Proceedings Published by the LSA (Linguistic Society of America)
Not a member yet
4323 research outputs found
Sort by
Gender assignment strategies in Spanish-English mixed noun phrases
Full text linkThis study explores how Spanish-English bilinguals in California assign grammatical gender when code-switching between a Spanish determiner and an English noun. In bilingual speech, these moments of cross-linguistic contact offer a window into how speakers navigate competing grammatical systems. Using data from a semi-spontaneous picture description task, the analysis draws on 649 mixed noun phrase tokens to examine whether bilinguals rely on analogical gender, semantic cues, phonological endings, number, or determiner type when choosing between masculine and feminine determiners. While masculine determiners were strongly favored overall, the results show that analogical gender plays a meaningful role—feminine determiners were more likely to appear when the English noun had a grammatically feminine Spanish translation. Other factors did not reach statistical significance. These findings support the view that gender assignment in code-switching reflects structural sensitivity to the grammatical systems of both languages, as well as variability shaped by individual bilingual experience
Phonetic duration is more variable than phonological duration
Full text linkDuration may be phonologically meaningful, as with contrastive segment length or as a cue to stress, or it can be purely phonetic, as with final lengthening. This paper explores the amount of durational variation present in the articulation of rimes that have different sources of duration. We show evidence to support the hypothesis that phonological duration is more stable and phonetic duration is more variable. We present evidence from new analyses of two English production studies (Lunden 2016, 2017) with nonce words that show significantly greater durational variability due to final lengthening than durational differences due to vowel quality or stress. Variability in duration was calculated by looking at the residuals from Generalized Linear Models of raw duration. Subsequent homogeneity of variance tests (Levene\u27s tests) were performed as part of one-way ANOVAs on the saved residuals as the dependent variable. We suggest that the greater variability of phonetic duration is plausibly responsible for weakening effects such as the avoidance of stress and avoidance of long vowels and geminates word-finally, as it is perceptually difficult to signal duration-based phonology in a position with highly-variable duration.
Finding ‘language’ in the Hebrew Bible
Full text linkIn the Hebrew Bible, lashon [לָשׁוֹן] and safah [שָׂפָה] are both translated as ‘language’. In Genesis 10, survivors of the flood went their separate ways each according to their lashon. In Chapter 11, the builders of the Tower of Babel all spoke the same safah. Focusing on lashon, it changes radically over the Biblical millennium. From ~1400 BCE, lashon describes ways in which individuals speak. By 550-450 BCE, it was synonymous with ‘nation’ or ‘people’ and used in a manner more familiar to us today
Distribution of evidential markers in a Cuzco Quechua corpus
Full text linkDue to contact, several varieties of Quechua are losing evidential markers. Using a Cuzco Quechua corpus (Macedo et al., 2022), we present the distribution of evidential markers -mi (-n,-ni), -si (-s, -sis), and -chá on discourse connectors (DCs) (Sanchez et. al., 2021) and sentence-level constituents (subject, object, verb, and adverb). To determine how DCs and constituents are marked with evidentiality, we analyze a corpus of semi-structured interviews in Quechua (N=29; 995 utterances total). 3.42% of evidential suffixes were found on DCs and 96.58% on constituents indicating that their syncretic nature (focus and evidentiality) is at play in constituents. A GLMM suggests that evidential type (-mi, -si and -chá) is predictors of DC and constituent marking. -mi is more likely to appear in both syntactic positions, while -si shows the greater shift from DC to constituents (p<0.05). Men are also more likely to mark DCs and less likely to mark constituents
Semantic change of female-denoting nouns in diachronic German corpora
Full text linkComputational studies of language change using word embedding models have focused on lexical semantic shifts. Current work in Natural Language Processing has revealed gender bias in word embeddings trained on internet data. However, little is known about what distributional methods can reveal about possible gender bias in diachronic corpora. This paper addresses this gap by examining gender-specific nouns denoting humans, e.g. Magd ‘maidservant’, Knecht ‘farmhand’ in a historical corpus of German (1350-1899). We show that word embedding models provide evidence for gender bias and negative evaluations associated with female-denoting terms in diachronic German corpora
A more comprehensive view on the production-perception link of cue weighting in sound change: The case of vowel length contrast in Long’an Zhuang
Full text linkThis paper aims to provide a more comprehensive view on the production-perception link of cue weighting in sound change by further examining the relationship between the weights of different cues in production and perception, as well as comparing two ways of cue weight calculation (raw vs. relativized). Production and perception experiments were conducted on Long’an Zhuang (a Tai language in Guangxi, China) vowel length contrast, which involves both duration and spectral cues. The results showed a negative correlation between duration and formant cue weights in production but a positive correlation in perception. Additionally, the production-perception link was found to be more robust regarding relative cue weights rather than raw weights. These findings suggest that sound change may be related to asymmetry in production and perception, weakening the production-perception link in absolute cue contribution but not in relative dominance.
Historical productivity of VERB-NOUN compounds in English
Full text linkEnglish shows an exocentric verb-noun compound type with an uninflected verb followed by a noun object, e.g. pickpocket, where noun pocket is the object of verb pick. These “pickpocket compounds” first appeared in English under French influence post-1066 (Marchand 1960:37–39). We delve into English pickpocket compounds with the OED as our primary source. Despite fluctuations in frequency for individual forms throughout their history, we demonstrate pockets of productivity analogically involving compounds with similar semantics and compounds with the same verb as first member, or similar nouns as second. These were productive enough to compete directly with and even be preferable to more usual noun-verb compounds with similar meaning and components; for instance, sweepchimney (1657) predates chimneysweep (1709) by decades. Particularly telling regarding productivity into the 20th century are over 20 pickpocket coinings by humorist James Thurber (1951). These examples show that although pickpocket compounds have always been rare, they have enjoyed sustained productivity in English. Viewing these compounds through an analogical lens, versus a general compounding rule, better predicts the pockets of observed productivity
The strength of conservativity: evidence from learnability experiments
Full text linkOne of the best-known semantic universals is determiner conservativity, roughly, the idea that the truth of sentences like most fish swim depends only on the things named by the determiner’s first argument (fish). Conservativity has been argued to reflect a fundamental property of grammatical architecture, a conclusion bolstered by evidence that adults (and children) can pair novel determiners with conservative meanings but not with minimally-different non-conservative meanings. Here we extend this work to ‘weakly’ conservative meanings (ones for which the truth of the quantificational sentence depends either on the determiner’s first argument or on its second argument). Such meanings are classically non-conservative but would be permitted under a weakened version of the generalization, which was designed to accommodate meanings like those of only and even. We find that adults cannot learn these kinds of novel determiner meanings either. This result suggests that the classical understanding of conservativity (on which ‘weakly’ conservative meanings are ruled out by virtue of being non-conservative) better describes the constraint that learners embody
On gender stereotypicality in nouns and adjectives: Comparing humans, large language models and text-to-image generators
Full text linkBoth humans and large language models (LLMs) are known to exhibit effects of gender stereotypicality. We conducted a series of studies to systematically assess to what extent humans’ and LLMs’ interpretational patterns align, how different kinds of linguistic expressions (role nouns vs. adjectives) contribute, and to what extent these patterns extend to text-to-image models. Experiments 1 and 2 test how gender-biased role nouns (e.g. plumber, nurse) and adjectives (e.g. powerful, kind) influence humans’ and GPT-4o’s assumptions about gender in a fill-in-the-blank task. Experiment 3 tests how role nouns and adjectives influence images created by the image generator DALL-E 3 (a text-to-image model). Our results show that humans, LLMs and text-to-image models’ outputs are all influenced by gender stereotypes but diverge in unexpected ways
The semantics of multi-headed wh correlatives in Georgian
Full text linkExisting accounts of Georgian relative clauses mention two relativization strategies, one involving the complementizer rom, and the other involving the relativizing clitic -c. While Foley (2013); Bhatt & Nash (2023) provide a syntactic account of these structures, their semantics remains unstudied. In this paper I present novel data on preposed relative clauses in Georgian formed via the relativizing clitic -c, and argue for a correlative-like semantics of these structures along the lines Dayal 1996’s analysis of Hindi correlatives. Moreover, contra Bhatt & Nash (2023) I argue it is these multiple wh-romelic relative clauses rather than the multiple headed rom relative clauses that pattern like canonical correlatives in terms of definiteness effects, and their ability to license a pair-list interpretation