Differences between signed languages and spoken languages

[Daniel]

This is just silly. I've stated my position a few times now, and it is of course none of these things. If that's not apparent by now, I'm not sure what more I can do.

Ok. What's your practical suggestion then? I want to investigate conjunctions in ASL and other signed languages. What's the first step?
Based on everything you've said, I can't imagine it's anything other than "ask a different question". Am I wrong? I'm not trying to be difficult, so please let me know if I am wrong.

I don't doubt that you experience what I say as as "redefining everything along the way", but the irony is that it's all fully canonical in the literature.

That's fine. I'm happy to redefine. But I don't think we should start by redefining everything-- I like to start by exploring the data.

In other words, I'm sorry to contradict your intuitions, but this is well researched, well documented stuff. If you'd like me to pass some references, I'd be happy to.

References on what? I'd love some on conjunctions (or lack thereof) in signed languages. And possibly on other topics. I'll certainly consider your suggestions.

In the meantime, I've made my point and repeated it a few times. I don't need to make it again.

I don't really understand your point. In a sentence or two, what is it? That these questions are difficult? I agree.

Are you reading my posts thoroughly? I feel like I've expressed a lot of this and that we don't really disagree, yet you still seem to think we do. Perhaps I'm misunderstanding something. Anyway, if you don't have the time to read my long posts in detail I can (try to) make my points more succinctly, or at least pick a few short items to focus on.

For one: you seem to have an inaccurate impression of my beliefs/intentions. That's probably my fault. Despite your impression of me, try to not assume I'm assuming too much. But let's talk about the data rather than what I might be assuming.

I hope you enjoy your ASL class, and I hope you stick with it long enough to grok the language.

Thanks! Honestly I doubt that I will continue to high proficiency though. I have limited time and at this point I tend to stick with a language enough to get a basic feeling for it only (a semester? a year?), because I know I won't actually reach fluency and usually I gain the most from the earliest exposure-- one year or three is about equivalent if I don't continue to use the language often. But regardless I know that I'll have some foundation from which to understand signed languages that I completely lack now. I'm looking forward to that.

[Daniel]

Alright, so I just went back through your posts to see if I missed anything (the forest while paying attention to the trees).
As I see it, you do overall seem to be saying: constituency is uncertain, so therefore analysis of coordination in SL is uncertain.
You might be right. Unfortunately that's both inconclusive and uninteresting (not that I'm claiming that's your fault).

[Daniel]

MalFet, I have to ask... what do you think of the collection of articles in this journal volume?
https://benjamins.com/#catalog/journals/sll.16.2/toc

[MalFet]

MalFet, I have to ask... what do you think of the collection of articles in this journal volume?
https://benjamins.com/#catalog/journals/sll.16.2/toc

I have no idea, I'm afraid, as I haven't read them. I know most of the authors if there's something in particular you're asking about.

[Daniel]

They're appying Generative syntax, in some very technical ways, and certainly relying on analyses/assumptions of constituency. What do you think of that? If you browse through the journal there are a lot of very specific claims made about things like the CP domain or c-command, as well.
Based on what you've said, I'm curious what you think of that extensive literature.
Is it founded on highly uncertain assumptions?
Could coordination be studied at least as well as the items in those articles?
Or was your point that coordination in particular relies on assumptions more extreme than what is made for those articles.

In short: researchers seem to be doing exactly what you suggested wasn't possible or realistic. Is that a problem?

[MalFet]

They're appying Generative syntax, in some very technical ways, and certainly relying on analyses/assumptions of constituency. What do you think of that? If you browse through the journal there are a lot of very specific claims made about things like the CP domain or c-command, as well.
Based on what you've said, I'm curious what you think of that extensive literature.
Is it founded on highly uncertain assumptions?
Could coordination be studied at least as well as the items in those articles?
Or was your point that coordination in particular relies on assumptions more extreme than what is made for those articles.

In short: researchers seem to be doing exactly what you suggested wasn't possible or realistic. Is that a problem?

Ronnie Wilbur is probably the most prolific of that bunch, and certainly she doesn't make many assumptions constituency in the way that you're suggesting. Her article here appears to be a reformulation of a talk she's been giving over the last year, and in that talk at least the problem of constituency is central. As for the rest, I'd have to read the articles.

So, yes, walking in blindly and trying to decide whether or not there *is* coordination relies on assumptions more extreme than what appears to be in this journal.

[Daniel]

How can one apply Generativism without constituency? I'm at a loss.

Perhaps those articles (which I haven't read) aren't the best examples. Here are a couple that do seem to make claims that would involve structure and constituency:
https://benjamins.com/#catalog/journals/sll.10.1.05zha/details
It discusses the "universal" about the order of D, Num, Adj and N elements. Certainly that must rely on constituency, right?
(For example, we could coordinate adjectives and then find a coordinated structure, assuming we can identify adjectives in the first place.)

Or this dissertation (announced in the journal) about passive constructions:
https://benjamins.com/#catalog/journals/sll.10.2.09han/details
Passives are a purely structural tool, given that the semantics are the same as the active sentences. So there must be a structural and I'd assume constituency-based analysis to make such an argument.


Anyway, there are others.


These researchers are applying generative models to sign language, using phrases like "CP" and "c-command". One of the only things that seems uncontroversial and reliable in generativism is constituency, and the rest basically follows from that with some work. Certainly the models require hierarchically aligned separable chunks (constituents) for formal reasons.
If they can do that, I genuinely don't understand why I can't at least attempt to look at coordination. It seems like it would be about as accurate as their results, however accurate they might be. (Of course this isn't practical for me at the moment with my lack of experience this field, and I'm not claiming otherwise.)

[MalFet]

How can one apply Generativism without constituency? I'm at a loss.

Ronnie Wilbur's work is concerned with a very particular register of sign. In the rough categorization I offered before, she is interested in lexical-sequential structures with citational stability. In her work, she gathers sets of utterances and then applies a generative analysis to them. She is ultimately interested not in typological classifications but rather the cognitive implications of speech acts.

That interest produces, as I hope is obvious, a very different kind of claim than the broadly synthetic typological ones you're describing about whether a signed language *has* conjunctions. I have many professional and theoretical disagreements with Prof. Wilbur's work, but ultimately her methodology is sound. She is doing something very different than what you're talking about, and that's very clear if you take the time to engage this literature seriously.

[Daniel]

So sometimes constituency is ok, if you look at the right data?
That's fine then. Looking at that data, are conjunctions observed?

whether a signed language *has* conjunctions

I don't understand the emphasis. I don't mean any particularly deep or exclusive sense of "have". Are, sometimes, conjunctions observed?

[MalFet]

So sometimes constituency is ok, if you look at the right data?
That's fine then. Looking at that data, are conjunctions observed?

whether a signed language *has* conjunctions

I don't understand the emphasis. I don't mean any particularly deep or exclusive sense of "have". Are, sometimes, conjunctions observed?

Certainly. My first post in this thread was a link to the ASL word glossed "and", no?

[Daniel]

Then my questions reformulated:
1. Why don't we observe conjunctions as often in signed languages as we do in spoken languages? You've probably answered this already-- because constituency is different.
2. When conjunctions are observed, has any substantial descriptive work been done? I'd be interested in that.

Of course getting a paper specifically focused on this for less described languages is rare, but here's an example of a fun one:
Wilkins, David. 1984. How, and how not, to say "and" in Mparntwe Arrernte (Aranda). Language in Central Australia. 2. 23-30
What's nice about that is that it describes strategies (eg, via translation) including ones that are formally similar (with conjunctions).

But maybe this has just not caught the attention of any signed language researcher(!) -- something to add to my "if I have a chance" list of topics

[MalFet]

Then my questions reformulated:
1. Why don't we observe conjunctions as often in signed languages as we do in spoken languages? You've probably answered this already-- because constituency is different.

Actually, no, what I've been saying again and again is precisely the opposite. Because we don't have an adequate model of constituency, we cannot meaningfully make or evaluate claims like "We don't observe conjunctions as often in signed languages as we do in spoken languages". These types of broad, typological claims are just not empirically sound.

[Daniel]

After my first week of class, a few thoughts:

1. ASL doesn't seem too difficult in many ways, and I feel somewhat comfortable with most of it (the very basics anyway).
2. Keeping up with the manual articulations is very difficult for speed, and recognizing meaning at a fast pace is challenging too. (I know we're nowhere near the full speed of fluent signers anyway, though.)
3. Studying a language that can't be written is a strange challenge, but forces me to pay attention rather than writing notes, which the instructor pointed out was a good thing with a visual language that you can't listen to while looking down and writing.

4. The most unique aspect for me is handedness, something I hadn't thought of before. I'm left handed, so I'm doing everything as the mirrored version of what others are doing. Needless to say, at least for now, that makes certain things a little confusing. Only one other student in the class is left handed, and it's very different when I do one of the practice conversations with her, so I'm sure everyone notices it with me too, haha.

[MalFet]

After my first week of class, a few thoughts:

1. ASL doesn't seem too difficult in many ways, and I feel somewhat comfortable with most of it (the very basics anyway).

Strange! I've been signing 10-20 hours a week for a decade, and I don't feel comfortable with any of it. Deixis alone is mind-bogglingly difficult, I find.

[Daniel]

I don't mean that I'm good at it. I mean that it feels generally natural to learn. It might be because I'm a visual learner. But it's certainly not difficult in the same ways that learning a spoken language is, and not more so because it's a novel medium for me. It's a challenge, and I'm not claiming anything like proficiency yet, but for the bits that do make sense it doesn't feel too hard.

(For unrelated reasons I may not be able to continue taking the class this semester, due to my schedule. But I'll try to get back to it when I can! It was a fun start. Definitely on my list.)

And for the record, I'm beginning to understand a lot of what you're talking about with the difficulty of segmentation. I'm not entirely convinced about constituency (chunks seem generally separable, at least most of the time), but I can't imagine figuring out a system of basic contrasts or a way to write it.