Ch. 6 & 7: the game & its trump-order, 145-159, 98-9, 174-9

Dummett's next chapter, number 5, is on Divination and Occult Meanings in the early history of the tarot. It is mostly negative, in the sense of his not finding any evidence of divination or clues as to occult meanings.. I am going to skip that chapter for now, for two reasons. First, I have already been dipping into chapters 6-10 in evaluation of claims in chapters 1-4. I might as well have done with those chapters. Second, it is frequently advanced that occult meanings and divination explain nothing that the tarot's use in games cannot explain as well or better. To evaluate that claim, it is necessary to see what the use in games does explain and to see if anything is left. For that we need to look at chapters 6-10 at least, and probably more.

CHAPTER SIX: THE GAME OF TRIUMPHS AND RELATED GAMES

Dummett devotes chapter 6 to a reconstruction of the basic rules of the game of tarot. First he discusses bidding, a practice that he says was introduced in the 18th century (p. 145) from the Spanish game of Ombre and applies to partnership games, which can be for as few as 3 players, one playing against the other 2. Then he moves to the nature of trick-taking, which of course was part of the game from the beginning. He observes that the direction of play is another indication of its Italian origin (p. 145)

Nei giochi di Tarocchi l’ordine ciclico è quasi sempre antiorario, elemento che costituisce un ulteriore indizio della loro origine italiana poiché, quando i giochi di carte furono introdotti, l’ordine antiorario prevalse in Italia e nella Penisola Iberica, l’ordine orario negli altri paesi.

(Tarot games are almost always in a counter-clockwise cyclic order, an element that is a further indication of their origin as Italian, when card games were introduced, the counterclockwise order prevailed in Italy and the Iberian Peninsula, the clockwise order in other countries.)

Each player contributes a card from the suit led, or if he has none, then a triumph. The trick is won by the player with the highest card. For the suits, the King was the highest card, followed in sequence by the other courts. Then in Swords and Batons, the order went from 10 to Ace, but in Cups and Coins, it went from Ace high to 10 low. "This picturesque feature was abandoned only in Sicily and France," he adds. This feature is of historical interest. In other games, using the normal pack, the Ace was sometmes promoted as the high card among the numerals. That is not hard to understand as a variation on the usual order, but the complete reversal of the order is "more bizarre". It also occurs in Ombre, which suggests that it precedes the 16th century (p. 147). John of Rheinfelden says that "some suit signs are considered good and others evil, which seems like an allusion to the practice. It is also a feature of Marziano's game (p. 148):

Nel gioco di Marziano, nel quale si usava il mazzo non standard dipinto da Michelino da Besozzo per Filippo Maria Visconti, le carte numerali erano gerarchizzate in ordine discendente, dal 10 all’Asso, nei semi di aquile e tortore, ma nell’ordine inverso, dall’Asso al 10, nei semi di fenici e colombi.

(In Marziano's game, which used a non-standard pack painted by Michelino da Besozzo for Filippo Maria Visconti, the pip cards were hierarchical in descending order, from 10 to Ace, in the suits of Eagles and Turtledoves, but in the reverse order, from Ace to 10, in the suits of Phoenices and Doves.)

Dummett notes that in India the games of Ganjifa also had this unusual practice. So no doubt the Europeans learned it from the Muslims, and tarot simply kept it.

I would note also that the heraldics in the Cary-Yale are consistent with this practice, Swords and Batons having Sforza heraldics, and Cups and Coins having Visconti. Presumably the more severe suits (Virtues, Virginities: Swords, Batons) go from 10 to Ace, the more pleasurable (Riches, Pleasures: Cups and Coins) from Ace to 10, as though severities were of higher value than pleasures.

France did not retain the practice, but I would observe that even there we see black suits, Piques and Trefles, and red ones, Coeurs and Carreaux.

In tarot the object is to score points, not simply win tricks. I have already mentioned the difference between Milan and Bologna, that in Bologna the Angel got points (as well as the three cards in Milan), there were points for the last trick, and also points for combinations. Unlike trumps, which Dummett thinks was an innovation of the tarot (disregarding the case of Emperors, which I discussed in the last post), the idea of cards with different point values was not original with tarot, but was present in many games by the early 15th century (p. 151f:

Il gioco dei Tarocchi non [end of 151] fu certamente il più antico gioco a prese europeo. Certamente esso introdusse un’idea completamente nuova — quella del Matto. Se supponiamo che abbia anche introdotto l’intero concetto di briscola, è più che sufficiente: è estremamente improbabile che ne abbia anche introdotto un terzo, quello di carte con punteggio diverso. Quest’idea si ritrova in moltissimi giochi di carte europei, inclusi alcuni che non hanno alcun possibile rapporto diretto con i Tarocchi, come il gioco spagnolo di Malilla (noto in Francia come Manille e nei Paesi Bassi come Manilla). È probabile che fosse già diffusa all’inizio del XV secolo e che i Tarocchi se ne siano appropriati derivandola da forme esistenti di giochi a prese complessi senza briscole

(The game of Tarot [end of 151] was certainly not the oldest European trick-taking game. Certainly it introduced a completely new idea - that of the Fool. To assume that it also introduced the whole concept of Trumps is more than enough: it is extremely unlikely that it also introduced a third, that of cards with different scores. This idea is found in many card games in Europe, including some that can have no direct relationship with Tarot, like the Spanish game of Malilla (known in France as Manille and the Netherlands as Manilla). It is likely that it was already widespread at the beginning of the fifteenth century, and that the Tarot appropriated it, deriving it from existing forms of complex trick-taking games without trumps.)

By a "complex trick-taking game" Dummett means any such game where the score is not simply a matter of the number of tricks won.

In the above quote, Dummett asserts that the Fool, like the idea of trumps, was an original contribution of the tarot. He does not find it in earlier card games or in Asian games. Since both trumps and wild cards were a feature of Chinese games, according to "Andy's Playing Cards" (and trumps a feature of Karnoffel), it might be wondered how he knows that they weren't a feature of Muslim games. Has he looked at the early rules of Muslim games? It might simply be lack of evidence one way or the other. The matter does not seem to me very important.

The Fool can be played at any time in lieu of another card, although it can never win any trick. At the end of the game the player who played it exchanges one of the cards won in tricks with the other player to get the Fool back (if he has none, then he is out of luck). In scoring, the Fool not only has points, the same as the Kings and other "tarots", but can substitute for any card missing to make combinations, in games that count combinations. Dummett assumes that this was an original property of the Fool, as opposed to being the lowest trump, but it really isn't known. We have discussed this matter in the "Dummett and Methodology" section.

Dummett describes other rules as well: how at the end of the deal, the cards left over go to the dealer, who discards them face down before play starts but gets to count them as his in the scoring. Also, points accrue at the beginning of play as well as the end, if a player so chooses. But then he has to show his cards to the other players. Dummett does not say when these rules came into effect or how they differed from place to place, so I will not go into detail. Since the first written rules are of 1637 France, it is not easy to say.

Other games did develop which had trump suits, and the word "triumph" came to be applied to the cards of the suit so designated (usually by turning over a card at the beginning of play). Dummett spends considerable time discussing whether the word "triumph" came originally from Tarot or from some other game. The problem is that the name also designated the game. Why would it do that if some other game also had cards named "triumphs"? Also, Moakley's hypothesis that the iconogaphic model for some of the cards of the game was somehow modeled on those of triumphal processions is an attractive one (p. 154), which would explain how the name "triumphs" got applied to the game. I would add here that while the idea that the imagery of such processions corresponded to that of some of the cards is not verified, it surely is that it fits the Trionfi of Petrarch plus the similar poem by Boccaccio.

Dummett hypothesizes that when the word "triumphs" came to be applied to cards in other games, and even to the games themselves, it was necessary to use another word for the original game. And thus "tarocchi" came into use. That hypothesis is universally accepted these days.

I would add that before "tarocchi" the word "minchiate" may have been used as as the name of a game using the tarot pack in Florence. The word "tarocchi" has a similar meaning to words of the "minchiate" family of words, that of Fool. Later, an expanded version of tarot in Florence came to be called "Minchiate". It was also called "germini", after the highest numbered trump, the Gemini. There were also "Ganellino", in Liguria, and "Gallerini", in Sicily, for the same game, from a name for the first triumph, i.e. the tarot's Bagatto (http://en.wikipedia.org/wiki/Minchiate). If the game of tarot was still unique in having a wild card, then "the game with the fool" would have been an appropriate designation. Naturally, those who thought playing tarot was evil could also use this term to their advantage, calling it the game of fools (although such people tended to consider all card games in this light). This thesis, as well as the other (a term of abuse), is apparent in the course of several essays by Andrea Vitali. But Dummett does not speculate on why the name "tarocchi" became the preferred name, or where the word comes from.

Of special interest here is the first French use of the word "triumphe" for a card game, in 1482. Dummett discusses whether it was a tarot game or a game with the normal deck (pp.158-160; for the sake of ease in reading, I have included the footnotes only in the Italian)

Il gioco francese di Triomphe o Triumphe è di particolare interesse perché è l’unico per il quale esistano tracce anteriori al XVI secolo. Il primo riferimento compare in una lettre de rémission (grazia del re per un reato commesso) datata 1482 8; essa descrive una lite scoppiata nella città di Béthune, nei pressi di Lille, a causa di una partita di triumphe fra quattro membri dell’aristocrazia, uno dei quali aveva ucciso un altro durante l’alterco 9. Ora, per quanto ne sappiamo, la parola «tarocchi» non era ancora stata introdotta in Italia. Il più antico uso documentato risale, come abbiamo visto, al 1516; ma la versione francese, nella forma taraux, compare fin dal 1507 10 e possiamo quindi presupporre che fosse già nota in Italia a quell’epoca.
______________________
8. Archives Nationales, Parigi, serie JJ [?], libro 206, lettera 828, folio 181 recto. Alla lettera si fa riferimento in F. Godefroy, Dictionnaire de L'ancienne langue française, Vol. X, Complement, Parigi, 1902, s.v. *triomphe’, e in Du Cange (sieur Charles du Fresne), Glossarium Mediae et Infimae Latinitatis, Vol. VI, Parigi, 1845, s.v. *triumphus'; ma Thierry Depaulis ha visto la lettera stessa e a lui sono debitore di dettagliate citazioni da essa.
9. «... se esbatirent a jotter aux quartes aujeu du triumple pour le vin»; «triumple» è ovviamente un errore per «triumphe». L’uccisore era Jehan de Tremons e la sua vittima Alixandre de Lassisse.
10. Cfr. H. Chobaut, Les Maitres-Cartiers d’Avignon du XVe siècle à la Revolution, Vaison-la-Romaine (Vaucluse), 1955, estratto dalle Mémoires de l’Académie de Vaucluse, année 1955, p. 25. Questa fonte verrà nuovamente discussa nel capitolo XIII. [end of 158]

D’altra parte, abbiamo trovato l’espressione «triumphos cum chartis» in uno statuto di Reggio Emilia del 1500 in un contesto in cui deve far riferimento al gioco dei Tarocchi, per analogia con ordinanze simili di Salò, Brescia e Bergamo risalenti rispettivamente al 14S8, 1489 e 1491. Resta quindi aperta la possibilità che il gioco citato nella lettre de rémission non fosse quello che verrà più tardi praticato con il nome di Triomphe e un mazzo normale, ma un gioco di Tarocchi. Non possiamo escludere completamente questa possibilità. È tuttavia molto probabile che il gioco dei Tarocchi fosse introdotto in Francia e vi acquistasse rapida popolarità in seguito alle guerre francesi in Italia, iniziate nel 1494, e all’occupazione francese di Milano dal 1499 al 1512 e dal 1515 al 1522. Inoltre, nella lista che, nel 1534, Rabelais fornisce dei giochi praticati da Gargantua, la triumphe e le taraut sono elencati separatamente. È vero che passa più di mezzo secolo fra la lettre de remission e il Gargantua di Rabelais, ma nel frattempo compaiono altri tre riferimenti al triumphe. Nei libri contabili della corte di Lorena, ci sono due voci che registrano denaro usato dal duca Renato II nel 1496 per giocare a le triumphe 11. Due anni dopo, nel primo dei due riferimenti scoperti da Thierry Depaulis, un documento dell’ufficio dell’arcidiacono di Parigi accenna a la triumphe come a un gioco proibito 12; il secondo è un sermone di Michel Menot del 1518 in cui la triomphe è citata con altri giochi di carte 13. La fonte del 1518 deve far riferimento a un gioco praticato con il mazzo normale, dal momento che a quel tempo era già in uso la parola «tarot» o «tarau». Si può pensare che, nel corso dei venti anni che separano il documento del 1498 dal sermone di Menot, l’uso della parola «triumphe» fosse cam-[end of 159]
______________________
11. Citati in H.-R. D’Allemagne, Les Cartes à jouer du XVe au XX' siècle, Vol. Il, Parigi, 1906, p. 212: «Au Roy, le 29 armi pour jcruer au triumphe à Vézelise deux francs». »Encore audit seigneur roy le V mai pour jouer audit triumphe à Vézelise deux florins d'or».
12. Citato in L. Pommeray, L’Officialité archidiaconale de Paris aux XVe et XVe siècle, Parigi, 1933, p. 422, da un documento latino negli Archives Nationales, Z 1°-21, folio 33 verso: la frase che ci interessa dice «lusisse ad ludum prohibitum videlicet a la triumphe a la carte». Notare che il genere del nome sembra essere cambiato fra il 1496 e il 1498.
13. Michel Menot, Sermons choisis, a cura di Joseph Neve, Parigi, 1924, secondo sermone di Quaresima: «... qui ludit at ludum chartarum, du glie, du flus, de la triomphe».

biato proprio come quello di «trionfi» in Italia più o meno nello stesso periodo. Tuttavia, le probabilità sono complessivamente a sfavore di quest’ipotesi. Sembra più probabile che in tutti questi passi, dal 1482 in poi, si faccia riferimento a uno stesso gioco, quello citato da Randle Cotgrave nel suo dizionario francese-inglese del 1611 come «triomphe» 14 e descritto con quel nome nella Maison académique del 1659 e con il nome di «French-Ruff» nel Compleat Gamester di Charles Cotton del 1674 15.

È vero che, nell’ipotizzare che il gioco di Triumphe a cui si fa riferimento nel 1482 fosse praticato con un normale mazzo, noi distruggiamo le nitide linee di demarcazione del quadro che altrimenti otterremmo: la parola «tarocchi» introdotta dopo il 1500 e prima del 1507; la parola «trionfi» usata esclusivamente per il gioco dei Tarocchi per tutto il XV secolo e, in seguito, assieme ai suoi affini, per giochi con briscole praticati con il mazzo normale. Si potrebbe osservare che esistono buone probabilità che uno statuto come quello entrato in vigore a Reggio Emilia nel 1500 facesse uso di un linguaggio antiquato; ma nel 1492 il cardinale Ippolito d’Este scriveva per ringraziare la madre per l’invio di «triumphi dorati»16 e la parola «trionfi» — e non «tarocchi» — è usata da Pier Antonio Viti nella sua Illustrazione del poema del Boiardo che deve risalire al decennio 1490-1500 17. E la precisione stessa del quadro, tuttavia, a renderlo sospetto. Quando vengono introdotte nuove parole, esse spesso coesistono, per qualche tempo, con le vecchie; pertanto, la parola «tarocchi» potrebbe essere più antica di quanto possiamo verificare con prove documentarie e molti potrebbero aver continuato a preferire ancora il termine più [end of 160]
______________________
14. Randle Cotgrave, A dictionarie of the French and English tongues, Londra, 1611.
15. Charles Cotton, The Compleat Gamester, Londra, 1674, pp. 121-2. Il resoconto nella Maison académique del 1659 è alle pp. 181-4, dove la versione principale è la variante nota alle volte come L’as qui pitie', in essa, l’Asso vale più del Re e può essere preso dal mazziere in cambio di un’altra carta se viene scoperto per fissare il seme di briscola.
16. Cfr. Giulio Bertoni, Poesie, leggende, costumanze del medio eoo, Modena, 1917, p. 218.
17. II titolo ‘I Tarocchi’ fu dato alla composizione del Boiardo da un curatore dell’Ottocento.

antico «trionfi». Il nostro problema immediato comunque non è questo, ma piuttosto se esistessero prima del XVI secolo giochi a briscole con mazzo normale. Se il Triumphe giocato nel 1482 è un gioco di questo tipo, la risposta è affermativa. Infatti, sarebbe una coincidenza straordinaria se il nome di questo gioco non avesse alcun legame con la parola italiana «trionfi»; almeno l’idea centrale del gioco deve essere di provenienza italiana e quindi giochi del genere dovevano essere praticati in Italia prima del 1480. Questo, di per sé, è perfettamente coerente con la nostra ipotesi generale. Verso, diciamo, il 1475, il gioco dei Tarocchi doveva esistere da almeno quarantacinque anni — tempo più che sufficiente perché a qualcuno venisse l’idea di adattare il concetto di briscola a un gioco con il mazzo normale. Forse per qualche tempo i due tipi di gioco coesistettero in Italia, entrambi conosciuti come «Trionfi»; se essi erano diffusi in città e regioni diverse, questo fatto non avrebbe causato troppa confusione, soprattutto perché ci sono ben pochi elementi per pensare che il gioco con il mazzo normale abbia mai raggiunto grande popolarità fra gli Italiani, nonostante gli entusiasmi che suscitava in Cardano.

(The French game of Triomphe or Triumphe is of particular interest because it is the only one for which there are traces prior to the sixteenth century. The first reference appears in a lettre de rémission (grace of the King for an offense committed) dated 1482 (8); it describes a fight that broke out in the town of Bethune, near Lille, due to a game of triumphe among four members of the aristocracy, one of whom killed another during the altercation (9). Now as far as I know, the word "tarocchi" had not yet been introduced in Italy. The earliest recorded use dates back, as we have seen, to 1516; but the French version, in the form taraux, appears as early as 1507 10, and we can therefore assume that it was already known in Italy at that time.[end of 158]

On the other hand, the expression "triumphos cum chartis” is found in a statute of Reggio Emilia of 1500 in a context in which reference must be made to the game of Tarot, by analogy with similar ordinances of Salò, Brescia and Bergamo dating back to 14S8, 1489 and 1491. The possibility therefore remains open that the game mentioned in the lettre de rémission was not one that will be later practiced with the name of Triomphe and a normal deck, but a game of Tarot. We can not completely exclude this possibility. However, it is very likely that the game of Tarot was introduced into France and there acquired rapid popularity following the French wars in Italy, starting in 1494, and the French occupation of Milan from 1499 to 1512 and from 1515 to 1522. Moreover, in a list that, in 1534, Rabelais gives some games practiced by Gargantua, triumphe and taraut are listed separately. It is true that there is more than half a century between the lettre de rémission and the Gargantua of Rabelais, but in the meantime three other references to triumphe appear. In the account books of the court of Lorraine, there are two entries that record money used by Duke René II in 1496 to play at le triumphe (11). Two years later, in the first of two references discovered by Thierry Depaulis, a document of the office of the Archdeacon of Paris mentions la triumphe as a prohibited game (12); the second is a sermon by Michel Menot of 1518 in which triomphe is cited with other card games (13). The 1518 source must refer to a game played with a standard deck, since at that time the word "tarot" or "tarau" was already in use. You might think that, in the course of the twenty years between the document of 1498 and the Menot sermon, the use of the word "triumphe" had changed [end of 159] just like that of "triumphs" in Italy at more or less the same time. However, the odds are in total disagreement with this hypothesis. It seems more likely that all these passages, from 1482 onwards, refer to the same game, the one cited by Randle Cotgrave in his French-English dictionary of 1611 as "triomphe” (14), described by that name in the Maison Académique of 1659 and with the name "French–Ruff” in the Compleat Gamester by Charles Cotton of 1674 (15).

It is true that, in assuming that the game Triumphe to which reference is made in 1482 was practiced with a regular deck, we destroy the sharp lines of demarcation in the picture that otherwise we would get: the word "tarocchi" introduced after 1500 and before 1507; the word "triumph" used exclusively for the game of Tarot throughout the fifteenth century and later, together with its cognates, for games with trumps played with the normal deck. It could be observed that there is a good chance that a statute such as that entered into force in Reggio Emilia in 1500 made use of an antiquated language; but in 1492 Cardinal Ippolito d'Este wrote to thank his mother for sending “golden Triumphs" 16 and the word "triumph "- and not "tarocchi" - is used by Pier Antonio Viti in his Illustrazione of the poem by Boiardo one must go back to the decade 1490-1500. And the accuracy of the picture, however, has to make it suspect. When new words are introduced, they often coexist with the old for some time; therefore, the word " tarot" might be more ancient than we can check with documentary evidence, and many might have still continued to prefer the older term, [end of 160] “trionfi”. Our immediate problem, however, is not this, but rather whether there existed before the sixteenth century games with trumps with a normal deck. If the Triumphe played in 1482 is a game of this type, the answer is yes. In fact, it would be an extraordinary coincidence if the name of this game did not have any link with the Italian word "triumph"; at least the central idea of the game is to be of Italian origin, and therefore such games had to be practiced in Italy before 1480. This, in itself, is perfectly consistent with our general hypothesis. In, say, 1475, the game of Tarot had to be at least forty years old - more than enough time for someone to come to the idea of adapting the concept of a trump to the game with the normal deck. Perhaps for some time, the two types of game coexisted in Italy, both known as "Triumphs'’ if they were common in cities and region, this fact would not cause too much confusion, especially since there are very few reasons for thinking that the game with the normal deck ever achieved great popularity among Italians, despite the enthusiasm it aroused in Cardano.

Well, in France there are two possibilities, it seems to me. So far nothing is proven one way or the other. The odds that the altercation among nobles in Lille involved a tarot deck, as opposed to a regular one, are considerably enhanced if one bears in mind Marcello's letter of 1449, which assumes that his reader, the Duchess of Anjou, knows the game, that in the years before 1461, the future Louis XI (for whom the Goldschmidt deck might have been done) was just across the border in Flanders, and that in around c. 1488 the Flemish artist Hieronymus Bosch will start using tarot-like images ("The Stone Operation" and "The Wayfarer" both around then, according to Laurinda Dixon in Bosch, and "The Conjurer" around 1496).

In this case, a game with a normal deck could be called "triumphs" without necessitating a name change for the other game, because in France the other game did not exist. It has been argued that another game using a normal deck with a trump suit was played in Italy, namely, Ronfa. And if so, the name change should have been earlier than it was. Dummett argues successfully that there is no grounds for that assertion about Ronfa; it may not even have been a trick-taking game, but merely one where different cards had different points in combinations, like Primiera.

Finally, Dummett considers the game of Karnoffel, also called Kaiserspiel, from which I gave lengthy quotes in my last post. He argues that this game, even if invented at the same time as the tarot, did not co-exist with the tarot in any widespread way anywhere, so that "truimphs" would not distinguish tarot from it. The problem is that if it did exist in Italy, or something like it with special cards, it would have been in exactly some areas where tarot developed, notably Florence and Ferrara, because that is where a game named "Imperatore" is mentioned. Dummett says that there is no evidence that this game of "Imperatore" had any relationship to the German game of "Kaiserspiel". And if tarot is a descendant of "Emperors", why are there no descendants?

The first point seems to me unclear, since the names and places of the trump cards in that game correlate well with those of the tarot. Also, the Italian courts had their authority from the north (the Emperor) and had many connections there.It seems to me that if the two games co-existed in Ferrara and elsewhere, there are two ways in which "triumphs" would fit the game later known as "tarocchi" better and more naturally than that called "Emperors". First, there are many more of these special cards than in "Emperors", in fact forming a fifth suit; and second, the iconograpy is suggestive of Petrarch and Boccaccio.

As to why there are no signs peculiar to "Karnoffel" in any Italian games, my answer is first, "Emperors" never was very widespread; and second, it got eclipsed very quickly by the game later known as "tarocchi". Perhaps it only held on where there were people, such as Borso d'Este, who had played it in their youth, and then it died out.

CHAPTER SEVEN: THE ORDER OF TRIUMPHS

Dummett's chapter 7 is on the order of the triumphs. Here he presents his well-known division into, first, three groups within the hierarchy--beginning, middle, and end--and second, his division into three types of orders, A, B, and C.

Again, it seems to me that he is putting the cart before the horse, logically speaking (although not rhetorically). He bases his argument here on what he will develop in the next six chapters or so, namely, a large number of lists of the order of trumps in different places that come from poems, prose accounts, and the cards themselves, when numbers are on them. However I will proceed as he does.

But first I need to go back to Dummett's chapter 4. It is not about the invention of the game and the deck, but rather about when the triumphs became standardized as to the subjects and their number, as he makes comments on that issue. This point is of course relevant to the division into sections and orders.

In his conjectured timetable, he puts 1444 as the time of standardization (p. 106):

1444: la composizione del mazzo di tarocchi diventa standardizzata dappertutto.

(1444: The composition of the tarot deck becomes standardized everywhere.)

As far as I can tell, Dummett picks 1444 because that is the approximate date of the Brera-Brambilla, which has the standard 14 cards per suit. If it follows the 3:2 principle of triumphs to cards per suit, it will have 21 triumphs. As I have said, that is a big "if"; not only are there different possible principles, including that of no principle, but that there was a state of war among some of the regions affected.

Elsewhere in the chapter he makes other statements. On p.98f, we read:

Il mazzo Visconti di Modrone fornisce una prova che il mazzo dei tarocchi subì una certa evoluzione, come era da attendersi. Quest’evoluzione deve aver toccato senza dubbio i soggetti dei trionfi, e forse anche il loro numero. Poiché la serie dei trionfi è estremamente incompleta in tutti i gruppi di carte da tarocchi dipinte a mano, a parte il mazzo Visconti-Sforza e i tarocchi ‘Carlo VI’, si possono avanzare ipotesi di vario tipo. E nondimeno probabile che, a partire dal 1450, fosse ormai fissa la composizione standard di un mazzo di tarocchi, per quanto riguarda sia il numero delle carte che i soggetti dipinti sui trionfi.

In ogni caso, quella che è di gran lunga la più dettagliata fonte quattrocentesca sui tarocchi, un sermone ‘De Ludo' contro il gioco d’azzardo, tratto da un volume manoscritto anonimo di sermoni, conferma che i soggetti dei trionfi erano già stati [end of 98] standardizzati negli ultimi due decenni del secolo 3. In questo sermone, il predicatore elenca tutti i soggetti normali dei trionfi, compreso il Matto. La maggior parte del sermone fu pubblicata dallo studioso inglese Robert Steele in un articolo del 1900 4. In esso data il volume fra il 1450 e il 1470. In un articolo dell’anno seguente, egli è più cauto nella datazione, suggerendo il periodo 1450-80 5. Ricerche più recenti di Ronald Decker suggeriscono una data più tarda per lo stesso volume, perché alcuni fogli hanno filigrane del 1500 circa. Ovviamente la scrittura del libro può essere stata di molti anni posteriore alla predica del sermone, che è perciò da datare fra il 1480 e il 1500.

(The Visconti di Madrone pack provides evidence that the tarot pack underwent a certain evolution, as was to be expected. This development undoubtedly must have affected the trump subjects, and perhaps even their number. Since the set of trumps is extremely incomplete in all groups of hand-painted tarot cards, apart from the Visconti-Sforza pack and 'Charles VI' tarot, one can advance hypotheses of various types. It is nevertheless likely that, beginning in 1450, it the standard composition of a tarot pack was now set, as regards both the number of cards that the subjects painted on the triumphs.

In any case, what is by far the most detailed source on the fifteenth-century tarot cards, a sermon 'De Ludo' against gambling, taken from an anonymous manuscript volume of sermons, confirm that the subjects of the trumps had already been standardized by the last two decades of the century (3). In this sermon, the preacher lists all the normal subjects of the trumps, including the Fool. Most of the sermon was published by the English scholar Robert Steele in an article of 1900 (4). In it, he dates the volume between 1450 and 1470. In an article the following year, he is more cautious in dating, suggesting the period 1450–80 (5). More recent research by Ronald Decker suggest a later date for the same volume, because some papers have watermarks circa 1500. Of course, the writing of the book may have been many years back to the preaching of the sermon, which is therefore to be dated between 1480 and 1500.)
____________________________
3. Property for a time of Robert Steele, the volume is currently preserved at the Museum of the United States Playing Card Company of Cincinnati, Ohio; it was already in the Museum of Art in the same city.
4. R. Steele, 'A Notice of the Ludus Triumphorum and some early Italian card games', Archaeologi, Vol. 57, 1900, pp. 185-200.
5. Id., 'Early playing cards, Their design and decoration', Journal of the Royal Society of Arts, Vol. 49, 1900-1, pp. 317-23.

Dummett's 1450 date here seems to be based on the estimated date for the PMB (i.e. Visconti-Sforza), from which 19 out of 21 triumphs survive, plus the Fool. Given that 6 of these triumphs are by a second artist in the style of around 1475, and that none of the Milan decks based on the PMB have either the Devil or the Tower cards, it seems to me that the assumption that the deck was standardized everywhere by then is unclear on that basis. But certainly by the time of the Boiardo trionfi poem (he died in 1494), which has 21 subjects plus the Fool, as well as the "Steele Sermon", we can say for sure that this was true in Ferrara, probably by the 1480s. And before that, by the time of the d'Este wedding in 1473, there was already the Sun and Star, and in Florence of around the same time the Moon and the Tower. Also, in a painting by the Bembos in Cremona, which the Denver Art Museum dates at 1455-1460 (although Longhi to 1462, per Kaplan vol. 2 p. 132), there is a detail on the upper right (http://www.denverpost.com/entertainment/ci_17029148), of the three magi very much like the two figures the Star card of the d'Este deck (or the three of the Rothschild sheet). So I think we can say that by 1455-1460 there were, in one place or another, at least 19 subjects and perhaps all. However it still might be that some subjects were included in some places and others in other places, so that in any given place there might be few than 21 in all. In any case, there were enough around that we can start theorizing about the order of the triumphs.

To get the three groups, Dummett performs two operations. First, he removes the virtues from consideration at this point.That is because they have the feature of being widely various among all the different lists that have come down to us (including the numbers on the cards).

He does not reflect on why this might be, as he is doing a purely formal operation. It seems to me that this shows that the virtues were less important in defining the sequence than the other cards. More important were the "triumphs" of Petrarch, Boccaccio, and perhaps Emperors, and then the virtues and other "triumphs".

In excluding the virtues, he does so on the basis of how they move around from one standard order to another in the different lists. Now he extends that principle to the other cards, resulting in the definition of three sections to the hierarchy. Here only one thing, it seems to me, is critical, namely, the definition of the middle group. If you have that clearly defined on purely formal grounds, the other two fall into place. Here is what he says about that middle group (p. 174):

Il segmento intermedio comprende cinque carte, il cui ordine tipico, dalla più bassa alla più alta è: l’Amore, il Carro, la Ruota, l’Eremita, l’Impiccato. In tutti i casi in cui l’ordine interno di queste carte è diverso, la differenza è il risultato di uno scambio di posizione fra una coppia di carte adiacenti: l’Amore e il Carro; il Carro e la Ruota; la Ruota e l’Eremita; oppure l’Eremita e l’Impiccato. Almeno due virtù, e qualche volta tutte e tre, saranno intercalate in questo segmento.

(The intermediate segment includes five cards, whose typical order, from lowest to highest is: Love, the Chariot, the Wheel, the Hermit, the Hanged Man. In all cases in which the internal order of the cards is different, the difference is the result of an exchange of position between a pair of adjacent cards: Love and the Chariot; The Chariot and the Wheel; The Wheel and the Hermit; or the Hermit and the Hanged Man. At least two virtues, and sometimes all three, will be interspersed in this segment.)

Since Love is the lowest card of the segment, that defines its lower boundary. Since the Hanged Man is the high card of that series, that defines the upper boundary. And since no other cards except the virtues are ever in between, we have a clear separation into groups. It might help here if we saw the various orders in their respective charts.

Southern of type A (Florence, Bologna), the Eastern of type B (Ferrara, Venice), and the Western of type C (Milan, France).
Type A, southern (Florence, Bologna): https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhOnSEWZKJHsXEES3yu5FtdUTPPD5V5j1P7ULWJ7jAHSp-rxX48ngrPPIPJksce2kLKxLB7CgikFCVKhYHX2CBJixmVStiim41KbwrCN9Tc5FH1qTTtkw1VcnLMe2Fy6ij-HVr2rpxyN3k/s600/Screen+Shot+2013-11-03+at+3.47.35+PM.png
Type B, Eastern (Ferrara, Venice): https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEj2VfUWhZDIlgTINCN6jc0TF_KdeDjZHa-bsXQy5lH3jtsVtLlRyG-AhxNx7jpuClc53EnAHbrO_tB977RStdyhI5yhhOH1u_oo7hFk3KallZXXeh-k9N5Rkvxr0xlR-axZcVDz2SpE7Wo/s1600/Screen+Shot+2013-11-03+at+3.47.56+PM.png
Type C, Western: https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhQp-glovdoYEeFpSYqP-G99gJ0RqrWvHcsDC0omgsfw39Mva6XMX4udQEOkh9pxq8gcUwhivhv3RHAK1gN0DfE3YKFyrKD0OT_6xr8fuzLueoDCLJbGOYhaYsLr56HsiEyWi7j_7QTACI/s1600/Screen+Shot+2013-11-03+at+3.48.16+PM.png
In addition, Marco Ponzi did another C chart adding a couple not in Dummett:
https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhH7y-xy4UFBoHJpGZPyDo5F2ZeTtQPTCZzfApa9qVDYB4VkbwAC3kn4I_BukZYQ85TG0THsaXiDq710oJM_tB1nI6SLjwdOJEk8ySzrfqXaQTXQZnWEJGd66DAEGArneBnq5keY8sCEdk/s800/westernorders2.jpg
Visually, A, B, and C::

In the case of the first group, cards 1-5, the interchange of cards (i.e. Popess being 2nd or 4th) does not define the boundaries. The Pope is never interchanged with any other card. So it could just as well have been put in the second group. Likewise, Death is never interchanged with any other card above it. So it could just as well be in the second group. In fact, in his 1986 FMR article, he did put Death in the middle group. But it is the fact of interchange within the second group that defines all of them. So the Pope is clearly in the first group and Death clearly in the third group. I had not appreciated this point until I read this chapter. Some followers of Dummett ignore this point and put not only Death but even the Devil in the middle group. But that is ignoring the principle upon which he divides the groups.

I have one criticism of this division into sections. If the basis is the variability of different triumphs, then there are actually four groups, because there is one long sequence that doesn't vary at all among the lists, namely, from Death through the Sun, followed by two that do. He says (p. 174):

Il segmento finale è formato da otto trionfi — la Morte, il Diavolo, la Torre, la Stella, la Luna, il Sole, l’Angelo e il Mondo. Trascurando ancora una volta il possibile inserimento di una delle virtù, queste carte si presentano esattamente nell’ordine indicato, con la sola, ma importantissima, eccezione che talvolta le posizioni del Mondo e dell’Angelo sono rovesciate e l’Angelo diventa così la carta più alta.

(The final segment consists of eight triumphs - Death, the Devil, the Tower, the Star, the Moon, the Sun, the Angel and the World. Neglecting once again the possible inclusion of one of the virtues, these cards appear exactly in the order listed, with the single but important exception that sometimes the positions in the World and the Angel are reversed and the Angel becomes the highest card.)

There are actually two parts here: an invariant one, Death through Sun, and a part where the triumphs switch with each other, the last two triumphs. I will explain the importance of this distinction at the end of this post.

Then we come to the orders. Putting the virtues back in, there are three ways it is done. In one, type A, they are all one right after the other in the middle section. In another, type B, they are separated, and Justice is next to Angel. And in type C, the virtues are separated by 3 other cards in between, and Temperance is put just above Death..

Finally, there is a "mixed A and C". This has the Angel above the World, as in type A, as well as four "papi" of the same rank, as in some of type A, but the virtues are arranged in order of type C. This is the order found in Piedmont, which evidently had influence from both A and C regions. This can be seen in Marco Ponzi's chart, https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEhH7y-xy4UFBoHJpGZPyDo5F2ZeTtQPTCZzfApa9qVDYB4VkbwAC3kn4I_BukZYQ85TG0THsaXiDq710oJM_tB1nI6SLjwdOJEk8ySzrfqXaQTXQZnWEJGd66DAEGArneBnq5keY8sCEdk/s800/westernorders2.jpg, under "Piscina":

There are other things that can be said about the three orders, but these are the criteria he uses in differentiating them.

Looking at the places associated with the various lists, he also notices a geographic designation that can be put in the three orders. A is in Bologna and Florence, and later also in Rome and Sicily, and thus "Southern". B is in Ferrara and probably Venice, and so "Eastern", and C is in Milan and France, and so "Western".

How are we to account for the differences among the three orders? The main thing is that there were no numbers on the cards. Producers or designers of decks in one region, early on, did not know, or perhaps chose to ignore, the order of triumphs in other areas. With no numbers, either they or the first players themselves were free to improvise.. And then the improvizatons stuck. Players in one locality only cared that there be uniformity within that locale and not other regions of which they may know nothing. Therefore, he says, referring to this second element (p. 179):

E questo elemento, più ancora delle differenze fra i modelli standard usati nelle diverse aree, a fornire la discriminante principale per distinguere tre diverse tradizioni di Tarocchi, la cui origine risale ai primi stadi dello sviluppo del gioco. Non siamo in grado di stabilire se i diversi ordini di trionfi furono adottati come deviazioni intenzionali dalla pratica dei giocatori di altre città, o semplicemente come conseguenza di un imperfetto ricordo di tale pratica; ma è evidente che almeno le caratteristiche principali di ciascuno dei vari ordini possono essere state fissate solo nel primo momento in cui il gioco fu introdotto nell’area che osserva quel dato ordine. Vedremo che l’ordine di tipo A rappresenta la pratica dei giocatori di Bologna, quello di tipo B la pratica dei giocatori di Ferrara e quello di tipo C la pratica dei giocatori di Milano.

(It is this element, even more than the differences between the standard models used in different areas, that provides the main discriminant to distinguish the three different traditions of the Tarot, whose origin dates back to the early developmental stages of the game. We are not able to determine whether the different orders of triumphs were adopted as intentional deviations from the practice of players to other cities, or simply as a result of an imperfect recollection of this practice; but it is evident that at least the main features of each of the various orders can only have been laid down the first time the game was introduced in the area that observes the given order. We will see that the order of type A is the practice of the players of Bologna, one of type B the practice of Ferrara players and type C the practice of the players of Milan.)

And with that, the chapter ends. It seems to me that any deviation in the order of the cards done after numbers had been put on the cards (i.e. after the early 16th century at the latest) would have had to be intentional, e.g. in the Tarot de Marseille and Sicilian cards.

My main area of doubt about this argument is that he has lumped into his various lists, resulting in around 21 different orders, cards of widely different historical periods: they go from Ferrara c. 1480 to France of c. 1650 and Sicily of the same period. These last two are definitely not from the earliest period, nor can they even be used to infer what was present there in the time before a relevant document survives, as the cards were clearly, acccording to his own research, imported there from outside at a late date. It seems to me that the time at which a particular deck was established is an important factor is knowing what to count as part of the "primitive" period. He dates this period as ending at 1480, the time of the "Steele Sermon". However it is not known when this sermon was preached, up to 1500. Also, even in 1521, two players had to have a discussion in order to agree on an order of trumps (according to a document found by Pratesi that I quoted in my previous section). Moreover, the distinction between two of the groups may have been drawn before the distinction between either of them and the third. Whether any more precision may be obtained is a matter of looking carefully at the dates and places of the lists, the titles used, other associations with the titles, the iconography of the cards, communication between regions, likely sponsors, and any other relevant considerations.

Another consideration is that while Dummett's explanation accounts for variability, there is also invariability to account for. In Europe north of the Alps, for example, entire countries had exactly the same tarot order, and even among different countries, despite the same lack of awareness between one region and another pertained. And even in Italy, there is an invariant section within the cards, Death through Sun, to be accounted for. Why there and not in the other sections? It seems to me that the answer is the same to both questions. The reason for the invariability among countries north of the Alps is that by then a definite order had been formed in one place that spread by contact everywhere else. That condition didn't exist earlier. People from north of the Alps weren't coming to a particular part of Italy in any great numbers. In the case of the Death-Sun section, I would say that the same condition existed, namely, lack of large-scale population movements among different regions. Only after the Treaty of Lodi in 1454 and the resulting Italian League was there a state of peace among city-states in Italy sufficient to allow that. Also, trade among regions in mass-produced decks could happen. In that situation, a sub-sequence developed in one place could come not only with new subjects but also in a definite order. In other regions, these would replace or add to whatever was already there. At the same time, there could be respect for the existing order of subjects where the subjects themselves had not changed. Variability is a product of unstable times when regions are in isolation from each other Standardization happens when masses of people go between regions formerly isolated, or there is easy trade among regions. That happened within Italy from 1454 to 1494. Then in 1494 the French invasions started, and an Italian invention as standardized in the particular place of invasion could move north.

This result, that 1454 would have been an important date for standardization, agrees with the 1455-1460 date I began this post with. That is not to say that there would not have been tarot packs with all the standard subjects before 1454-1460. It is just that they probably wouldn't have been a repeated phenomenon everywhere before that period. That the Devil was a card in Karnoffel/Kaiserspiel and the Tower on the "Charles VI" and Rosenwald speaks in favor of those cards being part of the standardization, which the PMB-style hand-painted decks perhaps chose not to include.

One group of cards affected by these considerationsare Devil through Sun. It is possible that cards went from simpler to complex. In that sense the Rosenwald, with just the celestial bodies, might be an example of their first version, and the hand-painted variations we see on the "Charles VI", d'Este, and PMB second artist as variations from those, which then got incorporated into the mass produced designs. The Beaux Arts/Rothschild sheet of Bologna is also more complex than the Rosenwald. The same would likely apply to the Devil and Tower. Considering the time between the Giusti note and a known 1450 Florentine edict prohibiting playing card production (indicating the the availability by then of inexpensive decks), some woodblock version in Florence might go back as far as 1444, in time to influence or be influenced by the Brera-Brambilla (although I don't know how). But what would have been on the woodblock sheets at that early date is another question.

Removing Devil-Sun from consideration, there are still cards that are more variable and cards that are more fixed. By Dummett's principle--giving up, to be sure, his principle that the special cards were always 22--the more variable the order, the earlier the card. The virtues are variable, hence early. The Popess is variable, hence early. Love, Chariot, Wheel, and Old Man are variable. Judgment and World are variable. So these are all early cards. Empress and Emperor are not variable, but we know from the Cary-Yale that they were there early on. Other cards are invariable: Fool, Bagatella, Pope, Hanged Man. The problem is that we don't know whether it is because they were added later or because they marked a division between sections that had somehow been transmitted among regions and kept in mind. All we know is that they were not in the earliest known tarots, the CY and BB. Also, the variabiity in the Popess may not be an indication that it was early, but rather that the B order itself is late, since that is the only place where she is put next to the Pope.