Lexicon-phonology relationship

A forum for all topics related to constructed languages
Post Reply
Alessio
sinic
sinic
Posts: 386
Joined: 03 Sep 2012 21:27
Location: Modena, Emilia-Romagna, Italy

Lexicon-phonology relationship

Post by Alessio »

I'm trying to create a conlang in which the lexical meaning of each word is somewhat encoded into the word's pronunciation. My goal would be that if two words sound alike, there is a high chance that they mean something similar; it should thus be possible to understand what a word means just by hearing how it sounds, without having memorized its meaning prior to that moment.
So far I've tried several solutions, but none of them really seems to float my boat.

My first try was - some may say obviously - triconsonantal roots, a concept borrowed from Semitic languages. For example, I established that the root for "dwelling" would be <d-s-t> and applied the pattern CaCC to get <dast> (house), or CiCāC to get <disāt> (houses). This is nice and all, but it's somewhat limited. I couldn't find a satisfying way to derive a big enough word pool from a single root. How do I turn "house" into "mansion" or "hut"? Sure, I could define a pattern for "big things" and one for "small things", but then I'd need a pattern for the plural of each of those, possibly one to turn them into verbs or adjectives, and then maybe those patterns would be unused for most other roots. Do I make another root then, maybe by changing only one consonant so that it stays similar enoguh? With three consonants being the only distinctive element between roots, it's easy to run out of combinations if you go down this path, or end up having two different concepts whose roots differ only by one consonant (you only have so many after all).

Second try - one-syllable roots. This is the one that worked best for me so far. The problem is that you must find a balance between how many roots you make and how much you break down a concept before stopping. You can have a root for "crab", or you can analize it as "claw" + "fish", or you can break "fish" into "water" and "animal", or maybe "claw" into "pinch" and "tool". For every division you make, you need to use one more syllable, so your word for crab might end up being six or seven syllables long. Which leads us to the next option.

Third try - oligosynthesis. Definitely one of the coolest approaches to lexicon I've ever tried. My take on it is that you have one consonant and one vowel/diphthong for each semantic prime, and build words by alternating them (CvCvCv...). Words are a lot shorter because you now have (mostly) a single letter per morpheme as opposed to a whole syllable. The big problem with this approach is that the semantic primes are about 50, so you'd need both 50 consonants AND 50 vowels/diphthongs to have a reliable way to form new words. The result is that you need to distinguish, for example, aspirated and labialized consonants, short/long vowels, tones and such to have a big enough phonemic inventory without filling it with exotic sounds (that I probably can't pronounce properly anyway). Thus the words risk not sounding all that different even when they mean very different things - /tʰaːkʷ/ would be made up of three completely different roots from /tak/. Also, have you ever tried breaking down complex concepts to the semantic primes that make them up? It's not so easy. For example, I'd render "computer" as "calculating machine"; "calculate" would be "math" + "do", and "math" would be "number" + "study", where "study" might be "know" + "become" or something like that. How many levels are we down the graph? Good, now get back up and do the same for "machine". Also, find a way to mark the boundary where the part that means "calculating" ends and the one that means "machine" begins.

Fourth try - one I'm particularly proud of. Lohet Silgeh, the language of an alternate reality I'm currently working on where Antarctica drifted off to the Indian Ocean rather than to the South Pole and thus never froze over, uses bisyllabic roots where the syllables can be swapped to form antonyms or other kinds of derivations. For example, "lohet" means "language", and "hetlo" means "gibberish". "Silgeh" means "Antarctica", and "Gehsil" means "the rest of the world". You can then squeeze infixes between the two syllables, for example if -sö- is the infix for adjectives "losöhet" means "linguistic, pertaining to languages", and "Silsögeh" means "inhabiting Antarctica; Antarctican". I love Lohet Silgeh, but it only adds one "free" pattern to the game and requires infixes to form any other concept. Might as well go back to monosyllabic roots at this point.

So, after this big wall of text I wrote, what do you think? Do you have any suggestions on how to build a reliable and regular system to create lexicon in such a way that similar concepts are represented by words that sound similar?
:ita: :eng: [:D] | :fra: :esp: :rus: [:)] | :con: Hecathver, Hajás, Hedetsūrk, Darezh...

Tin't inameint ca tót a sàm stê żōv'n e un po' cajoun, mo s't'armâgn cajoun an vōl ménga dîr t'armâgn anc żōven...
User avatar
Creyeditor
MVP
MVP
Posts: 5091
Joined: 14 Aug 2012 19:32

Re: Lexicon-phonology relationship

Post by Creyeditor »

This looks a lot like what people habe tried with philosophical languages. Wilkins' Real Character is a good example, IIRC.

Natlangs seem to do this only to a limited extent, in sound symbolism and the like. On the other hand there is a recent paper in the journal Linguistic Typology that languages do a bit more than sound symbolism in non-arbitrary phonology-meaning relationships. Here is the link: https://www.degruyter.com/view/journals ... y=abstract

What we learn from sound symbolism is that the connection is often between broader phonological categories and more general semantic fields. Front vowels indicate smallness, sibilants relate to water. This can be fruitfully used in a naturalistic artlang, but I don't know if it is geberalizable to an engelang-y approach.
Creyeditor
"Thoughts are free."
Produce, Analyze, Manipulate
1 :deu: 2 :eng: 3 :idn: 4 :fra: 4 :esp:
:con: Ook & Omlűt & Nautli languages & Sperenjas
[<3] Papuan languages, Morphophonology, Lexical Semantics [<3]
User avatar
Pabappa
greek
greek
Posts: 577
Joined: 18 Nov 2017 02:41

Re: Lexicon-phonology relationship

Post by Pabappa »

https://en.wikipedia.org/wiki/Ro_language is another example of a language with a similar idea. also, credit to http://www.ygyde.neostrada.pl/ for seeing the major flaw in Ro and rectifying it, though Ygyde also changes a lot of other things.

Your first three options could be combined .... there's no reason a language has to stick to only one strategy of building words. Even in an oligosynthetic language, there's no reason the morphemes have to occur in a strict order ... so you could have C_C shells where the vowel that comes between carries the most salient meaning. e.g. if you associate colors with vowels, your color words could just be /tan ten tin ton tun tən/ for the basic six, or however many basic terms you want. And use infixes, prefixes, etc to build words for finer shades of colors.

Your fourth idea, on the other hand, seems like it would be harder to combine with any of the others, since the order of the morphemes in a word becomes a source of meaning in itself. I like all of these ideas, but I think it might be difficult to win if you stick with just one.
Kavunupupis, šiŋuputata.
When I see you pointing at me, I know I'm in trouble. (Play)
Tanni
greek
greek
Posts: 600
Joined: 12 Aug 2010 02:05

Re: Lexicon-phonology relationship

Post by Tanni »

Alessio wrote: 12 Jun 2020 21:32 I'm trying to create a conlang in which the lexical meaning of each word is somewhat encoded into the word's pronunciation. My goal would be that if two words sound alike, there is a high chance that they mean something similar; it should thus be possible to understand what a word means just by hearing how it sounds, without having memorized its meaning prior to that moment.
This describes the usual derivational morphology with affixes quite well.
My first try was - some may say obviously - triconsonantal roots, a concept borrowed from Semitic languages.
Even Semitic languages use Pre- and Suffixes.
Second try - one-syllable roots. This is the one that worked best for me so far. The problem is that you must find a balance between how many roots you make and how much you break down a concept before stopping. You can have a root for "crab", or you can analize it as "claw" + "fish", or you can break "fish" into "water" and "animal", or maybe "claw" into "pinch" and "tool". For every division you make, you need to use one more syllable, so your word for crab might end up being six or seven syllables long. Which leads us to the next option.
The problem here is that the analysis of "claw" + "fish" is or at least can be quite misleading. This can be a good starting point, of course. Then apply some thousand years of language development, and the word respectively will have become a unique root.
Third try - oligosynthesis. ... For example, I'd render "computer" as "calculating machine"; "calculate" would be "math" + "do", and "math" would be "number" + "study", where "study" might be "know" + "become" or something like that. How many levels are we down the graph? Good, now get back up and do the same for "machine". Also, find a way to mark the boundary where the part that means "calculating" ends and the one that means "machine" begins.
This could also be a normal pocket calculator. A computer could also be perceived as a "graphics machine" or a "sound machine" depending on its primary usage in the culture in question.
My neurochemistry has fucked my impulse control, now I'm diagnosed OOD = oppositional opinion disorder, one of the most deadly diseases in totalitarian states, but can be cured in the free world.
Post Reply