Friday, February 10, 2006

Doubly Maximal Lipograms

Cup of java in my hand, so warm,
A sunny, pinkish glow proclaims a morn.
Across a Sound, Olympic pyramids
Command a land from shiny, snowy lids.

About five years ago, I wrote a few examples of what I call "doubly-maximal lipograms"; writings missing the two most-common letters in English, "e" and "t". I misplaced my notebook containing them, though.

Yesterday I decided to start anew, one or two a day, and with a goal in mind for each piece. Yesterday's was: "If any of you, man, woman, or child, can show a fair claim for avoiding a union of him and his woman, you should say so now." Today's piece is the trochaic pentameter poem at top; today was the first day in a (rainy) month or so that I could see sunrise off the Olympics.

After my effort in 2001, I came across a comment in Douglas Hofstadter's "Le Ton beau de Marot" (p. 107), where he notes that in 1983 he and a few friends started speaking in plain maximal lipograms (speaking without words with "e"). They then tried speaking in doubly-maximal lipograms (though he did not give it a name). He wrote:

We soon discovered that whereas in an "e"-less mode we felt hampered but were still able to say pretty much anything we wanted, in the new doubly deprived mode, we were virtually immobilized -- not a single pre-formed thought could be expressed without lengthy mental struggle, and usually not even then. Deleting just the one letter of highest frequency was debilitating but survivable; deleting the two letters of highest frequency brought us to our knees.


I'm going to withhold judgment on how difficult it is after practice. It certainly is hard when starting out. I suspect it's not so much the time you need to become fluent at single or double lipography, but rather the volume of output. That is, it may take having a few dozen lipograms under your belt to get the tricks of that lipogram form. If you're avoiding "e" and "t", that number of successes may take ten or twenty times as long as if you're just avoiding "e", but currently I do suspect that the tricks are out there. You will, however, probably find that you can be nowhere near as precise as you want. For example, I haven't thought of a good replacement for "three" (or tre-/tri- words), other than "four minus ... a singular?".