Lexical ambiguity—the phenomenon of a single word having multiple, distinguishable senses—is pervasive in language. Both the degree of ambiguity of a word (roughly, its number of senses) and the relatedness of those senses have been found to have widespread effects on language acquisition and processing. Recently, distributional approaches to semantics, in which a word's meaning is determined by its contexts, have led to successful research quantifying the degree of ambiguity, but these measures have not distinguished between the ambiguity of words (...) with multiple related senses versus multiple unrelated meanings. In this work, we present the first assessment of whether distributional meaning representations can capture the ambiguity structure of a word, including both the number and relatedness of senses. On a very large sample of English words, we find that some, but not all, distributional semantic representations that we test exhibit detectable differences between sets of monosemes (unambiguous words; N = 964), polysemes (with multiple related senses; N = 4,096), and homonyms (with multiple unrelated senses; N = 355). Our findings begin to answer open questions from earlier work regarding whether distributional semantic representations of words, which successfully capture various semantic relationships, also reflect fine‐grained aspects of meaning structure that influence human behavior. Our findings emphasize the importance of measuring whether proposed lexical representations capture such distinctions: In addition to standard benchmarks that test the similarity structure of distributional semantic models, we need to also consider whether they have cognitively plausible ambiguity structure. (shrink)

Biases and misunderstanding stemming from pre-training in Generative Pre-Trained Transformers are more likely for users of underrepresented English varieties, since the training dataset favors dominant Englishes (e.g., American English). We investigate (potential) bias in GPT-4 when it interacts with Trinidadian English Creole (TEC), a non-hegemonic English variety that partially overlaps with standardized English (SE) but still contains distinctive characteristics. (1) Comparable responses: we asked GPT-4 18 questions in TEC and SE and compared the content and detail of the responses. (2) (...) Accurate translation: we assessed how accurate and authentic 29 TEC and 34 SE translations were. (3) Language knowledge and attitudes: we asked what language the prompts were written in and categorized the responses and examined any language attitudes that were exhibited. Content and detail in prompts were comparable. The model was proficient at translating TEC pronouns and many grammatical categories. It was weaker at processing spelling and vocabulary items. In addition, it produced several inauthentic features. Only 39% of TEC-generated sentences were fully grammatical. While GPT-4 was perfect at identifying SE, it was 21% accurate at identifying TEC, which it sometimes classified as English with “errors” and “corrected”. GPT-4’s scope of use is limited for non-hegemonic English users. It is problematic that some of its analyses perpetuate bias against underrepresented Englishes. Increased research on lesser-documented Englishes is necessary and we anticipate that this problem affects dialects of other languages. We intend to partner with Trinidadian stakeholders to train GPT-4 in the future. (shrink)

We explore the following two cognitive questions regarding crosslinguistic variation in lexical semantic systems: Why are some linguistic categories—that is, the associations between a term and a portion of the semantic space—harder to learn than others? How does learning a language‐specific set of lexical categories affect processing in that semantic domain? Using a computational word‐learner, and the domain of color as a testbed, we investigate these questions by modeling both child acquisition of color terms and adult behavior on a non‐verbal (...) color discrimination task. A further goal is to test an approach to lexical semantic representation based on the principle that the more languages label any two situations with the same word, the more conceptually similar those two situations are. We compare such a crosslinguistically based semantic space to one based on perceptual similarity. Our computational model suggests a mechanistic explanation for the interplay between term frequency and the semantic closeness of learned categories in developmental error patterns for color terms. Our model also indicates how linguistic relativity effects could arise from an acquisition mechanism that yields language‐specific topologies for the same semantic domain. Moreover, we find that the crosslinguistically inspired semantic space supports these results at least as well as—and in some aspects better than—the purely perceptual one, thus confirming our approach as a practical and principled method for lexical semantic representation in cognitive modeling. (shrink)

In this paper, we examine a specific type of Relative Clause. We look at the construction consisting of a ‘light noun’, that is, a noun with highly non-specific lexical content which does not do referential work, plus a relative clause. It has generally been assumed that the functional contribution of RCs is to narrow the set of referents of the head noun to only those for which the predicate of the RC holds true. However, the ‘Light Head RC construction’ has (...) various interactional affordances that mostly revolve around characterizing referents with discourse-relevant properties rather than establishing reference. We argue that these various functions of LHRCs revolve around participants’ orientations to categoryhood. Data are in English and Dutch. (shrink)