Notes on Oulipo

In 1961, Raymond Queneau's book A Hundred Thousand Billion Poems (Cent mille milliards de poèmes) offered the striking but now well-known proposition that vast amounts of poetry could be generated from simple rules--that is, not from randomness but rather from constraints that gave definition to a space of possibilities. Queneau's book contained ten sonnets, all having the same rhyme scheme, but each page was cut into horizontal strips allowing each line of each sonnet to be exchanged with any line on the same row in any other sonnet. So the rule for generating other poems in this case would be to iterate through every possible exchange of lines combinatorically to yield a hundred thousand billion poems.

A Hundred Thousand Billion Poems was a foundational text for the Oulipo or "workshop of potential literature," a group that Queneau co-founded to explore the generation of literary texts from well-constrained rules. The Oulipo often looked back on the history of poetry for inspiration, and they took special interest in older forms that had been considered to work more like jokes. For example, the "Second Manifesto" by Oulipo's co-founder François Le Lionnais says, "Most writers and readers feel (or pretend to feel) that extremely constraining structures such as the acrostic, spoonerisms, the lipogram, the palindrome, or the holorhyme (to cite only these five) are mere examples of acrobatics and deserve nothing more than a wry grin, since they could never help to engender truly valid works of art."

However, a poem with 268,435,456 possible combinations actually had been treated as a valid work of art in the 17th Century, and it wasn't the only playfully generative poetry that long pre-dated Queneau's poem.

"Vœux pour sa Maiesté" by Honorat de Porchères Laugier

               V O E V X  P O V R  S A
                    M A I E S T É.

                     S O N N E T

               Du Sieur DE PORCHERES.

La grandeur       (&) l'amour  le destin, la victoire 
D'un Dieu,        d'une beauté du ciel &) des ſoldats
Conduiſe,         enflãme, anime (&) pouſſe en mille
                                                (parts
Tes pas,          tõ cœur tõ ame (&, ta vertu notoire.
                                                (moire
Iunon,            Pallas Cypris (&) la vielle me-
De ſe biens,      de ſes dons, de ſes ris, de ſes arts
Rempliſſe,        orne, contente (&) chãte tes hazars
Ta maiſon,        tes beaux ans, tõ eſprit (&) ta gloire
Que le printẽps,  l'Eſté, que l'Autõne (&) l'Hyuer,
De ſes fleurs,    de Zephirs, de ſes fruits, de ſon air
Te parfume,       t'eſuente (&) t'honore (&) t'agrée,
Bref que l'air,   que le feu, que la terre (&) que l'eau
Souffle,          eſchauffe, nourriſſe (&) raconte à
                                               (Nerée
Ton les,          tõ ſein, tõ corps (&) ton renom plus
                                                (beau.

• In several respects, it's a typical French sonnet with 12-syllable lines and an abbaabbaccdede rhyme scheme. The rhyme on line 2 is only a near rhyme with respect to lines 3, 6, and 7, but the overall form is too close to normal for its place and time to see that as a problem.
• The lines are not typical French alexandrines. They're also not alexandrins ternaires. The sub-division of each line seems irregular.
• Each line is instead composed of a list of four items, and the items within a line are similar parts of speech--and a bit repetitive to read in order.
• Also the ampersands are in parentheses, as if they could be omitted, but where present, the ampersands are essential for filling out the syllabic structure of each line. What are the parentheses communicating?
• Moreover, what is the typography communicating by setting the beginning of each line apart from the rest of the line?

La grandeur        (May) The greatness
D'un Dieu,         Of a God,
Conduiſe,          Guide,
Tes pas,           Your steps, [but the full line ends in a period]
Iunon,             (May) Juno,
De ſe biens,       With her blessings,
Rempliſſe,         Fill,
Ta maiſon,         Your house,
Que le printẽps,   May the Spring,
De ſes fleurs,     From its flowers,
Te parfume,        Perfume you,
Bref que l'air,    In short may the air,
Souffle,           Breathe (upon),
Ton les,           Your flank, [archaic but consistent with the line]

Let's suppose the parentheses around the ampersands indicate that the ampersands are optional too. That is, in spite of the ampersands' importance to the poem's measure, there is some way to read the poem such that they don't matter. Taking as a model the way in which the single typographic breakpoint in each line showed it can be read as a column, the optional ampersands suggest to the reader that the four segments of the line are not necessarily chained together and may be read as columns too. Here's a cleaner, modernized way to look at it that also backfills other punctuation to match the grammar:

Here's an English translation that also interpolates a "May ..." in a couple of places because the verb form that follows is in the subjunctive in French:

Note that Cypris is another name for Aphrodite or Venus, so a reading of Old Lady Memory as the goddess Mnemosyne or Moneta would be consistent with the line. Each segment within a row is essentially the same kind of thing, and the columns line up meaningfully too: Spring brings flowers, fire warms, water speaks to Nereus (the Old Man of the Sea), etc., etc.

But it's also remarkable how well each line segment follows any other segment from the previous line. Among the most challenging counterexamples, winter has some kinds of flowers too, there's no reason elements other than water can't speak to Nereus, and all we need to accept water breathing on someone or blowing over someone is to know that the context is a poem--the verb "souffler" has a ton of figurative meanings and possibilities. Achieving that level of cross-column coherence and so few rough spots across so many lines purely by chance seems unlikely.

Moreover, this isn't just a poem: it is a collection of wishes for a king, queen, queen regent, etc. The ease with which all four segments of each line recombine with all four segments of the line that follows should be interpreted in light of the poem's rhetorical aim: to offer impressive and bountiful well-wishes. What poem ever expressed something more bountifully than one encouraging the reader to recombine its line segments in every possible way?

Incidentally, de Porchères was a favorite of the Princess de Conti who evidently gave him the title "steward of nocturnal pleasures." French Wikipedia links the relevant Princess de Conti to Marie de' Medici, but that's more than I know. What I can say is that some Princess de Conti from around this time remained a legend in the history of parlor games, because she became the star of Charlotte-Rose Caumont de la Force's 1701 parlor game manual / historical novel Les Jeux d'esprit.

Anyway, with 4 options to select 14 times, the number of possible wishes within this sonnet is 4¹⁴ or 268,435,456--let's just say over 250 million. That's a lot less than a hundred thousand billion, but it's still a lot.

Well, the facts about this poem's typography are so glaring they probably meant something--they're not typical for the collection the poem is in. But is it reasonable to read a 17th Century poem this way? Was anyone else at the time using this kind of combinatoric potential to generate meaningful texts? It turns out the answer is yes, and it turns out yet more recombinant mythological hymns lay ahead in the history of 19th Century parlor games too.

Polygraphia by Johannes Trithemius

a Deus           a clemens
b Creator        b clementiſſimus
c Conditor       c pius
d Opifex         d pijſſimus
e Dominus        e magnus
f Dominator      f excelſus
g Conſolator     g maximus
h Arbiter        h optimus
i Iudex          i sapientiſſimus
k Illuminator    k inviſibilis
l Illuſtrator    l immortalis
m Rector         m eternus
n Rex            n semipiternus
o Imperator      o glorioſus
p Gubernator     p fortiſſimus
q Factor         q ſanctiſſiumus
r Fabricator     r incõprehenſibilis
ſ Conſervator    ſ omnipotens
t Redemptor      t pacificus
v Auctor         v miſericors
x Princeps       x miſericordiſſimus
y Paſtor         y cunctipotens
z Moderator      z magnificus
w Salvator       w excellentiſſimus

I'm not aware of a source linking Trithemius's cipher to poets or parlor game players from de Porchères's time, but I'm aware of a source that links Trithemius to them in general. Trithemius's other famous work was Steganographia, and it was definitely known to and even copied by Queen Elizabeth's court astrologer / advisor John Dee. In 1705, Robert Hooke speculated that Dee actually used Trithemius's methods to send secret messages, which I point out not because it's necessarily true but because what's necessarily true is that at least a couple of well-known figures from the Elizabethan and Baroque eras understood that combinatoric rearrangements of text were potentially meaningful.

Incidentally, there was another interesting book in John Dee's personal library: Hubert Philippe de Villiers's 1555 text, Cinquante Jeux divers d'honnete entretien, a French translation of fifty games taken from the one hundred games in Innocenzo Ringhieri's parlor game manual, Cento giuochi liberali, et d'ingegno, published in 1551. See my own partial translation of de Villiers's text, focused on games in it with mythological themes.

So secret messages and parlor games were both part of the life of one well-known person alive during de Porchères's lifetime--John Dee--illustrating not that de Porchères knew anything about either topic but rather that it's not anachronistic to suppose he could have independently come up with the idea of playing with language in a combinatoric way.

There are probably many ways of making the same point, e.g. via acrostics or other wordplay. But I want to make the point this way, because Trithemius definitely had a direct role in inspiring combinatoric literary games in the 19th Century.

"Ave Maria, Hymne mythologique et mystérieux" by Antoine Fabre d'Olivet

For one thing, Fabre d'Olivet's interests in languages, hermeneutics, and esotericism put him on the list of authors who could have written this text in the game manual, introducing a variant of Trithemius's cipher and attributing it to an unknown young man:

Madame B*** ne le nomme pas. Elle dit seulement que c'est un jeune homme, qui, deux ou trois longues soirées d'hiver, s'est amusé à composer cet Ave Maria, d'après l'idée prémitive qui en avait été conçue par l'abbé Tritême. Cet illustre abbé a effectivement composé un Pater Noster en latin; mais il suffit de le voir pour sentir que l'ordonnance lourde et monotone de son ancien Pater n'a rien de comparable avec l'élégance et la variété qu'on remarque dans notre moderne Ave Maria. Quelle différence sur tout de la langue latine à la française et d'une langue morte à une langue vivante? Combien il a fallu plus de travail, de goût et de richesse dans l'imagination, pour trouver vingt-cinq synonimes, dans cell ci, qui correspondissent parfaitement avec vingt-cinq autres qui les les précedent et vingt-cinq qui les suivent! -- Je ne sais, dit Madame B*** mais il me semble que cette bagatelle, quelque frivole qu'elle puisse paraître à certaines personnes est une des plus fortes preuves matérielles qu'on ait donné jusqu'ici de la richesse de la langue française et de sa merveilleuse flexibilité. Que les traducteurs après cela, ne viennent plus se plaindre devant moi de sa pauvreté et de la rareté des synonimes, car je les renvoye impitoyablement à cet Ave Maria, dont les vingt-cinq traductions différentes, pouvant se confondre entr'elles, forment dans leurs innombrables combinaisons, une série d'Hymnes, tous parfaitement cadencés et fidèles, dont la somme totale est faite pour écraser l'imagination des plus robustes Calculateurs.

Madame de B*** doesn't name him. She says only that it's a young man who, during two or three long winter evenings, amused himself by composing this Ave Maria, based on an initial idea conceived by the abbot Trithemius. This illustrious abbot did compose a Pater Noster in Latin; but it suffices to see it to sense that the heavy and monotonous structure of his old Pater has no comparison with the elegance and variety found in our modern Ave Maria. What a difference above all between the Latin and French languages and between a dead language and a living one? How much more work, taste, and richness of imagination were necessary to find twenty-five synonyms in the latter that correspond perfectly with the twenty-five that precede them and the twenty-five that follow! -- I don't know, said Madame B***, but it seems to me that this bagatelle, however frivolous it may appear to some people, is one of the strongest material proofs given so far of the richness of the French language and its marvelous flexibility. May translators after that not come complaining to me any longer about its poverty and the rarity of synonyms, because I will pitilessly refer them to this Ave Maria, whose twenty-five different translations [that is, substitutions of each letter in the alphabet], able to be merged with one another, form in their innumerable combinations, a series of Hymns, all perfectly cadenced and faithful, the total number of which is enough to break the imagination of the most robust Calculators [that is, people making calculations].

          A V E  M A R I A,

  Hymne mythologique et mystérieux,

  Imité de la Salutation Angélique.

I.                         II.
A  Je te salue,            A  Marie,
B  viens,                  B  Pallas,
C  vole,                   C  Jsis,
D  accours,                D  Astarté,
E  salut                   E  Vénus,
F  parais,                 F  Thétis,
G  descends,               G  Flore,
H  écoute,                 H  Eleusine,
I  ô                       I  Uranie,
J  auguste                 J  Vesta,
K  hélas !                 K  Pomone,
L  chaste                  L  Cypris,
M  céleste                 M  Cibele,
N  divine                  N  Hébé,
O  oh !                    O  Thémis,
P  sublime                 P  Cythérée,
Q  puissante               Q  Aphrodite,
R  tendre                  R  Diane,
S  belle                   S  Astrée,
T  sensible                T  Egérie,
U  ô toi,                  U  Junon,
V  montre toi,             V  Jris,
X  écoute nous,            X  Céres,
Y  entends nous,           Y  Minerve,
Z  exauce nous,            Z  Rhéa,

Well, but is it a poem generator? I'd say so: Fabre d'Olivet obviously admires how "cadenced" the output is. But it doesn't rhyme, so there's that. Fortunately someone else came along pretty soon to fix that.

"La Tragédie Burlesque," "Le Jeu des Oracles," and "Couplets" by Charles Jean Rougemaître

Oh, except it begins with a little role-playing game called "La Cour de France." This game works a bit like GutsMuths's 1796 "Avocat" variant called "Parlament" in that players take on roles of heads of state and whatnot and talk about things related to ruling the country, but the mechanics are also a bit like the various games of "The Butterfly"--that is, the gameplay is less about role-playing than making players pay for mistakes in how they respond.

Getting back to the more Oulipian games in the book, Rougemaître has one called "La Tragédie Burlesque" in which the players are all given numbers plus an assignment to write down a four-letter word (e.g. "Marc, jour, père, mère, etc."). In number order, the parlor game manual itself passes to each player to look up the four letters of their word on a series of tables that turn each letter into a longer line of text. Each player writes down their lines, and when they're all done, each player in order speaks their lines to the next. The tables look pretty much like the tables given above from Trithemius and Fabre d'Olivet, so to illustrate this game, I'll generate the outputs that would result from its example inputs:

Player 1 ("Marc"):

M - Es-tu fous
A - Vil aspic
R - Chanteur
C - Aux vilains yeux

Player 2 ("jour"):

J - Employons
O - C'est certains
U - Un biais
R - Mielleux

Player 3 ("père"):

P - Mon sang bout
È - Et je vais
R - Je t'occis
E - Pour bien faire

Player 4 ("mère"):

M - Arrêtez
È - Fier-à-bras
R - Car je crains
E - Ma Glycère

Next, Rougemaître gives "The Game of Oracles" in two variants, one for men and one for women. This is interesting on several levels, because it ties substitution-cipher games to fortune games that have a longer history (see the many fortune books that I've mixed in with a bibliography of old parlor game manuals), and it foreshadows the fact that the first 'choose your own adventure' gamebook, Consider the Consequences! (1930) was written by Doris Webster and Mary Alden Hopkins, who had published the fortune book Tell Your Own Fortune in 1929 and may have also been inspired by the parlor game "Consequences."

To play Rougemaître's fortune game for men, each player writes down a question that has exactly 16 alphabetic characters. The example he gives is "Quand me marîrai-je?" or "When will I be married?" For each letter, there's a corresponding table and row converting the letter into a line of text, just as in all these other substitution ciphers. The output text is a short and absurd poem, often but not necessarily suggesting what "She" will do or be like, and Rougemaître assures the reader that the answers will always be in well-measured and well-rhymed verse that can be sung to the tune of two evidently well-known airs. The oracles for women are very similar, but each player writes down a question with 18 alphabetic characters and the result can be sung to a variety of apparently well-known airs.

Finally, Rougemaître gives the game of "Couplets," and although it's very similar to the other games, he frames it in a super interesting way. What's typical about it is it can take a 17-character string of alphabetic characters as input in order to generate a metered poem that rhymes, and Rougemaître mentions that process could be used to encipher a secret message. But what's unusual is he also gives an alternative way to generate the poem:

on pourra se contenter de prendre une letter au hasard dans chaque colonne, en commençant par la première et en finissant par la dernière; c'est que, dans le premier avertissement, nous avons nommé: faire des vers sans aucun but déterminé.

one could be contented to take a letter haphazardly from each column, starting with the first and ending with the last; this is what, in the first avertissement, we had called: to create verse without any pre-set goal.

I define a gamebook as a narrative printed on paper and divided into sections connected by links. By following the links, the reader experiences alternative narrative paths which lead to multiple endings.

And what about Raymond Queneau?

I would be pretty surprised if Queneau's literary experiments like "A Hundred Thousand Billion Poems" or "A Story as You Like It" (a tiny gamebook-style story) had any direct relationship to historical parlor games. Of course, Queneau had been a Surrealist before co-founding Oulipo, and I've previously documented how earlier Surrealists definitely were influenced by at least one actual historical parlor game that was sometimes--as in several examples from 1812--just as surreal as the Surrealists. And I'd guess anyone born in France in 1903 knew more about historical parlor games than pretty much anyone born since WWII or so--mainly as folklore, like when Simone Collinet (André Breton's ex) remembered the game of "Consequences" simply as petits papiers traditionnels. But that's a far cry from direct experience with any of these very old poems and very old parlor game manuals.

On the other hand, would I say a lot of Oulipian constraints are literary parlor games? Sure and why not--in the same sense that François Le Lionnais said that gimmicky old forms can give rise to works of art, those ways of making art seem like pretty good games too.

As a final note, I'd like to acknowledge clavdivs on Metafilter, whose post about Su Hui's "Star Gauge" (a classical Chinese reversible poem) finally motivated me to expand on this material. "Star Gauge" seems so rich maybe anyone with an interest in Oulipian texts would want to know about it.