Formal Linguistics

>>8427864
yes. Can we talk about the similarities between baltic languages and the dead Language: Dalmatian.

>>8427864
I have a question, Mr SHOA-msky, Sir. Yes, you see, I have this conundrum, that has been, as you say, keeping me up all night, ehehe, if you understand heheheh.

ANYWHO. the question, dear Rabbi, is that, if grammar is genetically imprinted into the human brain (as you insist it is) instead of being learned during childhood.. heehe, well then, how come there still be monkeys?

czech m8 Professor Left-winged-intellectual

>linguistics
>science

>>8427984
prove your point

What's the deal with Chomsky? Is he a legit scientist or a new-age woo peddler? For years I thought I had two Chomskies mixed up or something.

>>8427864
Yes, formal linguistics is great. There are two unfortunate things though.
1. almost nobody outside of linguistics knows what formal linguistics is about.
2. linguistics departments don't just employ people who work on phonetics, phonology, syntax, and semantics, they also have tons of people who would probably be better in a department for gender studies, african studies, etc.

>>8427984
depends what you mean by science. though it does not enjoy the same degree of consensus that the physical sciences have, the approach in formal linguistics is scientific, at least in the so-called generative tradition. the object of inquiry for formal linguistics is the human language faculty, not any particular language or languages. the main question that guides research in formal linguistics is what the language faculty is as a formal computational system. that leads to questions about how much (if any) of that system is innate, how and how recently it evolved, as well as questions about the actual contents of the system.

Currently writing my masters on Pirahã syntax, I made a thread on /his/ because I thought I'd get bullied here, thread went p well. There are probably other 5 linguisticsfags lurking 4chan.

>>8428359
Where'd you get your data from?

>>8428363
My own fieldwork, I'm at Unicamp (same Brazilian university Everett went for his Masters) so I get funds to pay for my field trips and all that it's p good

>>8428373
that's sick. so did everett get it right, do you think there's no embedding?

>>8427864
Gnome.

>>8428373
Olá. :)

>>8428381
I honestly think Everett's assumption about Pirahã not being recursive (as all other human natural languages) is bullshit

>>8429282
Ok. I've suspected that for a while but I don't really know anything about piraha syntax outside of the paper by nevins and pesetsky responding to everett's claim.

>>8428359
Are you going to go for a phd after your masters?

>>8428065
what makes you think he's pedalling woo? he's a scientist who does science, and also a political activist.

Should I study Latin

>>8428359
How much Piraha do you know? Could you buy bread in a shop if you went to Pirahaland, for instance?

I was flying to Italy from the states last year and got seated next to a PhD in linguistics. He was such a swift guy and we talked for hours while getting drunk off shitty, complimentary wine. I knew virtually nothing about the field but it's incredibly logical and I would consider it a science at this point. We got to talking about trends that are consistent among all spoken languages, and talked about deriving a crude formula for understanding sentence structure/syntax in alien languages (not necessarily ET languages, but ones with no known meaning).

The fact that there are emergent consistencies in spoken language, regardless of geographical location , is a testament to the processing methods of the human mind.

what do you guys think about cartography stuff, like finely articulated functional hierarchies? it seems to me like it's a little far-reaching

File: 12515.png (2MB, 1920x1080px) Image search: [Google]

2MB, 1920x1080px

Is linguistics worth self studying? It's interesting, but I don't want to take it as a minor because of an obligatory seminar where I'd have to have a presentation in front of people.

>>8428314
This.
Also, the situation in my uni (Unicamp) is that the department is divided between like 70% discourse analysts and gender studies hacks and 30% formal linguists and actually competent researchers.

>>8428381
I believe Everett got some things right, and he deserves respect for his huge contributions in the field. He also started a decent debate in linguistics by meming super hard against chomsky. I'm on the generativist side, I don't buy most of Everett's arguments, particularly his whole "everything is parataxis" strategy, feels cheap.

>>8429657
My thesis advisor (don't know if this is the correct word, it's like, the professor who helps me write my dissertation and stuff) wrote some articles last year with Rodrigues about Pirahã syntax, it'll be published by Cambridge Press next year I think. There's a lot of new things being discovered about Pirahã by like, 5 researchers (including me), it's been fun.

>>8429734
They don't really use money between themselves, but I can ask for stuff.

>>8429744
Some of the most talented linguists I've met were italian, there are some amazing guys over there. Chances you met one of them.

>>8430216
You should give Chomsky's original paper on Skinner's theory of language a go, it's a nice introduction to contemporary linguistics. If you enjoy it, go for syntactic structures and afterwards pick a language you like and read papers about it. If you feel you need more updated syntax theories look for some decent handbooks on minimalist syntax (chomsky's original paper on the minimalist program is a clusterfuck tho, but you can try it if you're feeling adventurous). Chomsky, Hauser & Fitch is a must read too.

File: 1459638396783.jpg (743KB, 1000x1300px) Image search: [Google]

743KB, 1000x1300px

>>8430675
Thanks for the response. I saved those names and will see if I end up reading what you recommended some day.

>>8427864
I've seen threads on /his/ that have turned out pretty gud

If anyone wants to read my original thread on /his/ I've answered a bunch of question there, http://archived.moe/his/thread/1787152/#q1787152

[DP that [NP feel [CP when [NegP no NP GF]]]]

>>8430675
what do you think about the chomsky hauser & fitch view of language evolution? even among generative syntacticians who generally follow chomsky's stuff it's really controversial.

>>8428359
hey, do these people have skin problems like acne?

>>8427864
I'm a mathfag and don't really know much about linguistics, but this sounds interesting. Any recommended reads?

>>8427864
https://www.youtube.com/watch?v=THDNsB-vtFo

https://www.youtube.com/watch?v=P78SJf8NL2k

two cool fables of reconstructed proto-indo-european.

>>8427864
Very, but I don't have the requisite knowledge to discuss this subject in any real depth.

>>8431773
There's a pretty cool paper by Collins & Stabler that just came out this year in the journal Syntax called "a formalization of minimalist syntax." If you have access to Syntax through a library you can get it online

abstract: The goal of this paper is to give a precise, formal account of certain fundamental notions in minimalist syntax. Particular attention is given to the comparison of token-based (multidominance) and chain-based perspectives on Merge. After considering a version of Transfer that violates the No-Tampering Condition (NTC), we sketch an alternative, NTC-compliant version.

>>8430712
>http://archived.moe/his/thread/1787152/#q1787152
this is great. you could do a media tour with this stuff

>>8432166
haha I am using C&S in my thesis to show that they are effectively admitting that set theory by itself is not going to cut it without a bunch of other assumptions about the grammar.

Chomsky is a smart guy, and a lot of MP is probably even on the right track, but BPS is just not a very robust mathematical framework, and I think if you read C&S closely, you'll find that they have to invent idiosyncratic conversions of sets into other objects of particular kinds and use extra annotation to accomplish all sorts of goals. It is highly nontrivial to remove those annotations and idiosyncratic conversions, which is something left undiscussed.

>>8432297
>set theory by itself is not going to cut if without a bunch of other assumptions about the grammar
For sure. I think the guiding principles of the theory are right though. If I had to hazard a guess, my intuition is that the minimalist treatment of the lexicon is too reductionist.

>>8430712
If you're still around could you post more Pirahã jokes?

>>8428065
Possibly confusing him with Deepak Chopra? Chopra is the woo-peddling "quantum mysticism" guy; Noam Chomsky is pretty left-wing politically, but doesn't peddle pseudoscience. He's done good work in linguistics and cognitive science, but obviously opinions will vary as to the rest of his work in the humanities.

>>8432400
Yh. I think the guiding principles are right, but that the amount of other stuff you have to say is just guiding us towards the fact that the set-theoretic representation Chomsky has cooked up is entirely a red herring. In fact, C&S are sort of accidentally showing us that set theory is not the right tool for describing the properties Chomsky talks about (eg note that they have to annotate which features are checked at each stage off to the side. See if you can even state the extension condition if you didn't do that, but rather updated features in the syntactic objects themselves. You'll find its really tricky, probably impossible without adding gizmos).

But, I think contra PoP, Chomsky is wrong about trees/dependencies being the red herring, and sets being correct. I think C&S actually shows the opposite. The derivation procedure and the interactions of features and the syntactic object as some sort of structured object like a tree or graph are way more relevant and useful properties of grammar. It's even more the direction Stabler has gone himself.

>>8432578
Maybe. The whole set-theory move seems to me to be motivated by the fact that Chomsky's conception of the lexicon is pretty set-theory-friendly. A few things do fall out of it (eg displacement/trace theory), but the way the derivations end up definitely doesn't have the aesthetic appeal of GB, and it doesn't yet have as wide empirical coverage. I think allowing the syntactic representations to have more intrinsic properties could resolve some of the unwelcome complexity and would certainly make it easier to cover more data, but it would also require a blurred grammar-lexicon distinction, otherwise you just end up relocating the extra assumptions and annotations to another part of the grammar.

File: syntax_tree.png (4KB, 205x516px) Image search: [Google]

4KB, 205x516px

>>8430910

>>8432719
you mean it "falls out" assuming that we

a) have first indexed all of our lexical items prior to insertion (if we did not, a sentence might have one "masculine phi feature" which is simultaneously in a bunch of different lexical items)

b) never change features within the syntactic objects at all/or alternatively change them in such a way which affects all "occurrences" (then, C&S's label trick is a no-go)

c) reconstruct the different positions like C&S do with paths

Effectively, it only "falls out" insofar as you decide to treat a set as indicating a multidominant graph, and choose conventions about how multidominance will arise (and instate indices to make sure it doesn't where you don't want it to) and how to interpret it as constituent-with-copy data.

>>8428065
He spent his youth and middle age developing brilliant and field-defining linguistic theories. Then he got old and became a self-flaggelation machine blaming the West for everything savages do in third world countries.

>>8428065
Chomsky is a kike so everything he says is garbage

>>8434367
That makes sense, but how can we get away with, say, not specifying the position of a syntactic object by a root-to-label path? It seems like even if we use directed tree structures we would have to specify the position of vertices in about the same way. As for indexing the items in a lexical array, what alternative do you have in mind? Could you do away with lexical arrays entirely?

>>8434421
I'm more saying that if it's these relative dependency/positional/or phrase containment properties we are interested in, we should axiomatise THAT and study THAT (graphs of dependencies, partial orders of constituents by containment, a derivational history complete with showing the operations that mapped input phrases to output phrases, whatever), not some auxiliary object which we need extra conventions for just to make sense of it. that's just bad application of occam's razor (i could also gödel number every lexical item and phrase in the language according to some convention, it just wouldn't be very useful for studying language). i think the set theoretic version of BPS is like that, or trying to write in ASCII even though we really cared about symbols. the set theoretic object alone doesn't even have the properties we want it to have - so why not just study an object that does?

conjectures about evolution and computational theory of mind aside- the formalism is there to help us reason about language and it's properties, so handicapping yourself with ill-behaved notation that requires a bunch of auxiliary parts to get off the ground doesn't do much to help you in that regard.

e.g. trees were nice mathematical objects. they are easy to prove things about, and understanding abstract structure of trees does map on to corresponding linguistic notions (like the fact that binary ordered trees can be put in correspondence with a free binary algebra, and stuff like that).