From Anatole France’s The Queen Pedauque (La Rôtisserie de la reine Pédauque) (ch17):

“That’s well,” sail M, d’Anquetil. “You have some wine, I have dice and cards in my pocket. We can play.”

“It is true,” said my good master, “that is a pleasant pastime. A pack of cards is a book of adventure, of the kind called romances. It is so far superior to other books of a similar kind that it can be made and read at the same time, and that it is not necessary to have brains to make it, nor knowledge of reading to read it. It is a marvellous work, also, in that it offers a regular and new sense every time its pages are shuffled. It is a contrivance never to be too much admired, because out of mathematical principles it extracts thousands on thousands of curious combinations, and so many singular affinities that it is believed, contrary to all truth, that in it are discoverable the secrets of hearts, the mystery of destinies and the arcanum of the future. What I have said is particularly applicable to the tarot of the Bohemians, which is the finest of all games, piquet not excepted.

It’s probably either a trope I don’t recognize or a chance convergence of thought but this reminds me much of the notion behind Italo Calvino’s The Castle of Crossed Destinies. Either way, Calvino’s explanation of the formation of that work – in which characters, unable to speak, tell their tales through setting out Tarot cards and are given verbal interpretations by the narrator – is curious enough to deserve quoting. At bottom is an image of how this process played out:

I publish this book to be free of it: it has obsessed me for years. I began by trying to line up tarots at random, to see if I could read a story in them. “The Waverer’s Tale” emerged; I started writing it down; I looked for other combinations of the same cards; I realized the tarots were a machine for constructing stories; I thought of a book, and I imagined its frame: the mute narrators, the forest, the inn; I was tempted by the diabolical idea of conjuring up all the stories that could be contained in a tarot deck.

I thought of constructing a kind of crossword puzzle made of tarots instead of letters, of pictographic stories instead of words. I wanted each of the stories to have a coherent significance, and I wanted them to afford me pleasure in writing them—or in rewriting them, if they were already classic stories…. And so I spent whole days taking apart and putting back together my puzzle; I invented new rules for the game, I drew hundreds of patterns, in a square, a rhomboid, a star design; but some essential cards were always left out, and some superfluous ones were always there in the midst. The patterns became so complicated (they took on a third dimension, becoming cubes, polyhedrons) that I myself was lost in them.

To escape from this impasse I gave up patterns and resumed writing the tales that had already taken shape, not concerning myself with whether or not they would find a place in the network of the others. But I felt that the game had a meaning only if governed by ironclad rules; an established framework of construction was required, conditioning the insertion of one story in the others. Without it, the whole thing was gratuitous.
There was another fact: not all the stories I succeeded in composing visually produced good results when I set myself to writing them down. There were some that sparked no impulse in the writing, and I had to eliminate them because they would have lowered the tension of the style. Then there were others that passed the test and immediately acquired the cohesive strength of the written word which, once written, will not be budged.

Suddenly, I decided to give up, to drop the whole thing; I turned to something else. It was absurd to waste any more time on an operation whose implicit possibilities I had by now explored completely, an operation that made sense only as a theoretical hypothesis. A month went by, perhaps a whole year, and I thought no more about it. Then all of a sudden, it occurred to me that I could try again in a different way, more simple and rapid, with guaranteed success. I began making patterns again, correcting them, complicating them. Again I was trapped in this quicksand, locked in this maniacal obsession. Some nights I woke up and ran to note a decisive correction, which then led to an endless chain of shifts. On other nights I would go to bed relieved at having found the perfect formula; and the next morning, on waking, I would tear it up. Even now, with the book in galleys, I continue to work over it, take it apart, rewrite. I hope that when the volume is printed I will be outside it once and for all. But will this actually happen?