ELTE SEAS

Discussion Forum
It is currently Sat Nov 18, 2017 10:27 pm

All times are UTC + 1 hour




Post new topic Reply to topic  [ 30 posts ]  Go to page 1, 2  Next
Author Message
 Post subject: Re: ANN-341.23 VC phonology
PostPosted: Sun Dec 23, 2007 3:31 am 
Offline

Joined: Fri Apr 27, 2007 9:37 pm
Posts: 17
I'm gonna think this thing over in the break and stand forward with a bulletproof proposal in the spring (if I finish my paper on aphasics). In the meantime, happy holidays to You and Your family.

_________________
Prince & Smolensky 2008 - Optimalizing America!


Top
 Profile  
 
 Post subject:
PostPosted: Sat Dec 22, 2007 11:02 pm 
Offline

Joined: Wed Nov 10, 2004 6:15 pm
Posts: 808
huncut felhocske wrote:
In a theory where everybody has the same number of elements and sonority is not encoded in complexity the stop paradox would vanish
I'm not so sure of this. The stop paradox is not about sonority, it is about lenition trajectories and markedness/complexity. One could separate markedness and complexity, but this entails an undesirable separation of the two types of markedness, empirical and representational.
huncut felhocske wrote:
I've been thinking about this: an obvious choice is van der Hulst's Radical CV, problem is he has Vc (V=>C) "major class"* as approximants. (Otherwise the top level of the RCV representation could be the skeleton itself) The Gordian knot is cut if we give a V(i.e. Vc) to approximants on the skeleton, in which case branching onsets would vanish - would you think this possible? (*actually it's a stricture value, but the point is the same)
I don't think I understand your story at this point. As much as I do, it sounds similar to what Sašo Živanović is saying.

_________________
szp


Top
 Profile  
 
 Post subject:
PostPosted: Fri Dec 21, 2007 3:02 pm 
Offline

Joined: Fri Apr 27, 2007 9:37 pm
Posts: 17
As a now acknowledged authority on the subject I might add that a j or a w (or a r/l, for that matter) roaming around governing things would not be fortunate - but we could say that their little C bereaves them of their vocalic power (the argument of course wouldn't go in this upside-down, "association lines do not cross" manner). Something of the like could be conjured up for affricates, which are Cv in RCV.

_________________
Prince & Smolensky 2008 - Optimalizing America!


Top
 Profile  
 
 Post subject:
PostPosted: Fri Dec 21, 2007 2:48 pm 
Offline

Joined: Fri Apr 27, 2007 9:37 pm
Posts: 17
Complicating things to the utmost my learned colleague Mr. Szeredi measured vowel reduction in Hungarian and found some context (a TVR sequence if I remember well, but I'm quite unsure about its exact nature) where vowels are especially susceptible to become schwas.

Touching codas, and the issue of V2V gov. or C2C licensing in the case of VCC in general, this seems to be mostly decided by the sonority sequencing of the consonants - yet your theory won't say what nature this sequencing might be as the melody part of it is, forgive me for saying so, in shambles (not that ET in general isn't - like Scheer has something to say about this and the stop paradox, but only in exchange to having this awkward setup of elements with B-s and crossed I-s and having more elements in sonorants and infrasegmental government and whatnot).

In a theory where everybody has the same number of elements and sonority is not encoded in complexity the stop paradox would vanish - I've been thinking about this: an obvious choice is van der Hulst's Radical CV, problem is he has Vc (V=>C) "major class"* as approximants. (Otherwise the top level of the RCV representation could be the skeleton itself) The Gordian knot is cut if we give a V(i.e. Vc) to approximants on the skeleton, in which case branching onsets would vanish - would you think this possible? (*actually it's a stricture value, but the point is the same)

_________________
Prince & Smolensky 2008 - Optimalizing America!


Top
 Profile  
 
 Post subject:
PostPosted: Fri Dec 21, 2007 2:08 pm 
Offline

Joined: Wed Nov 10, 2004 6:15 pm
Posts: 808
Note that vowel strength is dependent not only on stress: a vowel can reduce to zero (syncopate) only in open syllables, it can reduce to schwa only before a ``light'' coda, and before a ``heavy'' coda it will remain full (e.g., the underlined vowel in October cannot reduce to schwa). (Before you ask, I am not prepared to define the difference between light and heavy codas.)

_________________
szp


Top
 Profile  
 
 Post subject:
PostPosted: Thu Dec 20, 2007 9:36 pm 
Offline

Joined: Wed Nov 10, 2004 6:15 pm
Posts: 808
huncut felhocske wrote:
What would VCP say to vowel reduction?
VCP is a theory of consonant lenition and phonotactics. It is indeed missing the insight of Licensing Inheritance wrt vowels.
huncut felhocske wrote:
but stress is not playing any role otherwise
is stress playing any further role in lenition empirically? If yes, the relevant observations have to be incorporated into the theory. If they cannot, the theory fails.

_________________
szp


Top
 Profile  
 
 Post subject: nem értem #5
PostPosted: Thu Dec 20, 2007 4:00 pm 
Offline

Joined: Fri Apr 27, 2007 9:37 pm
Posts: 17
What would VCP say to vowel reduction? As V-s aren't affected by licensing or guvmint that way (well, they are, but government seems to just turn them off and licensing as far as I can see only connects long vowels or diphtongs) - there is an anti-penetration constraint stopping unwelcome lenition but stress is not playing any role otherwise or am I wrong here?

_________________
Prince & Smolensky 2008 - Optimalizing America!


Top
 Profile  
 
 Post subject:
PostPosted: Tue Dec 18, 2007 12:21 am 
Offline

Joined: Wed Nov 10, 2004 6:15 pm
Posts: 808
It is Harris who has to restrict the domain of licensing paths to the foot: i.e., whether the head of that foot is ``primarily'' or ``secondarily'' (or ``tertiarily'') stressed is immaterial.

_________________
szp


Top
 Profile  
 
 Post subject: nem értem #4
PostPosted: Mon Dec 17, 2007 10:14 pm 
Offline

Joined: Fri Apr 27, 2007 9:37 pm
Posts: 17
The VC frame does not handle domains of primary and secondary stress differently, whereas the Harris Lic Inh model does (more arrows!). This can be a problem for the former if consonants tended to behave differently in the domains of secondary stress as opposed to primary stress. Is such a phenomenon known? (Not that words large enough to illustrate Harris's powerhouse would swarm in the world's languages.)

I would like to add a short poem by one of my colleagues as well: it is useful to show to ignorant foreigners who blur the distinction between CV and VC phonologists (though it wasn't written with such a purpose).

leszögezi ez epilógus,
nem vagy cv-fonológus,
c-k és v-k rád nem hatnak,
egymás közt kormányozgatnak,
bocsásd meg a vádat hát,
s élj vidáman, jóbarát!

_________________
Prince & Smolensky 2008 - Optimalizing America!


Top
 Profile  
 
 Post subject:
PostPosted: Mon Dec 17, 2007 12:37 pm 
Offline

Joined: Fri Apr 27, 2007 9:37 pm
Posts: 17
I agree. Monosegmentality has POTENTIAL.

_________________
Prince & Smolensky 2008 - Optimalizing America!


Top
 Profile  
 
 Post subject: Re: nem értem #3
PostPosted: Sun Dec 16, 2007 11:40 pm 
Offline

Joined: Wed Nov 10, 2004 6:15 pm
Posts: 808
huncut felhocske wrote:
if we explain CC# as consonants burying the vowel our souls shallow due to use of metaphors, I mean we'll have a falling sonority slope in a burying domain in the word /tabl/ - which would cry for syncope. (Though admittedly the majority of CC#-s will be handled by this approach.)
Sure, C-to-C government can explain why CC# is typically the same type of cluster that you call coda--onset in GP, which it is. /tabl/ is a pain in the ass. Monosegmental?

_________________
szp


Top
 Profile  
 
 Post subject:
PostPosted: Sun Dec 16, 2007 10:01 pm 
Offline

Joined: Fri Apr 27, 2007 9:37 pm
Posts: 17
In the meantime, I've summarized all the compulsory readings to the VC phonology lecture since I had nothing better to do. I thought I might as well share the material with the public. Nevertheless, I'd strongly advise anybody against using them for preparation. They are sketchy, haphazard, biased, and generally worthless. In any case, they are here:

http://www.soksokkep.fw.hu/vaudeville_cakewalking.pdf

_________________
Prince & Smolensky 2008 - Optimalizing America!


Top
 Profile  
 
 Post subject: nem értem #3
PostPosted: Sun Dec 16, 2007 9:11 pm 
Offline

Joined: Fri Apr 27, 2007 9:37 pm
Posts: 17
Thank you for your previous answers.

What I meant was the notion of segments in general, neatly detached from the others. Then again, your explanation of prints with the spreading meathook* is convincing. (Although it would be hard to pinpoint when it spread when comparing a speaker with another. This is like Trubetzkoy's dialectal maps where he argues that the boundaries of a phonemic difference between dialects is clear-cut. The same way a difference between the prince=prints and the prince/=prints people would be clear-cut, an association line here or there. Hesitant answers or variation in the perception of an individual speaker could not be accounted for.)

My nouvelle question concerns VC phonology (as a matter of fact, you do have the possibility of going silent at any point): if we explain CC# as consonants burying the vowel our souls shallow due to use of metaphors, I mean we'll have a falling sonority slope in a burying domain in the word /tabl/ - which would cry for syncope. (Though admittedly the majority of CC#-s will be handled by this approach.)

*Though I like the sound of it, I suppose meathook is derogatory, however, I'd welcome the appearance of insulting theoretical terms as the symptom of the broadening of the scientific discourse.

_________________
Prince & Smolensky 2008 - Optimalizing America!


Top
 Profile  
 
 Post subject:
PostPosted: Sun Dec 16, 2007 8:43 pm 
Offline

Joined: Wed Nov 10, 2004 6:15 pm
Posts: 808
For one thing, Lowenstamm does not need to assume nonlocal government to get #TR: if TR is one segment, it acts like T in word-initial position. The simplification of Tr in Greek reduplication neatly parallels that of Th, so that's not a real counterargument.

Ls theory is then better in that we do not need to produce fancy explanations for the behaviour of TR clusters, if they are not clusters.

I don't know what they would do with it, I would say [ns] > [nts] involves the passing of ? form the coda to the onset. [ns] is a marked cluster, we either get strengthening of the following fricative (intfluence, empfasis, etc.), or the ? is lost from the coda, yielding a nasalized vowel (Hu eleg~aS).

I guess you mean coarticulate segments or segments with secondary articulations. I agree, some branching onsets are difficult to imagine as doubly articulated segments.

_________________
szp


Top
 Profile  
 
 Post subject: nem értem #2
PostPosted: Sun Dec 16, 2007 8:22 pm 
Offline

Joined: Fri Apr 27, 2007 9:37 pm
Posts: 17
Lowenstamm (2003) on the unsplitability of ostensible branching onsets (p9 onwards)

M. Lowenstamm daintily demonstrates that segments sometimes act as if there were two of them (in a branching onset) and sometimes as if they were only one (a segment with a secondary articulation). In the Greek example he goes further to show that not as if it were bad enough that we need to store TR and Tr separately in the phonemic inventory of a language, we cannot even separate them clearly, as not only do they collapse into the same thing in the surface (if we don't start to measure them, that is) but also sometimes the Tr gets simplified (as in the Greek reduplication example), and sometimes the TR does (as wherever else).

In what respect is the Lowenstammian approach (breeding the phonemic inventory) better than the Scheerite one (infrasegmental government)?

A bit more far-fetched question: what would the above authors do with the thing going on in English that prince /prin(little t)s/ and prints /prints/ are getting to be homophones? - Lowenstamm is aware of the problem in Czech but he is pretty much mute about it. Is not the question hovering in the reader's mind while reading his article not the one asked by Lowenstamm ("Do onsets ever branch?") but rather whether the notion of articulate segments should be discarded?

_________________
Prince & Smolensky 2008 - Optimalizing America!


Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 30 posts ]  Go to page 1, 2  Next

All times are UTC + 1 hour


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group