UC Berkeley PhonLab Annual Report

Rhythmic Repair of Morphological Accent Assigned Outside of a Metrical Window

Most current models of grammar assume that syntax has no sensitivity to phonological information (Pullum and Zwicky, 1986, 1988). Phonologically determined agreement, also called alliterative concord, challenges the assumption that syntax is phonology-free, because it appears that phonological form determines morphosyntactic agreement. Here I present a pattern of phonologically determined agreement from Gu´ebie, an endangered Kru language spoken in Cˆote d’Ivoire, assessing whether phonologically determined agreement is, in fact, phonologically determined. I show that with a combination of Distributed Morphology operations (Halle and Marantz, 1994) plus category-specific phonological grammars (Smith, 2011) via Cophonology Theory (Orgun, 1996; Anttila, 2002; Inkelas and Zoll, 2005), we need not modify our model of syntax as phonology-free. In addition to accounting for phonologically determined agreement in Gu´ebie and across languages, the proposed analysis includes a formal account of ellipsis via constraints at PF.

Lusoga Noun Phrase Tonology

In this paper we investigate the distribution of vowel systems in the Macro-Sudan Belt, an area of Western and Central Africa proposed in recent areal work (Güldemann 2008, 2011; Clements & Rialland 2008). We report on a survey of 615 language varieties with entries coded for two phonological features: advanced tongue root (ATR) harmony and the presence of interior vowels (i.e. non-peripheral vowels, such as [ɨ ɯ ɜ ə ʌ …]). Our results show that the presence of ATR harmony in the Macro-Sudan Belt is limited to three separated zones: an Atlantic ATR Zone, a West African ATR Zone, and an East African ATR Zone, all geographically unconnected to one another. We additionally show that between the West and East African ATR Zones is a geographically extensive, genetically heterogeneous region of Central Africa where ATR harmony is systematically absent which we term the Central African ATR-less Zone. Our results also show a large region where phonemic and allophonic interior vowels are disproportionately prevalent, which we term the Central African Interior Vowel Zone. This zone noticeably overlaps with the Central African ATR-less Zone, suggesting that ATR and interiority have an antagonistic relationship. Chi-squared tests support the presence of a strong relationship between the two types of vowel contrasts.

Phonetic accommodation is the process in which a speaker becomes more phonetically similar to his or her interlocutor over the course of a conversation. This experiment investigates phonetic accommodation in the English speech of Mandarin speakers after exposure to a model speaker who shares their language background. The results show that when including phones, tasks, and conditions as dependent variables, there are statistically significant differences across tasks and phones. Phonetic accommodation is observed in all shadowing tasks and the effect remains in post-shadow task in some dependent variables. The social manipulation of this study is only statistically significant in formant durations and word-final consonant clusters durations and the pattern suggests that subjects who were in the condition that encouraged speakers to achieve a closer social distance with the interlocutor accommodated more than subjects in the condition that were encouraged to be native-English like. This experiment thus contributes to the understanding of phonetic accommodation in second language English speakers under different conditions.

Sound symbolism is the hypothesized property for sounds to convey semantic meaning. Shinohara and Kawahara (2010) proposed that features of vowels (frontness, height) and obstruents (voicing) cause listeners to perceive words as either larger or smaller. Study 1 firstly replicates the original experiment then repeats the experiment using a speech perception paradigm. The speech perception experiment assesses whether listeners perceive sizes differently between spoken language and visual reading. The results from Study 1 were consistent with Shinohara and Kawahara (2010) except that words with /u/ were perceived as smaller in our results. We hypothesized that this result may be due to u-fronting which is an iconic feature of Californian English so we repeated both the written word and speech perception experiments in Study 2 with non-Californian English speakers. Our results support Shinohara and Kawahara’s claims and suggest that speakers perceive dialect-specific phonetic properties from written word.

Bantu Tone Overview

Word prosody II: Tone systems

In Search of Prosodic Domains in Lusoga

In this paper I present and update some of the major points Russell Schuh and I made in our 1974 Linguistic Inquiry paper concerning universals of tone rules. Emphasis is on the distinction we made between synchronic and diachronic naturalness. Any diachronic change can be a synchronic rule while the reserve is not the case. We suggest(ed) that it is profitable to talk about natural synchronic rules that could not be (phonetically motivated) sound changes. This includes tone shifting, tonal polarity, and tonal downstep among possibly other commonly occurring tonal phenomena.

In this paper I consider two analyses of the possessive pronoun tonal paradigm in Tswefap, a Bamileke language spoken in Batoufam, Cameroon. As in the case of related languages that have been previously described, Tswefap has a rather complex tone system that involves multiple tone heights, tonal contours, and tone alternations. Although simplified, it also maintains several of the inherited noun class distinctions. In this study attention is on the tones of possessive pronouns and their effects on a preceding modified noun. I first present a paradigmatic account as one might find in a descriptive or pedagogical grammar indicating which possessive pronouns receive which tones. I then turn to a more traditional Bamileke and Grassfields Bantu analysis in terms of underlying representations and floating tones. It is argued that all possessive pronouns are preceded by a floating L tone which affects a preceding mid tone noun in one of two ways, depending on the syllable shape of the pronoun: (i) if the pronoun begins with a consonant, the mid of the noun becomes a mid to low contour tone; (ii) if the pronoun has consists solely of a vowel, the mid is raised to a high tone. Although I argue for the latter analysis, I conclude by demonstrating that alternate tonal variations indicate on-going change which may ultimately undermine the more abstract phonological analysis in favor of a considerably simplified paradigmatic tone assignment.

Costs and Cues to Code-switched Lexical Access

Recordings of read speech in Korean and English were made by native South Koreans and Korean Americans of varying generational status (“second-generation” American-born or “1.5-generation” foreign-born) and analyzed for differences in usage of VOT and fundamental frequency to contrast production of Korean lenis and aspirated stops and affricates. The speech was then played back to listeners of Korean heritage and judged metalinguistically regarding proficiency in Korean and other attributes relevant to the speech and demographics of immigrant speakers. Results show that second-generation Korean speakers, especially females, are not showing the collapse of VOT contrast found in the other two groups, one part of the “tonogenetic” sound change nearing completion in Seoul. Female second-generation speakers are also not using f0 to differentiate between the stops to the extent that first- and 1.5- generation speakers are. These second-generation speakers were easily identifiable as having been born in the United States, but the correlation with their generational identification and use of VOT and f0 to contrast lenis and aspirated stops and affricates is mild. It is concluded that because second-generation Korean Americans vary in their production of Korean, there is no clear sociophonetic marker for a Korean American “variety” of the language. The most proficient second-generation speakers closely resemble native speakers and do demonstrate the tonogenetic sound change, but the least proficient second-generation speakers diverge from this norm in a variety of ways. Second-generation Korean American speakers are easily identifiable as not speaking in the same way as native South Korean speakers, although this does not hinge on their use of the f0 contrast. The analysis makes a stronger case for applying new models of language acquisition, speech production, and identity formation to heritage language speakers that differ from those used for bilingual speakers.

Frequency can often predict when children will acquire units of language such as words or phones. An additional predictor of speech development may be a phone’s functional load (FL), or the contrastive work that a sound performs in a language. A higher FL may correlate with earlier phone emergence in child speech as children selectively converge upon highly meaningful contrasts in their input. This hypothesis is tested across five typologically-diverse languages that vary by phone inventory size and structure as well as word composition. Consonant FL was calculated over more than 390,000 words of child-directed speech. Results demonstrate that FL correlates positively with earlier consonant emergence in all languages. Models fit to bootstrapped corpus data include both FL and frequency as predictors, but suggest that frequency may be the stronger of the two. A need to complicate assumptions on the relationship between environmental effects and phonological development is discussed.