Re: NATLANG: Vowel harmony rules?

On 18 Jun 2004 David Peterson <ThatBlueCat@AO...> wrote:

> I just wrote my final morphology paper on just this
> phenomenon.
  It will be very hard for me to argue with you, I cannot write papers
on Turkish phonology :))).


> That aside, though, have you read anything anywhere about this
> phenomenon? If so, can you let me know?   I was trying to find something
> on this, but couldn't find anything.
  I wrote that I borrowed this idea from the Hungarian grammars. In
Hungarian this phenomenon is much more frequent. If we can apply this
also for Turkish, we can maintain a common vowel harmony modell for the
whole Ural-Altaic group.

  I do not know any references about lowering stems which is available
outside Hungary in English. I can only propose e.g. Miklós Törkenczy:
Practical Hungarian Grammar (A compact guide to the basics of Hungarian
Grammar). Published by Corvina, Hungary in 2002. ISBN 963 13 5131 9


> First of all, are you suggesting that this is phonological in some way?
> Because if so, consider:
>
> yen = "to overcome" > yenar = "to overcome (aorist)"
> yen = "to be eaten" > yenir = "to be eaten (aorist)"
  It is not phonological, I wrote "[it] is a _lexical_ feature". In
your examples there are two different lexemes "yen(1)" 'to overcome'
and "yen(2)" 'to be eaten'. "Yen(1)" has a [+lowering stem] attribute
in the lexicon and "yen(2)" has a [-lowering stem].

  I used the word "predict" when I wrote about the phonological
circumstances of the lowering stems. This simply means that "it is very
likely that verbs with the described structure belong rather to the
lowering stem type than to the non-lowering one; but this is a
simplifying prediction." Thus this is not a phonological rule but an
illustration, a first aid for the learners. Of couse it has a number of
exeptions but it helps us to draft the sphere of the affected lexemes.

  (The actual text of the "prediction" was cited from my Osmanli
Turkish phrase book. This book does not deal with phonemes at all but
it is good in useful practical hints for learners.)

  (N.B. We have no also exact phonotactic rules for the valency of anti-
harmonic stems in Hungarian: it is a lexematic feature. Therefore, in
this case, we must assign the [+-back] feature on the lexematic level.
However, we may draft some non-100% "brute force" rules to predict
whether an anti-harminic stem governs back or high suffixes in
Hungarian.)


> As you might be able to tell, I'm basically going through my dictionary
> from A to Z.   Since I've just now gotten to C and I've already found
> two, I'm going to assume there are more.
  Well, you have found two counterexample, but the question is how many
of the browsed verbs meet "my" requirements? If there were 8 verbs
meeting the requirements in addition to the two counterexample then I
have a hit ratio of 80%. IHMO this is quite good for a prediction --
note that I used the word "predict"!

  On the contrary you used the statement "You get [...] /-Ar/ after
most monosyllabic stems". This is a worse (i.e. less selective) test
than mine one.

  I suppose the difference between you and me is that you seem to treat
the <-Ar> ~ <-Ir> alternation as an irregularity. On the contrary I say
that it is not an irregularity but an inherent lexical attribute of the
stems. We say the same but I think my solution is a bit systemic.

  However, both you and me must say something practical about this
"irregularity". But this practical information is not the rule itself.