Zompist Bboard Again

bradrn wrote: ↑Mon Aug 23, 2021 7:01 pm Optimality theory, on the other hand…

...would be the COBOL of phonology if it paid

I had no idea there was so much to it.

Ketsuban wrote: ↑Mon Aug 23, 2021 5:19 pm Yeah, I overstepped the mark and came off a bit too accusatory there, and I'm sorry about that.

I sentence you to handing out tea & pickles to the next 10 new board members.

Nortaneous wrote: ↑Mon Aug 23, 2021 7:04 pm

bradrn wrote: ↑Mon Aug 23, 2021 7:01 pm Optimality theory, on the other hand…

...would be the COBOL of phonology if it paid

I was under the impression that, of all the purportedly universal formal linguistic theories, OT worked the best, aside from its little uncomputability problem. (I’ve never understood why they don’t just add input constraints.) Certainly it seems more effective than any of Chomsky’s attempts at formal syntax.

bradrn wrote: ↑Mon Aug 23, 2021 7:31 pm

Nortaneous wrote: ↑Mon Aug 23, 2021 7:04 pm

bradrn wrote: ↑Mon Aug 23, 2021 7:01 pm Optimality theory, on the other hand…

...would be the COBOL of phonology if it paid

I was under the impression that, of all the purportedly universal formal linguistic theories, OT worked the best, aside from its little uncomputability problem. (I’ve never understood why they don’t just add input constraints.) Certainly it seems more effective than any of Chomsky’s attempts at formal syntax.

Uncomputability isn't the only problem - AIUI orthodox OT holds that the only differences between languages are differences in the ranking of constraints, which doesn't sound like it's panned out

I've learned to eschew all formal linguistic theories as probably wrong, whether it's turning all languages into English or deciding that all languages, as mentioned, have the same rules and differ only in their priority.

I was writing a reply and the forum ate it. Bleh. As a result, I'm just posting the barebones of my reply here, so it may come off as terse.

bradrn wrote: ↑Mon Aug 23, 2021 7:01 pm stuff

I disagree. Wikipedia is a highly useful resource and many conlangers have said stated such. Academic literature is good, but Wikipedia often has the advantages of time (pages are generally a good summary) and accessibility (no paywall: if you can't access a resource, it's useless).

I was thinking of your ergativity thread and Zompist's Syntax Construction Kit. Both are very good resources, but I think could be made more digestible. For the ergativity thread, I would have stuck much closer to McGregor's "buckets" of morphological, syntactic, and semantic ergativity. These give users a simple framework to help organize the topic and make it easier to understand.

Gathering & organizing lists of various resources would be a valuable endeavor in the conlang community. Most such lists tend to be random, with no structure, which limits their usefulness. It would allow community members to reduce the amount of redundant searching & reading of materials others had to do to create quality conlangs.

Vardelm wrote: ↑Mon Aug 23, 2021 8:49 pm I was writing a reply and the forum ate it.

Aargh! Sympathy.

bradrn wrote: ↑Mon Aug 23, 2021 7:01 pm stuff

I disagree. Wikipedia is a highly useful resource and many conlangers have said stated such. Academic literature is good, but Wikipedia often has the advantages of time (pages are generally a good summary) and accessibility (no paywall: if you can't access a resource, it's useless).

Paywalls are certainly an issue. But if you have access (which, um, there are ways of getting even if you don’t belong to a library, which luckily I do), academic resources are far, far better. There are certainly occasional cases where Wikipedia gives an OK summary, but as I said, its best linguistic articles still tend towards incomprehensibility and/or uselessness, and are far inferior to the best academic articles; while its worst are simply wrong.

Or let’s take a more concrete example. Say I want to know more about comparatives. (Not a cherry-picked example; I was trying to find information about comparatives a while ago.) First up, Wikipedia. I won’t quote the whole thing, but it has an… um, interesting structure. First, it says ‘the syntax of comparative constructions is poorly understood’. Next, a whole section on ‘absolute and null forms’, which I’ve never seen mentioned anywhere else, and are a very minor part of comparisons for conlanging purposes. The rest of the article is a bunch of syntactic examples (in English, of course) to illustrate specific points on the syntax of comparison — all very interesting in its way, but again, somewhat useless for conlanging. There is nary a non-English example to be seen in the whole thing. This is sadly typical of Wikipedia articles, especially on matters syntactic.

Next, I’ll have a look at my favoured academic resource, Dixon’s Basic Linguistic Theory. If I open it to the chapter 26 ‘Comparative constructions’, I see the following:

Dixon wrote: Comparison, in general terms, involves examining two or more items in order to note similarities and differences between them. Many languages include grammatical means for coding comparison; however, not all do so. §28.2.1 considers the rationale for this.

The prototypical comparative scheme in a grammar involves comparing two participants in terms of the degree of some gradable property relating to them, as in the English sentence John is more handsome than Felix. The property is typically expressed by an adjective, in a language with a large open
class of adjectives; or else by a stative verb (with an adjective-like meaning).

The prototypical comparative scheme is characterized in §26.1. In §26.2 we see how it may be realized through various types of mono-clausal construction. The discussion is extended, in §26.3, to bi-clausal constructions, and to languages which do not have a dedicated comparative construction as such but instead employ what we can call a comparative strategy (and there is mention of languages which have available a combination of means). §§26.4–6 provide brief discussion of ‘less’ and ‘the same as’, of superlatives, and of inherently comparative lexemes.

And then we get a systematic overview of every single mono-clausal and bi-clausal construction which has been used for expressing comparison, organised by the possibilities for Parameter, Index and Mark, with copious glossed examples ranging from Jacaltec to Dhimal. This is perfect for conlanging — I can look through the whole chapter, compare the examples, and eventually decide that e.g. a biclausal comparative strategy (his ‘Type S’) would work well for Wēchizaŋkəŋ.

Vardelm wrote: I was thinking of your ergativity thread and Zompist's Syntax Construction Kit. Both are very good resources, but I think could be made more digestible. For the ergativity thread, I would have stuck much closer to McGregor's "buckets" of morphological, syntactic, and semantic ergativity. These give users a simple framework to help organize the topic and make it easier to understand.

Thanks for the feedback! Though I wish you had told me while I was writing the thing. I did try to organise it as you said — I started with an overview, then morphological ergativity (including split ergativity), then syntactic ergativity, then diachronic ergativity. Semantic ergativity I left out because I couldn’t find anything on it.

(Also, if you want to reply to this, maybe you should do so in the ergativity thread itself. No need to clutter up this thread with ergativity.)

Gathering & organizing lists of various resources would be a valuable endeavor in the conlang community. Most such lists tend to be random, with no structure, which limits their usefulness. It would allow community members to reduce the amount of redundant searching & reading of materials others had to do to create quality conlangs.

Yes please! I’ve seen some attempts, but nothing comprehensive.

Vardelm wrote: ↑Mon Aug 23, 2021 8:49 pm

bradrn wrote: ↑Mon Aug 23, 2021 7:01 pm stuff

I disagree. Wikipedia is a highly useful resource and many conlangers have said stated such. Academic literature is good, but Wikipedia often has the advantages of time (pages are generally a good summary) and accessibility (no paywall: if you can't access a resource, it's useless).

wikipedia imo is good for looking up phoneme inventories and sometimes basic conjugation and declension charts, not great for much else. which is possibly relevant to how conlangers make phoneme inventories and sometimes basic conjugation and declension charts and usually not much else

Nortaneous wrote: ↑Mon Aug 23, 2021 11:24 pm wikipedia imo is good for looking up phoneme inventories and sometimes basic conjugation and declension charts, not great for much else. which is possibly relevant to how conlangers make phoneme inventories and sometimes basic conjugation and declension charts and usually not much else

There are plenty of articles that are more than just inventories & charts, and do well at at least introducing a topic: morphosyntactic alignment, tense & aspect, moods, consonant gradation, tones, & more. Yes, if you want to add detail of a particular feature to your conlang, you'll probably want to look up more resources.

Even (especially?) if I grant your assertion, it still indicates the conlang community hasn't done a great job of organizing lists of quality, accessible resources.

Vardelm wrote: ↑Tue Aug 24, 2021 9:07 am

Nortaneous wrote: ↑Mon Aug 23, 2021 11:24 pm wikipedia imo is good for looking up phoneme inventories and sometimes basic conjugation and declension charts, not great for much else. which is possibly relevant to how conlangers make phoneme inventories and sometimes basic conjugation and declension charts and usually not much else

There are plenty of articles that are more than just inventories & charts, and do well at at least introducing a topic: morphosyntactic alignment … consonant gradation

OK, I’ll give you this — those articles are pretty good. (I hadn’t even seen the one on consonant gradation before; I’ll have to read it through carefully.) But I think these topics benefit from being more niche: only people who actually know their content are likely to contribute.

tense & aspect, moods

You must be joking. Wikipedia’s articles on TAM are some of the most confusing documents I’ve ever read, in any subject. When I say it’s incomprehensible, useless and wrong, well, I’m mostly thinking about the way Wikipedia ‘explains’ aspect.

tones

This article is interesting. Some parts are excellent (particularly the coverage of tonogenesis); some are OK if you know nothing about the subject, but superfluous or useless for conlanging (e.g. ~5 pages on transcribing tone, half a page of tongue-twisters, little discussion on plausible tone systems); one or two parts are actively wrong, even contradicting the rest of the article (‘the majority of tone languages belong to the Niger-Congo, Sino-Tibetan and Vietic groups’, ‘It was not until recent years that tone was found to play a role in inflectional morphology’). I’d say this mix is generally the norm for Wikipedia’s linguistics articles.

bradrn wrote: ↑Tue Aug 24, 2021 9:21 am You must be joking. Wikipedia’s articles on TAM are some of the most confusing documents I’ve ever read, in any subject. When I say it’s incomprehensible, useless and wrong, well, I’m mostly thinking about the way Wikipedia ‘explains’ aspect.

This article is interesting. Some parts are excellent (particularly the coverage of tonogenesis); some are OK if you know nothing about the subject, but superfluous or useless for conlanging (e.g. ~5 pages on transcribing tone, half a page of tongue-twisters, little discussion on plausible tone systems); one or two parts are actively wrong, even contradicting the rest of the article (‘the majority of tone languages belong to the Niger-Congo, Sino-Tibetan and Vietic groups’, ‘It was not until recent years that tone was found to play a role in inflectional morphology’). I’d say this mix is generally the norm for Wikipedia’s linguistics articles.

Again, they are **A** good resource, not **THE** resource. I haven't said you can learn everything about a topic from just Wikipedia. Pages serve as an overview of and/or intro to a topic. Just because academics poo-poo Wikipedia as a site for references (which is what they SHOULD be doing; it's not fully peer-reviewed, academic level writing!) doesn't mean that it's useless. I have found a few things sorely lacking or confusing on Wikipedia, but in general I've learned a lot from it.

Additionally, I'm not trying to point to specific pages I think are great, but to show how broad Wikipedia's linguistics coverage is compared to Nortaneous saying that it's only useful "for looking up phoneme inventories and sometimes basic conjugation and declension charts".

Vardelm wrote: ↑Tue Aug 24, 2021 9:43 am

bradrn wrote: ↑Tue Aug 24, 2021 9:21 am You must be joking. Wikipedia’s articles on TAM are some of the most confusing documents I’ve ever read, in any subject. When I say it’s incomprehensible, useless and wrong, well, I’m mostly thinking about the way Wikipedia ‘explains’ aspect.

This article is interesting. Some parts are excellent (particularly the coverage of tonogenesis); some are OK if you know nothing about the subject, but superfluous or useless for conlanging (e.g. ~5 pages on transcribing tone, half a page of tongue-twisters, little discussion on plausible tone systems); one or two parts are actively wrong, even contradicting the rest of the article (‘the majority of tone languages belong to the Niger-Congo, Sino-Tibetan and Vietic groups’, ‘It was not until recent years that tone was found to play a role in inflectional morphology’). I’d say this mix is generally the norm for Wikipedia’s linguistics articles.

Again, they are **A** good resource, not **THE** resource. I haven't said you can learn everything about a topic from just Wikipedia.

…ah, OK, thanks for clarifying this point. Because when I started conlanging, Wikipedia was basically my only resource aside from zompist’s books, so anything confusing there had a massive effect on me. And I’m sure I couldn’t be the only conlanger relying entirely on Wikipedia. (I know I see it quoted here every now and again, even when it’s wrong.) If I’d known about the depth and quality of those typological books I keep on talking about, I’m sure my conlangs would have gotten much better much earlier.

Additionally, I'm not trying to point to specific pages I think are great, but to show how broad Wikipedia's linguistics coverage is compared to Nortaneous saying that it's only useful "for looking up phoneme inventories and sometimes basic conjugation and declension charts".

Yeah, I agree that’s far from its only use.

I personally find Phonology to be one of the less interesting part of language and I find it tedious to come up with extremely detailed "descriptions" of phonetics of a language that isn't actually spoken by anyone. For most of my recent languages, I keep the phonology chapter very simple and move on to what I find more interesting. That's not to say interesting things can't be done with phonology (and in particular, as some of you have pointed out, the intersection of phonology and other aspects of grammar) but I'm super invested in coming up with cool ways to treat cross reference, indexing, case alignment, phrase structure, algorithmically determined word order, differential object marking, hierarchical index marking etc. etc. Just a lot more meat on the bones, and more thought provoking (to me!) than descriptions of very particular tongue movements I may or may not be capable of.

vegfarandi wrote: ↑Tue Aug 24, 2021 2:12 pm I personally find Phonology to be one of the less interesting part of language and I find it tedious to come up with extremely detailed "descriptions" of phonetics of a language that isn't actually spoken by anyone. [snip]

Yeah, what he said! My feeling is, what can we do about "the phonology hurdle"? Um... walk around it and go on your way?

fusijui wrote: ↑Tue Aug 24, 2021 4:53 pm Yeah, what he said! My feeling is, what can we do about "the phonology hurdle"? Um... walk around it and go on your way?

Toss the phonology stuff to Nortaneous to play with in the corner while the rest of us move on to interesting things!


I say that as someone jealous of Nort's phonology prowess.

vegfarandi wrote: ↑Tue Aug 24, 2021 2:12 pm I personally find Phonology to be one of the less interesting part of language and I find it tedious to come up with extremely detailed "descriptions" of phonetics of a language that isn't actually spoken by anyone. For most of my recent languages, I keep the phonology chapter very simple and move on to what I find more interesting. That's not to say interesting things can't be done with phonology (and in particular, as some of you have pointed out, the intersection of phonology and other aspects of grammar) but I'm super invested in coming up with cool ways to treat cross reference, indexing, case alignment, phrase structure, algorithmically determined word order, differential object marking, hierarchical index marking etc. etc. Just a lot more meat on the bones, and more thought provoking (to me!) than descriptions of very particular tongue movements I may or may not be capable of.

This is, I think, a good encapsulation of the attitude I find frustrating: that phonology is just consonants, vowels, syllable structure and allophony, and grammar is where all the interesting stuff is. And I’ll admit to finding grammar intensely interesting — I can spend literally months reading about the intricacies of verb serialisation or split intransitivity — but there are just so many more interesting things that can be done with phonology. For one thing, unless you make a language with no morphophonology whatsoever (which is basically unattested on Earth), morphology is intensely dependent on phonology: stuff like infixes and mutation are impossible to specify without going back to the phonology, and that in turn has consequences elsewhere. (e.g. One reason we know that Nias has the intensely rare marked-absolutive alignment is the exact pattern of initial consonant mutation in the absolutive.) Or you can have stuff like syncope or harmony rules or tone sandhi, which can mess up morphology in all sorts of fascinating ways. I personally find it really thought provoking to see how messed-up I can make my morphology, which ends up with things like e.g. Hlʉ̂ verb alternation between singular and plural objects (actually an infix fused with syllable nucleus), or the unholy mix of syncope, lenition and mutation in Wēchizaŋkəŋ.

bradrn wrote: ↑Sun Aug 22, 2021 8:01 pm

even phonosyntax (burmese developing tonal word class distinctions from phrase-final creaky voice)

Sorry, what‽

Can't find citation, possibly I hallucinated this

bradrn wrote: ↑Tue Aug 24, 2021 9:20 pm

vegfarandi wrote: ↑Tue Aug 24, 2021 2:12 pm I personally find Phonology to be one of the less interesting part of language and I find it tedious to come up with extremely detailed "descriptions" of phonetics of a language that isn't actually spoken by anyone. For most of my recent languages, I keep the phonology chapter very simple and move on to what I find more interesting. That's not to say interesting things can't be done with phonology (and in particular, as some of you have pointed out, the intersection of phonology and other aspects of grammar) but I'm super invested in coming up with cool ways to treat cross reference, indexing, case alignment, phrase structure, algorithmically determined word order, differential object marking, hierarchical index marking etc. etc. Just a lot more meat on the bones, and more thought provoking (to me!) than descriptions of very particular tongue movements I may or may not be capable of.

This is, I think, a good encapsulation of the attitude I find frustrating: that phonology is just consonants, vowels, syllable structure and allophony, and grammar is where all the interesting stuff is. And I’ll admit to finding grammar intensely interesting — I can spend literally months reading about the intricacies of verb serialisation or split intransitivity — but there are just so many more interesting things that can be done with phonology. For one thing, unless you make a language with no morphophonology whatsoever (which is basically unattested on Earth), morphology is intensely dependent on phonology: stuff like infixes and mutation are impossible to specify without going back to the phonology, and that in turn has consequences elsewhere. (e.g. One reason we know that Nias has the intensely rare marked-absolutive alignment is the exact pattern of initial consonant mutation in the absolutive.) Or you can have stuff like syncope or harmony rules or tone sandhi, which can mess up morphology in all sorts of fascinating ways. I personally find it really thought provoking to see how messed-up I can make my morphology, which ends up with things like e.g. Hlʉ̂ verb alternation between singular and plural objects (actually an infix fused with syllable nucleus), or the unholy mix of syncope, lenition and mutation in Wēchizaŋkəŋ.

Note that I didn't say phonology was not interesting, just less interesting. A more succinct view of what I was saying is that I find con-phonetics super boring. And as I did say, where phonology interersects with morphology and syntax – that stuff is super interesting.

vegfarandi wrote: ↑Thu Aug 26, 2021 10:26 am

bradrn wrote: ↑Tue Aug 24, 2021 9:20 pm This is, I think, a good encapsulation of the attitude I find frustrating: that phonology is just consonants, vowels, syllable structure and allophony, and grammar is where all the interesting stuff is.

Note that I didn't say phonology was not interesting, just less interesting. A more succinct view of what I was saying is that I find con-phonetics super boring. And as I did say, where phonology interersects with morphology and syntax – that stuff is super interesting.

Also, it's perfectly reasonable to find various aspects of a topic more or less interesting.