New CBB

Another time to drink is when disagreements break out about how to pronounce a linguistic term.

_________________
death to vowels
/ / / /

I'd like to discuss the question brought up in the show: "Why SVO and not some other order"?

I used to have the same intuitions as George, that it would - for some reason - be easier to tell the words apart with SVO. But intuitions do not always work as a window to the truth. Let's think about various cases in which ambiguity could occur in an analytical language.

It was claimed that SOV could possibly be interpreted as SAdjV, in a language where nouns could also serve as modifiers of other nouns (as adjectives).

Could another ambiguity occur with SVO? Suppose verbs could also have an attributive/modifying/adjectivish use. And that nouns could also serve as verbs. Then SVO could possibly be interpreted as SAdjV

Suppose you have a sentence like "man walk shop" in an isolang. An SVO interpretation could be "The man walks to the shop". An SAdjV interpretation could be "The walking man is shopping".

So, can we for sure say that one word-order would be more or less ambiguous?

_________________
constructedlanguages.net

That's almost how my main conlang works. It has no compounding at all, and almost no derivational morphology (maybe 5 affixes), but has a lot of inflectional morphology (about 130 affixes last time I counted).

The language is highly synthetic and highly analytic (e.g. it's wh-in-situ).



Yes, we can. However, "SVO" isn't the complete word order of any language. One SVO language may be ambiguous, while another SVO language may be unambiguous, because the order of subject, verb, and object is just a tiny fraction of the word order of a language (and word order isn't all there is to a language, either), and no matter which of the six possibilities we choose the language may be either ambiguous or unambiguous.

However, from my experience (and I intend my conlang to be completely unambiguous), it's a lot easier to make a VSO or SOV language that is expressive, unambiguous, simple and otherwise having pleasing design, than to make such a language with SVO order. My conlang uses a pure head-initial word order (even with prefix notation for conjunctions, e.g. "X and Y" is expressed as "and X Y", a feature which possibly is unattested among natlangs, I'm at least not aware of any natlang doing it).

The idea, which I've heard several times, that SVO is superior to VSO and SOV when it comes to distinguishing between subject and object is bollocks, though. If we have VSO and the clause "verb noun noun", it's just as unambiguous which noun is the subject and which is the object as if we used any other of the five possibilities. However, if we had the clause "noun1 verb1 noun2 verb2 noun3", SVO order, and the subject or object could be a clause, we don't know if "noun1 verb1 noun2" is the subject of verb2, or if "noun2 verb2 noun3" is the object of verb1, while in a VSO or SOV language it would be unambiguous.

I use VSO or SOV for all my recent conlangs, except for one sketch that uses SVO: a Simlish inspired simple engelangy artlang for which I use the commonalities of English and Mandarin as a basis.

Hmm, good points, though I think an SVO language is more likely (but not destined) to be head initial, which makes that ambiguity moot. there could still be a case where a verb is mistaken for a noun, though that seems like it would be less common than confusion about nouns being attributive or free.



This is the wrong way to approach this. All languages have ambiguities. Even engelangs and loglangs that try to eliminate ambiguity entirely likely end up with a few cases. As William said in our episode, ambiguity is non-fatal. My thoughts were that the observed tendency that highly analytic languages tend to be SVO might be explained by an avoidance of an ambiguity -- but the observed tendency is just a tendency, and there are certainly some analytic languages of other word orders.



Latin had "X Y-que" where "que" is a clitic meaning "and". I believe there are other languages that do this as well. There are some languages that don't use a conjunction here at all, in fact, just juxtaposition.



Clauses used as subject or object are a much less common occurrence than a simple sentence. Also, your "verb noun noun" fails to address my example of nouns used attributively. Does your language not allow this?

_________________
George Corley
Producer and Moderating Host, Conlangery Podcast

Of course much will depend on other features of the languages' grammar. If you have, for example, obligatory determiners preceding noun phrases, and/or obligatory TAM particles preceding verbs, it may help one to discern phrase boundaries and avoid some ambiguities.


How do you know that? Can you in some way measure the degree of ambiguity? And what do you mean by "ambiguity"? "Ambiguity" in the strict, logical sense? Or "ambiguity" in the sense that it's likely to cause confusion for actual flesh and bone-speakers?



But if millions, or even billions, of speakers of analytical languages over millennia tend to favour SVO word order, it might be because they somehow find it easier to process that word order. Even if we perhaps don't know why. (We don't know everything about human cognition and language processing.) My spontaneous reaction to the idea that the subject of one clause may be confused with the object of a preceding clause, is that it looks like a typical drawing board-ambiguity, that's unlikely to pose real problems for flesh and bone-speakers. I could be wrong though.

_________________
constructedlanguages.net

lmao "You don't want George or Bianca mentioning your language in the same breath as Klingon." Indeed.

_________________
I speak English and a touch of Gàidhlig.
I am creating a conworld, which I refer to as the Carrion Series, that will contain three languages, Iriex, Dvoen and Maxna.

_________________
George Corley
Producer and Moderating Host, Conlangery Podcast

Good Episode!

_________________
"Wait for the starvation of glaciers, chicken hamburger Noel Weber Arthur."

= | = | =

Aye, a very good episode. And I'm proud to have the rare /ɢ/ in Hra'anh.

_________________
Hra'anh, Ponese

I'd like to create a language with either only one phonemic vowel, or no phonemic vowel at all.

_________________
constructedlanguages.net

The wrong way to approach what? Just to be clear: I wasn't talking about "the observed tendency" at all.



I know that. Do you know any natlang that uses the word order "conj X Y"?



My intention is that my conlang should handle any sentence of any complexity without upper limit unambiguously.

Why address a problem that isn't there? Yes, my conlang doesn't allow that.



The reason I wrote "and otherwise having pleasing design" is that it's easy to make an unambiguous language with SVO word order, the solutions that I've so far found, however, doesn't please me aesthetically.

I mean "ambiguity" in "the strict, logical sense".

I was referring to you saying a language is "... either ambiguous or unambiguous". It's a little like saying "shades of gray are either black or white". Perhaps there is some engelang or loglang that has entirely eliminated ambiguity, at least syntactic ambiguity. But outside of extremely careful construction of a language that humans probably couldn't use effectively, I can't see anyone creating a totally unambiguous language.



No, though I wouldn't be surprised if it existed somewhere.



OK. So you are creating a loglang.



I didn't know the problem wasn't there until you told me. That's why I asked the question.



Fair. You are entitled to your own aesthetic opinions.

_________________
George Corley
Producer and Moderating Host, Conlangery Podcast

*implosion of the mind*

Hey, wait. I did have an idea for an entirely vowel-less language. It would be tonal, mostly voiceless, full of plosives and ejectives, a couple of voiceless ingressive, inspirated stops (the totally reverse airflow of an aspirated stop, I don't think quite the same as "voiceless implosives"), and it would be stress-timed and often metronomic. Think of beatboxing as a language. And I may have been inspired by that one scene in Men in Black II. I'd call it pʼː tʼ pʼː kʼ or something.

_________________
Hra'anh, Ponese

Could it be tonal of it's mostly voiceless? Or are you referring to the few voiced sounds that actually exist?

When you say 'vowel-less', do you mean 'no vowel phonemes' or 'no vowel sounds at all'?

_________________
constructedlanguages.net

uh, hi. Is anyone else having issues playing these podcasts?

_________________
"We are like butterflies who flutter for a day and think it's forever."

A number of people seem to be, yeah.

_________________
I speak English and a touch of Gàidhlig.
I am creating a conworld, which I refer to as the Carrion Series, that will contain three languages, Iriex, Dvoen and Maxna.

I suppose all its vowels would be completely devoiced. And the tonal system is just an illusion based on the way things are articulated. Lots of non-pulmonic consonants with tense (resulting in "high tone") and lax (resulting in "low tone") articulations.

_________________
Hra'anh, Ponese

Super awesome to get Jim Henry on the show: Conlangery #49: The Noun Phrase

_________________
George Corley
Producer and Moderating Host, Conlangery Podcast

Whether a language is unambiguous or not is a binary opposition, even if it is a matter of degree how ambiguous an ambiguous language is.



My point is that having VSO word order and allowing nouns to be used attributively isn't a problem to begin with. Only if you make a lot of assumptions about the rest of the language will it be a problem.

Please show me a totally unambiguous language. If any of the loglangs and engelangs out there have achieved zero ambiguity, I highly doubt it's a user-friendly one.



This confuses me. Are you saying that it wouldn't be a problem even if it did have attributive nouns? Or are you suggesting that I assumed that the language had attributive nouns a priori (which I didn't, I asked you whether it had them)?

_________________
George Corley
Producer and Moderating Host, Conlangery Podcast