Print

Print


On 2 December 2012 15:13, Alex Fink <[log in to unmask]> wrote:
> On Sat, 1 Dec 2012 10:20:15 -0700, Logan Kearsley <[log in to unmask]> wrote:
>
>>On 30 November 2012 00:07, Alex Fink <[log in to unmask]> wrote:
>>> On Thu, 29 Nov 2012 01:42:13 -0600, George Corley <[log in to unmask]> wrote:
>>>
>>>>I have heard
>>>>that some syntacticians believe there are maybe a half dozen types of
>>>>adverbs that can have different syntactic restrictions, though AFAIK that
>>>>theory requires you to go down a particular generative rabbit hole that I'm
>>>>not sure I want to mess with myself.
>>>
>>> I should like to hear a bit more of this, if only to judge how moored it is to the world outside the rabbit hole!  But there are definitely some ways we should break down the adverb: as Stevo points out, disjuncts should well be something entirely separate.
>>
>>Disjuncts could be very fun if there are ways of constructing complex
>>ones, a la relative clauses.  If adjectives and adverbs are treated
>>separately, one could have a total of four different modifier classes
>>(or more with more different kinds of adverbials split out);
>>otherwise, just adjunctive vs. disjunctive modifiers. I would expect
>>disjunctive and adjunctive adjectives to end up grouped differently in
>>a configurational language- e.g., all disjunctives come first, then
>>all adjunctives, or adjunctives before a noun and disjunctives after.
>>Disjunctive relative clauses would provide the restrictive vs.
>>non-restrictive distinction that we have in English relative clauses.
>>And compound adverbials could be formed with some kind of "like x"
>>structure, with different kinds of "like" to indicate a disjunctive
>>phrase vs. an adjunctive phrase.
>
> Hm.  I do like the idea of complex disjuncts.  But to me it seems the case for establishing these various classes as distinct parts of speech, rather than just different syntactic constituents, would be best made by the underived forms in them.

Oh yes, certainly. "Heavy" constituents can do all sorts of weird
things to complicate proper analysis. I just think, having introduced
disjunctive modifiers as separate parts of speech, it would be cool if
there were complex forms of them, as well.

> I would expect a disjunct modifying a noun to be something like "I'm going to meet the hopeful president", with "hopeful" read as 'who I hope will be'.

That doesn't feel like a disjunct to me. According to The Fount of All
Knowledge, "a disjunct is a type of adverbial adjunct that expresses
information that is not considered essential to the sentence it
appears in". We can ignore the "adverbial" bit since we're making up
new parts of speech, but "hopeful" in that sentence is definitely
essential to the meaning of the sentence.

> I guess I can sort of see how to interpret it as a restrictive vs. nonrestrictive distinction, but I'd be inclined to take the 'commentary on the described state of affairs' function to be more central.  (Or is this a clever way to increase the distinctiveness of the classes that I haven't cottoned onto?)

I don't see the difference. "Commentary on the described state" is
exactly how I use non-restrictives.

> On the point of splitting adverbs from adjectives, my thoughts are that, at least in an SAE-esque language, the underived adverbs do fall into some heterogeneous classes of the sort which would be conducive to further splitting.  For instance, time adverbs could split off, and their distinctiveness maintained by folding them into the tense system.

What I would do with temporals and the tense system is make temporals
obligatory as the means of tense marking. Maybe requiring some kind of
agreement with verbal inflection for tense anyway. These would be
distinguished from run-of-the-mill tense marking particles by being an
open class that can be added to by regular derivational processes. So,
there might some "default" simple temporals with meanings like "now",
"before", "in a while", etc., but then one could also have
derivational processes that turn nouns into temporals with meanings
like "during the time of"; names of royalty might be derived to
produce temporals that correspond to "during the reign of" or some
such thing. Intransitive verbs might be derived to indicate "at the
same time", "after  doing this", "before doing this", etc., while
there might also be a means of producing relative temporals that allow
an entire clause describing a time period to fill that syntactic slot.
Temporals would be further distinguished by demonstrating that other
adverbs are disallowed from filling that particular role, and that
some adverbial must exist in every clause (or maybe just in every
top-level clause).

>>On the other hand, one could make an
>>argument that proper nouns really are a different class than regular
>>nouns because of, e.g., not being able to take adjectives, and run
>>with that. Now this is reminding me of my sketch of a language that
>>requires tons of voice transformations because of proper nouns not
>>being allowed to occur in any syntactic position except subject.
>
> Hm, what else could proper names do distinctly?


> On Fri, 30 Nov 2012 21:49:28 -0800, Leland Paul Kusmer <[log in to unmask]> wrote:
>
>>On Thu, Nov 29, 2012 at 11:07 PM, Alex Fink <[log in to unmask]> wrote:
>>>
>>> Hey, wait a minute, that physical / mental split of the verb class is what
>>> Elkarîl does!
>>>   http://zompist.com/elkaril.htm#Case
>>> So, I take it back, there's at least one language that has gone for
>>> weirdness in this fashion.  And delightfully so.
>>
>>Ooh – this is lovely. My personal aesthetics make me want to see what would
>>happen if you made this language more inflecting and then allowed word
>>order variations for information structure. Or, more generally, gave
>>significantly different morphology to the four classes (two verb-like, two
>>noun-like) outlined in that section. Mental verbs, i.e., might not inflect
>>for tense. Or, hey, you could have evidential systems on both, but
>>different systems – given that the mental verbs would be unlikely to be
>>direct-observation knowable, anyway.
>
> I hadn't gotten as far as thinking about information structure and word order variations, but similar thoughts had already been percolating in my mind about the differing morphology.
>
> As I've said above, I am taken by this system of positionals = your depictives, with the other physical verbs subordinate to them.  (A fragment of a kinda mimetic system of positionals, for static arrangements, came to me already; perhaps I'll post that.)
> Regarding tense, I had the idea that for positionals one could try to do a bit of unification of space-time and thereby fold the tense marking in, along the lines of the way andatives often grammaticalise into futures.  You know, "dog is on a trajectory towards fish, and eats it" [with reference point here-and-now] for 'dog will eat fish'.  If one adopted that, mentals could by contrast have a more conventional tense system (even if just a past vs. present).

Celimine folds position & time into a single space-time category, but
the inflections are not very complicated, basically boiling down to
"here and now", "either not here or not now", and "not real in any
case". It doesn't really produce any new parts of speech.

>>Depictives that have no immediately obvious physical interpretation often
>>metaphorically take on aspectual interpretations.
>>
>>6) distributed.stochastically.through.volume me
>>"I'm scatter-brained."
>
> How's that aspectual?
>
>>Depictives are the one syntactically necessary category in every sentence,
>>and select either a noun or a verb phrase as argument. (6) is thus a
>>minimal sentence in tükwäi.
>>
>>(7) me-TOP horizontal.to.goal-Pt teacher
>>"I'm becoming a teacher." (or "became" – unmarked for tense)
>>
>>An example of a nominal predicate. The marking on the depictive can change
>>interpretation – if it were marked agent, this might be something more like
>>"I'm trying to become a teacher." It wouldn't be "I'm going towards a
>>teacher" – that requires a verb.
>
> Oh, that answers my question above.  So I wonder how much this proliferate language should use motion metaphors.

I'm still not sure I get it 100%. And I want to, because I suspect it
would help me in getting unstuck on one particular point in revising
Celimine. Just as the tense and positional deictic systems are merged,
I wanted to find some way of expressing "spatial aspect" in
conjunction with temporal extent. But I have not yet been able to
figure out how to make that make sense.

-l.