Three-Way Contrast of Secondary Articulations in PIE

[David L. White]

> On 25 Mar 2001, David L. White wrote:

>> My point is that to speak of "plain" consonants existing in such a system is
>> in a sense non-sensical, because "plain" in effect means "as in a language
>> without secondary articulations", which in turn means "as in a language with
>> ordinary co-articulation", but a language with secondary articulations could
>> not possibly have ordinary coarticulations, since overriding ordinary
>> coarticlation for phonemic effect is what languages with secondary
>> articulations do.

> This indicates that we have encountered a terminological whirlpool.  In the
> phonological theory to which I most strongly subscribe, the term "plain"
> means simply "not having secondary articulations" when used to describe a
> sound in a language which has same.  It makes no reference to other languages
> or features to be found therein.

        I am not concerned with phonological theory (and running out of
energy, under the delusion that I am perhaps getting a life).  And if I have
adopted a phonological theory that does not allow negatives, it is news to
me.  "Not having secondary articulations" to my mind means "having normal
coarticulation".  By "normal" is meant what is most energetically efficient,
as in fronting velars to some (variable) extent before front vowels, doing a
bit of anticipatory rounding before round vowels, etc.  (The lips take a
while to get going, which is why u-umlaut in Norse has a longer leftward
range than i-umlaut.)
        But to back up a bit, and hopefully make matters clear, it is best
to consider the case of the labials.  Here, unlike in the case of velars and
dentals, it is clear that the primary place of articulation cannot in any
way be shifted, which simplifies life considerably.   Thus [i]-quality in
such cases can only be realized by putting the tongue in position for [i]
while making the labial in question, and leaving it there long enough for
some effect on adjacent vowels to be perceptible.  (An [i] gesture that was
somehow magically confined to the period the labial closure would not be
perceptible.)  Likewise with [u]-quality, though matters are complicated
somewhat by having to throw in a little extra labial something (in Russian I
have heard it called "lip-protrusion") to make the theoretical lip-rounding
perceptible.  (Otherwise, we would have merely a velarized labial.)  In
languages without secondary articulations, and therefore presumably with
"plain" labials, during the production of these the tongue assumes whatever
position is most convenient for associated (most often following) vowels.
This fact has even been used to develop a theory of how consonantal place is
heard, since the transition from a labial to any vowel is fairly rapid.  (It
is, I think I recall, slowest in the velars.)
    So the question is, in a language with palatalized, labio-velarized, and
"plain" consonants, what would a speaker's tongue be doing as the syllable
/pi/, with "plain /p/", was produced?  It would have to be doing something
other than anticipating [i] in the normal manner, or confusion with
palatalized /p/ would be inevitable.  Putting the tongue in the position for
[a] would result in the clearest contrast.  Or, to put it (yet) another way,
with labials secondary articulations can hardly be realized as anything
other than super-short diphthongs (rising or falling).  Thus /ap'i/ with
palatalized /p/ would have to sound something like [aip] (with super-short
[i]), or the palatalization would not be perceptible.  /ap/ with "plain" /p/
would sound like [ap], and /ip' with palatalized /p/ would sound like [ip].
But /ip/ with "plain /p/ would have to sound something like [iap].  And /pi/
with "plain" /p/ would have to sound something like [pai] (with rising
"diphthong"], otherwise normal anticipation of /i/ would make it sound like
/p'i/.
        I mean, we are deaing with a very small window here, outside of
which vowel-like sounds will inevitably be taken as vowels.  (This process
has happened repeatdly in Irish, where former secondary articulations have
become vowels, and vice versa.)  But the small window means that 1)
confusion is a danger, and 2) very few distinctions can be made.  Beyond a
point, what was meant to be palatalized /p/ before a vowel begins to sound
like /py/, and is likely to be re-analyzed as such, and after a vowel begins
to sound like /-ip/, which again is likely to be re-analyzed as such.  In
order for such systems to be preserved, the times involved have to be short,
and the distinctions made few.
        To sum up, for the labials at least, secondary articulations must be
realized as de facto super-short diphthongs, rising or falling, since no
displacemnt of place is possible.   Where [i]-quality and [u]-quality are
already in use, a third quality, and therefore a third type of diphthong,
would have to use [a].  Anything else, or normal coarticulation such as
occurs in languages without secondary articulations, would be too
confusable, given the severe constraints such systems necessarily operate
under.  It is not as if anything (abstractly conceivable) goes.
Phonological systems must be implemented, after all.  And all phonemes must
to some extent stay out of each other's way, or run the very real risk of
merger.  Even if Old Irish had a three-way contrast, most observers admit
(so it seems) that this was inherently unstable, for the reasons which have
been indicated, and there is no denying that whatever labio-velarized and
"plain" series may have existed merged, seemingly not too long after the
system was created in the first place.  (The fact that Old Irish had "real"
diphthongs surely did not help matters, as it would have reduced the
"window" for moraic consonants even further.)
        Anyway, that's my understanding.  It could be wrong.  The real
question, hopefully a fairly straightforward empirical one, is whether it is
a fact the Caucasian languages have systems with three-way contrasts.
Catford would be the one to go to, I think, but I am too lazy to do more
than give refs:

Catford, J.C.  1972.   "Labialization in Caucasian Languages,  with
Special Reference to Abkhaz"; in Proceedings of the Seventh
International Congress of Phonetic Sciences, ed. by A. Rigault and     R.
Charbonneau.  Mouton:  the Hague; 679-82.
__________. 1977.  Fundamental Problems in Phonetics.  Indiana
University Press: Bloomington.
__________. 1977.  "Mountain of Tongues:  the Languages of the
Caucasus"; in Annual Review of Anthroplogy 6:  283-314.

        Nothing in the phonetics books I have available here describes any
such thing, and the silence, while not defeaning, is (to my mind)
suspicious, especially given the practical considerations note above, not to
mention the tendency of the exotic and extraordinary to attract notice.  But
I do not have Maddieson's "Patterns of Sounds".  Perhaps it is in there.

Dr. David L. White