[YG Conlang Archives] > [engelang group] > messages [Date Index] [Thread Index] >
OK, I'm off to a conference for a few days -- giving a paper on the syntax--phonology interface, ironically! --, out of the country so maybe without unextortionately-priced net access, but when I get back I'll write out what I think the few rules of xorban syntax are. In the meantime, if Mike were to find the time to augment his BNF rules with some illustrative examples, I could see if I could do a syntax for that too, tho I am anticipating some problems with that. --And. John E Clifford, On 09/10/2012 23:15:
Well, I was born in the 1930s and do a bit of syntax and have no particular problem with syntax working more or less like BNF, though I prefer slightly different approaches of a more dynamic sort. I guess my problem is that I have no idea of what you think syntax is like, how you would state the rules, what matrix you would embed them in, and so on. A short course would be helpful here, I think -- for everyone involved.
*From:* And Rosta <and.rosta@hidden.email>
*To:* engelang@yahoogroups.com
*Sent:* Tuesday, October 9, 2012 4:57 PM
*Subject:* Re: [engelang] reformulating the core grammar
John Cowan, On 09/10/2012 18:43:
> And Rosta scripsit:
>
>> However, we may mean different things by FOPL, especially in the
>> context of loglanging. I understand it to be the minimal syntax of
>> propositional meaning. I expect others understand it to be a fairly
>> standardized combinatorics of symbols supplemented by a body of
>> interpretation rules. These different understandings may well give
>> different answers in edge cases such as the one under discussion.
>
> In addition, as has been painfully established, what you mean by "syntax"
> is not quite what the rest of us mean by the term.
Yes, I have a syntactician's understanding of "syntax" and most of my colleagues here appear to have a computer scientist's understanding of "syntax". I concede that because we were talking about language, I was simply assuming that the relevant understanding was the syntactician's, but I was probably being presumptuous.
As with understandings of FOPL, I guess each has its rationale and motives.
I presume (if I may venture to try to think myself into Mike's POV) the attraction of standardized FOPL and computer science "syntax" and model-theoretic semantics, is that most of this stuff is well-understood, ready-formalized, and those who favour it want a speakable version of it, and maybe it doesn't matter if the formal rules don't themselves constitute human-usable rules (so long as speakers can induce some tacit human-usable rules).
I, on the one hand, want to make something that unambiguously encodes logical form (the minimal syntax of propositional meaning) and is as ergonomically usable as possible, and is a language (which means at least by default giving it the mechanics and architecture of language).
> (I note that not all linguists agree with the autonomous-syntax view
> you have been presenting as How Language Works.)
What do you mean by autonomous syntax? I'm often unclear what people mean by it, but it seems to boil down to whether syntax has ('constructional' meaning) or whether all semantically interpretable meaning comes from insertion of lexical items (-- the autonomist view). I happen to be fine with there being constructional meaning, so am not autonomist, but anyway, (i) on autonomist approaches there are ways to approximate constructional meaning, (ii) Generative Grammar, which is the bastion of autonomism, sees logical form as part of syntax (-- which looks like a kind of constructional meaning), and (iii) I don't think any of this is relevant to a loglang (certainly not at this inchoate stage).
As for syntacticians' view of syntax, I am not reporting anything controversial or nonmainstream, I believe. Or rather, I take it as a given, from our agreed goals and discussion, that we don't require phonology--syntax homomorphism, because we want the phonology to be as brief as possible -- if one phonological segment could stand for a whole phrase of a dozen syntactic words, we'd welcome that -- and we all agree that parts of logical form can be implicit. With that given, the rest of what I say is normal.
No syntactician that I know of has had any time for anything like our BNF rules since the 1950s, when syntax was barely at the toddler stage. In my time I've talked to a couple of descriptive grammarians who assumed syntax worked like that, but they were born around 1920 and didn't actually work on syntax.
--And.