LINGUIST List 3.569

Tue 14 Jul 1992

Disc: Assimilation, Dissimilation, Natural Phonology

Editor for this issue: <>


  1. , Polish voicing assimilation
  2. , Dissimilation
  3. John S. Coleman, Natural Phonology

Message 1: Polish voicing assimilation

Date: 12 Jul 1992 20:13:27 CDT
From: <>
Subject: Polish voicing assimilation

Re Polish voicing assimilation: Alexis Manaster Ramer
writes (3.511): "However, for [obstruent] clusters whose
second element is a voiced fricative, it appears that within
morphemes assimilation is progressive (...)". In support he
cites Hebrew _titrakzi_ 'you will concentrate' which a Polish
person pronounced _titraksi_. Relevant data, as he notes,
are hard to come by, but _gz/lo_ is a dialectal word meaning
'shirt' (/l is "barred l", pron. w). S/lawski (S/low. Etym. j,ez.
polsk.) says the etymology is obscure, but Vasmer (Russ.
Etym. W"orterbuch, at chexol) derives it from *kuzlo. The lax
u dropped and the resulting -kz- assimilated to -gz-
regressively, as is the rule for obstruent clusters. It is
sonorant consonants that assimilate progressively.
After a voiceless obstruent within a morpheme they are
devoiced , e.g., in _teatru_ 'theater (gen.)' but not in
_bobra_ 'beaver' (gen.). This helps explain why the Polish
cognates of Russian _knjaz'ja_ 'princes' and _kniga_ 'book'
are _ksi,e.za_ and _ksi,ega_. The _n_ was devoiced and
lenited to palatal s and its nasality spread to the following
vowel (the nasal e, which I have spelled ,e). The labial
fricative _w_ was a sonorant in Common Slavic and still
patterns like one in Polish. Thus in _krewny_ 'relative'
_w_ is [v] while in _krwi_ 'blood (gen.)' it devoices to [f].
The alternations with _r_ are more complicated. That _r_ is
less voiced in _teatru_ than in _bobra_ is no big
deal; an r is an r. But when _r_ palatalizes to _rz_ it becomes
an obstruent and falls together with _.z_ ("dotted z", the
voiced hushing continuant) when voiced and with _sz_
when devoiced. Thus while the locative of 'beaver' is
_bobrze_ (all voiced), the locative of 'theater' is _teatrze_
with _rz_ pronounced "sz" as the result of progressive
assimilation. On the phonetic surface Alexis is right: _rz_ is
a voiced continuant obstruent that undergoes progressive
devoicing. But so is _z_ and it doesn't. I think a better
description is possible if _rz_ and _w_ are classified as
sonorants for purposes of the voicing rules. The Polish
Hebrew speaker may have said _titraksi_ because (s)he knew
_titragzi_ would be wrong and unified the voicing of the
cluster the other way.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 2: Dissimilation

Date: Sat, 11 Jul 92 12:31:24 EDDissimilation
From: <>
Subject: Dissimilation

It may be only me, but I think it would be desirable to distinguish
different classes of productive rules, in part., what Bloomfield
as far back as the 1920's (presumably echoing the more complex
typology of alternations offered by Baudouin a generation or two
earlier) called 'automatic' from those which involve only a single
affix or morphological class.
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue

Message 3: Natural Phonology

Date: Mon, 13 Jul 92 09:40:46 EDNatural Phonology
From: John S. Coleman <>
Subject: Natural Phonology

It seems that Eric Schiller, like Richard Goerwitz, missed the beginning
of this discussion. I did not criticize any aspect of Natural Phonology
in the way his comment seems to suppose, so a review of the literature
is not called for. The objection I raised was this: a proponent of
Natural Phonology (can't remember who it was now) recently claimed
that Natural Phonology was the ONLY theory to address phonetic data.
I think this claim is ignorant and false. To refute it it is not
necessary to examine the Natural Phonology literature. It suffices
to find an example of just one other phonological theory which
also attends to phonetic data. Firthian prosodic phonology is
one such case: I could cite others. But in my original comment
I made the stronger claim that
Natural Phonology addresses the same pool of phonetic "data" as
other phonological theories. Even if Natural Phonology placed some
new "data" into this pool, proponents of other theories of phonology
have also taken them up in the meantime. So whatever the value of
Natural Phonology may be as a contribution to the field, it is not
distinguished by the data which it addresses.

Could I respectfully suggest that LINGUIST contributors read
postings carefully before jumping in aflame.

--- John Coleman

Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue