Spring 2021
February 5
Maksymilian Dabkowski (UC Berkeley): Stress in Paraguayan Guaraní
I describe and analyze stress assignment in Paraguayan Guaraní (Tupian, ISO 639-3: gug). In PG, final stress predominates at the level of prosodic words and phonological phrases alike. However, non-final stress is also attested in exceptional roots, morphologically complex words, and multi-word phonological phrases. I propose that there is one basic mechanism for stress assignment in PG: In the course of prosodification, stress targets the right edge of a prosodic constituent. This captures PG's predominantly final stress. Finally, I propose that non-final stress is a consequence of extrametricality, extraprosodic suffixation, and stress clash avoidance, which correctly restrict deviations from stress finality observed in Paraguayan Guaraní.
February 12
Florian Lionnet (Princeton University): Downstep in Paicî: between accent and tone
In this talk, I propose a description and analysis of tone in Paicî, a language of New Caledonia, and one of the rare Oceanic languages to have developed a tone system without external influence. Building on Jean-Claude Rivierre's (1974, 1978) work and drawing from fieldwork data, I show that the tonal system of Paicî is best described with three underlying primitives: two tones (High vs. Low), and a downstep /ꜜ/. The Paicî downstep is particularly interesting for the empirical documentation as well as the typological and theoretical understanding of downstep, because it combines many rare properties, including the following:
(i) it only affects L tones, and is attested only after a L-tone or utterance-initially;
(ii) it is its own phonological object, a register feature separate from tone and only indirectly interacting with it;
(iii) it is not triggered by a floating tone;
(iv) it is “total” in Meeussen’s (1970:270) terms, i.e. it lowers the register in such a way that the new “ceiling” corresponds to the former “floor”;
(v) it is culminative and demarcative within the prosodic word and is partly conditioned by metrical structure.
Many of these properties have a clear accentual flavor. The Paicî prosodic system thus appears to consist of two subsystems: a non-ambiguous tone system (with a H vs. L lexical contrast), with no accentual properties (hence not “pitch-accent”) – and a parallel system best described as a defective accentual system in the early stages of tonologization, marked by downstep. Preliminary comparative evidence from neighboring and closely related (non-tonal) Xârâcùù shows that the Paicî downstep pattern is indeed very likely to be the tonologized descendant of a former accentual system which historically predates the innovation of tonal contrasts in Paicî.
February 19
Hannah Sande (UC Berkeley)
Check-in meeting facilitated by HannahSande: come share current and upcoming projects and get to know our newest faculty member!
February 26
Susannah Levi (NYU): Talker familiarity helps speech perception. Does the benefit stop there?
To understand speech, listeners must parse the highly variable acoustic signal into appropriate, language-specific phonological categories and generalize these categories to novel stimuli in order to perceive words correctly. Previous research has found that listeners are better at processing speech from familiar speakers, typically measured in a spoken word or sentence recognition task. Work in the lab extended these findings to children with a range of language skills. The next stage of our research will test whether the benefits at the level of speech perception impact higher-level processing.
March 5
Katie Russell (UC Berkeley): Grammatical Tone and Length in Gã
In Gã [Kwa; Ghana], many tense, aspect, mood and polarity (TAMP) distinctions are exponed tonally. When the subject of a clause is pronominal, these grammatical tones are realized on a subject pronoun prefix: with non-pronominal subjects, they are exponed on verbal TAMP prefixes instead. Prior literature centers around two main analyses to account for these alternations. Kropp Dakubu (2002) describes the processes as resulting from the deletion of the TAMP prefix after a pronoun, in which the tone of the pronoun delinks and reassociates to the pronoun. Paster (2003), on the other hand, argues that the process of prefix deletion after pronouns is not part of the synchronic phonology, instead arguing for an analysis in pronouns are analyzed as portmanteau STAMP (subject TAMP) morphs. In this talk, I present new data collected with a native speaker of Gã that sheds light on these processes, specifically through examples involving intervening PP constituents, which have not previously been discussed in relation to grammatical tone in Gã. I argue that this data provides evidence that a portmanteau analysis, as proposed by Paster, is not tenable for all contexts. Instead, I take the position that subject pronoun prefixes undergo fusion with aspect and mood prefixes. Additionally, I present novel data connecting progressive marking, which is exponed suprasegmentally through vowel length, to grammatical tone.
March 12
Anna Mai (UCSD): On the Relationship Between Interaction and Complexity in Phonology
In this talk, I will present work done jointly with Eric Meinhardt, Eric Bakovic, and Adam McCollum on classes of regular functions at the outer edge of the complexity expected for phonological patterns. In particular, I will focus on the class of weakly-deterministic functions, a class of regular functions previously hypothesized to represent an upper bound on phonological complexity. Building on Elgot & Mezei's (1965) demonstration that regular functions can be decomposed into two contradirectional subsequential functions, I will present a definition of weakly deterministic functions as those for which the two contradirectional subsequential composands do not interact, in a way that will be formally clarified. In doing so, this work makes connections with more familiar notions of interaction in the phonological literature and provides greater insight into what distinguishes phonological patterns that require the greatest expressivity of regular functions (i.e., non-determinism) from those that require properly sub-regular expressivity (i.e., weak determinism).
April 2
Karee Garvin (UC Berkeley): Syllabification, Stress, and Gestural Diagnostics in English
Phonological theory makes strong predictions about syllabification in English, where phonological models typically syllabify a sequence of VCCV as V.CCV, so long as the cluster is a permissible onset. This is supported by typological data, which suggests that codas are more marked than onsets. Models of gestural coordination, like the coupled oscillator model, similarly lend support to the principle of onset maximization, arguing that onsets are synchronously planned with the following vowel, global timing, whereas codas are timed sequentially, local timing; as a result, onset+nuclei are more stable than codas. While literature on gestural coordination has demonstrated this pattern of coordination for onsets in English, studies have focused on word initial and stressed syllables, and thus, whether these coordination patterns extend to medial and unstressed syllables has yet to be determined. Furthermore, phonological models predict that word medial sequences should follow word boundary syllabification; however, typologically common phonological patterns like lenition and distributional patterns like edge-effects and extrametricality suggest that word medial syllable margins are distinct from word boundary syllable margins. In this presentation, I will present Electromagnetic Articulography (EMA) data to illustrate the effect of stress and word position on articulation and syllabification in English to show that our current phonological and gestural models are insufficient for predicting and modeling word medial syllabification. Instead, I propose that the coordination of jaw oscillation to segmental sequences serves as a diagnostic for syllabification that can be leveraged in both gestural and phonological models to better capture the attested patterns of syllable structure cross-linguistically.
April 9
Maho Morimoto (National Institute for Japanese Language and Linguistics [NINJAL])
April 16
Natasha Warner (University of Arizona): Perception of all English sound sequences: The roles of probability and duration (with collaboration from Seongjin Park)
April 30
Yuni Kim (University of Essex): Phonologically conditioned allomorphy in Amuzgo inflectional tone
In Amuzgo (Oto-Manguean; Mexico), tonal alternations contribute to person/number marking. Although there is a large number of lexically arbitrary tonal inflection classes (Kim 2016, Palancar 2020), in certain grammatical contexts the inflectional tones become predictable based on the underlying lexical tone and presence/absence of a final glottal stop on the verb stem (as briefly alluded to in Sande 2019 and Kim & Sande 2020). This is a case of phonologically conditioned allomorphy, since no plausible phonological operation can relate lexical tones to inflectional ones. Yet, the pattern does not sit easily with standard approaches to phonologically conditioned allomorphy such as subcategorization or markedness-based phonological output optimization. Rather, I analyze the pattern as involving the interaction of input-output faithfulness with arbitrary (but independently motivated) morphotactic constraints. I observe that while faithfulness is not normally involved in allomorph selection, tone is more prone to replacive morphology than other types of exponents, opening the door to faithfulness as a mechanism. Furthermore, the Amuzgo pattern supports a cophonology-based approach, since the crucial interaction between faithfulness and morphotactics cannot be accomplished with concatenation and interleaving (pace e.g. Bye & Svenonius 2012). I conclude that tone is very special (Hyman 2011, inter alia).
Fall 2021
September 3
Round robin
Sepember 10
David Gaddy (UC Berkeley): Decoding silent speech with electromyography
September 17
Jon Rawski (San Jose State): Abductive Learning of Phonotactic Constraints
September 24
AMP 2021 practice talks
October 1
Yevgeniy Melguy (UC Berkeley): Mechanisms of listener adaptation in perceptual learning for speech
October 8
Zion Mengesha (Stanford): A Social Meaning Perspective on Vowel Trajectories: The Realization of FEEL and FILL Among African Americans in California
October 15
Evelin Balog (Erlangen-Nuremberg): Entrenchment revisited: Old and new concepts and their empirical validation
October 22
Natalie Weber (Yale): Correspondence of prosody and syntax by phase in a polysynthetic language.
October 29
Chloe Willis (UC Santa Barbara): The Theoretical and Methodological Implications of Bisexuality in Language and Sexuality Research
Voices communicate not just what we say, but also who we are. The past three decades have seen an abundance of research on how sexuality is indexed through the voice. This work mostly focuses on stereotypes about sounding gay, especially as they relate to the pronunciation of /s/ and the “gay lisp” (e.g., Munson et al. 2006a,b; Campbell-Kibler 2011; Zimman 2017), whereas lesbian-sounding voices are less represented (e.g., Van Borsel et al. 2013; Barron-Lutzross 2015). Bisexuality is conspicuously absent in this literature. In this talk, I first overview my previous work on bisexuality and /s/ production, namely that bisexual women and men produce /s/ in a way that is distinct from their lesbian, gay, and straight counterparts (Willis 2021). Next, I present preliminary findings that directly expand upon this work. These findings suggest that ethnoracial identity—which is typically not considered or even reported in previous research—is a significant predictor for variation in /s/ production. Finally, I discuss the theoretical and methodological implications of these two analyses for the broader study of language and sexuality and identify how my current research addresses these issues.
November 5
Katie Russell (UC Berkeley): Nasal harmony and interactions with lexical strata in Paraguayan Guaraní
November 19
Caitlin Smith (UNC Chapel Hill) and Charlie O'Hara(University of Michigan): Learning Derivationally Opaque Patterns in the Gestural Harmony Model
In this talk, we examine the learnability of two apparently derivationally opaque vowel harmony patterns: attested chain-shifting height harmony and unattested saltatory height harmony. We analyze these patterns within the Gestural Harmony Model (Smith 2018) and introduce a learning algorithm for setting the gestural parameters that generate these harmony patterns. Results of the learning model indicate a learning bias in favor of the attested chain-shifting pattern and against the unattested saltation pattern, providing a potential explanation for the differences in attestation between these two derivationally opaque patterns. Furthermore, we show that feature-based learning models of these patterns show no such learning bias and provide no account of the typological asymmetry between chain-shifting and saltatory height harmony.
December 23
Zachary O’Hagan (UC Berkeley): Verbal Reduplication in Caquinte
Reduplication in the strongly headmarking Nijagantsi Arawak languages of western Amazonia -- namely in Asheninka -- has figured prominently in the theoretical literature (e.g., Spring 1990, McCarthy and Prince 1993:25-108, Downing 2005), based on original data and analysis only from Payne (1981:143-152; see Martel Paredes 2012 and Mihas 2015:86-90 for some data from the Perené dialect). While multiple patterns of reduplication are attested in Asheninka, comparable data from related languages is generally lacking (cf. Beier 2010 and Michael 2008 for Nanti, and Snell 2011:832 for Matsigenka). In this presentation I profile and augment Swift's (1988:126-131) data on reduplication in Caquinte, introducing patterns not attested in Asheninka and proposing some ways in which possible analyses contrast with those argued for for Asheninka.
In the basic case, the Caquinte reduplicant is a trimoraic suffix following the stem (i.e., the verb root and possibly an epenthetic segment and/or the reversative suffix -rej). In the simplest case, the reduplicant is composed of the first non-initial CV.CV.CV (e.g., imatsagamatsagatakaro, with the root underlined), so long as the stem -- most often the root -- is at least three moras, such that suffixes are not included in the reduplicant (n.b., with the exception of -rej) . If the stem is fewer than three moras, a suffix -i is added to the reduplicant (e.g., nokenakenaibaepoji); and if the root begins with a vowel, the vowel is not included in the reduplicant (e.g., nanijinijiitanake). If the addition of -i is insufficient for reaching three moras, then a subject prefix and (when relevant) the initial vowel of a root are included in the reduplicant (e.g., notenoteitanaka). Other patterns include ones where -i is absent, where it applies unnecessarily given the above formulation, where -a seems to have the same function as -i, where there seems to be a lexicalized reduplicative stem (especially with roots ending in velars), and where the reduplicant appears to be prefixal. I conclude by discussing whether the root augmentation present in Asheninka (e.g., McCarthy and Prince 1995:46-59) is present in Caquinte.