LINGUIST List 4.1083

Tue 21 Dec 1993

Sum: Yers

Editor for this issue: <>

Directory

1. , Russian yers (aka "jers")

Several weeks ago I posted a query concerning yers. Having now
finished my Christmas shopping (well, most of it :-), here's a
summary. I've considerably edited replies, since many were quite
similar. Thanks for replies to Harry Bochner
(bochnerdas.harvard.edu), Ursula Doleschal
(ursula.doleschalwu-wien.ac.at), Steve Seegmiller
(seegmillerapollo.montclair.edu), Arkady Borkovsky
(arkadydnt.dialog.com), David Pesetsky (pesetskMIT.EDU), Rob
Pensalfini (rjpensalMIT.EDU), George Fowler
(GFOWLERucs.indiana.edu), Irina Sekerina (saicunyvms1.gc.cuny.edu),
?? (gsavpiASUCLA.UCLA.EDU), Bill Idsardi (idsardichopin.udel.edu),
and Morris Halle hallemit.edu.

THE ORIGINAL QUERY:

In Phonologica 1988 (= Proc 6th Intl Phonology Mtg), Jonathan Kaye
describes the behavior of yers (underlyingly high, probably lax
vowels) in Russian as follows (pg. 149):

(1) Yers are nonhigh...before a syllable whose head vowel is a yer.
(2) Yers are deleted.

...What happens when three or more yers appear in sequence (in stems
and/or suffixes, but not in prefixes, since the latter introduce
complications that I won't go into here)?

The replies fell into two categories: what happened diachronically,
and what happened synchronically.

DIACHRONICALLY:
>Originally in sequences of yers deletion occurred in alternate
>syllables; this is known among Slavicists as Havlik's Law. E.g. in
>Old Church Slavonic mss "dnes" < "dInIsU". So the [synchronic]
>pattern you refer to in stems & suffixes is historically the result
>of paradigm levelling. The complications in cases with prefixes
>reflect the earlier state of affairs, preserved relatively well in
>conservative languages like Russian, but completely levelled in
>Serbo-Croatian, for instance. (--Harry Bochner)

>The common position taken on the falling of the yers is:
>1) you count the syllables starting from # or $V
>2) yers (both back and front) are deleted in odd and strengthened in
> even syllables

>Some examples taken from a standard historical survey of Russian
>(you might have a look at any, this problem has been studied
>extensively), I'll transcribe yers as ?:
> kus?k?m? > kuskom (bite-Instr.Sg.)
> zh?n?nc?>zhnec (harvester, zh is voiced sh)
> d?n?c?>dnes' (daily adv.)
>There are exceptions, but this is the general rule. References are
>e.g. Valentin Kiparsky 1963. Russische historische Grammatik Band 1.
>Heidelberg; Frantisek Mares 1969: Diachronische Phonologie des
>Urund Fruhslawischen. Munich, but there must be things in English,
>too. (--Ursula Doleschal)

SYNCHRONICALLY:
>It's my understanding that when we look at yers across suffix
>boundaries, all but the last get realized as vowels:
[NB: My original posting was mistaken in its description--MM]
> dEn-Ek-U 'day (dim. nom.)' --> dene"k /denjok/
> dEn-Ek-a 'day (dim. gen.)' --> denka
>also:
> dEn-Ek-Ek-U 'day (2xdim. nom.)' --> dene"chek
> dEn-Ek-Ek-a 'day (2xdim. gen.)' --> dene"chka
>note:
> dEn'-U 'day (nom.) --> den'
> dEn'-a 'day (gen.) --> dn'a
>Janis Melvold did an MIT dissertation on yers and their interaction
>with the stress system in 1990, which you can get from MIT Working
>Papers in Linguistics here (MITWPL). (--David Pesetsky)

>In a sequence of more than two yers, generally all will be vocalized
>before the last one, which will be deleted. This occurs mainly in
>multiple diminutives, with two or more diminutive suffixes. For
>example (in a kind of pre-vowel-reduction transliteration; yer is
>#):
> ogon' 'fire', Gen. sg. ogn'a, from og#n'-#, og#n'-a
> ogon'ok 'little fire [dim.]', ogon'ka, from og#n'-#k-#, og#n'-#k-a
> ogon'ochok, 'itsy-bitsy flame', ogon'ochka, from og#n'-#k-#k-#,
> og#n'-#k-#k-a
>A good place to read about this is in David Pesetsky's unpublished
>100-page 1979 ms. "Russian morphology and lexical theory", which
>does this stuff via cyclic rules and bracket erasure in lexical
>phonology. So yer vocalization ("Lower", in his terminology) is
>cyclic, and occurs at the 2nd, 3rd, and 4th cycles in the last word
>above. Yer erasure is post-cyclic, and wipes out the
>last one after the derivation is complete. This creates some
>problems when a yer is vocalized between a prefix and root. For
>example,
> podzhog 'set fire to [masc. past]' but podozhgla 'ditto
> [fem. past]'
>These putatively come from pod#-zh#g-l# (with a rule that deletes
> l in this environment after certain obstruents) and pod#-zh#g-la.
>Here, it looks like the rule doesn't work as with diminutives,
>since in the sequence of three yers in the masculine form the first
>doesn't vocalize. In his 1985 LI paper "Morphology and Logical
>form", Pesetsky argues for a bracketing paradox: the prefix is
>added at the last cycle, AFTER the inflectional ending -l#.
>This makes yer application occur inside-out even with prefixes.
>(--George Fowler)

gsavpiASUCLA.UCLA.EDU (sorry, I don't know his real name) mentioned
that "...there are problems with positing underling jers in
synchronic descriptions of Slavic languages. Ultimately it reduces
to personal belief in whether it's less messy to assume them or to
assume epenthetic vowels, and to what extent synchrony recapitulates
diachrony. Underyling jers buy you a lot less in Russian phonology
than they do in Ukr., Pol., Cz., and maybe others, because Russian
doesn't have changes in vowel quality before jers, historical or
underlying."

Some other references:
Anderson, Stephen. The Organization of Phonology, pp. 243ff.
Stephen Wilson's 1991 Berkeley PhD thesis, "Patterns of Change in
Prosodic Systems." and references cited there. (--Bill Idsardi);
Jolanta Szpyra's article on yers in Language 1992 (--Morris Halle).

Thanks to all!