I don't know if this has been done - I assume since Dummett doesn't mention it he didn't - but certain figures could be estimated from the Fermiers Generaux archives and other taxes to know at the moment taxes were imposed how many decks were created, I believe this is Depaulis methodology.
Regarding those figures, I'd like to make clear an important detail, following a remark by Robert :
If a scientist were to carry out an experiment with a sample of 4 (or 40!) out of a pool of 1,000,000, I wonder how reliable the research would be?
which sounds quite confusing.
If the production was of a million, however accurate that figure may be which is not a big deal anyway, we speak about the number of copies of X decks, X being the figure to which we shall compare the remaining sample of "4 or 40".
Given the known numbers of cardmakers in France, given that not all of them made Tarot decks, given that some of them used maybe four molds in their career while other unique molds were used by several cardmakers, the X - number of different
decks - is far from a million. As I see it - and I may be wrong but would love to see this estimation corrected - a thousand different decks from XVIth and XVIIth centuries would be an extremely optimistic estimate of the number of different decks. Hence the sample of known decks may be more of 1 to 10% than 40 to on million, which makes quite a huge difference.
Moreover concerning the french parisian decks from the XVIIth century, the fact that those decks were part of Michel de Marolles collection, and as such were selected by him, should be taken in consideration when we wonder if this small sample is representative of the period production.
I don't mean here that we have a clear picture of the french cardmakers production, but maybe it is more accurate than is usually thought when we compare for instance the number of known different decks to the number of produced decks (each copy), a comparison that appeared here and there and though apparently convincing may be a bit irrelevant.
Again this is just my two cents, and I may be missing important points (maybe am I lacking some of those essential bricks we talked about earlier), so I eargerly expect your comments and corrections