Lexical diffusion and phonemicity

Discussion of natural languages, or language in general.
Post Reply
Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Lexical diffusion and phonemicity

Post by Travis B. »

One thing I have noticed is that with regard to flap elision (i.e. elision of unstressed intervocalic /t d n nt nd/) in the dialect here is that it is very much subject to lexical diffusion - some words are more likely to have it that other words, with some words almost always having it, and some words almost never having it - and the exact lexical distribution depends heavily on the individual speaker (e.g. I have it in many words, yet many people only have it in a small set of very common words). Yet at the same time, this presents a problem because flap elision combined with cheshirization has the consequence that flap elision requires phonemic vowel length and phonemic vowel nasality while without flap elision neither vowel length nor vowel nasality are phonemic. So in essence there are two coexisting vowel phonologies, which depend upon the word in question and the particular individual speaking. But how can this be - should there not be one phonology for all words aside from interjections? So how do we resolve this?
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

Sumelic
Avisaru
Avisaru
Posts: 385
Joined: Sat Mar 28, 2015 7:05 pm

Re: Lexical diffusion and phonemicity

Post by Sumelic »

Travis B. wrote:One thing I have noticed is that with regard to flap elision (i.e. elision of unstressed intervocalic /t d n nt nd/) in the dialect here is that it is very much subject to lexical diffusion - some words are more likely to have it that other words, with some words almost always having it, and some words almost never having it - and the exact lexical distribution depends heavily on the individual speaker (e.g. I have it in many words, yet many people only have it in a small set of very common words). Yet at the same time, this presents a problem because flap elision combined with cheshirization has the consequence that flap elision requires phonemic vowel length and phonemic vowel nasality while without flap elision neither vowel length nor vowel nasality are phonemic. So in essence there are two coexisting vowel phonologies, which depend upon the word in question and the particular individual speaking. But how can this be - should there not be one phonology for all words aside from interjections? So how do we resolve this?
I'm not sure whether the variation that you describe has to be explained as a consequence of lexicalization of a phonological contrast. That would be a more convincing scenario if the lexical distribution didn't depend so heavily on the individual speaker—it seems more likely that speakers would vary in application of synchronic lenition processes than that they would vary in the distribution of lexically stored, phonemicized variants. So I would lean towards assuming that the elision is not lexicalized, and what appears to be lexical diffusion may in fact have other synchronic explanations, like differences between speakers in tendencies to elide based on word frequency (perhaps), phonetic environment, that kind of thing.

I found a paper that has a postscript that seems like it might be relevant to your second-to-last question: https://web.stanford.edu/~kiparsky/Papers/labov.pdf

Travis B.
Sumerul
Sumerul
Posts: 3570
Joined: Mon Jun 20, 2005 12:47 pm
Location: Milwaukee, US

Re: Lexical diffusion and phonemicity

Post by Travis B. »

The only real solution I have come by for my dialect is that the consonants are never elided in the underlying form, but rather are only elided in the surface forms. But this is unsatisfactory because there are many words for which I elide about 95% of the time, and only do not elide when I am deliberately speaking carefully. Also, this only works because of the nature of this phenomenon in my dialect; different cases involving sound change and lexical diffusion in other language varieties may not be amenable to this kind of solution.
Dibotahamdn duthma jallni agaynni ra hgitn lakrhmi.
Amuhawr jalla vowa vta hlakrhi hdm duthmi xaja.
Irdro. Irdro. Irdro. Irdro. Irdro. Irdro. Irdro.

Sumelic
Avisaru
Avisaru
Posts: 385
Joined: Sat Mar 28, 2015 7:05 pm

Re: Lexical diffusion and phonemicity

Post by Sumelic »

Travis B. wrote:The only real solution I have come by for my dialect is that the consonants are never elided in the underlying form, but rather are only elided in the surface forms. But this is unsatisfactory because there are many words for which I elide about 95% of the time, and only do not elide when I am deliberately speaking carefully. Also, this only works because of the nature of this phenomenon in my dialect; different cases involving sound change and lexical diffusion in other language varieties may not be amenable to this kind of solution.
Well, I'm not sure the fact that elision occurs around 95% of the time necessarily makes it unsatisfactory to analyze the words as still having consonants in an underlying form at some level. As for other varieties, it's possible that they might pose challenges, but maybe not.

One comparison that came to my mind is lenition of intervocalic obstruents in Western Romance; from what I understand, the elision of intervocalic /d/ [ð] in present-day Spanish (mainly from Latin intervocalic singleton /t/), is somewhat variable and seems to affect some specific words more than others (sometimes /b/ and /g/ can also be elided), and in French intervocalic /v/ (from Latin /b/, /p/ and /w/) can be elided in some contexts (I don't know the details), but I don't know of any Romance language that shows frozen, lexicalized diffusion effects for lenition processes like this (all the cases I am aware of where there is apparent inconsistency in the reflexes seem like they can be explained by analogy and borrowing between dialects, not word-by-word diffusion of a sound change within a single dialect). This example at least suggests to me that such variability isn't likely to develop into an actual dichotomous phonemic split.

Then again, my impression that mixed results are mainly due to dialect mixing might be unfounded; lexical diffusion has apparently been proposed as an explanation of unexpected voiced reflexes in some Italian words: https://books.google.com/books?id=XZAMH ... on&f=false

Post Reply