Earlier relativist notions about color naming have yielded to the recognition that color categorization is a linguistic universal. The first comprehensive argument for universality is made by Berlin and Kay (1969), who propose a total possible inventory of eleven basic color categories. Subsequent work in bilingualism and prototype theory has led to refinements of Berlin and Kay's original thesis. This paper, which includes a formal color experiment, examines the treatment of the Russian color terms sinij 'dark blue' and goluboj 'light blue' within the framework of this research. The experiment includes informants from four groups: 1) Soviet Russians; 2) adult emigre acutes; 3) young adults who emigrated during childhood; and 4) Americans tested in English. Results suggest that sinij and goluboj are bona-fide basic terms in standard Russian and that this treatment is fixed by adulthood. Among the younger emigre acutes, however, there is definite evidence of semantic shift, the result of interference from English blue. The experiment helps confirm the theory of basic color categories as well as its addenda and revisions.
andrewsd@guvax.acc.georgetown.edu

An analysis of clitic word order is proposed, based on the division of Bulgarian verb-complex clitics into core and peripheral with respect to the clitic cluster formation. Taking into account inherent prosodic properties, the treatment of the "movable" core clitics is separated from that of the peripheral strictly proclitic and strictly enclitic elements, which allows for attribution of apparently problematic clitic placements to the interaction of the two types.
tania@coli.uni-sb.de

Inflectional endings are assigned in languages by general principles, but these can come into conflict. We address the question of how such conflict is resolved. A particularly complex example is the Russian genitive plural, where we find that with soft-stem nouns there is a conflict between exponent assignment according to declension class and a default exponent assignment for soft-stem nouns. What is specially interesting is that the conflict here can be resolved by reference to subsystems over and above the paradigm, such as stress. We present an explicit account of the conflict and its mediation by basing our study on default inheritance. For this purpose we make use of the lexical knowledge representation language DATR. This allows us to demonstrate in the output provided that the correct forms are indeed predicted by our theory.

d.brown@surrey.ac.uk
a.hippisley@surrey.ac.uk

Slavic, as represented by Old Church Slavonic, exhibits a curious parallelism of "hard" and "soft" declensions based on the final consonant of the stem, which may be neutral or palatal. Many endings then begin with back versus front vowels. This is a most un-Indo-European feature, for IE is supposed to have had only one set of endings per declensional type, and suggests some strong phonetic influence on the emerging Slavic language, which is most likely to have come from the Huns or Avars, probably Turkic -speaking peoples, who dominated the Slavs between ca. 400-800 A.D. In their agglutinative language, front or back vowels in the stem require corresponding front or back vowels in all suffixes, and the process of attachment also affects the intervening co nsonants. In some consonants, such as velars and laterals, this effect is particularly marked, and there is a curious back counterpart of front /i/, a vowel like the Russian /y/, which is quite un-Indo-European. Its source as well as that of the three suc cessive palatalizations which set off Slavic from its Baltic matrix is probably to be sought in an Altaic influence which asserted itself in Slavs seeking to imitate the speech habits of their Altaic masters and military commanders. The grammatical system was not imitated on anything like this scale, but more words than commonly realized were borrowed, including the very name of the Slavs.

Institut für Slawistik, Liebiggasse 5, Wien 1010, Austria

This paper examines the structure of li yes-no questions and the distribution of focused elements in them. Li is a clitic complementizer which assigns a focus feature. If Spec-head agreement occurs, a maximal projection moves to SpecCP, where it is the focus of the question and hosts the clitic. If no maximal projection moves to SpecCP, then the verb in I^0 undergoes head-movement to C^0 in order to host the clitic. In these verb-initial structures, the entire clause is questioned. If the clause contains a focused constituent marked by stress, then that constituent is the focus of the question; the resulting reading is similar to what would result if the focused constituent had moved to SpecCP. However, if there is no stressed, focused constituent, the result is a "simple" yes-no question.
thking@csli.stanford.edu

The article examines the syntactic phenomenon of na-drop, its distribution, and its implications for the nature of object doubling in Bulgarian. Na-drop is the optional omission in colloquial Bulgarian of the dative marker na from the object NP in a dative reduplicative sentence. That the dative pronominal clitic (PC) in such constructions operates as the sole dative marker for the reduplicated object NP suggests that Bulgarian doubling PCs in general may have a strong case-marking function. Testing with 23 native speakers shows that na-drop is tolerated well beyond its historical environment (doubling of 1sg and 2sg long-form pronouns). The subjects as a group found na-drop acceptable, to varying degrees, throughout the personal pronoun paradigm and with reduplicated object nouns and personal names. A major factor influencing acceptability was the position of the reduplicated object NP in the sentence. Tentative results also suggest a higher tolerance of na-drop in impersonal sentences.
vakarel@oregon.uoregon.edu

English translation of a sample entry from the Dictionnaire des mots du discours en russe contemporain, providing a full description of the discouse functions of the collocation eshche raz.
remi.camus@univ-caen.fr

Three examples from East Slavic which have been cited as evidence that analogy can produce new phonemes are reexamined. It turns out that in each case the forms in question can be naturally explained as borrowings from a dialect in which the "new" phonemes had arisen by regular sound change into dialects without these phonemes.