LCC Digital 2021

This is supporting documentation and some additional notes and comments for my LCC 2021 talk, Magn(idea) = big: Collocations for Lexicon Building.

A few things I didn't really have time to get into in the talk...

How Many?

Not agreed upon, as far as I can see. I've seen suggestions that the collocational lexicon is about half the size of our single word lexicon — which is quite large — but Mel'čuk states that there are millions of collocational entries.


There hasn't been a great deal of research into why collocations are such a rich part of our vocabularies. I've seen arguments that it allows for greater specificity. I myself wonder if it isn't a kind of redundancy coding. Just as things like agreement provide a little extra bit of information in your communication that can help you recover what was said if you mishear, collocations might provide similar help. Finally, I refer you to Decorative Morphology in Khmer, in which the author (John Haiman) argues that some parts of a language might be purely decorative, kept for esthetic reasons.

Lexical Explorations

For MTT linguists, lexical functions are descriptive, but for conlangers there are interesting creative possibilities. For example:

Result(buy X) = own X.

Here we have a telic verb, buy, that leads to a reasonable resulting transitive stative verb, own. What other sorts of telic verbs might have interesting results?

Result(buy X) = own X.
Result(kill X) = ??? X.

There are various paraphrases I can use to express this notion of Result(kill), but there's not to my knowledge a transitive stative verb that captures this meaning. It's not clear how useful or necessary such a term would be, but it's interesting to think about.

Copyright (c) 2006-2022 William S. Annis