum at Athens there is a 3 of Cups identical
with that shown on I-1010, and therefore
probably from an identical pack; it appears to
have been discovered in Egypt, and hence to be
from an Italian pack exported there during the
Mamluk period. (18) Also in the Cary Collection is
17. Melbert B. Cary, Jr., 'A stencil sheet of playing cards of
the late 15th century with two related uncut sheets of cards',
The Print Collectors' Quarterly, vol. 26, 1939, pp. 392-423.
18 See M. Dummett, 'A note on some fragments in the
Benaki Museum', Art and Archaeology Research Papers (AARP),
no. 4, December 1973, pp. 93-9.

The Order of the Tarot Trumps 405
another sheet (I-1008), discussed by Cary in his
1939 article, and again with another copy then in
Budapest and with a duplicate in the
Metropolitan Museum. This is not a sheet for the
same pack, the sizes of the cards being different,
and the sword on the Ace of Swords being held
by a lion instead of by a human hand as on
I-1010; but there are close similarities. I-1008
shows only numeral cards, of Swords and Batons,
and includes 10s of both suits. Finally, there are
four sheets, all for the same pack, in the Fournier
Museum in Vitoria (no. 1 in the Italian section of
the catalogue). Although a few cards are
damaged, they together make up a virtually
complete regular pack, again of only 48 cards.
The Kings are again highly similar both to those
on I-1009 and to those in the Metropolitan
Museum Tarot pack, without being identical
with either; the Jacks and Cavalli bear no
resemblance to those of the other two packs; the
Ace of Swords seems to be identical with that on
1-1010, and those of Batons and Cups highly

There are two distinctive features on the
numeral cards of Swords and Batons in all these
packs, save for the Metropolitan Museum Tarot
pack, of which, as remarked, no numeral cards
survive. In all three of the other packs, the
Swords, though curved (save for the odd straight
one) and mostly extending the length of the card,
are arranged so as not to intersect, being concave
towards the nearest edge of the card; on the
higher-numbered cards, one or more swords are
often placed horizontally at top or bottom. On
I-1008 and in the Fournier pack, the swords are
encircled by a crown; on I-1010, they are tied
together by a scarf. In all three packs, the
numeral cards of the Batons suit bear a scroll on
which is written in full the number of the card.
The pack in the Fournier Museum is dated 1462
in the first edition of the catalogue, but this is
presumably due to a misreading of the scroll on
the 2 of Batons, which should probably be read
duobs: on I-1008 the form used is duobs, and on
1-1009 duos. A safer dating would seem to be 1490-

The similarities between all four packs,
including the Tarot pack at the Metropolitan
Museum, imply an origin from the same locality;
and the use of identical designs for some of the
cards in different packs must surely indicate that
they all came from the same studio, one
employing a selection of alternative designs for
the court cards and Aces. In his article, Cary
proposes Venice as the place of origin of the three
sheets he is discussing. He quotes the Budapest
museum as describing their copies as Venetian,
and cites two authorities on prints, Campbell
Dodgson and Franz Schubert, as concurring with
his attribution of them to Venice; the catalogue
of the Fournier Museum also assigns the pack
there to Venice. The only reason given for this
attribution by any of these writers is, however,
one advanced by Cary himself, namely that the
form diexe used for the word 'ten' on the 10 of
Batons of I-1008 belongs to the Venetian dialect,
the modern Venetian form, in use at least by the
sixteenth century, being diese. Though linguistic
evidence of this kind is perfectly valid, it would be
pleasant to have a broader basis for the
attribution, especially as some of the linguistic
forms used seem distinctly odd (one cannot in
general expect very accurate spelling from
cardmakers). In default of any other evidence,
however, we may resonably fall in with the
prevailing opinion, and agree in regarding all
these packs as Venetian; it is very plausible that
the pack exported to Egypt, from which the
Benaki card comes, should have been made in

We may thus tentatively assign the
Metropolitan Museum pack, no. (21), to Venice,
and regard it as exemplifying, if not exactly a
standard pattern, at least a general style of Tarot
pack in use there at the end of the fifteenth
century. There is a curiously persistent tradition
in the literature on playing cards of referring to
78-card Italian Tarot packs as 'Venetian Tarots',
as distinguished from Bolognese Tarots and
'Florentine Tarots', i.e. Minchiate packs; often
the term 'Venetian Tarots' is applied even to
packs made in or after the eighteenth century
with designs derived from the Tarot de
Marseille.(19) Venice is definitely not among the
places in which the game of Tarot was played
after its reintroduction from France in the
eighteenth century, and there is no clear evidence
19 The latest to follow this tradition is Mr Stuart R.
Kaplan, op. cit. On p. 49, under the heading 'Tarocchi of'
Venice', he speaks, rather oddly, of 'the so-called Tarocchi
of Venice or Lombard pack ... more commonly known as
Piedmontese tarot', and illustrates this type on p. 48 with a
pack in the Fournier Museum made in Gorizia, captioning
it 'Piedmontese or Tarocchi of Venice Cards'. The pack is
no. 12 in the Italian section of the Fournier catalogue, where
it is also described as 'tarocchi de Venecia' and assigned the

406 Part III: Italian Games and Italian Cards
that it was ever very popular there in the fifteenth
and sixteenth centuries, so the tradition of
speaking of 'Venetian Tarots' is a thoroughly
misleading one; in the Metropolitan Museum
sheets we have the one example of a popular pack
that may plausibly be described as a Venetian

We have now established one tentative
association of the type B orders, namely with
Venice. This is corroborated by. the Rouen pack,
which also has the trumps arranged in a type B
order, and has also been associated with
Venice. (20) There remain the three literary
references. Unfortunately, I know no way of
assigning a geographical origin to the Steele MS.
volume of sermons. Tomaso Garzoni's Piazza
Universale was published in Venice, but that is not
very significant, since Venice was a great
publishing centre. He himself, though he studied
law at Ferrara and Siena before joining the
Lateran Congregation in 1566, was a native of
Bagnacavallo, near Ravenna, and also died there.
As for the Bertoni tarocchi appropriati, they relate
to the ladies of the court of Ferrara. Ferrara was
never within the Venetian dominions, though it
was very close to their border; but the Sola-Busca
tarocchi, made by a Ferrarese artist but dated
from the foundation of the city of Venice, testify
to a link between the traditions of Tarot play in
the two cities. Ferrara was, of course, one of the
principal centres where the game of Tarot was
played in the early period, and very likely the
birthplace of the game; the Bertoni poem
provides incontrovertible evidence that the type
B order prevailed there. It seems quite possible,
therefore, that the type of design exemplified by
the Metropolitan Museum sheets, which we have
assigned to Venice, was equally characteristic of
Ferrara; indeed, we should keep open the
possibility that those sheets are not from Venice
but from Ferrara. In any case, it seems safe to
assign type B orders to Venice as well as to
date 1650; Kaplan more cautiously says 'circa late 17th to
mid-18th century'. It has inscriptions in French, and
represents that adaptation of the Tarot de Marseille used in
Lombardy, but made in a wide range of areas, different
from the adaptation characteristic of Piedmont; it can
hardly be earlier than 1740.
20 See Pompeo Molmenti, La Storia di Venezia nella vita
, vol. II, Bergamo, 1906, p. 525. Francesco Novati, on
p. 19, fn. 1, of the first of his articles cited in footnote 22,
expresses the same opinion. How well founded it is, I am not

Ferrara, and they probably prevailed in the
whole of Emilia except for the city of_
Bologna. On the evidence of the Steele sermon,
type B orders go back to at least about 1470-80;
since they are associated with Ferrara, they quite
possibly go back to the original invention of the
Tarot pack. On the evidence of Garzoni's book,
they lasted until at least the late sixteenth

That leaves us with the type C order and the
Cary tarocchi sheet,- no. (24) of Chapter 4. Susio's
poem, which is our only source for type C orders
in Italy, concerned the ladies of the court of
Pavia. (21) This city was in the dominions of both
the Visconti and Sforza dukes of Milan, who
styled themselves Princes of Pavia and for whom
it was a second capital; it contains a great
Visconti castle, begun by Galeazzo II and
completed by Giangaleazzo Visconti, who also
founded the Carthusian monastery near Pavia
and began the building of Pavia's Cathedral, and
was the first to bear the title of Duke of Milan.
We may therefore reasonably assume that it was a
type C order which prevailed not only at Pavia
but at Milan, the second great early centre, after
Ferrara, for the game of Tarot. Now despite the
variations that occurred, all the trump orderings
used by French and Swiss cardmakers were of
type C. If we had no evidence that an order of
this type was ever used in Italy before the
eighteenth century, we should most naturally
infer that it was invented in France or
Switzerland. But, as it is, we are forced to
conclude that either the French or the Swiss, or
both independently, picked it up from Italy. Even
if we did not know of Viévil's pack, the
occurrence in both Susio's trump order and in
the Tarot de Marseille of the intrinsically rather
implausible placing of Temperance between
Death and the Devil would seem unlikely to be a
coincidence. Indeed, this is another case in which
a conjecture made in the original version of this
book received additional confirmation while it.
was in proof: for the almost exact agreement
between Vi6vil's trump order and that given by
21 In my earlier discussion of this subject, 'The Order of
the Tarot Trumps', Journal of the Playing-Card Society, vol. II,
no. 3, February 1974, pp. 1-17, no. 4, May 1974, pp. 33-50,I
made the mistake, for which I cannot now account, of
saying that the Susio poem was about the ladies of the court
of Mantua. S.R. Kaplan, op. cit., pp. 30, 373, also cites
Susio's poem, though he gives no reference, and also makes
the same error; possibly mine was the source of his.

The Order of the Tarot Trumps 407
Susio, presumably used in Milan in the sixteenth
century, ^converts a plausible hypothesis into a
certainty. The modification of the Milanese order
which resulted in that used in the Tarot de
Marseille had obviously occurred by the mid-
sixteenth century, without gaining universal
acceptance in France for another hundred years:
whether it originated in France, in Switzerland
or even in Milan it is hard to judge.

If the French and Swiss did pick up the type C
order in Italy, by far the likeliest place for them
to have done so is Milan. Charles VIII of France
invaded Italy in 1494, originally on the invitation
of Lodovico Sforza (il Moro), Duke of Milan.
Louis XII, the grandson of Valentina Visconti,
claimed Milan by right of succession, and
launched a second invasion in 1499; the city was
then under French rule until 1512, when Louis
was defeated at Novara by an alliance which
included the Swiss. Up to 1515, when Francis I
secured their exclusive services, Swiss
mercenaries played a prominent role in these
wars; and from 1512 to 1515, Duke Massimiliano
Sforza was maintained in power by Swiss arms.
In 1515 the French, under Francis I, again
invaded Italy, defeated the Swiss at Marignano,
and once more occupied Milan until 1522.
Especially during the reign of Francis I (1515-
1547), there was a great vogue in France,
centring upon Lyons, for Italian culture. The
period of the French incursions into Italy, from
1494 to 1525, may therefore well have been the
time when the game of Tarot first entered
France. It may have reached Switzerland
independently in the same period, for, as we saw
in Chapter 10, certain details of Swiss Tarot play
suggest a direct derivation of the game from
Italy. If this is true, of France alone or of both
France and Switzerland,, it must surely have been
a Milanese style of Tarot game, and a Milanese
version of the Tarot pack, that were adopted;
with them would naturally go the order of the
trumps observed in Milan. If the game of triumphe
played by Duke Rene II of Lorraine in 1496 were
truly one played with the Tarot pack, his
knowledge of the game may have been due to his
contact with the Swiss, with whose help he had
achieved his great victory in 1477 outside the
walls of Nancy against Charles the Bold, Duke of
Burgundy, who was killed in the battle. Even if
this is so, the adoption in the sixteenth century of
the French version tarots of the new term tarocchi
argues a continued contact with Italy in the
domain of card play as in other spheres.

On this theory, then, a type C trump order,
and, specifically, that given by Susio, was in use
in Milan; and it was from Milan that the French,
and probably the Swiss, first learned the game of
Tarot. They must, therefore, originally have used
whatever type of design was standard in Milan:
but we cannot immediately deduce what this
was, since, as we saw in Chapter 9, there were
two distinct traditions of design for the Tarot
pack in the French-speaking lands, one culminating
in the Belgian Tarot and the other in the
Tarot de Marseille. We have, however, still
to determine the geographical origin of the
sheet in the Cary collection (no. 24). After
what has gone before, this affords us very
little difficulty, because of the close_resemblance
of certain of the cards, despite their lack of
numerals or other inscriptions, to those of
the Tarot de Marseille. Specifically, the Sun
(of which the left-hand part is missing) is, so
far as can be seen, exactly like its Tarot de
Marseille counterpart; a small naked boy is to be
seen at the bottom of the card, the sun has a face
and rays and sheds the characteristic Tarot de
Marseille droplets. The Chariot, incomplete at
the top, is likewise, so far as can be seen, exactly
like that of the Tarot de Marseille. The Moon
resembles the Tarot de Marseille one, save that
there are no dogs, and the buildings are much
smaller: there is the same pool with a lobster or
crab in it in the foreground. The Tower again has
a close similarity to the Tarot de Marseille
Maison Dieu: round thunderbolts are falling
about a round, bricked tower, though no
lightning is apparent. The Star resembles the
Tarot de Marseille one in general conception,
though not in detail: a very large star,
surrounded by four smaller ones, shines on a
naked girl pouring water into a stream. The
Emperor has the same general pose and
appearance as in the Tarot de Marseille, though
the positions of his shield and sceptre are
different. The Bagatto has the same posture as
the Tarot de Marseille one, though he does not
face in the same direction, and his hat and table
are differently shaped. The Love card, of which
only the lower half survives, is particularly
interesting. Only two figures can be seen,
corresponding perhaps to the left-hand and
central figures of the Tarot de Marseille card:
but they are highly reminiscent of the couple on
the corresponding card in the Visconti di

408 Part III: Italian Games and Italian Cards
Modrone pack (no. 1 in Chapter.4). The Wheel
of Fortune, another card whose top half is
missing, has the same orientation, and the same
handle, as in the Tarot de Marseille; and the
Empress, though again with shield and sceptre
reversed, has the same chair-back that, on several
cards, keeps threatening, in Tarot de Marseille
derivatives, to turn into a pair of wings. Not
much can be seen of the Fool, save that, as in the
Tarot de Marseille, he is striding off to the right,
a staff over his shoulder; no dog is to be seen,-
however. The two suit cards, the 7 and 8 (or 9) of
Batons, are exactly like those of the Tarot de
Marseille, save for the lack of inscribed
numerals; the Batons have just the same flat
appearance, with widened ends, found both in
the Tarot de Marseille and in Viévil's pack.
Other cards, however, have little or no similarity
to their Tarot de Marseille counterparts:
Temperance, the Devil, Fortitude and the Popess
(on this sheet, a Bishopess); of the Pope, not
enough can be seen to be sure.

These resemblances cannot possibly be
coincidence: it is evident that the French
cardmakers borrowed these designs, which thus
became ancestral to the Tarot de Marseille
pattern. It follows that the origin of the Tarot de
Marseille goes right back to the first introduction
of the Tarot pack into France, and perhaps also
into Switzerland, around the beginning of the
sixteenth century. It also follows that we can
firmly identify the Cary sheet as being from
Milan, and hence as exemplifying the standard
pattern employed in the late fifteenth or early
sixteenth century for popular Tarot packs in the
city from which came the finest of the early hand-
painted cards; that city where, as we have
argued, the French and probably also the Swiss
first encountered the game of Tarot and took it
back with them to their home countries. This
hypothesis fits well with the fact that the trumps
on the Cary sheet are unnumbered. It was stated
earlier that, of all the literary references, only the
Steele sermon assigns numbers to the trumps,
but this was not strictly accurate: the lines of the
Susio poem are also numbered, but in the reverse
order, the World being numbered 1, the Angel 2
and so on. This is the most explicit testimony
possible to the fact that the trumps of the pack in
which this type C order was used were not
numbered: if they had been, it is inconceivable
that the poem could have given the numbering in
the reverse order. In whichever form of pack
the ordering of the trumps given by Susio was
employed, the players accustomed to pjajy with it
must have had to memorise the trump sequence,
just as the Bolognese ones had to do. Because of
the affinities which both the trump order and the
designs of the Cary sheet have to the cards used
in France, it is highly probable that the kind of
pack for which the type C order was used was
that of which the Cary sheet is an instance.

These conclusions are corroborated by the set
-of six-Tarot- cards (no. 29 above) found at the
Castello Sforzesco in Milan during the
restoration work of 1908, and by others found
there at the same time; again, these were not
known to me in detail until this book was in
proof, and provided strong confirmation of the
hypotheses set out above concerning Milan. The
cards found at the Castello were cursorily
described by Francesco Novati in two articles of
1908.22 His datings tend to be uniformly too
early. He considered all the cards to date from
the fifteenth, sixteenth or seventeenth century,
whereas some demonstrably exemplify the
eighteenth-century Lombard variation on the
Tarot de Marseille. One such example is a set of
five cards (Cavallo of Swords, torn at the bottom,
5 and 7 of Coins, and 6 and 8 of Swords, the last
four all having Roman numerals at the sides)
made by the cardmaker who used the tradename
'Al Soldato' and operated in Bologna
during the eighteenth century. Another is a
Cavallo of Batons, bearing a legible French
inscription at the bottom. There is also a set of
three cards, which appear to have been trimmed
at top and bottom, and are probably from a
French Tarot de Marseille pack of the
Revolutionary period: they consist of the Sun,
numbered XVIIII, the lower inscription having
been trimmed off, the Cavalier of Cups, of which
the same is true, and the 4 of Swords, with
Roman numerals at the sides" There are also a 10
of Coins, with the trade-name 'Al Leone' on the
back, and a 5 of Cups, with the trade-name 'Al
Mondo'; both these names signify other
Bolognese cardmakers of the eighteenth century.
However, besides the set of six tarocchi
classified as no. (29) above, there are also other
22 Francesco Novati, 'Carte da giuoco dei secoli XV, XVI
e XVII rinvenute nel Castello Sforzesco', Bullettino dei civici
musei artistico ed archeologico di Milano
, anno III, num. 3, 1908,
pp. 17-20, and 'Per la storia delle carte da giuoco in Italia:
appunti', Il Libro e la Stampa, anno II (n.s.), 1908, pp. 54-69;
see pp. 65ff.

The Order of the Tarot Trumps 409

cards evidently dating from before the eighteenth
century, none-of-which we have any reason to
regard as being from a Tarot pack. Just as in
Bologna the same standard pattern was used for
the regular Primiera pack and for the suit cards
of the Tarocco pack, so the same may well have
been true in Milan. One of these earlier cards is a
5 of Batons, exhibiting the usual fiat shape, with
widened ends, of Tarot de Marseille Batons, but
without numerals, but only floral decoration, at
the sides. Another is a much damaged Cavallo
(of Swords?), not coinciding in design with any
known standard Tarot pattern. This may come
from the same pack as the 5 of Batons, but from
both the back design has peeled off; they may be
of the sixteenth or seventeenth century. A set of
nineteen cards, some very fragmentary, all of the
same size (102 x 69 mm.) and presumably by the
same maker, though not from the same pack, is
assigned by Novati to sixteenth-century Venice.
The back of each card depicts a classical deity,
whose name is inscribed on a scroll; nine
different deities appear on different cards. Each
back design is surrounded by a wide border with
lozenge-shaped dots, which fold over to form a
border for the face of the card.23 Of these,
fourteen are numeral cards, of all four suits; in
every case, they tally precisely in design with the
corresponding cards of the Tarot de Marseille,
save that they lack Roman numerals at the sides.
The remaining five, all fragmentary, are court
cards. They are: the top half of a King (of
Batons?), the back showing SATURNO; the top
half of a Jack (of Coins ?) and the bottom half of a
Jack (of Cups?), both backs showing
PROSERPINA; and the top half of a Cavallo of
Cups and the bottom half of a Jack of uncertain
suit, both backs showing IOVE. These court
cards do not show the same close correspondence
with Tarot de Marseille designs. In particular,
the Cavallo of Cups holds in his left hand a Cup
shaped like that on the Tarot de Marseille Jack of
that suit, instead of the Spanish-style Cup held in
the right hand of the Cavalier in the Tarot de
Marseille; while the King, who is bearded, holds,
with his left hand, a Baton (or sceptre) over his
left shoulder. Nevertheless, the Jack of which we
have the top half wears the celebrated widebrimmed
hat found in the Tarot de Marseille on
23 Compare the backs of seventeenth-century Italian cards
from the Correr Museum, Venice, shown by Hoffmann, op.
cit., as plate 7b.

the Jacks of Coins and Swords and the Cavalier
of Batons; and the King likewise wears the same
hat surmounted by a crown, as do all four Kings
in the Tarot de Marseille. These nineteen cards,
from nine distinct packs, could be either of the
sixteenth or of the seventeenth century. There
seems no reason, however, to regard them as
Venetian rather than Milanese; Swords and
Batons on the numeral cards of the Venetian
standard pattern, and on older cards assignable
to Venice, have a different shape from those
found here.

One of the oldest cards found at the Castello is
a King of Cups of which, again, the top half is
missing. The posture of the King resembles that
of the King of Coins in the supposedly Spanish
fifteenth-century pack which, in Chapter 2, we
tentatively assigned to Naples, and, more
generally, of several early German Kings; it may
be of the early sixteenth or even of the fifteenth
century. The most interesting of all the cards
found at the Castello is a 2 of Coins which bears
the scroll in the shape of an inverted S which, in
Chapter 9, we noted as always occurring
(sometimes not inverted) in Tarot de Marseille
packs, and also in the pack of Jacques Viévil. As
on the French cards, the scroll is inscribed with
the maker's name and the date; the inscription
1499. Novati cites documentary evidence of the
presence in Milan in 1508 and 1513 of a
cardmaker by the name of Paolino di Castelletto.
There is no reason to regard Paolino's 2 of Coins,
any more than the other cards just discussed, as
having belonged to a Tarot pack: it nevertheless
provides good evidence that a distinctive feature
of French and Swiss Tarot card design was
borrowed from Milan. The S-shaped feature is
found to this day on the 2 of Coins in certain
standard patterns for the Italian-suited regular
pack, the Primiera Bolognese and Brescia
patterns; in these however, it no longer has the
character of a scroll; and does not bear the
maker's name.

We have no pre-eighteenth-century Milanese
example of the 2 of Cups, which is so distinctive
of French Tarot packs, including Viévil's and de
Hautot's as well as the Tarot de Marseille. But
the curious and prominent feature of the two
dragon-heads on the French versions of the card
provide a particular reason for considering the
design to have been derived from some very early
prototype. As first remarked by Mr Jan

410 Part III: Italian Games and Italian Cards
Bauwens, these dragon-heads have an
extraordinary similarity to the dragon-heads in
which certain of the suit-signs terminate on some
cards of the Polo-Sticks suit in the Istanbul
Mamluk pack.24 It looks as if we had here a
detail, faithfully copied for centuries, that had
originally been borrowed from Islamic cards; if
so, the design of at least that particular card must
go back to a period when Islamic cards were
familiar to European players or cardmakers.
This, too, fits very well with our conclusion thatthe
designs on which the Tarot de Marseille was
based were ones used in Milan, the birthplace of
that Valentina Visconti whose inventory, upon
her death in 1408, had listed ung jeu de quartes
along with unes quartes de Lombardie. (25)

One's first impression, looking at the various
cards found at the Castello Sforzesco together, is
of their uniformity of style. It would be a great
mistake to suspect them for this reason of all
dating from after 1700. The 2 of Coins by Paolino
proves incontrovertibly that some are much older
than that; and there is a clear criterion of
24 Jan Bauwens, Muluk wa Nuwwab, Aurelia Books,
Leuven (Louvain), 1972, booklet issued with a reproduction
of the Istanbul Mamluk pack, pp. 36-7, figs. 8-11. (It should
be noted that the reproduction pack is not a faithful copy of
the original, but involves a good deal of 'reconstruction', to a
large extent unsound: see the review by me in the Journal of
the Playing-Card Society
, vol. II, no. 2, November 1973, pp. 15-
26. This does not, of course, affect the present point, which
is a very interesting one.)
25 F.M. Graves, Deux inventaires de la Maison d'Orleans, Paris,
1926, pp. 49, 134; see Chapter 3. On p. 84 it was strongly
argued that the references to triumphe of 1482 and 1496 must
relate to Tarot games. Such a hypothesis would contradict
the idea that Tarot first entered France during the French
occupation of Milan, since, as explained in footnote 2 to
Chapter 9, the 1482 reference must concern a game played in
France proper. But there is no difficulty in supposing the
game to have spread to different parts of France at different
times and by different routes; possibly it was the non-
Milanese ancestor of the Paris/Rouen pattern that was the
earlier arrival, and would have been used in 1482. If the game
of 1482 was that played with the regular pack, the only
reasonable hypothesis is that it also had an Italian origin, in
which case an Italian game, involving trumps but played
with the regular pack and known as trionfi, must have been in
existence well before the end of the fifteenth century, and
even longer before the earliest recorded use, in 1516, of the
term tarocchi; if so, the reference to triumphi cited by W.L.
Schreiber from the statutes of Bergamo, Brescia, Salo and
Reggio nell'Emilia may not have been to Tarot games. On
balance, this does not seem to me very likely; but without
doubt the 1482 reference generates perplexity. An
examination of the document at the Archives Nationales
might yield further clues.

distinction between post- and pre-eighteenth-
century cards, the presence, or absence, of
numerals on the numeral cards of the suits, and
of names on the trumps and court cards. By this
criterion, the set of six tarocchi is to be assigned to
the earlier category, probably to the seventeenth
century. What the cards found at the Castello
together demonstrate is the absolute constancy of
the Milanese designs for the numeral cards of the
Italian-suited pack, and the equal constancy of
the-same designs as borrowed by the French
cardmakers and employed in the Tarot de
Marseille. If you ignore the inscription on
Paolino's 2 of Coins, you might think that you
were looking at an eighteenth-century card made
in Marseilles. In the same way, apart from the
absence of the Roman numerals at the sides, the
numeral cards from the set of six tarocchi, those
with classical deities on the backs, and the odd 5
of Batons are virtually indistinguishable from the
eighteenth-century Lombard pattern. This may,
at first sight, raise a doubt whether our original
hypothesis, stated in Chapter 8, was after all
correct, namely that the Lombard variant on the
Tarot de Marseille signalised the reintroduction
of the game of Tarot into Lombardy after a
period in which it had been defunct there.
Perhaps, we may now think, the Lombard
pattern was a direct continuation of the standard
pattern always used in Lombardy for Tarot
cards, and for Italian-suited ones generally, and
represented no more radical a change than a new
vogue for putting names on the trump and court
cards and numerals on the numeral cards. But
reflection shows that we have no reason to doubt
our original hypothesis. Whether that hypothesis
is sound or unsound, the cards found at the
Castello prove conclusively that the Tarot de
Marseille designs for the numeral cards had
faithfully preserved the Milanese prototypes on
which they must have been modelled in the early
sixteenth century, and that these same designs
remained unchanged in Milan itself. Given this, a
close resemblance between those made in pre-
eighteenth-century Milan and those of the later
Lombard pattern is precisely what we should
expect, even if our hypothesis is true; that we find
just such a resemblance is therefore no argument
against that hypothesis. Our original ground for
the hypothesis remains as suasive as ever,
namely that, if inscriptions had been added to an
existing pattern for a pack used to play a living
game,' they would have been in Italian; the fact

The Order of the Tarot Trumps 411
that, for several decades, they remained in
French clearly indicates that the Tarot de
Marseille designs were being introduced from
France; and that can have happened only if the
indigenous tradition had, however little time
before, died out. What the Cary sheet and the
Castello Sforzesco cards make clear is that, in
adopting a form of pack derived from the Tarot
de Marseille, the players of tarocchi in Lombardy
were welcoming home a descendant of the type of
pack with which their ancestors had played, and
a close relative of that used by their fathers.
How close a relative? The World card in the
Castello Sforzesco set suggests that, as on the
Cary sheet, the trump cards were without
(visible) numerals, just as were those of the
Tarocco Bolognese at that date; but the
conclusion is uncertain, since there could have
been numerals on all the trumps except the
highest, as in the Metropolitan Museum pack. In
any case, it is a presumption, though not a
certainty, that the trump order was as given by
Susio rather than as in the Tarot de Marseille.
Indeed, were it not for the Cary sheet, we might
suspect Viévil's pack, rather than the Tarot de
Marseille, to preserve the Milanese tradition of
design; for it will be recalled that Viévil's World
resembles that of the Tarot de Marseille, and
hence also the Castello Sforzesco card, much
more closely than it does that of de Hautot or of
the Belgian Tarot. The few pre-eighteenthcentury
court cards from the Castello Sforzesco
show that the Milanese designs of the time were
far from wholly identical with those of the Tarot
de Marseille. The French cardmakers introduced
Batons of Spanish type for the Cavalier and Jack
of that suit, and a straight-sided Cup of Spanish
type for the Jack of Cups. Probably they
departed in many other respects as well from the
Milanese prototypes, and the same may easily be
true for the trump cards. But, at the same time,
the wide-brimmed hats found in two of the court
cards of the classical deities set corroborate our
general conclusion that the Tarot de Marseille
was of Milanese origin.

We have, then, a surprising result: the Tarot
pack entered France with a pattern of design
ancestral to the Tarot de Marseille but with a
trump order almost identical with that of Viévil's
pack. Viévil's trump order was thus not, in the
first place at least, especially associated with the
Paris/Rouen standard pattern. Rather, it must
have been a survival from an earlier epoch:
perhaps the modified order which is found in
Geoffroy's pack of 1557, in the anonymous
Parisian pack and in the Tarot de Marseille
existed alongside the original Milanese one for a
considerable time. What, then, can have been the
origin of the Paris/Rouen pattern which finished
its career as the Tarot pattern proper to
Belgium? It has certain particular affinities with
Italian Tarot cards. The man with the compasses
found on the Star appears on the Moon card in
the Minchiate pack and the Tarocco Bolognese,
and also in the hand-painted 'Charles VI' set
(no. 4 in Chapter 4); and the woman with the
distaff on the Moon card appears in the Tarocco
Bolognese and the 'Charles VI' set on the Sun.
There is also a resemblance between the World
in de Hautot's pack and in the Belgian Tarotand
the World in both the Minchiate pack and the
Tarocco Bolognese, and also in the 'Charles VI'
set and the Catania set (no. 7 in Chapter 4). We
cannot be sure how significant this last point of
affinity may be, since the Rouen and Belgian
World seems to have been copied from the
anonymous Parisian pack, Viévil's version being
quite different. It is unlikely that any French or
Belgian cardmaker would know anything about
Florentine, Bolognese or Ferrarese cards. The
likeliest hypothesis seems to be as follows. It has
been argued that a Tarot pack with a type A
order for the trumps, such as prevailed in
Bologna and Florence, must, before the
eighteenth century, have been in use in
Piedmont. The game presumably spread there
from somewhere like Bologna, Florence or Rome,
just as the game of Minchiate had by the
seventeenth century spread from Florence to
Genoa. The designs for this former Piedmontese
pack may, then, have become the original of the
Paris/Rouen pattern; they could easily have
passed via Savoy into France, existing French
tradition being too strong to allow this type of
Tarot pack to retain its type A trump order.
(Possibly-the employment of a shortened 62-card
pack at Chambery points to Bologna as the most
likely place of origin.) On this hypothesis,
therefore, the ancestor of the Belgian Tarot was
this lost Piedmontese pattern. A small piece of
evidence in its favour is the preservation in Savoy
into the 1900s of the form Baga employed by
Viévil to name the trump I.

The Playing-Card Society is engaged upon a
definitive classification of all standard patterns
that can be indubitably recognised as such, and

412 Part III: Italian Games and Italian Cards
for this purpose assigns numbers to each pattern:
thus the Tarot de Marseille is IT-1, the Tarocco
Bolognese IT-2 and the Belgian Tarot IT-3, the
letter I indicating an Italian-suited pattern and
the letter T a Tarot pack.(26) Numbers with more
than one digit indicate derivatives from the
pattern whose number is found by deleting the
last digit: thus the Tarot de Besançon, regarded
as a modification of the Tarot de Marseille, is IT-
1 -4, and the romanticised nineteenth-century
version, still used in Switzerland, that is its only
modern descendant is IT-1-41. The Tarot de
Marseille has lasted down to the present day,
though now mostly used for fortune-telling. The
Tarot de Besançon, which was used, though not
of course exclusively, almost everywhere in
Europe except Belgium and Italy, died out, save
for the IT-1-41 form, in the nineteenth century. (27)
Of the Italian derivatives from the Tarot de
Marseille, the earliest form of the Tarocco
Piemontese receives the number IT-1-2. It was
subsequently modified, making it less precisely
similar to the Tarot de Marseille, and replacing
the French inscriptions by Italian ones; this
second stage is labelled IT-1-21. A further
modification resulted in the double-headed
version of the Tarocco Piemontese used today,
labelled IT-1 -211. The Lombard pattern,
whether with French or Italian inscriptions, is
designated IT-1-1. In the early nineteenth
century it was replaced by a romanticised
version, sufficiently different to merit a distinct
numeral after the decimal point, and thus
designated IT-1-3, subsequently succeeded by a
slightly modified double-headed version, IT-
1-31. The IT-1-31 tradition, sometimes
inaccurately called the Tarocchino Milanese, in
its turn died out, leaving the modern Tarocco
Piemontese, IT-211, as the sole form of 78-card
pack now used in Italy. (The term tarocchino has
26 The Society issues a four-page sheet for each standard
pattern, illustrating characteristic cards, giving its history
and listing prominent makers past and present.
27 The Playing-Card Society sheet on IT-1-4 says that,
although 'it was not until about 1800 that any quantity of
these cards were made in Besançon, ... the pattern can
barely have survived the early part of the 19th century,
being replaced in most areas by the French-suited Tarot
packs'. This latter remark is certainly true of Germany and
the Austrian Empire; but, as reported in Chapter 15, a
booklet on the game first published in Besançon in 1880 still
describes it as played with an Italian-suited pack; so the
Tarot de Besançon probably survived in Besançon itself
until the end of the century or later.

been applied to IT-1-3 because of the small size
of the cards; properly it-should relate only to a
reduced number of cards in the pack.)

What, then, of the subsequent history of our
conjectural four early Italian standard patterns?
The Bolognese one presents no problems. We
know the stages through which it went: the
replacement of the old design for the Devil by a
new one, say around 1600; the substitution of the
Moors for the Papi in 1725; the introduction of
numerals on the trumps and the change to a
double-headed form in the later eighteenth
century. The Metropolitan Museum pattern,
which we have taken to characterise Venice and
perhaps Ferrara, has a single possible surviving
later exemplar, the isolated Love card in the
Museo Nazionale delle Arti e Tradizioni
Poplari in Rome (no. 28 above). The design is
more complex than that of the corresponding
card on the Metropolitan Museum sheet, of
which only the top half remains. The latter shows
only a Cupid in the air aiming his bow, with a
man below on the left and a girl on the right. The
later card has an additional putto in the air, and,
below, an embracing couple on the left and, in
front of them, two musicians, one in a feathered
hat playing a viol da gamba and the other
playing a lute. This could, however, be seen as a
development of the earlier design: what makes it
probable that it represents a descendant of the
same pattern, or comes at least from a pack with
a type B order, is its being numbered VIII, like
the Metropolitan Museum card (as it would
presumably also have been numbered in the
Rouen pack and in any with the Garzoni/Bertoni
version of the type B order). The Museo
Nazionale card is cited by the anonymous editors
of Antiche Carte de Tarocchi as Venetian, of the late
sixteenth century; it. may be as late as the
seventeenth century. By Garzoni's testimony, the
type B order was still known in the 1580s. It
seems likely, however, that the game of Tarot
suffered a general loss of popularity, in Venice
and Ferrara, during the sixteenth century. The
verse diatribe by Lollio, a Ferrarese author
whose poem was published in a collection
printed in Venice, was obviously less than half
serious, like Berni's earlier derisory remarks; but
it evidently indicates a decline in the esteem with
which the game was regarded in the midsixteenth
century. Probably we shall not be far
wrong if we see the Venice/Ferrara pattern, and,
with it, the type B order, as having died out not

The Order of the Tarot Trumps 413
long after 1600.

From the set no. (29) of tarocchi found at the
Castello Sforzesco, it seems likely that the
Milanese pattern, of which the Cary sheet is the
earliest example, survived into the seventeenth
century, although the dating of that set is far
from certain. The reintroduction in about 1740 of
the 78-card pack in its IT-1-1 form represented a
revival of an ancient tradition; but it is difficult to
guess how long the interval had been during
which Tarocco had no longer been played. As for
the ancient Piedmontese pattern whose
existence, as the ancestor of the Paris/Rouen
pattern, we have conjectured, that can hardly
have died out much more than thirty years before
the introduction of IT-1-2, if there were to be
players who still remembered that the Angelo
used to be superior to the Mondo.

The Florentine pattern represented by the
Rosenwald sheets had, by contrast, a very
eventful history. At some time before the
invention of the Minchiate pack in the first half
of the sixteenth century, the pattern must have
assumed a partly Portuguese type of suit-system,
by changing the shape of the Swords from curved
to straight. It is, presumably, this type of 78-card
pack that is represented in seventeenth-century
Florence by the Orfeo packs. Save for the
Marchese di Villabianca, there is no known
literary reference alluding to any type of Tarot
game other than Minchiate played in Florence or
in Rome during the sixteenth and seventeenth
centuries. Yet the Orfeo and alia Colonna packs
testify to the continued existence in both cities of
the 78-card pack. On Villabianca's testimony,
the 78-card pack was introduced into Sicily by
the Viceroy in 1663. Since the Minchiate pack
was, according to him, introduced at the same
time, it must have been in Florence or in Rome
that the Viceroy had become acquainted with
these games, and he is much more likely to have
visited Rome than Florence. This is confirmed by
the fact that the 78-card pack introduced into
Sicily must have had a type A order, and,
moreover, as we have seen, probably had
'Portuguese' suit-signs. From the alia Colonna
pack, it is apparent that in Rome the Florentine
type of Tarot pack had evolved into one using a
fully-fledged 'Portuguese' suit-system; and it
must have been a pack of this kind which was
ancestral to the Tarocco Siciliano. We know, in
broad outline, the later history of the Tarot pack
in Sicily. The Minchiate pack died out in Sicily
during the eighteenth century, but continued to
flourish in Florence until almost the end of the
nineteenth, and, in Genoa, until the 1930s. As for
the Portuguese or Italo-Portuguese versions of
the 78-card pack, descended from the Florentine
pattern, and used in Rome and Florence, we must
suppose them to have become obsolete some time
in the second half of the seventeenth century.

Though we have been forced to rely on a good
deal of conjecture, we have been able with its
help to reconstruct in outline the entire history of
the Latin-suited Tarot pack, which can be
illustrated by a diagram. For this diagram, the
Playing-Card Society numbers for standard
patterns have been supplemented by some
additional ones. A zero after the decimal point
may be used to indicate an ancestor of a singledigit
pattern: thus the postulated Milanese
pattern represented by the Cary sheet, ancestral
to the Tarot de Marseille, is designated IT-10.
The original and the later Minchiate patterns
are designated IPT-1 and IPT-11 respectively in
the P.C.S. system, the letters IP, which refer to the
suit-system, standing for 'Italo-Portuguese', while
IPT-2 is used for the early version of the Tarocco
Sicilian/), IPT-2-1 representing the later form
(from the Fortuna pack down to Modiano). It
seems better to alter these numbers so as to be
able to indicate the relationship with patterns for
78-card packs. Thus the Orfeo packs may be
designated IPT-5, the Florentine pattern
represented by the Rosenwald sheets IT-5-0, and
the Minchiate patterns redesignated IPM-5 and
IPM-5-1. It seems better to indicate a fully-fledged
Portuguese pattern by the single letter P: if the
two stages of the Tarocco Siciliano are then
redesignated PT-6 and PT-61, we can use
PT-6-0 for the alia Colonna pattern. The
Rothschild/Beaux Arts sheets agree with later
Tarocco Bolognese designs closely enough to
justify the use for them of the straight designation
IT-2, while the Venetian pattern represented by
the Metropolitan Museum sheets can be called
IT-4. The pack made by de Hautot in Rouen
obviously exemplifies just the same pattern as the
later Belgian Tarot, IT-3 in the P.-C.S.
numbering; although Viévil's pack has many
differences, in particular having a different trump
order and including the Pope and Popess, we may
use the same number for it, keeping IT-3-0 for the
conjectural Piedmontese ancestor of this pattern.
We thus arrive at the following code, where an
asterisk denotes a number not used by the

414 Part III: Italian Games and Italian Cards

IT-1-0*: Milanese pattern (Cary sheet and Castello Sforzesco set)
IT-1: Tarot de Marseille
IT-1-1: Lombard version of the Tarot de Marseille
IT-1-2: Tarocco Piemontese (early form)
IT-1-21: Tarocco Piemontese (intermediate form)
IT-1-211: Tarocco Piemontese (modern doubleheaded form)
IT-1-3: romanticised Milanese version of IT-1-1
IT-1-31: doubleheaded version of IT-1-3
IT-1-4: Tarot de Besançon
IT-1-41: romanticised Swiss version of IT-1 -4
IT-2: Tarocco Bolognese
IT-3-0*: conjectural early Piedmontese pattern
IT-3: Paris/Rouen pattern and Belgian Tarot
IT-4*: Venetian/Ferrarese pattern (Metropolitan Museum sheets)
IT-5-0*: early Florentine pattern (Rosenwald sheets)
IPT-5*: later Florentine pattern (Orfeo packs)
IPM-5*: original Minchiate pattern (P.-C.S. number IPT-1)
IPM-5-1*: later Minchiate pattern (P.-C.S. number IPT-1-1)
PT-6-0*: Roman pattern (alia Colonna sheets)
PT-6*: Tarocco Siciliano, early form (P.-C. S. number IPT-2)
PT-61*: Tarocco Siciliano, later form (P.-C. S. number IPT-2-1)

The diagram is principally concerned to show
where the different patterns were used, rather
than where they were made; thus, although
Bologna cardmakers produced both Minchiate
cards and Lombard pattern packs, neither is
shown under Bologna because neither was used
there. A solid line indicates that cards of the
given type and made in or for the given area
survive to us from the given period; a short line
indicates a single surviving set, a longer one
the existence of packs sufficiently close in date
to warrant a presumption of continuous
manufacture and employment. In some cases, of
course, the dates are only approximate. A dotted
line indicates the conjectural use of the given type
of Tarot pack; the grounds for such conjectures
vary in strength, and have been set out in this
chapter. Where a given pattern has travelled
from one area to another, this is indicated by a
nearly horizontal line with an arrow; where one
pattern has developed out of an earlier one which
then continued to co-exist with it, this has been
shown by a nearly horizontal line without an
arrow, whether the new pattern was used in the
same area or another; where a pattern was
replaced by a new one that had developed from
it, this is shown by a small horizontal bar across
the vertical line. French-suited Tarot patterns
are not shown. The part of the diagram from
1700 on is not open to doubt; that before 1700 is
highly conjectural. Further research may yield a
different picture. The diagram will be found on
the end-papers.

It may be asked which-of the trump orders was
the original one, that which was intended when
the Tarot pack was first devised. If we knew
nothing about the geographical associations of
the different orders, it would be natural to guess
that a. type A order was the original one, for
several reasons. First, it seems more intelligible
that the Angel, which undoubtedly represents
the Last Judgment, should be placed at the end
of the sequence than in the penultimate or ante-
penultimate position. Secondly, save for the
association of Justice with the Judgment, it seems
difficult to discern any appropriateness in the
scattering of the Virtues through the sequence in
orders of types.B and C. And thirdly, if a type A
order were the original one, it would be possible
to explain the invention of types B and C as
devices for bringing the Death card to position 13
once the practice of numbering the trump cards
had been introduced. The evidence of
geographical association shows this guess
unlikely to be right. Of the two possible
claimants for the birthplace of the Tarot pack,
Ferrara certainly observed a type B order and
Milan very probably a type C order;
furthermore, we saw that it is likely that, in Italy,
the trumps in packs with a type C order were not
numbered, so that we cannot explain the genesis
of this order in the way suggested. In the present
state of knowledge, the question concerning the
priority of the three types of order does not seem
to be a fruitful one. What the variations do
strongly suggest is that there was never any very
great symbolic significance in the precise order in
which the trump subjects were arranged. It will
be recalled that the Visconti di Modrone pack,
the earliest that has survived.to us, differed from
all later ones in having six court cards, and thus
sixteen cards altogether, in each suit, and also in
containing Faith, Hope and Charity, as well as
Fortitude, among its trump cards, and therefore,
probably, all seven Virtues. It was suggested in
Chapter 4 that it may have had as many as

The Order of the Tarot Trumps 415
twenty-four trumps, the constant factor being the
3-2 ratio of trumps to cards per suit. There is, of
course, no way of being sure of its exact
composition. It is possible that the Visconti di
Modrone pack was no more than a freak, and
that what was later the standard composition of
the Tarot pack was standard from the time of its
first invention. But it is also possible that the
Visconti di Modrone pack represents the original
form of the Tarot pack, and that the 78-card pack
as we know it is the result of a modification
adopted early in its history. If so, the standard set
of twenty-one trumps must itself be the slightly
mutilated remnant of the original, and possibly
larger, set. In that case, we could not expect any
ordering of the trumps in the standard set to
make perfect sense; even if there was any
particular symbolic intention underlying the
original sequence of Tarot trumps, which there
may not have been, we could expect fully to
understand it only if we knew which subjects the
original set contained and in what order they
were arranged. It is unlikely that we ever shall.

This chapter has attempted a reconstruction,
no doubt to be improved as further evidence is
uncovered, of the history of the Latin-suited
Tarot pack. Despite two centuries of research on
playing cards, it was only very recently that this
history began to be investigated. Until then,
writers on playing cards were content to rely on a
standard traditional account, amounting only to
a static classification into types (Tarocchino,
Minchiate, etc.), as if these had all come into
existence on the eighth day of creation; among
such types, the so-called Venetian Tarocchi
formed a mere ragbag comprising all 78-card
Italian Tarot packs. The reason for this failure
has been the lack of any clear concept of standard
patterns: without this theoretical tool, a historian
cannot set aside luxury packs and other
obviously non-standard ones, and hence can
make only the crudest distinctions within the
heterogeneity of the data that then confront him.
As in other areas of the subject, such as
Portuguese-suited cards, the first steps towards
an analysis of the evolution of Latin-suited Tarot
cards were taken by Sylvia Mann. As explained
in the introduction, I have refrained from
cluttering up the preceding exposition with
repeated acknowledgments to her. Her
contribution has, however, been so substantial
that it requires more than a generalised
recognition.. This is particularly so for two
reasons: academics working in the field are prone
to underestimate the contributions of a non-
academic; and the magnitude of her contribution
cannot be estimated from her published writings.
Probably the most enduring monument to her
work will be the Playing-Card Society's
anonymous classificatory sheets, mentioned
above, a project inspired, and in considerable
part executed, by her. A great many of her ideas,
freely offered, have been incorporated into the
work of others and first expressed in print in their
writings; of that process, this book contains many
examples. I have been happy to be able to work in
this field as a member of the school of which she is
the leader.

A number of observations by her formed the
basis for my own work on the subject. She first
drew attention to the importance of the Orfeo
packs, which had escaped the attention of
everyone else, and proposed that, rather than very
incomplete Minchiate packs, they were nearly
complete 78-card ones. She also identified the
Rothschild/Beaux Arts sheets as from what she,
probably mistakenly, described as a 'Tarocchino'
pack, but at any rate from Bologna. She
emphasised the problem posed by the Belgian
Tarot, and remarked on the affinity between
certain of its cards and some of the Italian ones.
She also noticed that the Tarot de Marseille was
restricted to French-speaking regions, and that
the Tarot de Besançon was originally used in
German-speaking ones.

All these observations are, in my opinion,
sound and illuminating. But the most important
thesis advanced by her, concerning the Italian
Tarot de Marseille-derived packs, proved more
problematic. She was the first to distinguish
between the two standard patterns, that
ancestral to the Tarocco Piemontese (IT-1-2),
restricted to Piedmont, and that - with narrow
cards and fold-over backs (IT1-1), made in
many parts of Italy. (The ground for describing
the latter as 'the Lombard pattern' rests on
information from literary sources concerning
where it was used, evidence which, whenever
possible, ought to supplement that based on
place of manufacture.) She remarked that these
patterns appear to have been introduced only in
the eighteenth century, and proposed that this
was to be explained on the hypothesis that the
78-card pack, and the game played with it, had
died out in Italy during the seventeenth century,
and had been reintroduced from France.

416 Part III: Italian Games and Italian Cards
Principally because of the French inscriptions
originally used on these two patterns, I continue
to view this thesis as essentially correct.
Nevertheless, as first propounded, it was
misleading in two respects. First, it rested on the
idea that there was something describable as 'the
Italian 78-card pack', and something describable
as 'the Italian game with 78 cards'. In fact, there
has been much more interplay between the
designs of Tarot cards and the modes of play
with the 78-card pack between countries~other~
than Italy than there has ever been within Italy.
The distinct traditions of design and of play
evidently established before the close of the
fifteenth century appear thereafter to have
remained without influence on one another, and
to have evolved, in so far as they did evolve, quite
independently. What first made this clear was
the investigation into the different trump orders.
The various orders could have been established
only at a date at which there were as yet no
numerals on the trump cards; and, in each of the
different centres, the order there observed could
have been fixed only at the moment of the first
introduction of the Tarot pack, since, once a
trump order had been agreed on, utter confusion
would have resulted among players if it were to
be changed, at least so long as each trump card
was identified primarily by its subject rather
than by an inscribed numeral. But, even if Miss
Mann's thesis, as originally stated, incorporated
an error in this regard, it was she herself who
prompted its correction; for, although it was I
who carried out the investigation into the trump
orders, it was she who saw the importance this
might have and first suggested to me that I look
into it.

Secondly, the thesis suggested, what I at first
assumed, that no designs resembling the Tarot
de Marseille were known in Italy before the
eighteenth century; from this it appeared to
follow that the Tarot de Marseille was a purely
French invention. As explained above, this is not
so at all: the Cary sheet and the Castello
Sforzesco cards show that very similar designs
were in use in Milan from the fifteenth to the
seventeenth century. It therefore appears that the
Tarot de Marseille was derived from a Milanese
prototype; and this conclusion weakens, though it
does not destroy, the case for assuming that the
introduction of the Lombard pattern occurred
only after an interval during which the 78-card
pack had been defunct. Moreover, the natural
presumption is that the form of Tarot game
played outside Italy withjthe 78-card pack was
descended from the Milanese game. On this
assumption, it ceases to be of such importance
whether or not there was an interval before the
adoption of the Lombard pattern: the game
played after the interval would still have been, in
broad outline, similar to that played before it.

In the end, therefore, it appears that Sylvia
Mann's thesis rests on a slenderer base of
evidence, and has less far-reaching consequences,
than at first appeared. That is not to disparage
the importance of the step taken by her in
advancing it. For one thing, it served as a
surrogate for the realisation of the independence
of the four Italian traditions of Tarot play, by
making what is probably to be regarded as the
Milanese tradition appear as an eighteenthcentury
importation from abroad. For another, it
was a first attempt to make sense of the very
confusing evidence, which those who spoke about
'the Venetian Tarot' as a single type alongside
the Tarocchino and Minchiate packs had simply
declined even to try to interpret. At the present
stage, the history of Tarot cards in Lombardy
before the eighteenth century stands in need of
further investigation. We cannot be sure that
there was indeed an interval during which they
were no longer used. We cannot be sure that the
trump order used in the Tarot de Marseille was
really a French invention, and that, before the
eighteenth century, players in Lombardy
remained faithful to the Susio/Viévil order. In
both cases I have suggested affirmative answers;
but more evidence is desirable. We do not know
just how the indigenous Milanese trump designs
evolved, or how close they came to be to those of
the Tarot de Marseille; nor do we know whether,
before 1700, Milanese players continued to use
trumps without inscribed numerals. As for
Piedmont, the uncertainty concerning it is
greater still. We have sufficient reason to assume
that the game was known there before the
eighteenth century, and that there was indeed
some break in continuity; but it is conceivable
that this break occurred earlier, and was ended
by the introduction of Milanese designs, some
time before the latter were in turn replaced by
the earliest form of the Tarocco Piemontese.
Until these problems, and those relating to the
exact evolution of the 78-card pack in Florence,
Rome and Sicily, have been resolved, there can
be.-j no assurance that we have successfully

The Order.of the Tarot Trumps 417

reconstructed the history of Latin-suited Tarot
cards. But the first progress towards such a
reconstruction was made by Sylvia Mann, and it
would have been improper to have ended this
chapter without making clear that this was so.

Note on classical deities set

As stated on p. 409, there are now nineteen cards, of
those found at the Castello Sforzesco with back designs
of classical deities, in the Raccolta delle Stampe
Achille Bertarelli. Novati, in his article in the Bullettino
dei civici musei
, gives the number found of this type as
twenty-one; there is some discrepancy between the
numbers cited by him of cards with particular back
designs and those at present in the Raccolta. In detail,
the numbers are as follows (the figures in brackets
being those given by Novati): love - 4(7); Mercurio -
1(1); Proserpina - 4(2); Ercule-1(3); Marte - 2(0);
Pluto - 2(0); Venere - 1(1); Veritas - 3(3); Saturno -
1(1); Diana - 0(2); unidentified - 0(1). It will be seen
that the discrepancies are not all in the same direction,
which makes them very puzzling.

Game of Tarot Chapter 21, The early Italian game

The Early Italian Game

We can construct a detailed history of Tarot
games, in almost all countries, from the mid-
eighteenth century to the present day, with no
more than the unavoidable minimum of lacunas.
Of the first three hundred years of their history,
on the other hand, we have only the patchiest
evidence. From the seventeenth century we have
only two explicit accounts of Tarot games of any
kind: the description of French and Swiss Tarot
in the 1659 Maison académique, and Paolo
Minucci's account of Minchiate in 1676; to these
we may add the citations in Carlo Pisarri's
Instruzioni of 1754 from his 'very old' manuscript
relating to Tarocchino. For the rest, we must
rely on indirect evidence and backwards

The diversity of orderings of the trumps which
we have seen to prevail in different regions of
Italy points to an early diversification in the
manner of play. Variations in the local design of
cards using the same suit-signs are not always
closely associated with variations in the games
played; but a difference in the order of the
trumps affects the actual practice of play. It is
therefore likely in principle that, where different
trump orderings were observed, there were also
concomitant variations in the rules of play. The
variations in the trump order must have
developed at an early date, before it had occurred
to anyone to put numerals on the trumps. Since
there are such numerals on all but one of the
trumps on the Metropolitan Museum sheets, and
on more than half of those on the Rosenwald
sheet, this means that the differentiation in the
type of Tarot game played in the various regions
probably occurred well before the end of the
fifteenth century.

[new column]
We should beware of distorting the manner in
which Italian players perceived Tarot cards and
the games played with them by the use of such
expressions as 'the game of Tarot' and 'the Tarot
pack'. Outside Italy, even when Tarot has been
extremely popular, the Tarot pack has,
naturally, always been seen as something quite
special and exceptional. Until the Italian suit-
signs were replaced by the French ones, even the
suit-system was quite unfamiliar, except to those
who knew that other oddity, the Trappola pack.
But in Italy, at least during the sixteenth century,
when Tarot cards had existed far too long to be a
novelty, but Tarot games were still sufficiently
popular and widespread for the cards not to
appear a mere local peculiarity, they do not seem
to have been regarded as forming a special pack.
Rather, the Tarot trumps were simply special
cards which you needed to add to the ordinary
ones in order to play certain games. Originally
called triumphi, a set of cards including them were
carte da triumphi, cards with trumps. Thus, in the
passage of his Caos del Triperuno leading up to the
five sonnets on the Tarot trumps, Teofilo Folengo
has Triperuno speak of being led to a room where
there were carte lusorie de trionfi, 'playing cards
with trumps'; and Sperone Speroni, in his brief
tract on games, says that to the cards of the four
suits one sometimes adds certain other cards
called tarocchi, so that the first distinction to be
made concerning the cards used is whether they
are with tarocchi or without tarocchi. Something
viewed as being a special kind of pack, quite
different from ordinary playing cards, for
instance the Aluette pack, the Trappola pack or
the Tarot pack itself outside Italy, is likely to be
treated as an instrument of just one game, even if

The Early Italian Game 419
one played with a number of variations in its
rules. But, where the Tarot trumps are regarded,
not as a part of any such special pack, but merely
as cards which, for some games, are added to the
ordinary pack, one would no more expect them
to be used in just one game than one would
expect ordinary playing cards to be used for only
one game. This expectation is corroborated by
the remark made by Pier Antonio Viti at the end
of his commentary on the special Tarot pack
devised by the poet Matteo Maria Boiardo: with
these cards, he says, many other games can be
played, just as many as is continually done with
the ordinary (Tarot) pack. (1) The point should not
be overstated: Francesco Berni and others refer
to Tarocchi as a specific game. Nevertheless, it
would almost certainly be wrong to suppose that
there was just one such game known in each
region of sixteenth-century Italy where Tarot
games were played. The Tarot trumps were
introduced in order to fulfil a novel function in
card play, and were used in games in which they
fulfilled that function; but there was probably a
considerable variety of such games.

We know of four main centres where Tarot
games were played in fifteenth- and sixteenth-
century Italy: Ferrara, Milan, Bologna and
Florence. It was also played in Reggio nell'
Emilia, Salo, Brescia, Bergamo and Urbino, and
very probably at Mantua and Venice. By the
beginning of the seventeenth century
it had spread to Rome, and, in the course of
that century, it spread to Sicily and presumably
to Piedmont. Of the mode of play in Bologna
we can feel fairly certain. We know that the
game was, in all its essential features, played there
in the seventeenth century in the same manner as
it is played now. The extreme conservatism of
Bolognese players, and the great complexity of
the seventeenth-century game, in which various
additional features that have since fallen away
were superimposed on the basic rules, make it
very likely that those basic rules go back to the
first half of the sixteenth century, when the
shortening of the pack probably first occurred;
and they may well date back to the time before
the pack had been shortened. Minchiate, too, we
know to have been played in the seventeenth
century in essentially the same manner as it was
played in the eighteenth; it is highly probable
1 'Cum li quali molti altri se fariano, e tanti, quanti con il
commune di continuo se fa.'

[new column]
that the fundamental principles of the game were
established when it was first devised.
Of the Milanese manner of play, we have no
direct evidence whatever. Since, however, we
have concluded that it was from Milan that the
game of Tarot first spread to France and
Switzerland, from those countries to reach
Germany and other parts of Europe, we have to
regard Milan as the ultimate source of all the
Tarot games played outside Italy, as well as
those played in present-day Piedmont and
Lombardy. This means that the original
Milanese games were probably very close to
what, at the beginning of Chapter 11, were
described as being the fundamental three- and
four-handed forms of Tarot game, ones in which
all the cards are dealt out, save for three or two
additional ones taken by the dealer, and in which
the counting cards are the court cards and the
XXI, Bagatto and Fool, the Fool serving as
Excuse. The very existence of the early
Piedmontese game rests only on a hypothesis,
and our principal evidence to its nature consists
of the later Piedmontese games of Sedici and
Trentuno, which look like a survival from before
the introduction of games after the French
fashion in the eighteenth century. These have
two striking characteristics, namely that the Fool
does not serve as Excuse, but is the lowest trump,
and that not all the cards are dealt out and there
is no discard. The French games described in the
Maison académique have the second of these
characteristics, though not the first, unlike the
Swiss game described there, which lacks both.
Since we have conjectured that the French games
were played in those areas, Paris and Normandy,
where the Paris/Rouen pattern was used, and
that, in the East of France, Tarot play was more
akin to that described as prevalent in Switzerland,
and since we have also conjectured that the
Paris/Rouen pattern derived from Piedmont, this
affords a partial confirmation of the idea that
Sedki and Trentuno represent some earlier form
of Piedmontese Tarot play. As for the other cities
mentioned above, probably the mode of play in
each of them conformed to that of one of the four
main centres; if any of them was a centre for some
other tradition of Tarot play, it is now

Minchiate was, of course, a deliberate
invention, a conscious variation upon existing
Tarocco games; but we know that in both
Florence and Rome the 78-card pack continued

420 Part III: Italian Games and Italian Cards
to be in use down to the seventeenth century.
Our best clue to the nature of such games
consists in those played in Sicily; the Sicilian
game was probably first imported from Rome,
and the traditions of Tarot play in Rome seem to
have been strongly linked with those of Florence.
If this hypothesis is sound, then what Minchiate
shared with the game played in Florence and
Rome with the 78-card pack was not the versicole
but the fact that, in addition to the Bagatto and
the Fool, all five top trumps, the arie, were counting
cards; the versicole, special combinations
of cards for which one scored not only when they
were contained in a single hand at the start of
play, but also when they were included, in the
cards won by partners at the end of play, formed
a point of resemblance with Tarocchino as
played in Bologna.

What emerges as the principal point of
distinction between the different local traditions
of play is the number of counting cards. In all
games other than Minchiate, the court cards are
always counting cards. In Milan there can have
been only three others, the XXI, Bagatto and
Fool, as in every Tarot game played outside
Italy. In the Bolognese form of Tarocchi, there
are four, the two top trumps, Angelo and Mondo,
the Bagatto and the Fool. In Sicilian Tarocchi,
and therefore, on our hypothesis, in the 78-card
game as formerly played in Florence and Rome,
there are seven, the five top trumps, the Bagatto
and the Fool; and in Minchiate there are twenty-
one, the eleven top trumps, the five lowest ones,
four others and the Fool. On the other hand,
both in Sicilian Tarocchi and in Minchiate,
though not in Tarocchino, there are signs that
the three usual counting cards have a special
role. In Sicilian Tarocchi this lies in the fact that
these cards have a point-value twice that of a
King or of one of the other four arie; in
Minchiate, it consists in the existence of a versicola
consisting of the Trombe, Uno and Matto,
corresponding to the Angel, Bagatto and Fool.
Tarocchino, Minchiate and Sicilian Tarocchi
also all have a feature presumably absent from
the Milanese game, since scarcely a trace of it is
found in any Tarot game played outside Italy: a
fixed score in points for winning the last trick.
This is, indeed, a well-known feature of many
card games played with a regular pack; it is
found, in particular, in the Venetian game of
Trappola. On the other hand, no Italian Tarot
game which we know to have been uninfluenced

[new column]
by foreign modes of play has a vestige of that idea
which became so characteristic an element of
Tarot games as played elsewhere, the special
bonus for winning the last trick with the Bagatto.
In Tarocchino and in Sicilian Tarocchi, and
presumably also in the early Milanese game, the
whole pack is distributed; even in Minchiate,
there is a device for ensuring that all the counting
cards are in play and that the players know how
many trumps are not in play. Virtually the only
-varieties of Tarot that form exceptions to this
general rule are some of the minor Bolognese
games, the Piedmontese games of Sedici and
Trentuno and their hypothetical predecessors,
and the French games described in the Maison

We have, thus, some idea of the type of Tarot
game played in early times in Bologna, in
Florence and in Milan. We cannot in the same
way use the method of backwards extrapolation
to deduce the kind of game played in Ferrara,
and probably also in Venice and other cities
where the type.B order prevailed, because this
game appears to have died leaving no progeny
behind. Our only, uncertain, clue of this kind is
the Venetian game of Trappola, which may have
been invented by card players familiar with
Tarocco. The only direct evidence we have
consists of two literary items, of which,
unfortunately, neither proves as illuminating as
we should wish. The first is the commentary by
Pier Antonio Viti already mentioned. Matteo
Maria Boiardo (1441-1494) was the celebrated
author of the verse epic Orlando Innamorato which
preceded the even more famous Orlando Furioso of
Ariosto. Boiardo composed two sonnets and five
capitoli which were first published posthumously
in 1523 (2) and have subsequently become known
as I Tarocchi, though not so entitled by Boiardo.
They were included in a selection of Boiardo's
work edited by G.B. Venturi in 1820, and were
accompanied, in Angelo Solerti's collected
edition of Boiardo's poetry of 1894, by an
Illustrazione or Commentary by Pier Antonio Viti
da Urbino (c. 1470-1500) not previously
published and taken by Solerti from a
2. In a volume containing poems by various authors and
entitled Amore di Hieronimo Beniueni Fiorentino, Alio Illustris. S.
Nicolo da Correggio. Et una Caccia de Amore bellissima & cinq:
Capituli, sopra el Timore, Zelosia, Speranza, Amore, & uno
Trionpho del Mondo, Composti per il Conte Matteo Maria Boiardo
et altre cose diverse
(Venice, 1523). The volume was reprinted
five more times in the years up to 1537 and again in 1808.

The Early Italian Game 421
manuscript. (3) Viti's commentary is addressed to a
lady of the court of Urbino, though not to the
Duchess Elisabetta herself; Ridolfo Renier
suggested as the recipient Emilia Pia, a close
friend of the Duchess. (4) Viti explains that the
poems are intended to be engraved on the cards
of a specially designed pack, and expresses the
hope that his patroness will order such a pack to
be made. There are to be eighty cards altogether,
two of which are to bear nothing but the two
sonnets: the remainder are to comprise four suits,
twenty-one trionfi and the Macto (i.e. Matto). The
composition of the pack, apart from the two
sonnet cards, was thus to be standard for a 78-
card Tarot pack; but the suit-signs and the
trump subjects were to be non-standard. The
suits were to be Arrows, Vases, Eyes and Whips;
each of the trionfi was to symbolise a quality,
represented by a historical or mythological
character. The trump subjects for the most part
show no correspondence with those of the
standard set, and, when they do, occupy a
different position; for instance, the highest trump
was Fortezza (Fortitude), represented by
Lucretia; in a reversal of the usual arrangement,
the Macto, called Folle by Boiardo himself, was
designated Mondo (World), and shown as a fool
riding a donkey. The figures on the court cards
were to be identified with characters from
classical mythology. Each of the seventy-eight
cards was to bear a tercet from the appropriate
capitolo: those of the Arrows suit from the capitolo
on Love, of the Vases suit from that on Hope, of
the Eyes suit from that on Jealousy, and of the
Whips suits from that on Fear; the tercets for the
trionfi and the Macto were to be taken from the
fifth capitolo, the Triumph of the World. Wood-
engraved packs were actually made to Viti's
precise specifications. One example, missing the
Matto, the trumps and twelve other cards is
mentioned by Merlin in his book of 1869; (5) what
3. Le Poesi Volgari e Latine di M.M. Boiardo, ed. by A. Solerti,
Bologna, 1894, pp. 313-38, with notes pp. xxxii-xxxv. Poems
and commentary can also be found in Tutte le opere di M.M.
ed. by A. Zottoli, vol. 2, Milan, 1937, pp. 702-16,
with notes pp. 748-9.
4. See R. Renier, 'Tarocchi di M.M. Boiardo', in Studi su
Matto Maria Boiardo
, ed. by N. Campanini, Bologna, 1894.
Elisabetta Gonzaga, Duchess of Urbino, sister of Francesco
Gonzaga, Marquis of Mantua, and so sister-in-law of
Isabella d'Este, is, like the Lady Emilia Pia, famous from
Castiglione's The Courtier.
5. R. Merlin, Origine des cartes à jouer, Paris, 1869, pp. 94-6
and plate 28.

[new column]
was probably the same copy was sold in London
in 1971 to Signor Carlo Alberto Chiesa of Milan.
Another example, missing the Matto and the
court cards, but including the trumps, was
mentioned by Carlo Lozzi in 1900. (6)

Viti gives some instructions for playing with
Boiardo's pack. Although Viti himself was from
Urbino, and although his commentary is
addressed to a lady at the court of Urbino, it is
not unreasonable to connect this piece of
evidence with Ferrara, at least if Viti be assumed
to be setting out the intentions of Boiardo
himself. Boiardo was brought up in Ferrara and
studied there; he spent some time at court, and
was highly esteemed by Borso d'Este, Marquis of
Ferrara from 1450. Borso was made Duke of
Modena and Reggio by the Emperor Frederick
III in 1452, and Duke of Ferrara by Pope Paul II
in 1471, the year of his death. Boiardo was in the
service of his successor, Ercole I, the father of
Isabella and Beatrice, acting as ducal captain of
Modena from 1480 to 1482 and as governor of
Reggio from 1487 until his death there in 1494. It
must, therefore, have been the Ferrarese type of
Tarocco game with which he was familiar; it is
possible that they played in the same manner at
Urbino. Viti's instructions, though interesting,
are of far less value for us than we might have
hoped, because, as is plain from the remark
quoted above, Viti takes it for granted that his
reader will know all about the games normally
played with the ordinary Tarot pack, and
restricts himself to an account of a special game
to be played with Boiardo's cards. This game
falls into four parts. The two cards bearing the
sonnets are first set aside, and then each draws a
card to determine the dealer, who is the one
drawing the highest card. When the cards have
all been dealt out, the first part of the game
consists simply in each player's reading out the
verses on his cards, from which, Viti naively
remarks, much amusement may be had. It is not
explained what happens when the number of
players is such that the cards cannot be
distributed equally and exhaustively. The second
part of the game consists in playing out the cards
in tricks: it is necessary to follow suit when
possible, and, when not possible, to play a
6 Carlo Lozzi, 'Le Antiche Carte da Giuoco', La Bibliofilia,
vol. I, Florence, 1900, pp. 37-46 and 181-6. For more
detailed discussion of this pack, see M. Dummett, 'Notes on
a 15th-century pack of cards from Italy', Journal of the
Playing-Card Society,
vol. I, no. 3, February, 1973, pp. 1-6.

422 Part III: Italian Games and Italian Cards
trump. Nothing is said about the way the Matto
is played. Among the numeral cards, the higher-
numbered beat the lower-numbered in the suits
of Arrows and Vases, but, in those of Eyes and
Whips, the lower-numbered beat the higher-
numbered; among the trumps, the higher-
numbered win. At the end of the play, each
player is paid by every other player who has won
fewer tricks than he has as many scuti as the
difference in the number of tricks they have won.
At this stage, those who have won no tricks fall
out of the game, the third part of which now
takes place. Using the cards they have won in
tricks, each player reckons up points as follows:
he counts one positive point for each card he has
in the suits of Arrows and Vases, and one
negative point for each card in the suits of Eyes
and Whips; presumably trumps do not count
either way. When the points have been thus
reckoned, each player may demand from any
other who has a lower score that he hand over to
him any card of the demander's choice.
Presumably it is meant that one card is to be
surrendered for each point of difference in the
scores, and presumably also, if a player asks for a
card which the other does not have, he cannot try
again. In choosing which card to ask for, a player
will have an eye on the fourth part of the game:
this consists in each player's putting together the
longest sequence of consecutive cards which he
now has in his hand. The winner - the player
with the longest such sequence - can demand
from the loser (presumably the player with the
shortest sequence) everything that he has about
his person; clearly, one needs to deposit one's
valuables elsewhere before sitting down to play
this game.

This somewhat jejune entertainment is hardly
a serious card game; it is more consonant with
the kind of society game described in various
works such as Ringhieri's Cento Giuochi Liberali.
There is no reason for assuming that any of its
special features derive from already existing
games played with the ordinary Tarot pack,
although we cannot exclude such a possibility;
the main value of the account, for our purpose, is
in respect of those few fundamental features
which it does share with Tarot games as we know
them. It is a trick-taking game; the trionfi serve as
permanent trumps. It is obligatory to follow suit
if one can, and to play a trump if one cannot. Not
only do all the court cards beat the numeral

[new column]
cards of a suit, but the order of the numeral cards
runs in one direction in two suits and in the other
direction in the remaining two. Without_Viti's
commentary, we should have assumed that these
were all features of Tarot games from the first;
but it is pleasant to have confirmation of that
assumption, and, although Viti is not purporting
to describe ordinary Tarot games, the fact that
his, or Boiardo's, game has these features is some
confirmation. On the other hand, there is not a
trace of the principle of assigning different point-
-values-to the court-cards and to some of the
trump cards; the second part of the game is a
simple trick-taking game in our technical sense,
where only the number of tricks count. This can
hardly be called evidence that simple trick-taking
games were ever played with the ordinary Tarot
pack; but it hints at the possibility.

The most frustrating feature is Viti's silence, in
the section on how to play with the cards, over
the Matto. Possibly he simply assumed that
everyone knew how to use the Matto; but the
other possibility is that he intended it to be
treated as the lowest trump, as in the eighteenth-
century Piedmontese games of Sedici and
Trentuno. The allusions to it in Boiardo's verses
and in the parts of Viti's commentary relating to
the cards are maddeningly ambiguous. In the
fifth capitolo, which deals with the trionfi, the tercet
to be inscribed on the Matto stands first, the
subject of the card being given as Mondo; there
follow the tercets on the twenty-one trionfi in
ascending sequence. From the two lines in the
first sonnet relating to the trionfi, it is difficult to
tell whether Boiardo is counting the Fool as a
trionfo or not: he writes: 'Con vinti et un
Trionfo e al piu vil loco/E un Folle, poi che'l
folle el mondo adora' ('With twenty-one
trionfi, and in the lowest place there is a Fool,
because the fool adores the world'). In Viti's
commentary the Matto is not spoken of as a
trionfo in the paragraph devoted to it, and the
next card, Ozio (Idleness) is called the first trionfo,
each of them being given a number up to the last
one, Fortitude, said to be in the twenty-first
place. At the end of the section on the trionfi, Viti
speaks of Boiardo's capitolo as divided into
twenty-two tercets about 'twenty-two trump
cards, with the Matto' (in vintidue carte de Trionfi,
con el Matto
); the phrase could not have been
better designed to leave us in uncertainty
whether Viti regards the Matto as a trump or
not. Clearly, the Matto is regarded as in some
way. different from the twenty-one ordinary

The Early Italian Game 423
trionfi: but whether that is because it plays a quite
different role in the game, as it does when it
serves as Excuse, or merely because, like the
Miseria of the Sicilian pack, it does not bear a
number, it is impossible to tell.

The only other early source giving any
information about how Tarocco was played may
be definitely associated with Ferrara. It is not,
alas, an account of the rules, but a verse diatribe .
against the game by Flavio Alberto Lollio (1508-
1568), entitled Invettiva contra il Giuoco del Taroco
and first published in Venice in 1550. (7) Lollio was
born in Florence, but lived most of his life in
Ferrara, where he died, and he is referred to in
the by-line of the poem as 'Flavio Alberto Lollio
Ferrarese'. He assumes the game to be known to
his readers, but describes a game in progress; this
description is both illuminating and puzzling.
The one thing that is absolutely clear is that
three players take part in the game. The cards
are dealt out in several rounds almost certainly
five at a time; the poem follows the fortunes of a
single player, referred to, save in an occasional
line, in the second person singular. Evidently,
after each round of the deal there was a pause,
during which it was possible for a player to
increase the value of the round, and also for one
to propose going a monte, that is, abandoning the
deal. This latter was, as we have seen, a very
general custom in Italian games; but normally it
required the agreement of all the players for the
deal to be abandoned at any stage. The player
whose fortunes are followed does not appear to
be the dealer, for it is said at the outset only that
the deal 'is begun', not that 'you begin' it. But his
first five cards are good ones; the result is stated
7 The Invettiva appears on pp. 272-82 of Rime Piacevoli,
Venice, 1550, a collection of verses by various authors. A
second edition was published in Ferrara in 1590. A selection
from the poem as printed in the 1590 edition was included
by Samuel Weller Singer as appendix XIII to his Researches
into the History of Playing Cards
, London, 1816, pp. 354-6. This
selection includes most of the part describing the play, but,
unfortunately, not quite all, in part because Singer evidently
wished to spare the blushes of his readers by suppressing
lines containing words like orinal and cacapensieri. He does
not give a translation. A translation of part of the poem,
without the original, is given by Stuart R. Kaplan, The
Encyclopedia of Tarot
, New York, 1978, pp. 29-30. The
selection coincides very nearly, but not quite, with that
made by Singer, but, unfortunately, no indication is given of
the places where some lines have been omitted, and the
omissions occasionally affect the sense; a reader with no
other source to refer to would suppose that the whole of the
poem had been translated.

[new column]
to be that 'you hold the invitation, and make it
again' (tu tien I'envito, & lo rifai). The word I
have here rendered 'invitation' is actually
printed enutto in the first edition, a non-existent
word. In the second edition of 1590 it appears as
invitto ('unconquered'), which makes little sense;
the most likely reading is envito or invito =
'invitation'. Presumably the invitation is to
continue the deal for an increased value; our
player not only accepts this invitation, but
proposes a further increase. The player's next
five cards are bad ones. Our player then goes a
(vai à monte). This cannot mean, as it is
rendered by Mr. Kaplan in his partial and largely
accurate translation, that he throws in his cards,
because the deal continues, and we find him,
after four rounds of dealing, with twenty cards;
he is merely proposing an abandonment of the
deal, but in vain, since another player is going to
increase the stake, that is, the value of the round.
There follow two more rounds of the deal, one
good for our player, the other bad. He then
expects some help from the last cards; but when
they arrive, they are terrible, the opposite of what
he needs. Presumably, then, he is dealt a final
five cards; unfortunately, there is here an obscure
line. When you are expecting the last cards to
help you (quando / L'ultime aspetti, che ti dian
), Lollio says, and now comes the obscure
line: 'Having already invited them from the
stock' (?); you then, he continues, see arrive
hideous cards to make you die, etc. Tu ti vedi
arrivare (oh dolor grande)/Carte galioffe da farti
morire,/Totalmente contrarie al tuo bisogno
). The
obscure line contains a questionable reading: in
the 1550 edition it is printed Havendogli invitata già
dal resto
, which is'ungrammatical. The grammar
is corrected in the 1590 edition by the reading
Havendola invitata già dal resto - ('Having already
invited it from the stock'), but there is nothing for
the singular pronoun to refer to. Probably
Havendogli was correct, the gli, in what would
now be considered a grammatical error, referring
back to the ultime [carte] (last cards), and invitata
should be invitate. Assuming this to be so, we are
left unclear in what sense the player has incited
these last five cards (if there are five of them):
perhaps by his invitation after the first round of
the deal, perhaps by his acceptance of the
increase in the stake after the second. Now, Lollio
says, enraged by the miserable cards he has
received at the end, the player hurls reproaches
at 'the remainder of his cards, which are twenty'

424 Part III: Italian Games and Italian Cards
(Lo avanzo de le Carte, che son venti); presumably
these are the cards he received on the first four
rounds of the deal, which is why I said that the
dealer gave each player five cards at a time.

Our problems of interpretation are by no
means over. The player now sorts his hand into
the trumps {trionfi) and the four plain suits. Now
comes the remark, 'quindi s'hai quattro,/O cinque
Carte di Ronfa, tu temi/Che non ti muoia il Re, con le
: 'if you then have four or five carte di ronfa,
you are afraid that your King and the court cards will
die'. We know from accounts of Minchiate
that a King 'dies' when it is captured by another
player's trump; but what does Lollio mean by
carte di ronfa? We saw in Chapter 7 that the term
ronfa had a curious double meaning, sometimes
referring to the play of a trump and sometimes to
a long holding in a suit. This line of Lollio's is the
only known instance of the use of the word in
connection with a Tarot game. At this stage of
his description, actual play has not yet started,
and the reference to the King and the other court
cards makes it evident that he is speaking of a
holding in a plain suit. But that does not resolve
all uncertainty. Why should our player be so
anxious at having four or five cards in a suit, even
if they are all high ones? If we are to understand
that each player finishes with 25 cards (perhaps
with the dealer taking the last three and
discarding three - nothing is said about this),, an
average holding in a plain suit will be about 4-6.
If there was no discard, it will be very unlikely
that either of the other two will have a void in a
suit in which one holds four or five cards; even if
the dealer discards three cards, there is no
especial danger that he will have created a void in
that suit. Possibly four or five carte di ronfa are the
top four or five cards of a suit in which the player
has much greater length, say seven or eight
altogether, though that seems a complex
meaning for the simple phrase carte di ronfa to
bear. Probably there was a penalty for losing a
King; possibly there was a rule, as in Minchiate,
obliging a player to surrender his King if a trump
had been played to the first trick in that suit. But,
with all these suppositions, the anxiety displayed
- 'your heart aches and your mind is racked' -
seems disproportionate to the actual danger.

It will be recalled that the French Tarot games
described in the 1659 Maison académique were
ones in which not all the cards were distributed
among the players, and there is no mention of a
discard. On the other hand, the Swiss three-

[new column]
handed game there described resembles later
three-handed games played with the 78-card pack
and without bidding in that the dealer gives
25 cards to each of the other two players and 28
to himself, discarding three. We saw that the
Swiss manner of playing Tarot appears to have
changed little between the seventeenth century
and the twentieth; and two details of the modern
game strongly suggest that the Swiss mode of
play is based very directly upon Italian models.
One is the option of abandoning the deal, by
agreement, after each round of the deal, that is,
in effect, of going a monte. The other is the signal,
made by striking the table with the fist while
leading trumps, given by a player to convey that
he has the XXI. This signal, unknown except
in Italy and Switzerland, is the batto of
Sicilian tarocchi and of the old Bolognese game,
and is related to the busso of the modern
Bolognese game and also of Tressette, which is
popular in Italian-speaking areas of
Switzerland. The signal probably represents a
feature of Italian Tarot play already present by
the mid-seventeenth century. Both the Bolognese
and Sicilian games bear witness to the same
general principle, namely a complete distribution
of the pack between the players save for two or
three extra cards, going to the dealer except
when there is some form of bidding. If it were not
for the passage in Lollio's poem about the carte di
, the most natural interpretation of the game
he describes would be similar. The player
receives twenty cards, and then, after that, some
last ones, on which he has counted to redeem his
hand; the obvious assumption is therefore that he
receives 25 cards altogether, perhaps with the
last three going to the dealer, who has to discard.
This is reinforced by Lollio's later complaining
about the tedious necessity of counting every
trump that is played. When all the trumps have
been distributed, or when, as in Minchiate, the
players know just how many remain undealt, the
counting of trumps is an utter necessity for
successful play, above all for a player who wants
to save or to capture the lowest trump. In a game
in which a large number of cards remain undealt,
it loses something of its importance.

Even with the mysterious passage about the
carte di ronfa, this remains the most probable
interpretation; but that passage does raise a
doubt. The most likely meaning of 'four or five
carte di ronfa' is the four or five top cards of a suit,
irrespective of whether or not the player has

The Early Italian Game 425
other cards in that suit. If so, the alarm of Lollio's
player at finding himself with such a holding
remains exceedingly overwrought, if we suppose
every player to have 25 cards. The Swiss game is,
if our analysis is correct, in the Milanese
tradition; Bologna had its own tradition, and the
Sicilian game probably ultimately derives from
Florence. But Lollio was surely describing
Tarocco as played at Ferrara; and it is possible
that the game played there differed from that
played in, say, Milan in respect of the number of
cards each player had when play began. How
could such a supposition fit with what Lollio
says? We have seen that it is unlikely that, in
Lollio's game, there remained any undealt cards,
and that, after the first four rounds of the deal,
the player received some last cards; but we may
ask why Lollio draws explicit attention to the
number of cards - twenty - dealt out in the first
four rounds of the deal, without mentioning,
what is of greater importance, the number of
cards the player held when the deal was
completed. This question suggests the answer
that 20 was the number of cards each player held
before play began: that each player discarded as
many cards as the 'last' ones he received, after
the original twenty cards had been given out.
Perhaps, on this last round of the deal, each
player received six cards, discarding six; or
perhaps the dealer received eight and the other
players five each, making corresponding
discards. True, Lollio says nothing at all about
any discard; but, if we make such a supposition,
we can at least make sense about the anxiety a
player might feel if he had all four court cards in
some suit, for then he would be in serious danger
of losing them.

I offer this suggestion only as a possibility; I do
not consider it the most likely one, though I shall
explore possible consequences of it below. For
the present, let us return to Lollio's account.
Actual play now begins: our unfortunate player
finds himself having to keep following suit to the
leads of the other two players, evidently unable to
win a trick. It is, of course, perfectly evident from
the mention of following suit, as from the earlier
reference to the King's dying, that we have here a
trick-taking game, in which the trionfi serve as
trumps. When our player revokes, the voices of
the others are raised in protest; otherwise, one
must keep even more silent than when attending
Mass. Ferrarese players plainly took the game
very seriously. From the remark, in the carte di

[new column]
ronfa passage, about the King dying, with the
court cards, it is apparent that all the court cards
were counting cards, and possible that, as in
Minchiate, one scored not only for having them
among one's tricks, but additionally for
capturing them. But Lollio remarks grumblingly
that it is necessary to take note of every least card
that is played. This suggests that, as in the
Bolognese game, players may have scored not
only for individual cards won in tricks, but also
for particular combinations of cards among those
so won; otherwise, it would be a grave
exaggeration to say that a player had to take note
of every single card. If it should sometimes
happen that you have a good hand, Lollio says,
you play it so badly that you lose one or two
dozen of them, sometimes ail (ne perdi/Una dozzina
ò due: tal hora tutti
). If Lollio here means that you
lose one or two dozen of your cards, this would
rule out an interpretation under which the
players received less than 25 cards each; but it
seems more likely that dozzine are here some type
of scoring device or token. And now comes a
remark that is really baffling. How many times,
Lollio asks, are you unable to cover the Matto?
(Quante volte non puoi coprire il Matto?) As a result,
you unwillingly find yourself robbed of the good
you have gained {Onde mal grado tuo, spogliar ti
senti/Del buon c'havevi
). What is it to 'cover' the
Matto, and why does it have such disastrous
consequences? One might conjecture that
'covering' the Matto consisted in giving a card in
exchange for it, the player unable to do so having
to surrender the Matto. But this can hardly be
made to fit the context: a player in such a
position is one who has won no tricks, and such a
player has no other good to be robbed of. The
most natural interpretation of the word 'cover' is
'to play a higher card', i.e. to capture. If this is
right, then, in the Ferrarese game, the Matto
cannot have served as Excuse, as in the
Bolognese and Sicilian games and in Minchiate,
but must have been a trump, presumably the
lowest one, as in Sedici or Trentuno. Lollio does
not mention the Bagatto or Bagatella, save in a
general list of trionfi after the description of play;
the Matto is here mentioned in just the context
where we should expect the Bagatto to be spoken
of. The Matto must surely have been a counting
card; but that by itself does not explain why a
failure to capture it (or, possibly, a failure to
bring it home) should have spelled such ruin.
Perhaps there was a high premium for bringing it

426 Part III: Italian Games and Italian Cards
home, or a high penalty for failing to do so; or
perhaps, if our earlier supposition that a player
scored for particular combinations of cards won
in tricks is correct, the Matto could, as in
Minchiate, be added to such combinations, or
functioned as a wild card, like the two contatori in
the Bolognese game, when such combinations
were being scored.

At this point, Lollio leaves his description of
the play to engage in a general diatribe against
the game and the childish nature of the cards
used. From his poem, I have extracted more
problems than solutions; but the poem is
precious for anyone interested in the history of
the game of Tarot, since it is the only account of
the game, of any kind, from before the
seventeenth century. I therefore give in the
appendix the whole text of the relevant passage,
with the best attempt at a translation I have been
able to make, in the hope that someone will be
able to arrive at a surer interpretation than I have
been able to do.

Now, with very few exceptions, it is
characteristic of later Tarot games, of the most
diverse kinds and whatever the size of the pack,
that there is no stock of undealt cards, but that
every card in the pack counts, at the end of the
round, to one or another player or side. It is
almost equally characteristic that there is a small
residue of undealt cards forming a talon, which,
in games without bidding, goes to the dealer, and
whose distribution, in games with bidding,
depends upon the final bid; in each case, those
receiving additional cards must discard an equal
number, under certain nearly constant
constraints. This practice is not a mere device for
handling the situation when the number of cards
in the pack is not exactly divisible by the number
of players, since it is observed in three-handed
games, even without bidding, played with the 78-
card pack; the most striking example is that of
Tarok-Quadrille, played by four players with
only 76 cards, in which the dealer still takes four
extra cards. Even Minchiate is not a genuine
exception to this rule, because, although the
mechanics are different, there are still discards,
and the principle is upheld that the players
should know how many cards of each suit are in
play, even though not every card belongs at the
end of the round to one side or the other. Both
the Bolognese and Sicilian games incorporate the
standard practice. It therefore seems
overwhelmingly probable that the practice is one

[new column]
going back to an early stage in the history of the
Tarot games: the fact that it is found in both
Bolognese and Sicilian Tarocchi debars us from
supposing that it was invented outside Italy and
introduced there only at the time of the invasion
of the Tarot de Marseille pattern in the
eighteenth century. This conclusion is reinforced
by an etymological consideration. In the
terminology used in Germany, the discard was
almost always referred to as the Scat, and this
name was also used, by transference, for the
talon; the exception is the game of Cego, in
which the talon is called the Blinde and the
discard the Legage. In Austria, too, the term Scat,
sometimes in the form Scar, was originally used,
although it was later dropped in favour of the
French word Talon for the talon, with no separate
noun being used for the discard. The words Scat
and Scar are obviously corruptions of the Italian
word scarto, meaning 'discard', and it therefore
seems likely that the practice itself is of Italian
origin. It might be objected that the borrowing of
the term Scat may have occurred only after the
reintroduction of the 78-card game from France
into Italy in the eighteenth century; we know
that Viennese Tarot players of the mid-
eighteenth century borrowed a specific form of
play from Lombardy, and, with it, an Italianate
vocabulary, and that this Viennese/Lombard
game spread into Germany and as far as the
Netherlands. But this objection appears
unsound: classic Tarot games, in which the word
Scat was used, were being played in Germany
before the spread of the Viennese/Lombard

A possible source for the practice may have
been a card game called Scartino, of which we
hear much from a brief period around 1500:
there are over a dozen references to it between
1492 and 1517. (8) We have no idea how Scartino
8 For references to Scartino, as played by Beatrice,
Isabella, Ercole, Ippolito and Alfonso d'Este, Lodovico il
Moro, and others, see: F. Malaguzzi-Valeri, La corte di
Lodovico il Moro
, vol. 1, Milan, 1913, p. 575; A. Venturi,
'Relazioni artistiche tra le corti di Milano e Ferrara nel
secolo XV', Archivio Storico Lombardo, anno XII (pp. 225-80),
1885, p. 254; A. Luzio and R. Renier, Mantova e Urbino,
Turin and Rome, 1893, pp. 63-5, especially fn. 3, p. 63; the
same two authors, 'Delle relazioni di Isabella d'Este
Gonzaga con Lodovico e Beatrice Sforza', Archivio Storico
, anno XVII (pp. 74-119, 346-99, 619-74), 1890, p.
368, fn. 1, and pp. 379-80; A. Luzio, I precettori d'Isabella
, Ancona, 1887, p. 22; G. Bertoni, 'Tarocchi
versificati' in Poesie, leggende, costumanze del medio evo, Modena,

The Early Italian Game 427
was played, although it appears to have
demanded a special type of pack; for instance,
Lodovico il Moro wrote in 1496 to Cardinal
Ippolito d'Este complaining that the latter had
not sent him the carte de scartino that he had
promised, and there are other references to
orders for packs of Scartino cards. The game
seems to have originated from Ferrara: it was a
favourite game both of Beatrice d'Este, wife of
Lodovico il Moro, Duke of Milan, and of Isabella
d'Este, wife of Francesco Gonzaga, Marquis of
Mantua. The name Scartino is presumably
connected with the verb scartare, 'to discard', and
games are often named after their most
characteristic or novel feature. It is therefore a
possibility that this was a trick-taking game in
which a new practice was introduced, namely
that the dealer took some extra cards and
discarded a corresponding number. If so, it could
be that it,was from Scartino that this practice
was taken over into Tarocco games, in which it
had been previously unknown, and that Scartino,
after its short-lived popularity, died out, having
made a lasting contribution to card play. This, of
course, is the merest guess: Scartino may not
have been a trick-taking game at all, but, say, one
in which the winner was the player who first
contrived to get rid of all his cards after the
fashion of a Stops game.

If Scartino did influence Tarocco, it is possible
that the practice whereby the dealer took extra
cards and made a corresponding discard was not
an original feature of Tarot games, but was
incorporated into them about the beginning of
the sixteenth century, in time for it to be
imported, as a feature of the game, when Tarot
arrived in Switzerland. In that case, the French
1917, p. 219; and the Diario Ferrarese of 1499 in Muratori,
Rerum Italicarum Scriptores, vol. 24, p. 376. A letter of August
1493 quoted by Malaguzzi-Valeri and by Luzio and Renier
appears to imply that Scartino was a three-handed game.
The earliest reference is from 1492; one is from 1509, one
from 1517, and all the rest from the 1490s. Several concern
the obtaining or ordering of packs of Scartino cards (para de
carte da scartino
or para de scartini), which appear all
to have come from Ferrara; what was special about these cards
there is no way of telling. It is just conceivable that Scartino
was itself a particular type of Tarot game, and that these
were therefore Tarot packs of a special type; but, unless
they were very special, it does not seem very likely that
Lodovico Sforza should have been having to obtain Tarot
packs from elsewhere. Most of the references are about
games of Scartino being played.

[new column]
games described in the Maison académique may
represent a yet more ancient tradition; if our
conjecture that they were derived from Piedmont
is correct, Tarot playing in that region may go
back to the very earliest times, the players
remaining exceptionally conservative. But the
hypothesis that the important feature of the
discard was borrowed from Scartino should be
treated with great caution, since it implies a
continued mutual influence between the style of
Tarocco play in all four great early centres,
Ferrara, Milan, Bologna and Florence. However
this may be, we have every reason to assume that
the principal games played with the Tarot pack
were from the start trick-taking games in which
the trionfi were permanent trumps, and in which
it was obligatory to follow suit-v/hen-possible
and, when one could not, to play a trump if
possible: the only doubt is as to the original use
of the Matto. There is also every reason to
assume that they were from the start complex ones
in our technical sense, i.e. with different cards
carrying differing point-values. There is scarcely
any indication of a Tarot game in which only the
number of tricks taken is significant. What we
have throughout called the 'original' method of
reckoning points, namely by counting 1 point for
each trick, with the point-values of counting
cards reckoned separately, as in the Spanish
game of Malilla, must indeed have been the
original one: on no other assumption can we
explain the genesis of the distinctive method of
computing the point-totals found in almost all
Tarot games. In every Tarot game except
Minchiate, the counting cards of the plain suits
are the court cards, as Lollio's poem suggests;
and, as in many games with the regular pack,
they always have point values increasing by 1
with each card. So constant is this feature that we
can take it as certain that it was original, their
values under the original system being 4 for the
King, 3 for the Queen, 2 for the Cavallo and 1 for
the Jack.

Everywhere outside Italy, the three remaining
counting cards are the XXI, the Bagatto and the
Fool or Matto. Since the French and the Swiss
probably imported the game of Tarot from
Milan, and the Germans imported it from
France, the selection of these three counting
cards, in addition to the court cards, is likely
to be the Milanese practice. In Bologna, as we
have seen, the four counting cards other than
the court cards were the top two trumps, the

428 Part III: Italian Games and Italian Cards
Bagatto and the Matto; and we should take the
evidence of the Bolognese game particularly
seriously, since the indications are that
Bolognese players have always observed an
extreme conservatism in their manner of play,
which, in essentials, probably therefore faithfully
represents a sixteenth-century style of game.
Nevertheless, as already noted, although both
Minchiate and Sicilian Tarocchi admit several
other counting cards, both games retain features
which set the top trump, the Bagatto~and~The-
Matto aside from the other cards. It therefore
seems likely that the system found in all non-
Italian Tarot games, whereby these three cards
equal the Kings in value, and all the other
trumps are low cards, was the original one. We
still have to account for the fact that, in all the
Italian Tarot games surviving from before the
eighteenth century, this system has been
replaced by one under which at least one other
trump has a point-value.

It may be that no further explanation is
required than that this was a means adopted to
make the game more complicated and more
interesting. One fact, however, suggests the
possibility of a more specific explanation. This is
the coincidence that, when the standard manner
of counting points is being used, the cards being
counted in threes, the total number of points,
when the full pack is employed, is the same as the
number of cards in the pack, namely 78; and this
is the same as to say that, when the original
method of counting points was used, the total
came to 78 when the game was a three-handed
one. Perhaps this is merely accidental. But there
is a further coincidence: when there are no
bonuses, the total number of points in Trappola
is also 78; it is hard to avoid thinking that this
was arranged to agree with the point-total in
Tarocco games.

The aggregate number of points gained by all
the players together is not constant, and is
therefore not significant, except in games in
which all the counting cards are in play, or at
least contribute to some player's point-total. If
our supposition that this was not true of the
earliest Tarocco games is correct, the agreement
between the number of cards in the pack and the
total number of points in a three-handed game
cannot have existed until the new style of play, in
which the entire pack was distributed, came into
vogue; it is therefore likely to have been
accidental in the first instance. But, once noticed,

[new column]
it is likely to have seemed a desirable and
amusing feature: the coincidence of the 78 points
and 78 cards strikes me as just the sort of conceit
that would appeal to men and women of
Renaissance times.

If so, some method would probably have been
devised for extending it to games played with
more than three players. One possibility is that
this was the genesis of the score for the last trick.
It is only because there is such a score in
Trappola that the total number of points comes
to 78; without it, it would amount only to 72. In
assuming that the total number of points was 78
in early, as in later, three-handed Tarocco
games, we have in effect been assuming that in
those games there was no score for the last trick.
But, when the game is played by four players,
each must start with 19 cards, the dealer having
had an extra two and having discarded two; at
least, this is how it must be if we assume that, as
in the later games, only the dealer received extra
cards, and that he received as few as possible. On
the original method of reckoning points, 1 point
is allowed for each trick of four cards. In
supposing the total number of points in the three-
handed game to be 78, we were tacitly assuming
that 1 point was awarded to the dealer for his
discard of three cards; so let us assume that, in
the four-handed game, the dealer was awarded 1
point for his two discarded cards. With no
further adjustment, the total number of points in
the four-handed game will then be only 72:
namely 52 for the nineteen counting cards, 19 for
the nineteen tricks and 1 for the dealer's discard.
This could be brought up to 78 once more by
awarding 6 points for the last trick - its value
both in Trappola and in the Bolognese form of
Tarocco. We may, admittedly, be on quite the
wrong track: but it is conceivable that this was
the original motivation for the score for the last
trick. Five-handed Tarot games are rare; but, if
the painting of 'The Tarocchi Players' at the
Casa Borromeo in Milan deserves its name, they
are as old as the Tarot pack itself. Five players
can receive at most 15 cards each, with a discard
of three cards for the dealer, making a total of 68
points on the cards themselves: this would then
require a score of 10 points for the last trick - its
value in Minchiate - to bring it up to 78.
The fact that a score for the last trick is
virtually unknown outside Italy makes it
probable that it was not awarded in the early
Milanese game, or at least not in the three-

The Early Italian Game 429
handed form of it. It may nevertheless be thought
•to-be more likely that, in other areas of Italy, it
was an original feature of the game, independent
of the number of players. Without altering our
other assumptions, we cannot reconcile this
supposition with an aggregate total of 78 points
in the three-handed game. It will be recalled,
however, that our inconclusive study of Lollio's
poem yielded a possible alternative
interpretation of the deal, one according to which
each player had an opportunity to discard and
ended with a hand of 20 cards. If points were
awarded, as usual, for counting cards, whether in
tricks or in the discards, but otherwise only for
actual tricks taken, the aggregate total would be
72 points (52 for counting cards + 20 for tricks);
with a score of 6 points for the last trick, this
would yield a total of 78.

As already remarked, the interpretation of
Lollio's account here invoked is not the more
probable one, though it does help to resolve the
problem about the carte di ronfa. But suppose that
it were correct: how, then, could the aggregate of
78 points be maintained in a four-handed form?
We already know one answer to this - 19 cards to
each player, a discard of two cards for the dealer,
counting 1 point, and 6 points for the last trick.
Such a game would, however, lack sufficient
analogy to the three-handed form we are
hypothesising. If each player were to be able to
make a discard, no points being awarded for the
discards save for the counting cards contained in
them, the aggregate total could be maintained
only by assigning a point-value to another trump.
The most likely assignment of this kind would be
one of 4 points to the second highest trump; the
aggregate total of 78 points would then be
ensured if each player held 16 cards at the start of
play (56 points for counting cards, 16 for tricks
and an extra 6 for the last trick). In such a game,
the remaining 14 cards would be distributed
among the four players, with corresponding
discards (say five to the dealer and three to each
of the other three players). Such a game, played
with the full pack and without cricche or
sequences, might be the remote ancestor of the
Bolognese game. After the cricche and sequences
had been introduced, the aggregate total would
again become variable and hence unimportant;
there would then have been no obstacle to the
subsequent shortening of the pack. In a five-
handed game devised according to the same
principles, yet one more trump would have to

[new column]
acquire a point-value: the players would then
have to begin play with hands of 12 cards each,
with a total discard of 18 (say five by the dealer
and three by each of the others).

It would be wrong to give much weight to
either of these alternative speculations. Whether
the variations in the set of counting cards came
about in this way or in some other, the Sicilian
game, presumably going back to the seventeenth
century, in which, in addition to the Bagatto, the
Matto and the court cards, all five top trumps are
recognised as counting cards, testifies to a
freedom felt in Italy about the selection of the
counting cards never exercised in other countries.
The arie of the Sicilian game may have been
imitated from those of Minchiate; more
probably, those of Minchiate were borrowed
from some earlier Florentine game, played with
the 78-card pack and directly ancestral to
Sicilian Tarocchi. The fact that, nevertheless, the
three cards which, in games played outside Italy,
are always the only ones beside the court cards to
possess point-values seem in Italy also to have
been regarded as in some manner special may be
due to their having been the only original
counting cards, or to their being the only ones
recognised in Milan, or possibly to the three-
handed game's having acquired, save in Bologna,
an ascendancy over other forms. This effect
would have been reinforced if there were, in some
three-handed games, a practice of making
declarations of combinations of cards held in
hand before the beginning of play, and among
such combinations were the set of the three
counting cards of the tarocchi (the tous les trois of
other lands). Such declarations are not
mentioned in Lollio's poem, but, in view of the
importance that they assumed both in
Tarocchino and in Minchiate, as well as their
presence in non-Italian Tarot games, it is
extremely likely that they figured in some pre-
eighteenth-century Italian games played with the
78-card-pack. If a declaration of the set consisting
of the highest trump, the Bagatto and the Matto
were carried over into games with more than
three players, despite there being one or mpre
other trumps with point-values, this would be
enough to preserve a sense that these three cards
were especially important and in some way
belonged together.

It is in any case likely that it was in its three-
handed form that the game of Tarot first
travelled from Italy to France and Switzerland,

430 Part III: Italian Games and Italian Cards
and from there to other countries; and, as far as
the records show, it was as a three-handed game
that it first appeared both in Switzerland and
Germany. But it did not take with it any score for
the last trick; and this makes it likely that such a
score was absent from the game as played in
Milan. We can associate it firmly with the games
played with the trumps arranged in a type A
order, those of Bologna, Florence, Rome and,
eventually, Sicily. But, in view of its presence in
Trappola, it is very likely to have figured in the
Tarot games played in Venice and Ferrara,
observing a type B order for the trumps.

There is one striking similarity between
Tarocchino and Minchiate. In both of them, the
special combinations of cards for which points
are earned, the cricche and sequences of
Tarocchino and the versicole of Minchiate, play
the predominant role, primarily because they are
scored for when they are contained within the
tricks won by a pair of partners, and in
consequence overwhelm the points made on
individual cards, particularly in Tarocchino. We
have noted that there is some oblique indication
in Lollio's poem that the game he was describing
had the same feature, although we cannot be at
all sure of this. It is a feature that does not appear
in the Sicilian game. It evidently represents one
line along which Tarot games developed in Italy
during the sixteenth century; but either that was
not the only line of development, or it had, by the
seventeenth century, been generally abandoned.

But for the single line in Lollio's poem, 'How
many times are you unable to cover the Matto?',
the only reasonable assumption would be that
the Matto served from the beginning as Excuse;
the Piedmontese games of Sedici and Trentuno
would then have to be regarded as freaks, with no
historical importance. It seems hard to interpret
Lollio's question, however, save on the
supposition that, for him, the Matto was a low
trump which it was of the highest importance to
capture; and since, as observed, the poem is the
sole direct piece of evidence that we have
concerning the way in which Tarot was played in
the sixteenth century, we must take it seriously.
If, as I have argued, the invention of the Tarot
pack represented, if not the first, at any rate an
independent invention of the idea of trumps, then
it would be a little surprising if, from the first, the
Matto had, in play, its special role of Excuse: for
that presupposes the simultaneous introduction
of two radically new ideas into trick-taking play,

[new column]
whereas, in games, as in other fields, it is rare for
anyone to have more than one entirely_new idea
at a time. The possibility thus arises that
originally the Matto was the lowest trump and
the Bagatto only the second lowest; if, as seems
likely, those two cards were both counting cards
from the start, then, at least in the three-handed
game, the counting cards in trumps consisted
originally of the highest one and the two lowest.
We have seen that Lollio's remarks about the
-consequences- of-failing to 'cover' the Matto
suggest that it was not merely one among several
counting cards with a high point-value; and I
suggested that perhaps it functioned like a
contatore in Tarocchino, that is, as a wild card
able to fill gaps in special scoring combinations of
cards won in tricks. If so, it would have been seen
as a card having a function different from all
others, even though it behaved in actual play
simply as the lowest trump; and this might serve
to explain why it was never numbered or
regarded as occupying a numerical position. On
this hypothesis, the. invention of its special role in
play, as Excuse, must have occurred at some
stage after the original invention of the game. If
this is correct, this was only the first of two
changes of role to which the Fool was subjected
in the history of Tarot; as we shall see in Part IV,
it later became a trump once more, this time the
highest trump, beating even the XXI.

Thus the position of the Bagatto in Sicilian
Tarocchi, as bearing the number 1 and as yet not
being the lowest of the trump cards, may not
have been wholly unparalleled. It was suggested
in the last chapter that the practice, in packs in
which the trumps are arranged in an order of
type A, of starting the numbering after the
Bagatto may have been a device for ensuring that
Death received the number 13. If, however, the
Matto was originally the lowest trump, the
numbering nevertheless always starting after it,
an alternative interpretation becomes possible. If
the original order was of type B or type C, then
type A orders may have been a later
development, which took place only after the
Matto had acquired its new role of Excuse; the
practice of starting the numbering after the
Bagatto may then have been an imitation of the
previous practice of starting it after the Matto
when that card was the lowest trump. On this
interpretation, our previous idea should be
inverted: it would in this case have been the
displacement of the numbering one place

The Early Italian Game 431
upwards that prompted the rearrangement of the
ordering- all three Virtues having to be brought
below Death in order to ensure to that card the
number 13.

I do not pretend to be advancing a satisfactory
theory on this question. On the contrary, it is
definitely a puzzle; and perhaps we lack the
means of solving it. If we knew only of the games
with the 54-card pack, to be described in Part IV,
in which the Fool plays an entirely different role,
and had no more than its name, der Sküs, and one
cryptic line from a poem to hint to us of its former
role as Excuse, it is improbable that we should
ever guess how it had been used in other games;
or, conversely, if we knew nothing of the games
with the 54-card pack save the allusions in some
poem like Lollio's, we should never deduce the
function of the Sküs in those games. The only
piece of evidence we have for the Matto's having
served as the lowest trump are the passage in
Lollio's poem and the games of Sedici and
Trentuno, which look like a survival from an
earlier epoch, but in which the Matto does not
have a high point-value; to these we may add, as
a very tenuous clue, Pier Antonio Viti's failure to
specify the role of the Matto in the game with
Boiardo's pack, although he specifies almost all
other particulars. Everything else that we know
would fit more easily with the supposition that
the Matto served as Excuse from the first
invention of the game of Tarot. Even Sedici and
Trentuno fail to fit perfectly with our hypothesis,
because, when it is assumed, it is most natural to
think of the type A orders as coming into
existence after the introduction of the new way of
using the Matto, as Excuse, whereas the fact that
in Piedmont the Angelo was treated as higher
than the Mondo suggests that it was a type A
order that had been familiar there before the
game died out. If the present hypothesis is to be
entertained at all, the invention of the new role
for the Matto cannot be dated later than the
1470s. The reason for this is that, whereas in
Garzoni's book, in Susio's poem and in the
Bertoni poem, the trumps are listed in sequence,
beginning with the highest, il Mondo (the World),
with the Matto coming at the end, the
arrangement in the Steele sermon, written at
least seventy years before Lollio's poem, is
different. In the sermon the trumps are listed in
ascending order, beginning with el bagatella, and
the Matto, cited as el matto sie nulla (the Fool or
zero), still comes at the end, after the 21, which is

[new column]
oddly cited as '21. El mondo cioe dio padre' ('21.
The World, that is, God the Father'). Moreover,
the entry for the Bagatella reads in full 'Primus
dicitur el bagatella: et est omnium inferior'
('The first
is called the bagatella; and it is the lowest of
all'); (9) this explicit statement is immediately
preceded by a sentence stating that there are
twenty-one trumps (Sunt enim 21 triumphi ...). All
this is quite conclusive evidence that the preacher
did not regard the Matto as a trump in the
proper sense, and did not take it as ranking
below the Bagatto. Now Lollio was writing some
seventy or eighty years later. Admittedly, it may
have been some time since he had played
Tarocco, since, earlier in the poem, he describes
himself as having formerly been an enthusiast for
it, and as only later having come to detest it. But
he was not an old man when he wrote the
Invettiva; he was only 42 when it was published.
This may indicate, therefore, that his line about
the Matto should be understood in some quite
different way from that adopted here. If not, it
must be that the Ferrarese game was in this
respect of a more ancient type than that known to
the preacher of the Steele sermon. This is not
impossible. Games in which the Fool played its
new role as the highest trump and ones in which
it served, in traditional fashion, as Excuse,
coexisted in Germany for a considerable period,
and both are played, in different countries, to this
day: But, as remarked, the whole question is a
puzzle; and I regret having to leave it unresolved.

If there is a problem about the Matto, there is
a surprise concerning the Bagatto, or, more
exactly, concerning the bonus for the Bagatto
ultimo, for winning the last trick with the Bagatto.
Throughout the Austrian Empire, this Pagat
ultimo bonus was to become the most celebrated
feature of Tarock in all its forms. The use of the
Italian term ultimo naturally suggests that the
practice was of Italian origin; and an equally
natural conjecture is that it was an adaptation of
the Trappola idea of a bonus for winning the last
trick with the lowest card of a suit. Trappola as
described by Cardano was a game without
trumps; but it may well be that it had
incorporated the idea of trumps before it
migrated from Venice to more northerly regions.
These speculations, natural as they are, do not
9 The words are here given in full, as in Steele's article,
rather than in the severely abbreviated form in which they
appear in the manuscript.

432 Part III: Italian Games and Italian Cards
square with the facts. There is no indication of a
bonus for the Bagatto ultimo in Bolognese or
Sicilian Tarocchi or in Minchiate, nor any hint of
it in Lollio's poem or in any of the Italian literary
sources which refer to the individual Tarot
trumps. As for the term ultimo, this was used in
nineteenth-century Vienna, but there is no
evidence for its use anywhere outside Italy before
the nineteenth century; the bonus itself was a
feature of German Tarok games played in the
eighteenth century, before the term ultirmo had
been adopted. The bonus does not seem to have
been imitated from Trappola, but, as we saw in
Chapter 11, to have originated as a correlative to
a penalty for losing the Pagat, doubled when this
happened in the last trick; and it was probably in
Germany that it made its first appearance.

We saw that the Bagatto ultimo bonus was
well established in Piedmont by 1787,
presumably having reached Italy from Germany
via France. It may well be that it arrived in
Austria, not directly from Germany, but by the
route Germany-France-Italy-Austria, bringing
the term ultimo with it from Italy. Milan was
awarded to Austria under the Peace of Utrecht in
1713. The award was contested by the Spanish
Bourbons, in alliance with the French, but,
under Maria Theresa, Milan came firmly under
Austrian control at the Peace of Aix-la-Chapelle
in 1748, and remained so until 1859, save for the
Napoleonic interlude from 1796 to 1814. The
bonus for winning the last trick with the Bagatto
or Pagat presumably reached Vienna from Italy
at some time during the later eighteenth century,
the Italian term ultimo being adopted for a
practice not in origin Italian.

As we shall see in the next chapter, what may
have been copied from Trappola games was the
practice of announcing at the start of play an
intention to gain the bonus for Pagat ultimo. But
this development did not take place until the
early nineteenth century, and then not in Italy,
where Trappola had long been forgotten, but,
almost certainly, in Austria, as an imitation of
the announcement of intention to make a twenty-
six in Hundertspiel; we find it in Austria by 1821,
and in Piedmont by 1830. However the Bagatto
ultimo bonus and announcement originated, we
find no trace of the former before the eighteenth
century or of the latter before the nineteenth: in
both cases the evidence is clearly against an
Italian origin.

We succeeded, in the last chapter, in

[new column]
reconstructing, from a meagre supply of clues,
the history of the-Tarot pack in Italy from
the fifteenth century to the eighteenth.
Unfortunately, no comparable success can be
claimed for the attempt in this chapter to
reconstruct the early history of the game in Italy.
There are as many clues in this case as in the
other; but some are hard to interpret, and some
point in opposite directions. I hope that someone,
reflecting on the evidence, will be better able to
discern a pattern than I have succeeded in doing;
or perhaps new evidence will come to light.


The following is the text of that part of the poem from
the last line of p. 276 to the last line of p. 279 in the
1550 edition; I believe that this includes all that bears
in any degree on the way the game was played. The
footnotes give readings from the 1590 edition, as
quoted by Singer, that represent significant
emendations, together with three other emendations
in square brackets.

[Transcriber's note: in what follows, most of the lines are left-indented, but a few are not. Since such formatting is not possible here, I have indicated those lines by putting in a line space before them, assuming that a break is intended.]

Ecco che s'incomincia a dar le carte:
La prima man ti fa vna bella vista,
Tal, che tu tien l'enutto, (1) & lo rifai:
Quelle, che vengon dietro, altra facenda
Mostrano hauer: ne più de' casi tuoi
Tengon memoria alcuna: onde tu stai
Sospeso al quanto: & di vada: quell'altro
Ilqual par che il fauor lor si prometta,
Ingrossera (2) la posta: allhor trafitto
Da vergogna, dolor, d'inuidia, e d'ire,
Ten vai à monte, co'l viso abbassato.

Non à si gran cordoglio vn Capitano,
Quando si crede hauer la pugna vinta,
E mentre ei grida vittoria, vittoria;
Da nuouo assalto sopragiunto vede
Andar la gente sua rotta, o dispersa,
Quando ha costui. Vengon dapoi quell'altre

Due man di carte, hor liete hor triste: & quando
L'vltime aspetti, che ti dian soccorso
Hauendogli (3) inuitata gia dal resto,
Tu ti uedi arriuare (oh dolor grande)
Carte galioffe (4) da farti morire,
Totalmente contrarie al tuo bisogno.
Onde di stizza auampi; e tutto pieno
Di mal tal talento, (5) rimbrottando pigli
Lo auanzo de le Carte, che son uenti.
1 l'invitto
2 [ingrossera]
3 Havendola
4 gaglioffe
5 mal talento

The Early Italian Game 433

Queste t'empien (6) le mani, & buona pezza
Ti dam irauaglio e briga, in rassettare.
Dinar; Coppe; Baston, Spade; e Trionfi.
Però che ti conuien ad una, ad una,
Metterle in ordinanza; & far di loro,
Come farebbe il buon pastor, che hauesse
Di molti armenti; apparecchiando mandre
Diuerse per ciascun Quindi s'hai quattro
O cinque Carte di Ronfa, tu temi
Che non ti muoia il Re, con le figure:
Onde si strugge il cuor, spesma, la mente, (7)
Stando in bilancia fra speme, e timore.

Quello e lo isfinimento e'l creppacuore,
Che sei sforzato a tener per tuo specchio
Certe cartaccie che ti fan languire:
Et, come se tu fussi un'Orinale,
Seruir conuienti à gli altri due compagni,
Rispondendo, à ciascun giuoco, per giuoco:
Et se per ignoranza, ò per errore,
Da in una Carta, (8) che non uada à uerso,
Tu senti andar le uoci infino al cielo.

Ne ti pensar che quiui sian finite
Le pene tue: bisogna tener conto
D'ogni minima Carta, che si giuochi,
Altramente ogni cosa va in ruina.
Pero tu brami spesso la memoria
Di Mitridate, di Cesare, ò di Ciro.

Et s'egli auien tal hor c'habbi un bel giuoco,
T'andrà si mal giocato, che neperdi
Vna dozzina ò due: tal hora tutti.

Quante volte non puoi coprire il Matto?
Onde mal grado tuo, spogliar ti senti
Del buon c'haueui: & sembri la cornacchia,
Che resto spennacchiata infra gli vccelli,
Allhora se tu fossi uno Aristide,
Vn Socrate, un Zenone, un Giobbe, un fasso,
Tu sprezzaresti il fren della patienza,
Stracciaresti i Tarocchi in mille pezzi,
Maladicendo il primo che ti pose
Mai carte in mano, e t'insegnò à giocare.

Doue lasso quel numerar noioso
D'ogni Trionfo, ch'esca fuori? o quanto
Fastidio hai tu di questo, che non puoi
Pur ragionar, pur dire una parola:
Anzi seruar conuien maggior silentio
Che non si fa alia Predica, o la Messa.

Ei mostro ben d'hauer poca facenda,
Et esser certc vn bel cacapensieri
Colui, che fu inuentor di simil baia:
Creder si dè, ch'ei fussi di pintore
Ignobil, scioperate, (9) e senza soldi,
Che per buscarsi il pan si, mise (10) a fare
6 t'empion
7 spasma la mente
8 Dai una carta
9 [scioperato]
10 [pan, si mise]

Cotali filostroccole da putti.

Che vuol dir altro il Bagatella, e'l Matto,
Se non ch'ei fusse vn ciurmatore, e vn barro?
Che significar altro la Papessa,
II Carro, il Traditor, la Ruota, il Gobbo;
La Fortezza, la Stella, il Sol, la Luna,
E la Morte, e l'lnferno; e tutto il resto
Di questa bizaria girando l'esca,
Se non che questi hauea il capo suentato,
Pien di fumo, Pancucchi, e Fanfalucche?
Et che sia ver, colei che versa i fiaschi,
Ci mostra chiar ch'ei fusse vn ebbriaco;
E quel nome fantastico, e bizarro
Di Tarocco, senz'ethimologia,
Fa palese à ciascun, che i ghiribizzi
Gli hauesser guasto, e zorpiato il ceruello.

Dr. Lorenzo Minio-Paluello very kindly went through
this passage with me; on the basis of his help, and I
hope without importing any errors of my own, I
append the following translation.

See, the cards are beginning to be dealt. The first
hand looks good to you, so that you hold the invitation
[reading I'invito in view of the invitata that comes later],
and make it again. Those [cards] that come next show
a different state of affairs; they no longer have your
chances in mind. You therefore remain in suspense;
and on it goes[?]. That other [player] who seems to
expect their favour [the favour of the cards] will
increase the stake; then, wounded by shame, pain,
envy and anger, you go a monte, with face downcast.

A Captain who thinks he has won the battle, and,
while he cries, 'Victory, victory!', sees his people
crushed or dispersed by an unexpected new attack,
does not feel anguish so great as does this [player].

There then come two other hands of cards, first
fortunate, then miserable: and when you are
expecting the last [cards] to give you some help,
having invited them [reading Havendogli invitate] from
the stock [?], you see arrive (oh, bitter pain) hideous
cards to make you die, quite the contrary of what you
need. You are therefore inflamed with vexation, and,
full of an evil disposition, you begin scolding the
remainder of your cards, which are twenty.

These fill your hands, and for a long time they give
you trouble and worry to arrange Coins, Cups,
Batons, Swords and Trumps, because it is useful to
put them in order one by one; to do with them as a
good shepherd would do, if he had many flocks,
preparing different folds for each. Then, if you have
four or five carte di ronfa, you fear that the King will
die, with the court cards; and therefore your heart
aches and your mind is racked, standing in balance
between hope and fear.

That is the exhaustion and the heartbreak, when
you are forced to keep before your eyes, as in a mirror,

434 Part IIII: Italian Games and Italian Cards
some low cards that make you droop. And, as if you
were a urinal, you must be at the disposal of the two
other players, having to respond suit by suit to each;
and if, by ignorance or mistake, you play a card which
does not fit, you hear their voices go up to the skies.

Do not think that your sufferings are at an end here.
You must take account of every least card that is
played, otherwise everything goes to ruin. You
therefore often long for the memory of Mithridates, of
Caesar or of Cyrus.

And if it sometimes happens that you have a good
hand, you will play it so badly thaf you lose one or
two dozens; sometimes you lose them all.

How many times are you unable to cover the
Matto? Whence you feel that, against your will, you
are deprived of the good you have; and you look like
the crow that lost its feathers, among the other birds.
Therefore, if you were an Aristides, a Socrates, a
Zeno, a Job, a stone, you would despise the curb of
patience, and would tear the tarocchi in a thousand
pieces, cursing the first person who ever put the cards
in your hands and taught you to play.

Where do I leave that tedious counting of each

[new column]
trump that comes out? How much annoyance you
have from being unable to discuss anything or say even
a single word: one must keep greater silence than
one does at a sermon or at Mass.

He who invented such nonsense showed himself to
have little to do, and truly to have diarrhoea of the
mind; we must suppose that he was a worthless
painter, out of work and penniless, who, in order to
earn his bread, started making such childish

What else do the Bagatella and the Matto mean,
'save that he was a trickster and a cheat? What else
can they signify, the Popess, the Chariot, the Traitor,
the Wheel, the Hunchback, Fortitude, the Star, the
Sun, the Moon, Death, Hell and all the rest of this
revolving bizarrerie, save that he [the inventor] had an
empty head, full of smoke, caprices and idle tales?
The woman who empties wine-bottles [Temperance]
clearly shows that it is also true that he was a
drunkard. And that whimsical and bizarre name
'Tarocco', lacking an etymology, makes it manifest to
everyone that his fantasies had damaged and ruined
his brain.

Chapters 7 and 8 of Dummett's Game of Tarot, 1980


The Game of Tarot

With a pack of cards, you can play games; you
can tell fortunes; you can do conjuring tricks; or
you can build a house of cards. For the fourth
and most trifling of these uses, it matters not at
all whether the faces of the cards bear anything
on them or not. For the third, all that is necessary
is that your audience can readily identify a card
shown to them. The second of these practices we
have seen to have been virtually unknown until
the eighteenth century. It follows that a special
type of pack, invented around 1440, can have
been invented for one purpose only: for play.

We have no actual description of any game
played with the Tarot pack from before the mid-
seventeenth century. We do, however, have
sufficient evidence from which to extrapolate
backwards, so as to determine, with high
probability, at least the main features of the
original game of trionfi, the game played with the
Tarot pack at the time of its first invention and
that for which it was devised. All later forms of
the game have certain constant features, which
may be presumed to have belonged to it from its
inception. We know how certain forms of the
game, played with variant types of Tarot pack,
were later played, and we have excellent grounds
for thinking that these games changed little in the
course of the centuries; two of these variant packs
— the Minchiate pack and the Tarocco Bolognese
pack (sometimes called the Tarocchino pack) —
originated in the sixteenth century, and so we
have a good idea of the sixteenth-century
Italian style of play. We do have a
description of the game intended to be
played with a very special form of Tarot pack
invented at the very end of the fifteenth century
by the poet Matteo Maria Boiardo. We cannot
place a great deal of weight on this, because the
description ends with the remark that one can
also use this pack to play the usual sort of game;
but at least it seems probable that those features
which the game played with Boiardo’s pack
shared with those later played with the standard
Tarot pack were, from the first, characteristic of
Tarot games.

In asking for what purpose the Tarot pack was
invented, we are asking why the twenty-two
additional cards — the Matto and the twenty-one
triumphs — were added to the regular pack. We
shall not gain any enlightenment if we study the
iconography of the Tarot pack. That study is, in
itself, fascinating, and has absorbed many
people, not all of them with occultist leanings; it
is, no doubt, worth pursuing for its own sake. But
it is highly improbable that, by this means, we
shall learn anything relevant to the game played
with Tarot cards, or, therefore, to the primary
purpose for which the pack was originally
devised. Whoever invented the pack wanted to
add, to the regular playing-card pack, two new
elements: one consisting of a single card, the
Matto or Fool; the other consisting of a sequence
of twenty-one picture cards. The members of the
latter sequence had to be memorable and readily
recognisable. One way of achieving this would
have been to inscribe large numerals on the
cards; the subjects would then have been
comparatively unimportant, just as it is
essentially a matter of indifference to a modern
Bridge or Poker player what the court cards look
like, so lgng as they have clear indices to give
their rafik and suit. This was, indeed, the method

The Game of Tarot 165
adopted, from the end of the fifteenth century
onwards, in many, possibly in most, Tarot packs;
and we shall see that, in the type of Tarot pack
that, from the later eighteenth century on,
superseded the traditional form in Germany and
Central Europe, the subjects on the triumph
cards really did come to be entirely irrelevant.
But, when the Tarot pack was invented, such a
device would not have seemed natural. In the
mid-fifteenth century, indices of any kind were
unknown: card players were used to identifying
each card — court card or numeral card — from
the whole design, not from one special ingredient
of it. It was therefore necessary to choose, for the
new picture cards, a series of very definite
subjects, that could be easily distinguished, and
with each of which-could be associated a name
that could be used to refer to the card. The
subjects on the Tarot triumphs — the Sun, the
Devil, Death, the Virtues, the Wheel of Fortune,
Love, the Pope and all the rest — served just this
purpose. By and large, they were standard
subjects of mediaeval and Renaissance
iconography, to be met with in many other
contexts; they were precisely the sort of subjects
which any fifteenth-century Italian, faced with
the problem of devising a sequence of twenty-one
picture cards, would have been likely to select.
We can derive some entertainment from asking
why that particular selection was made, and
whether there is any symbolic meaning to the
order in which they were placed; and we may or
may not come up with a plausible or illum-
inating answer. (If we do not, that may not
indicate that we have failed to solve the riddle;
there may be no riddle to solve.) But our answer,
though it may throw light on what the original
designer of the pack, or the Duke or other noble
who ordered it to be made, had in mind, is
unlikely to throw any on the way in which an
average fifteenth-century player of the game
would have viewed the cards. For him, they were
simply a set of picture cards arranged in a
particular sequence and having a particular role
in the game; he would be as unlikely to take any
special interest in the selection of subjects or
possible symbolic significance in their order as a
modern Bridge player is to be able to tell you
which Jacks have moustaches.

We can see how little importance was attached
to the precise subjects represented on the
triumph cards from the variations introduced at
various times and places. The presence of the
Pope and Popess has perennially given offence to
some, and various replacements have been found
for them. In that variation on the Tarot de
Marseille pattern known as the Tarot de
Besançon, originating in the seventeenth century
and widely used in eighteenth-century Germany
and Switzerland, Jupiter and Juno were
substituted for the Pope and Popess. In the quite
different pattern used in Belgium for the Italian-
suited Tarot pack during the eighteenth century,
the Pope and Popess were replaced, respectively,
by Bacchus and the Spanish Captain (the latter a
character from the Commedia dell’Arte). In
Bologna after 1725, in deference to Papal
displeasure (Bologna being within the Papal
States), the Pope, Popess, Emperor and Empress
were all removed in.fayour of four Moorish kings
or satraps. It would be a mistake to read into
these substitutions any subtle symbolic equation
of the new figures with the old ones: all that was
being sought was a choice of new and easily
recognisable figures, with obvious labels, in place
of the old ones that caused offence. In a similar
way, when a Sicilian duchess objected to the
presence of the Devil in the pack, her wishes were
accommodated by replacing him with a Ship, a
card borrowed from the Minchiate pack and
there representing the element of Water. It could
be objected that all that these examples show is
that, by the seventeenth and eighteenth
centuries, the original symbolism had been
forgotten or .was no longer regarded as
significant. But the fact is that the same attitude
was already clearly displayed in the sixteenth
century. The Minchiate pack invented in that
century in Florence is the sole form of Tarot
pack — with the possible exception of that
represented by the Visconti di Modrone pack —
having a different number of triumph cards from
the otherwise invariable twenty-one: it has as
many as forty. To find new subjects for this
greatly increased number of triumph cards, the
devisers of the pack added the four Elements, the
four missing Virtues (Prudence, Faith, Hope and
Charity) and the twelve signs of the Zodiac. It
has seldom been suggested that the resulting,
sequence of forty triumph subjects has any
special symbolic significance: it is obvious that
those who devised the Minchiate pack simply
looked around for a convenient method of adding
new and memorable subjects to the existing ones.
If this was true in the sixteenth century for the
inventors of the Minchiate pack, there is little

166 Part I: History and Mystery
reason to suppose that it was not also true a
hundred years earlier for the inventors of the
Tarot pack itself. At any rate, if there was any
special symbolism underlying the sequence of
Tarot triumphs, as originally devised, this played
no role in the games played with the cards, and
hence disappeared from the consciousness of
card players.

It might be thought that the introduction,
towards the end of the fifteenth century, of the
practice of inscribing numerals on the triumph
cards marked a change in the manner of play
with those cards: that, originally, they did not
form a sequence with a definite order, but that
the association of a numbering with them
represented the imposition of such an order for
the first time. If so, the original game played with
the Tarot pack can have had no resemblance to
those later played with it; and the suggestion is
weakened by the grave difficulty of making any
plausible proposal about what kind of a game it
can have been in which there were twenty-two
distinct picture cards, all with a special role in
the game, but none ranking higher or lower than
any other. The suggestion is, however, quite
superfluous. It is founded on the fact that we
should find it very troublesome to have to
memorise an ordering of twenty-one different
picture cards without the aid of a numbering
appearing on the cards themselves. But, as we
saw in Chapter 4, players of the special form of
the game of tarocchi peculiar to the city of Bologna
and its environs continued to do just that until
the later eighteenth century, when for the first
time makers of the Tarocco Belognese pack
adopted the practice, universal elsewhere, of
putting numerals on the triumph cards — more
exactly, on those of the middle range, from 5 to
16. The eighteenth-century card-game books
that give the rules of the game as played in
Bologna simply list the names of the triumph
cards in order: anyone wishing to play the game
had to commit this sequence to memory. As
noted in Chapter 4, there can be no doubt that
this ordering was already a feature of the game
in the seventeenth century; the intense
conservatism of Bolognese players, in respect
both of the rules.of the game and of the designs of
the cards, makes it overwhelmingly probable that
it goes back to the beginning of the sixteenth. Of
that which, for two and a half centuries,
Bolognese card players were able to do, there is
no reason to declare those of the fifteenth-century
courts of Milan and Ferrara incapable.

In any case, there is positive evidence that,
from the start, the triumph cards were regarded
as forming an ordered sequence. The earliest
detailed reference to the Tarot pack is the
sermon, already mentioned, by an anonymous
Dominican friar included in the volume of
sermons formerly belonging to Robert Steele,
and dating from between 1450 and 1480. In this,
the triumph cards are not merely listed in
sequence, but actually numbered. No hypothesis
makes any kind of sense, or has any evidence in
its favour, other than that the additional cards of
the Tarot pack, the Fool excepted, formed from
the outset a series with a determinate ranking
which originally had to be memorised by the
players and subsequently came to be indicated
by numerical indices on the cards.

To understand the purpose for which the
Tarot pack was invented, we have, therefore, to
ask for what reason an ordered sequence of cards,
of different length and composition from the
ordinary suits, was added to the regular pack:
the particular subjects depicted on these cards
can, in this context, be entirely neglected.
Obviously, to find the answer, we have to look: at
the role that these cards play in the game, on the
reasonable assumption that the essential features
of the game, in the various forms in which it was
later played, belonged to it from the start. This is
easily stated: the game of Tarot is a trick-taking
game, and the twenty-one triumph cards are
permanent trumps.

Probably most readers of this book are already
familiar with trick-taking games such as Whist,
Bridge or Pinochle. It ray be worth while,
however, to make. the mechanics of such games
quite explicit at this point, particularly because
we ordinarily take it for granted that a number of
distinct features go together.

Trick-taking games

In any trick-taking game, it is essential that, at the
beginning of actual play, the players should ail start
with an equal number of cards in their hands. In each
game, there is a cyclic direction of play - clockwise or
counter-clockwise. In Tarot games, the direction is
usually counter-clockwise: this is because Tarot is an
Italian game in origin, and in Italy and Spain the
standard direction was originally counter-clockwise,
while north of the Alps and the Pyrenees it was
clockwise. The cards held by the players are played

The Game of Tarot 167
out in tricks, each trick consisting of as many cards as
there are players, each player contributing one. Each
player plays his card to the trick face upwards in the
centre of the table. At the beginning of each trick, one
of the players has the lead, that is, the duty to play the
first card to the trick: the other cards are then played
successively by the other players in rotation,
according to the designated direction, clockwise or
counter-clockwise. If, at the beginning of play, each
player holds 7 cards, there will therefore usually be n
tricks, though in certain games, so long as the stock of
cards not dealt out before play is not exhausted, each
player replenishes his hand by drawing one card from
the stock after each trick; this is primarily a device for
making the game one of imperfect information when it
is played by only two players. The rules of the game
will provide a way of determining which is the
winning card in any trick; the player who played that
card is said to have won the trick, and he gathers up
all the cards of the trick and places them face down
beside him. The cards of a trick already playéd are
out of play, and cannot be used again in any later
trick; but the outcome of the given round of play
depends on what cards, and how many, have been
won in tricks by the various players. (A ‘round’ of
play is to be understood as that segment of the game
that lasts from one deal to the next.) When two or
more players play in permanent or temporary
partnership, one partner usually collects all the cards
won in tricks by himself or any of his partners.

The foregoing description may be taken as a
definition of a ‘trick-taking game’: there are some
games having some of those features but not others,
for example ones in which the player who has won the
trick takes the cards of that trick into his hand for
later use, but they are based on an essentially different
principle. In some trick-taking games, a good deal
takes place in any round before actual play begins, for
example bidding: but there are three main features to
be specified before the mechanics of the actual play is
determined. These are: (i) who has the lead to each
trick; (ii) which is the winning card in any trick; and
(iii) what constraints restrict a player’s freedom to
choose the card he will play to a trick.

In trick-taking games — not only European ones,
but those played in China and in India and Persia as
well — it is a virtually universal rule that the winner of
any trick before the last has the lead to the next trick.
This leaves only the player who leads to the first trick
to be determined by convention. The rule varies in
different games: but, in European games without
bidding, it is usually the player next in rotation to the
dealer — the one on the dealer’s right if the direction of
play is counter-clockwise, and the one on his left if it is
clockwise — who leads to the first trick.

Winning a trick

In trick-taking games played without trumps, the rule
almost always holds that a card of any suit other than
that of the card led to the trick is devoid of trick-taking
power: the trick is therefore won by the highest-
ranking card of the same suit as the first card played
to it. This rule, complicated in most games by the
presence of trumps, is so nearly universal that most
card players would probably regard it as one of the
defining characteristics of trick-taking games. An
alternative principle, however, is not merely
abstractly conceivable, but is actually realised in a few
deviant trick-taking games, and serves to point up
how dominant is the usual rule that (trumps
excepted) only a card of the suit led can win the trick.
The alternative is to disregard suit altogether in
determining who has won the trick, and count as the
winning card that of highest rank irrespective of
suit. Such a rule prevails in the French game of le
Truc and its now obsolete English cousin Putt. In a
modified form, it also governs the still popular games
of Aluette and el Truco, the former played on the
West coast of France with a special type of 48-card
Spanish-suited pack, the latter in Spain and South
America with the 40-card pack. Under this rule, it can
of course happen that two cards of equal rank, higher
than any other card in the trick, are played. by
opponents to the same trick. Such a situation can also
arise, even under the usual rule, in any game played
with two or more packs simultaneously; and it is then
sometimes resolved by counting the card played later
as beating that played earlier, sometimes by the
opposite rule. In all the four deviant trick-taking
games mentioned above, however, such a trick is not
considered as won by any of the players. Yet another
conceivable alternative would be to regard the entire
pack as forming a single sequence, so that a trick
would be won by the highest-ranking card of the
highest suit; but, as far as I am aware, there is no
game governed by a principle of that kind. Even the
first alternative rule for determining which card wins
the trick holds good only in a very small minority of
unusual games, just those four already: mentioned: in
the overwhelming majority of trick-taking games the
rule is that, unless trumps are admitted and a trump
is played to the trick, the winning card is the highest
one of the suit led.

Following suit

In European trick-taking games, there is seldom any
constraint on what card may be led to a trick. There
are a few exceptions to this, but they are infrequent
and minor: usually, the player with the lead is legally
free to lead any card of his choosing from his hand. It
is very rare, however, for there to be no constraints on
the cards played by the other players, those who play

168 Part I: History and Mystery
to a trick to which the first card has already been led.
Among European games, those in which there is no
constraint at all are mostly two-handed games in
which each player draws a card from the stock after
each trick; and then, as soon as the stock is exhausted,
constraints come into force. The most familiar
constraint is that requiring each player to ‘follow suit’
if he can: that is, to play a card of the suit led if he has
one. Less usual, but far from uncommon, is the
additional requirement that any player who is able to
play a higher-ranking card of the suit led than any so
far played to the trick must do so. In games with
trumps, the constraints vary a great deal: but,
although there is much variation in detail, which
naturally has a profound effect upon the strategy of
the game, almost all European trick-taking games
leave the player with the lead free to’ play any card
that he chooses, and almost all circumscribe in one
way or another the freedom of the other players.


The presence of trumps in a trick-taking game
introduces an entirely new additional feature into the
determination of which player has won a given trick.
When such games are played with the regular pack,
some means or other is used to select one of the four
suits as the trump suit for a given round. When there
is bidding, the nomination of one of the suits as trumps
may form an ingredient of the bidding itself, as in
Bridge; an older practice, exemplified in Skat as now
played, is for the declarer - the player making the
highest bid - to announce the trump suit, at his
choice, after the bidding is completed. When there is
no bidding, the trump suit is determined
by chance: when-not all the cards are dealt out,
the first undealt card may be turned face up,
and its suit taken as the trump suit; when the
whole pack is dealt out, the last card dealt may be
turned up for the same purpose, or the pack may
be cut before the deal begins. The cards of the
trump suit behave in a quite different way from those
of the other suits (the ‘plain’ suits). Any card of the
trump suit, even the lowest, beats every card of a plain
suit, even the highest; and it does so irrespective of the
suit led to the trick. The rule governing which card
wins the trick thus takes the form that the trick is won
by the highest trump played to the trick, or, if the
trick contains no trump, then by the highest card of
the suit led.

In games played with the Tarot pack, there is no
need of any device for determining a suit as trumps,
for the twenty-one triumph cards always fulfil this
function, while the four ordinary suits are always
plain suits. The triumph cards are, however, trumps
in precisely the sense just given: that is to say, each
trick is won by the highest triumph card played to it,
or, if none is played, then by the highest card of the
suit led. For this-purpose, the triumph cards normally
rank in descending numerical order: that is, the
highest is the XXI, followed by the XX and so on
down in sequence to the I (the Bagatto or Pagat).
Within each suit, the ranking is in accordance with
the ancient principle observed in Italy and Spain.
That is to say, in every suit, the four highest-ranking
cards are the court cards: the King is the highest,
followed by the Queen, then the Cavalier and then the
Jack. In Swords and Batons, the numeral cards follow
in descending numerical order: the 10 ranks next
below the Jack, followed by the 9, and so on down to
the Ace, which is the lowest card of the suit. In Cups
and’ Coins, however, they rank in ascending
numerical order: in these suits, the Ace ranks next
below the Jack, followed by the 2, then the 3, and so
on, with the 10 as the lowest card. It is obligatory to
follow suit, if possible: This means that, if a card of a
plain suit is led, then every player who still has any
cards of that suit must play one to the trick; and, if a
triumph card is led to a trick, then likewise any player
who still has some triumphs in his hand must play
one. In Whist and Bridge, as in many other trick-
taking games, a player who has a void in the suit led
(has no cards in that suit), and who therefore cannot
follow suit, is at liberty, when it was a plain suit that
was led, either to play a trump or to ‘throw away’ a
card of another plain suit, thus denying himself the
chance to win the trick: ‘throwing away’ is, of course,
tactically important as a means of creating a void in
one’s hand, so as subsequently to be able to trump a
trick to which that suit is led, or simply to get rid of
losing cards. In Tarot, however, there is not this
opportunity: when a plain suit is led, any player
unable to follow suit has the obligation to play a
trump if he has one. This obligation holds good even if
the player’s partner has already won the trick, and
even if it means playing a lower trump than some
other player has already played. The reason for this is
that there are, in games played with the 78-card pack,
half as many trumps again as there are cards in any
one plain suit; in Minchiate, and in the many games
played with shortened packs, the proportion is even
higher. If a player with a large number of trumps were
allowed to hoard them to use on valuable tricks, he
would have too great an advantage. In games like
Whist and Bridge, the only way to force one’s
opponents to play their trumps is by leading trumps
oneself: in Tarot, it can alternatively, and more
economically, be done by. repeatedly leading a plain
suit in which they are short.

The Matto

This explains the role of the triumph cards in the
game. The role of the Fool or Matto is entirely
different. To speak more exactly, it was

The Game of Tarot 169
entirely different in the earlier of those forms of the
game of which we have records. In those forms of it
now played in’ Austria, Hungary, Slovenia,
Czechoslovakia and the Black Forest, it has been
transformed into the highest trump card, beating even
the XXI: but, previously, it had a quite distinct role,
which it retains in the forms of the game played in
Italy, France and Switzerland. In games of this kind,
the player holding the Matto can play it to any trick
he chooses: it has no trick-taking power, but it
releases the player who plays it from the obligation to
follow suit or to trump. He may thus use it to avoid |
losing a valuable card which would otherwise be
captured in that trick. There is some variation of rule
about whether it is permissible to lead the Matto as
the first card of a trick; in any case, it is usually
inadvisable. When it is allowed, the player next in
rotation is usually at liberty to play any card of his
choosing; and the card that he plays is treated as if it
has been led to the trick, so that the remaining player
or players have to follow suit to it if they can and, if
they cannot, to trump if they can.

The object of the game

In order fully to explain the role of the Matto, it is
necessary to say something about how the winner of a
round is determined. So far, we have merely said that
the winning player or side in a trick-taking game is
determined by the cards won by the various players in
tricks. Trick-taking games may be classified into
those, like Bridge and Ombre, in which what
determines which player or side has won a given
round is simply the total number of tricks won by
each, and those in which what matters is which
particular cards each player has won in tricks. There
is no serviceable accepted terminology to distinguish
these two types: I shall therefore call those of the
former type simple trick-taking games, and those of the
latter type complex ones. These are, of course, intended
merely as technical terms: no suggestion is being
made that Bridge, or Ombre for that matter, is an
easy game to play. In complex trick-taking games,
different cards carry different point values, according
to their rank; and the winner or loser is determined by
the total number of points each player has on the
cards he has won in tricks. Originally, the points won
by a player constituted his score for the round. In
many games, however, the points won on the cards
captured in tricks serve merely to determine which
player or side has won the round, and the scoring is
done on a different principle. In order, then, to avoid
confusion, it is necessary to keep the two kinds of
points carefully distinguished: we may therefore
speak of card points to mean those that go to
determine which player or side has won the
round, and of game points to mean those that
are actually’ marked on the score sheet (or
immediately paid in cash or chips). In most complex
trick-taking games, such as Skat and Jass, the
national card games of Germany and Switzerland,
only certain cards carry any point-values at all: the
others contribute nothing to the total of card points,
but are merely of. strategic significance, as
determining who wins the trick. Ia a few such games,
however, such as Manille, every card contributes to
the card-point total, though some make a much
higher contribution than others.

Trick-taking games may be classified in another
way as well, namely as positive or negative. In positive
games, the object is to win as many tricks, or as many
card points, as possible, or to win at least a certain
number. In negative games, the object is to win as few
tricks, or as few card points, as possible, or to win no
more than a certain number. In certain games with
bidding, the object may be positive or negative
according to the bid made, and thus vary from round
to round. Nevertheless, most trick-taking games can
be classified as primarily positive or primarily
negative. (1) It seems impossible to say whether simple
or complex trick-taking games are the older in
Europe: it may be that both are as old as card playing
itself. By contrast, it is clear that negative games, of
which the. best-known to-day is Hearts and its
variants, and the earliest really successful one was
Reversis, which originated in Spain in the sixteenth
century, are a later development: the positive games
came first, and subsequently the idea occurred of
playing a game with essentially the same mechanics,

but with the aim of the players reversed, (This is
evident enough from the very name of Reversis — in
Spanish Revesino — which was played in England
under the name of Losing Loadam, and is probably
the game referred to by Rabelais as Coquimbert, qui
gaigne perd

Card points in Tarot

In this terminology, Tarot is a positive complex trick-
taking game. It resembles Manille rather than such
games as Skat and Jass in that, every trick won
contributes to the card-point total, even if it contains
none of the high-scoring cards. The method usually
1There is a small modern category which escapes this
classification for a different reason. One example is the
simple trick-taking game of Oh, Hell, in which each player’s
object is to make exactly the number of tricks he declared
that he would, no less and no more; the excellent game
played with the Taotl pack, a modern invention with five
suits and various extra cards, works according to the same
principle. Likewise, in a modern member of the Jass family of
complex trick-taking games played in Switzerland,
Differenzlerjass, each player predicts how many points he
will make, and obtains a negative score consisting of the
difference, by defect or excess, between his actual and
predicted total.

170 Part I: History and Mystery
adopted for computing the total of card points won by
each player is rather complicated, and will be
described later. For the present, I give here a
simplified description, which probably corresponds
rather well to the original method: it must be
understood that some particular forms of Tarot
diverge very considerably from the methods here
described. In this simplified method, each player
scores 1 card point for every trick he wins. In addition,
he scores 4 card points for each King he wins in tricks
(whether originally played by himself or another), 3
points for each Queen, 2 points for each Cavalier and
1 point for each Jack. He also scores 4 points for the
highest trump (triumph XXI), 4 points for the lowest
trump (triumph I, the Bagatto) and 4 points for the
Matto or Fool. The player holding the XXI cannot
lose it, since it is the highest card in the pack. The
player holding the Bagatto, on the other hand, can
very easily lose it, since it can be beaten by any other
trump: so trying to bring the Bagatto home, if you
hold it, and trying to capture it, if you do not, are
important subsidiary objects of the game. Now, as
remarked, the Matto has no trick-taking power. But,
while it cannot take a trick, it cannot normally be
captured either. When the player who holds the
Matto plays it to a trick, it is not taken away by the
winner of the trick: the player who originally held it
takes it back and puts it with the cards he has won.
Usually, he is required to give another card in
exchange for it from among the cards he has already
won; he will, naturally, select a low card (a card
without special point-value). If he has not yet won any
tricks, he waits until he does in order to give a card in
exchange for the Matto to the player who won the
trick to which he ‘played it. But (whether the rule
about offering a card in exchange is in force or not) if
the player who played the Matto never wins any tricks
at all, he must, at the end of the round, surrender the
Matto to the player who won the trick to which he
played it. Normally speaking, therefore, a player
holding the Matto can count on gaining 4 card points
with it. The rule governing the Matto is usually
summarised by saying that it can neither capture nor
be captured: it would, of course, have appeared
senseless to state the rule that it is not captured before
the system of card points had been explained. Because,
by playing the Matto, a player is released from the
obligation to follow suit or to trump, the card is
sometimes known in Italian as the Scusa, and
frequently, in French, as the Excuse. This term was
corrupted in German into Skis or Skys, and it is as der
Skis that it is now regularly referred to in German,
despite the fact that, in the games played in Austria
and Baden, it has long lost the role that has been here
described, and has, in effect, been converted into
trump XXII.

Now, it may be said, this tells us how the
triumph cards and the Matto were used in play:
but it fails to explain why the Tarot pack was
invented in the first place. If, in the plain suits, it
is the court cards that are going to carry high
point-values, we can readily understand the
addition of a fourth court card: but that, as we
have seen, was by no means unique to the Tarot
pack, being a frequent feature of fifteenth-
century German packs, particularly of the
costlier sort, hand-painted or copper-engraved.
The special role of the Matto in play also made
this a comprehensible addition to the pack:
although there appears to be no actual
connection between the Matto and the later
Joker, and although the function of the two cards
is different, we may naturally compare the
invention of the one with that of the other. But
that does, not account for the principal
distinguishing feature of the Tarot pack, the
sequence of the twenty-one triumph cards. If, in
a game played with the regular pack, the trump
suit is determined by chance at the beginning of
each round, it would clearly make no essential
difference to the game if, instead, some suit were
designated a permanent trump suit, as indeed
happens in certain games. Thus, when there is no
bidding, there is no disadvantage in having
special cards as permanent trumps; but,
conversely, there appears to be no particular
advantage in it either, since it is so easy to confer
that role on one of the suits by convention. It is
true that, in Tarot, there are in effect five suits —
the trumps in addition to the four plain suits: but
five-suited packs were perfectly well known in the
fifteenth century, at least in Germany, and it
would seem much easier to introduce an
additional suit than to devise a special set of
picture cards. It is also true that the sequence of
triumphs is longer than any of the suits proper.
This, if desired, would. have been harder to
achieve by adding extra cards to one suit of a
regular pack: numeral cards higher than 10
would be very difficult to identify without
indices, and the addition of any large number of
court cards takes us precisely in the direction of
the Tarot triumphs. However, the effect can be
attained by convention, namely by removing
certain cards from the suits to which they
properly belong, and treating them as permanent
trumps; this is a feature of several games,
including Ombre, Schafkopf and Skat. The role
of the triumph cards as permanent trumps in the
games pla¥ed with the Tarot pack simply does

The Game of Tarot 171
not, at first sight, provide an adequate motive for
the invention and production of a special pack of
cards not readily adapted to the playing of games
of other kinds.

The puzzle is solved once we drop the
assumption that, at the time when the Tarot
pack was first devised, the idea of trick-taking
games with trumps was already familiar. Our
difficulties were caused by taking it for granted
that card players of the time were already
acquainted with games played with the regular
pack, in which some one suit would be
designated as trumps, permanently or for a
round at a time: we then could not see why
anyone should go to the trouble of inventing a
new form of pack in which a quite special set of
cards were to serve as trumps. But, if we assume
that the idea of trumps was not already familiar,
the aspect of the matter is quite altered. In that
case, the invention of the Tarot pack must have,
at the same time, constituted the introduction of
the idea of trumps into trick-taking games, one of
the great inventions in the history of card play;
and the question, ‘Why bother with a special set
of picture cards when one of the ordinary suits
would do?’, loses most of its force. We must, of
course, assume that, when the Tarot pack was
invented, trick-taking games in general, and,
more specifically, positive complex trick-taking
games, were already well known. In games, as in
all other fields, human inventiveness usually
proceeds by a step at a time; we cannot expect
more than one, or at the most two, new ideas
from the same source simultaneously. The game
of Tarot cannot possibly have been the first trick-
taking game known in Europe: the first
realisation of the fundamental conception of
trick-taking play must have assumed a very
straightforward form, without the complication
of trumps or that of the Matto. Indeed, we know
for certain that trick-taking games are older in
Europe than the Tarot pack, since the German
game of Karnöffel was such a game, and is
referred to as early as 1426; and it had
complicated rules, under which the natural
ranking of the cards is violently disturbed, so that
it is as incapable of having been the earliest
known trick-taking game as is Tarot itself.

For similar reasons, it is not thinkable, either,
that Tarot was the first known complex trick-
taking game, that is, the first in which different
point-values were attached to the various cards.
On our hypothesis, the invention of the Tarot
pack was the invention of the idea of trumps. It is
not quite certain that the Matto was an original
member of the pack; it may have been added at a
slightly later stage, when the form of pack
represented by the Visconti di Modrone one was
replaced by what became the standard 78-card
form. There is also one piece of evidence to
suggest that the classic manner of using the
Matto may have been a later development, and
that it was originally the lowest trump.2 But,
whether or not the invention of the Tarot pack
represented the introduction of the Matto in its
classic role, one radically new and fundamental
idea, that of trumps, is as much as we can
suppose to have been introduced at one time; it is
implausible to suggest that the game of Tarot
simultaneously embodied the innovation of
assigning to the cards not merely different trick-
taking powers but different point-values. It is
true that we do not have such clear positive
evidence to the contrary as in the case of the
trick-taking principle itself: Karnöffel was, in our
technical terminology, a simple trick-taking
game, and no game can be cited that is known
both to have been older than Tarot and to have
been a complex game in our technical sense.
There are, however, several complex trick-taking
games played with the regular pack in which the
point-values obey the same general principle as
that displayed by the court cards in the plain
suits in Tarot: that is to say, the high-scoring
cards are precisely those with the greatest trick-
taking power, and these carry point-values which
diminish by 1 point as one descends through the
sequence. Among such games is the 'sixteenth-
century Italian game of Trappola, in which the
highest-ranking cards in each suit are, in
descending order, Ace, King, Cavalier and Jack,
these being also the high-scoring cards, carrying
point-values of 6, 5, 4 and 3 card points
respectively. Other examples are provided by the
Spanish game of Malilla, sometimes asserted to
be an invention of the nineteenth century, but
demonstrably as old as 1776,3 and quite possibly
2 The evidence is from the Invettiva of Lollio: see Chapter

3 Gégé, Historique ei régle du jeu de la Manille, Enghieri-les-
Bains, 1883, maintains that the game originated only in
1865, and spread from the South of France to the rest of the
country only after the Franco-Prussian War of 1870. This
last remark may well be correct, but, for the rest, Gégé is
completely confused. He fails to distinguish between French
Manille and Spanish Malilla, apparently not realising that

172 Part I: History and Mystery
far older, together with its French and Dutch
descendants Manille and Manilla, and the rather
similar Spanish game of Solo (not to be confused
with several other games of the same name). In
all these last four games, the five highest-ranking
cards of each suit carry, from the highest to the
lowest, point-values of 5, 4, 3, 2 and 1: in Malilla,
these cards are, in descending order, the 9, Ace,
King, Cavalier and Jack; in Manille they are the
10, Ace, King, Queen and Jack; in Manilla, the
Ace, 10, King, Queen and Jack; and in Solo, the
7, Ace, King, Cavalier and Jack. Yet another
example is the negative complex game of
Revesino or Reversis, originating in Spain no
later than the early sixteenth century, in which
the highest cards of each suit are Ace, King,
Cavalier or Queen and Jack, bearing the point-
values 4, 3, 2. and 1. A not very well-known game,
played both in France and Germany under the
names Quarante des Rois [/i]and [Vierzig vom
König, has the King, Queen and Jack as the
highest cards of each suit, carrying point-values
of 5, 4 and 3 respectively. Another game of this
kind is Fünfzehnern, popular in Germany
throughout the nineteenth century, in which the
top five cards of each suit were Deuce (Ace),
King, Ober, Unter and 10, once more with point-
values of 5, 4, 3, 2 and 1. A long obsolete member
of the genus is the Alsatian game of Trümpfspiel
described by Daniel Martin in 1637, in which the
top cards of each suit were Ace, King, Queen and
Jack, with point-values of 4, 3, 2 and 1
the Spanish pack lacks 10s, and describes Manille as being
popular in Catalonia, Navarre and Andorra as well as the
South of France. Having discovered that the game played in
France under the name of Manille in the eighteenth-century
was an entirely different one, namely a Stops game, later
called Cométe, he quite mistakenly takes early Spanish
references to Malilla to be alluding to the Stops game. An
anonymous pamphlet of 1776, Le Jeu de la Malille, perhaps
published in Amsterdam, informs the reader that the game
has long been popular in Spain, and hopes that it may be
introduced into France. It describes the Spanish game, as
played with the 48-card pack, and the author states that he
is abridging the earliest written account, written by a Don
Antonio, whose date he does not give. Malilla therefore
probably dates from the first half of the eighteenth century
or earlier, while Manille may be an adaptation from the
early nineteenth century.

4 Daniel Martin, Parlement nouveau, Strasbourg, 1637, ch,
74, p. 635. This is a French/German reader, with parallel
texts on a variety of subjects, intended for those learning one
or other language. I have not seen the original, but the
French text of the. entire book is reprinted in Collection

None of these games shows any sign of being
directly derived from or influenced by Tarot: the
mark of such an influence is a special value or
role for the lowest trump. Indeed, Fünfzehnern
and Reversis are played without any trump suit,
and the same appears to have been true of
Trappola in its original form. (Fünfzehnern has,
indeed, a curious affinity to the Indo-Persian
Ganjifa games, which is presumably a result of
coincidence.) Probably none of these games is as
old as Tarot; but they collectively bear witness to
what is very likely to have been a principle of
assigning point-values in complex trick-taking
games that antedated the invention of the Tarot
pack, and may go back to the introduction of
card games into Europe. This very
straightforward principle is that underlying the
assignment of point-values to suit cards in Tarot:
that governing their assignment to the additional
cards of the Tarot pack is considerably more
subtle. It seems natural to suppose that the game
of Tarot was founded upon some existing game
or games played with the regular pack and
without trumps, in which the high-scoring cards
in each suit were just the court cards.

Now, under the supposition that the Tarot
pack was devised in order to embody what was
then a completely new idea in trick-taking play,
that of trumps, its invention makes perfect sense.
Trump cards play a completely different role in a
trick-taking game from cards of the plain suits:
they can win a trick to which a different suit was
led, and they beat all other cards regardless of
their rank. If card players had long been
accustomed to trick-taking games played without
trumps, in which the principle held without
exception that a‘trick could be won only by a
card of the suit led, then, when the idea of trumps
first occurred to some ingenious individual, they
would as much be thought of as cards fulfilling a
completely new and special function as would the
Matto, when that was first invented. It would
thus be entirely natural to realise the new idea,
not by adopting a convention under which one
of the suits behaved quite differently from the
others, but by adding a new sequence of cards to
the pack. There would have been no apparent
reason why this new sequence should have the
same structure, or the same length, as the
existing suits; nor would it seem natural to assign
a new suit-symbol to it, as if it were equivalent to
the ordinary suits. These would be cards
intended to play an entirely new role in the game,

The Game of Tarot 173
and therefore naturally represented by designs of
a new kind: like the important cards of the four
suits, they would be picture cards, but they
would no more need a suit-sign than would the
Matto, or the Joker in a modern pack.

If this hypothesis is correct, the invention of
the Tarot pack was one of the great moments in
the history of card play. One of the reasons why a
serious history of card games has never been
written is that people tend not to take card games
seriously, in the way that they are prepared to
take chess. But another reason is the complexity
of the subject: card games proliferate far more
rapidly than do board games. When such a
history comes to be written, it will therefore have
to concentrate, not so much on the detailed
evolution, spread and fortunes of individual
games, as on their constituent ideas. New card
games are constantly being invented, and old
ones modified: but it is much more rarely that
anyone comes up with a radically new principle
of play. The evolution of a card game is primarily
the history of its adopting or adapting principles
already practiced in other games; the invention of
new card games primarily a matter of a new
combination of existing ideas.5 In he course of
this process, the new or modified games present
tactical problems that may not be able to be
exactly paralleled in older games; even an
alteration in the method of scoring may radically
affect strategy, as in the change from Auction to
Contract Bridge. Nevertheless, human inventive
genius is best displayed, in the field of games,
when someone devises a fundamentally new
principle of play; and, in the history of card
games, that has not happened often. If we restrict
ourselves to trick-taking games, then, since the
principle of trick-taking play was first invented,
only two outstanding new principles have been
alsacienne, Dantel Martin ou la vie & Strasbourg au commencement
du XVIII siécle
, ed. Charles Nerlinger, being fascicule 49 of
the Publications de la Faculté des Lettres de (Université de
Strasbourg, Strasbourg, 1900, the section ‘Du Ioüeur de
cartes’ being on pp. 216-18, while the German text of the
account of Trümpspiel is quoted in Heinrich A. Rausch,
Das Spielverzeichnis, im 25. Kapitel von Fischarts
, Strasbourg, 1908, pp. xxxvii-xxxviii.

5 Alain Borvo, Anatomie d’un jeu de cartes: 1’Aluete ou le Jeu
de Vache
, Nantes, 1977, p. 18, cites Pierre Berloquin as
advancing the same conception of a card game as a
compound of constituent principles of play, each of which
may be found in many other games, and as coining the term
ludémes for these constituent principles.

introduced: the idea of trumps, and the idea
of bidding.

We are able to say precisely where the idea of
bidding comes from: it comes from the Spanish
game known in Spain as Tresillo, and
everywhere else as Ombre or l’Hombre. This
game is comparatively little played to-day: but,
in its time, it was the most successful card game
ever invented, enjoying, at rather different dates,
an enormous vogue in France, in England, in
Germany and in Italy; it was carried by
Portuguese sailors to Indonesia and Japan. It
occupied the same position as Bridge does to-
day, as the fashionable card game and as, par
excellence, the intellectual card game: and it
occupied it for a far longer period than that for
which Contract Bridge has yet existed. Its history
in France can be traced through the series of
card-game books published every few years in
that country from the mid-seventeenth century
onwards. From looking at these books, it
becomes apparent that, at any given time, a card
game may be in a stable state or in a nova state.
A game in a stable state is usually:accorded only
a brief description, perhaps incorporating a few
hints on strategy: and this description will, very
often, go on being reprinted, perhaps without any
verbal changes, for many decades. These are the
games played in the family circle, or by the
peasantry or the bourgeoisie. Their devotees are
likely to be conservative in their habits of play,
and not much given to verbal analysis of strategy,
although they may be highly skilled in play. A
game goes into a nova state, on the other hand,
when it is taken up by the Haut monde: and the
manifestation of this in the card-game books is
extremely striking. As the game becomes more
and more fashionable, the accounts given in the
books become longer and longer: there are
detailed discussions of strategy and analyses of
sample hands. This is partly because the game
has caught the attention of intellectuals and
people given to making explicit analysis, and
partly because proficiency at the game has
become a requirement for social success in a
milieu that prizes such success highly. At the
same time, the game tends to undergo rapid
evolution, as it is modified with the intention of
making it more entertaining, more skilful or
more varied; and it may also bud off a whole
progeny of variant forms. All this is exemplified
by Ombre, which was the greatest nova ever
seen, and, at the height of its popularity in

174 Part I: History and Mystery
France, was occupying about half the space in
the card-game books, at least if its variants are
reckoned in.

It is easy to understand the reason for the
enormous popularity of Ombre: it introduced
card players to an entirely new and exciting
principle, that of bidding. There can be no doubt
that it was from Ombre that this principle
entered card play: not only is it the earliest game
to have involved bidding, but, as soon as the

principle of bidding was adapted to many other
trick-taking games, Tarot included, which had
not formerly incorporated it; and the form which
bidding first took in these games was patently
borrowed from the type of bidding used in
Ombre, often not very well adapted to the games
for which it was being borrowed. Ombre needed
bidding because it was a three-handed game in
which, in each round, one player, - the hombre or
man - played on his own against the other two in
temporary partnership. Until the invention of
Ombre, the most usual ways of playing trick-
taking games were with four players, forming two
fixed partnerships, as in Whist, or with three
players, each playing for himself. Now, obviously,
Ombre would have made a very poor game if,
say, the player to the right of the dealer had
always to play on his own against the other two:
he would often have been resoundingly defeated,
and the game would have been boring. Hence
each player in turn had an opportunity to say
whether he wished to undertake the role of the
hombre or solo player in that round. This was, in
origin, the simplest form of bidding imaginable:
a player could either pass or elect to play on his
own, and, as soon as anyone so elected, the
bidding stopped. However, the principle once
introduced, it was natural to make it more
interesting by allowing different bids to be made,
so that a player who had made a positive bid
could be overbid. As time went by, an ever more
elaborate code of possible bids was devised. It is
worth noting, however, that the different bids did
not relate to what the bidder committed himself
to doing during the actual play. Once actual play
began, the task of the declarer — the hombre — was
always the same: he had to win more tricks than
did either one of his opponents, taken separately.
The bids related, instead, to the preliminaries
preceding the actual play: to whether or not the
declarer could discard worthless cards from his
hand and replenish it from the stock of undealt
cards, to how the trump suit was to be
determined, and so on. Thus what the declarer
had to do in order to win was fixed: the bids
affected only the conditions under which he tried
to do it. The idea of variable contracts — different
things that the declarer committed himself to
doing during actual play, if he was to count as
winning — which figures so prominently in
Bridge, did not make its appearance until Boston
Whist came on to the scene at the end of the
eighteenth century. It was even later that the
practice of bidding was adopted in four-handed
games with fixed partnerships, like Bridge:
Boston Whist, whose simplified modern
descendant is Solo Whist, is the only branch of
the Whist family in which the principle of fixed
partnerships is abandoned. For a long time, the
principal point of bidding was taken to be as a
means of deciding, in four-handed as well as in
three- or five-handed games, which player or
players should, in any one round, play against
the rest in temporary partnership. Indeed, so
long as the object of actual play, the way of
deciding which side had won, remained constant,
there could be no point in bidding in a game with
fixed partnerships. Thus, surprising as it may
seem to those who associated bidding primarily
with Contract Bridge, trick-taking games with
fixed partnerships were the last to adopt the
principle of bidding introduced by Ombre; and
they could do so only after that principle had
been modified in Boston Whist by admitting
variable contracts.

The idea of bidding is thus a European, and,
specifically, a Spanish invention. According to
our present hypothesis, the other leading idea,
that of trumps, was invented in Italy in about
1440, with the invention of the Tarot pack. In
any case, trumps, like bidding, were certainly a
European invention: trumps are unknown in
Oriental card games that do not derive from
European ones. By contrast, the basic idea of
trick-taking play itself is not of European origin.

As we have seen, even when we ignore the use of
trumps, trick-taking play incorporates several
distinguishable elements that are nevertheless
always, or almost always, associated. In Chapter
3 we imagined different conventions determining
which player should lead to each trick. One
was that the lead might simply rotate from one
trick to the next, irrespective of who won the
trick: this would parallel the invariable custom in
European games concerning the deal for each

The Game of Tarot 175
round. Another was that the lead might remain
with one player so long as he went on winning
tricks, but then, as soon as he lost one, :pass to
the next player in rotation, regardless of who
won the trick: this would be the analogue of the
usual Chinese principle for selecting the dealer.
There is no reason to think that games played
in accordance with either of these rules would
be strategically trivial: yet no such rule is to be
found in any trick-taking game, European or
Oriental; with negligibly rare exceptions, the
winner of a trick always leads to the next one. We
noted, also, the possibility, realised only in a tiny
minority of unusual games, of alternative rules
for determining by which card a trick is won: as I
remarked before, these exceptions underline how
predominant is the principle that (trumps_
excepted) only a card of the suit led can win the
trick. These facts strongly suggest that trick-
taking play has a single origin: that all, or nearly
all, trick-taking games have evolved from some
one common ancestor.

This prototype of trick-taking games, of which
Bridge, Tarot, Ombre, Skat and all the rest are
the descendants, must have been created before
playing cards ever reached Europe. One of the
leading uses for playing cards has, of course,
always been as instruments of pure gambling
games like Baccarat, Faro and Thirty-One,
where the cards become a mere randomising
device like dice or the roulette wheel. But it can
hardly have been in order to play such games
that playing cards were invented in the first
instance, since they often ignore suit altogether,
or fail to discriminate between the different ranks
of court cards, and, in any case, dice are
incomparably older than cards and already
served perfectly as a randomising device; the
pure gambling games do not exploit the true
essence of playing cards as an instrument of play.
If we set them aside, and consider only those
games that require at least some strategic skill,
we find that European card play has always been
dominated by trick-taking games. If you look at
any card-game book — French, English, German,
Italian, Dutch or any other — published in
whatever century, you will find that at least three-
quarters of the games described are trick-taking
ones. Almost all the card games that have
enjoyed a great vogue at one time or another —
almost all those that have entered or passed
through a nova state — have been trick-taking
games; and the same is true, with a very few
exceptions like Scopa in Italy and Cribbage in
England, of almost all those that have had a
widespread popular following. Apart from the
pure gambling games, perhaps 75 per cent of the
man-hours spent in playing cards have in Europe
been expended upon trick-taking games. Despite
the great success of Bridge, this is less likely to be
true of the twentieth century than of any period
in the past, because of the popularity in modern
times both of Poker and of the many forms of
Rummy and its offspring Canasta: but, taken
over the whole history of card games, it would be
a fairly safe bet.

This makes it highly probable that trick-taking
play was not independently invented in Europe,
but entered, very likely in various forms, with
playing cards themselves. Once it is accepted
that playing cards entered Europe from outside,
it is not to be supposed that they would have
entered merely as instruments of possible games
still to be thought up, any more than the
chessboard and chess pieces would have spread
without being accompanied by information
about the game played with them; playing cards
would have come into Europe together with a
game, or, more likely, many different games, that
they could be used to play. In view of the
subsequent predominance of trick-taking games,
it is surely to be presumed that such games were
among those that entered Europe along with
playing cards themselves.

We know, moreover, that by 1426 the game of
Karnöffel was already being played in Germany.
This was half a century after the first
introduction of playing cards: but, as already
remarked, it subjected the cards to an
extraordinary rearrangement of their normal
order; and, as we shall see, it overlaid the
basic principle that only a card of the suit led can
win the trick with many and complicated
exceptions. For these reasons, it cannot have
been invented before the fundamental idea of
trick-taking play had already long been familiar:
its existence presupposes general acquaintance
with that idea.

We know very little about the card games
played in Europe in the fourteenth and fifteenth
centuries. We do know, however, that there was a
clearly accepted notion of the order of the cards
within each suit: this is already apparent in John
of Rheinfelden’s treatise, when he speaks of the
two ‘marshals’ as being under the Kings. A little
reflection shows that this conception of itself

176 Part I: History and Mystery
implies the existence of trick-taking games. It is
true that, in some games of other kinds, we need
the idea of a cyclic order: for example, in any
game in which a significance is attached to
having three or more cards in sequence. But it is
only, or at least primarily, in trick-taking
games that we need an idea of the direction of
the ordering, that is, of one card’s being higher or
lower than another. It might be objected that,
when two sequences of equal length, held by
different players, have to be compared; this can be
done only by taking that sequence to beat the
other which is headed by the higher-ranking card.
We have seen, however, that, until about the end
of the fifteenth century in Italy, and until well
into the sixteenth century in Spain, it was the
normal practice. to treat the ranking of the
numeral cards as running in one direction in
Swords and Batons (the higher numeral beating
the lower), and in the other direction in Cups
and Coins (the lower numeral beating the
higher). This practice shows that what was
required was to compare the ranks of any two
cards of the same suit, but not those of cards of
different suits: it is easy to remember that the 4
of Cups beats the 7 of Cups, and that the 6 of
Swords beat the 3 of Swords; but, as remarked,
it is not so easy to remember straight off that the 4
of Cups has a higher absolute rank than the 6 of
Swords, and that therefore the sequence 4-6 of
Cups is superior to the sequence 4-6 of Swords.
The need for a ranking within each suit, but not
as between cards of the different suits, irresistibly
suggests trick-taking play. It is true that we have
no evidence of the practice of reversing the order
in different suits in France, Germany or
Switzerland (John of Rheinfelden’s text is rather
ambiguous, but suggests that the Ace was the
lowest card of every suit); but, since it is likely
that it was from Italy that playing cards spread
to the rest of Europe, trick-taking games, if they
had arrived together with the playing-card pack,
would have followed wherever it spread.

As we have seen, this practice of reversing the
order of the numeral cards in half the suits is
observed, without exception, in the trick-taking
games played in India under the general name of
Ganjifa. This unquestionably establishes a link
between the Indian games and those of Europe;
and trick-taking games hold an even more
predominant position among Indian than among
European card games. As we have seen, the
connection between Indian and European card
games cannot have been direct. Direct contact
between Europe and India was made for the first
time since the rise of Islam by Vasco da Gama’s
arrival in 1498. By then, both trick-taking games
and the practice of reversing the order of the
numeral cards in two of the suits had
demonstrably long been known in Europe, and
hence cannot have been learned from Indian card
players. By the early sixteenth century at the
latest, and possibly two decades before the end of
the fifteenth, the use of trumps in games played
with the regular pack was being practised by
European players, whereas, in the Ganjifa
games, there is neither a trump suit nor bidding.
Furthermore, as we noted before, although the
Ganjifa games are full-fledged trick-taking
games, with a trick being won by the highest card
of the suit led and the winner of a trick
(almost always) leading to the next one, they
exhibit a number of radical differences from the
trick-taking games played in Europe. There are
severe and often complicated constraints upon
the leads that may be made; there is the basic
principle that the holder of the’ highest
outstanding card of any suit must play it when
that suit is led; there is the general rule that the
lead to the first trick belongs to the player
holding the King of a particular suit (usually
varying according to whether the game is played
in the daytime or in the night-time); and there is
the frequent practice of playing multiple tricks —
the player with the lead puts down several cards
simultaneously, and the other players have each
to put down an equal number of cards. The
Indian games and the European ones must have
developed from some common ancestor; neither
can directly have influenced the other.

It is highly probable, as we saw, that Indian
playing cards, and the Indian Ganjifa games,
were derived from those of Persia, where the
eight-suited pack, associated in India with the
Moghuls was in use at least by the sixteenth
century, and where similar and perhaps identical
games were being played at least by the
seventeenth century. There was, of course, close
cultural contact between Persia and the Moghul
court: Mirza Sadiq, who was born in 1018 AH.
(1610 A.D.) in Isfahan and died in 1061 A.H. (1653
A.D.) in India, devoted the 64th chapter of his
Encyclopaedia to ganjifeh, describing various
games including Hamrang, known from later

The Game of Tarot 177
Indian sources.6 The eight-suited pack, and
probably also the Perso-Indian type of game,
date back to the early sixteenth century. Further
back we cannot at present go: but we saw that
etymological evidence makes it likely that
playing cards and card games reached Mamlük
Egypt from Persia, and that the 52-card Mamlük
pack and the 96-card Persian and Moghul pack
probably have as their common ancestor a 48-
card pack with four suits and two court cards per
suit. The special type of game played in Persia
and India therefore presumably developed in
Persia at some date after playing cards first
reached Egypt. This early history is, on present
evidence, highly conjectural: but, if the re-
construction of history suggested here is correct
in outline, the occurrence in both India and
Europe of the practice of reversing the order of
the numeral cards in half the suits can be
explained only by supposing it to have been
followed before the eight-suited pack was
invented, by players in Mamlük Egypt as well as
in Persia, and to have entered Europe along with
trick-taking games and with playing cards
themselves. We have already noted that the
disposal of the upper inscriptions on the Istanbul
cards gives partial support to this hypothesis.
The predominance of trick-taking games in
Europe and India is not paralleled in China.
There, as we saw, setting aside pure gambling
games as before, two quite different categories of
card game have predominated: draw-and-
discard games and what we called ‘fishing’
games. Draw-and-discard games were virtually
unknown in Europe until this century, when
Rummy games were imported from the United
States, and, subsequently, Canasta from South
America, and achieved great popularity. Fishing
games are played in Europe in a different manner
from that customary in China and Japan: there is
no turning over the top card from the stock after
playing a card from the hand, but, instead, there
are repeated deals in the course of play;
moreover, it is usually allowable to capture more
than one card from the pool by the play of a
single card, which never happens in the Chinese
° The Encyclopaedia is called Shahid-i Sadiq, and its ch. 64
is discussed by Ali Naqi Manzavi in 'Ganjafeh-ye Shahid-i
Sadiq’, Danesh, no. 8, October 1954, pp. 459-60. Hamrang is
described in ‘A note on Ganjpa’, by the Maharaja of
Sonepur, Journal of the Bihar and Orissa Research Society, vol.
10, 1924, pp. 221-6.

and Japanese games. In England the best-known
fishing game is Casino; by far the most successful
of all European fishing games is the Italian game
of Scopa, with its derivatives. Fishing games have
not, however, played any great role in the history
of European card games; and, though I know
little of their history, Scopa appears to have been
already well known in Naples by the early
eighteenth century, but may possibly have arrived
from the Middle East.

Although there are few Chinese trick-taking
games, one, the domino game T‘ien chiu
(‘Heavens and Nines’, so called from the name of
the highest domino of the civil suit, the double six
or Heaven, and of the highest dominoes of the
military suit, the Nines, i.e. the 6-3 and the 5-4),
is very celebrated and very ancient. The primary
object of the game is to win the last trick; but
how much is paid by the other players to the
winner depends on how many tricks they have
won. There are a few European trick-taking
games in which the object is to win the last trick,
for example the Swedish Femkort and the
German Letzter Stich; and it is a persistent
feature of many games, going back at least to
Trappola and some forms of Tarot in early
sixteenth-century Italy, that the winner of the
last trick gains a bonus. In the present state of
knowledge, we can only speculate whether T‘ien
is a remote ancestor of the trick-taking games
played in Persia, India and Europe, or whether
the principle of trick-taking play was invented
twice independently.

Whatever the exact truth on these questions
may be, it seems certain that trick-taking
games are older than the use of playing cards in
Europe, and perhaps as old as the four-suited
pack itself; and, since trumps are an unknown
phenomenon in any game not demonstrably
derivable from a European forebear, we may
confidently assert that they, like bidding and
unlike the general trick-taking principle,
represent a European contribution to the art of
card games. As for the idea of complex trick-
taking games, it is impossible to be certain. No
known trick-taking game of non-European origin
is complex in our technical sense, although it is 4
regular feature of fishing games that different
cards captured bear different point-values; we
might, therefore, conclude that the trick-taking
games that entered Europe along with playing
cards were all simple ones, and that the idea of

178 Part I: History and Mystery
assigning different point-values to cards
according to their rank first arose in Europe. On
the other hand, it is highly tempting to associate
the distinction between high-scoring cards and
low cards with that between court cards and
numeral cards. Familiar as we are with the latter
distinction, we forget how curious it is. But, when
we reflect, when, say, we recall the Eton platoon,
we are impelled to ask why anyone should have
thought of continuing a numerical sequence with
three social or political ranks. It would be a-
satisfying solution if we could argue as follows. A
difference in the type of design which a card
displays is likely to point to a difference in its role
in the games for which it was originally devised;
and an elaborate design is likely to point to an
important role. In a simple trick-taking game,
and, indeed, in most card games, the court cards,
even when they head the suit, differ from the
numeral cards only in respect of their higher
rank: it would be in no way inappropriate to
their function if they were replaced by numerals
from 11 to 13. It is only in complex trick-taking
games that the higher cards of each suit differ in
role from the lower ones otherwise than just by
being higher. If, therefore, we hypothesise a
prototypical complex trick-taking game in which
the court cards were the highest cards in each
suit and the only cards with high point-values, as
in fact they are in Quarante des Rois, we have an
explanation for the differentiation between court
cards and numeral cards: court cards were given
special designs because they played a special role
in the game. This idea is enticing; unfortunately,
what little evidence there is tells against it. If that
were ever the rationale for the distinction
between court cards and numeral cards, it had
been completely forgotten by the time that the
eight-suit pack and the Indo-Persian Ganjifa
games played with it were invented.

Furthermore, we have seen it to be probable that
there were originally only two court cards per
suit, and possibly, in the first instance, only one:
and a complex game with only two high-scoring
cards out of twelve, and no trumps, would be
distinctly tame. We have, therefore, reluctantly
to abandon this attractive idea. Complex trick-
taking games may have entered Europe together
with playing cards, or they may have been
invented in Europe at an early date, perhaps
prompted by the distinction between court and
numeral cards; but, in either case, they probably
do not have an antiquity equal to that of simple
trick-taking games. This conclusion tends to
reinforce the hypothesis that, in the beginning,
the King was the only court card, acquiring first
one subordinate and then a second: it would be
less bizarre, at the time when the four-suited
pack was first devised, to head every suit with a
single figure card than with a sequence of them.

We may regard it as certain that trumps are a
European invention; it does not follow that it was
with the game of Tarot that they made their first
appearance. Is there any positive evidence for
this hypothesis, apart from the plausible
explanation that it provides of why the Tarot
pack was invented at all? Here etymology proves
once more a clue. As has been said, the name
used in Italy throughout the fifteenth century,
right down to 1500, for Tarot cards and for the
game played with them is trionfi or triumphi. In the
sixteenth century, however, from at least 1516,
this name was replaced by the word tarocchi,
whose etymology is unknown and already
perplexed people in the sixteenth century.7 From
early in the same century, however, we find
names cognate with ‘triumph’ being used, in
several countries, for games played with the
regular pack. In France there is Triomphe, a
game still played, or at least described in the
card-game books, to this day.8 In England there
is Triumph, first referred to in 1522, and next in
a sermon by Latimer preached in 1529.9 In Spain
there was Triunfo, of which the first mention was

7 Lollio, in his Invettiva of 1550, spoke of ‘quel nome
fantastico, e bizarro/Di Tarocco, senz’ethimologia’ (that
fantastic and bizarre name of ‘Tarocco’, devoid of

8 The earliest French reference which is certainly to this
game, and not to Tarot, is, as so often, in Rabelais’s list of
the games of Gargantua, Gargantua, 1534, book I, ch. 22;
Rabelais spells it triumphe, and includes tarau separately in
the list.

9 The earlier reference is from Henry Watson; The chirche
of the euyll men and women/ whereof Lucyfer is the heed/ and the
membres is all the players dyssolute and synners reproued
. This was
a translation of La Petite Dyablerie dont Lucifer est le chef by
Thomas Varnet and Noel Beda which had previously been
published in Paris, itself an abridged and adapted
translation of the sermon De alearum ludo by St Bernardine of
Siena. At the end of The chirche of the ewyll men and women the
date is given as 22 August 1511, though it is not quite clear
to me whether this is meant to be the date of the translation
or of the work translated. The British Library gives the date
of La Petite Dyablerie in its catalogue as ‘1520’, and that of
Henry Watson's translation, of which it has a microfilm of a
copy in the Cambridge University Library, as 1522. There is
another, anonymous, translation, differing only in trifling
variations in wording, entitled The churche of yuell men &

The Game of Tarot 179
by J.L. Vives in 1539.10 The German game of
Trumpffen, perhaps the same as the Trümpfspiel
described by Daniel Martin in 1637, appears in
the list of games given by Fischart in his
translation of Rabelais in 1575.11 As for Italy

women| wherof Lucyfere is heed. And the membres is all the players
dissolute/ and synners reproued.
The first page of this says that it
has been ‘traslated out of frenche in to Englisshe/ at the
instaüce of Charles erle of Worcester/ and chaberlayne to
our soueraygne lorde the Kyng’. Charles Somerset, who was
Chamberlain to the Household of Henry VIII, was created
Earl of Worcester in 1514, and died in 1526. Both the
French and the English versions substitute different games
for those mentioned by St Bernardine. On p. 10 of Watson’s
book there occurs the passage ‘there is seuen specyalles/
whiche is/ the momon/ the gleke/ the flusshe/ the torment/
the regnet/ one and thyrty/ and ye tryumphe’. The names
of all these card games are known from other sources. The
French original of this passage gives the seven games as ‘le
mommo/ le glic/ le flux/ le tourment/ la regnette/ Trente
et Vng/ & la ronfle’. It would therefore appear that Watson
is equating ‘ronfle’ with ‘tryumphe’, but this impression is
contradicted by an earlier sentence on the same page: ‘the
players shall saye/ playe we at the romfle/ the other shal
saye playe we at the tryumphe’; at this point in the original
French, different games are mentioned (‘les ioueurs diront
iouds au glic. Lautre dira iouons aux flux’). Although this is
only a translation, it seems clear that ‘tryumphe’ must have
been known in England as the name of a card game. For
Latimer’s sermon, in which allegorical use is made of the
words ‘triumph’ and ‘heart’, see Hugh Latimer, Sermons on
the Card,
London, 1886, p. 27, and John Foxe, Actes and
1563, pp. 1297-8 and 1300 (col. 2): ‘the game
that wee wyll playe at, shall bee called the triumphe.’

10 Linguae latinae exercitatio F.L. Vivis Valentini, Basle, 1541,
Dialogus xxi, ‘Ludus Chartarum seu foliorum’. The first
edition of this book, of which there are very many, was
published in Basle in 1538. S.R. Kaplan, The Encyclopedia of
New York, 1978, states that ‘Ludovico Vives, Writing
in 1545 in Ludus Chartarum, Dialogus, describes the method of
playing the game of tarocchi’ (p. 28) and that this work
‘contains one of the earliest published references to the
game of tarocchi’ (p. 374). This is an error: Vives died in
1540, and the dialogue refers to a game with the regular
pack, as might be expected of a Spaniard. In the Spanish
dictionary of John Minsheu, one of the meanings of the
word triumfo or triumpho is given as ‘a game at cards so
called’: see A Dictionarie in Spanish and English, by Richard
Percivale, enlarged by I. Minsheu, London, 1599, reprinted
1623, s.v. ‘triumfo’. In the Dialogues in Spanish and English
appended to the dictionary, s.v. triunfo figures as the name of a
card game and is translated ‘Trump’ (p. 25, towards the
end of the 3rd dialogue). The dialogue later refers to a
distinct game called triunfo callado, translated as ‘still
trump’; I do not know what this is. The Tesoro de la Lengua
Castellana, o Espanola
of Don Sebastian Cobarruvias,
Madrid, 1611, gives un juego de naipes, que llaman triunfo (a
game of cards which they call triunfo) as one of the meanings
of triunfo.

11Johann Fischart, Geschichiklitterung, Gargantua, ed. U.
Nyssen, Düsseldorf, 1963, gives the text of the 1590 edition.
See also the work by Rausch cited in footnote 3.

itself, Francesco Berni alluded to Trionfi as a
game played with regular cards, and distinct
from Tarocchi, which he refers to under that
name, in 1526, and it has other mentions in
sixteenth-century Italian sources.12. These
games were by no means all identical; on the
contrary, they differed greatly. Daniel Martin’s
Trümpfspiel was a complex game, whereas the
other games are all simple ones: in Triomphe
each player received only five cards, whereas the
English game of Triumph was the ancestor of
Whist, and the entire pack was distributed.
What, then, explains this allocation to a variety
of trick-taking games played with the regular
pack of a name formerly applied to the game
played with the very special pack of cards now
known as tarocchi?

The explanation is very simple. What connects
all these different games is that they were games
played with trumps: their names refer to this
outstanding and, at the beginning of the
sixteenth century, novel feature of them. The
word ‘trump’ is simply a corruption of ‘triumph’,
a corruption that appears almost as early as the
parent word ‘triumph’, as applied to cards, and
soon supersedes it. Similarly, Trumpf is an
ordinary German word for ‘trump’, as triunfo is in
Spanish, although in French the word most
commonly used has long been atout. In Italy, the
word trionfi continued very often to be applied to
the triumph cards of the Tarot pack, as well as to
trump cards in other games, although it was no
longer used for Tarot cards as a whole, or for the
game played with them: in fact, it simply
acquired the meaning ‘trumps’, and, now that we
know for what purpose the triumph cards are
used in play, we may as well simply call them
‘trumps’ henceforward. Even before the word
tarocchi had been introduced, when the word
trionfi was reserved for Tarot cards, it had a
certain ambiguity; it could be applied to the
cards of the Tarot pack as a whole, but it denoted
more particularly the twenty-one cards we have
been calling triumph cards, or those together

12F. Berni, Capitolo del Gioco della Primiera, Venice, 1526.
The game is also mentioned by Girolamo Cardano in ch.
XXV of his Liber de Ludo Aleae, written in 1564, published in
Hieronymi Cardani Mediolanensts Opera Omnia, tom. 1, Lyons,
1663, pp. 262-76; see also S.W. Singer, Researches into the
History of Playing Cards,
London, 1816, pp. 328-31, and
Oystein Ore, Cerdano, the Gambling Scholar, Princeton, 1953,
appendix. It was also discussed in a lost Italian work of
Cardano’s, De Ludis, of 1524; see his Liber de Libris Propriis,
in Opera Omnia, tom. 1, pp. 60-95, particularly p. 61 (col. 2).

180 Part I: History and Mystery
with the Matto; a set of Tarot cards was more
properly called carte da tnonfi, ‘cards with
triumphs’. The word tarocchi inherited the same

These facts all but compel us to make the
following supposition. Somewhere around the
end of the fifteenth century, it occurred to some
Italian card players that the idea of trumps,
which hitherto had been the distinguishing
feature of the game of Tarot, in which special
cards were used for the purpose, could be
adapted to games played with the regular pack
by simply designating one of the suits as trumps.
This idea then spread rapidly to many other
countries, travelling much faster and more
widely than the game of Tarot itself, just because
it did not require a special pack of cards not
everywhere readily obtainable. The word used
particularly for the trump cards in Tarot was
very naturally transferred to trump cards in these
new games, necessitating the introduction of a
new word for Tarot cards as such; and, since
the new idea of trumps appeared to card players
the distinctive feature of the various games, the
same word was employed as the name of the
games themselves. This hypothesis is reinforced
if we accept Miss Moakley’s conjecture about the
origin of the word trionfi or triumphi as applied to
the Tarot trumps. Many have suggested that the
word was adopted because a trump card
‘triumphs’ over any card of a (plain) suit; but, on
Miss Moakley’s view, the word had originally an
iconographical significance, deriving from a
conception of the figures on the Tarot trumps as
belonging to one of those processions of
allegorical or mythological figures on floats that
delighted Italian Renaissance courts; such
processions, whimsically modelled upon the
triumphal processions of ancient Rome, were
ordinarily referred to as ‘triumphs’. I view this
conjecture as plausible, not as established; but, if
it is correct, it would follow that, when the Tarot
pack was invented, the word triumphi was not yet
in use for trump cards in other games. If the idea
of trumps had at that time been embodied in
games with the regular pack, it would have been
surprising if whatever word were then used for
‘trumps’ in such games had not, at least
sometimes, been applied to the trump cards of
the Tarot pack; yet there is no trace of any such
word. In any case, it is beyond question that, at
around 1500, a word formerly reserved
exclusively for Tarot cards was borrowed for
games played with the regular pack, and rapidly
came to mean simply ‘trumps’; and this would be
virtually inexplicable if trick-taking games with
trumps had been played with the regular pack
since before the Tarot pack was first invented.
Even if we accept this hypothesis, there is a
problem in determining the date at which the
idea of trumps was first transferred from Tarot to
games played with the regular pack. Were it not
for some fifteenth-century French references to
the game of triomphe or triumphe, we should
naturally place the innovation in the first decade
of the sixteenth century. We have already
mentioned a French reference to le triumphe as a
card game dating from 1482, however, as well as
Duke René II’s game of triumphe in 1496.13 Now
France was the first country to which the game of
Tarot spread from Italy, and we know from
Rabelais’s references to it in 1534 that it was
already quite well known by then. Rabelais also
refers to la triumphe, which thus obviously was for
him a game played with the regular pack; there
are other sixteenth-century French references to
Triomphe, and, since the game is described in
the seventeenth-century card-game books, we
can be confident that what Rabelais is referring
to is that game of Triomphe which has survived,
with very little change, down to the present. But,
in 1482 or even 1496, the game of Tarot was still
being referred to in Italy as trionfi, and the word
tarocchi was, so far as we know, not yet in use. The
fifteenth-century references to triumphe could,
therefore, equally well refer to Tarot. We have to
choose between the hypotheses, both mildly
surprising, that the Tarot pack and the game of
Tarot had already reached France by 1482, and
that, by that date, the crucial step had already
been taken of adapting the idea of trumps to
games with the regular pack; this latter step must
originally have been taken in Italy rather than in
France, unless, indeed, we are prepared to
assume that both hypotheses hold good.

Our theory that the triumph cards of the Tarot
pack were devised in order to play what was then
a wholly new role in trick-taking games, that of
trumps, and that this idea was subsequently
borrowed from the game of Tarot and adapted to
13 Frédéric Godefroy, Dictionnaire de Uancienne langue
francaise, Complément,
Paris, vol. X, 1902, s.v. ‘triomphe’,
quotes the phrase jouer au jeu du Triumphe from a manuscript
of 1482, JJ 206, folio 181 r° in the Archives Nationales. For
René II’s game, see H.-R. D’Allemagne, Les Cartes à jouer, vol.
II, Paris, 1986, p. 212.

The Game of Tarot 181
games played with the regular pack, can be
sustained only in the absence of any evidence for
games, of equal antiquity with Tarot, played
with the regular pack but involving trumps. We
must therefore consider carefully two possible
claimants for such a position. The first is the
Italian game of Ronfa. In his book Gambling, Mr
Alan Wykes informs us that “Triumph, ruff, ombre
and honors were all built round the idea of
trumping, and are all developments of ronfa, an
early sixteenth-century Italian game in which the
leading player could decide to his own
advantage which should be the superior or
trump suit in play ... Ronfa, like ombre, was
usually a game for two.’14 If this information
were sound, Mr Wykes, or the ‘professional
researchers’ who assisted him, might have
succeeded in identifying that Italian game in
which the idea of trumps was first adapted from
Tarot to play with the regular pack.
Unfortunately, the game of Ronfa is older than
the early sixteenth century, probably far older.
The first reference known to me from an Italian
source occurs in the same sermon by an
anonymous Dominican friar; dating from
between 1450 and 1480, which we have already
cited as giving the first listing of the Tarot
triumphs.15 There are here two mentions of
Ronfa, as if to two separate games: one is
specifically to a card game (‘Ronfa, ludus
cartularum’), and the other to a game bearing
the alternative name of Buffa Aragiato, which
may equally well have been a dice game or game
of tables (i.e. backgammon game);16 the whole
passage is one in which names of card games,
dice games and backgammon games are
indiscriminately jumbled together as being,
14 Alan Wykes, Gambling, London, 1964, p. 165. I know of
no game called ‘Honors’ or ‘Honours’; but a stage in the
evolution of Triumph into modern Whist was a game called
Ruff and Honours, described in Charles Cotton, The
Compleat Gamester
, London, 1674, pp. 114-20. The sentence
should therefore probably be emended to begin ‘Triumph,
and Honours and Ombre were all ...’.

15 Robert Steele, ‘A notice of the Ludus Triurnphorum’,
Archaeologia, vol. LVII, 1900, pp. 185-200; see also the same
author’s ‘Early playing cards’, Journal of the Royal Society of
, vol. XLIX, 1901, pp. 317-23.

16 'The famous manuscript compiled in 1283 at the order
of King Alfonso X the Wise of Castile on chess, dice,
backgammon and board games mentions buffa cortesa and
buffa de baldrace as backgammon games: see Alfonso el Sabio,
Libros de Acedrex, Dados e Tablas (Das Schachzabelbuch König
Alfons des Weisen
), ed. and trans. by Arnald Steiger, Geneva
and Zurich, 1941, pp. 324-5, 328-9,

according to the preacher, names of devils. Of the
game of Buffa Aragiato or Ronfa, the preacher
comments that ‘it is a cruel game, that has led
many to poverty’; so it was certainly a game
played for high stakes. Whether or not there was
also a backgammon game called Ronfa, the card
game was well established in Italy by the 1490s:
in 1491 ronfae was prohibited by municipal edict
in Bergamo;17 in 1492 a letter of Ippolito d’Este
acknowledged the receipt, among other things, of
a pack of cards for playing the game (carte da
);18 in 1499 the Diario Ferrarese mentioned
Rompha as among the games played at the court
of Ferrara.19 There are frequent references to it in
sixteenth-century Italian sources.20

If we take the date of the anonymous sermon
against gaming to be at the extreme end of the
interval, 1450-80, to which Robert Steele dated
it, that could just be consistent both with our
theory that the idea of trumps was taken over
from Tarot into other games, and with Mr
Wykes’s contention that Ronfa was a trick-taking
game with trumps. However, although
everything concerning the game of Ronfa is
obscure, it seems highly probable that we should
identify it with the game played in France under
the name of Ronfle. But, if so, it goes back to the
beginning of the fifteenth century: the earliest
reference to Ronfle in Godefroy’s dictionary is
from 1414, with another from 1464.21 If Ronfa
and Ronfle are to be equated, and if Mr Wykes is
correct in saying that Ronfa was a game
involving a trump suit, our theory about the
origin of the Tarot pack is in ruins.
17 See W.L. Schreiber, Die ältesten Spielkarten, Strasbourg,
1937, p. 79.

18 See Giulio Bertoni, ‘Tarocchi versificati’ in Poesie,
leggende, costumanze del medio evo,
Modena, 1917, p. 218.

19 See Muratori, Rerum Italicarum Scriptores, vol. XXIV, p.

20 See the works of Berni and Cardano cited in footnote

21 Frédéric Godefroy, Dictionnaire de l’ancienne langue
, Paris, vol. 7, 1892, s.v. ‘ronfle’. The 1414 reference
is from a manuscript, cited merely as Arch. JJ. 189, piéce
266: ‘Lesquelz compaignons commencerent à jouer au jeu
de ronfle.’ The 1464 one is cited from the Lettres de Jehan de
, in Cabinet historique, 1875, p. 241: ‘Comme l’on dist,
Pon joue aux cartes pour passer le temps, est a savoir a le
roufle, a XXXI, au ghelicque, au hanequin et au france ju.’
There are other references without precise dates: one of
1537, given as ‘Chicheface’, Poésie des XVe et XVIe siécles, vol.
XI, p. 290, employs the term roffle not as the name of a game
but as a technical term in another card game, la Picardie:
‘J’avais cinquante et cing de roffle/ En jouant a la Picardie.’
I am afraid I have not checked these references.

182 Part I: History and Mystery
As we saw concerning the alleged fourteenth-
century Persian references to As-Nas, and the
supposed entry in the diary of one of Columbus’s
seamen,22 Mr Wykes is better at making
assertions than at substantiating them. It would
be a matter of the greatest interest for the history
of card games to find out with certainty how
Ronfa (or Ronfle) was played: but a request to
Mr Wykes for references to support his confident
statement on the subject failed to elicit the most
tenuous ground for it. The fact of the matter is
that the evidence that has come down to us on
the question is confusing, and points in opposite
directions. The word ronfa, and the word ronfle as
used in card play, are certainly connected with
the English word ‘ruff’, still in regular use by
Bridge players as a verb meaning to play a trump
card to a trick to which a plain suit has been led.

By the seventeenth century, Triumph had
evolved into two forms, described by Charles
Cotton, one called Ruff and Honours, or,
alternatively, Slamm, and the other called Whist.
In Whist, as then played, the 2s were omitted
from the pack, and every player received twelve
cards, the last being turned to determine the
trump suit. In Ruff and Honours, each player
again received twelve cards, but the last four
were dealt to the centre of the table to form a
stock, the top one being turned to determine
trumps; the player who held the Ace of trumps
was entitled to take the four cards of the stock
into his hand, discarding four at his choice. The
procedure of taking the four cards of the stock
was called ‘ruffing’. Cotton does not use the
word ‘ruff? As meaning ‘to trump’; but he does
describe the French game of Triomphe under
the name ‘French-Ruff’.24 John Florio, in his
Italian/English dictionary of 1598, gives ‘Ronfa’
as meaning ‘a game at cardes called ruffe or
trumpe’, and the verb ‘ronfare’ as meaning ‘to
snort, to snarle (the 1611 edition adds ‘in sleep’)
...; also to ruff or trump at cards’.?5 Ronfare is not
in use in modern Italian in either sense, but the
more general meaning given by Florio confirms
the identification of Ronfa with Ronfle, ronfler
being the ordinary French word for ‘to snore’.
Randle Cotgrave, in his French/English
dictionary of 1611, gives ‘Ronfle’ as meaning
2? See Chapter 3, p. 46 and fns 41 and 42, and p. 51.
23 Charles Cotton, op. cit., pp. 114-20.

24 Ibid., pp. 121-2.

25 John Florio, The Worlde of Wordes, London, 1598.

‘Hand-Ruffe, at Cards’. All this naturally
supports the view that Ronfa or Ronfle was a
trick-taking game whose outstanding feature was
the use of a trump suit.

The word ‘ruff’ was, however, also used in card
play in a quite different sense. In Piquet and
some other games there is a score for the foint,
that is, for having the greatest number of cards of
any one suit, or, in case of equality, that having
the highest total value (the Ace counting 11, the
court cards 10 each and the others their face
values). The word ronfle regularly appears in the
older French dictionaries as meaning the ‘point’
in Piquet, though later supplanted by the word
point; and Cotton uses ‘Ruff’ in this sense too.27
The eighteenth-century French writer on playing
cards, Bullet, uses the term ronfle only in this
sense, without mentioning the game of Ronfle.28
A Spanish encyclopaedia of 1926 assigns the
same general meaning to the word runfla, though
without a specific mention of Piquet.29 Boiteau
d’Ambly, in his book of 1854, discussed the game
of Ronfle, and said that it was still played in the
Vosges: according to him, it was, as played there,
a very simple two-handed game, in which the
sole object was to have the highest ronfe or point,
taken as consisting in the highest sum of'values of
cards of any one suit held. in hand.30 The
authoritative Italian dictionary, the Vocabolario
della Crusca
of 1612, gives as a proverbial phrase
accusare la ronfa: accusare is a verb commonly used
in card play to mean ‘to declare’ some
advantageous combination of cards held in hand,
and the entry goes on to say that Ronfa is a card
game similar to Primiera.31 Primiera was a game
very popular in Italy in the early sixteenth
century, originating, like so many card games, in
Spain, where it was called Primero, under which
name it also enjoyed great popularity in
26 Randle Cotgrave, A dictionarie of the French and English
London, 1611.

27 Op. cit., p. 82.

25 Jean Baptiste Bullet, Recherches histortques sur les cartes à
, Lyons, 1757, p. 144.

29 Encyclopedia Universal Ilustrada Europeo- Americana, Bilbao,
Madrid and Barcelona, vol. LII, 1926, s.v. ‘runfla’: ‘reunién
de muchas cartas de un mismo palo’.

30 Boiteau d’Ambly, Les Cartes à jouer et la cartomancie,
Paris, 1854, p. 162.

31 Vocabolario degli Accademici della Crusca, Venice, 1612, s.v.
‘accusare’: ‘E proverbialmente, Accusar la ronfa giusta, che
é confessar la verita per |’appunto. E Ronfaé giuoco di carte,
come Primiera e simili’ (Proverbially, ‘to declare the true
ronfa’, that is, to confess the truth exactly. Ronfa is a game
of cards like Primiera and similar ones}.

The Game of Tarot 183
England. It was not a trick-taking game, but
belonged to the same general category as Poker,
being a gambling game whose outcome
depended on the particular combinations of
cards each player held in his hand.

It is plain that two totally different ideas, that
of trumps and that of the point, became
associated with the words ronfa and ‘ruff’. The
difficulty is to know which of these two
associations is the original one; which one
characterised the game of Ronfa as it is referred
to in the fifteenth and sixteenth centuries. Here
Randle Cotgrave’s translation of ‘Ronfle’ as
‘Hand-Ruffe’ may give us a clue: his entry for
‘Triomphe’ is ‘The Card-game called Ruffe, or
Trump; also, the Ruffe, or Trump at it’. By the
time he compiled his dictionary, the confusion
between the two senses of ‘ruff’ had already set
in: and his phrase ‘Hand-Ruffe’ may well have
been intended to differentiate a game in which
the ruff was the point on cards held in hand from
one in which it consisted in the play of a trump
card. In 1534 Rabelais mentioned both Ronfle
and Triumphe in his list of games: and his
English translator, Urquhart, of 1653 likewise
rendered the former as ‘handruff’, while he
translated the latter as ‘trump’.32 That, in
England at least, the name ‘Ruff’ originally
denoted a game in which ‘ruff’ meant ‘point’
seems to be clinched by one of the earliest
English references to Ruff as a card game, in the
contribution of 1589 to the Martin Marprelate
controversy called Martins Months minde and
dubiously attributed to Thomas Nashe. The
author says, in a passage whose metaphorical
intention escapes me, “They are now in hande to
shuffle the Cardes (as ill as they will seem to love
them) and to confounde all, to amende their
badde games, having never a good Carde in their
handes, and leaving the auncient game of
England (Trumpe) where everie coate and sute
are sorted in their degree, are running to their
Ruffe where the greatest sorte of the sute carrieth
away the game’.33 This passage is far from clear:
but it is apparent that the writer is not using
‘Ruff’, as was later done, as an alternative name

32 See F. Rabelais, The heroic deeds of Gargantua and
trans. Sir Thomas Urquhart and P. le Motteux,
Everyman edition, vol. I, London, 1929, p. 50, or many
other editions.

See The Complete Works of Thomas Nashe, ed. by
Alexander B. Grosart, vol. J, 1883-4, pp. 141-205; the
quoted passage is on p. 161.

for the game of Trump or Triumph, but as
standing for a quite different game; and the
concluding remark fits well with the idea that the
game referred to was one in which the winner
was he who held the longest single suit in his

It thus seems more probable than anything
else that Ronfa was not originally a trick-taking
game at all, or, at any rate, not one played with
trumps, but one in which the point - as
determined either by the mere number, or by the
sum total of the values, of cards of one suit held in
hand - played an important role; and that the
verbs ronfare and trionfare, as applied in card
games, later became confused owing to their
similarity of sound, leading to a similar confusion
between the English terms ‘ruff’_and ‘trump’
derived from them.34 This confusion can be
clearly seen in John Florio’s dictionary: the two
verbs seem to have been still further confused
with the verb tronfiare, meaning.“to puff oneself
up’. Florio gives tronfio as meaning ‘puffed,
swolne; ... also a trump at cards, a game at cards
called trumpe’, and tronfare, also spelled tranfiare,
as meaning ‘to puffe, to swell ...; also to trump at
cardes’, with the 1611 edition giving as a further
meaning ‘to snort’ (the basic meaning of ronfare).
The confusion seems to have become pretty
thorough. A game called Trionfetti was already
in existence in Italy in 1564, when, as Triumfeti,

‘it was mentioned by Cardano in addition to the
game of Triumphi, both being distinct from
Tarocchi, which is referred to by Cardano under
that name;35 Trionfetti may be the game
34 An important, but puzzling, use of the term ronfa, not as
the name of a game, but as a technical term in play, occurs
in the Invettiva contra il Giuoco del Taroco of 1550 by Alberto
Flavio Lollio (1508-1568) of Ferrara; this poem will be
discussed in detail in Chapter 21. Speaking of the moment
before play opens when you inspect your hand, Lollio says,
‘s’hai quattro,/ O cinque Carte di Ronfa, tu temi/ Che non
ti muoia il Re, con le figure’ (‘if you have four or five cards of
ronfa, you are afraid that you will lose the King, with the
court cards’). Carte de Ronfa does not mean ‘trumps’, for
which Lollio uses the term trionf. He is speaking of a three-
handed game with the Tarot pack, in which each player has
a hand of twenty cards. The most natural interpretation is
that Lollio is speaking of a long holding in a plain suit,
which is called ronfa; it you have the top four or five cards of
that suit, you are afraid that they may be trumped by
another player who has a void in that suit. I do not feel sure,
however, that this is the correct interpretation.
33 Cardano, Liber de Ludo Aleae, cap. XXV; see footnote
11. The game is also mentioned by Tomaso Garzoni, La
Piazza Universale,
Venice, 1585, as trionfitti; see p. 574 of the
1586 edition. John Florio, A Worlde of Wordes, London, 1598,

184 Part I: History and Mystery
mentioned by Francesco Berni in 1526 under the
name Trionfipiccoli, again as distinct both from
Trionfi and from Tarocchi.36 From one later
account, however, it appears that Trionfetti was
not a trick-taking game at all; played by four
players in fixed partnerships with a full or a 32-
card pack, it allows scores for pairs, for threes of
a kind and for the point (punto), taken as the sum
of values of the cards held in any one suit.?37

It is noteworthy that we have no evidence for
the use of the French word ronfle, or of Spanish
runfila or Portuguese rufa, with the meaning
‘trump’: Godefroy cites two sixteenth-century
metaphorical uses, one from Rabelais, of the
phrase en ronfle vue to mean ‘in a good position’,
which would tally well with its literal meaning
being that of a successful point displayed to the
opponents. In view of the priority of the French
references over the Italian ones, it is likely that
the game was of French origin and was imported
into Italy, the name ‘Ronfle’ being there
converted into ‘Ronfa’. Ronfle need not be
thought to have been so unsophisticated a game
as that mentioned by Boiteau d’Ambly: it could
even have been ancestral to Piquet. Piquet is
referred to by Rabelais under its older name of
Cent (the game being played 100 up), which in
England was corrupted to ‘Sant’. Being a game
played with a shortened pack (originally with the
36-card pack), it is unlikely to have originated
before 1500, and certainly cannot be as old as
defines Trionfeiti as ‘a game at cardes as our trump’; in his
Second Frutes of 1591 he had translated triumphetto as ‘trump’.
In the 1659 edition of Florio’s dictionary. A Dictionary Italian
and English,
revised by Giovanni Torriano, the entry for
Trionfetta is ‘the game ‘Trump, or Ruff, or Whisk at Cards’;
‘Whisk’ is just an alternative form of “Whist’, but it seems
very unlikely that Whist was played in Italy in the
seventeenth century.

36 F. Berni, op. cit.

37 Aquarius (i.e. Louis d’Aguilar Jackson), Italian Games at
Cards and Oriental Games
, pp. 49-52. Aquarius equates
Trionfetti with Gilé, a game separately referred to by
Garzoni, gilé being the technical term for a pair. Aquarius,
or L. d’A. Jackson, was an engineer who was an enthusiast
for foreign card games, and published a number of small
booklets about them, which, unfortunately, are far from
always accurate or clear; he frequently follows very closely a
German or Italian text, which he does not in all cases
interpret correctly. In this case, I do not know what his
source was, and have been unable to find any Italian-
language account of Trionfetti. I have no doubt that, in
some source available to Aquarius, Trionfetti was described
as a non-trick-taking game of the kind he describes; but it
may well be that the game referred to by Cardano, Garzoni
and Florio was a trick-taking game with trumps.

Ronfle. Piquet is a simple trick-taking game
without trumps; in which the importance of the
play in tricks is overlaid by that of the
declarations that can be made from hand before
play begins: these features are entirely consonant
with its being, or having been descended from, a
game going back to a very early stage in the
history of European card play, since trick-taking
games without trumps are probably as old in
Europe as playing cards themselves, and
declarations made from hand seem also to be
very ancient. Ronfle may well have been a game
of the same general type.

However this may be, it is evident that,
although at first sight Ronfa poses a problem for
our theory about the origin of the Tarot pack,
that problem evaporates upon examination. The
discovery of some really solid evidence about how
Ronfle and Ronfa were played in the fifteenth
century is, indeed, much to be desired; but, in
default of any more convincing reasons than we
have at present, we cannot with any assurance
classify them as trick-taking games, let alone
ones played with trumps.

A more serious problem is raised by the game
of Karnöffel, which was a celebrated game in
fifteenth- and sixteenth-century Germany. It goes
back to quite early in the fifteenth century,
having been listed in a municipal ordinance of
Nerdlingen in 1426 as among the games that
could lawfully be played at the annual city fête.38
A thorough investigation of the history of the
game has been carried out by Dr Rudolf von
Leyden, who has collected some thirty references
to it from the fifteenth and sixteenth centuries.
Of these, eight are from the works of the famous
Catholic preacher Bishop Johann Geiler of
Kaisersberg. For him it provided an apt
illustration of the disruption of the social order.
In the usual card games, the cards in each suit
ranked in a natural and determinate order, with
the King at the head, and the higher-numbered
cards beating the lower-numbered ones, the
Deuce ranking last. But in Karnöffel, which
Geiler mistakenly regarded as a new game, all
was topsy-turvey. During the Reformation,
Karnöffel, containing, as it did, cards known as
the Pope, the Emperor (Kaiser) and the Devil,
became the source for a more substantial allegory
by Protestant propagandists and satirists.
38 W.L. Schreiber, Die Altesten Spielkarten, Strasbourg,
1937, pp 42-3.

The Game of Tarot 185
After the sixteenth century, Karnöffel lost its
celebrity, and passed out of the consciousness of
most card players. But it by no means died out: it
continued to be played, in varying forms, in
isolated rural areas of Germany, Switzerland and
the Netherlands, and is still played in
Switzerland to this day; of all European card
games still living, it must be that with the longest
demonstrable history. The earliest detailed
account of the manner of play comes from an
article published in a German periodical of 1783,
describing Karniffel as then played amongst the
Thuringian peasantry.39 Dr von Leyden has
unearthed three others: one published in
Lucerne in 1841, describing Karnöffel or
Kaiserspiel as then played in the German-
speaking Catholic cantons of Switzerland; one
from a German periodical of 1924, describing the
Karnüffeln or Knüffeln then played in Friesland;
and a book of rules for the game of Kaiser-Spiel
or Kaiserjass (Chaisere) as now played in the
half-canton of Nidwalden, together with a very
recent description by Herr Hansjakob
Achermann published in a local journal.40 The
original game was played with a full 48-card
German-suited pack. In Thuringia it was played
with a 36-card one, and in Friesland with a
French-suited pack of 48 cards. In Switzerland it
is played with the Swiss-suited pack. In the
nineteenth-century Lucerne version, all 48 cards
were used; this must have been the last game for
38 Beytrag zur Geschichte der Kartenspiele, Teutscher
, 1783, pp. 62-87.

39 Das uralte edele so genannte Karnöffel- oder Kaiserspiel,
Lucerne, 1841; J.P. Berhard, ‘Das Karmüfeln (Knüffeln).
Ein friesisches Kartenspiel’, Die Heimat: Monatsschrift des
Vereins zur Pfelege der Natur- und Landeskunde in Schlesung-
Holstein, Hamburg, Liibeck und dem Fürstentum Lübeck
, March
1924, reprinted (pp. 22-4) at the end of G. Erhardt, ‘Das
Karniffel- oder Karnöffelspiel’, Der Alte, 5th Jahrgang,
Altenburg, 1930, pp. 20-2; and Jassreglement ftir das Kaiser-
n.d., issued by J. Müller & Cie. of Schaffhausen,
together with Hansjakob Achermann, ‘Der,Kaiserjass, wie
er heute in Nidwalden gespielt wird’. Dr von Leyden has
found references to Kaiserspiel or Karnöffel being played in
Switzerland from the early seventeenth century (Schweizer
Idiotikon, Zurich, 1895, vol. 3, p. 514); in Lübeck in the early
eighteenth century (Sporham-Krempel, Ein Handvoll Glück,
Munich, 1958, p. 49); in Switzerland and Thuringia in 1810
Solothurnisches Wochenblatt, no. 29, 21 July 1810, p. 222); and
in Iserlohn, Westphalia, in the nineteenth century (Kluge-
Gétze, Etymologisches Worterbuch der Deutschen Sprache, Berlin,
1951, p. 366). It is not known whether the game survives in
Germany or in Friesland. For all this, see R. von Leyden,
Karnöffel: das Kartenspiel der Landsknechte, Vienna, 1978, a
booklet issued by Piatnik & Söhne with a reproduction pack.

which they were required. But the modern game
uses a 40-card pack, missing the 8s and 9s; for
the few places in which this ancient game
survives, the 48-card Swiss-suited pack continues
to be manufactured. Dr von Leyden, Mr John
McLeod and Mr David Baird have all visited
Stans to witness this venerable game being

The games described in these various sources
differ considerably from one another, but share
several common features, All are simple trick-
taking games with trumps. The Swiss games are
those of which we have the most detailed
descriptions: the modern game as played in
Nidwalden differs from the nineteenth-century
one in certain ways, but the essential features are
in common. The Thuringian and Frisian versions
have the surprising feature of having two trump
suits, while the Swiss forms have only one. In the
Thuringian and Frisian games, the trump suits
are determined by turning the first two undealt
cards; if both are of the same suit, another is
turned, and so on until two suits have been
exposed. In the Swiss games, each player receives
one card face up in the first round of the deal,
and the lowest card (according to the ranking ‘in
plain suits) so dealt (in case of equality, the
earlier) fixes the trump suit. The game is
characteristically played with four or with six
players in two fixed partnerships; usually each
player is dealt five cards, and that side wins
the round which makes three tricks (once either
side has done so, the remaining tricks are not
played out). In the Frisian game, there are four
players in two fixed partnerships, but each player
receives nine cards, the object of each side being
to make five tricks; the nineteenth-century Swiss
form provides versions for three players or for
four, each playing for himself, and in these cases
the winner of a round is the first to obtain three
tricks, each player receiving as many cards as
required to ensure that someone wins (seven in
the three-handed version, nine in the four-
handed one). In no form of the game is there an
obligation to follow suit, nor a prohibition on
trumping although able to follow suit. In the
Thuringian game and in the Swiss nineteenth-
century one when played with three players or
with four, each playing for himself, it was
obligatory for each player to beat the highest
card yet played if he could (with a card of the suit
led or with a trump): in Swiss partnership games
and in the Frisian game, there are no constraints

186 Part I: History and Mystery
on play to a trick. In all forms, open discussion
between the players is lawful, or each side has a
director who instructs his partner or partners
what to play.

As in most trick-taking games with trumps, a
trick may be won only by a trump or a card of the
suit led. In all the forms other than the
Thuringian one, the ranking of the cards in the
trump suit or suits is quite eccentric, and differs
wholly from that in the plain suits. In the
Thuringian game, the same eccentric ranking
obtains in the plain suits as well as in the two
trump suits (unless, indeed, the author of the
1783 account had made a mistake in this regard).
This difference in the order of the cards in
trumps and in the plain suits is but an extreme
example of something that happens in several
other games, notably in Ombre; but, again with
the exception of the Thuringian form, Karnöffel
has an absolutely unique feature. This is that
only the highest of the trump cards are fully
fledged trumps, that can beat any card of a plain
suit. The lowest cards of the trump suit have no
power as trumps at all, while the intermediate
ones are what may be called partial trumps: they
can beat cards of a plain suit below a certain
rank, but not those above it.

In detail, the ranking of the cards differs
considerably in the various forms: but it is worth
describing, together with the names used for
certain cards, for the light that is thrown on some
of the fifteenth- and sixteenth-century references.

(1) As remarked, the nineteenth-century Swiss
game was played with a full 48-card Swiss-suited
pack. The Banners (10s) were called Kaisers; a player
holding one or more could declare it before the start of
actual play, thus increasing the value of the round and
making the Kaiser a trump; if not declared, it did not
become a trump and belonged to its proper suit. In
the plain suits, the cards ranked, in descending order:
King, Ober, Unter, Banner (if not declared), 9, 8, 7, 6,
5, 4, 3, Deuce (called Sau = Sow). The most
remarkable features, both of this game and of the
modern Swiss version, are, as just observed, that only
certain cards of the trump suit have any power as
trumps, those that do being called Stecher, and that
only some of the Stecher function as fully fledged
trumps; those that do are called Königstecher or
Kingstecher (King-beaters), because they beat even the
King of a plain suit. In the nineteenth-century game,
the Königstecher were, in descending order: the Unter
of the trump suit, called the Jos; the 6 of the trump
suit; the Deuce or Sau of the trump suit; and the
Kaiser of Roses (if declared), called Blass (pale). The
remaining six Stecher had restricted powers as trumps.
If a plain suit were led, they could beat certain cards
of that suit, but were beaten by others. The 3 of
trumps and, below it, the Kaiser of Acorns (if
declared), called Grün (green), were Oberstecher, and
could beat the Ober of a plain suit and any lower
card, but not the King. Next came the two Unterstecher,
the 4 of trumps followed by the Kaiser of Shields or
Tätsch: these could beat an Unter of a plain suit or any
lower card, but not the Ober or King. Last came the
two Farbstecher, which could beat any numeral card of
a plain suit (including an undeclared Banner), but
not a court card: these were, in order, the 5 of trumps
and the Kaiser of Bells or Fugel. Thus, if Shields were
trumps, and the 9 of Acorns were led, followed in turn
by the 5 of Shields, the Unter of Acorns and the 4 of
Shields, each card in succession would beat the
preceding one; the Unter would beat the 5 of trumps
because the latter was only a Farbstecher; but would be
beaten by the 4 of trumps because it was an

The 7 of the trump suit, called Sibbile or Babeli,
behaved in a peculiar way, as in all Karnöffel games.
When led to a trick, it could be beaten only by the
highest card of all, the Jos or trump Unter; but when
played to a trick otherwise than as the first card, it
lacked all trick-taking power. The remaining four
cards of the trump suit; the King, Ober, 9 and 8, had
no power as trumps, but formed, in effect, a separate
plain suit, ranking in that order. Thus, if one of the
Stecher were led to a trick, it could be beaten only by a
higher Stecher; e.g. the 3 of trumps, if led, could be
beaten only by a Königstecher, and not by the King of
trumps. But if one of the four cards of the trump suit
which were not Stecher were led, it could be beaten
only by a higher such card or a Stecher of the
appropriate rank, the Ober by an Oberstecher and the
King by a Königstecher. Thus, if the Ober of the trump
suit were led, it could be beaten by the trump 3, and
that in turn by the King of the trump suit.

(2) The modern Swiss game is played without the
8s and 9s; with this exception, the cards rank in the
plain suits as before, the Unter being called Bauer and,
as usual nowadays, the Deuce being called Ass (Ace).
There are now six Aingstecher, which are, in
descending order: the Kaiser of Bells, called Mugg;
the 5 of trumps; the Bauer (Unter) of trumps, called
Joos; the 6 of trumps; the Ace (Deuce) of trumps; and
the Kaiser of Roses, called Blass as before. (It will be
seen that the former Farbstecher have been promoted.)
The Oberstecher are the 3 of trumps followed by the
Kaiser of Shields, called Oberkaiser; the Unterstecher are
the 4 of trumps, followed by the Kaiser of Acorns,
called Wydli or Gyger. The 7 of trumps, with no special
name, has the same peculiarity as before: played to a
trick, it has no power at all, but, led to one, it can be
beaten only by the Joos. The Joos, being no longer the

The Game of Tarot 187

highest card, can then in turn be beaten by the Mugg
or the 5 of trumps, neither of which could have beaten
the trump 7 directly. This leaves only the King and
Ober of the trump suit, called, respectively, der Fuil
and hiratä; they in effect form, as before, a diminutive
plain suit of their own, having no power as trumps or
when a Stecher has been led.

(3) The Frisian game is or was played with 48
cards, comprising a full French-suited pack from
which the 10s have been removed. The cards in the
plain suits rank, in descending order, Ace, King,
Queen, Jack (Bauer), 9, 8 and so on down to 2. Five
cards, called Alten (old ones), are permanent trumps,
and rank above all the others: they are, in descending
order, 2 of Hearts, 4 of Clubs, 8 of Spades, 8 of Hearts
and 9 of Diamonds.41 Then follow, in descending
order, the Aces, the Jacks, the 6s (called Papen, Pope),
thé 2s (called Twist), the 3s (called Drist) and the 4s
(called Wagen, Carriage, or Viereck, Rectangle) of the
two trump suits, save for any that are Alten. Only the
trumps down to the 6s or Papen inclusive are fully-
fledged trumps able to beat any card of a plain suit;
the 2s can beat Kings or lower, but not Aces, and the
3s Queens or lower, but not Kings or Aces. It seems
likely that the 4s of trumps beat Jacks or lower,
though this is not explicitly stated. The 7s of the
trump suits are even more highly privileged than in
the Swiss games: played to a trick, they have no trick-
taking power, but, led to a trick, they can be beaten
by nothing whatever, not even the Alten. The 8s and 9s
of the trump suits are also privileged: called Freikarten,
when led to a trick they can be beaten only by the
Alten or the Ace or Jack (Bauer) of a trump suit.
Presumably the King, Queen, 9, 8 and 5 of the trump
suits rank in that order and have no power as trumps,
though this is not made plain: any player who holds a
trump 5, called Hunde (Dog), may exchange it for one
of the exposed trumps before the start of play, and this
is a privilege accorded in many games to the lowest
card of the trump suit. The trump suits are called the
gewahlte Farben (selected suits), and the plain suits
buterwahlte (unselected).

(4) In the Thuringian game, there were again some
permanent trumps, ranking above all others: in
descending order, the 8 of Leaves, called der Tolle (the
Madman) or das alte Thier (the Old Beast); the 9 of
Hearts, called das rothe Thier (the Red Beast); and the
9 of Bells, called das gelbe Thier (the Yellow Beast).
The pack used was the 36-card German-suited one,
which thus lacked 3s, 4s and 5s. The highest cards of
the two trump suits, ranking below the three
permanent trumps, were the Ober, called Oberkarniffel

41The article reprinted in Der Alte is in German, and,
though it quotes some specimen remarks by players in
Frisian, it for the most part gives the names of the cards in
German; I cite them in the forms given in the article.

or Landsknecht (Footsoldier), followed by the Unter,
called Unterkarniffel or Büttel (Beadle), and then the 6,
called the Papst (Pope). The 7s of the trump suits,
called die böse Sieben (the evil 7s), had the same
privilege as in the Frisian game; led to a trick, they
automatically won. They could not be led to the first
trick, however. The 8s and 9s of the trump suits (other
than the three Beasts) were called Freykarten; their role
is not explained very clearly, but it appears that,
though without power as trumps, they could be
beaten only by the three Beasts, the Karniffels and the
Popes. The Deuce, the King and the 10 of each trump
suit are said to be the lowest cards and to be merely
for throwing away: evidently they too had no power as
trumps. A player holding a trump 10 could exchange
it for one of the exposed trumps at the start of play.
There is no indication that any of the cards acted as
partial trumps: evidently the cards of the trump suits
were divided into fully fledged trumps and those
without any power as trumps at all. The order in the
plain suits imitated that in the trump suits, namely, in
descending order, Ober, Unter, 6, 8, 9; Deuce, King,
10, 7. An Ober or Unter of a plain suit was called
fauler Schlingel (lazy rascal). The plain suits were
called ungewählte Farben (unselected suits). In the
game as played in the nineteenth-century at Iserlohn,
Westphalia, the 9 of Hearts was the permanent
highest trump, but the böse Sieben, when led to a trick,
could not be beaten even by it; this form therefore
appears to resemble the Thuringian and Frisian more
than the Swiss ones, but no further details are
available. Much of the interest of the Frisian and
Swiss games lies in the system of scoring; in the latter,
there is, during the deal, a pause after the delivery of
cards in each round to two successive players, during
which either side may offer to increase the value of the
round and the other may either accept or surrender.
The details are of no importance for the history of the

Despite the differences between the various
forms, there is a strong family. resemblance
between them. All are simple trick-taking games,
without the obligation to follow suit, in most of
which there are only five tricks per round; in all
there are trumps, and in all the trump 7 has the
peculiarity that it is powerless when played to a
trick, but cannot be beaten, or can be beaten by
only one card, when it is led. Above all, all having
the feature that the ranking of the cards in the
trump suit or suits is eccentric, and that only the
highest cards: actually function as trumps. The
Thuringian game is exceptional in having an
unusual order in the plain suits; and in all forms
but the Thuringian one there is the feature,
unique in trick-taking play, of partial trumps.

188 Part I: History and Mystery
Games can change greatly in the course of their
history, and, in particular, may borrow
interesting features from other games. Trappola,
for example, began as a trick-taking game
without trumps, and later imported this feature;
and dozens of games, including Tarot games,
adopted the practice of bidding which did not
originally belong to them. But, in general, a
game preserves throughout its history that
feature which was characteristic of it and gave it
its particular flavour and identity. It looks very
much as though the special features of Karnöffel
which distinguished it from other games were a
trump suit in which only some of the cards
functioned as trumps, the special role of the
trump 7 and, probably, the presence of partial
trumps, together with the rearrangement of the
ranking in the trump suit. Even if we had no
indication of how the game was played in the
fifteenth and sixteenth centuries, we should have
to regard it as probable that these were original
features of it, and to assume that it was from the
beginning a game played with trumps.

The earliest substantial reference to Karnöffel
discovered by Dr von Leyden is a poem by
Meissner written in or before 1450:42 from it, he
has conjecturally reconstructed the ranking of
the highest cards as being, in descending order:
the Unter, called the karnöffel; the Deuce, called
the süw (Sow); the 3, called the babst (Pope); the
5, called the keyser (Kaiser); and the 4, called the
tüfel (Devil). The term ‘Sow’ for the Deuce was
a common one, since, in many early German and
some Swiss packs, a sow was depicted on each
Deuce. The other four names, however, are
peculiar to the game of Karnöffel, and, though
later attached to cards of other ranks, they were
evidently used from an early stage in its history.
The next references of any substance are from
Bishop Geiler. There are two principal passages,
both somewhat confusing. The first comes from a
sermon originally preached in 1496. Geiler is
comparing the social order with the order of the
cards in play, and begins by remarking that in
(ordinary) card games, there is a fixed order, in
which the King beats the Ober, the Ober the
Unter and so forth. In the Latin version, he then
continues: ‘But now a game has been invented

42 Eyn suberlich höfflich spruch von dem spiel
karnoffelin’, ed. by J.C. von Fichard, Frankfurter Archiv für
altere Deutsche Literatur und Geschichte
, Frankfurt am Main,
1815, part III, chap. 4, ‘Altdeutsche Lieder aus der ersten
Halfte des XV Jahrhunderts’, no. LXI, pp. 293-7.

which is called Kaiserspiel or Karnéffel in which
everything is turned upside down: thus “the 3s
beat the Ober, the 2 beats the King, and so on;
and there occurs a wonderful transformation
(vicissitudo) of Kaisers, as in this game the Kaiser
is made by chance now from this set (cetu), now
from another.’ (The meaning of this last sentence
is obscure, and I have translated it as best I
could.) After speaking of the perversion of the
social order, he comes back to his analogy, and
says further: ‘But now another game has been
invented with the ranking of the cards such that
the Unter beats the King, and the 2 and the 6
likewise, the 3 beats the Ober and the lower-
numbered beats the higher-numbered; and it
may happen that from this number some are by
chance made Kaisers.’44 The German text differs
slightly. In it, the first passage runs: ‘But now we
have a game called Karniffelspiel in which
everything is turned upside down: the 3s beat an
Ober, the 4 beats the Unter, the 2 and the 6 beat
a, King; and a card is turned over, so that now
one is Kaiser, now another becomes Kaiser, as
luck will have it.’The subsequent passage runs:
‘But now there has been invented another game
with the ranking of the cards such that the Unter
beats the King, the 2 and the 6 beat the King, the
3 beats the Ober, and the lower-numbered beat
the higher-numbered, when they are in the game
of Kaiserspiel.’44 The other principal mention of
the game by Geiler is in a sermon preached in
1515. The German text of the crucial passage
runs: ‘The game of Karnöffel has been invented,
in which the lower-numbered cards beat the

43 Sed nunc ludus inventus gui appellatur (keiserspil)
ludus Caesaris: vel (karniffelius) in quo hec omnia
pervertuntur: ita ut tria vincant superiorem: duo regem etc.
fitque mira vicissitudo cesarum: ut in hoc ludo iam de hoc
cetu/ iam de alio fiat cesar ad fortunam ... Nunc aut est
alius ludus inventus in cartas potestatum ut inferior vincat
regem/ et duo et sex etc. tria superiorem et minor maiorem:
si fuerit ex horum numero qui forte fortuitu facti sunt
cesares’, Sermones Prestantissimi Doctoris Fohannis Geilert
Keisersbergtt de Arbore Humana
, Strasbourg, 1515, pp. 132b-

44 Aber jetz so hat man ein spil/ heisset der karniffel spil
karniffelius/ da seint alle ding verkert/ die drü stechen ein
ober/ die fier den undern/ zwei und sechs stechen ein
künig/ und so schlecht man umb/ ietz so ist einerlei keiser/
darnach so würt ein anderer lei keiser/ wie das glück gibt ...
Aber ietz so ist ein ander spil funden in dem regiment/ auff
der karten/ das der underbüb sticht den künnig/ die zwei
und sechs ein künnig/ die drü den obern/ und der minder
den merern/ wan sie von dem keiser spil sein’, Geiler von
Kaisersberg, Das Buch de Arbore Humana. Von dem menschlichen
, Strasbourg, 1521, pp. 138-40.

The Game of Tarot 189

higher-numbered, and the Unter beats the Ober,
and a Kaiser is made by turning up a card; the 2s
beat a King, and the 6 and the 2 beat the Ober
(or: the 6 beats the 2 and the Ober), and the
Carnöffel beats all of them.’ Later he says, of
‘those who play Karnöffel, when they wish to do
so, they can draw a card out, and then shuffle the
cards, so that the same one is exposed and is
Kaiser’. A little later yet in the same sermon, he
speaks of throwing all the cards, the King, the
Kaiser, the Ober, the Banner and the Devil, on to
the fire.

These passages leave a very confused
impression. There is a card called the Devil, but
we are never told which it is. There is also a card
called the Karnöffel, which beats all others, but,
again, Bishop Geiler does not further identify it:
it does not seem on the face of it that it can be the
Unter, since it is stated that the 4 beats the
Unter. There is no mention of a card called the
Pope; and apparently a card, or perhaps two
cards, are selected as Kaisers by a random draw.
More light is thrown on the matter by examining
the Protestant sources. A work of 1537 asks, with
allegorical intention, the following questions.
Why does the Karnöffel beat the Kaiser and the
Pope, when he is a simple footsoldier
(Landsknecht) and the Ober is a cavalryman? Why
is the Pope called the ‘Six’, and why does he beat
the Kaiser with all the cavalrymen and
footsoldiers, with the exception of the Karnöffel,
that is, the selected (erweleten) footsoldier? Why is
the Devil, or the evil 7 (die böse sieben), free as the
devil (teufelsfret), so that neither the Kaiser, the
Pope nor the Karnöfffel can beat him? Why is the
selected (erwelete) Deuce, the lowest and weakest
of all the cards, called the Kaiser? To this last
question the answer is given that many believe
that the Pope has stolen so much from him that,
although he is still called Emperor, he has
become a beggar; the selected (erwelete) 6 has

45 ‘man hat erdacht karnöffelspil/ da stechen die
mindern die merern/ unnd die under die obern/ und macht
man einerley keiser/ die schlecht man umb/ als da die zwei
stechen ein künig/ und die sechs die zwei/ den oberman
und das Carnöffel sticht es alssamen .., die/ die karnöffels
spil spilen/ wenn sie gern einerlei hetten/ so künnen sie es
ziehen daruff/ unnd also mischen/ das der selbenlei uff ligt/
und keiser ist ... Wenn denn das spil uf ist unnd man lang
gekurtzweilt hat mitt der kart/ so zerreift man sie/ unnd
würfft man den Künnig unnd Keiser/ den Oberman das
paner und den tüffel als miteinander in das feuer und
verbrent es’, Die brösamlin doct. Keiserspergs uffgelesen von Frater
Johann Paulin
, Strasbourg, 1517, pp. 109-10.

three times as much as the Deuce, and so it is no
wonder that the triple crown beats the single
crown of the Kaiser. A further question is why
the lazy Fritz (der faule Fritz) beats the 10 or
Banner.46 Another satirical work of 1546 takes
the form of a dialogue between the Pope and the
Devil. From this we learn that neither of the
Devil and the Pope beat the other; that the Pope
beats all the cards, including the Kaiser and the
Kings, with the exception of the Karnöffel; that
the Karnöffel beats the Pope, the Devil and all
other cards; that the Karnöffel is an Unter; that
it is the 2, 3, 4 and 5, and only they, that are
called Kaiser; that the 2 beats the King, the
Obers and the other Kaisers; and that the 5 is
beaten by all other Kaisers, and by the King,
Ober, Pope and Karnöffel, but beats only the 10,
9, 8, etc.47

From Geiler’s muddling remarks it might seem
that we were dealing with a game in which there
was a perverse ranking of the cards in every suit;
but, although these Protestant sources seen to
differ in the way the term ‘Kaiser’ is used, the
earlier one reserving it for the Deuce alone, it is
apparent from them that only one of the Unters is
chosen to be the Karnöffel, the highest card in
the game, only one 6 is chosen to be the Pope,
only one Deuce becomes a Kaiser; presumably
also there is only one evil 7 or Devil. It thus
becomes natural to suppose that these are the
Unter, 6, 2 and 7 of some one suit selected by a
chance procedure. It would nevertheless be very
difficult to reconstruct from these passages alone
the precise order of the cards. With a knowledge
of the nineteenth-century Swiss game, however,
the remark that the Deuce, 3, 4 and 5 are all
Kaisers, and that the Deuce beats the King and
the Ober, whereas the 5 beats only the numeral
cards (those from 10 down), becomes extremely
suggestive: we appear to have here the same idea
of partial trumps that is found in the Swiss game.
With the same hindsight, we can make sense of
46 ‘Eine Frage des ganzen heiligen Ordens der
Kartenspiler von Karnöffel an das Concilium zu Mantua
1537’, in J. Voigt, ‘Über Pasquille, Spottlieder und
Schmhäschriften aus der ersten Hälfte des 16.
Jahrhunderts’, Historisches Taschenbuch, ed. Fr. von Raumer,
9, Jahrgang, Leipzig, 1838, p. 418. This was reprinted in
Cyriacus Spangenberg, ‘Wider die böse Sieben ins Teufels
Karnöffelspiel’; see C. Fr. Flögel, Geschichte der komischen
, vol. III, Leipzig, 1786, p. 321. The alternative
name die böse Sieben for der Teufel is Spangenberg’s addition
to the original text.

47 Pasquillus, Newe Zeytung vom Teüffel, 1546.

190 Part I: History and Mystery

Bishop Geiler’s observation that the 6 and the 2
beat a King, the 3 beats an Ober and the 4 an
Unter. What we seem, then, to arrive at is
essentially the Swiss game, without the special
role of what, in it, are called Kaisers, namely the
Banner 10s. Within some one suit, the Unter
becomes the Karnöffel, the highest card in the
game; it is followed by the 6 or Pope, and then
the Deuce or Kaiser; next come the 3, 4 and 5 of
that suit, whether or not they are called Kaisers;
and the Devil or evil 7 of that suit plays the
special role that it plays in all the subsequent
forms of the game, with some variation about
whether or not, when led, it can be beaten by the
Karnöffel. The Karnöffel, Pope and Deuce are
unqualified trumps; but the 3 can beat plain
cards only from the Ober down, the 4 those from
the Unter down, and the 5 only numeral cards. It
follows that, in the plain suits, the cards have
their usual ranking, with the King Highest and
the Deuce lowest. Presumably the remaining
cards, King, Ober, 10, 9 and 8, of the selected
suit had no power as trumps.

Evidently Meissner’s poem reflects an earlier .
stage of the game; but it is unlikely that the
special names of cards already used in it, the
Kaiser, the Pope and the Devil, applied to all
those of the relevant denominations: they would.
surely have been names of individual cards of
some one selected suit. We are thus driven to
conclude that a trump suit was a feature of
Karnöffel from the beginning: and since
Karnöffel’ was in existence by 1426, whereas we
have seen reason. to put the origin of Tarot no
earlier than about 1435, this destroys the claim of
Tarot to have been the very first card game
played with trumps. Was there any connection
between the two? Dr von Leyden, impressed by
the presence of the Pope, the Emperor and the
Devil in both games, is persuaded that there
must be. The reason is unconvincing: all three
figures were very familiar to men of the fifteenth
century, and their occurrence in both a German
and an Italian game requires no hypothesis of
influence either way to explain it. But, before we
go further into this question, let us assume that
there was such a connection, and ask, further,
whether this invalidates our hypothesis about the
origin of the Tarot pack.

On the face of it, it does: we can no longer say
that the invention of the Tarot pack represented
an independent invention of the idea of trumps.
But on reflection we see that we have here two
rather dissimilar ideas. In the later forms of
Karnöffel, there is a large number of trumps,
owing to some cards being permanent trumps
and, in some versions, to there being two trump
suits. But, in the earlier form, there are just three
cards of one suit, the Karnöffel (Unter), Pope (6)
and Kaiser (Deuce), lifted out of their natural
order to become fully-fledged trumps, three more,
the 3, 4 and 5, which serve as partial trumps, and
one, the Devil (7), which plays a special role. Let
us suppose that the inventors of the Tarot pack
were familiar with Karnöffel and got their idea
from it: what they did with it was still something
radically new, namely to add to the pack twenty-
one special cards to play the role that the three
top cards played in Karnöffel. They might have
thought, instead, of making all the cards in some
one suit behave in that way; but, on the
assumption that Karnöffel was the only game in
which trumps had hitherto appeared, it would
have been in no way unnatural for them to have
done what we are supposing that they did. The
etymological evidence makes it clear that it was
from tarot, the game originally: known as
Triumphs, that the idea of trumps was taken into
card play generally. In the Swiss Karnöffel rules,
the word Trumpf is used for the trump suit as a
whole, those members of it which actually
function as trumps being called Stecher. But in the
accounts of the Thuringian and Frisian games,
we find instead the adjective gewählte or erwählte
(selected), as we do in the early sources, which
never use Trumpf or any cognate word; here the
original technical term, antedating the term
Trumpf, which derived ultimately from the Italian
word trionfi as used of the Tarot trumps, has
been transmitted over the years right down to
the twentieth century. That the game of Tarot,
rather than that of Karnöffel, must have been the
original source of the idea of trumps as it features
in countless European trick-taking games played
with the regular pack is also evident from the
form which that idea took. If Karnöffel had been
its source, we should expect that, at least in some
cases, the special features of Karnöffel would be
found in other games, namely that not all the
cards of the trump suit function as trumps, and
that some are merely partial trumps. Yet these
features are quite unknown outside Karnöffel: in
all other games, a card either has unlimited
power as a trump or none, and every card of the
trump suit, or in sequence with it, is a genuine
trump. The strongest hypothesis concerning the

The Game of Tarot 191

relation of Karnöffel to Tarot does not impugn
our thesis about the origin of the Tarot pack and
its later consequences for card play; it merely
diminishes the originality of its inventors.

But was there any actual connection between
Karnöffel and Tarot? There is some reason to
think so. The preferred Swiss name for Karnöffel,
Kaiserspiel, was already familiar to Bishop
Geiler, who, in one passage quoted above, uses
the term Keiserspil, or, in Latin, ludus Caesaris. A
Latin text from Würzburg contains a passage
relating to the period 1443-1455 concerning unus
guidam ... ludens ad cartas ludum vocatum imperatoris

(a certain man ... playing at cards the game
called the Emperor’s game).48 Ludus Imperatoris
would be quite a natural Latin rendering of
Kaiserspiel, and may therefore denote the game
called by Geiler Ludus Caesaris, i.e. Karnöffel. In
this case, the game may also have been known at
Ferrara. In 1450 a certain Andrea di Bonsignore
of that city was paid 2 lire for painting two packs
of carte da Imperatori;48 around 1454 Borso d’Este
played at cards dette dell emperatore (called ‘of the
Emperor’), and an account-book for the years
1452-7 records two payments, at 12 soldi a pack,
for carte da imperaturi or de imperatore, 49 Since
Ferrara was a principal early centre for the game
of Tarot, and may have been its birthplace, it is
hard not to suspect that, if Karnöffel were known
there soon after the invention of the Tarot pack, it
was known before that event, and gave the original
idea which led to it. If so, Karnöffel is the remote
ancestor of Whist, Triomphe, Ombre and all the
rest, by a line that leads, surprisingly, through the
games which the Tarot pack was invented to play.

This conclusion is far from certain. If the carte
da imperatore
made in Ferrara were cards intended
for playing Karnöffel, they were presumably
48 W.L. Schreiber, op. cit., p. 52; the text is a manuscript
Tractatus de contractibus by Paulus Wann, Cod. lat. man. 4695
p. 37 and Cod. lat. 12 730 p. 56b in the Staatsbibliothek in

49 G. Campori, ‘Le Carte da Giuoco dipinte per gli
Estensi nel Secolo XV’, Atti ¢ Memorie delle Reali Deputazioni
di Storia Patria per le provincie modenesi e parmensi
, vol. III,
Modena, 1874, pp. 123-32; see p.127.

50 See G. Bertoni, op. cit., p. 218, for Duke Borso’s game,
and fn. 3 for the entries in the account-book of the court.
W.L. Schreiber, op. cit., p. 96, mentions carte dei Imperatori as
being painted in Ferrara in 1450, but gives no reference; he
may be alluding to Gampori’s article.

German-suited cards, and there is no other
indication that German-suited cards were
known, let alone made, in Italy during the
fifteenth or even the sixteenth century. More-
over, Karnöffel appears to have been from the
start what it was in later times, a game of the
common people; it is therefore difficult to
imagine its coming to be known in the
elegant court of Ferrara. More probably, the
two games of Karnöffel and Tarot were uncon-
nected, and represented different independent
approaches to the idea of trumps as we are
familiar with it in games like Bridge. In
Karnöffel, one of the ordinary suits is used, but
only some of the cards in it are given total or
partial power as trumps; in Tarot, a whole new
series of picture cards is added to the pack. Just as
bidding became part of many positive trick-taking
games that originally lacked that feature, so the
idea of trumps was incorporated into almost all
such games, from some of which, such as
Trappola, it is known originally to have been
absent: the most notable such games to resist the
idea of trumps are Piquet and those of the popular
Italian Tressette family. It happened to be Tarot
from which the idea was taken over into other card

Whether or not Karnöffel served as an
inspiration to those who invented the Tarot pack,
the reason for its invention cannot be doubted,
namely to embody the new idea of trumps in a
special additional sequence of cards. It was
motivated, therefore, by considerations that had
to do, not with magic, or the occult, but with the
only use to which playing cards were then put,
that is, to play card games; that has been the
principal contention of this chapter.


General Features of the Game

We have now to describe the very varied and
numerous forms of game that have developed
from the original Italian game of trionfi since its
beginnings in the fifteenth century, many of them
still played in different parts of Europe, some
long obsolete. It would be agreeable if we could
simply take them in historical sequence, tracing
their evolution over the course of the centuries.
Unfortunately, this is difficult to do, because of
the gaps in our knowledge. The earliest explicit
description of any form of Tarot game is from a
French card-game book of 1659, which describes
several French versions and one Swiss one.
Despite some lacunae which have to be filled by
guesses, we can trace the evolution of the game in
France, Germany, the Habsburg dominions and
elsewhere, from the eighteenth century onwards,
from numerous descriptions in German, French
and other languages; its history in the French-
speaking cantons of Switzerland is uniform with
that in France itself, but, in the rest of
Switzerland, it appears to have changed very
little from the seventeenth century to the present
day. Assuming the game to have been introduced
into France and Switzerland in about 1500, and
into Germany in about 1600, we are, indeed, left,
in both cases, with about a century and a half of
unrecorded history; but it seems likely that there
had been comparatively little evolution before the
earliest descriptions in French and German were

In order to provide a strictly chronological
account, however, we should have to start with
Italy, where the game began. There we find, not
one continuous tradition of play, but four distinct
ones, corresponding to four main centres where
the game had taken root before the close of the
fifteenth century — Ferrara, Milan, Bologna and
Florence. The type of game played in Bologna
has always been peculiar to that, city and its
immediate neighbourhood, and was formerly
known as Tarocchino, from being ‘played with a
shortened pack of only 62 cards. .This game is
still very much alive, and we have descriptions of
it going back to the eighteenth century, one of
which gives information about the mode of play
in the preceding century. It is reasonable to
assume that the general characteristics of the
game go back to the beginning of Tarot play in
Bologna in the fifteenth century, although the
shortening of the pack probably did not occur
until the first half of the sixteenth; but we have no
ground to draw inferences from this to the kind of
Tarot game played in any other part of Italy.
Florence was responsible for another, even more
unusual, form of Tarot, Minchiate, played with a
specially devised expanded pack of 97 cards.
Minchiate, as played at Florence, came to be
very popular at Rome as well, and also, in a
different form of which we have no record, at
Genoa. We have many descriptions of Minchiate,
one from the seventeenth century; but, again, we
are dealing with a very special form of Tarot
game. In Bologna, the shortened 62-card pack
entirely superseded the original 78-card form,
but in Florence and Rome the 78-card pack
survived, alongside the more popular Minchiate
pack, until the seventeenth century. We have no
direct information about the type of game played
in those cities with the 78-card pack, but it is
probable that the game still played in Sicily is a
descendant of it: though the Sicilian pack was
reduced to 63 cards in the course of the
eighteenth century, the game had been

196 Part II: Games with 78 cards

introduced into Sicily, almost certainly from
Rome, in 1663, and was originally played there
with a full pack of 78 cards. This represents a
tradition of play which may well go back to quite
early times in Florence, but probably has little
connection with the style of play in more
northern parts of Italy.

To the type of game played in Ferrara, and
probably also in Venice, we have only a few
clues, and cannot reconstruct it with any
certainty. But it is Milan that poses our greatest
problem. We have eighteenth-century
descriptions both of the game of Tarocco, as
played with the 78-card pack in Milan and in
Lombardy generally, and of the distinct, but
broadly similar, games played in Piedmont. The
probability is, however, that these games do not
represent continuous local traditions of play.
From at least the beginning of the eighteenth
century, French cardmakers were exporting
Tarot packs to Savoy, then an independent state
comprising Piedmont as well as Savoy proper,
which became part of France only in 1880. Pro-
bably the earliest surviving pack of this kind is one
made by Jean Dodal of Lyons, who was active
there from 1701 to 1715; this pack, of which
one example is in the Bibliothéque Nationale
in Paris, and another in the British Museum,
bears, on several cards, the inscription F.P.LE
TRENGE, ie. ‘made for export’1 The type of
design which it exemplifies is a slight variant on
the celebrated standard pattern known as the
Tarot de Marseille, and is ancestral to the
modern Tarocco Piemontese. It differs from the
Tarot de Marseille only in a few details; among
these are the appearance of a face on the belly of
the Devil (trumps XV), a full face rather than a
profile on the Moon (trump XVIII), the presence
on the Judgment (trump XX) of the droplets
that appear on several other cards and the use of
the name LE FOL or LE FOU instead of LE
MAT for the Fool. By the second quarter of the
century this earliest form of the Tarocco
Piemontese was being: made in Piedmont itself;
the earliest example known to me is one by
Giuseppe Ottone of Serravalle made in 1736 and
now in the Museo de Naipes Fournier in Vitoria.2

1 The British Museum example is F-5 in O’Donoghue’s
catalogue. The 2 and Cavalier of Cups of the Bibliothéque
Nationale copy are illustrated in H.-R. D’Allemagne, Les
Cartes à jouer
, vol. 1, Paris, 1906, p. 192.

2 No. 25 in the Italian section of the catalogue. A ‘Spanish
Tarot’ was issued by the firm of Fournier a few years ago,

A letter of Thomas Gray testifies to his having
seen Tarocco played inTurin in1739.4

All this strongly suggests that the game of
Tarot was introduced into Piedmont in the
eighteenth century from France via Savoy; and
there is no direct evidence of the game’s having
been played in Piedmont before that date. There
is, however, one curious detail of Piedmontese
play that suggests that it was: namely that the
Angel or Judgment (trump XX) is always treated
as superior to the World (trump XXI), an
eccentricity that continues to this day. In the
Tarot games played, now or formerly, in Bologna,
Florence, Rome and Sicily, the Angel, or the card
corresponding to it, does indeed rank higher than
the World; it is difficult not to suppose that, in
this detail of the Piedmontese games, we have a
survival from a previous period when the game
was played with some quite different type of
pack. However this may be, the type of pack used
in Piedmont from the eighteenth century down to
the present day was certainly of French origin,
and this may well be true of the games played
there also; we cannot extrapolate with any
assurance from the Piedmontese games, as
described in the earliest printed sources, to
games played in Italy before the eighteenth

From the 1740s onward, cardmakers in various
parts of Italy began producing Tarot packs with
78 cards whose design constituted another slight
variation on the Tarot de Marseille; since, so far
as we know, these cards were used only in
Lombardy, it may be called the Lombard
pattern.5 The actual designs again differ only
slightly from the Tarot de Marseille; the chief

based on this pack, but with inscriptions in English at the
top of each trump and in Spanish at the bottom of each
trump and court card; it cannot be relied on as otherwise a
faithful reproduction. An illustration of it is‘in S.R. Kaplan,
The Encyclopedia of Tarot, New York, 1978, p. 277.

3 See The Letters of Thomas Gray chronologically arranged from
the Walpole and Mason Collections
, vol. 1, London, 1827, p. 54.
Gray, writing from Turin on 16 November 1739, speaks of
his having seen taroc played there.

4 The date 1650 given in the first edition of the catalogue
of the Fournier Museum for no. 12 in the Italian section, a
pack of Lombard type made in Gorizia, is certainly about
100 years too early. For illustrations of the eighteenth-
century Lombard pattern, see C.P. Hargrave, A History of
Playing Cards
, New York, 1930, 1966, pp. 226, 232. For
illustrations of nineteenth-century versions, see Roger Tilley,
Playing Cards, London, 1967, pp. 16-17, and Kaplan, op. cit.,
pp- 154,160, 164.

General Features of the Game 197

divergence in this respect from the ‘Tarocco
Piemontese is that the Devil (trump XV) has no
face on his belly but wears a pair of furry
trousers. The principal difference between the
two forms of pack is that those of the Lombard
type had the narrow format generally typical of
Italian cards, the backs being folded over to form
borders for the faces of the cards, whereas the
Piedmontese cards are quite wide and do not
have the backs folded over. On the testimony of
Joseph Baretti, the game of Tarocco was highly
popular both in Lombardy and Piedmont by

The Lombard variation on the Tarot de
Marseille appears to have been of Italian origin,
but obviously testifies to strong French influence.
We have only a very little information about the
design of Tarot cards used in Milan during the
two hundred years or so before 1740, and no
direct information about the type of Tarot game
played there during that period; but it is
probable that neither the Tarot de Marseille-
derived designs nor the type of game that became
popular in Lombardy represented so abrupt a
departure from tradition as seems to have
occurred in Piedmont. In Lombardy, the World
was, and still is, given the rank appropriate to its
number (XXI), as superior to the Angel or
Judgment (XX); and this accords with what
appears to have been the continuous Milanese
tradition from the earliest times. Moreover, the
Tarot de Marseille designs themselves are likely
to have had a fairly close resemblance to those
long traditional in Milan. Nevertheless, the
probability is that the introduction of the
Lombard pattern signalises a historical
discontinuity, just as does that of the Tarocco
Piemontese. Both the Lombard and Piedmontese
patterns share with the Tarot de Marseille two
striking features: the particular order and
numbering of the trumps; and the inscriptions at
the bottom of the trumps and court cards of their
names in full (in addition to the numerals at the
top of the trumps and at the sides of the numeral
cards of the suits). The order which the trump
subjects have in the Tarot de Marseille goes back
to the mid-sixteenth century in France. Although
the order observed in Milan was. certainly
similar, we have no reason to suppose that the
precise order employed in the Tarot de Marseille

5. Joseph Baretti, An Account of the Manners and Customs of
vol 2, London, 1768, pp. 219-21.

was ever known in Italy before the eighteenth
century. We cannot rule it out as impossible that
this impression is due only to our ignorance; but
the evidence from the inscriptions appears
decisive. Nowhere in Italy was it the regular
practice to inscribe their names on the trump
cards. We have specific evidence that it was not
done in Milanese packs before the eighteenth
century; before that time, such names are found
in Italian packs only of an obviously non-
standard type, such as the Sola-Busca farocchi. In
France, however, the practice goes back to the
seventeenth century, and, so far as we know, was
always observed in the Tarot de Marseille. What
appears to clinch the case is that, in the earliest
versions both of the Tarocco Piemontese and of
the Lombard pattern, the inscriptions were in
French. Moreover, these French inscriptions
employ precisely those names used for the trump
cards in the Tarot de Marseille; particularly
noteworthy is the use for the trump XVI of the
surprising name La Maison Dieu, meaning ‘The
House of God’ or, possibly, ‘The Hospital’, a
name by which it is always known in the Tarot
de Marseille, but for which no equivalent term
has ever been used in Italian. It was only at a
later stage that the inscriptions in both types of
Italian Tarot de Marseille-derived pack were
rendered into Italian, La Maison Dieu then

becoming La Torre (the Tower). In the same way,
“the early packs follow the usual Tarot de
Marseille custom of omitting the name of Death
(trump XIII), inscribed La Morte when the
inscriptions were later translated into Italian;
and, in the early packs, the third court card of
each suit is labelled Cavalier or Chevalter, although
in Italian this card has always been known as
Cavallo (Horse), which name was used when the
inscriptions went into ltalian.

Now we can, no doubt, imagine a vogue for a
French type of design, and, with it, for the
French practice of inscribing their names an the
trump cards, affecting the designs employed by
Italian cardmakers without any other break in
continuity. But this leaves the use of inscriptions
in the French language unexplained. There was
not in Lombardy, as there was in Savoy, ‘a
substantial French-speaking population; if it had
been a matter merely of the introduction of a new
style of design to replace an old one, we should
surely expect that, from the start, the inscriptions
would have been in Italian, using the names
customary among players of the game. The fact

198 Part II: Games with 78 cards
that the cardmakers not only took over, almost
intact, the Tarot de Marseille designs, but at the. .
same time adopted the French inscriptions,
strongly suggests that the introduction of the
Lombard pattern represented the reintroduction
into Lombardy of the Tarot pack itself and of the
game played with it, that game having died out
there at some earlier date, perhaps by the end of
the seventeenth century; this may well have
happened in Piedmont also, although there the
gap between the demise of the indigenous game
and the reintroduction of Tarot games from
France must have been short enough for some
players to have remembered the traditional
ranking of the Angel and the World relative to
each other. On this hypothesis, it is entirely
plausible that it should have been the Tarot de
Marseille designs that were borrowed: by the
early eighteenth century, there was probably no
form of 78-card pack surviving in mainland Italy.
The Bolognese pack had long ago been reduced
to 62 cards; in Florence and Rome, the 78-card
pack had probably succumbed to competition
from the 97-card Minchiate pack; in Ferrara and
Venice, all forms of Tarot game seem to have been
long defunct.

It follows that we can no more assume that the
form of Tarot game we find played in Lombardy
in the eighteenth century represented a contin-
uous indigenous tradition than we can make the
parallel assumption for Piedmont. As a matter of
fact, we shall later see that, if we did make this
assumption, our deductions would probably not
be seriously in error; but this is a point that needs
to be argued for, certainly not to be assumed.
The general upshot is this. We can, with fair
confidence, reconstruct the history of Bolognese
Tarocchi, and can even speculate from what
game played with the full 78-card pack the game
of Tarocchino might have been derived. We
know how Minchiate was played, and may
reasonably presume that its main features were
present from the start. We can make only vague
conjectures about the game played in Florence
and Rome with the 78-card pack, conjectures
based on the Sicilian game and on Minchiate. All
of these, however, are forms of Tarot game which
had no influence on the games played with the
78-card pack in other parts of Europe. We have
little direct evidence for the type of Tarot game
played in Ferrara and Venice, and none for that
played in Milan before the eighteenth century;
any conclusions we reach about the latter can be
based only upon a presumed relationship with
games played in other countries. It has therefore
seemed better to abandon any attempt at a
strictly chronological account. Instead, I shall, in
Part II, describe those Tarot games played with
the 78-card pack outside Italy, together with
those played in Piedmont and Lombardy from
the eighteenth century onwards. Part III will be
exclusively devoted to Italy; the games of which
we have actual accounts, or which are still
played, will first be described, after which it will
be possible to speculate about those games for
which we have no conclusive evidence. Part IV
will describe that great family of Tarot games in
which the card originally known as the Fool has
assumed an exalted rank.

In our survey of all the various forms that the
game of Tarot has assumed at different periods
and in different countries, we shall encounter
certain features again and again; and, to avoid
pointless repetition, it is well to adopt a constant
terminology for these at the outset.

General terminology for Tarot games

Stages of the game

All card games take place in stages, punctuated by a
dealer’s distributing the cards to the other players and
to himself; and, as stated in the last chapter, we shall
use the term round to apply to the period of play
between one such deal and the next. Card players are
more accustomed to use ‘hand’ or ‘deal’; but these
can easily lead to ambiguity when an unfamiliar game
is being explained. In some games; the players simply
play as many rounds as they feel like, adding the score
for each to the curnulative total. In others, several
rounds in succession may be required before the score
can be settled; for instance, the winner or winning
side may be stipulated to be the one first reaching, in
several rounds, a certain fixed cumulative total. In
Italian, such a set of several rounds is usually called a
partita, in French a partie: in this book, it will be called
a game, following the usual English nomenclature. At
the beginning of each round, the deal may be said, in
a different sense, to consist of several rounds, one for
each time that the dealer comes back to the starting-
place; but this should cause no confusion.

Ranking of the cards

As already explained, Tarot games are trick-taking
games; any reader who does not clearly understand
what this means is referred to the explanation in

General Features of the Game 199
Chapter 7. The trick is always won by the highest
trump played to it, or, if no trump is played, by the
highest card of the suit led. The trumps usually:rank
in sequence from the XXI (high) down to the I or
Bagatto (low): any exceptions to this will be explicitly
stated. The top four cards of each suit are invariably
King, Queen, Cavalier and Jack (or Maid), in that
order. When the numeral cards rank, in Swords and
Batons, from 10 (high), 9 and so on down to Ace
(low), but, in Cups and Coins, from Ace (high), 2 and
so on down to 10 (low), it will be said simply that the
suit cards rank ‘in their original order’. Sometimes,
however, this practice is abandoned, and the numeral
cards rank in every suit, after the court cards, in the
order 10 (high) down to Ace (low); in such a case, it
will be said that the suit cards rank ‘in the simplified
order’. Play is usually counter-clockwise, but sometimes
clockwise; it will be taken for granted that the dealer
deals in the same cyclic order as that in which play
proceeds, and that the deal passes, from one round to
the next, in the same direction. A player with the
lead is free to lead any card, and the subsequent
players to any trick must follow suit if they can,
including playing a trump if a trump is led, and, if
they cannot follow in a plain suit, must play a trump if
they have one regardless of any other considerations.
Apart from these constraints, a player is usually free
to play any card that he likes; the exceptions will of
course be noted. All this will be expressed by saying
that ‘the usual constraints’ are observed in play.

In almost all Tarot games played with the 78-card
pack, the Fool is treated in the manner already
explained. It has no trick-taking power, and may be
played at any time, irrespective of any obligation to
follow suit or to play a trump. It cannot, however, be
captured, save in exceptional cases, and is therefore
taken back from the centre of the table, and placed
with the other cards won in tricks by that player or
side; very often, it is not the custom to play it to the
centre of the table, but simply to show it to the other
players before placing it with the cards won in tricks.
More often than not, another card must be given in
exchange for the Fool, from among those won in
tricks, either immediately or at the end of the round;
but since this is not an invariable rule, it will be
expressly noted. Whenever the Fool is treated in this
general manner, this will be conveyed by saying that
‘the Fool serves as Excuse’: in French Tarot, the card is
actually called Excuse, because it excuses the one
who plays it from following suit or trumping, and this
provides us with a convenient term for indicating this
special role of the card in play. Unless otherwise
stated, the Fool may be led ta a trick, and the next
player may then play any card, which is treated just
as if it were the lead, other players having to follow
Suit to it if they can, or, if they cannot, to trump if they
can. Usually, if the player or side which had the Fool
makes no tricks at all, the Fool must be surrendered at
the end of the round to the player who won the trick to
which it was played.


As we have seen, the object of the game is almost
always to make as many points as possible on the
cards won in tricks, different cards having different
point-values. The cards are divided into those having
no special point-value, which will be called ‘low. cards’
(a term which will always refer to their poini-values, not
to their trick-taking power), and ones having a high
point-value, which will be called ‘counting cards’. In a
great many games, the only counting cards are the
XI and the I or Bagatto, that is, the highest and the
lowest trumps, the Fool, and the court cards of the
four suits; all the numeral cards, and the trump cards
from II to XX, are then low cards, In some games,
however, there are other counting cards. In games in
which the counting cards are the nineteen
enumerated above, the most usual point-values are as
follows: .

The XXI: 5 points
The trump I: 5 points
The Fool: 5 points
Each King 5 points
Each Queen 4 points
Each Cavalier 3 points
Each Jack 2 points

In certain games, however, even when there are no
other counting cards, the point-values differ from
these standard ones. When there are only these
nineteen courting cards, and they have the values just
stated, this will be expressed by saying that ‘the
counting cards have their standard values’.

What needs considerable explanation is the
method, common to most, though not to all, Tarot
games, of reckoning the total point-value of the cards
won in tricks by a player or side. This is apt to strike
someone who first encounters it as eccentric and
puzzling; but the rationale of it is soon understood.
The cards are counted out in sets, which may be sets
of two, of three, of four or even of some larger number,
according to the particular game being played. Let us
suppose-first that they are counted out in pairs. Then,
normally, the pairs will be so arranged that each
counting card is accompanied by a low card: the value
of such a pair is precisely the number of points given
as the point-value of the counting card. For instance,
when the point-values are as above, a pair consisting
of a Queen and a low card will together be worth 4
points, the point-value of a Queen being 4. Each pair
of low cards will be worth 1 point. If, however, the
cards cannot be arranged so as to pair off a low card
with every counting card, there will be some pairs
consisting of two counting cards: the value of such a

200 Part II: Games with 78 cards
pair is 1 less than the sum of the point-values of the
two counting cards. For instance, with the above
point-values, the value of a pair consisting of a King
and a Cavalier will be 7 points, one less than the sum
of the values of the King (5) and of the Cavalier (3). It
must be stressed that it will make no difference to the
total of points on a player’s cards how he groups
them. Suppose a player has won in tricks just six
cards, the XXI, a Queen, a Jack and three low cards.
If he groups each counting card with a low card, he
will have one pair worth 5 points (for the XXI), one
worth 4 points (for the Queen) and oné worth 2 points
(for the Jack): his total will be 5 + 4 +2 =11 points.
Suppose that instead he takes the XXI with the
Queen. That pair will then be worth 5+ 4 - 1=8
points. He will also have a pair consisting of the Jack
and a low card, worth 2 points, and another
comprising two low cards, worth 1 point, making a
total of 8+ 2+ 1=11 points, the same as before.

When the cards are counted in threes instead of in
pairs, a set consisting of a counting card and two low
cards will be worth the poirit-value of the counting
card; e.g. a Cavalier and two low cards will together
count 3 points. A set consisting of three low cards will
count 1 point. A set consisting of two counting cards
and a low card will be worth one less than the sum of
the point-values of the two counting cards; e.g. the
Fool, a Jack and a low card will, on the above values,
be worth 5 + 2 – 1 = 6 points. A set consisting of three
counting cards will be worth twe less than the sum of
the point-values of the three cards: e.g. the trump I, a
King and a Cavalier will be worth 5 +5+3 - 2=11
points on the above values. Again, it does not in the
least matter to the total score how the cards are
grouped. If, instead of saying that the low cards have
no value, we say that each has a point-value of 1, then
it will be a general rule that each set of three cards has
a combined value 2 less than the sum of the point-
values of the three cards; it is then obvious that it will
make no difference how the cards are distributed
among the sets of three.

The same goes if the cards are counted in fours. A
set of four consisting of one counting card and three
low cards will be worth the point-value of the
counting card. A set consisting of four low cards will
be worth 1 point. A set consisting of two counting
cards and two low cards will have a total value one
less than the sum of the point-values of the counting
cards. One consisting of three counting cards and one ,
low card will be worth two less than the sum of the
values of the counting cards. And a set of four
counting cards will be worth three less than the sum of
their point-values. Once again, if we were to regard
each low card as having a point-value of 1, we could
say generally that each set of four had a combined
value of 3 less than the sum of the values of the
individual cards.

This method of reckoning the points won in any
round by each player or side is so extremely
common, and so perplexing at first sight, that I have
thought it right to labour it here, so that it will be
quite clear before particular games are described.
Having thus explained it generally, I will, in
particular cases, say merely that the cards are counted
in pairs, in threes or in fours, and leave it to be
understood that the method just described is
intended. The only question that remains to be settled
is what to do with any odd cards left over: with a
single card when the cards are counted in pairs, with
one or two cards when they are counted in threes, or
with one, two or three cards when they are counted in
fours. In some games this cannot happen, in others it
can; and, when it can, there are different ways of
dealing with it, which will be explained for each
game. The matter is simplest when the cards are
counted in threes. In this case, the usual practice is
that two odd cards that remain over after the cards
have been counted out in threes count as if there were
three of them: if they are both low cards, they count 1
point, if they include one counting card, they count
the value of that counting card, and, if they are both
counting cards, they count one less than the sum of
their values. A single odd card that remains over then
counts nothing if it is a low card, and one point less
than its value if it is a counting card. A player will
never have so many counting cards that he cannot so
arrange his cards that the odd ones reniaining over
are all low cards. Hence, when the method just
described is adopted, this may be succinctly indicated
by saying that ‘two odd low cards count 1 point and one odd
low card counts nothing’.

Readers who did not skip the last chapter will at
once see the rationale of this method of reckoning
points, which otherwise may at first be obscure.
Suppose we have a three-handed game, in which the
cards are counted out in threes; there will then be no
cards left over when a player counts his cards, because
there must be three cards in every trick. Now it is
plain that an apparently quite different method of
reckoning the total number of points from that
described above would yield the same result. Suppose
that, instead of counting out the cards in threes in the
above manner, a player were to reduce the point-value
of every counting card by one, and simply reckon up
the total of these reduced point-values, adding 1 for
every set of three cards: a moment’s reflection will
show that he will arrive at the same total as by the
other method. Since, in a three-handed game, each
trick consists of three cards, adding one point for
every set of three is tantamount to adding one for
every trick taken. And, indeed, in very exceptional
cases, we find the rule for reckoning points given in
precisely this way: the point-value of each counting
card is cited as being one lower than those usually
given, and there is stated to be an additional point for
each trick won. Just this is the method of reckoning

General Features of the Game 201

points employed in the Spanish game of Malilla and
its more famous French offspring Manille: and there
can be no doubt that it was the original rule in Tarot.
Presumably the nearly universal method that was
described above arose simply as a matter of
convenience. Players found it tedious to carry out two
operations, one of counting the number of tricks and
one of adding up the point-values of the counting
cards, and fused the two together by adopting the
foregoing procedure. At first, the combined operation
would be thought of as consisting of adding 1 to the
sum of the point-values of the counting cards in each
trick. But, in the course of time, since there would
most often be at most one counting card in a trick,
and since it might, in any case, be easiest to count by
rearranging the cards so that no more than one
counting card was included in a trick, the point-values
associated with the counting cards must have come to
be increased by one; and thus the method of
reckoning that has been described would have been
arrived at.

It is, however, evident that, very often, the original
rationale of the procedure was forgotten. Let us say
that, when the cards are counted out in pairs, 2 is the
‘base number’ in that game, and, when they are
counted out in threes or in fours, that 3 or 4 is the
‘base number’. Then the original rationale of the
method requires the base number always to be the
same as the number of players, and thus as the
number of cards in each trick. But, as we shall see,
this is not always the case: there are three- or four-
handed games in which the base number is 2, and
four-handed games in which it is. 3. In all such cases,
the probable explanation is that, at some former time,
the most popular form of the game was one in which
there were that number of players given by the base
number: that the practice of counting the cards out in
threes is testimony to the former predominance of a
three-handed game, and that of counting them out in
pairs to the popularity of a two-handed form. The
rationale of the method of reckoning points being
forgotten, the practice of counting the cards out in
pairs or in threes (of using 2 or 3 as the base number)
was then thoughtlessly transferred to games with a
different number of players. It is obvious that, the
lower the base number, the greater becomes the
comparative importance of winning low cards in
tricks: of course it is always of importance to win
counting cards.

Any beginner at Tarot must accustom himself to
counting out his cards in this manner. At first, the
adding and subtracting seem to involve a great effort;
but, as with everything of the sort, it quickly becomes
second nature.

Card points and game points

A frequent source of confusion in descriptions of card
games is the use in different senses of the word
‘points’. Very few card games are, like chess, simply
won, lost or drawn: in virtually all the score is
reckoned in points, and, when the play is for money,
settled at so much per point, per hundred points, or
the like. In some games, such as simple trick-taking
games like Bridge, the points in which the score is
kept are the only ones involved, since the outcome of
each round depends upon the number of tricks made
by each side. But, in complex trick-taking games like
Tarot, it is often important to maintain a sharp
distinction between the points won on cards taken in
tricks in any one round and the score for that round.
In the simplest cases, indeed, there may be no such
distinction: at the end of each round, a player may
simply write down as his score the number of points
that he won on the cards that he captured in play. It is
reasonable to suppose that this was the original
system. But very often, it is not like this. Very often,
the points won on the cards taken in tricks serve
simply to determine which player or side has won the
round: the points then allocated to:each player as his
score are then computed according to some different
system, To avoid ambiguity, I shall, as stated in
Chapter 7, use the expression ‘game points’ to mean
those that are written down on the score card and
used at the end of play to determine the final money
settlement between the players, or that are translated
immediately into cash terms at the end of each round.
The unqualified term ‘points’, on the other hand, will
be used exclusively for those by reference to which it is
determined, at the end of each round, which players
have won and which have lost, for instance the points
on the cards won in tricks. It would be possible to be
even more explicit, and use the term ‘card points’ for
these, as we did in Chapter 7, But this would not only
be cumbersome, but also, in some cases, inaccurate;
for, in some games, the points that go to determine
win or loss of a round will include others besides those
won on individual cards, for instance a bonus for
winning the last trick or for having had some special
combination of cards in one’s hand at the start of
play. The term ‘card points’ will, however, sometimes
be used in this sense, when’ it is necessary to
emphasise the distinction from game points, and no
inaccuracy is involved. Very occasionally, there is also
an intermediate stage between the card points and the
game points: some system of reckoning that is one
level above the points won in each round, but does not
yet represent the final score on which the settlement is
based. In such a case, I will avoid confusion by not
using the word ‘points’ at all for this intermediate
level, but employ some other suitable term.

With these preliminaries, we are now ready to
embark on our study of the various different
games that have been and are played with the
Tarot pack.

Re: Dummett's Game of Tarot, 1980, a few chapters

More recently, Mr Jan
Bauwens has claimed that a pack of playing
cards recorded in the Register of Duke
Wenceslas of Brabant
as having been bought for
the Duke and Duchess was a Tarot pack, on the
ground that it contained 78 cards;(3) but a
reference to the original entry reveals that neither
it nor any of the numerous later similar entries
contains any mention of the number of cards in
the packs bought or played with, nor anything
else to suggest that these were not
straightforward regular packs. (4)

3. In a booklet accompanying a reproduction of the
Mamluk pack from Istanbul published in 1973 by S.A.R.L.
Aurelia Books, of Louvain and Brussels.
4. The entry is cited in A. Pinchart, Recherches sur les cartes à
jouer et leur fabrication en Belgique, Brussels, 1870.

Forget the ludicrous tarot claim, just having playing cards in the Low Countries at this early point is fairly eye-opening, for the owner of the cards was dead by 1383 (Wiki:):

Wenceslaus I (often called Wenceslaus of Bohemia in chronicles) (25 February 1337 – 7 December 1383) was the first Duke of Luxembourg from 1354. He was the son of John the Blind, King of Bohemia, and Beatrice of Bourbon.

In 1352, Wenceslaus married Joanna (1322 – 1406),[1] daughter of John III, Duke of Brabant and Limburg, and Marie d'Évreux. In 1354 Charles raised Luxembourg to the status of a duchy. In 1355, Joanna inherited Brabant and Limburg. In order to guarantee the indivisibility of Brabant, Wenceslaus signed the Joyous Entry, but had to fight against his brother-in-law Louis II of Flanders, who asserted his share of the duchy. He failed to prevent the seizure of Brussels by the Flemings, but a certain Everard 't Serclaes succeeded by an audacious coup in driving them out of the city. Thereafter, Wenceslaus had to face primarily internal disorders. In 1371, he overestimated his military capacities and waged war with William II, Duke of Jülich, resulting in humiliating defeat at the Baesweiler, losing a part of his army, and several noblemen.[2] He was captured and suffered 11 months of captivity.[2]

Burial Place in Abbaye d'Orval, Belgium
Wenceslaus died in Luxembourg,, leaving Joanna as sole ruler of Brabant, and was succeeded by Wenceslaus II (Wenceslaus IV of Bohemia) as duke of Luxembourg. There are speculations that he might have died of leprosy. His last wish was his heart to be displaced from his dead body and sent to his wife (Joanna stayed in Brussels). He is buried in a crypt at the now-ruined Orval Abbey in Belgium.

Wenceslaus I of Luxembourg wrote the lyric poetry interpolated in Jean Froissart's Méliador, which was identified as his by Auguste Longnon in the 1890s (Wenceslas was a patron of this chronicler). His lyric output comprises 79 poems (11 ballades, 16 virelais, 52 rondeaux).

I don't subscribe to to a single point of Mamluk/Saracen playing card diffusion into Europe theory (multiple points of ingress is more than likely), but if 1370s is when all of the earliest evidence appears, then this is damn early and about as far removed from direct contacts with "Saracens", unless of course it came from the duke's Bohemian connections. Perhaps this was simply one of the earliest German luxury playing decks, although those date from 1430. French encroachments into the Low Countries accelerated in 1364 when the French Duchy of Burgundy was created. Again, just wondering why Marziano's project was in the medium of cards - whence the impetus? More on that elsewhere...


Re: Dummett's Game of Tarot, 1980, a few chapters

Phaeded wrote: 31 May 2020, 18:46 Again, just wondering why Marziano's project was in the medium of cards - whence the impetus?
Isn't your question sufficiently answered by Filippo Maria's particular love of playing cards?

it's also important to remember the purpose of it - a refreshing pastime, not too taxing or involved. This is different from chess, which was more likely to tire the mind than to refresh it. The concentration involved in chess - against a suitable opponent, naturally - is too much for the purpose. And chess was already suitably moralized in any case; there was nothing Marziano could add.

A game that was partly chance and partly skill was one that did not demand perfection in every move in order to win; sometimes you got lucky, sometimes you were dealt a bad hand, an overall evenness that meant that players of differing abilities could play together convivially. Moralists like Pier Paolo Vergerio in De liberalibus studiis (the text bound with De deificatione in BnF latin 8745), written 1402, approved of chance-skill games like this, unlike dice:
Palamedes, as the oldest authors say, invented the game [of tables=chess] during the Trojan War in order to occupy his soldiers with this diversion and keep his army, thus distracted, from mutiny. The game of dice on the other hand either fosters a greed unsuited to a free person, or a softness unbefitting a man. Those who play dice for the sake of gain might be better off engaging in more profitable businesses, but those who hunt pleasure here are slow-witted, as they cannot fina any more honorable pleasure. Pleasure is most suitably taken in games that require some or even great skill, and as little chance as possible.
Craig Kallendorf, Humanist Educational Treatises (2002), p. 87.

(an older translation is more clear, but less literal: "Dice-playing is to be utterly condemned. It is either a base form of money-getting, or an effeminate excitement; though a game of skill, in which chance plays but a small part, is allowable.”

William Harrison Woodward, Vittorino da Feltre and Other Humanist Educators: Essays and Versions. An Introduction to the History of Classical Education. Cambridge University Press, 1897, p. 117. )

Sterner theologians might condemn all card games because they could be used for simple gambling like dice, but most authorities took a more nuanced approach, allowing some while forbidding others, or the betting of too much money.

In any case, Filippo Maria liked playing cards, so it was his choice that mattered. He just wanted a highly moralized kind of cards, which Marziano supplied. Obviously not Christian catechetical morality, sin and redemption, but Virtue-based. This recalls the Senecan stoic oratory of Barzizza, with, in Marziano, a nod to the refined Epicurus.

Re: Dummett's Game of Tarot, 1980, a few chapters

Ross G. R. Caldwell wrote: 05 Jun 2020, 09:54
Phaeded wrote: 31 May 2020, 18:46 Again, just wondering why Marziano's project was in the medium of cards - whence the impetus?
Isn't your question sufficiently answered by Filippo Maria's particular love of playing cards?
To me its a chicken and egg question, in the sense that luxury decks had to exist before Filippo would likely come to love playing them...and the earliest luxury deck I'm aware of is a German deck with a hunting theme from 1430. Decembrio's comment was post-mortem so there is no reason to assume Filippo loved playing cards as early as 1412. We do know of 'Lombard' and 'Saracen' decks among the possessions of the Duchess of Orleans Valentina Visconti in 1408 (inventoried by her son right after she died; and note this is Filippo's half-sister), and I think the 'Lombard' designation allows us to posit cards existing in Milan. But all of this is still extremely early - for Marziano to have invented an entirely new game within this new "genre" is startling to me; I don't think the novelty of this can be over-stated. Filippo's love of cards may have begun with the Marziano deck.

On the flip side, 'Saracen'/Mamluk decks may have made their way into Visconti hands when they married off a different/older Valentina Visconti (perhaps the Duchess of Orleans was named after this one) who married the son of the King of Cyprus who lead the raid on Alexandria in 1365 that may have resulted in the initial introduction of Mamluk cards into Europe. https://en.wikipedia.org/wiki/Valentina ... _of_Cyprus

And to add to you collection of quotes of leisure and gaming in a sentiment fairly close to that of Marziano, c. 1400, written by a friend of Duchess Valentina (in fact the Othea was dedicated to her husband Louis, Duke of Orleans, brother of the mad king, Charles VI):
Text 83
You may well amuse yourself
With Ulysses games every now and then
For they are subtle and honorable
In times of truce and celebrations.

Gloss 83
Ulysses was a Greek baron of great cleverness. And during the time of the long siege before Troy which lasted ten years, when there was a truce, Ulysses invented subtle and beautiful games to entertain the knights while they were resting. And some say that he invented the game of chess and other similar games. For this reason Othea tells the good knight that in appropriate moments he can well entertain himself with such games. And Solinus says, 'Anything subtle and honest is an honorable thing to do.'

Allegory 83

Ulysses games can be understood to mean that when the good spirit is weary of praying and devoting himself to contemplation, he can entertain himself by reading the Holy Scriptures. [more moralizing a'la Ovide Moralise follows - I'll spare you; the Allegory portion of the Othea is almost always disappointing to read but Pizan obviously felt the need for an interpretatio Christiana in order to absolve herself as a suspect woman writer, but she was certainly a believer]

(Othea's letter to Hector / Christine de Pizan; edited and translated by Renate Blumenfeld-Kosinski and
Earl Jeffrey Richards, 2017: 115)

Pizan's Othea illuminated in Harley 4431 f. 133 Ulysses playing chess


Re: Dummett's Game of Tarot, 1980, a few chapters

Palamedes, as the oldest authors say, invented the game [of tables=chess] ...
quote from Craig Kallendorf

That's confusing ... "tables" meant usually games played on a backgammon board in medieval times (games with dice). I guess, that Palamedes was never accused to have invented chess (he invented dice, and is also related to a game called "pessoi"). The invention of chess was usually given to Persian persons.
German wiki: "Unter Schachspielern galt Palamedes bis ins 19. Jahrhundert als Erfinder dieses Spiels; daher hießen Schachklubs früher gelegentlich Palamedes, und die erste europäische Schachzeitschrift nannte sich Le Palamède." That says, that Palamedes was considered by chess players (whom ? when?)
as inventor of the game till 19th century; therefore chess clubs were called Palamedes, and a first European chess magazine was called Le Palamède. But this didn't change history. Chess wasn't invented in Greece or Phoenicia, but much later in Asia, plausibly in India/Persia.

Palamedes was in mythology an inventor of many things comparable to the Egyptian Thot, but he was human, not a god. His father was Nauplios, a seafarer. There are suspicions, that Nauplios and Palamedes stood for Phoenicians, who brought progressive technologies to Greece.

Phaeded wrote:
To me its a chicken and egg question, in the sense that luxury decks had to exist before Filippo would likely come to love playing them...and the earliest luxury deck I'm aware of is a German deck with a hunting theme from 1430.
The deck, that John of Rheinfelden described in 1377, was obviously a deck of high quality with persons on each card, which on the number cards presented professions (as the later Hofämterspiel, which was made for a young Bohemian king). The context of the time allows the suspicion or conclusion, that this was made for the Bohemian/Luxembourg court of emperor Charles IV.
It had 60 cards, as reported by John.
With 4 Kings, 16 gods, and plausibly 40 number cards the Michelino deck also had 60 cards. The Bohemian court in Prague and Milan had close contact in 1395, when the German king Wenzel sold for very much money the duke title to Giangaleazzo Visconti. The Milanese delegation in Milan used likely the opportunity to get some decks of the Bohemian production, which reached the Milanese court and caused there a young Filippo Maria Visconti having cards in his youth.
There is another intensive report about a luxury declk produced in Mantua in 1386 (?), if I remember correctly. It involved the participation of a few artists to make the production. No, it was 1388 ... the document was reported by IPCS in 2006.

Re: Dummett's Game of Tarot, 1980, a few chapters

Huck wrote: 06 Jun 2020, 06:07 Phaeded wrote:
...luxury decks had to exist before Filippo would likely come to love playing them...and the earliest luxury deck I'm aware of is a German deck with a hunting theme from 1430.
The deck, that John of Rheinfelden described in 1377, was obviously a deck of high quality with persons on each card, which on the number cards presented professions (as the later Hofämterspiel, which was made for a young Bohemian king). The context of the time allows the suspicion or conclusion, that this was made for the Bohemian/Luxembourg court of emperor Charles IV.
It had 60 cards, as reported by John.
With 4 Kings, 16 gods, and plausibly 40 number cards the Michelino deck also had 60 cards. The Bohemian court in Prague and Milan had close contact in 1395, when the German king Wenzel sold for very much money the duke title to Giangaleazzo Visconti. The Milanese delegation in Milan used likely the opportunity to get some decks of the Bohemian production, which reached the Milanese court and caused there a young Filippo Maria Visconti having cards in his youth.
There is another intensive report about a luxury deck produced in Mantua in 1386 (?), if I remember correctly. It involved the participation of a few artists to make the production. No, it was 1388 ... the document was reported by IPCS in 2006.

Thank you for dredging up that pertinent old post. I recognize that the 1388 letter from the Gonzaga court has to refer to a luxury deck, but the only attribute of the deck they are careful to note, more than once, is that the pigment blue be used, specifically in the first instance of a (cheaper?) replacement ultramarine.
The second instance specifies ultramarine as we find on the Mamluk deck in Topkapi museum in Istanbul; so perhaps instead of one variation among many lost examples, perhaps blue was the distinguishing characteristic of Mamluk decks? This study of Muslim manuscript pigments notes: "The dominant blue identified on folios from all periods was ultramarine (lazurite or lapis lazuli)" (https://heritagesciencejournal.springer ... 018-0217-y Knipe, P., Eremin, K., Walton, M. et al. "Materials and techniques of Islamic manuscripts," Herit Sci 6, 55 (2018): no on-line pagination, but quote is above figure 12)


Furthermore, while the Visconti certainly intermarried with the German princes as they did with French, there is no historical record of the commissioning or importation of German cards before the record of Visconti having decks in their possession before c. 1412. Nor can the mere mentioning by Rheinfelden of persons on cards mean a luxury deck - its simply the West's translation and supplement of the Mamluk deck's King, Lieutenant and the Second Lieutenant. And assuming anything from the Hofämterspiel deck of 1453 (or later) for whatever Rheinfelden knew is simply too much of a temporal gap. In the The World in Play - Luxury Cards, 1430-1540, Timothy B. Husband only notes the record of manufacture "remain unspecified until a mention in 1422 of a card maker in Nuremberg, who apparently made woodblock cards and was patronized by the Tucher family" (2016: 14). The earliest German luxury deck is the The Stuttgart Playing Cards from 1430 - but where is there a record of hand-painted luxury German deck before then? Husband also notes that "in German-speaking lands, the development of playing cards was inextricably linked with that of copperplate engravings and, in particular, of woodblock prints" (ibid). Not luxury decks in this earliest phase of development.

For some reason you left out the other major lead in your link, Valentina Visconti, which I mentioned in my post above. You noted in that link the Gonzaga commission may be seen in the light of Valentina's wedding to the Duke of Orleans:

As a note of the authors this document is noted after 3 bans (Florence 1377, Siena 1377, Sicily 1377-1391) and chronicles (Viterbo) as the 5th oldest Italian document. Also it's the first document, which shows playing card interests of Italian nobility. The action is rather near in time to the wedding of Valentina Visconti, from which it is assumed, that Valentia brought a Lombard playing card deck to France. So perhaps this was a time of special interest or playing cards.
[and a little later]
The date of the playing card document 1388 is rather near to 1389, the year of Valentina Visconti's marriage to Louis d'Orleans and Valentina is said to have had brought a worthwhile playing card deck to France. Indeed Francesco Gonzaga accompanied the bride journey and the festivities in France.

None of these earliest references in Italy rely on anything German and the first two bans are exactly contemporary with Rheinfelden. In fact Italy - and the Visconti in particular - could have received Mamluk decks directly "from the source" as early as 1365. More on that date below, but first I would hasten to add my own note again here that we do know of 'Lombard' and 'Saracen' decks among the possessions of the Duchess of Orleans Valentina Visconti in 1408 (inventoried by her son right after she died; and note this is Filippo's half-sister). If the Gonzaga deck was a gift to Valentina was it considered 'Mamluk' or 'Lombard'? The focus on blue pigment might allow us to hazard a guess that it was an early European recreation/adaptation of a Mamluk deck, but nothing about either type of deck allows us to posit anything conclusive here. Dummett has this sparse info about the inventory commissioned by her son:
….mentions ung jeu de quarte sarrasines as well as unes quartes da Lombardie. Louis and Valentine were married in 1389, and she could have brought the cards with her from Milan then. (114) [footnote source: F. M. Graves, Deux inventaires de la Maison d’Orleans, Paris 1926; see pp 49 and 134]. (Michael Dummett and Kamal Abu-Deeb, “Some Remarks on Mamluk Playing Cards”, Journal of the Warburg and Courtauld Institutes, Vol. 36 (1973), pp. 106-128, 114).
A "Saracen" deck does seem to indicate early possession of a Mamluk deck and there is indeed a fairly direct path that explain that acquisition: among other royal marriages, Bernabò Visconti was keen to marry his children to the King of Cyprus. One of his five sons, Carlo (Sep 1359-Aug 1403), was made lord of Parma in 1364 and betrothed (1376) to Marguerite of Cyprus, daughter of Pierre I Lusignan, King of Cyprus & his second wife Infanta doña Leonor de Aragón, per a charter dated 4 April 1376 that notifies the marriages by proxy - "Karolum natum…Bernabos" and "dominam Margaritam domini regis prelibati sororem" (Documenti Diplomatici Milanesi, Vol. I, CXXIII, p. 180). This marriage never happened. However, in 1363, when Peter I of Cyprus was visiting Milan, Valentina's was promised to Peter;s son, the future Peter II of Cyprus; the same charter notes "domini Petri Jerusalem et Cipri regis" and "Valenziam natam…Bernabos" (ibid). In September 1377, the marriage was performed by proxy, but the bride left Milan the following year and they were wed in Nicosia, Cyprus. There is some nonsense on Wiki about her returning to Milan and marrying "her second husband a Count Galeazzo" but in fact she died on Cyprus in 1393.

Whence Mamluk cards in all of this Visconti marriage with Pierre II Lusignan of Cyprus? His father, Pierre I, the son of Hugh IV to whom Boccaccio dedicated his famous Genealogy of the Pagan Gods, was in Milan as part of his tour of Europe in raising knights for his crusade, holding jousts everywhere he went as the most renowned of jousters of his age. This is when the betrothals were arranged. After Italy he was in France for the coronation of Charles V, and then from there recruited in eastern Europe. Jean, Duke of Berry, was to lead the crusade but died, so Pierre - his Lusignan family originally hailing from the county Poitiers (and connected to other French royals) before inheriting Cyprus from convoluted dealings resulting from earlier crusades - took the lead. So with an international force of knights, Pierre I lead the most daring crusade in the 14th century in temporarily taking Alexandria in 1365, in which the city was looted and slaves taken. He was assassinated a few years later; and few years after that, c. 1373, the greatest French poet of the age, Guillaume Machaut penned his most famous work celebrating this crusade:

Along with Froissart’s Chroniques and long Arthurian romance Meliador, among works by others, the Prise d’Alixandre [The Taking of Alexandria] launched a literary movement during the closing years of the century that can be appropriately be terms the ‘chivalric rival.’ This amazingly detailed and accurate verse biography of Pierre I, king of Cyprus, celebrates knightly honor and valor. Its centerpiece is a detailed account, based on eyewitness testimony, of Pierre’s improbably capture of Alexandria, the richest and most populous city known to the West. The twin virtues of knightly courage and religious fervor were quickly losing prominence in a time of profound social change. Yet late fourteenth century writers like Froissart and Machaut, and the noble patrons that supported them, were eager to celebrate the aristocratic code steeped in the Christian zeal of the High Middle Ages. (Guillaume de Mauchaut: La Prise D'Alixandre [The Taking of Alexandria], ed. and tr., R. Barton Palmer,2002: 2)

Back to the chicken and egg question: Did the West acquire Saracen/Mamluk cards first and then create a market or did the Mamluks allow heathens to play cards among them and that in turn created the market? We don't know. But a 1365 raid of the richest and most populace Mamluk city, Alexandria, immediately preceding the earliest evidence in the West points to that event as the accelerant. An international body of knights disbursing from the 1365 expedition to all points back in Europe would explain the early diffusion. And what could be more portable than cards,ransacked from the most prosperous Mamluk city? Painted in costly lazurite/lapis lazuli and perhaps gilt, these were obviously luxury items. Notably Venice refused to ship the knights back since the event had negative impacts on their trade with the East, so the return through other ports such as Genoa and Pisa and quick diffusion to places like Viterbo and Florence would explain early card-playing laws in those cities (on Venice: "...in 1366 [the year after the sack of Alexandria] Venetians forbade their ships to carry men and war materials to Cyprus" (Peter W. Edbury, The Kingdom of Cyprus and the Crusades, 1191-1374; 1991: 178). .

As the leader of the crusade, naturally Pierre I returned to Cyprus with the choice spoils from Alexandria. When Valentina Visconti was escorted by her kin to Cyprus eight years after Pierre I's death, its not hard to imagine the Visconti entourage being given easily moved gifts, "Saracen" decks among them, thus returning with perhaps an original Mamluk deck to Milan in 1377. Gonzaga, themselves intermarried to Visconti, gifting a luxury deck to the latter Valentina, daughter by the first wife of Giangaleazzo Visconti (who had united the Lombard duchy after getting rid of Bernarbo, father of the Cypriot Valentina). We still don't know what the difference is between a 'Lombard' deck and a 'Saracen' deck but presumably the former added persons to the "court cards" - something Muslim law forbid in all art.

I'll go into the ramifications of Machaut's poem for Marziano elsewhere (it at least provides cultural context; the pagan gods attending on Pierre's conception is certainly interesting, making him no less a 'semideus' than Filippo), but the bottom line is that even though the fortunes of Cyprus declined precipitously after the death of Pierre I, the Visconti could proudly claim a direction connection to this paragon of knighthood via the older Valentina, married as she was to his son - certainly key in their relations with France, who fetishisized Pierre as a "10th Worthy". Playing cards, specifically linked to the spoils of Muslim lands as "Saracen" cards, were not just a luxury novelty but had specific crusade connotations, something that came to the fore with the crusade literature boom in the late 14th century. This more than anything explains the adoption and rapid spread of playing cards.

Like the Gonzaga gift to them, the Visconti in turn may have given luxury decks for the wedding of the younger Valentina to the Duke of Orleans in 1389. Furthermore there was another opportunity for gift-giving in 1391:
In March of that same year [1391] a group of French nobles, among them Louis of Orleans (husband of Valentina Visconti) and the duke of Burgundy, were guests of Giangaleazzo in his castle at Pavia. Accompanying them was Eustache Deschamps, the poet disciple of Machaut (Giangaleazzo would later try to bring the poetess Christine de Pizan to his court as well). ( F. Alberto Gallo, Music in the Castle: Troubadours, Books, and Orators in Italian Courts of the Thirteenth, Fourteenth and Fifteenth Centuries, tr. Anna Herklotz and Kathryn King, 1995: 55).
The reason the Duke or Orleans was in Milan is that he wanted to expand his dowry territory of Asti, located between Genoa and the Duchy of Savoy, into a new duchy of "Adria" (a dominion completely made up for Louis). Giangaleazzo wanted this to keep Savoy out of Lombardy and Louis was in an expansive mood as was the Duke of Burgundy for his own lands. To the point: Valentina's death in far off Blois in 1408 where her card decks were inventoried was in reality not some remote event given the House of Orleans and Visconti were intimately connected politically.

I do believe that if the German luxury decks existed before 1412 and featured the animal suits (e.g., ducks and hawks), that would partially have given Marziano the idea for the bird suits, but the main impetus in connection with gods came from the chivalric vowing made on symbolic birds (more on that later - this is already too long). If Marziano made his deck as early as 1412 the French influence in Milan would have still been keenly felt, hence why I see the influence of the likes of Machaut and Pizan on Marziano.