Will my language have an ambiguity problem?

A forum for discussing linguistics or just languages in general.
xBlackWolfx
banned
Posts: 179
Joined: 16 Aug 2010 04:30

Will my language have an ambiguity problem?

Post by xBlackWolfx » 07 Sep 2010 02:55

I dont intend on using homophones. I can easily avoid that.

The problem is with my re-design I've considered moving back to a simplistic phonology. There's 8 consonants and 3 vowels and the language allows a (C)[l]V[n] syllable structure. Though it will give me enough possible words. I think it may still cause problems since inevitibly one string of syllables can easily have multiple meanings depending on which syllables in the string you count as a word together. for example, 'pa tasi' can just as easily be interprested as 'pata si'. And it only gets worse as the number of syllables increase. 'Pa ta si la' could be interpreted as 'pata sila', 'pata si la', 'patasi la', 'pa tasi la', 'pa tasila', 'pa ta sila', 'pa ta si la', and probably more.

I could switch to a (C)[l]V(C) structure since my latest conscript was being designed for such a language anyway, but I dont know if that would help avoid this problem.

Any help?

User avatar
MrKrov
banned
Posts: 2414
Joined: 12 Aug 2010 01:47
Location: /ai/ > /a:/
Contact:

Re: Will my language have an ambiguity problem?

Post by MrKrov » 07 Sep 2010 02:58

OH MY CTHULHU. Do you really have this problem in English? Prosody. Learn it.

User avatar
Ilaeriu
sinic
sinic
Posts: 313
Joined: 12 Aug 2010 01:42
Location: Canada, eh?

Re: Will my language have an ambiguity problem?

Post by Ilaeriu » 07 Sep 2010 04:30

...

Anyway, yeah, prosody would probably help. And I'm no expert in it, but check out what they do in Hawaiian since you'd probably be in a similar situation.
Image

(in order of proficiency from greatest to least)
Native :eng: | Fluent :tgl: | Learning :esp: · :kor: · :zho: · :qbc: | Want to learn :lat: · :jpn: · :ara:

Etherman
cuneiform
cuneiform
Posts: 97
Joined: 14 Aug 2010 00:50

Re: Will my language have an ambiguity problem?

Post by Etherman » 07 Sep 2010 05:03

Context also will be a big help. On the flip side ambiguity, or near ambiguity, is good for comedy and rhetoric.

goneriku
rupestrian
rupestrian
Posts: 16
Joined: 31 Aug 2010 05:01

Re: Will my language have an ambiguity problem?

Post by goneriku » 07 Sep 2010 07:14

Maybe you should think about using tones.
Also what's wrong with homophones anyways?
L1: :eng: Learning: :por: :esp:

xBlackWolfx
banned
Posts: 179
Joined: 16 Aug 2010 04:30

Re: Will my language have an ambiguity problem?

Post by xBlackWolfx » 07 Sep 2010 07:42

i really really hate amibguity. i hated ppl making fun of me if i just so happened to say something that sounded like something else

i had problems learning to talk when i was a kid, had to take speech therapy classes throughout elementry school and even in middle school i still had issues getting ppl to understand what i was saying. and in middle school i always talked in my famous monotone simply bc i didnt understand tone and never bothered with it, as far as i could hear i sounded like anyone else. even when they tried to imitate me i still heard no difference, i do pay attention to tone now and can imitate it (at least it sounds identical to me) but i'm not used to it and i cant use it to interpret anything.

not to mention i'm a really abstract person with very precise thinking and english's inprecisiveness drives me insane. the fact that english doenst have a seperate collective plural annoys me for example. the fact that english relies on context to determine if by 'his desk' you mean 'the desk that he uses' or 'the desk that he owns' or 'the desk that is associated with him aka it's a gift he gave to someone'. heck in my conlang there will be eight cardinal directions instead of four (left, right, infront of, behind of and all four areas inbetween). and to make it more complex there may be even more than that considering that we live in a four dimensional world (besides up there may be above and infront of, above and behind, above and to the left etc...) but that may even be too much for me. but nonetheless now that i thought of i probably wont be able to resist the urge of putting them in. my preposition list btw is probably going to look similar to esperanto's (outside the fact that there'll be a partitive preposition along with several more i'd like to have). the whole thing's essentially going to be a loglan when i get done funny enough. a language based on my normal thought patterns turns into a loglan? lol

Systemzwang
greek
greek
Posts: 615
Joined: 15 Aug 2010 14:48
Contact:

Re: Will my language have an ambiguity problem?

Post by Systemzwang » 07 Sep 2010 09:21

xBlackWolfx wrote:there may be even more than that considering that we live in a four dimensional world (besides up there may be above and infront of, above and behind, above and to the left etc...)
"up"... the fourth dimension!

xBlackWolfx
banned
Posts: 179
Joined: 16 Aug 2010 04:30

Re: Will my language have an ambiguity problem?

Post by xBlackWolfx » 07 Sep 2010 09:24

Systemzwang wrote:
xBlackWolfx wrote:there may be even more than that considering that we live in a four dimensional world (besides up there may be above and infront of, above and behind, above and to the left etc...)
"up"... the fourth dimension!
that was a typo

Golahet
sinic
sinic
Posts: 212
Joined: 12 Aug 2010 15:01

Re: Will my language have an ambiguity problem?

Post by Golahet » 07 Sep 2010 14:31

I scream "ice cream"!

It sound like you need a self-segregating morphology!

xBlackWolfx wrote:the fact that english relies on context to determine if by 'his desk' you mean 'the desk that he uses' or 'the desk that he owns' or 'the desk that is associated with him aka it's a gift he gave to someone'.
That's not ambiguity. That's vagueness. A language works well without ambiguity, in fact it works better without ambiguity, but it can't work without vagueness. But ofcourse there's no problem in having a language being specific by default, like how English likes to indicate tense and number by default, and if you don't want to specify you need to paraphrase it. You just need some way to optionally be vague, that's all. And remember the difference between ambiguity and vagueness, the difference between polysemy and hypernymy.



For this suggestion I will assume that your eight consonants are /p t k m n l s h/ and your three vowels are /a u i/. If not, you could just replace them. I will assume the syllable structure is C(l)V(n). The optionality of the C is replaced with /ʔ/, which is a ninth consonant, but /ʔl/ is prohibited. /hl/ is possibly pronounced as [xl] or [ɬ].

Imagine there are two sets of phoneme blocks (I will call them "(phoneme) blocks" because they doesn't quite correspond to syllables), a set P and a set S. The following are the members of those sets:

P = pu, pi, ta, ti, ka, ku, ma, mi, na, nu, lu, li, su, si, ha, hi, ʔa, ʔu, pla, plu, tlu, tli, kla, kli, mlu, mli, nla, nli, sla, slu, hlu, hli, m.pa, n.tu, n.ki, n.sa, n.hu, n.ʔi, m.pli, n.tla, n.klu, n.sli, n.hla

S = pa, tu, ki, mu, ni, la, sa, hu, ʔi, pli, tla, klu, mla, nlu, sli, hla, m.pu, m.pi, n.ta, n.ti, n.ka, n.ku, n.su, n.si, n.ha, n.hi, n.ʔa, n.ʔu, m.pla, m.plu, n.tlu, n.tli, n.kla, n.kli, n.sla, n.slu, n.hlu, n.hli

The basic idea is that every morpheme has the shape P{P}{S}S (where {} indicates zero or more times). Thus, every morpheme is at least bisyllabic. But it's more complex than P{P}{S}S: first, the initial P of a morpheme can't be one containing a coda (those blocks marked with periods). Then, those P that contain codas may act as S, provided that they are immediately preceded by a "true" S, or they are immediately preceded by a "false" S that is in turn immediately preceded by a "true" S. Phonetically, /ml/ and /nl/ may act as codas followed by onsets if they occur within a word, but they are not considered such for this "complication".

This allows for words like tiʔi, slunsu, ʔani, nanlu, plahu, takintu, nani, tapa, etc, and morpheme boundaries are never ambiguous. Morphemes always end in vowels, which is necessary I think if you don't want geminates at morpheme boundaries or limit the possible onsets, or use "i" as a morpheme final epenthetic vowel and prohibit "ni" from occurring elsewhere.

Morphemes are always at least bisyllabic, which could be changed with another structure.

Basically, if you instead of the pattern above have all morphemes to be of the shape {P}S, that is: an optional string of P followed by a single "true" S, then you will have monosyllables too (note that if you do this, only a S that doesn't contain a coda will be acceptable as a monosyllabic root). The drawback is that the number of morphemes with three or more syllables will be less. On the other hand, maybe you use those monosyllabic morphemes as suffixes (including those that contain codas).



So, no, your phonology doesn't inevitibly lead to ambiguity problems.

Golahet
sinic
sinic
Posts: 212
Joined: 12 Aug 2010 15:01

Re: Will my language have an ambiguity problem?

Post by Golahet » 07 Sep 2010 15:33

You could add allophony to /ʔ/ to make it less salient. This is a suggestion:

First, [ʔ] is always allowed in all cases. It is in free variation with whatever the rest of these rules dictate.

Secondly, /ʔ/ is always [ʔ] word initially, as well as everywhere it doesn't occur between two vowels.

If it occurs between two vowels and those vowels are different, and one of them is /u/, then /ʔ/ is [w], and if one of them is /i/ then /ʔ/ is [j]. If one of the two vowels is /u/ and the other /i/, then it is [w] if the preceding vowel is /u/, and [j] if the preceding vowel is /i/. If the two vowels are identical, then /ʔ/ is [ʔ] no matter what those vowels are.

Furthermore, /ʔ/ doesn't need to be spelled word initially or between vowels in the romanization (nor in the native orthography, I suppose).

xBlackWolfx
banned
Posts: 179
Joined: 16 Aug 2010 04:30

Re: Will my language have an ambiguity problem?

Post by xBlackWolfx » 07 Sep 2010 15:39

the phonology is p f m t s n l k. you got the vowels right. but an L may not occur infront of all consonants, it cant occur infront of n or l and it probably wont occur infront of t. when i calculated the number of possible syllables i counted the Cl onsets along with the C onsets. this came out to 8+5=13, plus 1 for a 'glottal stop' or 'null consonant'. 14 times 3 is 42 and multiplied by 2 (half the syllables can end in N) we get 84 possible mono-syllables. to count the number of bisyllabic words, i removed two consonants (n and the null consonant) and recalculated. 12x3x2=72. thus 84x72=6048 possible legal bisyllabic words. since most compounding auxlangs get away with around 1000, and some compounding natlangs get away with about 5000 that should be more than enough words. but if i was to disallow [n] at the end of a word, that would reduce the number of possible words literally by half. 42 possible mono syllables to make pronouns prepositions numbers affixes and some probably basic adjectives and verbs with 3042 bi-syllabic words for root words.

this is why i abandoned the idea of such a limiting phonology in the first place. in my last conlang attempt (for which i actually made 200 words, making it the first conlang where you can actually say a sentence) the phonology was far larger, i believe it was p b f v m t d s z n l r j k g x a e i o u au ai oi. the syllable structure was (C)V[f v m s z n l r](C), where V also includes the diphthongs au ai and oi. the only other phonological constraints would've been that s could only occur before a voiceless plosive, a fricative could only occur next to a consonant of the same voicing (so dva and tfa are legal but dfa and tva are not) but this didnt apply across syllalbes. i never bothered to calculate how many words this would allow, it must be quite immense. infact i was planning on making the language mostly monosyllabic with only a scattering of bisyllabic words (the only one i can remember is vakva, water)

i loved how the language sounded, i thought it was beautiful. if you want a sample sentence:

men kar spedat asunden da mobi: my dog went underneathe that car there.
my dog go(perfective) to-under that(overthere and not by you) car.

the vocabulary is obviously based mostly on indo-european languages. i had completely ditched the idea of an auxlang and was just trying to make one that sounded pretty. but after taking a break from it for a while, i now think it sounds unpoetic and bland. no idea why.

i'm really leaning towards going back to a complex syllable structure and larger phoneme list. but for some reason i dont want to. i hate how simply phonologies sound but i dont like the idea of making more complex ones. it drives me insane.

User avatar
MrKrov
banned
Posts: 2414
Joined: 12 Aug 2010 01:47
Location: /ai/ > /a:/
Contact:

Re: Will my language have an ambiguity problem?

Post by MrKrov » 07 Sep 2010 15:57

Moderation can be fairly awesomesauce.

xBlackWolfx
banned
Posts: 179
Joined: 16 Aug 2010 04:30

Re: Will my language have an ambiguity problem?

Post by xBlackWolfx » 07 Sep 2010 19:18

MrKrov wrote:Moderation can be fairly awesomesauce.
what's that mean?

goneriku
rupestrian
rupestrian
Posts: 16
Joined: 31 Aug 2010 05:01

Re: Will my language have an ambiguity problem?

Post by goneriku » 07 Sep 2010 19:32

I'm thinking he means to compromise on your phonology- there's no need for it to be very complex or simple, you got like Ubykh on one side and Rotokas on the other and thousands of languages that fall somewhere in between.
And why do you have an IE vocabulary? Unless its a posteriori, that's pretty yawn.
L1: :eng: Learning: :por: :esp:

User avatar
MrKrov
banned
Posts: 2414
Joined: 12 Aug 2010 01:47
Location: /ai/ > /a:/
Contact:

Re: Will my language have an ambiguity problem?

Post by MrKrov » 07 Sep 2010 20:20

That and I find phonological processes like unconventional consonant mutations more intriguing than a mere do-nothing inventory.

Golahet
sinic
sinic
Posts: 212
Joined: 12 Aug 2010 15:01

Re: Will my language have an ambiguity problem?

Post by Golahet » 07 Sep 2010 21:45

xBlackWolfx wrote:but if i was to disallow [n] at the end of a word, that would reduce the number of possible words literally by half.
True, but regardless method you can't get rid of ambiguity for free.


Using the phoneme inventory

/m n p b t d k g ʔ f v s z x l r j/
/a e i o u au ai oi/

and the syllable structure

C(l)V(n)

(fun fact: WALS label everything beyond (C)V as "moderately complex")

I get

P = ma, me, mi, na, nu, ni, pa, pe, pi, ba, bi, bo, ta, tu, te, du, di, do, ka, ku, ko, ga, gi, go, ʔi, ʔe, ʔo, fu, fi, fe, va, ve, vo, sa, su, so, zu, zi, ze, xi, xe, xo, lu, li, lo, ru, re, ro, ja, ju, ji, mla, mli, mlo, pli, ple, plo, bla, blu, bli, klu, kli, kle, gla, glu, glo, flu, fle, flo, vlu, vli, vlo, sla, sli, sle, zla, zlu, zle, xla, xle, xlo, mau, mai, nai, noi, pau, poi, bai, boi, tai, toi, dau, dai, kau, kai, gau, goi, ʔai, ʔoi, fau, foi, vau, vai, sau, soi, zai, zoi, xau, xoi, lai, loi, rau, rai, jau, joi, mlai, mloi, plau, plai, blai, bloi, klau, kloi, glai, gloi, flau, flai, vlai, vloi, slau, slai, zlau, zloi, xlau, xloi (=135) m.pu, m.po, m.bu, m.be, n.ti, n.to, n.da, n.de, n.ki, n.ke, n.gu, n.ge, n.ʔa, n.ʔu, m.fa, m.fo, m.vu, m.vi, n.si, n.se, n.za, n.zo, n.xa, n.xu, n.ra, n.ri, n.je, n.jo, m.pla, m.plu, m.ble, m.blo, n.kla, n.klo, n.gli, n.gle, m.fla, m.fli, m.vla, m.vle, n.slu, n.slo, n.zli, n.zlo, n.xlu, n.xli, m.pai, m.bau, n.tau, n.doi, n.koi, n.gai, n.ʔau, m.fai, m.voi, n.sai, n.zau, n.xai, n.roi, n.jai, m.ploi, m.blau, n.klai, n.glau, m.floi, m.vlau, n.sloi, n.zlai, n.xlai (= 69)

S = mu, mo, no, ne, pu, po, bu, be, ti, to, da, de, ki, ke, gu, ge, ʔa, ʔu, fa, fo, vu, vi, si, se, za, zo, xa, xu, la, le, ra, ri, je, jo, mlu, mle, pla, plu, ble, blo, kla, klo, gli, gle, fla, fli, vla, vle, slu, slo, zli, zlo, xlu, xli, moi, nau, pai, bau, tau, doi, koi, gai, ʔau, fai, voi, sai, zau, xai, lau, roi, jai, mlau, ploi, blau, klai, glau, floi, vlau, sloi, zlai, xlai (= 81) m.pa, m.pe, m.pi, m.ba, m.bi, m.bo, n.ta, n.tu, n.te, n.du, n.di, n.do, n.ka, n.ku, n.ko, n.ga, n.gi, n.go, n.ʔi, n.ʔe, n.ʔo, m.fu, m.fi, m.fe, m.va, m.ve, m.vo, n.sa, n.su, n.so, n.zu, n.zi, n.ze, n.xi, n.xe, n.xo, n.ru, n.re, n.ro, n.ja, n.ju, n.ji, m.pli, m.ple, m.plo, m.bla, m.blu, m.bli, n.klu, n.kli, n.kle, n.gla, n.glu, n.glo, m.flu, m.fle, m.flo, m.vlu, m.vli, m.vlo, n.sla, n.sli, n.sle, n.zla, n.zlu, n.zle, n.xla, n.xle, n.xlo, m.pau, m.poi, m.bai, m.boi, n.tai, n.toi, n.dau, n.dai, n.kau, n.kai, n.gau, n.goi, n.ʔai, n.ʔoi, m.fau, m.foi, m.vau, m.vai, n.sau, n.soi, n.zai, n.zoi, n.xau, n.xoi, n.rau, n.rai, n.jau, n.joi, m.plau, m.plai, m.blai, m.bloi, n.klau, n.kloi, n.glai, n.gloi, m.flau, m.flai, m.vlai, m.vloi, n.slau, n.slai, n.zlau, n.zloi, n.xlau, n.xloi (=115)

(the members of the sets are chosen such that phoneme frequencies and semantic load should be more or less evenly distributed)


And with a {P}S structure, there are 196 monosyllabic morphemes, 81 unbound morphemes (i.e. those that doesn't start with codas), and 115 suffixes (i.e. those that start with codas). (In the language itself you could ofcourse use any of those "unbound morphemes" as suffixes, or any of those "suffixes" as postclitics.)

There are 39984 bisyllabic morphemes, 26460 (135*196) unbound morphemes and 13524 (69*196) suffixes.

And there are more trisyllabic morphemes than you will ever need.

xBlackWolfx wrote:but if i was to disallow [n] at the end of a word, that would reduce the number of possible words literally by half. 42 possible mono syllables to make pronouns prepositions numbers affixes and some probably basic adjectives and verbs with 3042 bi-syllabic words for root words.
Note that prohibiting /n/ word finally wasn't the deal. That alone isn't enough to completely eliminate ambiguity. My calculation is more like 1110 bisyllabic unbounded morphemes, or 1517 bisyllabic morphemes when including words that start with codas, half or a third of your 3042.

Curlyjimsam
sinic
sinic
Posts: 255
Joined: 01 Sep 2010 14:31
Location: UK

Re: Will my language have an ambiguity problem?

Post by Curlyjimsam » 07 Sep 2010 22:53

MrKrov wrote:OH MY CTHULHU. Do you really have this problem in English? Prosody. Learn it.
What MrKrov is trying to say here is something along the lines of: "If you have a predictable stress rule or similar, the problem you are worried about will not arise." If, say, you have a rule that says the first syllable of every word is stressed, you will also be able to tell when there's a new word because there'll be a stressed syllable. There are other ways of achieving something similar - this is probably the easiest.
Twitter: @jsbaker750
Website: seven-fifty.net

goneriku
rupestrian
rupestrian
Posts: 16
Joined: 31 Aug 2010 05:01

Re: Will my language have an ambiguity problem?

Post by goneriku » 08 Sep 2010 09:23

MrKrov wrote:That and I find phonological processes like unconventional consonant mutations more intriguing than a mere do-nothing inventory.
Definitely, I think OP should research more about consonant mutations and other phonological phenomena to enrich his language a bit.
L1: :eng: Learning: :por: :esp:

User avatar
Czwartek
greek
greek
Posts: 453
Joined: 12 Aug 2010 14:50

Re: Will my language have an ambiguity problem?

Post by Czwartek » 08 Sep 2010 13:56

How about simply by context? Mahal mentioned I scream/ice cream. In everyday speech we don't have any trouble interpreting the meaning of this homophonal phrase because each would make no sense in the context of the other (except for a situation where any random context-less word(s) are possible, for example in graffiti or a password.) English has many such homophonal phrases, but virtually all are automatically disambiguated through context.

And your example of 'his desk' has nothing to do with homophonal phrases, it's simply due to the multiple meanings of the possessive case in English. I agree with you about that point; I'd much rather speak a language which distinguishes these.

Systemzwang
greek
greek
Posts: 615
Joined: 15 Aug 2010 14:48
Contact:

Re: Will my language have an ambiguity problem?

Post by Systemzwang » 08 Sep 2010 18:02

Slight allophonic variation for onsets (possible or even likely for some consonants), maybe glottal stops before word-initial vowels (in a language that lacks glottal stops as a phoneme - this is not an unusual approach even in real languages)
not to mention i'm a really abstract person with very precise thinking and english's inprecisiveness drives me insane. the fact that english doenst have a seperate collective plural annoys me for example. the fact that english relies on context to determine if by 'his desk' you mean 'the desk that he uses' or 'the desk that he owns' or 'the desk that is associated with him aka it's a gift he gave to someone'.
it sounds like you're misusing the word abstract. Abstraction always is 'more generalized', whereas you want concrete precision.

As for 'his', think of it from Gricean maxims, especially the maxims of quantity (don't be needlessly informative, don't omit important information) and the maxim of relation ('relevance', hence, say stuff that's relevant - if saying 'by his desk' in an office where no one per se owns their desks, it's irrelevant to say 'he is by the desk he usually sits at (but doesn't own)' or any such, the shortest MOST ABSTRACT way is to just go straight ahead with a generalized 'his' as in 'the desk that somehow is associated to him by some manner of association').

the impreciseness is really an optimization. live with it. It's the most optimal way of doing language. Because we are smart enough to do very abstract pattern-matching.
Last edited by Systemzwang on 08 Sep 2010 19:34, edited 1 time in total.

Post Reply