The Universe of Discourse
Sat, 26 Apr 2014

My brush with Oulipo

Last night I gave a talk for the New York Perl Mongers, and got to see a number of people that I like but don't often see. Among these was Michael Fischer, who told me of a story about myself that I had completely forgotten, but I think will be of general interest.

Oulipo Compendium
Oulipo Compendium
with kickback
no kickback

The front end of the story is this: Michael first met me at some conference, shortly after the publication of Higher-Order Perl, and people were coming up to me and presenting me with copies of the book to sign. In many cases these were people who had helped me edit the book, or who had reported printing errors; for some of those people I would find the error in the text that they had reported, circle it, and write a thank-you note on the same page. Michael did not have a copy of my book, but for some reason he had with him a copy of Oulipo Compendium, and he presented this to me to sign instead.

Oulipo is a society of writers, founded in 1960, who pursue “constrained writing”. Perhaps the best-known example is the lipogrammatic novel La Disparition, written in 1969 by Oulipo member Georges Perec, entirely without the use of the letter e. Another possibly well-known example is the Exercises in Style of Raymond Queneau, which retells the same vapid anecdote in 99 different styles. The book that Michael put in front of me to sign is a compendium of anecdotes, examples of Oulipan work, and other Oulipalia.

What Michael did not realize, however, was that the gods of fate were handing me an opportunity. He says that I glared at him for a moment, then flipped through the pages, found the place in the book where I was mentioned, circled it, and signed that.

The other half of that story is how I happened to be mentioned in Oulipo Compendium.

Back in the early 1990s I did a few text processing projects which would be trivial now, but which were unusual at the time, in a small way. For example, I constructed a concordance of the King James Bible, listing, for each word, the number of every verse in which it appeared. This was a significant effort at the time; the Bible was sufficiently large (around five megabytes) that I normally kept the files compressed to save space. This project was surprisingly popular, and I received frequent email from strangers asking for copies of the concordance.

Another project, less popular but still interesting, was an anagram dictionary. The word list from Webster's Second International dictionary was available, and it was an easy matter to locate all the anagrams in it, and compile a file. Unlike the Bible concordance, which I considered inferior to simply running grep, I still have the anagram dictionary. It begins:

aal ala
aam ama
Aarhus (See `arusha')
Aaronic (See `Nicarao')
Aaronite aeration
Aaru aura

And ends:

zoosporic sporozoic
zootype ozotype
zyga gazy
zygal glazy

The cross-references are to save space. When two words are anagrams of one another, both are listed in both places. But when three or more words are anagrams, the words are listed in one place, with cross-references in the other places, so for example:

Ateles teasel stelae saltee sealet
saltee (See `Ateles')
sealet (See `Ateles')
stelae (See `Ateles')
teasel (See `Ateles')

saves 52 characters over the unabbreviated version. Even with this optimization, the complete anagram dictionary was around 750 kilobytes, a significant amount of space in 1991. A few years later I generated an improved version, which dispensed with the abbreviation, by that time unnecessary, and which attempted, sucessfully I thought, to score the anagrams according to interestingness. But I digress.

One day in August of 1994, I received a query about the anagram dictionary, including a question about whether it could be used in a certain way. I replied in detail, explaining what I had done, how it could be used, and what could be done instead, and the result was a reply from Harry Mathews, another well-known member of the Oulipo, of which I had not heard before. Mr. Mathews, correctly recognizing that I would be interested, explained what he was really after:

A poetic procedure created by the late Georges Perec falls into the latter category. According to this procedure, only the 11 commonest letters in the language can be used, and all have to be used before any of them can be used again. A poem therefore consists of a series of 11 multi-word anagrams of, in French, the letters e s a r t i n u l o c (a c e i l n o r s t). Perec discovered only one one-word anagram for the letter-group, "ulcerations", which was adopted as a generic name for the procedure.

Mathews wanted, not exactly an anagram dictionary, but a list of words acceptable for the English version of "ulcerations". They should contain only the letters a d e h i l n o r s t, at most once each. In particular, he wanted a word containing precisely these eleven letters, to use as the translation of "ulcerations".

Producing the requisite list was much easier then producing the anagram dictionary iself, so I quickly did it and sent it back; it looked like this:

a A a
d D d
e E e
h H h
i I i
l L l
n N n
o O o
r R r
s S s
t T t
ad ad da
ae ae ea
ah Ah ah ha
lost lost lots slot
nors sorn
nort torn tron
nost snot
orst sort
adehl heald
adehn henad
adehr derah
adehs Hades deash sadhe shade
deilnorst nostriled
ehilnorst nosethirl
adehilnort threnodial
adehilnrst disenthral
aehilnorst hortensial

The leftmost column is the alphabetical list of letters. This is so that if you find yourself needing to use the letters 'a d e h s' at some point in your poem, you can jump to that part of the list and immediately locate the words containing exactly those letters. (It provides somewhat less help for discovering the shorter words that contain only some of those letters, but there is a limit to how much can be done with static files.)

As can be seen at the end of the list, there were three words that each used ten of the eleven required letters: “hortensial”, “threnodial”, “disenthral”, but none with all eleven. However, Mathews replied:

You have found the solution to my immediate problem: "threnodial" may only have 10 letters, but the 11th letter is "s". So, as an adjectival noun, "threnodials" becomes the one and only generic name for English "Ulcerations". It is not only less harsh a word than the French one but a sorrowfully appropriate one, since the form is naturally associated with Georges Perec, who died 12 years ago at 46 to the lasting consternation of us all.

(A threnody is a hymn of mourning.)

A few years later, the Oulipo Compendium appeared, edited by Mathews, and the article on Threnodials mentions my assistance. And so it was that when Michael Fischer handed me a copy, I was able to open it up to the place where I was mentioned.

[ Addendum 20140428: Thanks to Philippe Bruhat for some corrections: neither Perec nor Mathews was a founding member of Oulipo. ]

[Other articles in category /lang] permanent link