Re: Ablaut and Infix Origins

On Feb 26, 2008, at 2:36 PM, Jeffrey Jones wrote:

> On Sun, 24 Feb 2008 01:29:30 -0600, Eric Christopherson
> <rakko@...> wrote:
>
>> On Feb 23, 2008, at 10:28 PM, Jeffrey Jones wrote:
>>
>>> I don't really understand how ablaut and infixing come about.
>>> I've been
>>> trying to find information online with good explanations without any
>>> significant success. I found one paper on the theory of infix
>>> origins but it
>>> was very Chomskyan. Another summarized the different types but
>>> didn't
>>> give a me "feel" for it. There seems to be even less satisfactory
>>> information
>>> on ablaut origins. Apparently all the existing ablaut systems
>>> came about
>>> thousands of years ago. Any ideas?
>>>
>>> Jeff
What was the Chomskyan paper on infix origins? I'd like to take a
look at it.

>>
>> Funny you'd mention that -- I finally got around to reading some of
>> _A natural history of infixation_, by Alan C. L. Yu (Amazon: <http://
>> www.amazon.com/Natural-History-Infixation-Theoretical-Linguistics/dp/
>> 019927939X/ref=sr_1_1?ie=UTF8&s=books&qid=1203837054&amp;sr=8-1>). It
>> talks about four origins: metathesis; entrapment; reduplication
>> mutation; and morphological excrescence and prosodic stem
>> association.
>
> Someone was kind enough to send me the whole text in a private email.
> Unfortunately it was pretty much incomprehensible, except for
> chapter 5; I
> suppose that's the relevent part.
Yes.

>
>> For ablaut, you might follow Guy Deutscher's hypothetical model of
>> some features of Semitic morphology in his _The unfolding of
>> language_ (Amazon: <http://www.amazon.com/Unfolding-Language-Guy-
>> Deutscher/dp/0099460254/ref=sr_1_1?
>> ie=UTF8&s=books&qid=1203837261&sr=1-1>). His scenario depends on
>> sound changes in vowels caused by adjacent consonants (such as
>> pharyngeals), which then spread by analogy.
>
> I've heard of that book, but I don't have a book budget.
Try a library. I recommend it.

>
>> Finally, I know of a short paper by Adrian Macelaru called
>> "Compensatory Metathesis as a Source of Nonconcatenative Morphology:
>> Semitic Evidence". There used to be a Google-cached copy of it
>> somewhere on the web, but unfortunately I can't find it now. As the
>> title suggests, it implicates metathesis, but in this case it's
>> compensatory -- where, e.g., the loss of a final vowel happens at the
>> same time that some echo of that vowel occurs inside the word. Maybe
>> if we can find him we can ask him for a copy; he seemed very nice,
>> but busy.
>
> I couldn't find it either.
I will look on my other computer and see if I can find it.

>
>> Infixes, ablaut, and nonconcatenative morphology are some of my
>> favorite morphological things.
>
> On Sun, 24 Feb 2008 02:42:12 -0600, Eric Christopherson
> <rakko@...> wrote:
>
>> On Feb 24, 2008, at 1:29 AM, Eric Christopherson wrote:
>>
>> Here's something else I just read, from Andrew Sihler's _New
>> comparative grammar of Greek and Latin_ (Amazon: <http://
>> www.amazon.com/New-Comparative-Grammar-Greek-Latin/dp/0195083458/
>> ref=sr_1_1?ie=UTF8&s=books&qid=1203841422&sr=8-1>; Relevant section
>> from Google Books: <http://books.google.com/books?
>> id=IeHmqKY2BqoC&printsec=frontcover&sig=0SXOYO1u9_WAGMKcWJ4Xu5Kge
> QY#PPA109,M1>):
>>
>> <quote>
>> Vowel alternations result from commonplace kinds of sound change.
>> English has vowel alternations which arose at various times from
>> various causes. Thus the alternations seen in NE [=New, or Modern,
>> English] _drink, drank, drunk; meet, met; blood, bleed; wise, wisdom;
>> revise, revision; efficient, effective_ represent six unrelated
>> patterns, that is, they arose via six different historical
>> developments. In addition, accidents (such as borrowing or chance
>> resemblance) on occasion create an appearance of alternation: _cat,
>> kitten; ill, ailing; choose, choice; bed, boudoir; strap, strop;
>> whole, hale_.
>> </quote>
>
> I suppose so. The question is, how can a _system_ of alternations
> involving
> the whole language, such as in PIE or Semitic, arise? Each of the
> English
> pattern examples applies to a small set of words and the accidental
> alternations are each one of a kind AFAICT. (I'm having trouble
> saying what I
> want to say, so I'll leave it at that)
You're doing fine!

The key, according to everything I've read, is analogy. People notice
a few forms (or maybe in some cases just one form) which are similar
except for internal alternations, and which seem to mean similar
things, and then they extend the pattern to other words that didn't
fall into that category originally.

>
>> I find it very interesting to learn that the alternations listed all
>> come from separate patterns, and am especially impressed with his
>> list of chance resemblances (which I had already read could lead to
>> novel alternations in the minds of speakers, but had never seen a
>> list of such words in English).
>>
>> AFMCL, I made up a list of root words with definitions once, and it
>> turned out that three of the roots for body parts ended in /?/, quite
>> by accident! -- a situation which could certainly cause speakers to
>> infer that /?/ is a body part suffix. If I had instead coined words
>> with /?/ inside, my speakers might instead have intuited that /?/ to
>> be an infix.
>
> I suppose they might, although three words doesn't sound like much
> of a basis
> for generalization (and I personally find infixes completely
> unintuitive).
Well, the way I see it, the set of words for body parts is probably
pretty small in most languages, so the occurrence of three roots
ending in /?/ out of this small set of vocabulary will seem more
significant than it would if they had unrelated meanings. Body part
words, I think, form a natural set.

Now, where they would go from there, I don't know. There was another
body part word ending in /h/, which is a similar sound to /?/, so
perhaps they would start by replacing the /h/ with /?/ in that word.
Then maybe they would replace final stops with /?/, which would be a
somewhat natural sound change anyway, at least in small quantities.

>
> Thanks for the detailed response(s).
You're welcome.

I've also come up with some more possibilities. As Daniel mentioned,
umlaut could contribute; the umlauted sounds could stay even after
the sound that caused the umlaut vanished.

There's also the phenomenon of vowel harmony. Especially useful for
producing ablaut within a stem is dominant/recessive vowel harmony,
where there is a class of vowels which is "dominant" and changes the
vowels of a word to conform with it, whether they occur in the stem
or in affixes. (The other kind I know of -- I don't know the name
offhand -- causes the vowels of affixes to conform to the vowel of
the stem. Oddly, I don't see these two types listed in the Wikipedia
entry on vowel harmony, but I thought they used to be there.)

So you could have a recessive set /i/ /1/ /u/ and a dominant set /e/ /
a/ /o/. For a stem /mikut/ plus different affixes you'd get:

mikut + -i > mikuti
mikut + -e > mekote

Then delete the final vowels (for example), and you'd end up with /
mikut/ : /mekot/.

Of course, you might want to explain then how vowel harmony arises. I
think it's similar to umlaut except that it has a longer range. In
some languages, post-velar (e.g. uvular and pharyngeal) consonants
tend to retract and/or lower the vowels in a word, not just the ones
adjacent to the consonant.

I've found that I sometimes have trouble pronouncing words or close-
knit phrases containing both /a/ and /O/; I tend to shift both to /O/
sometimes. It's more common for some words than others, e.g. with
<oddball> and <dodgeball> I don't have much trouble keeping the
vowels distinct, but <automaton> is more difficult. I've toyed with
using this tendency in a conlang, such that /a/ shifts to [Q] or [O]
when next to /w/, and takes the rest of the /a/s in the word with it.
It feels perfectly natural to me, although I don't know why, and I
don't think everyone would feel the same way.