Productivity in Polish o-Raising
Maciej Baranowski and Gene Buckley
University of Pennsylvania
Modern Polish shows raising of /o/ to [u] before a voiced word-final consonant, except nasals (Gussmann 1980): pora ~ pur ‘time’, moja ~ muj ‘my’; but dom ‘house’, lot ‘flight’. The alternation is now opaque in many words due to historically later word-final devoicing (Stieber 1973): raising occurs before underlying /d/ in broda ~ brut ‘beard’, but not before /t/ in lot ‘flight’. Kenstowicz (1994) gives a standard ordering analysis: raising before a voiced sound precedes the devoicing of final obstruents. Yet there are many exceptions that cast doubt on this generalization and indicate that lexical listing plays at least a partial role. In the fully lexicalized approach we adopt, the historical role of voicing has been replaced by selective vowel-feature prespecification, such that some tokens of /o/ alternate with [u] while other tokens are stable.
A lexical approach predicts that assimilation of new words will be significantly affected by existing patterns in the lexicon rather than by a categorical rule environment (Skousen 1989, etc.). Elicitation of nonce words from 16 adults produced strong evidence in favor of word-similarity, i.e. analogy, in the choice of whether to raise or not. Nonce forms similar to gondola ~ gondol were raised just 27% of the time, compared to 59% for words similar to topola ~ topul ‘poplar’, though both have the same stem-final consonant. Those similar to non-raising kaczor ‘gander’ raised less than 1%, while those similar to potwur ‘monster’ raised 53%, with the most similar words (with medial stop+w cluster) raising most of all (Skousen’s proximity). The similarity effect is also found for the irregular words with voiceless stops. Nonce words resembling sobota ~ sobut showed a small amount of raising (25%, kept small by the competing gang effect of the very numerous non-raising stems with voiceless stops), while those resembling zatoka ~ zatok ‘bay’ raised than 1% (since the gang effect and proximity agree on a non-raised output). That gondola-resembling forms raised at all is attributed to the competing effect of other analogs in the lexicon, such as soli ~ sul ‘salt’.
In the latest series of elicitations, twenty adults produced nonce forms with stem-final nasals. There is no raising to /u/ in such stems in actual words in standard Polish (e.g. wrona-wron), hence the expected level of raising was low (gang effect). However, most speakers show a degree of raising in forms closely resembling real “raising” words with a stem-final oral consonant, e.g. nagrona-nagroda (proximity). The fact that there is any raising at all in stems ending in a nasal consonant casts doubt on the traditional rule-based approach whereby raising may only occur before a word-final oral consonant.
Raising in Polish vocabulary (with positive and negative exceptions to the historical generalization) and the treatment of nonce forms by speakers provide strong support for a lexicalized and analogical approach.
References:
Albright, Adam, and Bruce Hayes. 2001. Rules vs. analogy in the English past tense: A computational/experimental study. Ms., UCLA.
Baranowski, Maciej, and Gene Buckley. 2003. Lexicalization and analogy in Polish o-Raising. LSA Annual Meeting. Atlanta, January 2-5.
Bethin, Christina. 1978. Phonological rules in the nominative singular and genitive plural of the Slavic substantive declension. Dissertation, University of Illinois at Urbana.
Booij, Geert, and Jerzy Rubach. 1987. Postcyclic versus postlexical rules in Lexical Phonology. Linguistic Inquiry 18, 1–44.
Brooks, Maria Zagorska. 1975. Polish reference grammar. Mouton.
Buckley, Eugene. 2001. Polish o-Raising and phonological explanation. LSA Annual Meeting, Washington, DC, January 4–7.
Gussmann, Edmund. 1980. Studies in abstract phonology. MIT Press.
Grzegorczykowa, Renata, and Jadwiga Puzynina. 1973. Indeks a tergo do Słownika języka polskiego pod redakcją Witolda Doroszewskiego. Państwowe Wydawnictwo Naukowe.
Kenstowicz, Michael. 1994. Phonology in generative grammar. Blackwell.
Kenstowicz, Michael, and Charles Kisseberth. 1979. Generative phonology: Description and theory. Academic Press.
Sanders, Nathan. 2001. Preserving synchronic parallelism: Diachrony and opacity in Polish. GLiP-3, University of Warsaw.
Skousen, Royal. 1989. Analogical modeling of language. Kluwer, Dordrecht.
Steele, Richard D. 1973. The segmental phonology of contemporary standard Polish. Dissertation, Harvard University.
Stieber, Zdzisław. 1973. A historical phonology of the Polish language. Carl Winter.
Tokarski, Jan. 2002. Schematyczny indeks a tergo polskich form wyrazowych. Edited by Zygmunt Saloni. Warszawa: Wydawnictwo Naukowe PWN.