Sunday, July 5, 2015

chapter 1 (pp. 12-47) & part of 4 (146)

The late Prof. Michael Dummett, when he wanted to take time out from his work in philosophy or civil rights, wrote some of the best books on the history of tarot there are, some by himself and some with co-authors. The first of his two main books without co-authors, Game of Tarot (GOT) (1980), had as its main priority the games played with the tarot pack, but also managed to include most other aspects of the history as well. The second, written in Italian only, focused on tarot as a series of packs of cards. As Dummett says in his preface (his words followed by my translation):
È una storia delle carte, non dei giochi, e, in primo luogo, delle carte usate per giocare; solo alcune sezioni sono dedicate alle carte destinate a scopi carto- mantici e occultistici e solo due capitoli sono dedicati ai giochi, nei quali mi limito ad illustrare lo stretto indispensabile per la comprensione della storia delle carte.

(It is a history of the cards, not the games, and in the first place, the cards used for playing; only some sections are devoted to cards intended for cartomantic and occultist purposes and only two chapters are devoted to games, in which I limit myself to illustrating that which is indispensable for understanding the history of the cards.)
To my mind this book represents a further development of his thinking beyond Game of Tarot. Yet it is not readily available, even in Italian, in fact out of print. Fortunately I managed to secure a copy. In the summer of 2014 I started discussing it on the Internet a Tarot History Forum, with ample translations of passages I thought said interesting things going beyond what he said in GOT. Now it is a year later; on this blog I am rewriting my posts there, to incorporate comments by others and also so as to make them more accessible to people not intimately versed in tarot history; here Dummett's book itself is a model, because it assumes no prior knowledge. I am going to go through the book chapter by chapter, with translations of some passages and summaries of others; however I will take up topics, and chapters, in a different order. At the same time I will be engaging with the text myself, adding information that has become available since its publication in 1993 and my own reflections. Since I am not an expert translator, I will include Dummett's relatively easy Italian for the passages I quote.

In the section headings for this blog I include page numbers. These are for what is translated in the section, not for the chapters that are the section's main focus, which typically are longer.


On the first page of Chapter One (p. 12 of the book), Dummett explains why he is writing the book. It is also why it needs to be read by anyone interested, for whatever purpose, in tarot decks before the 20th century:
Nelle note alla Terra desolata T.S. Eliot scrisse: «non conosco la costituzione esatta del mazzo dei tarocchi». Non esiste, in realtà, nulla del genere; esistono più forme distinte del mazzo dì tarocchi, ciascuna diversa dall’altra per composizione.

In the notes to The Waste Land ,T. S. Eliot wrote: "I do not know the exact constitution of the tarot deck." There is, in fact, nothing of the kind; there exist several distinct forms of the tarot deck, each different in composition
In other words, there is no such thing as "the" tarot pack. There are only particular tarot packs used at particular times and places, with a variety of imagery and sequential orderings of the 22 special cards and even some variations in the suit cards. Anyone who wants to understand the history of the tarot, including the symbolism of historic decks, has to understand it in relation to these historical contexts

Chapter One asks, "Che cosa sono i tarocchi?"--"What is the tarot?" After describing the composition of a typical deck used today in Italy, the Piedmontese, he states the main thesis of the chapter:
Il mazzo dei tarocchi fu inventato nell’Italia del Nord nella prima metà del XV secolo. Questa origine italiana sembra probabile dal numero relativamente alto di carte da tarocchi ita-[end of 13]liane del Quattrocento che ci sono pervenute, e dai numerosi riferimenti ad esse in documenti quattrocenteschi italiani. Sebbene esistano numerosi riferimenti ad esse nella Francia del XVI secolo, per il XV non ce n’è alcuno al di fuori dell’Italia. L’argomento davvero conclusivo è che il sistema di semi usato nel mazzo dei tarocchi è tipicamente italiano.

The tarot deck was invented in northern Italy in the first half of the fifteenth century. This Italian origin seems likely by the relatively high number of Italian tarot cards [end of 13] of the fifteenth century that have come down to us, and the many references to these documents in fifteenth-century Italian. Although there are numerous references to them in France of the sixteenth century, in the fifteenth there is none outside Italy. The really conclusive argument is that the system of suits used in the tarot deck is typically Italian.
The first actual report of a tarot deck so far known is in 1440 Florence, a hand-painted deck for the Lord of Remini, who was one of two main military leaders for Florence at that time. The documentation, published in 2002 in an article on a Tuscan diarist, came to the attention of readers of internet tarot forums only in 2012 (see When Dummett was writing in 1993, the first known reference was Ferrara 1442, the sale of a deck to the Ferrarese court from a Bolognese merchant. Two years makes little difference. Dummett assumes as a realistic possibility that tarot was invented several years before, even a few decades, but not as far back as the first reports in Europe of regular cards, which goes back to the 1370s. He says (p. 23, including the footnotes for completeness):
Le carte da gioco comparvero per la prima volta in Europa nel 1370 circa 6. Non ci è pervenuta alcuna carta del XIV secolo (con al massimo un’eccezione 7); dei molti riferimenti ad esse in documenti dell’epoca, uno solo ci dà qualche informazione sul loro aspetto. Si tratta del celebre Tractatus de moribus et disciplina humanae conversationis, scritto a Basilea, probabilmente nel 1377, da un frate di nome Giovanni (di solito indicato come Giovanni da Rheinfelden).
6. Il primo documento contenente la parola «naips», che significa ‘carte da gioco’, è il Diccionari de rims (Vocabolario di rime) del 1371 del poeta catalano Jaume March (pubblicato a cura di Antoni Griera, Barcellona, 1921; si veda p. 63). Cfr. Jean-Pierre Étienvre, Figures du jeu: études lexico-sémantiques sur le jeu de cartes en Espagne (XVIe-XVIIie siècle), Madrid, 1987, pp. 19, 68. Étienvre cita un altro riferimento alle carte da gioco nel Llibre de les dones di Francesc Eiximenis, «probabilmente dello stesso anno».
7. Si tratta di due fogli antichissimi stampati da matrici di legno e non tagliati nell’Instituto Municipal de Historia a Barcellona: si veda Simon Wintle, ‘A «Moorish» Sheet of Playing Cards’, The Playing Card, Vol. XV, 1987, pp. 112-22. È probabile che questi fogli risalgano al primo decennio dei XV secolo.

(Playing cards appeared for the first time in Europe in about 1370 (6). No card of the fourteenth century still exists (with a maximum of one exception (7)); of the many references to them in documents of the time, only one gives us some information on their appearance. This is the famous Tractatus de moribus et disciplina humanae conversationis, written in Basel, probably in 1377, by a monk named Johannes (usually indicated as John of Rheinfelden).
6. The first document containing the word ‘naips', which means 'playing card', is the Diccionari de rims (Dictionary of rhymes) by the 1371 Catalan poet Jaume March (published under the editorship of Antoni Griera, Barcelona, 1921; see p. 63). Cf. Jean-Pierre Étienvre, Figures du jeu: études lexico-sémantiques sur le jeu de cartes en Espagne (XVIe-XVIIIe siècle), Madrid, 1987, pp. 19, 68. Étienvre cites another reference to playing cards in Llibre Francesc de les dones by Francesc Eiximenis, “probably of the same year”.
7. There are two sheets printed from ancient wood matrices, uncut, in the Instituto Municipal de Historia in Barcelona: see Simon Wintle, 'A Moorish Sheet of Playing Cards', The Playing Card, Vol XV, 1987, pp. 112-22. It is likely that these sheets go back to the first decade of the fifteenth century.)
When the 22 special cards were added so as to make a tarot deck, the packs so formed reflected the same characteristics as the normal pack used in the region. For example, in Bologna a 40 card pack was used in many games. The tarocchi pack, called tarocchini, i.e. little tarocchi, used the same 40 cards with the addition of the characteristic special cards. So the tarot pack is simply the regular pack with the cards called "triumphs", later "tarocchi", added.

I should perhaps say something at the outset about the role of the special cards added to the regular deck. Although there is no documentation of the specific rules of the game with the special cards until 1659 (, the role of the special cards is clear, both from that document and others: they served as a permanent "trump" suit; that is, any card in that suit beats any card of the four other suits in a "trick", meaning a round in which each player puts in a card, following suit where possible, and the high card takes the others. This practice may be inferred from the account of another game that also had special cards of this nature, designed for Duke Filippo Maria Visconti in 1420s Milan, where there were 16 such cards, each representing a particular Greco-Roman god or demi-god, in a numbered hierarchy of precedence ( The main difference is that for purposes of "following suit" in a trick, 4 of each of the special cards were assigned to each of the four suit. This characterstic is not mentioned in 1659 or any other time in relation to the game of tarot, so if it persisted it cannot have been for long, because by the latter part of the 15th century there were 22 such cards, a number not divisible by 4.


The regular deck (i.e. without the special cards of the tarot) exists, and has done so almost from the beginning, in "standard models" (p. 14):
In ogni paese del mondo le carte da gioco prodotte per uso ordinario si uniformano a particolari modelli grafici stereotipi, con pochissime variazioni fra pn fabbricante e l’altro.

(In every country in the world playing cards produced for ordinary use will conform to particular stereotypical graphical models, with very little variation between one manufacturer to another.)
This is not only true country to country, but in some cases region by region within countries. It is because players need to be able to identify the cards at a glance, regardless of what is on them. They want the cards to be familiar. He continues (p. 14)
Pertanto, i modelli standard sono di solito estremamente conservatori. A volte cambiano, ovviamente; in qualche caso all’im- prowiso, più spesso così lentamente che i giocatori non se ne rendono neppure conto. La storia della grafica delle carte da gioco è quindi, prima di tutto, la storia dell’evoluzione e diffusione dei modelli standard.

Therefore, the standard models are usually extremely conservative. Sometimes they change, of course; in some cases unexpectedly, more often so slowly that the players are not even aware of it. The history of the design of playing cards is then, first of all, the history of the evolution and spread of the standard models.
There are French suits, German suits, and Latin suits, each with various standard models. Rather than go through  Dummett's detailed descriptions, I will post some examples, taken from the site
French suits, which are now international, started in France and have what in English are called Spades (French Piques, pikes), Clubs (Trèfles, i.e. flowers), Hearts (Coeurs), and Diamonds (Carreaux,  tiles).

German suits in the beginning had much variation, but settled on Leaves, Acorns, Hearts, and Bells. With their origin in Germany, they are used primarily in Central Europe.

Latin suits have Swords (in Italian, Spade or Espade) Staves (Bastoni), Cups (Coppe), and Coins (Denari). They in turn divide into Italian, Spanish, and Portuguese, each with a characteristic way of depicting the suit-objects. In this case, he says, these terms are of convenience and do not imply the country of origin.  It will turn out that while Italian suits did originate in Italy, but the others are more complicated.

Italian suits (directly left) are used in many parts of Italy, especially in the North, where they probably originated. Within Northern Italy, there are variations in suits among regions.

Spanish suits (lower left) are used in Spain and Spanish America, as well as central and southern Italy and the eastern part of the region bordering Spain in France. The Staves have knobs and neither Swords nor Staves intersect.
The Portuguese system (not shown) used to be the national system of Portugal and Portuguese colonies and trading partners, but is now almost extinct, used only in Sicily. However most Asian cards are descended from it. In it, Dummett says, the Swords are straight, as in the Spanish system, but intersect, as in the Italian. The Staves also intersect; they are smoother than the Spanish but not as smooth as the Italian.


Dummett argues that playing cards came to Europe from the Islamic world. First, the earliest documented mention in Europe is in Catalonia of 1371, followed in 1377 by Florence, Basel, Siena, and Paris. Several of these mentions stress that it is a new game. Speaking of playing cards (p. 26):
Frate Giovanni da Rheinfelden dichiara apertamente che esse furono introdotte nella regione di Basilea l’anno stesso in cui egli scriveva, il 1377; le Cronache di Viterbo parlano della loro introduzione nel paese nel 1379; un editto di Valenza del 1384 fa riferimento alle carte come a «un nuovo gioco»; e l’editto più antico, quello fiorentino del 1377, le descrive come «recentemente introdutte in queste parti»10.
10 Si vedano Ludovico Zdekauer, ‘Il giuoco in Italia nei secoli XIH-XIV e specialmente in Firenze’, Archivio Storico Italiano, ser. IV, Vol. 18, 1886, pp. 20- 74; Francesco Novati, ‘Per la storia delle carta da giuoco in Italia’, Il libro e la stampa, anno n, 1908, pp. 54-69; Wilhelm Ludwig Schreiber, Die àltesten Spielkarten, Strasburgo, 1937, p. 74; H. Rosenfeld, ‘Zur Datierbarkeit Friiher Spielkarten in Europa und im nahen Orient’, Gutenberg-Jahrbuch, 1975, pp. 353-71; F. Pratesi, ‘Italian Cards: New Discoveries’, n. 8, The Playing Card, Vol. XVII, 1989, pp. 107-11.

(Friar John of Rheinfelden says openly that they were introduced into the Basel region the same year in which he wrote, 1377; the Chronicles of Viterbo talk about their introduction into the country in 1379; an edict of Valencia in 1384 refers to the cards as "a new game"; and the oldest edict, that of Florence in 1377, describes them as "introduced recently in these parts" 10.
10 See Ludwig Zdekauer, ‘Il giuoco in Italia nei secoli XIH-XIV e specialmente in Firenze’, Archivio Storico Italiano, ser. IV, Vol. 18, 1886, ['Games in Italy of the XIH and XIV centuries -especially in Florence ', Italian Historical Archives, ser. IV, Vol 18, 1886], p. 20 -74; Francesco Novati, ‘Per la storia delle carta da giuoco in Italia’, Il libro e la stampa, anno n, 1908 ['For the history of card playing in Italy', The book and the print, year no, 1908], p. 54-69; Wilhelm Ludwig Schreiber, Die àltesten Spielkarten [The Oldest Playing Cards] Strasbourg, 1937, p. 74; H. Rosenfeld, 'Zur Datierbarkeit Friiher Spielkarten in Europa und in nahen Orient', Gutenberg-Jahrbuch, 1975, p. 353-71; F. Pratesi, 'Italian Cards: New Discoveries', n. 8, The Playing Card, Vol. XVII, 1989, pp. 107-11.)
If so, from where? He continues (still p. 26):
Affermare che qualcosa fu introdotto in Europa da qualche altro luogo nel tardo Trecento può solo significare che fu introdotto dal mondo Islamico: o dalla Spagna moresca in quella cristiana, o dall’Egitto mamelucco attraverso il grande porto di Venezia. Non c’è dubbio che le carte da gioco erano conosciute fra i Mamelucchi: i più antichi accenni in documenti, che risalgono all’anno 1400 circa, li troviamo negli Annali di Ibn Taghri-Birdi, in cui sono chiamate kanjifah.

(To say that something was introduced to Europe from some other place in the late fourteenth century can only mean that it was introduced by the Islamic world: from Moorish or Christian Spain, or Mamluk Egypt through the large port of Venice. There is no doubt that playing cards were known among the Mamluks: the oldest references, in documents that date back to 1400, we find them in the Annals Ibn Taghri-Birdi, in which they are are called kanjifah.)
The Mamluks had been imported from Central Asia as slaves of the Egyptians, used for military purposesm until they revolted successfully against their masters ( A set of 16th century Mamluk cards is in the Topkapi Museum in Istanbul.  They have 52 cards in 4 suits: swords, coins, cups, and polo sticks. Polo was unknown at that time in Europe. As for the courts (p. 29):
Nel mazzo di Istanbul ci sono tre figure per ogni seme: la più alta è il Re (malik) e, sotto di lui, ci sono il Viceré {na’ib ma- lìk) e il Secondo Viceré (thani na’ib).

(In the deck of Istanbul there are three figures for each suit: the highest is the King (malik) and, under him  there are the Viceroy (na'ib but-lik) and the Second Viceroy ( thani na'ib)).
Not coincidentally, the first term for playing cards was "naipes", by which they are still called in Spain (p. 29).
Inoltre, la più comune parola italiana tre-quattrocentesca per indicare le carte da gioco è «naibi», affine alla parola spagnola «naìpes», usata ancora oggi, e a forme analoghe che si trovano in antichi documenti francesi e catalani.

(Moreover, the most common Italian word to indicate playing cards is "naibi," cognate with the Spanish word ' NAIPES', still used today, and similar forms found in ancient French and Catalan documents.)
Similarly, in the document of 1440 published in 2002, the reference to the tarot deck does so with the phrase "un paio di naibi a trionfi", i.e. "a pack of triumph cards" (

Also there is no trace of the evolution of playing cards from something else in Europe. The Mamluks, however, called their cards kanjifàh, from the Persian ganjifeh, which itself is probably not an indigenous word, he says. So from there playing cards can be traced even further back.

In Chapter 4 Dummett tells us more, that playing cards originated in China and worked their way west. Here I want to give an exact quote, because I also want to indicate some problems with his account (p. 146):
Le carte da gioco non sono un’invenzione europea. Come suggeriscono i titoli di molti libri scritti sull’argomento, il problema della loro origine ha affascinato gli studiosi per duecento anni; ma, in realtà, esso è posto in termini ambigui. Le parole «carte da gioco» possono essere usate in senso generico, comprendendo quindi le carte per Cuccù e Mercante in Fiera; è praticamente certo che le carte da gioco intese in questo senso furono inventate in Cina, dove, dopo tutto, fu inventata la carta e scoperta la stampa. Ma le prime carte da gioco cinesi rappresentavano i Domino, anch’essi un’invenzione cinese. Possiamo dunque porci un problema diverso, quello cioè dell’origine del normale mazzo a quattro semi usato, in forme diverse, nei vari paesi d’Europa. Abbiamo visto che, quasi certamente, esso giunse in Europa dal mondo Islamico, nella forma del mazzo di cinquantadue carte. Fu inventato lì, oppure, come gli scacchi, venne da zone ancora più a Oriente? Come sorse l’idea di fare un mazzo di carte da gioco diviso in semi, contraddistinti da segni di seme? Perché le carte di ciascun seme sono divise in carte numerali e figure? Non siamo ancora in grado di rispondere con certezza a queste domande, ma mazzi di carte da gioco con le stesse caratteristiche erano noti in India fin dai tempi del primo imperatore mogol Babur, e vengono ancora usati in quel paese. Questi mazzi indigeni hanno dieci carte numerali e due figure per seme; nella forma Mogol originaria, ci sono otto semi, mentre le versioni Hindu hanno dieci o dodici semi. Analoghi mazzi a otto semi furono in passato usati anche in Persia, più o meno dalla stessa epoca (inizio del XVI secolo). Questi mazzi sicuramente non derivano da quelli europei, ma altrettanto sicuramente hanno in comune con essi l’origine, forse da un tipo scomparso con quattro semi e quarantotto carte. Secondo questa ipotesi, i Mamelucchi aggiunsero una terza figura ad ogni seme, mentre i Persiani raddoppiarono il mazzo: il prototipo a quarantotto carte sarebbe venuto da qualche luogo più a Oriente, forse dall’Asia Cen-[96]trale. Tutto questo è pura speculazione: ma è certo che il mazzo a quattro semi non fu ideato in Europa ma importato dal mondo Islamico.

(Playing cards are not a European invention. As suggested by the titles of many books written on the subject, the question of their origin has fascinated scholars for two hundred years; but, in reality, it is put in ambiguous terms. The words ' playing cards' can be used in a generic sense, thus including cards for Cuccù and Mercante in Fiera; it is is virtually certain that playing cards in this sense were invented in China, where, after all, paper was invented and printing discovered. But the first Chinese playing cards depicted dominoes, also a Chinese invention. So we can set ourselves a different problem, namely, of the origin of the regular deck in four suits used in different forms in the various countries of Europe. We have seen that, almost certainly, it arrived in Europe from the Islamic world, in the form of a pack of fifty-two cards. Was it invented there, or, like chess, came from areas further to the East? How did the idea arise of making a deck of playing cards divided into suits, marked by suit-signs? Why were the cards of each suit divided into numeral cards and figures? We are not yet able to answer these questions with certainty, but decks of playing cards with the same characteristics were known in India since the time of the first Mughal emperor Babur, and are still used in that country. These native decks have ten pip cards and two figures per suit; in the original Mogul iorm, there are eight suits, while the Hindu versions have ten or twelve suits. Similar decks of eight suits were formerly used in Persia, more or less from the same era (early sixteenth century). These decks certainly did not derive from European ones, but just as surely have a common origin with them, perhaps from a type that has disappeared with four suits and forty-eight cards. According to this hypothesis, the Mamluks added a third figure in each suit, while the Persians doubled the pack: the prototype to forty-eight cards would have come from somewhere further east, perhaps from Central Asia. [96] All this is pure speculation, but it is certain that the deck in four suits was not conceived in Europe but imported from the Islamic world.
There are really two theses here. First, that it is certain that the idea of four suits came from the Islamic world somewhere. . Second, more speculative, he is saying that perhaps the Malmuks, in the Mediterranean, added the third figure in each suit. In that case the suits would have come first to southern Europe, and then, in the suits there, to Northern Europe. I will address only the second point.


The problem is that there are other reasonable alternatives to the scenario Dummett sketches out. The main issue is whether playing cards came by only one route, from Islamic Persia or India to Egypt and the Islamic Mediterranean, then to Christian Southern Europe, and finally to Northern Europe, or perhaps by other routes as well, including an overland route from Central Asia directly to Central Europe. "Huck" on THF has brought our attention to the 19th century researcher of central European trade F. L. Hübsch, who stated, unfortunately without supporting references (he apparently didn't see what he was saying as unusual or controversial), that the first producer of playing cards in Bohemia was a certain card-painter Jonathan Kraysel from Nuremberg in 1354 ( He also mentions playing card production earlier in Poland, and in Bohemia by 1340, as well as some playing card prohibitions enacted by Charles IV (reigned 1346-1378).  "Huck" has more recently posted a passage from Hübsch in which he says that dice and playing cards were known in Bohemia since 1309 (  He has also found other publications, from 1599 and subsequent centuries, saying the same thing (starting at It seems to me possible that Hübsch was merely relying on one of them. Another historian, Schreiber, writes of three card-players struck by lightning in 1303. This, to be sure, is the kind of tale that often is invented for moralistic purposes. And even if playing cards were in Central Europe this early, they still could have gotten there from points south, such as Venice.

There is then the question of how much development had already occurred by the time playing cards reached Central Asia.

Here is "Andy's Playing Cards" (, on Chinese playing cards, after its account of domino cards. The "Wilkinson" mentioned was a late 19th century Sinologist, first initials "W. H.", who was a diplomat in China:
Another early pattern is known as Gun Pai ("stick cards" or "cane cards"), likely referring to their shape.Wilkinson maintained that these cards were created from early books, whose pages were made individually detachable for an easier reference; later on, their use as an amusement would have caused their size to be reduced. According to his theory, this kind of books came into use in China by the mid 8th century; if this proved correct, the Gun Pai might have been created before the domino cards.

The structure of this pattern was described as based upon three suits, whose cards featured signs from 1 to 9 (but one suit had numerals). The suits were identified by Wilkinson as Jian (or Qian) "coins, money", Tiao with a meaning of "long things, sticks", and Wan meaning "myriads, 10,000".

Three more subjects named Qian Wan ("Thousand Myriads", also known as "Old Thousand"), Hong Hua ("Red Flower") and Bai Hua ("White Flower") completed the set.

A full deck contained four duplicates of each subject (120 cards in total) plus, in some editions, a variable number of special loose cards, up to six per pack, whose function in play - according to Wilkinson's report - was the same as that of Western jokers, i.e. they acted as wildcards that could replace any standard subject.
Another type of deck had four suits and two of these special subjects, corresponding to "Red flower' and "White Flower", for 38 cards total. Still other decks had numeral cards and additional "honours" in each suit, plus special cards:
The main features of the group are: three suits, given the Western names of Coins, Strings and Myriads; values running from 1 to 9, plus a number of honour cards, of higher rank; some decks also have one or more special subjects.
There were as many as 5 of these special cards, not attached to any suit.

It seems to me reasonable to hypothesize that  some decks were developed that used two "honour cards" per suit, from which the Mogul decks descended, while other decks used three in each suit. The Malmuks had originally come from precisely Central Asia.

"Andy's Playing Cards" presents evidence that the Malmuks' suit signs come directly from Chinese characters for the suits (unfortunately I cannot post the Chinese characters on Andy's site (
But why the signs of the suit of Tûmân were chalices, or cups? Wilkinson suggested that the choice of cups as the suit's distinctive sign might have sprung from a misinterpretation of the Chinese and Manchu character which, turned upside down (), has indeed the shape of a chalice. A detail matching this theory is the position of these signs, always in the top part of the cards in Chinese patterns, while in Arabic courts they are featured below, as in a Chinese card turned upside down. Many scholars rejected Wilkinson's theory; however, in the case it was true, we should think that the earliest decks that reached the Arabs still had suits spelt with Chinese glyphs (not clearly understood by the Arabic players), thus the cards would have not come from Persia, but likely from a region further east.

Since all the suits of the Oriental system are related to money or coins, the symbolic meaning of Tûmân may have a similar relation, as well. For instance, more than the shape of a chalice, the meaning may be the metal which the cups featured in Mamlûk cards are likely made of, i.e. gold. It is probably not a coincidence that also the Persian Toman was a golden piece.

Also for the Chinese suit of Tens, Shi, only found in 4-suited patterns (see the Chinese gallery), it is impossible not to see this cross-shaped character as a stylized sword with its hilt. In fact the corresponding suit in the Arabic deck bears the name of Suyûf, whose meaning is "swords" or "scimitars".

A further coincidence seems to concern the three "special" cards of the money-suited packs, also called honours, and named Old Thousand, Red Flower and White Flower (see the Chinese gallery, page 1); they may have a relation with the three courts of the Arabic deck (king, deputy and second deputy).
So the three court cards may have come from China to Central Asia, and from there gone south to Egypt but also perhaps into North-Central Europe directly.

Even the idea of special cards not attached to any suit seems to be Chinese. It might be that they are just wildcards, but perhaps they were more than that. Further research is needed before we can say, with Dummett:
Il mazzo di carte normale non fu un’invenzione europea, ma il mazzo dei tarocchi indubbiamente sì. Non c’è la minima traccia di prova che qualcosa di anche vagamente simile al mazzo dei tarocchi sia stato conosciuto al di fuori dell’Europa prima del XIX secolo.

(The normal deck of cards was not a European invention, but the tarot deck undoubtedly yes. There is not the slightest trace of evidence that anything even vaguely similar to the tarot deck has been known outside of Europe before the nineteenth century.)
However it is certainly clear that even if the Chinese had a special suit more powerful than the others, there is no trace of its being used by Muslims or Europeans.


If playing cards entered Europe only via the Mediterranean, it would be expected that at first all playing cards would be Latin-suited. This is in fact a thesis Dummett defends. (p. 31):
(Sfortunatamente Frate Giovanni, nel suo Tractatus de moribus, mancò di indicare quali fossero i segni di seme; ma per quanto ne sappiamo oggi, è logico supporre che i segni ‘latini’ fossero usati in tutta Europa nei primi decenni di diffusione delle carte da gioco.)

Unfortunately, Brother John, in his Tractatus de moribus, failed to indicate what the suits signs were; but as far as we know today, it is logical to assume that the 'Latin' signs were used throughout Europe in the first decades of the spread of playing cards.
And later, more fully (p. 36):
Quale che sia la verità sulle diverse versioni del sistema di semi latino, è ampiamente dimostrato che carte con semi latini, nelle varie forme, erano l’unico tipo conosciuto in Italia, Spagna e Francia fin verso il 1470. Se accettiamo l’ovvia ipotesi che, quando le carte da gioco fecero la loro prima comparsa in Europa, esse avevano dappertutto lo stesso sistema di semi, ne consegue che questo poteva essere soltanto il sistema latino: si potrebbe giungere a questa conclusione anche senza saper nulla a proposito delle carte mamelucche.

Ci è pervenuta una notevole quantità di carte tedesche e svizzere del Quattrocento e primo Cinquecento, in maggioranza databili dopo il 1450. In entrambe le aree si producevano carte con semi latini e non solo per l’esportazione; erano tuttavia ben lungi dal predominare. In nessuna delle due aree, tuttavia, esiste un sistema alternativo che lasci in alcun modo supporre di essere stato il modello rispetto al quale gli altri siano deviazioni. Anzi, troviamo in Germania e, in misura minore, anche in Svizzera, tracce di frenetica sperimentazione con i segni di seme e altri tratti del mazzo di carte: innumerevoli oggetti diversi sono utilizzati come segni di seme nell’uno o nell’altro mazzo. Quello che doveva diventare il sistema di semi tedesco fu ideato intorno al 1460, e quello svizzero risale forse alla stessa epoca; ma fu solo verso la fine del secolo che il primo fu elevato a sistema standard — e quello svizzero qualche decennio più tardi. Niente di tutto ciò porta conferme definitive alla nostra ipotesi, ma tutto è coerente con essa. È probabile che le carte da gioco tedesche e svizzere del Trecento fossero a semi latini; nel Quattrocento iniziò una lunga ricerca di un sistema di semi più consono alle culture nazionali.

(Whatever the truth about the different versions of the Latin suit system, there is ample evidence that cards with Latin suits, in various forms, were the only type known in Italy, Spain and France after 1450. If we accept the obvious hypothesis that when playing cards made their first appearance in Europe, they were everywhere the same suit-system, it follows that this could only be the Latin system: one could come to this conclusion without knowing anything about Mamluk cards.

A significant quantity of German and Swiss cards of the fifteenth and early sixteenth century have come down to us, the majority dated after 1450. In both areas, cards with Latin suits, and not just for export, were produced; however, they were far from being predominate. In none of the two areas, however, is there an alternative system which lets us in any way suppose it to have been the model against which others are deviations. Indeed, we find in Germany and, to a lesser extent, also in Switzerland, traces of frantic experimentation with suit signs and other parts of the card packs: countless different objects are used as suit-signs in one or another pack. What was to become the German suit-system was invented around 1460, and the Swiss dates perhaps to the same time; but it was only towards the end of the century that it was elevated to a standard system – and that of the Swiss some decades later. None of this leads to final confirmation to our hypothesis, but everything is consistent with it. It is likely that German and Swiss playing cards of the fourteenth century were Latin-suited; in the fifteenth Century there began a long search for a suit-system best suited to the national cultures.)
So the argument seems to be lack of evidence to the contrary, including places (extant cards) where evidence of independent, non-Latin derivation would be expected to be found. If he is right, any spread of cards from a non-European source was so minimal as to be undetectable. However he seems not to have known about Hübsch.


A peculiarity of the tarot pack is that it always, so far as known, had Queens as part of its court cards. But at the time when the tarot was invented, this was generally not true of regular packs in Italy. In previous writings, he had made the presence of a Queen in an incomplete surviving deck a strong indicator that the deck is a tarot, because normal Italian decks didn't usually have Queens. His reasoning was that all the normal decks that have survived lack Queens; but since other countries did have such decks--German decks, before 1500, and French suits, from whenever French suits were invented--probably a few in Italy did as well, until after 1500. On this point he now cites two early documents that "suggest" [suggerita] by "hints" [accenni] Queens in normal decks (p. 21).
È tuttavia possibile che fossero occasionalmente usati nell’Italia del XV secolo mazzi normali contenenti le stesse quattro figure del mazzo dei tarocchi; anche se nessuna carta ci è pervenuta a sostegno di questa ipotesi, essa è suggerita da due accenni in documenti (5).

(However it is possible that there occasionally were used in Italy of the XV century normal decks containing the same four figures as tarocchi deck; even if no card has come down to us to sustain this hypothesis, it is suggested by two hints in documents (5)
Here is footnote 5:
5. Marzio Galeotti di Nami (morto nei 1478) usa l’espressione l’espressione «regum reginarum equitum peditumque potentiam», senza far cenno ai triumphi, in un passo relativo alle carte da gioco (De doctrine, promiscua, Firenze, 1548, cap. 36 in fondo). Analogamente, San Bernardino da Siena, nel suo sermone contro il gioco d’azzardo («contra alearum ludos») predicato nel 1423, nomina, a proposito delle carte da gioco, per primi «reges atque reginae» e poi « milites superiores et inferiores», ancora una volta senza alcun accenno a triumphi si veda S. Bernardini Senensis O.F.M. Opera Omnia, a cura dei PP. Collegii S. Bonaventurae, Vol. II, Firenze, 1950, Sermo 42, p. 23. È ovviamente possibile che l’uno o l’altro dei due autori avesse in mente le carte da tarocchi, ma non vedesse ragione di nominare i trionfi; questa è, tuttavia, un’ipotesi poco probabile.

(5). Marzio Galeotti di Narni (died in 1478) uses the expression "regum reginarum equitum peditumque potentiam," without mentioning the triumphi, in a passage relating to playing cards (De doctrina promiscuo, Florence, 1548, ch. 36 at the bottom). Similarly, San Bernardino of Siena, in his sermon against gambling (‘contra alearum Ludos’) preached in 1423, names in passing some playing cards, first "reges atque reginae" and then "milites superiores et inferiores', again with no mention of triumphi. See S. Bernardini Senensis, Opera Omnia, edited by PP. Collegii S. Bonaventurae, Vol II, Florence, 1950, Sermo 42, p. 23. It is of course possible that one or other of the two authors had in mind tarot cards, but did not see any reason to name the triumphs; this is, however, an unlikely hypothesis.)
In these quotes, "regum" means "king" and "regina" means "queen". These hints do not, however, prevent him later on from using Queens as a likely indicator that a deck is a tarot. I will get back to this point later in this post.

Another question, for me, is why the tarot, originating in a country where the normal deck usually didn't have Queens, managed so firmly and uniformly to have suits with Queens. I suspect influence from Germany, just as in France. But if so, why not normal decks as well? Here what Dummett says about the putatively 1377 Basel text Tractatus de moribus is relevant, with its discussion of female courts. I highlight the relevant parts in the quote below (although the whole passage is relevant to the issue of Queens) (p. 24f):
Il Tractatus de moribus ci è pervenuto solo in un manoscritto del 1429 e in altri tre, tutti del 1472 (8). Ci sono pochissime varianti fra questi quattro testi, ma, se l’originale è veramente del 1377, essi devono contenere interpolazioni, forse del copista del 1429. Da questi testi e da un certo numero di antichi mazzi tedeschi pervenutici, veniamo a conoscenza di un alto grado di sperimentazione nella composizione del mazzo normale nella Germania del Quattrocento, e fu in seguito a questi esperimenti che la Regina fece il suo primo ingresso nel mazzo di carte. Ai sostentori della liberazione della donna farà piacere sapere che essa fu originariamente introdotta non come inferiore al Re ma come di pari grado. Il Tractatus de moribus descrive mazzi in cui nei quattro semi, o in due su quattro, tutte le figure erano femminili. Un famoso mazzo quattrocentesco dipinto a mano, prodotto fra il 1427 e il 1431, è uno degli esempi a noi pervenuti di questo secondo tipo. In altri mazzi ancora troviamo Unter femminili. Attraverso quella che fu probabilmente una fase di sviluppo successiva, si giunse a mazzi in cui tutti e quattro i semi hanno Re, Regina, Ober e Unter. Esempio di questo è un mazzo dipinto a mano, datato 1440-5, noto come mazzo di caccia Ambraser 9; e ci sono molti altri mazzi tedeschi quattrocenteschi con quattro figure per seme. Il Tractatus non ne fa cenno; esso fa riferimento, tuttavia, con grande entusiasmo, a un tipo che non ci è pervenuto, con quindici carte in ciascuno dei quattro semi, compresi Re, Regina, i due Marescialli e una Servetta (ancilla) come carta più bassa fra le cinque figure. Quando, nel 1470 circa, i fabbricanti di carte francesi introdussero la loro grande innovazione, il sistema di semi francese, essi presero a prestito la Regina dai mazzi tedeschi con quattro figure, in sostituzione del Cavaliere del mazzo con semi latini; fu questa la sua comparsa insieme col Re in un mazzo con solo tre figure per seme.
8. Cfr. Sir Edward Augustus Bond, ‘The history of playing-cards’. Athenaeum, n. 2621, 19 gennaio 1878, da p. 87, col. 3, a p. 88, col. 2; Peter Kopp, ‘Die frühesten Spielkarten in der Schweiz’, Zeitschrift für Schweizerische Archäologie und Kunstgeschichte, Vol. 30, 1973, pp. 130-46; e Hellmut Rosenfeld, ‘Zu den frühesten Spielkarten in der Schweiz: eine Entgegnung’, ibid., Vol. 32, 1975, pp. 179-80. Più recentemente, Ronald Decker in ‘Brother Johannes and the Year 1377, The Playing Card, Vol. XVIII, 1989, pp. 46-7, ha proposto un’emendazione del testo, secondo la quale la data 1377 sarebbe quella in cui le carte da gioco arrivarono per la prima volta nella sua regione; il trattato quindi potrebbe essere stato composto un po’ dopo, forse intorno al 1400, oppure nell’anno 1429 della prima copia. Comunque, come rilevato da David Parlett nella sua lettera alla Playing Card, Vol. XVHI, 1990, p, 73, altri brani del trattato convalidano l’anno 1377 come data della sua composizione; si veda anche la risposta di R. Decker, The Playing Card, Vol. XIX, 1990, pp. 20-1.

(The Tractatus de moribus has survived in only one 1429 manuscript and three others all of 1472 (8). There are very few variations among these four texts, but if the original is really in 1377, they must contain interpolations, perhaps by the 1429 copyist. From these texts and a number of old extant German packs, we learn of a high degree of experimentation in the composition of the normal pack in Germany in the fifteenth century, and it was following these experiments that the Queen made her first entry into the card pack. Supporters of women's liberation will be pleased to know that she was not originally introduced as inferior to the King but as of equal rank. The Tractatus de moribus describes packs in whose four suits, or two out of four, all the figures were female. A famous hand-painted fifteenth-century pack, produced between 1427 and 1431, is one example presented to us of this second type. In which feminine Unters are found. Through what was probably a later stage of development, she entered packs in which all four suits have King, Queen, Ober and Unter. An example of this is a hand-painted pack dated 1440-5, known as the Ambraser hunting pack (9); and there are many other fifteenth century German packs with four figures per suit of which the Tractatus makes no mention; it refers, however, with much enthusiasm, to a type that has not survived, with fifteen cards in each of the four suits, including King, Queen, the two marshals and a Servetta (ancilla) as the lowest card among the five figures. When, in 1470, manufacturers of French cards introduced their great innovation, the French suit-system, they borrowed the Queen from German packs with four figures, in place of the Knight in the pack with Latin suits; this was an appearance together with the king in a pack with only three figures per suit.
8. Cf. Sir Edward Augustus Bond, ‘The history of playing-cards’, Athenaeum, no. 2621, 19 January 1878, from p. 87, col. 3, to p. 88, col. 2; Peter Kopp,‘Die frühesten Spielkarten in der Schweiz’, Zeitschrift für Schweizerische Archäologie und Kunstgeschichte, Vol. 30, 1973, pp. 130-46; and Hellmut Rosenfeld, ‘Zu den frühesten Spielkarten in der Schweiz: eine Entgegnung’, ibid., Vol. 32,1975, pp. 179-80. More recently, Ronald Decker in 'Brother Johannes and the Year 1377’, The Playing Card, Vol XVIII, 1989, p. 46-7, proposed an emendation of the text, according to which the date 1377 would be the one in which playing cards arrived for the first time in his region; so the treatise may have been composed a little later, perhaps around 1400, or in 1429, the year of the first copy. However, as noted by David Parlett in his letter to The Playing Card, Vol XVIII, 1990, p, 73, other parts of the treatise validate the year 1377 as the date of its composition; see also the response of R. Decker, The Playing Card, Vol XIX, 1990, pp. 20-1.)
However Dummett thinks that a corresponding change did not usually happen to normal decks in Italy, despite his quotes from St. Bernardino and Marzio Galeotti di Narni--except in the case of tarot decks, and then always. Dummett's basis is the lack of surviving normal decks with Queens. There are, to be sure, surviving Italian decks that have Queens and no triumphs, but they are "probably" tarots. But there are many more that don't. To evaluate Dummett's claim, we would have to know how many other surviving normal decks contemporary with the tarot decks there are, how many cards are extant in each deck. He says that surviving hand-painted tarot decks outnumber handpainted normal decks 2 to 1. So if there are around 27 surviving tarot decks, there are around 9 normal decks without Queens, vs. 2 with Queens and 2 documents. I have no idea how to compute the odds that the ones without Queens have survived by chance. But it does look like normal decks lacked Queens more than they had them.  IIf so, how did tarots get Queens but not normal decks?

In the passage I quoted, the parts I highlighted might have a bearing on what city the tarot originated in, or at least acquired Queens in. We know, if nothing else, that tarot decks had Queens. Some had other female courts as well. One whole subtype of normal decks, called "Portuguese" but probably originating in Spain, had female pages in two suits. In the tarot, we find that feature in the Minchiate and in the Budapest/Met woodblock sheets, so probably it was not common in Italy before the second half of the 15th century, perhaps via Aragonese influence (from normal suits in Aragon, where the "Portuguese" type is documented). Or vice versa. And where the Arogonese would have gotten it is another story. Given that Friar John had described a full complement of female courts in his 60 card pack, it is conceivable that it came from Germany. In the 15th century Spanish suited cards made in Germany were exported to Catalonia, just where Portuguese suits were also found.

The oldest surviving tarot deck, the so-called Cary-Yale, of c. 1442 Milan,  had a full complement of female courts in all four suits; this is almost as many as Frater Johannes has in the deck of 1377. Here I ask, which cities in Italy had the closest connection with Germany in the first half of the 15th century? Milan, Mantua, and Modena (and hence Ferrara) had court connections. Many cities, including Florence, in the first half of the century, made a point of distancing themselves from Germany, due to the old Guelph fear of the Emperors. There were, to be sure, connections via the printing of cards, both in Germany and by Germans moved south. But they normally printed whatever the local market wanted.

For this new experiment, whether in Milan or elsewhere, of adding a fifth suit, it seems reasinable, they also experimented with the German innovation of female courts, not only the Queen, but, in Milan at least, other courts. Such experimentation, however, requires an openness to German practices. 


Compared to the 1980 Game of Tarot, Dummett revises his estimate of when French suits were invented, from about 1480 to "probably not earlier than 1465" [probabilemente non è anteriore al 1465"] in 1993 (pp. 34f). In his footnote to that remark, he gives his reasoning:
20. Un mazzo dipinto a mano del XV secolo, venduto all’asta da Sotheby nel 1984, è probabilmente il più antico mazzo finora pervenutoci di carte da gioco prodotte in o per la Francia. Una descrizione dettagliata del mazzo è nell’articolo di Tom Varekamp, ‘A XV-Century French Pack of Painted Playing Cards with a Hunting Theme’, The Playing Card, Vol. XIV, 1985-6, pp. 36-45 e 68-79. Si tratta di un mazzo completo di cinquantadue carte con segni di seme non standard, tutti collegati alla caccia (comi da caccia, collari di cane, cappi doppi e rotoli di corda). Minuziose ricerche da parte di Varekamp e di altri hanno fissato, per questo mazzo, la data del 1470, con un margine d’errore molto ridotto. La sua importanza per noi è dovuta al fatto che la sua composizione è esattamente quella di un mazzo di semi francesi: ciascun seme ha dieci carte numerali e, come figure. Re, Regina e Fante. Ciò rende probabile, sebbene naturalmente non certo, che il mazzo di semi francesi esistesse già all’epoca in cui questo venne prodotto, perché non si ha notizia di un tale gruppo di figure in alcuna altra forma standard di mazzo normale.

(20. A hand painted pack of the fifteenth century, sold at Sotheby's auction in 1984, is probably the oldest pack of playing cards that has come down produced in or for France thus far. A detailed description of the pack is in the article by Tom Varekamp, 'A fifteenth-Century French Painted Pack of Playing Cards with a Hunting Theme', The Playing Card, Vol XIV, 1985-6, pp. 36-45 and 68-79. It is a full fifty-two card pack with non-standard suit signs, all linked to hunting (hunting horns, dog collars, double knots and coils of rope). Painstaking research by Varekamp and others have set a date of 1470 for this pack, with a very small margin of error. Its importance for us is the fact that its composition is exactly that of a pack of French suits: each suit has ten pip cards and the figures King, Queen and Jack, making it likely, though of course not certain, that the French-suited pack already existed at the time when this was produced, because there is no notice of such a group of figures in any other standard form of normal pack.)
So he is saying that French suits not only probably didn't exist before 1465, but probably did exist by 1470. On p. 37 (passage quoted below) he is even more precise: the French system "appeared around 1465".

His idea is that the French system got the Queen from the German system--as well as adapting the German suit-signs to the greater simplicity of the French. That system didn't become fully developed until "around 1460", he says. This point then leads to another, having to do with the tarot  pack.

The tarot would not have been commonly known in France until "after the Latin suit system was forgotten over time". First, as in GOT, he says it is a "reasonable assumption" that the Latin suit system was used all over Europe until replaced by other systems in Germany and France. In particular, the French system would have taken over in France very quickly after its invention (p. 38f): it could be produced much more cheaply, needing only stencils for the numeral cards, as opposed to woodblocks. Then he says, (I highlight the most important new inference) (p. 39):
Se il mazzo dei tarocchi fosse stato noto in Francia, Svizzera o Germania al tempo in cui il sistema di semi italiano, o una sua leggera variante, era d’uso comune in quei paesi, i segni di seme usati per i tarocchi avrebbero dovuto seguire la stessa evoluzione di quelli dei mazzi normali: avrebbero dovuto esserci mazzi di tarocchi con semi tedeschi e svizzeri e, già nel Cinquecento, mazzi con semi francesi. Solo se i mazzi dei tarocchi si diffusero in quei paesi quando ormai i segni di seme latini erano stati dimenticati da tempo è possibile spiegare la sopravvivenza di quei segni di seme per le carte da tarocchi in tutta Europa fino alla metà del Settecento.

(If the tarot pack had been known in France, Switzerland, or Germany at the time when the Italian suit-system, or a slight variation, was customary in those countries, the suit signs used by the tarot would have had to follow the same evolution as those of normal packs. There would have been tarot packs in Germany and Switzerland with German and Swiss suit signs, and then, as early as the sixteenth century, with French-suited signs. Only if the tarot packs spread to those countries when the Latin suit signs had already been forgotten over time [erano stati dimenticati da tempo], can the survival of Tarot cards with those suit signs throughout Europe until the mid-eighteenth century be explained.)
This of course gives him a lead-in for his next chapter, which discusses the earliest extant tarot decks, from Milan. It is also possible that he is providing ammunition for his later contention that the French learned about the tarot from the Italians at the time of their military forays into the Italian peninsula, and not before, i.e. not earlier than 1495.

In any case, Dummett's point seems to me rather speculative. Elsewhere in the book, he gives an examples of  a particular deck surviving because it is attached to a particular game, namely Aluette. He writes (p. 35):
L’ipotesi che, prima dell’invenzione del sistema di semi francese, si utilizzassero comunemente in Francia carte con semi spagnoli è confermata dal gioco di Aluette, ancora oggi fiorente sulla costa occidentale della Francia. Per il gioco di Aluette si usa un mazzo con semi spagnoli, in un modello standard che è diventato caratteristico di questo solo gioco; il mazzo è ancora prodotto dai fabbricanti di carte francesi. Il gioco è molto antico (le prime tracce risalgono al 1502), ma ogni tentativo di attribuirgli origine spagnola è fallito — non sembra che ci sia mai stato un gioco simile in Spagna. Non può quindi esserci altra spiegazione per l’uso di carte con semi spagnoli per questo gioco che il tradizionalismo dei giocatori; quando vennero introdotte le carte con semi francesi, essi si rifiutarono ostinatamente di smettere di usare il tipo di carte a cui erano abituati.

(The hypothesis that before the invention of the system of French suits commonly in France people were using cards with Spanish suits is confirmed by the game Aluette, still thriving on the west coast of France. For the game Aluette, using a deck with Spanish suits in a standard model thas become characteristic of this game only, the deck is still produced by manufacturers of French cards. The game is very old (the first traces date back to 1502), but any attempt to ascribe Spanish origin has failed – not it seems that there ever was a game similar in Spain. There can thenbe no other explanation for the use of cards with Spanish suits for this game than the traditionalism of the players; when they were introduced to cards with French suits, they refused stubbornly to stop using the type of cards to which they were accustomed.)
The tarot deck, with its customary 14 cards per suit, quaint designs, and aristocratic pedigree, is for tarot. The French deck of simple suit cards of 13 cards each is for other, more commonplace games.

There is also the issue of what designs might have inspired the French system, and why. Dummett says: (pp. 37-38):
Il sistema di semi francese, che comparve verso il 1465, va certamente considerato un adattamento del sistema tedesco, con le Picche (Piques) al posto delle Foglie, i Fiori (Trèfles) al posto delle Ghiande e, naturalmente. Cuori francesi (Coeurs) al posto di quelli tedeschi. Le forme dei segni di seme francesi, in tutti e tre i casi, sono versioni normalizzate di quelli tedeschi. L’unica corrispondenza mancante è quella fra i Quadri (Carreaux) e le Campane; e, anche in questo caso, uno dei primi mazzi francesi — prodotto da Francois Clerc di Lione fra il 1485 e il 1496 e di cui la Bibliothèque Nationale conserva un foglio non tagliato — presenta Mezzelune al posto dei Quadri, avvicinandosi così maggiormente alla forma rotonda delle Campane.

Questa variazione fu un colpo di genio commerciale da parte dei fabbricanti francesi. I segni di seme italiani, spagnoli, portoghesi, svizzeri e tedeschi sono tutti rappresentazioni, sia pur stilizzate, di oggetti concreti, di solito multicolori e di forma variabile a seconda di quanti devono comparirne sulla [end 37] carta; sulle figure, essi compaiono spesso come parte integrale del disegno. I segni di seme francesi sono sagome monocromatiche, di dimensioni costanti e forma semplice, e non sono mai tenuti in mano da alcuna delle figure, come accade frequentemente nelle figure degli altri sistemi di semi e come nei primi tempi doveva essere usanza universale. Questa innovazione rappresentò un enorme vantaggio commerciale. I fabbricanti francesi e spagnoli avevano per lo più già eliminato dalle carte numerali del mazzo di tipo classico ‘spagnolo’ la maggior parte delle decorazioni floreali o delle vignette superflue che abbellivano i mazzi italiani e tedeschi. I fabbricanti francesi le eliminarono completamente dalle carte numerali dei nuovi mazzi con semi francesi. Ne derivò un impoverimento estetico, ma fu un colpo commerciale di prima grandezza.

Per comprenderne la ragione, è necessario rendersi conto di quale fosse il metodo standard di fabbricazione delle carte da gioco fino all’inìzio dell’Ottocento, ovvero, la stampa su matrice di legno. Su una singola matrice di legno si incidevano i profili di un certo numero di carte; le matrici venivano inchiostrate e i disegni stampati su un foglio. La colorazione avveniva per stampinatura, con tanti stampini quanti erano i colori da usare. Alla fine, quando i colori si erano asciugati, si ritagliavano le singole carte del foglio e si formava il mazzo completo. Se i dorsi portavano un disegno, anziché essere lasciati in bianco, bisognava preparare fogli di disegni per dorsi che venivano poi tagliati e incollati dietro a ciascuna carta.

Per i mazzi che usavano sistemi di semi diversi da quello francese bisognava stampare una sagoma per ciascuna carta del mazzo; occorrevano quindi almeno due matrici di legno per ciascun mazzo. Una volta inventato il sistema di semi francese, tuttavia, non fu più necessario usare la matrice di legno per stampare le sagome delle carte numerali: ciascuna di esse, infatti, poteva essere prodotta usando stampini. Di conseguenza, una matrice di legno per stampare i disegni di un mazzo con semi francesi doveva contenere solo i disegni per le dodici figure.

(The French suit-system, which appeared about 1465, must certainly be considered an adaptation of the German system, with Spades (Piques) in place of leaves, flowers (Trèfles) instead of acorns and naturally, French Hearts (Coeurs) instead of the German ones. The forms of the French suit signs, in all three cases are normalized versions of the German ones. The only thing missing is a correspondence between Tiles (Carreaux) and Bells; and even in this case, one of the first French packs - produced by Francois Clerc of Lyon between 1485 and 1496 and of which the Bibliotheque Nationale retains an uncut sheet - has Crescents instead of tiles, thus approaching more the round shape of the Bells).

This change was a stroke of commercial genius on the part of French manufacturers. Italian, Spanish, Portuguese, Swiss and Germans suit signs are all representations, although stylized. of concrete objects, usually multicolored and of variable form depending on how many must be placed on the [start p. 38] card; on the figures, they often appear as an integral part of the design. French suit signs are monochrome silhouettes, of constant size and simple form, and are never held by the  hands of any of the figures, as frequently happens in the figures of the other suit-systems, and as in the earliest time had to be the universal custom. This innovation represented an enormous commercial advantage. French and Spanish manufacturers had mostly already eliminated from the  numeral cards of the classic 'Spanish' pack most floral decorations or superfluous vignettes that graced Italian and German packs. The French manufacturers eliminated themcompletely from the pip cards of the new packs with French suits. The result was an aesthetic impoverishment, but it was a commercial hit of the first magnitude.

To understand why, it is necessary to take into account the standard method of making playing cards until the early nineteenth century, i.e., printing on a wooden matrix. On a single wooden matrix are carved the profiles of a number of cards; the matrices were inked and the designs printed on one sheet. The coloring was done by stencil, with as many stencils as there were colors to use. At the end, when the colors had dried, the individual cards of the sheet were cut and formed into the completed pack. If the backs bore a design, instead of being left in white, they needed to prepare sheets of designs for the backs that were then cut and pasted  behind each card.

For packs that used suit-systems other than the French ones, it was necessary to print a card template for each card of the pack; thus it took at least two wooden matrices for each pack. Once the French suits were invented, however, it was no longer necessary to use the wooden matrix to print the silhouettes of each of the pip cards; they, in fact, could be  produced using stencils. Consequently, a woodblock print design for a pack with French suits had to contain only the designs for the twelve figures.)
I would observe only that Crescents do not look much like bells, and Flowers do not look much like Acorns. Where Crescents came from is anybody's guess, and what their relationship to Diamonds is, is another. It could also be that Carreaux (Diamonds) came from the shapes made by intersecting Italian Bastoni, and Flowers (Clubs) from the flower pattern inside Italian Coins. Below are examples from each: Bastoni from the "Catania" and "Pierpont-Morgan-Bergamo" tarot decks, both 1450s, and a 1499 Milanese 2 of Coins compared with Clerc's way of depicting the suits of Trèfles and Piques.

Chapter 2: Milan painted cards, 46-75

In Chapter Two of Il Mondo e L'Angelo (1993) Dummett discusses the cards he classifies as Milanese. This chapter and the next correspond to Chapter 4 of Game of Tarot (GOT), with many additions.

Before I begin, I need to say that he consistently calls the Cary-Yale deck (CY for me) by its older term, "Visconti di Madrone", and the PMB deck the "Visconti-Sforza". Also he consistently does not include the Fool among the triumphs. This is from its role in the game; it never takes, i.e. triumphs in, a trick. Also, it is normally not numbered, suggesting that it is not part of the trump sequence.

These later thoughts of Dummett's on the CY seem to me of extreme interest.

The CY is considered the earliest partial tarot deck known. 11 triumphs survive, not necessarily in this order: Empress, Emperor, Love, Chariot, Fortitude, Death, Hope, Faith, Charity, Angel (of the Last Judgment), and something else, a lady with two trumpets smiling down on a scene with knights and castles below. There are enough surviving suit cards to say with certainty that there were ten number cards plus 6 court cards in each of the usual standard Italian four suits. The six court cards were King, Queen, Male Knight, Female Knight, Male Page, Female Page.

Here are the issues of note that I see:

(1) The deck is frequently dated by means of the image of the coins in the suit of coins, which has   Filippo Maria Visconti's name in it, both in the CY and in the in the Brera-Brambilla (BB) suit of coins. It is said by some to be an image of the specific coin that came into use in 1442. But Dummett says (p. 46, note 7):
7. La tecnica di queste rappresentazioni del fiorino di Filippo Maria rimane un mistero. Pare evidente che sono state fatte per mezzo di un autentico conio, ma i numismatici ci assicurano che le immagini sulle carte sono più grandi della moneta stessa; forse l’artista usò il conio per una medaglia.

(It seems evident that they have been made by means of an authentic coinage, but the numismatic experts assure us that the images on the cards are larger than on the coin itself; perhaps the artist used the coin for a medal.)
I had not seen this bit at the end about the medal idea before. I assume he is implying that the medal, and not the coin, would have been what was imprinted on the coin. That would also explain the feature that Marco noticed, at viewtopic.php?f=11&t=917&p=13807&hilit=florins#p13807, that the words around the circumference aren't in the same place on the coin and the card.

But it won't help in dating the card, except to fix it sometime in Filippo's reign. The Visconti florins had had that same design through three prior Visconti rulers (see my post at viewtopic.php?f=11&t=917&p=13807&hilit=florins#p13797. Nonetheless the medal idea is interesting, because Pisanello was in Milan in 1440-41 and did strike medals, of both Francesco Sforza and Filippo Visconti. The device used for imprinting would not have been the actual medal, because the letters would have come out reversed (likewise for a coin). But the could have been made from the dye for such a medal--actually two dies, because both sides of the coin are represented. The technology was readily available even without Pisanello, for example, in making rings that left their imprint on sealed letters, i.e. a goldsmith's product, done to leave an imprint that looked like the coin. If so, it could have been done any time between 1412 and 1447. In fact Dummett here only uses the coin argument to establish an upper limit, i.e. no later than 1447, when Filippo Maria Visconti died. I can't argue with that. However the time Pisanello was in Milan, 1440-1441, is one reason, however good, for supposing that the cards were done not long after.

(2) Dummett no longer insists that Bonifacio Bembo painted the CY and BB. This is because, while the experts were unanimously for Bembo in 1980, some experts since have seen the hand of Zavatarri in the cards.

On this point, it should be said that only one expert has been for Zavatarri, namely, Algieri; the rest have been for Bembo, although not necessarily only Bonifacio. A collaboration with his brother Ambrogio was first proposed by Boskovitz in 1978. Evelyn Welch expressed support for this view in the Encyclopedia of Art, 1996. I have reproduced the encyclopedia entry at the end of viewtopic.php?f=11&t=917&p=13407&hilit=Welch#p13420. Most recently, the catalog to a 2013 Brera Gallery exhibition in Milan, I tarocchi dei Bembo by Sandra Bendera and Marco Tanzi, makes it clear that Bonifacio Bembo sometime in these years collaborated with his brother Ambrogio. Bandera and Tanzi assign a well known series of "Lancillotto del Lago" illuminations to both, for example, based on very evident similarities between them and certain designs attributed to Ambrogio alone (I Tarocchi dei Bembo, p 28ff). By many, these Lancelot designs had previously been attributed only to Bonifacio. The Brera-Brambilla (BB), however, Bandera and Tanzi do assign to Bonifacio alone, detecting his fine work throughout. But for the Cary-Yale, they only go so far as to attribute it to the workshop; whether this is because they have not had the opportunity to examine the cards, all of which are in New Haven, I do not know. But they say (p. 11):
Si conoscono tre mazzi di carte da tarocco esegiuti nella bottega cremonese di Bonifacio Bembo...

(Three decks of tarot cards made in the Cremona workshop of Bonifacio Bembo are known...)
After mentioning the Brera-Brambilla and Pierpont-Morgan-Bergamo, they add:
Il terzo, già nella raccolta milanese dei Visconti di Modrone, è ora nella Beinecke Library della Yale University, a New Haven.

The third, earlier in the Milanese collection of the Visconti di Madrone, is now in the Beinecke library of Yale University, in New Haven.
I can find no other description of the Cary-Yale. But they are careful not to ascribe it to Bonifacio alone. In the case of the PMB, in fact, they ascribe the execution of the deck to both brothers (p. 50). Whether one did the triumphs and court cards and the other the numeral cards is not clear.

(3)There is the issue of whether the deck commemorates a marriage. Many have thought so, only disagreeing about which wedding it was. In addition, Ronald and Charlotte Deckers proposed instead that it represents Tristan and Isolde. Dummett says (p. 48): I Decker ritengono che il mazzo non sia stato fatto per alcun matrimonio, e ipotizzano che le due persone davanti alla tenda rappresentino Tristano e Isotta.

The Deckers believe that the deck was not done for any wedding, and speculate that the two people in front of the tent represent Tristan and Isolde. I would guess that this hypothesis is based on the similarity of the cards to those for a Lancelot book that was done around 1446.

It seems to me that that in proposing a marriage commemoration, no one was saying it portrayed the actual act of getting married. That would have been done in a church, and at least with a priest present. So the card, on the "marriage" interpretation, would be a meeting between bride and groom before the wedding. This was a common situation among rulers then, who typically did not meet their politically chosen brides before the time of the wedding.

Dummett reviews and evaluates three different proposals for the marriage: the 1428 (or 1427?) wedding of Filippo with Maria of Savoy, based on the red and white shield alternating with the Visconti "viper" above the couple's heads; (b) the 1441 wedding of Bianca Maria Visconti and Francesco Sforza, interpreting the banner differently; and (c) the wedding of Galeazzo Maria Sforza and Bona of Savoy in 1468. The arguments against (a) and (c) are well known. Filippo kept his wife isolated, away from him, with no chance of contacting other men, and the marriage was childless, probably never consummated. Against both (a) and (c), the white cross on a red field, besides being the shield of Savoy, is also that of Pavia, where the Visconti had a second court. Against (c), the cards in the deck with the coins and Filippo's name would have to have been saved for 20 years, and the triumphs added in the same archaic style.

None of these arguments is very persuasive. Filippo is said to have been on good terms with his wife, and a marriage deck would have looked good to his in-laws, with whom he would have been desirous of keeping the peace. And while it is true that the red and white shield was the emblem of Pavia, its alternation with a Visconti emblem on a tent with the word "Love" on it certainly suggests the joining of a Visconti with some other family possessing that shield. However those two, if both were meant as Visconti, might have served together to balance the fountain heraldic on the man's chest, which was a Sforza emblem. Finally, although this is admittedly the weakest counter-argument, the Visconti dye could have been left unused then but not thrown away, and 30 years later used in CY; also, the Bembo were known for their ability to paint in an archaic style. Such a deck would make sense as propaganda for Galeazzo Maria's claim for "bon droyt" as Filippo's direct and loyal descendent. as well as being the son of the foremost condottiere of his age. But Galeazzo Maria was not known for his subtlety, and it is not likely that such dyes, with no foreseeable future use, would have been kept for 20 years.

Dummett also has a new argument, which he makes against the 1428 interpretation, but would work against the 1441 as well, unless Filippo really wanted his daughter's marriage to be a bad one (p. 48) :
Inoltre, la carta mostra un cagnolino fra le due persone; poiché l’abbaiare di un cane alla cerimonia nuziale era ritenuto di cattivo augurio da Filippo Maria, solo un artista della massima impertinenza o temerarietà ne avrebbe incluso uno su una carta destinata a rievocare l’evento.

(In addition, the card shows a dog between the two people; as the barking of a dog at a wedding was considered a bad omen by Filippo Maria, only an artist of the highest impertinence or recklessness would have included one on a card intended to evoke the event.)
Dummett gives no source for this alleged superstition. In any case, the dog is not barking, and in general, dogs were emblems of faithfulness.

In favor of the 1441 wedding, however, there remain the Sforza heraldics of the fountain and quince in all the figure cards of Batons and Swords, and the Visconti heraldics on the Coins and Cups. These are symbolically appropriate: Sforza is the military man, Bianca Maria the rich and religious maiden. And no such alternating symbolism occurs in the other two Bembo decks. Kaplan's proposal that the cards commemorate the Francesco-Bianca marriage is not eliminated. Kaplan did not address the difficulty of the banners on the tent; but as I have said, they could both be Visconti banners, one in Milan and one in Pavia, representing the bride and Lombardy, the groom's adopted home, balancing out the fountain on the man's chest, a Sforza emblem. Dummett concludes (p. 49):
Sia che sia stato un regalo nuziale oppure no, gli anni intorno al 1441 sembrano la data più probabile. Se si è trattato di un regalo, l’ipotesi di Kaplan è la più plausibile; non sarebbe stato un gran complimento per una donna della Savoia che emblemi savoiardi fossero dipinti su una sola carta, mentre le altre carte erano piene di quelli viscontei e sforzeschi. Anche se non si è trattato di un regalo, questa data rimane ugualmente probabile. Nei suoi ultimi anni, Filippo Maria era quasi cieco. Sappiamo che, quando era più giovane, il gioco delle carte lo aveva appassionato e che aveva commissionato un mazzo dipinto a mano a Michelino da Besozzo. E pertanto plausibile che egli abbia commissionato due mazzi di tarocchi a un pittore della fama di Bembo o di Zavattari, ma non è al tempo stesso plausibile che lo abbia fatto negli ultimi anni di vita. Il mazzo Visconti dì Modrone può essere ragionevolmente datato verso il 1441, e il mazzo Brambilla un po’ più tardi.

(Whether it was a wedding gift or not, the years around 1441 seem the most likely date. If it was a gift, Kaplan's hypothesis is the most plausible; it would not be a great compliment to a woman to have emblems of Savoy painted on a single card, and other cards were full of those of the Visconti and Sforza. Even if it was not a gift, this date remains equally likely. In his later years, Filippo Maria was nearly blind. We know that when he was younger, he was passionate about playing cards and had commissioned a hand painted pack from Michelino da Besozzo. It is therefore plausible that he commissioned two packs of tarot cards to a painter of the reputation of Bembo or Zavattari, but it is not very plausible that he did it in the last years of life. The Visconti di Modrone pack can be reasonably dated to 1441, and the Brambilla pack a little later.)
For myself, the 1441 marriage commemoration idea remains the most plausible, because of the heraldics on the court cards and the Love card. It need not have been made precisely in 1441, although relations between Sforza and Filippo Maria soon soured. Another occasion could be the birth of their first child, a son, in 1444; relations had temporarily improved.

Here I would add that there was a strong tradition among the Visconti of giving illuminated manuscripts featuring their own portraits and heraldics as marriage gifts, (as shown by Kirsch; see viewtopic.php?f=11&t=917&sid=126a9f155bea4789ececc09c61577b38;), of which illuminated cards might have been a fashionable extension. Francesco and Bianca themselves, much later, commissioned a marriage commemoration tapestry reminiscent of the scene on the card (

As to the banners, their both being Visconti is plausible, but the alternating banners still look like they are complementing the scene below and so the union of two houses, as early commentators said. It seems to me that they could refer back to a previous illumination, whether on a tarot card or somewhere else, of the Filippo-Maria marriage, perhaps commemorating the 1427 marriage. This is a standard technique in many works of art; it situates the work in a tradition. If done before the marriage, the work in question would have given a much needed (considering his previous wife's fate. of being beheaded) impression of a lover's devotion to the match--or done afterwards to give that impression to outsiders.

I have been criticized on THF for hypothesizing an actual tarot deck of 1427-1430, one without the Sforza stemmi but with an otherwise similar Love card. It was said that I was introducing a "Russellian teapot", i.e. something as unwarranted as hypothesizing a teapot floating in space somewhere between the earth and Mars. But the analogy does not hold. There is no reason to hypothesize a teapot floating in outer space between the earth and Mars, but plenty of reason for thinking that the Cary-Yale was not the first tarot deck in Milan. For one thing, there were already tarot decks in Florence, a city for which Francesco was working at that time. For another, card decks get used up with use. So it is to be expected that an earlier deck, perhaps not as expensive, would not have survived. Or Maria took it to Savoy after Filippo's death, where it was lost.

For his part Dummett makes no claim that that the Cary-Yale was the first deck in Milan. In fact, he hypothesizes just the opposite. On p. 106 Dummett himself hypothesizes the invention of the tarot deck in Milan around 1428:
Possiamo redigere dunque una cronologia provvisoria, basata di necessità su congetture; le date sono naturalmente approssimative:
1428: i tarocchi sono inventati alla corte viscontea.
1430: la corte estense di Ferrara conosce i tarocchi.
1435: i tarocchi si diffondono a Bologna.
1440: i fabbricanti di carte cominciano a produrre mazzi di tarocchi a buon prezzo, stampati da matrici di legno.
1442: i tarocchi si diffondono da Bologna a Firenze.
1444: la composizione del mazzo di tarocchi diventa standardizzata dappertutto.

(We can therefore draw up a provisional chronology, based of necessity on conjecture; the dates of course are approximate:
1428: Tarot is invented in the court of the Visconti .
1430: the Este court in Ferrara knows the tarot.
1435: tarot spreads to Bologna.
1440: card makers begin to produce decks of tarot cards at a good price, printed by woodblock.
1442 tarot spreads from Bologna to Florence.
1444: The composition of the tarot deck becomes standardized everywhere.)
This passage admittedly is out of date at least in one respect, in supposing 1442 as when the tarot spread to Florence: it is clear that tarot decks existed there in 1440 and probably before ( But this does not affect the hypothesized 1428 date for Milan. I will discuss these places and times when I get to chapter 6. My point is that Dummett hypothesizes decks not only in Milan but elsewhere considerably before the first recorded instance that he knew about, that of a court purchase in Ferrara from a Bolognese merchant in 1442. The additional datum of a deck made in Florence for one of its condottiere, Malatesta, makes no difference, especially when the other main condottiere was Francesco Sforza, who had previously served in Milan.

It might be argued that Dummett, unlike me, does not hypothesize any specific details about this deck. It seems to me that we can indeed hypothesize one specific detail of 1428, not with any certainty, to be sure, but with some weak claim to justification. The justification is precisely the presence of the banners in the later deck, the CY, which look so much like those of a courtly love situation particular to a Visconti and a Savoy. It is a detail, one may hypothesize, carried over from the previous deck. Given that Sforza was a commoner, putting alternating banners of the two houses being joined, as though they were equal, would not have been appropriate. So Filippo uses the old ones. It gives his wife some additional recognition, too.

(3) Another issue that Dummett takes up is that of what cards and how many are missing from the 11 that have come down to us. As to how many triumphs there were, he says various things. First he says:
E impossibile stabilire se il mazzo Visconti di Modrone sia stato un esperimento isolato, che si distaccava da una norma già stabilita, o se sia runico esempio superstite di uno stadio primitivo in cui il mazzo dei tarocchi non aveva ancora acquisito la struttura che doveva in seguito diventare canonica. Se esso rappresenta uno stadio primitivo, è altresì impossibile stabilire se si tratti di uno stadio in cui coesistevano notevoli variazioni nella composizione dei mazzi di tarocchi o in cui prevaleva una norma ben precisa, diversa da quella che sarebbe stata osservata in seguito.

(It is impossible to determine if the Visconti di Modrone pack was an isolated experiment, that departed from a norm already established, or it was the unique surviving example of a primitive state, in which the tarot pack had not yet acquired the structure that was to become canonical. If it represents a primitive stage, it is also impossible to determine whether it is a stage in which there co-existed considerable variations in the composition of Tarot packs or one in which a precise norm prevailed, different from that which would was observed afterwards.)
That seems to me very sensible.

However he also says what he said in GOT, against the idea that it had only 14 triumphs; now, however, he makes it it is "extremely likely", unless the deck was already standardized at 21, that it contained 24, despite the fact only 11 have come down to us; at the same time, this is "pure hypothesis", until more reliable information is at hand (in the lengthy quote that follows, I highlight these phrases). Although these statements "impossible to determine", "extremely likely", and "pure hyothesis", don't actually contradict each other, the middle one has a very different tone out of context than the other two At least one of these three statements is bound to sound right, if the truth is ever known. Here is what he says (pp. 51-52, basically what he said in GOT; but I give the 1993 statement as a groundwork for discussion:
L’ipotesi che il mazzo Visconti di Modrone non fosse una variante isolata ma rappresentasse uno stadio primitivo nell’evoluzione del mazzo dei tarocchi offre un valido motivo per ritenere che si tratti del più antico gruppo di carte da tarocchi pervenutoci e quindi sicuramente prodotto durante il regno di Filippo Maria Visconti e probabilmente prima del mazzo Brambilla: quest’ultimo, infatti, evidentemente aveva solo le solite quattro figure per ciascun seme, anche se, dal momento che ci restano solo due dei suoi trionfi, non possiamo sapere con certezza quali soggetti fossero ritratti su quelli andati perduti. Poiché una delle virtù di solito inclusa nel mazzo dei tarocchi, la Fortezza, compariva nel mazzo 52 Visconti di Modrone, è estremamente probabile che anche le altre tre virtù cardinali — Giustizia, Temperanza e Prudenza — figurassero originariamente fra i trionfi di quel mazzo, formando, con le tre virtù teologali, la consueta serie delle sette virtù maggiori. Una delle curiosità del gruppo standard dei trionfi nel mazzo dei tarocchi è che la quarta virtù cardinale, la Prudenza, non è inclusa a fianco delle sorelle; ma, quando incontriamo la Fortezza insieme a Fede, Speranza e Carità, è difficile non pensare che l'interò gruppo dovesse in origine essere presente.

Non sappiamo per certo se il mazzo Visconti di Modrone comprendesse il Matto, o quanti trionfi ne facessero parte. È stata avanzata l’ipotesi che, a parte le tre virtù mancanti, esso includesse solo quei trionfi che ci sono pervenuti, in tutto, quindi, quattordici. Quattordici trionfi e sessantaquattro carte dei semi porterebbero il numero di carte del mazzo alle consuete settantotto, seppure distribuite in modo diverso. Il totale delle carte in un mazzo di tarocchi non è, tuttavia, particolarmente significativo: ciò che conta davvero è il rapporto fra il numero dei trionfi (a parte il Matto) e il numero delle carte dei semi. È vero che questo varia notevolmente da un tipo all’altro di mazzi di tarocchi: ma è estremamente probabile che, in questa prima fase di evoluzione del mazzo, il rapporto fosse l’unico elemento costante. Nel mazzo a settantotto carte, il rapporto è 3 a 8: in altre parole, il numero dei trionfi è esattamente una volta e mezzo quello delle carte di ciascun seme. Se il rapporto nel mazzo Visconti di Modrone era questo, dovevano esserci ventiquattro trionfi. Esso avrebbe allora potuto includere tutti i soggetti standard e, in più, Fede, Speranza e Carità; oppure, se la Prudenza era uno dei trionfi, sarebbero stati presenti tutti i soggetti standard tranne uno. Tutto questo è, tuttavia, pura ipotesi: solo l’imprevista scoperta di nuovi elementi potrebbe darci informazioni sicure sui trionfi mancanti del mazzo Visconti di Modrone.

(The hypothesis that the Visconti di Modrone [i.e. CY] pack was not isolated but represents a variant of one primitive stage in the evolution of the tarot pack offers a valid reason for believing that it is the oldest group of tarot cards come down to us and thus definitely produced during the reign of Filippo Maria Visconti and probably before the Brambilla pack: the latter, in fact, clearly had only the usual four figures for each suit, although, since there are only two of its triumphs, we cannot know with certainty, of those subjects lost, which were portrayed. As one of the virtues usually included in the tarot pack, Fortitude, appeared in the 52 Visconti di Madrone pack, it is extremely probable that the other three cardinal virtues - Justice, Temperance, and Prudence - also originally figured among the triumphs of the pack, forming, with the three theological virtues, the usual series of the seven greater virtues. One of the curiosities of the group of standard triumphs in the tarot deck is that the fourth cardinal virtue, Prudence, is not included alongside her sisters; but when we encounter Fortitude together with Faith, Hope and Charity, it is difficult not to think that the entire group should have originally been present.

We do not know with certainty if the Visconti di Modrone pack included the Fool, or how many triumphs there were. It has been suggested that, apart from the three missing virtues, it included only those triumphs that have come down to us, in all, therefore, fourteen. Fourteen triumphs and sixty-four suit cards would bring the number of cards in the pack to the usual seventy-eight, although distributed differently. The total number of cards in a Tarot pack is not, however, particularly significant: what really matters is the ratio between the number of trumps (apart from the Fool) and the number of suit cards. It is true that this varies greatly from one type of Tarot pack to another, but it is extremely likely that, at this early stage of evolution of the pack, the ratio was the only constant. With seventy-eight cards in the pack, the ratio is 3 to 8: in other words, the number of triumphs is exactly one and a half times that of the cards in each suit. If this was the ratio in the Visconti di Modrone pack, there would have to be twenty-four triumphs. It would thus have been able to include all the standard subjects and, in addition, Faith, Hope and Charity; or, if Prudence was one of the triumphs, it contained all the standard subjects except one. All this is, however, pure hypothesis: only the unexpected discovery of new elements could give us reliable information about the triumphs missing from the Visconti di Madrone pack.)
Dummett's argument that Prudence, Justice, and Temperance would reasonably have been part of the deck, because there are already all three theological virtues and one cardinal virtue in the surviving cards, seems to me a good one. The seven virtues were a common theme in art. So likely, if the person who chose the subjects was thinking in a strictly traditional way, they all were there. Furthermore, the Popess in the PMB carries attributes of Prudence (also Wisdom), namely a book and a staff-cross (see my post at viewtopic.php?f=12&t=970&p=14202&hilit=prudence+staff#p14202). So perhaps the Popess substituted for Prudence and/or Wisdom. This is actually a suggestion that Dummet himself makes later in the book. Since Giotto used a cross-staff and a scroll for Faith, that is another possibility.

The rest of this passage, about the preferred principle being the ratio of triumphs to suit cards, seems to me more dubious. The problem is that we cannot say what principle governed triumphs vs. normal suits, if any, from looking at what happened later. We do not know "what really matters" at that early stage of development. What happened later, including decisions by people in other cities. was decided later, not at the earlier time.  That person, or those persons, may or may not have used the same principle used in the Cary-Yale, if there was one. The principle that Dummett enunciates is one possible way of continuing the tradition (which also helps define the tradition). But in fact we don't know. There are merely arguments for and against.

We do know that normal decks had a variety of numbers of suit cards, 40, 48, 52, etc. We know that later there were different tarot games for different numbers of suit cards, too. So another possibility is that the triumphs (excluding the Fool) were 21 in the CY, for whatever reason, and that whoever came next used the principle of "same number of trumps" and only the suit cards changed in number. In that case, the ratio would vary, but depending on the number of suit cards. These later decisions would then define the standard. It does not make the previous period, before such decisions, any the less "primitive".

Another realistic possibility for "same number of triumphs": it could have been 19 plus the Fool at the beginning, given that the PMB, also from Milan, has that number (and no hand-painted Milan deck has the Devil or Tower). The Star, Moon, and Sun of the PMB have visual similarities with Hope, Faith, and Charity, respectively and might have been replacements for them. That possibility is also suggested by the place of these virtues in Minchiate, which are exactly in the same spots as the Star, Moon, and Sun in the standard tarot order. Even though done much later, the Minchiate inventor might have been thinking about earlier decks with the theological virtues when he restored them to the sequence.

Another reasonable alternative that Dummett has not considered is that the number of triumphs equal the number of cards in each of the suits, whatever it was. Since suits typically had 13 or 14 cards,This is supported by various documents attesting to the making of "13 new playing cards" in 1422 (, "14 figures" in 1441 (, not known to Dummett in 1993), and decks of 70 cards in 1457 (; they might be 5x14-- or (4x12) + 22, as Franco Pratesi recently suggested). So in the CY, with 16 cards per suit, there would be 16 triumphs. If someone initially used "same number as the suits" as a principle, the next person could have changed it to "same number of trumps". Or vice versa.

I would also make an argument to literary and visual precedents for the sequence as a whole. Visual representations in games (including board games) devised deliberately rarely are just random; they have a rationale. In the triumphs as in the suits. it would have been easy just to number the cards, like pages in a book, and give free rein to the artists' imaginations for the pictures. But it didn't happen. You had to memorize the sequence, even after a few decks, e.g. the Sola-Busca, in fact were numbered. There was very likely a reason the sequence had to be memorized, something about life. The reason may have changed over time, too: from something satirical to something serious, for example.

Here I would build on an idea expressed by Dummett, the closest thing to an overall plan to the tarot that I find in his book, on pp. 106-107, immediately after the section I quoted earlier from p. 106). By "scope" in the first sentence I think he means whether divination or occult meanings were involved, which is the next subject he takes up. By the "carte supplementari" he means not the 6 second artist cards of the PMB, but rather the whole set of triumphs added to the normal deck.
Prima di considerare lo scopo dell’invenzione, vale la pena di domandarsi perché le carte supplementari furono chiamate ‘trionfi’.

Molti hanno cercato di spiegare il termine «trionfi» con l’uso dei ventuno trionfi nel gioco, cioè come quelli che ‘trionfano’ sulle altre carte; e non siamo in grado di dimostrare l’inesattezza di questa spiegazione. Più suggestiva è, tuttavia, una brillante ipotesi di Gertrude Moakley. La studiosa ritiene che il nome non abbia niente a che fare con l’uso delle carte, ma solo con ciò che vi è raffigurato: la serie dei trionfi rappresenterebbe una specie di corteo trionfale. Come è documentato da Burkhardt e Moakley, uno dei passatempi prediletti delle corti del Rinascimento italiano era proprio l’allestimento di tali cortei trionfali con carri addobbati di figure derivate dalla mitologia classica o raffiguranti astrazioni quali l’Amore, la Morte, ecc.: una metamorfosi del trionfo di un generale o imperatore romano in un elegante intrattenimento allegorico. Un elemento frequente di questi trionfi rinascimentali era l’idea che è al centro del poema petrarchesco I Trionfi, in cui ciascuna astrazione personificata trionfa, sconfìggendola, sulla precedente; così nel poema l’amore trionfa su dei e uomini, la castità sull’amore, la morte sulla castità, la fama sulla morte, il tempo sulla fama e l’eternità sul tempo. L’ipotesi sarebbe confermata se fosse possibile spiegare i soggetti dei trionfi del mazzo dei tarocchi come parte di un corteo trionfale di questo genere; ma, malgrado i notevoli sforzi di Gertrude Moakley, integrati successivamente da quelli di Ronald Decker, tale spiegazione, pur plausibile in linea di principio, è difficile da rendere convincente nei dettagli. Ciò nonostante, in mancanza di meglio, possiamo accettare come probabile, seppure assolutamente non come certo, che fu questa associazione di idee a ispirare l’uso della parola «trionfi» per le carte supplementari del mazzo dei tarocchi.

(Before considering the scope of the invention, it is worth asking why the additional cards were called 'triumphs'.

Many have tried to explain the word "triumph", with the use of twenty-one trumps in the game, that is, as those that 'triumph' over the other cards; and we are not able to demonstrate the inaccuracy of this explanation. More striking, however, is a brilliant idea of Gertrude Moakley. The researcher believes that the name has nothing to do with the use of the cards, but only with what is shown: the set of trumps would represent a kind of triumphal procession. As documented by Burkhardt and Moakley, one of the favorite pastimes of the courts of the Italian Renaissance was just the preparation of these triumphal processions with floats decorated with figures derived from classical mythology or depicting abstractions such as Love, Death, and so on. : A metamorphosis of the triumph of a Roman emperor, general in an elegant allegorical entertainment. A common element of these Renaissance triumphs was the idea that is at the center of the poem I Trionfi [The Triumphs] of Petrarch, in which each abstraction personified a triumph, triumphing over the previous one; thus in the poem, love triumphs over gods and men, the chastity over love, death over chastity, fame over death, time over fame, and eternity over time. The hypothesis would be confirmed if it was possible to explain the subjects of the trumps of the tarot deck as part of a triumphal procession of this kind ; but, despite the considerable efforts of Gertrude Moakley, supplemented later by those of Ronald Decker, such an explanation, though plausible in principle, is difficult to make convincing in detail. Nevertheless, in the absence of anything better, we can accept as probable, though not as absolutely certain, that it was this association of ideas inspiring the use of the word "triumph" for the additional cards of the tarot deck.
It seems to me that most of the Petrarchan triumphs (to which I think can be added one from Boccaccio's Amorosa Visione, and perhaps also a view of Love in a more positive sense than Petrarch expresses), are already in the CY surviving cards: Love, Chastity as the lady on the Chariot, Fortune (from Boccaccio, and seen in the nearly contemporary Brera-Brambilla deck by the same workshop), Death, Eternity (the one with the Angel of the Last Judgment). Probably Fame is there, too, in the so-called World card, with its trumpet-lady on top and a scene with at least one knight at the bottom.

If Dummett can extrapolate from four virtues the existence of the other three, surely I can extrapolate one or two more Petrarchan triumphs; Time was in the other early decks, usually called the Old Man. In the PMB deck, his first surviving appearance on a card, he carries an hourglass rather than a lantern. The virtues then take their places as tools for overcoming the negative and achieving the positive of the various triumphs. 7 triumphs plus 7 Virtues is 14. Since the Emperor and the Empress are among the surviving cards, perhaps coming from an earlier game documented as "VIII Emperatori" (, we can get as far as 16 in this way but no further.

We don't know what else was in the game of Emperors. Possibly there were other high and low members of society (e.g. Pope, Popess, Bagatto, Fool), as in the frontispiece to an edition of Petrarch's De Remediis done in Milan c. 1400, of which MJ Hurst posted a color reproduction, ... detail.jpg. If so, that would make 18 plus the Fool. Also, these cards were later all next to each other in the order. That is not true of the the 3 remaining. (The Hanged Man could conceivably count as a very low member of society; but he was not in any grouping of such figures that I know of, nor is he in that part of the sequence.) However it is at least as likely that the Emperor and Empress were at the beginning of the sequence by themselves, just to represent the card-player himself or herself, at the beginning of life, with material privileges but nowhere near the goal..

Another schema is the "chess analogy", of 16 pieces to 16 triumphs that "Huck" has worked out for this deck. Its results are consistent with the "triumphs plus virtues plus Emperor and Empress (only)" model. Over-determination is a valid principle for works of art.

Another fact is that when the Cary-Yale came to Yale, the existing triumphs had been assigned to particular suits (they still are, online; when I emailed Yale to ask the answer was "they came that way"). If this was original--a questionable assumption, I know--it would imply a number divisible by 4: i.e. 12, 16, 20, or 24, not including the Fool.

I am not proposing this as the "original" tarot, about which I have less of an idea. I am just proposing it as one reasonable alternative fitting the Cary-Yale.

I can also think of an argument that supports Dummett's proposal. The "Olympian gods" deck of Marziano, much like the tarot that followed, if it had 10 numeral cards per suit plus kings, i.e. 11 cards, the ratio of suit cards to triumphs is very nearly 3:2. If there were one less triumph and one less suit card, it would have been 3:2 (and very close to 3:8 as well in the other ratio) This deck, however, is also a precedent for 16 special cards in the CY: 16 in one, 16 in the other.

So there are several reasonable possibilities at that time, including "no principle" in that early "primitive" time, ranging from 13 to 24 triumphs (excluding the Fool), but with 16, 18, 19, 21, and 24 the most reasonable (with or without the Fool). I think we have to leave it at that, pending more information.


On the Brera-Brambilla, Dummett estimates that it, and the PMB as well, had 21 triumphs plus the Fool (p. 51).
A parte le Spade diritte nel mazzo Visconti-Sforza e le frecce sulle figure del seme di Bastoni nel mazzo Brambilla, la composizione e i disegni di questi due mazzi si accordano pienamente con l’ipotesi che siano stati, in origine, due mazzi di tarocchi del formato standard di settantotto carte: tutti i semi contenevano evidentemente le regolamentari dieci carte numerali e quattro figure; e i trionfi pervenutici, inclusi i sei del mazzo Visconti-Sforza, che sono opera del secondo artista, presentano tutti soggetti convenzionali.

(Apart from the straight Swords in the Visconti-Sforza pack and the arrows on the figures in the suit of Batons in the Brambilla pack, the composition and patterns of these two packs fully agree with the hypothesis that they were originally two tarot packs of the standard form of seventy-eight cards: all the suits demonstrably contain the regular ten numeral cards and four figure; and the triumphs that have survived, including the six in the Visconti-Sforza pack that are the work of the second artist, all show the conventional subject matter.)
As to the number of triumphs, it seems to me that there are other reasonable possibilities. It is so close in time to the Cary-Yale that it might have been part of the "primitive"--and experimental--time period.

The significant change in the BB compared to the CY is that its suit cards have been reduced to 14. If the designer of the deck had in mind keeping the same number of subjects in the fifth suit as in the others, that would have necessitated the removal of 2 cards. Which two might they have been? For clues, we need to look at the next deck, the PMB. In that deck, all 3 theological virtues are absent, not just two of them. Three of the four cardinal virtues are absent. But if one was turned into the Popess, that leaves two that might have been removed to make the subjects equal 14.

That the BB was done a little later than the CY ("il mazzo Brambilla un po’ più tardi") is not only by its 14 cards per suit, but because the cards are closer to the PMB than the CY in style, "severe" and sparser in extraneous details. Here is what he says, in relation to the three packs, p. 47 (here I have added explanatory notes to the translation, about which decks are which. In the beginning, he is speaking of the Cary-Yale):
Questo mazzo è collegato al mazzo Brambilla dall’impiego del fiorino di Filippo Maria e dall’altemarsi rovesciato di frecce e mazze nel seme di Bastoni. Ciò nonostante. Giuliana Algeri ritiene a ragione che il suo stile artistico differisca da quello degli altri due mazzi. Gli altri mazzi sono in un certo senso severi: dai trionfi e dalle figure è assente ogni dettaglio superfluo per la rappresentazione dei loro soggetti, mentre nel mazzo Visconti di Modrone le carte figurate mostrano spesso personaggi supplementari e in essenziali. Se fu dipinto da un altro artista, si trattò comunque di un artista della stessa scuola.

(This pack [the CY] is connected to the Brambilla pack by the use of the florin of Filippo Maria, and of alternating inverted arrows and clubs in the suit of Batons. Nonetheless, Giuliana Algeri rightly feels that the artistic style differs from that of the other two packs [the BB and the PMB]. The other packs are in a certain sense severe: every detail unnecessary for the representation of the subjects is absent from the triumphs and figures, while in the Visconti di Modrone pack [the PMB] the figure cards often show additional inessential personages. If it was painted by another artist, it was nevertheless an artist of the same school.)
Another reason for dating the BB later than the CY is that the 14 cards per suit reflect a less "primitive" period in the tarot's development. Later in the book, on p. 106, quoted in my last section, Dummett hypothesizes the time of the tarot deck's standardization to around 1444. This date  for the PMB is generally sustained (1442-1444, say Bandera and Tanzi, p. 36), as is the attribution to Bonifacio Bembo. Whether tarot decks in general were standardized at 21 triumphs, 56 suit cards, and the Fool, at that time is another matter. That one deck had 56 suit cards doesn't say much.

A major problem with the BB is that there are only two surviving triumphs, the Emperor and the Wheel of Fortune. It is hard to say much from that.We don't actually know for sure that it is a tarot deck, because of the pack known from documents as "VIII Imperatori" ( Whether it contained any other special cards besides Emperors is unknown. From Spain there is a report from c. 1450 of a proposed deck with four suits and just one special trump, called "Emperador". A Wheel of Fortune is not mentioned. However from the presence of that card, and the way in which the BB Emperor seems intermediate between the CY and PMB Emperors, and the Wheel so much resembles the PMB Wheel, it is reasonable to assume that the BB, like the other two, was a tarot deck. (For the Emperor, compare and the two at the top of For the Wheel, compare and


As to the PMB, which he calls the "Visconti-Sforza", Dummett says, e.g. in my quote above discussing the BB, that it, too, reflects the 21 plus Fool standard model, since there are 14 cards per suit and 19 surviving triumphs.

All this follows from and says no more than what he has already said in his "pure hypothesis" that the CY had 24 trumps: the principle is that of maintaining a ratio of 3 to 2. That such a principle was used is pure speculation, as I have already argued. Is there anything more substantial than that?

One can count the triumphs in the PMB. They are 19 plus the Fool, only two less than the standard 21. It is a considerably smaller jump from 19 to 21 than from 11 to 24. There are also 14 suit cards per suit.

One problem is that 6 of the 19 triumphs are by a different artist. in what he says art historians say is the style of a different period and place (Ferrara 1475 vs Milan 1450s). I will get into the reasons for this dating under another heading. We have to ask: are these 6 additions substitutions (meaning for subjects that are in the CY but not those of the PMB new cards), or replacements (meaning new versions of subjects in the PMB originally? (To view these cards, go to; they are called Fortitude, Temperance, the Star, the Moon, the Sun, and the World.)

Dummett argues strongly that the original PMB did not contain only the 13 surviving triumphs, plus the Fool, of the first artist. He says:
Se, dei sei trionfi dipinti dal secondo artista, nessuno era compreso nel mazzo Visconti di Modrone, sarebbe plausibile l’ipotesi che si tratti di addizioni successive. Le cose non stanno però cosi: versioni di due di essi, la Fortezza e il Mondo, sono incluse fra le carte Visconti di Modrone. Poiché la Giustizia ha fatto fin dall’inizio parte del mazzo Visconti-Sforza, non si può sostenere che le virtù fossero originariamente del tutto assenti da questo mazzo.È pertanto estremamente probabile che le versioni della Fortezza, della Temperanza e del Mondo ad opera dall’altro pittore siano o componenti del mazzo originario o sostituti di carte smarrite. Se le cose stanno così, si indebolisce notevolmente l’ipotesi che la Stella, la Luna e il Sole siano addizioni successive; anch’esse sarebbero o componenti originari o sostituti.

(If, of the six triumphs portrayed by the second artist, none were included in the Visconti di Modrone pack, it would be plausible to assume that they were succeeding additions. But things are not so: versions of two of them, Fortitude and the World, are included among the Visconti di Modrone cards. As Justice was part of the Visconti-Sforza pack from the beginning, it cannot be argued that the virtues were originally entirely absent from this pack. [54] It is therefore highly likely that the versions of Fortitude, Temperance and the World by the other painter were components of the original pack or substitutes for lost cards. If this is so, it weakens considerably the hypothesis that the Star, Moon and Sun are later additions; these also would be original components or substitutes.)
As to the necessity of the 2 missing virtues, that all depends on the BB having also followed the 3:2 ratio principle. If the principle was instead that of making the fifth suit equal to the other four, the absence of the 2 virtues can be explained as a cut needed to reduce the number of subjects to 14.

Just because the 3 later cards repeated old subjects (or apparently so, for the World) doesn't mean that the three replaced three others on the same themes. We have to remember that the CY had 3 subjects that the surviving PMB cards did not have, namely, the theological virtues of Hope, Faith, and Charity. We have no idea when they dropped out. Perhaps they existed in the original PMB, and someone, having visited another city in the 1460s-1470s, decided to replace them with the Star, Moon, and Sun. At the same time, that person decided that the Fortitude card should commemorate Francesco Sforza, perhaps even changing the name to "Forza", and the Temperance card--along with the Star and the Moon--should commemorate their sister Elisabetta, who died after childbirth in 1472. I have elaborated this argument at (search term "Elisabetta")

Dummett claims that "Forza" never appears in the 16th century literature as the title of a card (p. 71):
Nella letteratura italiana del Cinquecento, la carta che rappresenta la virtù cardinale è sempre chiamata ‘la Fortezza’, mai ‘la Forza’, come avviene nel moderno Tarocco piemontese;

(In Italian literature of the sixteenth century, the card that represents the cardinal virtue is always called 'la Fortezza', never 'la Forza', as in the modern Piedmont Tarot;)
"Forza" is in fact an alternate name for Fortezza in Folengo,, and Imperiali, ... a_1550_ca; and Alciato calls it "Forti",

Another possibility is that the theological virtues were removed for the first version of the PMB and replaced with 3 other cards, so as to keep the number of triumphs at 14. The new cards would have been the Bagatto, the Pope, and the Hanged Man. The Popess would have at some point replaced Prudence.

Another problem is with Dummett's leap from 19 to 21. There are numerous other examples of this "Milan standard pack" later in the century, and none of them have the Devil or the Tower. That could have been by chance or later intentional removal, but it most straightforwardly suggests that, whatever was the case in other cities, these two cards weren't present in the Milanese hand-painted decks.

In favor of Dummett's hypothesis, however, is that the Tower, at least, is present in another city's cards, those of the so-called "Charles VI" tarot, and that both Tower and Devil are in the Cary Sheet, of Milan c. 1500. But the "Charles VI', besides being from a different city, also is later than the original PMB. by 10 to 20 years; the Cary Sheet is much later, by around 50 years. Perhaps whoever added the 6 other cards didn't like the Tower card and thought it might bring bad luck.
As I have already said, Dummett gives two possible approximate dates for the PMB (which lacks the Visconti florins): 1450 and 1475 (p. 50). Here is what he says:
Se le sei carte secondarie del mazzo Visconti-Sforza rimpiazzano carte precedenti, il mazzo è databile intorno al 1450; ma se sono [49] opera di un collaboratore, secondo l’ipotesi di Ronald Decker, la data cade attorno al 1475, a causa dello stile delle sei carte.

(If the second six cards replace cards in the Visconti-Sforza pack earlier, the pack can be dated around 1450; but if they are [49] by an assistant, according to the hypothesis of Ronald Decker, the date falls around 1475, because of the style of the six cards.)
The 1475 date is on the assumption that the second artist cards were painted at the same time as the first artist cards, by an assistant. The 1450 is on the assumption that the second artist cards were painted later, i.e. 1475, as replacements for earlier versions of the same cards. Later he gives a lower limit for the date of the second artist cards. Speaking of the decks that derive from the PMB, he says (p. 55):
Quelli che contengono copie di una delle sei carte Visconti-Sforza dipinte dal secondo artista devono essere posteriori al 1470, che è la prima data plausibile per quelle carte;...

(Those that contain copies of one of the six Visconti-Sforza cards painted by the second artist must be later than 1470, which is the first plausible date for those cards
The reason for the late dating of these cards is their style (p. 45):
L’opinione corrente è che queste sei carte siano state dipinte circa vent’anni dopo da un ignoto artista di scuola ferrarese; ma Ronald Decker considera il mazzo frutto di una collaborazione ineguale fra due artisti del tempo del duca Galeazzo Maria 6.

(Current opinion is that these six cards were painted about twenty years later by an unknown artist of the School of Ferrara; Ronald Decker considers the deck the result of an unequal collaboration between two artists of the time of the Duke Galeazzo Maria 6.)
6. See R. Decker, 'Two Tarot Related Studies', Part III, Journal of the Playing-Card Society, Vol IV, 1975, pp. 46-52.
But most scholars, he adds, consider that the deck, except for those cards, was done during the reign of Francesco Sforza, "probably in the earlier years of his reign" [probabilmente nei primi anni del suo regno]. Bandera and Tanzi, 2013, date the PMB to 1455, based on similar stylistic innovations in other work by Bonifacio Bembo at that time, notably a Madonna and Child fresco recently uncovered in Cremona. See my discussion, with pictures, at I can't myself see the resemblances that Bandera argues for.

In 2007 (Artibus Historia 56, pp. 15-26) Dummett discarded the idea that the second artist cards were replacements and hypothesized that all the cards were painted around 1462-3 by Bonifacio and Benedetto Bembo, Benedetto as the second artist. He based this on stylistic similarities to ecclesiastical artworks done around that time attributed to Benedetto. I transcribed the relevant portion of his article at viewtopic.php?f=11&t=365&p=4912&hilit=Artibus#p4912.

I don't see the similarities to Benedetto's work, which exists in good reproductions. For examples, do a search under his name in Google Images.
I hypothesize that the first artist cards are from the early to mid 1450s. They may have been done for the Sforza children to use, since they show much wear. The Knight of Batons (at left) looks about the age Galeazzo Maria Sforza, who rode a white horse, would have been in 1455 or so. He is depicted on such a white horse in Benozzo Gozolli's famous "Procession of the Magi" painting of 1459, at the time he visited Florence (see

The six added cards, which show much less wear, in composition are somehat like two of the Belfiore Muses, probably among those done for Borso d'Este around 1449 (see viewtopic.php?f=12&t=463&p=5923&hilit=Belfiore#p5923). Ciriaco d'Ancona saw them then, just before moving to Cremona, where he died in the early 1450s. The style of these Muses later became that of the "Tarot of Mantegna" images, which seem to have influenced the Schifanoia Palace frescoes in c. 1470, a project on which many painters participated, including some from other cities. The 1475 date seems reasonable to me, both from the decreased wear and from my hypothesis that the lady on three of the cards (Temperance, Star, Moon) is likely Elisabetta Maria Sforza, Galeazzo Maria Sforza's sister, who died at age 16 as a result of childbirth, in 1472 (viewtopic.php?f=11&t=365&p=4821&hilit=Elisabetta#p4821).

Stylistically, however, they could be any time from 1450 to 1485, depending on when the artist and/or his patron decided to follow the model of these Belfiore Muses. After 1472 is most likely, because the Schifanoia project would have been the most logical way for a Lombard artist to learn Ferrarese styles. Dummett's 1462 date is probably based on the date of a Benedetto altarpiece now in the Sforza Castle. I see nothing Ferrarese about it in the reproductions, or similar to the 6 added cards ( La Spezia Museo Civico's Benedetto Bembo Madonna of Humility is more believably in the style of the Belfiore Muses, but it is not known when it was painted (,_1450-1455,_La_Spezia,_Museo_Civico_Amedeo_Lia.jpg). From reproductions, I see no point of resemblance in style to the six second artist cards.

The art historians are unanimous in attributing the second artist cards to the Cremonese painter Antonio Cigognara, who became active around 1480 (Bandera and Tanzi pp. 50, 52). The problem for the rest of us is finding work by him with which to compare the cards. I myself have only found one, in the Museo Civico in Cremona; there is indeed a stylistic similarity. I apologize for the blurriness; I did not have a tripod.


Dummett notes a possible third early deck from Milan, in a footnote on p. 51:
11. S.R. Kaplan riproduce nella sua Encyclopedia of Tarot, Vol. II, New York, 1986, p. 24, un’incisione di una Dama di Denari da Bertolo Belotti, La vita di Bartolomeo Colleoni, Bergamo, 1923. L’incisione rappresenta ovviamente una carta da gioco dipinta a mano e dorata, e corrisponde aìl’incirca alla carta Visconti di Modrone, ma non con esattezza completa; in particolare, la moneta della carta Visconti di Modrone porta l’emblema del sole raggiante, mentre quella dell’incisione porta il biscione visconteo. Il Belotti non cita la provenienza dell’incisione, e le indagini di Kaplan non l’hanno scoperta.

(11. S.R. Kaplan shows, in his Encyclopedia of Tarot, Vol. II, New York, 1986, p. 24, an engraving of a Queen of Coins, from Bertolo Belotti, Vita di Bartolomeo Colleoni, Bergamo, 1923. The engraving is obviously of a hand-painted and gilded playing card, and corresponds to a card in the ambit of the Visconti di Modrone, but not with complete accuracy; in particular, the coin on the Visconti di Modrone card bears a radiant sun emblem, while the engraving bears the Visconti viper. Belotti does not mention the origin of the engraving, and Kaplan’s investigation has not uncovered it.)
This card is also an example of how Dummett continues to use the presence of a Queen to conclude that the deck it comes from is probably a tarot.


There is a whole series of later hand-painted decks that seem influenced by the PMB. Dummett dates them to 1475-1510. The lower date is due to the fact that many of the decks copy the second artist designs. Given that he revised his dating of the second artist cards to c. 1462, he probably would have revised this lower date as well. 

In response to a query on THF, I made a comparison of Dummett 1993 (Il Mondo e L'Angelo) and Kaplan 1986 (vol. 2) regarding the various surviving decks of Milanese/Lombard hand-painted cards. Dummett lists 21 decks for Milan, of which two, as determined by analysis of the paint, are late 19th or sometime in the 20th century, but still of interest because the designs they copy are otherwise unknown. Kaplan has 15 decks, including all but two of the others in the category "other hand-painted cards". Of the two Kaplan does not mention, one is a group of two 20th century forged cards (number 21), the other (number 15) a group of five PMB-type cards that surfaced at an exhibition at the Estense Castle in 1987 Ferrara (catalog edited by Berti and Vitali). Here are the details, with some explanation of Dummett's reasoning for including them as he has them, and Kaplan when it is different.

I have left out many details of Dummett's presentation; hopefully I have not misread Dummett's relatively uncomplicated Italian.ImageDummett's book came out in 1993. It may be that there is more information since then. New cards surface periodically, although I can't think of any from Milan.

Dummett's numbers 1, 2, and 3 are the Cary-Yale, Brera-Brambilla, and PMB, in Kaplan vol. 2 as numbers 1,3, and 2 respectively. These are of course clearly 15th century and clearly Milanese.

Dummett's number 4 is what he calls the "Tozzi" pack, 13 cards, both triumphs and suit cards. It is the same as Kaplan's number 6, "Von Bartsch".
Dummett's number 5 is a group at the Fournier Museum, 5 out of the 6 cards held there, including only one triumph, an Emperor; it looks very much like the PMB Emperor (see Kaplan vol. 1 p. 103). It corresponds to Kaplan's number 7, who puts all 6 Fournier cards in his group.. Even though Kaplan does say that the Fournier Museum's Popess card is from a different set, because its back is red while the rest of the Fournier backs are black, he groups it with the others. Dummett calls this an error on Kaplan's part. Dummett puts this Popess in his group 7; for his reasoning, see my discussion of that group. Dummett thinks that his numbers 4 and 5 are from the same pack. as the dimensions and style are similar, both have black backs, and there is no duplication of cards.
(How do we know that group 5 is from a tarot and not an Imperator deck? Answer: Even without assuming it is of the same deck as Dummett's 4, the "a bon droit" on the 2 of Coins indicates a date after 1450, when Imperator decks cease to be recorded; the earlier cards all have "a bon droyt".)

Dummett's number 6 is a set at the Victoria and Albert Museum, 4 cards, Kaplan's number 8 (at left). The Page of Coins is very similar to that of the PMB. The Star card, oddly, has a crowned figure, unlike the PMB's.

Dummett's group 7, 2 cards, is the Fournier Popess (who looks very much like the PMB Popess except that her habit is much darker), plus the Page of Batons of Kaplan/Marzoli. This corresponds to Kaplan's number 13, 1 card. Both are the same size, style, and have red backs, whereas the rest of the Fournier have black backs.

(So we see here one way of associating a suit card with a tarot deck: if it looks like it belongs with a triumph in some other collection.)

Dummett's number 8 is what Kaplan calls the "Benomi", his number 10, 4 cards, including 2 triumphs.

Dummett's number 9 is that of the Pages of Swords and Coins of the Niedersachsisches Landesmuseum of Hannover. It is not necessarily a tarot pack. In vol. 1 p. 108, in the chapter on "other early hand-painted cards", Kaplan shows pictures of these two cards and says that they are of the Kestner Museum of Hannover, which is perhaps the same place.

Dummett's number 10 is Warsaw/Patocki, a Queen of Cups and a Knight of Coins. Dummett p. 60 says that the presence of a Queen suggests the probability that it is from a tarot deck. Kaplan does not consider these as "Visconti-Sforza"; he lists them under "other early hand-painted cards", vol. 1 p. 109.

Dummett's number 11 is two of the 4 Guildhall cards, the pair that is wider than the other two. One (below right) is an Ace of Cups similar to the Victoria and Albert Ace of Cups (the lower left in group 6, above). The other is either an Ace of Swords or a Sun (below left). These two cards correspond in Kaplan to two of three Guildhall cards listed under "other early handed painted cards", p. 111. 
(Why Kaplan considers one Ace of Cups as "Visconti-Sforza" and the other as "other" he does not explain; they look very similar. I assume it has to do with the decks he groups them with. )

Dummett's group 12 is the other two Guildhall cards, narrower in width than the first two (66 vs. 72 mm.). One (at left) is a "very true" copy of the PMB World card, the other (above middle) would appear, from the club in back of him, to be a Page of Batons. Kaplan groups the World card as from a "Visconti-Sforza" deck, Kaplan's number 11, and the Page of Batons with the cards of Dummett's group 11. Dummett says that Kaplan is mistaken in his grouping, owing to the differences in widths. (This difference is indeed apparent on the cards in Kaplan's pictures, vol. 1 p. 111, of the three he calls "other" as opposed to "Visconti-Sforza". He does not appear to have noticed this fact. For a good view of all four cards in color, see, near the beginning.)

Dummett's number 13 is the Andreolotti card, a Page of Coins. It is classified as a tarot because it is a copy of the corresponding PMB card. Its dimensions are within 1 mm. of those of group 12, Dummett says. This is Kaplan's group 12.

Dummett's number 14 is Biedack/Kaplan King of Cups, similar to the "Tozzi" king of Cups (group 4). This is Kaplan's group 14.

Dummett's number 15 are the Cocchi cards of the exhibition at the Estense Castle: Love, Sun, Page and 5 of Coins, and the 2 of Cups, all copies of the corresponding PMB cards. This group is not listed by Kaplan, probably because it wasn't publicized anywhere until Berti and Vitali's exhibition in Ferrara of 1987.

Dummett's number 16 is four cards of an anonymous collection, 2 triumph and 2 suit cards. They correspond to Kaplan's number 9, "Lombardy II". Three are poor copies of the PMB, plus a Death that is similar to the Victoria & Albert (group 6).

Dummett's number 17 is from another anonymous collection, a copy of the PMB Page of Coins. It corresponds to Kaplan's number 15, his "Lombardy III".
Dummett's number 18 (at left) is the nine Goldschmidt cards, which Kaplan classifies under "other hand-painted cards". One is an Ace of Cups similar to the Guildhall (group 11) and Victoria Albert (group 6), another is an Ace of Swords similar to the Guildhall Ace of Swords/Sun. Another is a Sun similar to that same Guildhall card. The others are harder to identify.

One might wonder if it was really a tarot deck, as opposed to a regular deck with odd figure cards. But the dimensions, some of the backgrounds, and a stave of the Goldschmidts correspond to that on the narrower Guildhall card (Dummet's number 12). Since the other Guildhall is a World, it follows that the Goldschmidt is also from a tarot deck, he says. Moreover, the iconography of some of the other Goldschmidts is similar to that of other decks related to the PMB. The man with the dog (upper right above) is similar to the Guildhall Falconer, as are the five clubs to that behind that Guildhall card. Other cards vaguely suggest other  themes: the third card in the top row resembles either the CY Hope card (related also in that the anchor is a conventional symbol of hope) or the PMB Star card. Likewise the lady at the kneeler is devout enough to suggest the Popess, and the crowned lady with a castle might be a variation on the Empress. In fact, Dummett goes so far as to suggest, due to the closely corresponding card dimensions, p. 72:
È pertanto plausibile congetturare che le carte Goldschmidt (18), la coppia Guildhall più stretta (12) e forse anche il Fante di Denari Andreoletti (13) provengano tutte da uno [72]stesso mazzo di tarocchi fortemente atipico 31.
31. R. Decker, nell’articolo ‘Early Tarots: Copies and Counterparts’, Journal of the Playing-Card Society, Voi. IX, 1980, pp. 24-31, fece indipendentemente la stessa congettura.

(It is therefore plausible to conjecture that the Goldschmidt cards (18), the narrower Guildhall pair (12) and perhaps also the Andreoletti Jack of Coins (13) all come from one and the same strongly atypical tarot deck 31.
31. R. Decker, in the article 'Early Tarots: Copies and Counterparts', Journal of the Playing-Card Society, Vol. IX, 1980, pp. 24-31, independently made the same conjecture.)
I would also think that group 11 is strengthened as a tarot. However a key part of Dummett's reasoning is that group 12's World card really belongs with the others in that group. It is a lot to pin on dimensions alone.

Dummett's number 19 is the 23 cards that Kaplan calls "Lombardy I" and constitute his group 4. They are all copies of the corresponding PMB cards, in fact exactly to the figures and triumphs of the group now owned by the Morgan Library. Dummett says that they may safely be considered 19th of 20th century copies of those cards, possibly copied in the period immediately before the Morgan Library purchased the originals in 1911, at a time when they had been out of the hands of their former owner since 1903.

Dummett's number 20 are the 23 cards that Kaplan calls "Rosenthal", his number 5, known only from photographs. Kaplan says they were in the possession of a collector named "Rosenthal" at least until 1939, when they were shown to another collector who declined to purchase them because of doubts of their authenticity. Dummett says (p. 67) that Albi Rosenthal, an Oxford antiquarian, reports that his father in the 1920s sold some hand painted Italian cards to a Swiss collector named Hardt, but has no idea where the cards are. Also, he had been shown some other cards in his office that were clearly false. The photographs (Kaplan vol. 1 p. 99) are nonetheless of interest; some of the cards resemble PMB cards and others more the Goldschmidt. Guildhall, Fournier, and Stozzi, but "are clearly different from the ones we have described so far", Dummett says. The Ace of Coins has a Cardinal's portrait in a circle on the card, quite unlike anything in any other deck. Dummett speculates that this was the forger's attempt to pretend that the deck is one recounted by the art historian (as opposed to the painter) Cigognara, since lost, as painted for Cardinal Ansconio Sforza. However the chronicle that Cicognara cited in evidence has no record of any tarot painted for the Cardinal or any reference to the alleged painter, Antonio Cicognara. (These issues are also discussed in The Game of Tarot, end of chapter 4.)

Dummett's number 21 is two cards at the Leinfelden Spielkarten-Museum, one a Falconer, another the Queen of Cups. They are "sicuramente dei falsi", definite forgeries. The Doemer Institute of Monaco showed the chemistry of the paint to be fairly recent. The Falconer resembles that of the Falconer of group 20. Assuming that the card is not a forger's invention, he speculates that the Falconer may correspond to the Magician. That is actually what Detlef Hoffman said of the Goldschmidt card (The Playing Card, 1972 p. 18):
The man with the little dog, bag on his belt and falcon on his hand, is certainly the sorcerer, the bateleur, the first card of tarots.
To me the Falconer fits in with the Page of Batons of Dummett's group 11. Perhaps both are variations on a "falconer" deck of normal cards.

So there you have it, Dummett's 21 vs. Kaplan's 15. The main difference is Kaplan's distinction between "Visconti-Sforza" vs. "other early hand-painted cards".
In "other" we have:

Dummett's group 9, classified under "other" as the Kestner cards.
Dummett's group 10, classified under "other" as the Warsaw cards.
Dummett's group 11, the Guildhall cards Kaplan classifies under "other".
Dummett's group 18, the Goldschmidt cards.
Also Dummett has a couple of groups not in Kaplan at all:
Dummett's group 15, the Cocchi cards.
Dummett's group 21, the two Leinfelden forgeries of non-standard cards.

And Kaplan has one group not in Dummett, namely the Fournier Popess card; even though he groups it with the other Fournier cards, it is for him from a different pack (or "set", as he calls it).

This discussion exemplifies how Dummett and others deal with suit cards that have been separated from the rest of their deck. They are joined with other cards most clearly by the dimensions of the cards and the color of the backs, but also in part by style, and in one case the presence of a Queen. Queens were often not part of regular decks, he says, although it was not unknown.

Whether there are indeed two Lombard groups, Visconti-Sforza (modeled on the PMB) vs. non-standard, is another question. As you can see, there is definite overlap, if the two Guildhall narrower cards really are from the same deck..

An interesting question is: how many PMB clones were there, i.e. tarot decks that closely resemble the PMB, not counting the PMB itself and done in the 15th or 16th centuries? If we start with Kaplan's number, 15, subtract the first 3 (CY, BB, PMB). then the Lombardy I (Dummett's group 19), and the Rosenthal (Dummett's 20), the first clearly a modern copy, the other probably), we are at 10. Kaplan's 6 (his Von Barsch; Dummett's 4, his Stozzi) and 7 (Fournier minus Popess; Dummett's 5) are probably from the same deck, so 9. But we have to add Dummett's number 15, the Cocci, not known by Kaplan in 1986. So probably 10 "PMB clones", of which 4 are represented by only one card, a court.

10 will be an underestimate if it turns out that the backs of Dummett's numbers 4 and 5 do not match (4, unfortunately, is inaccessible and known only by a reliable but incomplete description). Another possibility is that three of the groups are actually from the same deck. Here again is Dummett pp. 71-72 (never mind his reasoning):
It is therefore plausible to conjecture that the Goldschmidt cards (18), the narrower Guildhall pair (12) and perhaps also the Andreoletti Jack of Coins (13) all come from one and the same strongly atypical tarot deck (31).
Since 12 and 13 are both "Visconti-Sforza" in Kaplan's sense, that would reduce the number of decks by one again, down to 9. It seems to me that one way of testing this "conjecture" would be to turn the cards over and look at the backs, as none of Dummett's 18, 12, and 13 are listed as inaccessible. If that was done, neither Kaplan nor Dummett has recorded the result that I can find. Perhaps I have missed something, or it is in somebody's article somewhere.

To have made such an observation is important, because there is a more serious consequence. The Goldschmidt nine cards reflect a paradigmatic "other early hand-painted deck". But if groups 18, 12, and 13 are all from the same deck, Kaplan's distinction between the two types, "Visconti-Sforza" vs. "other Milanese hand painted decks" has utterly broken down, at least for decks that are like either the PMB or the Goldschmidt or both (or like the Cary-Yale but not the PMB, a possibility I will consider in a moment). That is probably why Dummett does not use Kaplan's distinction. So we are back to Dummett's 15 probably authentic groups after the 3 oldest decks are excluded. Since the 19th-20th century copies have really distinctive cards, it is possible they are copies from some deck since lost. So there are possibly 2 more (still not counting Dummett's "Lombardy I', which are most likely modern copies of PMB cards, unless there has been an astounding coincidence).

I question Dummett on his assertion that all the later decks do stem from the PMB and not either of the other two oldest decks. He says (p. 54):
Questi tre mazzi più antichi non esauriscono le carte dipinte a mano per la corte di Milano o almeno ad essa collegate. Molte delle altre hanno in comune un tratto interessante: alcune o tutte le carte sono copie, a volte esatte, a volte con volute variazioni, delle carte corrispondenti del mazzo Visconti-Sforza — comprese, a volte, le carte del secondo artista. Era sempre questo mazzo ad essere copiato, mai gli altri due, probabilmente dipinti per Filippo Maria. A causa di questa consuetudine, si può affermare che il mazzo Visconti-Sforza ha costituito un modello standard per i tarocchi dipinti a mano di origine milanese.

(These three oldest packs do not cover all the cards painted by hand for the court of Milan, or at least related to it. Many of the others have an interesting trait in common: some or all of the cards are copies, sometimes accurate, sometimes with unintended variations, of matching cards in the Visconti-Sforza deck - including, at times, the cards of the second artist. It was always this deck to be copied, never the other two, probably painted for Filippo Maria. Because of this custom, we can say that the Visconti-Sforza deck formed a standard model for the hand-painted tarot cards originally from Milan.)
I have already mentioned that the anchor on one Goldschmidt card links it to the Cary-Yale Hope card (at right), which in fact has an anchor on the bottom of the card.
Also, the Warsaw Queen of Coins (below) has a second person on the card, a girl, on our right, and so is more likely derived from the Cary-Yale ( than the PMB (, which has no such girl. The only effect this issue has on how many "Visconti-Sforza" decks there are, is to reinforce the point that Kaplan's method of separating the Milanese early hand-painted decks into two piles is untenable.


Dummett has more to say about the Goldschmidt cards, commenting on two articles in the 1980s by John Shephard ("The Goldschmidt Sun", The Ploying Card, Vol XVI, 1987, pp. 37-40; "The Lance and the Fountain: some Variant Forms of the World", The Playing Card, Vol XVII, 1988, pp. 54-7). He likes Shephard's idea that the subject of a card can be identified by what is shown on top. So for example when there are gravestones on the bottom and a sun at the top, the subject is the Sun. The idea also fits with the principle of "immediate recognition" for the players. He does not like Shephard's idea that in the Goldschmidt cards, all the trumphs have ceckered floors. That is because on one card, of a lady in prayer, there is no checkered floor and also no suit sign (p. 74 . Dummett can't imagine a hand-painted suit card without a suit-sign. Agaisnt Shephard here, Dummett makes the intersting proposal that the Goldschmidt cards might be the earliest instance of the suppression of the Pope and Popess cards. There is a bishop and a lady at prayer on a kneeler (images scanned from Shephard's article)
Dummett proposes that the Pope and the Popess have been downgraded, at the request of a pious family (p. 72).
Il vescovo [carta (c)] potrebbe sostituire il Papa e magari la dama all’inginocchiatoio [carta (e)] la Papessa; se così fosse, avremmo qui l’esempio più antico della frequentissima soppressione di quelle due carte molto spesso ritenute offensive: il mazzo Goldschmidt potrebbe essere stato prodotto per un ecclesiastico o per una famiglia nobile particolarmente devota. La dama all’ingi- nocchiatoio potrebbe essere un ritratto dal vero; questa ipotesi sembra particolarmente probabile per la dama con il castello in miniatura che, come avveniva per i ritratti di vescovi con raodellini di cattedrali, doveva essere la fondatrice di qualche famoso castello.

The Bishop [card (c)] could replace the Pope, and maybe the lady on the kneeler [card (e)] the Popess; if so, we would have here the oldest example of the frequent suppression of those two cards, very often deemed offensive: The Goldschmidt pack may have been produced for a clergyman or a particularly devout noble family. The lady at the kneeler could be a portrait from life; this hypothesis seems particularly likely for the lady with the miniature castle, who should be the founder of some famous castle. as was the case with portraits of bishops with reference to cathedrals.
Dummet's idea is that all these Milanese-based cards are painting the standard PMB subjects. In his view, the Goldschmidt cards are part of a deck to which two Guildhall cards also belong, with close to the same dimensions and same color of back. One of them has a hunter-type figure very similar to one of the Goldschmidt cards. The other is a very close copy of the PMB World card. So they would date somewhere around 1475-1500, or earlier if Dummett's 2007 suggestion of c. 1462 for the "second artist" cards is taken up.

 If, on the other hand, the World card is not part of the Goldschmidt deck, then the Goldschmidt cards could be considerably earlier than 1475.

Another issue regarding the Goldschmidt is the shape of the Baton suit-signs: gnarled cudgels in both the Guildhall and Goldschmidt, much like those of the "Spanish" suit style, which early on could have been either Spanish or French. He resists the idea that they are non-Italian, because the French only learned about tarot too late for these cards. However we have already seen the falsity of this idea, since Isabella of Lorraine knew about the game, and therefore, via her lady in waiting Agnes Sorel, the French royal family. Spanish suits could have been familiar by way of Naples, where Isabella had lived.

If the cards are indeed after the French incurion into Italy of 1494, there is also Shephard's proposal, which Dummett does not oppose, p. 73, that they were made for Cesare Borgia, of Spanish descent, with the Dolphin on one of the cards an allusion to the duchy in which he was a count, Delphinato (p. 73):
Shephard pensa anche che le figure umane dipinte nei tarocchi Goldschmidt siano quasi tutti ritratti dal vivo. A suo parere, il mazzo fu realizzato per Cesare Borgia nel 1500; ravvisa nel delfino incoronato un’allusione al ducato di Valentino nel Delfinato, e ritiene che la carta equivalga all’Asso di Denari, così come il biscione del gruppo Tozzi. Così si spiegherebbero, a suo avviso, i Bastoni di tipo spagnolo, poiché Cesare era, ben inteso, di lignaggio spagnolo. Questo tipo di Bastoni non era ancora molto diffuso in Spagna, ma sappiamo dal foglio di Barcellona che esisteva a quel tempo.

(Shephard also believes that the human figures painted on the Goldschmidt tarots are almost all portraits from life. In his opinion, the pack was created for Cesare Borgia in 1500; he sees in the crowned dolphin an allusion to the Duchy of Valentino in Delfinato, and believes that card is equivalent to the Ace of Coins, as well as the Tozzi group snake. This would explain, in his view, the Batons of the Spanish type, because Cesare was well known as of Spanish ancestry. This type of Batons was not still widespread in Spain, but we know from the folio in Barcelona that it existed at that time.)
On the other hand, this same dolphin might be an allusion to the Dauphin of France, who ruled, at least nominally, the whole of the Dauphiné. The stylized dolphin is exactly the usual impresa of the Dauphin.

It seems to me that these Goldschmidt cards deserve a closer look from that perspective.

I will start with  the lady at the kneeler, of which Shephard said (p. 4):
A Queen (non-trump) is shown kneeling in prayer, the colours of her robe perhaps acting as a suit-mark. A possibility here is Jeanne de France, Louis' first wife. She was very devout, and after the divorce she retired to Bourges where, with the permission of the Pope, she founded a new religious order.
The reason that Shephard thinks she is not a trump is that there is no checkered floor, unlike the cards that are clearly trumps. Here are four of the cards, with our lady on the far right; I get my scan of this row from Hoffman's The Playing Card, 1972.
Dummett mentioned another identification of the lady. After describing four other cards, which he designates a-d, he said (Il Mondo e l'Angelo, p. 66):
(e) L’ultima carta non presenta il pavimento a scacchi e raffigura una dama incoronata, genuflessa su un inginocchiatoio e assistita da una domestica; W.L. Schreiber la identifica come Dogaressa, non so in base a quali elementi.

((e) The last card does not show the checkered floor and depicts a crowned lady kneeling on a kneeler and assisted by a maid; W. L. Schreiber identifies her as a Doge's wife, I do not know on the basis of what elements).
So we may pass over this suggestion. Disagreeing with Shephard, Dummett found it too odd that there is no suit sign (p. 74). He had earlier argued that while a pair of cards now at Leinfelden have been securely established to be fakes, they may still be copies of originals now lost (p. 68). One is of a falconer and the other of a Queen of Cups. He now put that point to use (p. 74):
Se la Regina di Coppe a Leinfelden è una copia esatta di un prototipo autentico, la tesi che un pavimento a scacchi indichi sempre un trionfo sarebbe sbagliata; ma, in ogni caso, la tesi negativa che, nei tarocchi Goldschmidt, la mancanza di un tale pavimento indichi che la carta non sia un trionfo, è compietamente priva di fondamento. È improbabile che una Regina di qualsiasi seme sia priva del segno di quel seme; l’identificazione della carta (e) rimane un mistero.
(If the Queen of Cups in Leinfelden is an exact copy of an authentic prototype, the thesis indicates that a checkered floor is always a triumph would be wrong; but, in any case, the negative thesis in the Goldschmidt tarot, that the lack of such a floor indicates that the card is not a triumph, is completely without foundation. It is unlikely that the Queen of a suit would be devoid of any sign of that suit; the identification of card (e) remains a mystery.)
The Goldschmidt card is more likely a trump, he reasoned, and he proposed that she is the first known suppressed Popess, matching the Bishop as suppressed Pope (p. 73):
Il vescovo [carta (c)] potrebbe sostituire il Papa e magari la dama all’inginocchiatoio [carta (e)] la Papessa; se così fosse, avremmo qui l’esempio più antico della frequentissima soppressione di quelle due carte molto spesso ritenute offensive: il mazzo Goldschmidt  potrebbe essere stato prodotto per un ecclesiastico o per una famiglia nobile particolarmente devota. La dama all’inginocchiatoio potrebbe essere un ritratto dal vero; questa ipotesi sembra particolarmente probabile per la dama con il castello in miniatura che, come avveniva per i ritratti di vescovi con raodellini di cattedrali, doveva essere la fondatrice di qualche famoso castello.
(The Bishop [card (c)] could replace the Pope, and maybe the lady at the kneeler [card (e)] the Popess; if so, we would have here the oldest example of the frequent suppression of those two cards, very often deemed offensive: The Goldschmidt pack may have been produced for a clergyman or a particularly devout noble family. The lady at the kneeler could be a portrait from life; this hypothesis seems particularly likely for the lady with the miniature castle, who should be the founder of some famous castle. as was the case with portraits of bishops with reference to cathedrals.)
I do not wish to argue one way or the other on the issue of trump or non-trump. Instead, I propose another lady for the kneeler. In Manners, Customs and Dress During the Middle Ages and during the Renaissance Period,, the 1874 translation of Paul Lacroix's book Moeurs, usages et costumes au moyen age et a l'époque de la Renaissance (1871), I find a similar lady in fig. 427.

Lacroix said that the engraving was copied from a picture that was done around 1470 of Charlotte of Savoy, who was then Queen of France. The caption says,
Costume of Charlotte of Savoy, second Wife of Louis XI.--From a Picture of the Period formerly in the Castle of Bourbon-l'Archambault, M. de Quedeville's Collection, in Paris. The Arms of Louis XI and of Charlotte are painted behind the picture.
The original French, p. 382 at reads:
Fig. 427. Costume de Charlotte de Savoie, seconde femme de Louis XI. D'après un tableau contemporain provenant du château de Bourbon-l'Archambault, collection de M. de Quedeville, à Paris. (Les armoires de Louis XI et de Charlotte sont pientes derrière le tableau.
This type of subject, a stylish woman at a kneeler, is not one I have seen in art of that time otherwise.

A similar maid also appears on another card, that of a lady with a small castle, so the addition of a  maid is likely merely a feature of the deck. The castle on the card does not look like the one that Lacroix mentions (, nor are the ladies similar. I will get to the other lady later.

Charlotte seems more likely than Jeanne not only because of this picture, but because in the most famous paintings of Jeanne, she is presented standing and in a nun's habit, as she founded a religious order. I cannot find any of Jeanne at a kneeler.

Whether the card came first or the painting, there seems to me a connection, suggesting that in the court of Charlotte (who was neglected by her husband) and her son there was an appreciation of the tarot. Piedmont/Savoy, of course, was on the border of Lombardy and had long-standing relations with Milan. I would think that whoever did the second piece, card or painting, would likely have seen the first, or a sketch of it, because the facial resemblance is quite close. That the subjects are French is suggested by the fleur-de-lys I seem to see in the bishop's mitre. I will get to the bishop later.

Charlotte had married the then-Dauphin of France, the future Louis XI, in 1451; she was thus the Dauphine until 1461, when she became queen ( This is relevant because one of the Goldschmidt cards has the stylized dolphin that was the Dauphin's emblem. But the card is not likely of her in that early period, because by 1461 she was only 19; the card is of a mature woman. She died in 1483. Her son, the future Charles VIII, was Dauphin from his birth in 1470 until he  became king in 1483 (he later fought, coincidentally, at the Taro River). The year of the painting is the same as the year of Charles' birth.

So there are various possibilities for the card as Charlotte: 1461, looking older than her years; c. 1470, commemorating the birth of her son; and 1491, the year of the marriage of the new Dauphin, the future Charles VIII, to Anne of Brittany.

In the chapter on playing cards in his Arts of the Middle Ages and Renaissance, p. 165 of the English translation, Lacroix has an entry for a pack of cards for the Dauphin of France, 1454, i.e. the future Louis XI. The 1454 cards wouldn't have been the Goldschmidt, as the price recorded is too low, only 5 sous of Tours. But it shows that either the Dauphin, i.e. the future Louis XI, or his wife, Charlotte of Savoy, was interested in cards then. Savoy, of course, had close political and familial relations with Milan.

There is also a connection to the Dauphin's father, Charles VII. His mistress, starting in 1444, was Agnes Sorel, former lady in waiting to Isabella of Lorraine, whom we know from Marcello's letter was a player of triumphs at least by 1449 and probably by 1442 when she left Naples.

Detlef Hoffmann says that a pigment analysis done in the 1950s suggests "mid-15th century" for the cards ((p. 80, no. 19, of the English translation; I thank "Huck" for the reference to the German edition and Ross Caldwell for giving me the reference to the English translation, at and following).  If so, that would explain the Cary-Yale looking features, i.e. the attendant girl and the anchor, even if it doesn't fit the people who might be portrayed on some of the cards.


Shephard's theory about the lady with the castle was that she was Anne of Brittany, who in 1491, at age 14, became Charles VIII's wife. The lady on the card does not look that young, but Shephard put the deck later, c. 1500. The castle on the card does look somewhat like the castle at Nantes (, with its high central building flanked by two lower towers. Anne's importance was that with her, perhaps symbolized by the castle, Brittany could be made part of France. Making her a little older than she was on the card perhaps would be to reflect her importance and this outcome over time.

Nantes' relationship to Isabella of Lorraine's capital at Angers is that they are only 100 miles apart. Isabella would likely have visited the court at Nantes, where Anne grew up, or ladies from Angers visited Nantes. So Anne probably played triumphs, too.

Another possibility is Anne of France, Charles VIII's sister, who was regent during his minority. Her arrangement of the marriage between Charles and Anne of Brittany was aimed at making Brittany part of France. She would have been around 30 at the time of the marriage, 1491, about right for the card. She was never an actual queen, but as regent she was portrayed with a crown ( and ruled in such fashion

"Huck" suggested a date pf around 1461 and the castle where Louis, dauphin until March, was staying in Flanders; the town had a three-towered castle as its heraldic. That was also when a painter from Milan was sent to the region by Bianca Maria Visconti to learn from Flemish artists, who might have done the work. But that does not explain the castle lady or ladies. There were many such three-towered heraldics in Savoy-Piedmont, "Huck" says. But what is needed is a heraldic related however indirectly to Charlotte of Savoy. Savoy's main heraldic, of course, was the white cross on the red background. And that of course gives us another way in which the French royal family might have learned tarot, from Charlotte, who would have learned it from her aunt Maria of Savoy, who returned to Savoy after her husband Filippo Maria's death.


If the cards were given to Charles and Anne on the occasion of their marriage, then the Bishop could be the person officiating the marriage. I see a photo of an enactment of the marriage at with just such a bishop. But of course there were many bishops, for example the one that had married Charlotte and Louis in 1453, or some favorite bishop of Charlotte's. But he is definitely French, as indicated by the fleur-de-lys on his mitre.


Shephard had the Falconer card as a portrait of Cesare Borgia, created Count of Valentinois in 1498; Valentinois was part of Dauphiné.  But since Valentinois was such a  small part of the duchy, the dolphin is only marginally appropriate. Also, the Falconer does not look much like the portraits said to be of Cesare Borgia. The Falconer looks more like Charles or one of the  Louis, either the XI or the XII. Louis XII, born in 1462, would have been 29 in 1491. But the figure on the card  looks younger than that, and Louis XII was never a Dauphin. Louis XI would have had ample opportunity for falconing in Burgundy in 1461 and before. But he was 38 in 1461. He is "Huck's" candidate, however. Charles VIII was Dauphin until age 13. The coat of arms on the Falconer might help in identifying the Falconer, but I don't know it.

That the Batons on one of the other cards are in the Spanish style does not especially point to Cesare, who was of Spanish descent, as this style was also known in France. Dummett argued this point on p. 35f of Il Mondo e L'Angelo: the game of Aluette is played with Spanish suits, a game traced back at least to 1502 and still thriving on the west coast of France. There can be no other reason for this practice, he said, except the traditionalism of the players, who declined to switch to French suits.

I end up with the likeliest date, based on the castle lady, of 1491 or a little later or earlier, perhaps as early as the death of Charlotte in 1483. The "castle lady" would be the regent, Marie of France, and behind her Anne of Brittany. Based on the "paper research", the date of the deck would be earlier, so maybe 1470 or conceivably 1461, although Charlotte would have been younger than depicted on the card and then I have no idea who the castle ladies would be. I cling tenaciously to my identification of the lady at the kneeler as Charlotte.