Projects & Collaborators

Russian palatalization as incomplete neutralization.

Previous studies have found small but significant phonetic traces of underlying contrasts for phonologically “neutralized” contrasts. This phenomenon, often called incomplete neutralization, has been found for final devoicing in many languages, e.g. German (Port & O’Dell, 1985), flapping in American English (Herd et al., 2010), vowel deletion in French (Fougeron & Steriade 1997), vowel epenthesis in Levantine Arabic (Gouskova & Hall, 2009), etc. The present study examines whether Russian palatalization also results in incomplete neutralizations. Russian contrasts plain and palatalized consonants, e.g. /p/ vs. /pj/ with the “plain” stops possibly having a secondary articulation, involving retraction of the tongue dorsum (velarization/uvularization, see Litvin, 2014; Roon & Whalen, 2019; Skalozub, 1963). However, Russian also has stop-glide sequences that form near-minimal pairs with palatalized stops: e.g., /pjot/ ‘drink (3ps pres)’ vs. /pʲok/ ‘bake (3ps past).’ In the environment preceding palatal glides, the contrast between palatalized and plain consonants is neutralized, due to the palatalization of the plain stop: /pjot/à[pʲjot] (derived palatalization). The purpose of the study is to explore whether the neutralization is complete. To do so, we conducted an electromagnetic articulography (EMA) experiment examining temporal coordination and the spatial position of the tongue body in derived and underlyingly palatalized consonants.

Collaborators: Jason Shaw (Yale), Karthik Durvasula (Michigan State University) & Alexei Kochetov (University of Toronto).


Final devoicing in Bulgarian: Incomplete neutralization and L2 experience.

The present study examined the phonetic realization of word-final devoicing in Bulgarian. First, we asked whether final devoicing in Bulgarian was phonetically incomplete. Second, we asked whether the degree of any incompleteness was associated with greater L2 experience in a language without final devoicing (English), as such apparent L1 attrition has been reported previously for Russian. Production results from 34 native Bulgarian speakers indicated final devoicing to be (1) phonetically incomplete, at the group level, on two acoustic measures,  and (2) more severe in speakers with more L2 English experience on two (other) acoustic measures. However, the magnitude of the differences was very small, in line with previous findings on incomplete devoicing in other languages, and likely suggests variation in subtle subphonemic patterns rather than genuine attrition of a grammatical process.

Collaborators: Jason Bishop (CUNY Graduate Center) & Chen Zhou (CUNY Graduate Center) 


Vowel Reduction in Two Varieties of Bulgarian

The present study examines the realization of unstressed vowel reduction in contemporary Bulgarian, part of a larger phonetic investigation of (incomplete) phonological neutralization in monolingual and bilingual speakers of the language. Unstressed vowel reduction in Bulgarian targets non-high vowels, such that the six vowel inventory /e,a,o,i,ǝ,u,/ observable in stressed syllables is said to reduce via raising to a three-vowel /i,ǝ,u/ sub-inventory in unstressed syllables. The goals for this talk are primarily descriptive, and involve answering the following questions about vowel reduction by twelve monolingual Bulgarian male speakers. First, to what extent, overall, is this putative phonological process phonetically complete? Second, how uniform is the degree of reduction across the vowel space (testing front versus back vowels)? Third, how uniform is the degree of reduction across two major varieties of the language (comparing Western/Sofia speakers with Eastern/Balkan speakers)? We discuss preliminary answers to these questions and relate them to (a) previous empirical claims and (b) theoretical models of vowel reduction. 

Collaborators: Jason Bishop (CUNY Graduate Center)


Prosodic structure and Intonational Phonology of the Chungcheong dialect of Korean.

This study examined the underlying patterns of Accentual Phrase (AP) and the prosodic structure of the Chungnam dialect of Korean within the Autosegmental-Metrical phonology framework. Tonal contours from ten speakers of Chungnam dialect were analyzed. To determine the association of the AP-final low tone, the falling slope is calculated by measuring the f0 in the middle of each vowel and by examining the tonal pattern of IP-final AP when the IP ends with a H% tone. The results indicated that Chungnam has three prosodic units higher than a word: IP, ip, and AP. Furthermore, the underlying tonal pattern of AP is THLL (T=H when aspirated or tense consonants; otherwise T=L), where the initial two tones are associated with the AP-initial two moras and the final two L tones are associated with the AP-final two moras. The penult L tone was often undershot when an AP is shorter than 4 moras.

Collaborators: Sun-Ah Jun (UCLA)


Doubly Center-embedded Relative Clause Constructions in Korean

Doubly Center-embedded Relative Clause (henceforth, 2CE-RC) has been known to be remarkably difficult to process and understand (i.e. Chomsky & Miller, 1963) in head-initial languages. A structure of 2CE-RC consists of a main clause and 2 relative clauses where the main clause contains a relative clause (RC1) surrounded on both sides, with the second relative clause (RC2) surrounded on both sides inside it as shown in (1). Furthermore, Gibson and Thomas (1999) have found missing-VP illusion in English where a 2CE-RC sentence may be perceived, wrongly, as equally or more grammatical if VP2 is omitted. Previous studies, however, have found facilitating factors on the comprehension of 2CE-RC in head-final languages, i.e. optimal prosodic parsing (Fodor & Nickels, 2011). However, little is known about 2CE-RC in head-final languages. Thus, this study aims to provide a general understanding of 2CE-RC in Korean by examining roles of prosody and animacy on processing 2CE-RC, and the missing-VP illusion in Korean. To construct a 2CE-RC in head-final languages, objects should be relativized, not subjects, and RC1 must be the ditransitive construction, as shown in (2). 

(1) ENG: [Main NP1 [RC1 NP2 [RC2 NP3 VP1] VP2] VP3]               ex) The girl the man the cat scratched kicked died.

(2) KOR: [MainNP1 [RC1 NP2 [RC2  NP3 ti V1] Ni tj V2] Nj  V3 ]     ex) 빌의 여친은 빌이 그가 본 여자에게 준 명함을 찢었다.

Collaborators: Janet Fodor (CUNY Graduate Center) & Boram Kim (CUNY Graduate Center)