Klingon spelling checker

[pne]

After reading through the documentation on Vim's spelling checker interface (basically: it can read many input files designed for Myspell or hunspell), I thought I'd try a simple Klingon input.

However, after playing around a little, it seemed that the interface isn't really designed for a polysynthetic(?) language such as Klingon, where words can have dozens of affixes[*]. The documentation says that two suffixes and one prefix are supported, but I couldn't even get two suffixes in a row to work.

And while enumerating all possible suffixes is just about doable for nouns, you end up with several hundred combinations; for verbs, it simply isn't feasible, since you have nearly 200'000 combinations even before factoring in rovers or noun suffixes on nominalised verbs.

[*] I'd say that most words seen "in the wild" have no more than two or three suffixes and a prefix, though; words with more suffixes tend to be done for the sake of adding suffixes, rather than as part of natural speech.

For a word with lots of suffixes, consider something along the lines of {jIHoHHa''eghqangbe'qu'qa'moHlaHqu'be'bejtaHneSghach'a'meyqoqwIjmo'}... meaning, perhaps, something along the lines of "because of my so-called major certainly-not-BEING-ABLE-to-continue-to-make-myself-be-UNready-again-to-un-kill-myself-nesses". Rather silly, really.

FWIW, my favourite noun with the maximum number of suffixes is the canon {QaghHommeyHeylIjmo'} "due to your apparent minor errors".

Comments

[muckefuck.livejournal.com]

YM "highly agglutinative". Linguists disagree over precisely what degree of synthesis characterises "polysynthetic" languages, but a common shibboleth is object incorporation (http://en.wikipedia.org/wiki/Object_incorporation), a feature absent from Klingon AFAIK.

[pne.livejournal.com]

Thanks for the correction; the only criterion I had in mind was "many morphemes per word", but I could imagine that wouldn't be sufficient.

(Klingon indeed has no object incorporation.)

[ubykhlives.livejournal.com]

I'd say that most words seen "in the wild" have no more than two or three suffixes and a prefix, though; words with more suffixes tend to be done for the sake of adding suffixes, rather than as part of natural speech.

For what it's worth, that's true of many polysynthetic languages. I've found several instances of nine-morpheme words in Ubykh, and I think it's theoretically possible to construct a single phonological word of seventeen or eighteen morphemes, but the mean ratio of non-zero morphemes to phonological words in connected text seems to be a little over 3:1. (Apropos of not much, it's comforting to note that my own Ubykh writing seems to match that. The first two paragraphs of Philosopher's Stone average 3.19 non-zero morphemes per phonological word.)

I'd be interested to see how spell-checkers are implemented for highly agglutinative languages. With a language like Navajo, there's quite a bit of intermorphemic dependence that might make a spell-checking program practically feasible, but in Ubykh and Klingon the vast majority of morphemes are grammatically independent. But I would imagine that grammar checkers for agglutinative languages would be quite useful for learners when doing tasks like keeping track of Ubykh verbal agreement, say, or keeping nouns aligned with the right types of classificatory verbs in Navajo.