A Year’s Plan: 2021

I always have a lot of irons in the fire when it comes to creative projects. I tend to get a lot of ideas, usually more than I know what to do with; and of course I get very excited about them and I want to work on them immediately. While that’s undoubtedly a great thing, it means often try to work on a bunch of things at once and none of them get done. 

This year I’m going to try something a little different — not an earth-shattering idea but something I haven’t personally tried before. A sort of personal marshmallow test. I’ve picked eight of my top-priority creative projects and assigned each to a two-month period next year. Here’s my list for 2021:

DEC / JAN: The Voynich Manuscript

This is a (purportedly) 15th century document written in some completely unknown language / code / cipher. There are 232 pages now (originally ~270); each about 24×16 inches. The text is not in Roman characters, although they’re clearly influenced by them: 

No Roman or Arabic numerals appear in the text (except page numbers, added much later). And there were at least two separate authors (based on handwriting), and five sections, which have been labeled botanical, pharmaceutical, astrological, “balneological”, and textual. It appears to be a sort of cyclopedia of knowledge, not an uncommon kind of thing in the day. But the plants in it are unknown to science, and while the astrological section appears to match somewhat with our skies, a lot of it is inexplicable. Perhaps the oddest thing about the document is the balneological section, which has pages and pages of illustrations of naked women bathing in pools.

Normally trying my hand at translating something like this would be way outside my wheelhouse, but a recent paper by Bowern and Lindemann (Sept 2020) makes me think that the document might possibly be a language game based on a form Latin. If that’s the case, then it severely circumscribes the problem space, and makes it conceivable that I might be able to crack it.

In any case, I’ve been nattering away at it for a few months now, and although the majority of the things I’ve tried have not worked, I do I think I’ve shown that if it is a language game of this sort, it’s much more likely to be Latin or a Romance language than a Germanic language. I’ll be working on it the rest of January, but if I don’t get it cracked by then I’ll set it aside for a while and let it simmer while I go on to…

FEB / MAR Axon, Inc.

I do think a couple of months will be enough time to get this into a shape that I’d be comfortable sending out to agents again. The main issue is the beginning, which I’ve never been very happy with; but I have some ideas.

APR / MAY OT Syntax Paper

About 15 years ago I discovered a pretty remarkable pattern in the distribution of syntactic structures of the world’s languages — a pattern that predicted not only which kinds of structures would appear, but what their frequency would be (e.g. why prepositions are more common than postpositions, why SOV languages are often associated with postpositions, why VSO is less frequent than SVO or SOV, how this is related to adjective order, etc.), to a greater level of accuracy than has been achieved before (to my knowledge). I’ve presented this in a couple of places but I need to write it up properly and see if I can’t get it into a journal. 

JUN / JUL Sagaia

Return to Sagaia is a fantasy that I’ve been stuck on for some time, but I really would love to finish it properly. The last time I worked on it I was working on a chapter that featured the Ape; I found that I couldn’t write a convincing demagogue. Unfortunately I’ve had a lot more exposure to demagoguery in the meantime, and it may be that I can get unstuck now…

AUG / SEPT DagoBERT

Here’s another linguistics paper. A few years ago I developed a technique to predict which derivational affixes a root can combine with using  word embeddings. It’s an esoteric thing (and in some ways it’s really a solution in search of a problem) but I’ve seen a few published papers trying to achieve this, and I think my method would work at least as well as those. I was granted a patent on it a couple of years ago, also; so it’s worth writing up.

OCT / NOV Cretaceous Humanity

This is the project that I’ve been doing most of my recent conlang and world building work for. The idea is that a group of perhaps 200 scientists from the early 22nd century traveled back in time in order to research the Cretaceous period, targeting a time about 2500 years prior to the meteor impact that initiated the destruction of the dinosaurs. To be clear, time travel was a well-understood and widely used technology; the point of the expedition was to research the geology and biology of the period. They intended to stay for perhaps two months, and then return. But they were stranded in the past. They never found out why; no rescue or help was ever sent; and they had no way to get home. They and their descendants simply had to carry on as best they could, knowing that in 2500 years the apocalypse would arrive, probably killing all of them and destroying any lives they had managed to build in the meantime. This year, I hope to wrap up a sketch of the major languages and the history of the people, and maybe think about what kinds of stories (short stories or novels) I might want to write in this setting.

DEC Cobalt project

This is a project primarily for Cobalt (the company I work for) so I won’t discuss it here, just some low-priority research related to Cobalt’s dialogue engine I want to spend some time on to see if it’s worth pursuing further.

JAN / FEB 2022: UP

UP (or Universal Parser) is the latest iteration of a system I’ve been fiddling with for a couple of decades. The core intuition is that most linguistic knowledge is lexical knowledge; i.e., the lexicon is not just a list of words and meanings, but a network of meanings and associations, and this network consists of most of a person’s knowledge of a language. Syntax, on the other hand, is a rather simple, unimportant system. UP is somewhat like construction grammar but with two critical differences: (1) the lexicon is more structured, with strong associative and inheritance links between words, and (2) a lot of information within constructions such as element order are in fact not in the lexicon, but are derived from the “four rules of syntax” principles (see above). This is a big project and not going to be finished by Feb 2022, but I hope to use this time to reassess where I am with it (since I haven’t touched the code in a few years) and see if I can at least do some planning and make some more progress.