LINGUIST List 3.22

Wed 08 Jan 1992

Disc: Diachronic Lengthening

Editor for this issue: <>

Directory

1. , 3.21 Diachronic Lengthening
2. , Re: 3.21 Diachronic Lengthening
3. William Crossgrove, Re: 3.21 Diachronic Lengthening
4. "Larry G. Hutchinson", Re: 3.21 Diachronic Lengthening
5. John E. Koontz, Re: 3.21 Diachronic Lengthening
6. Jacques Guy, Diachronic lengthening
7. Richard Coates, Re: 3.21 Diachronic Lengthening

I think the original question was about words lengthening over
time as a result of PHONOLOGICAL change, which the recent
responses by Peeters and LaPolla do not address.
One example that is frequently cited is the addition of a final
vowel to certain consonant-final words between Old Tamil and
Modern Tamil.
Starting from the other end, so to speak, one has such examples
as the many varieties of Romance (Spanish, Catalan, Portuguese, French)
that add an initial vowel before ancestral (i.e., Latin) s+consonant
clusters.

Message 2: Re: 3.21 Diachronic Lengthening

The process known as the quantity shift in Germanic languages would be a
good example of genuine (non-compensatory) lengthening. In the Scandinavian
languages what is traditionally known as light stems, i.e. CvC monosyllables
and CvCv polysyllables, are lost as types, by virtue of the quantity shift.
Examples: Old Swedish vika 'week', gata 'street', bik 'pitch', tak 'roof'
show up as vekka <vecka>, gaata <gata>, bekk <beck> and taak <tak> in Modern
(central) Swedish.
Tomas Riad
Scandinavian languages
Stockholm university

Message 3: Re: 3.21 Diachronic Lengthening

It seems to me that when I was a graduate student many years ago, one of
the common speculations about pre-Indo-European was that it is likely to
have had post-postional forms that were later reduced to inflectional
endings, hence creating "longer" words. Of course, this is largely
speculative, but it suggests that people did not find it difficult to
envision this sequence of development. This may seem like a quaint
notion to today's Indo-Europeanists, but I offer it as an example.

Bill Crossgrove
Brown University

Message 4: Re: 3.21 Diachronic Lengthening

Phonological shortening is a mouth/time/familiarity thing. But semantics has
its own needs, and restoration of contrasts via things like preprefixes,
double negation, etc., etc. is widely attested. Longer expressions are the
 (temporary) result.

Message 5: Re: 3.21 Diachronic Lengthening

Don Webb asks
> 'Is there any language in which words have evolved into *longer* forms?'

In the Siouan family, most roots are monosyllabic or bisyllabic, but most
actual words are rather longer, at least bisyllabic. Part of this is due to
the accretion of clitic material and the presence of inflectional and
derivational affixes (including markers of oblique verb stems, reflexives,
etc.). However, even historically monosyllabic roots also tend to give rise
to bisyllabic stems. Several processes are at work.

One process is accretion of suffixes of the form *(r)V (R included after
final vowels). In some of the languages, e.g., in the Dhegiha branch, the
extended forms are the only ones attested, apart from a few fossils here and
there. In other languages, e.g., Dakotan, this process still operates more
or less productively to form independent (free-standing) stems from bound
stems. Although the process has generally been interpreted synchronically
as phonologically governed epenthesis, I think it makes more diachronic
sense with nouns to interpret it as fusing of stems with bleached anaphoric
material, after the fashion of Greenberg's analysis in (approx) "Where does
gender [marking] come from?" Examples: Da c^ha~l=wa's^te `pleasant' (heart
+ warm), vs. c^ha~te' `heart' (with -e) (the t => l shift is a low level
change); he'ya (*r => y in Dakotan) `louse', vs. hez^a~'z^a~=la `nit' (louse
+ pale (redup) + diminutive).

This process is comparable in general terms to processes in other languages
world-wide that have been called variously theme formation (cf.
Indo-European) or absolute formation (cf. Uto-Aztecan). The details vary,
but all such processes involve adding an empty or gender marking extension
to a stem in certain morphosyntactic contexts, especially when other markers
with more concrete meanings are absent, but often also with some markers,
but not all, or with all markers (in which case the extension has become
part of the root, for all practical purposes).

Other processes that extend stems in the Siouan family include reduplication,
cf. Winnebago ho'o `fish' (long monosyllable) vs. Omaha-Ponca hu'hu `fish'
(redup, presumably fish considered as a school of fish); appending of
bleached auxiliaries, cf. Omaha-Ponca ga~'=dha `want' (with coverb =dha of
unknown meaning); moribund unpossessed markers, cf. Omaha-Ponca watha~'
`squash', with wa- `unpossessed' or wahi' `leg'; pervasive compounding, cf.
Omaha-Ponca i'=ha `lip' (mouth skin), i'=dhe `to promise, to speak of'
(mouth make); pervasive use of instrumental prefixes with verbs to indicate
unfocussed means or instrument (effectively a system of manner classifiers),
cf. Omaha-Ponca gase' `to cut; sever with a blow', base' `to cut; sever with
a long tool', ma'=se `to cut; sever with a blade', etc.

Caution: All examples cited from memory!

There may be a case of diachronic lengthening in the northern dialect of Sakao
(Espiritu Santo, Vanuatu), and there certainly is, but artificially provoked,
in the variety of the southern dialect spoken in Hog-Harbour.

Sakao is an Austronesian language that has undergone drastic sound changes, e.g.

*vati --> jED (E = IPA epsilon, D = delta)

In brief, the article *na became fused to the noun, with its vowel partly
assimilating to the vowel of the following syllable, then word stress
was innovated and a sweeping vowel shift took place. Finally, in the
southern dialect, all unstressed vowels were lost. Thus:

*na gatsi (black ant) --> naR
*na matsi (fish, bird) --> nnEs
*na Gweleku-ku (my hand) --> nlkyG (G = IPA gamma)
*na patu-ku (my head) --> nBDyG (B = IPA beta)
*na toa (fowl) --> *notoa --> nCD (C = IPA mirror-image of "c")

The northern dialect, on the other hand, looks as if it had retained
the pretonic vowels lost by the southern dialect:

 North South

 aDalan nDlan "cloud"
 oeBDyG nBDyG "my head" (oe is the IPA oe, joined)
 Ette ntte "something"
 kElE klE "to paint, smear, write"
 mEkElE mklE "he paints, etc."
 mAnAs mnAs "he likes" (A = back, rounded, a)
 AGArAn nGrAn "its behind"
 oerymyG nRmyG 'my voice, thoughts"

>From even the very few examples above, it seems likely that what happened is
that consonant clusters were broken up in the northern dialect by epenthetic
vowels that only recently emerged as fully fledged phonemes (I'd rather say,
archiphonemes). Cases where pretonic vowels are not predictable (as they are
above) from the tonic seem to correspond to earlier vowel clusters (mostly
arising from the loss of fricatives). Thus:

olAn "its tail" <-- *ne-Gwele-na

as attested by:

AwAl- "tail of" <-- *ne-Gwele

(Gwele is attested in many other Santo languages by such forms as "fele, vele";
so what I reconstruct *Gw may have been [hw])

I used to believe that the Hog Harbour dialect behaved in a similar
manner, until I discovered what I had been told was Hog-Harbour dialect
was "church speech": the people of Hog-Harbour, in fact, spoke exactly
like the unchristianized villagers on the plateau above, except when in
the presence of church elders, or of outsiders. Digging through old
records and listening to stories from old men, I found that the
Presbyterian missionaries sent there had never managed to learn the
language properly, least of all pronounce 5 or 6 consonants in a row,
nor distinguish between its 11 oral vowels, so that they used a kind of
Sakao-Pidgin cum baby-talk, which they proceeded to enshrine in printed
translations of some psalms and parts of the New Testament. Thus they
wrote "nu-ruru-muc" or "nuru-rumuc" (they consistently used "c" for a
velar fricative) what was rightly /nRmyG/ (R = trill), and "vocovoc"
what was, and still is /BCGBCG/. The worst case I ever came across was
when I elicited a standard word list from the native Pastor: he gave me
/nCDCDCD/ for "butterfly"; much later I heard the proper word: /nDDAD/
(which is regularly derivable from *na pepepe, attested in other
languages of the region, whereas /nDDCD/ is not, let alone /nCDCDCD/).

I doubt that the epenthetic vowels of the northern dialect are due to
a similar influence for this simple reason that both pagan and christianized
northern Sakao speakers speak precisely the same language.

I vaguely remember having read somewhere that Russian did such a thing,
that is, break up consonant clusters. Could someone who reads Linguist confirm
or correct this, please?

Message 7: Re: 3.21 Diachronic Lengthening

Anyone interested in diachronic lengthening of words might note]
Schuhmacher's work "Wortverla"ngerung im Melanesischen und Man'czaks
zweites Gesetz", in Zeitschrift fu"r Phonetik 24 (1971).

Richard Coates