Wednesday, 3 June 2009

Date of Invention

I’ve been considering my plotting of the earliest evidence for tarot some more. Here are some conclusions.

The first consideration is the period between 1420 and 1460. All of the evidence is concentrated between 1442 and 1460. There is none between 1420 and 1442. Since, as far as I know, the total of all possible data is stable with regards to historical accident in the entire period 1420-1460 (i.e. it is not affected by a major change that guarantees a greater chance of survival, like the printing press) – that is, there was no greater loss, or gain, of writing or art, which has survived due to the vagaries of history, in the period 1442-1460 than in the period 1420-1442 – then the absence of evidence in the period 1420-1442 can be taken as evidence of absence.

In other words, if tarot cards and references to them had been made between 1420 and 1442, they are just as likely to have survived as those made between 1442 and 1460. There was no “great flood” which washed away all kinds of different potential sources prior to 1442, and no great invention which increased their chance of survival in the period after 1442. The data gives such a coherent picture that even if one whole data set were taken out – for example all the packs, or the iconography of players (all but one are uncertain) – or one whole city – for example, Ferrara – the overall pattern would remain undisturbed, and the same conclusion about diffusion and the inclusive range of absence of evidence would remain the same.

How long an absence? Can the absence be measured and the margin of error for the invention guessed at? I think so. From the half of this time that contains evidence, no datum is more than 3 years separated from another on the chart. For example, the earliest Ferrarese references, in 1442, are contemporaneous with the Brambilla and Cary Yale packs (1443-1445), and the Palazzo Borromeo fresco (circa 1445); at the other end, Bologna’s 1459 reference is contemporary to Ferrara’s 1457-1461 references, and Padua’s 1455 and c. 1460 reference. In a straight line therefore, taken as a whole, all data points from contemporary sources are within 3 to 5 years of one another in the first 18 years of tarot’s attested existence. Since the chance of preservation by historical accident remains constant in the whole 40 years 1420-1460, I think this observation of proximity gives us a control or statistical margin of error over the time prior to 1442 when we can expect tarot to have been invented – within 3 to 5 years of 1442, or between 1437/39 and 1442.

The chart also can be used to show that the place of the invention *cannot* be deduced solely by appeal to the chronological and geographical arrangement of the surviving data. The two closest towns on the chart are Bologna and Ferrara, with a mere 48 km between them. Yet the earliest references in these cities are separated by 17 years – the greatest distance in time between any first evidence in any two places in these 18 years. One of the longest distances between two towns is Ferrara and Milan (250 km), yet tarot appears practically simultaneously in both places. A final point to note is that tarot is recorded 9 years earlier in Florence than Bologna, when, any direct journey between Ferrara and Florence must pass through, or at least by, Bologna. Therefore the absence of any earlier record in Bologna must be an historical accident, and is not evidence of absence. Since the data cannot be relied upon for a period of at least 17 years, theoretically any place noted in the first 18 years after 1442 on this chart could be the birthplace of tarot.

In other words, if the chart is arranged by geographical proximity starting from any presumed diffusionary center, then the two closest towns are separated by the longest gap, of 17 years, while all the other more distant cities fill up the space with their own data, separated by no more than a few years each, which is counterintuitive for a diffusionary model (especially since the second earliest reference in Ferrara, also from 1442, is found in the hands of a Bolognese merchant, Marchione Burdochio.). This “margin of simultaneity” is therefore possible to within 250 km, which is essentially all of the cities noted in the first 18 years (definitely so if the distance is lengthened to 300 km). Thus it is impossible to use the chart to determine which is the most likely place of invention among Ferrara, Milan, Florence or Bologna.

However, historical and art-historical considerations make some places more likely than others, and other models (like a diffusionary model) might help make a decision easier. However, like the unlikelihood of deriving an accurate diffusionary model from the evidence, it is difficult, although perhaps easier, to trace iconographical developments and draw a genealogical tree. The great difficulty arises from the tendency of artists and patrons to invent, which means that trying to associate chronological and iconographical relationships among the earliest A, B and C families will still rest on no absolutely firm basis. In other words, if we limit ourselves to the fifteenth century, which includes documentary evidence as well as cards, it is still hard to make convincing arguments about some kind of relationship or evolution. The very earliest cards, like the Cary Yale, can be taken as the prototype of later standards - or an invented variation. The fragments as they are can be interpreted with equal parsimony in both ways.

I should note, finally, that this evidence is consistent with a scenario in which the 1440s were merely the time in which the wealthy became interested in tarot. The evidence of a popular game in this decade can be inferred from the Ferrarese note of the merchant Marchione Burdochio having a relatively cheap pack of carta da trionfi on hand to sell in July of 1442, as well as Marcello’s account of being given, probably in late 1448, a pack “not worthy” of being given to a Queen. But the range of the silence or negative evidence of this popular kind of tarot is harder to guess at than the evidence for a game played by the privileged. Personally I take Bernardino of Siena’s silence on these cards in his writings and sermons between 1424 and 1436, while mentioning and describing regular cards several times, to be evidence of absence; but as a solitary voice, however well-travelled and trenchant an observer of popular culture and sacrilege he may have been, his lack of testimony is admittedly thin evidence. Considering the absolute silence prior to 1442, the “courtly invention” scenario is therefore the most plausible, keeping in mind that, whether it went from artisan’s workshop to the streets and then to the courts, or from the courts to the streets, if we judge by the evidence both positive and negative that we have, the transfer must have been quick – in the range of about 2 years.


1442 – two records of payment for packs of carte da trionfi.
1450 – payments for carte da trionfi
1450-1455 – Chariot card, Issy-les-Moulineaux, school of Ferrara (other cards from this pack are in the Warsaw museum)
1451 – payment for a pack of triumph cards
1454 – payments for various packs of triumphs
1456 – Ugo Trotti recommends triumphs as a very good game
1457 – payments for packs of triumphs
1459 – printing block for triumph cards noted
1460 – payments for various packs of triumphs
1461 – payment for triumphs
1463 – payment for triumphs
1473 – earliest possible date for Ercole d’Este tarot

Cremona/Milan/Pavia/Masnago –
1443-1445 – approximate dates of Brambilla and Cary Yale tarots
1440s – dates of frescoes of wealthy card-players, traditionally considered tarocchi, in Palazzo Borromeo in Milan, and the Sala dei Svaghi of the castle in Masnago.
1448 – Marcello receives a standard tarot pack as a present while in the environs of Milan
1450 – Sforza writes a letter asking for his secretary to buy some packs of triumphs, which he appears to have received two days later
1452 – Sigismondo Malatesta writes the Duchess, Bianca Maria Visconti-Sforza, to commission for him some packs of triumphs
1455 – plausible date for the Visconti-Sforza pack (Dummett has recently argued for a date in the early 1460s)
1460s – fresco of card game between two players (male and female) in Roccabianca (now housed in Castello Sforzesco, Milan)
1468 – Galeazzo Maria Sforza commissions Bonifacio Bembo to paint a cycle of frescoes in the castle in Pavia, including one of “Ladies Bona and Isabeta (…) and her maids playing triumphs in the garden” (the explicit mention of “triumphs” gives weight to the impression that other contemporary “games frescoes” in Borromeo, Masnago and perhaps the Roccabianca card-scene are also intended to represent the playing of triumphs, although no tarot trumps can be seen in them).
1474 – two letters of GM Sforza requesting packs of triumphs

Florence –
1450 – permission of four card games, including triumphum
1450s – painted cards including “Charles VI”, Catania (Castello Ursino), Rothschild (attribution to Florence for these cards is a recent theory)
1463 – repetition of law permitting triumph to be played (with two additional games)
1466 – Minchiate is mentioned, presumably a card game (but there is no way to determine, and no consensus, that this is the same as the card pack and game that would be later known as Minchiate and Germini)
1471 – playing minchiate mentioned in Cortono (a city belonging to Florence)
1474-1478 – Florentine cards imported in Rome
1477 – Minchiate appears in a list of permitted games (but I cannot determine if trionfi is also listed, which might indicate that Minchiate was a distinct pack as well as game)

Siena –
1452 – record of importation of triumph cards from Florence (reported by Zdekauer, without transcription of original (hence the description “unconfirmed”))

Padua –
1455 – sermon of Roberto da Lecce Caracciolo, mentions triumph cards with popes and cardinals
1460 – approximate time Valerio Marcello is recorded to have played triumphs with his father’s friends in Monselice (he lived from 1452-1461, so I place it towards the end of his short life)

Bologna –
1459 – a stolen pack of triumphs is recovered
1477 – a Riminese man commissions a Bolognese cardmaker for several hundred packs of cards, including an unspecified number of triumph packs


Michael J. Hurst said...

Hi, Ross,

Very nice analytical treatment of the data and its temporal/spatial distribution. As with the introduction of playing cards to Europe, the fact that the earliest date remains stable despite decades of subsequent research is pretty convincing.

Why have some argued, as in WPC, that the date might have been decades earlier? Certainly that is possible, but it has seemed unlikely at least since Dummett published the Grand Collation of Evidence in 1980, when the basics of this temporal distribution became clear.

Regarding the question of courtly versus popular invention, it might be misguided. A cardmaker who sold cheap playing cards to the hoi polloi would probably also sell to the upper classes, and an invention like this would be almost immediately made known to both markets as well as in trade with nearby cities.

The key questions would be how popular it became in different locales, among what demographic groups it became popular, and therefore what the probability of some surviving evidence would be.

If, for example, Tarot was popular among average card players in Bologna while in Milan and Ferrara it also became popular among... well, among the people with whom we know it was popular, there might have been an equally thriving Tarot community in all three but only two of them left lots of early evidence.

Ross G.R. Caldwell said...

Hi Michael,

About the dating in WPC and other works, see The Visconti-Sforza Tarot Cards, page 6.

"If the invention occurred in Ferrara, it was probably only a short time before the first written record of the cards in 1442; the same is probably true if it was in Milan, though this is harder to substantiate. If it was in Bologna, on the other hand, it may have been as early as the second decade of the century, but the evidence for this is flimsy [he is here surely alluding to the Prince Fibbia legend - Ross]. It could not have been earlier because playing cards had been known for less than thirty-five years in 1410 which does not allow enough times for the invention of a radical innovation in the manner of play and in the cards used".

I don't find that reasoning for the low dating persuasive, and I find it hard to imagine the game flourishing in Bologna unmentioned for nearly 50 years. But I agree with the seriousness with which the Ferrara evidence should be considered - like the premise of "all things being equal" in the survival of evidence before and after 1442, the sudden appearance of the game in Ferrara should be very close to when it first became known there, whether invented there or not.

I don't know if he or any of us could have argued for a closer date from what was known in 1980, without presuming Ferrara or Milan as the place of invention. The earliest dates were Ferrara 1442 (and following), Milan mid-1440s, and Bologna 1459. Marcello's indications of the game in 1448, the Florence 1450 reference, Sforza's letters 1450, the Siena and Padua references - none of these were known to him then, and they tighten up the focus of the picture.

Dummett's opinion has appeared to change since WPC. In HGT page 1 he says "The tarot pack was invented in northern Italy in about 1425" (I assume Dummett for the historical assertions since from what I know of John McLeod he isn't very interested in the historical nitty-gritty, more the playing nitty-gritty).

In the Supplement just published, he opens instead with: "The invention in northern Italy, in the late 1430s, of the tarot pack..."

I like to think he has maybe drawn the same conclusions as I have from the chart.


Hi Ross and Michael,

I am a complete neophyte to blogging, but I can comment upon your remarks.

Yes, Michael Dummett has changed his mind, balancing between an ‘late’ birth and an earlier one. In 1980 he assumed Ferrara was a possible birthplace (whence the subtitle of GT: “From Ferrara to Salt Lake City”) and he believed the invention occurred “only a short time before the first written record of the [tarot] cards in 1442” (as in The Visconti-Sforza Tarot Cards of 1986).

However, some years later he had been convinced Milan was a more certain candidate. He also considered earlier dates, and concluded, in “Il Mondo e l’Angelo” (Naples, 1993), p. 106, that the Tarot was invented in Milan around 1428.

Here is how he presents this timeline (with which you may not be familiar):

1428: i tarocchi sono inventati alla corte viscontea.
(Dummett then supposed that the earliest tarot pack, the "Visconti di Modrone" tarot, had been painted to celebrate the wedding of Filippo-Maria Visconti with Mary of Savoy in 1428)
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 di matrici di legno.
1442: i tarocchi si diffondono da Bologna a Firenze.
1444: la composizione del mazzo di tarocchi diventa standartdizzata dappertutto.

He doesn’t seem to have changed since. HGT reflects this early dating.


Unknown said...

Thank you for joining in the discussion TD.

I was aware of Dummett's Il Mondo e l'Angelo chronology, but while he doesn't offer a dating theory in his Sylvia Mann lecture for the IPCS in 2004, he (or John McLeod) does say "late 1430s" in 2009 -
which is what made me think he had changed his mind for some reason.

(They credit me, erroneously, with Alciato's Lombardy order, on page S 6. I only brought it to the internet's attention - in fact it was Pratesi in the 1980s who rediscovered Zdekauer's 1890s reference - and Michael Dummett himself in 1993 mentions it (in Il mondo e l'Angelo pp. 327-328), which is where *I* found it.

There is even a 17th century reference to Alciato's discussion of tarocchi - obviously these things percolate slowly in historiography.)

My own methodology in this post has convinced me of a late 1430s date, at any rate.