LINGUIST List 3.1008

Tue 22 Dec 1992

Sum: Velar/Coronal Change

Editor for this issue: <>


  1. , Sum: velar/coronal change

Message 1: Sum: velar/coronal change

Date: Mon, 21 Dec 1992 22:03:36 Sum: velar/coronal change
From: <>
Subject: Sum: velar/coronal change

Back in mid-November i posted a query, seeking examples of syllable-final
velar stops changing into coronals (specifically, apicals), perhaps in the
environment of another coronal, e.g. /l/. This query was motivated (1) by
a desire to justify a proposed change in the pre-history of Uralic whereby
a hypothesized plural marker *-kl had become -t in many languages (2) by
arguments for the 'special (= unmarked, among other things) status of
coronals', many of which were gathered together in a recent book by that
title published by Academic Press.

I got about a dozen responses, some of which lead to further discussion at
a more personal level. Several people provided me with examples of the
sort of thing i was looking for:

Richard Coates of Sussex (<>) told me that the
Modern English word 'bat', meaning 'flying mammal', derived from a Middle
English form 'bakke'.
Jose Ignacio Hualde of Illinois (<>) reports that
in Basque morpheme-final velars often become /t/ in certain compounding
processes ('This is a productive process to some extent', he says): beg-
'eye' + azal 'skin' = betazal 'eyelid'. One example he gave me looks as
though it might involve assimilation: og- 'bread' + tarteko 'between' =
otarteko 'sandwich'. More on the subject of assimilation below.
Ari Pall Kristinsson of Reykjavik (<>) pointed out that the
Modern Icelandic middle voice marker -st is derived from an Old Norse form
-sk (i ought to have realized this myself). He made reference to a book by
Kjartan G. Ottosson, The Icelandic Middle Voice: the Morphological and
Phonological Development, published this year by the Lund University Dept.
of Scandinavian Languages, which i haven't been able to get my hands on
Fernando Martinez-Gil of Georgetown (<>) drew
my attention to the change, characteristic in some Spanish dialects, of
syllable-final /k/ to an interdental fricative when followed by /t/. He
referred me to T. Navarro Tomas' Manual de Pronunciacion Espanola which
describes this phenomenon.

Many of these velar -> coronal changes (especially those noted by
Kristinsson and Martinez-Gil) at least appear to be cases of assimilation,
which i originally suggested was the case in the prehistory of those Uralic
languages which use -t as a plural marker. Joe Stemberger of Minnesota
(<>) also flat-out suggested that assimilation would
account for what i was proposing, and that i didn't need to invoke the
'special status of coronals' to explain it. Some scholars, however, also
noted examples of dissimilation, of changes in the opposite direction,
along the line of Latin 'vetulus' -> LLatin 'vetlus' -> Italian 'vecchio'.

Fran Karttunen of Texas (<>) noted that the
lateral affricate /tl/ characteristic of Aztec is routinely borrowed into
Spanish as /kl/.
John D. Phillips of the National Language Research Institute, Tokyo
(<>) notes that the fate of word-final *-kt in
the various Celtic languages provides both examples and counterexamples to
the 'underspecification = coronal' hypothesis: In Welsh (and, as far as i
know, the other Brythonic languages, Breton and Cornish) it evolved into a
dental fricative, but in Goidelic (Irish, Manx, Scottish Gaelic) it evolved
to -xt (where 'x' represents a velar fricative); in Scottish Gaelic this is
now -xk, in Manx Gaelic simply -x.
Spike Gildea of Oregon (<>) referred me to a
recent ESCOL paper by Keren Rice of Toronto, in which she 'presented a
great deal of data showing both diachronic and synchronic alternations
between [velar and coronal segments], usually at the coda of a syllable,
usually in the direction coronal -> velar. Unfortunately, i have so far
been unable to get in touch with Keren to know what sort of data she is
using or what her argument specifically claims. I would be interested in
knowing how it affects the 'special status of coronals' hypothesis.

A couple of people brought up what might be examples of neutralization of a
velar/coronal distinction:

John S. Coleman (<>) notes that 'Middle Chinese
had distinct labial, coronal, and velar final stops that are preserved in
many modern dialects (e.g. Cantonese), but which distinctions have been
lost in some words in some dialects.' One example he gives has both velar
and dental final stops surfacing as dentals in Hakka. However, he notes
that in Mandarin, while historically labial and coronal final nasals both
surface as coronals, the velar nasals are retained in word-final position.
John Kingston of Massachusetts (<>) brought up a very
interesting fact from Southern Bantu, in which Proto-Bantu palatal stops
have evolved into lateral affricates. The stop portion of these affricates
varies freely from coronal and velar. He agreed with me that this might
well be an example of neutralization of the coronal/velar contrast, at
least in this environment. But he also raised an acoustic point: 'Laterals
tend to raise F3, and both F2 and F3 are high in coronals, but velars tend
to lower F3 while leaving F2 high.' So why would the presence of a lateral
induce a velar, as opposed to a coronal, manifestation of the stop. During
further discussion between us, he suggested that both assimilation and
dissimilation in this case might be the result of uncertainty as to where
the high F3 associated with the lateral component begins: if it begins with
the stop, then that stop is perceived as coronal, but if the high F3 is
perceived as being characteristic specifically of the lateral then the stop
will probably be perceived as a velar. He directed me to a paper by John
J. Ohala, 'The Listener as a Source of Sound Change' in Carrie Masek,
Robert Hendrick, & Mary Miller (eds.) Papers from the Parasession on
Language and Behavior, CLS 1981, pp. 178-203, in which this theoretical
issue is discussed in greater detail.
Brian Joseph of OSU (<>) also suggested
that, since both assimilations and dissimilations involving /kl/ and /tl/
clusters are attested, an acoustic account is probably the best

Finally, a couple of remarks from the field of child language acquisition:

Allan Wechsler (<>) notes that 'coronals
are acquired late ... I don't know if this argues against the status of
coronals you propose.' I told him i couldn't judge of the relevance of
child language-acquisition data for phonetic theory, but would pass the
question on to phoneticians. He acknowledged that, although one would
assume 'kids would tend to prefer "unmarked" segments in early stages of
acquisition', he had no strong theoretical reason to believe this is in
fact the case; other explanations of the data he mentioned are possible
(e.g., if +velar is indeed a 'marked' option, perhaps at a certain point in
language acquisition any word containing a velar will tend to be marked
+velar as a whole, with all obstruents in it being velars). But he went on
to express his suspicions of theories ranking basic points of articulation
in terms of markedness. Any phoneticians out there have anything to say
about this?
Joe Stemberger, on the other hand, noted that young children can handle
word-initial velars if they are alone, but pronounce them as alveolars if
they are parts of clusters (whether they pronounce the rest of the cluster
or not). He referred me to a paper by Alice T. Dyson, 'Development of
Velar Consonants among Normal Two-Year-Olds', Journal of Speech and Hearing
Research 29:493-498 (1986) that discusses this.

I will be including a version of this summary as an appendix of my paper on
the reconstruction of Dravidian and Uralic plural markers. Any further
comments are welcome.
Dr. Steven Schaufele c/o Department of Linguistics
712 W. Washington Ave. University of Illinois
Urbana, IL 61801 4088 Foreign Languages Building
 707 S. Mathews Street
217-344-8240 Urbana, IL 61801
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue