--- In
cybalist@yahoogroups.com, "Brian M. Scott" <BMScott@...>
wrote:
> E.g., Terence R. Carlton (Introduction to the phonological
> History of the Slavic Languages, 1990) gives it as follows:
>
> /i(:)/, /e,/ + /k/, /g/, /x/ + vowel (except /u(:)/) >
> /c/, /dz/ or /z'/, /s^/ or /s'/.
>
> Certainly what I've read gives me no reason to doubt that
> this is now the traditional formulation, even if it's not
> the original one. (In my own field the traditional and
> original definitions of 'compact space' are quite different,
> for instance.)
I think I was the one who introduced the term "traditional
formulation" and I owe you an explanation.
The problem is that the solutions to the challenge of what is now
usually called the Progressive Palatalization (Prog) fall into
various broad classes:
(1) Those attributing palatalization to a following *j.
(2) Those attributing it to a preceding front vowel.
(3) Those attributing it to morphological factors.
(4) The lunatic fringe.
The trouble about (1), (3) and (4) is that they fail to account for
the fact that in all (or nearly all) cases of palatalization a
preceding front vowel is actually present. (The first to make this
point explicit was Vatroslav Jagic in Archiv für slavische Philologie
10, 1887, 192.)
The trouble about (2) is that it massively overgenerates. Since the
role of the preceding vowel cannot be ignored (as argued by Jagic),
the problem becomes how to constrain the palatalization rule.
Basically the following approaches have been experimented with:
(2A) The place of the stress plays a role. This is Baudouin de
Courtenay's idea. It dominated the field between 1893 and 1910, after
which it virtually disappeared.
(2B) The presence or absence of a following *j plays a role. Actually
this combines the classical "class (1)" solution with the idea of a
progressive palatalization. It was first proposed by Karl Brugmann in
the Brugmann & Delbrück Grundriss 1897, pp. 291-292. It proved
remarkably tenacious, particularly in Scandinavia (Knutsson, Ekblom).
As far as I can see it was finished off by Maria Jezowa in her 1968
monograph.
(2C) The immediate phonological context of the velar plays a role.
This idea was first aired by S^axmatov in Izve^stija ORJS 1 (1896),
p. 703, but was more or less ignored while everybody was enthusing
about Baudouin de Courtenay's stress rule, without, however, anybody
taking the trouble to verify it on the basis of the available factual
material.
It is 2C that provides the class of solutions I have been
calling "traditional".
Carlton's version operates with the assumption that Prog was blocked
by the reflexes of following *u: and *u. Consensus about this point
has been virtually complete since it was first aired by Josef Zubatý
in 1910 and Tadeusz Lehr-Spl/awin/ski in 1911.
Carlton appears however to be oblivious of the fact that scholars
have in addition paid attention to the vowel triggering
palatalization. Baudouin de Courtenay may have seen already (his text
is somewhat ambiguous) that it is not just any *i, *I and nasal *e
that triggers palatalization but a subset. This idea was pursued with
particular persistence by Antoine Meillet in a series of
publications. Its most consistent variety holds that palatalization
is triggered only by *i, *I and nasal *e reflecting earlier high
front vowels, hence not by *i reflecting earlier *ei or umlauted *u:,
by *u reflecting umlauted *u, and by nasal *e reflecting a sequence
*eN. The implication is, of course, that Prog preceded the mergers
involved. Despite slight variations, Meillet's view of the matter has
become part of a broad consensus.
Scholars have disagreed notably about the effect of *ei and about
whether or not nasal *o blocked palatalization.
What I call a "traditional" formulation is any formulation that
assumes that palatalization was blocked by a following consonant or
high rounded vowel and that palatalization is not found after *I or
*i if reflecting high rounded vowels in umlaut position.
This leaves room for disagreement about such matters as preceding *ei
and following nasal *o, while keeping out such more excentric
formulations as Trubetzkoy's (1922), who assumed palatalization was
blocked by *oi but not by *U. (Neither assumption can be sustained.)
> Carlton also seems to look with some favor on the
> possibility that the progressive palatalization is actually
> very old, pointing out that this could explain a number of
> exceptions; I'd like to ask Miguel and Willem what the
> subsequent fate of this idea has been.
The idea has had a very interesting history.
If one reads between the lines in Roman Jakobson's "Remarques sur
l'évolution phonologique du russe comparée à celle des autres langues
slaves" of 1929, one can see the idea lurking there on p. 19, where
the possibility is implicit that the subphonemic phase of Prog
entered the system before the subphonemic phase of the First
Palatalization. Henceforth the idea that Prog preceded the First
Palatalization will be closely connected with the Jakobsonian school.
As far as I know the idea that Prog preceded the First Palatalization
was first made explicit by Martinet in 1952 in an article that was
later incorporated in his "Économie des changements phonétiques"
(1955, pp. 366-367). (In the early fifties Martinet was very close
with Jakobson and his reasoning is archetypically Jakobsonnian.)
The idea was subsequently developed by Robert Channon in 1965 in a
Master's Thesis that was published by Mouton in 1972 (p. 46). Since
the personal relationship between Jakobson and Martinet had long
since soured, it goes without saying that Channon does not mention
Martinet's name, although his reasoning is virtually the same.
By that time it had also been publicly mentioned by David Cohen in an
article I haven't seen (Papers from the 5th Regional Meeting of the
Chicago Linguistic Society, 1969, 306-313).
Channon's monograph was discussed (and gutted) in detail by Lars
Steensland in IJSLP 21 (1975), notably on pp. 95-97.
By that time, Horace Lunt, then of Harvard University, had embarked
on a series of publications intended to develop Martinet's idea
(again, it goes without saying, without actually mentioning Martinet,
who was a non-person, or, for that matter, Steensland, whose
criticism may have been felt to be unanswerable, and written in
German anyhow). This culminated in a monograph which appeared in 1981
and was followed by a series of articles intended to clarify his
position in the face of criticism by various investigators. Lunt's
monograph is useful for various reasons, notably because he discusses
most of the factual material.
It is my impression (perhaps wrongly) that for some time, it was not
considered politically acceptable in North America to reject
Martinet's chronology, linked as it was with Roman Jakobson and his
powerful pupils, and also because it was generally believed (wrongly)
to be based on generative reasoning, which gave it the additional
boost of modernity and political progressiveness.
Honour was saved by Alexander Schenker in his "The Dawn of Slavic"
of, I think, 1995 (pp. 91-92).
The most recent exponent of the Martinet chronology I'm aware of is
Geoffrey Schwartz, Speech perception, sound change, and the Slavic
palatalizations, FLH 22/1-2, 2001, 277-300. As was to be expected, he
pays no attention to the criticism that has been voiced in the course
of the decades.
Since in linguistics it is considered acceptable to ignore or
ridicule criticism, the Martinet chronology is probably here to stay
for a long time, whatever its merits.
Willem