Chaining and the growth of linguistic categories

Abstract

We explore how linguistic categories extend over time as novel items are assigned to existing categories. As a case study we consider how Chinese numeral classifiers were extended to emerging nouns over the past half century. Numeral classifiers are common in East and Southeast Asian languages, and are prominent in the cognitive linguistics literature as examples of radial categories. Each member of a radial category is linked to a central prototype, and this view of categorization therefore contrasts with exemplar-based accounts that deny the existence of category prototypes. We explore these competing views by evaluating computational models of category growth that draw on existing psychological models of categorization. We find that an exemplar-based approach closely related to the Generalized Context Model provides the best account of our data. Our work suggests that numeral classifiers and other categories previously described as radial categories may be better understood as exemplar-based categories, and thereby strengthens the connection between cognitive linguistics and psychological models of categorization.

Beyond the Benchmarks: Toward Human-Like Lexical Representations

“…For example, novel usages such as “I am a man (...usually all flannelled up )” tend to have high similarity to existing usages ( gear up, glove up, mask up, sweater up , ...) that form a broad and frequent class. Moreover, Watson et al ( 2021 ) find that novel usages cluster around other novel usages ( flannel is similar to sweater ), confirming that the exemplar-driven innovation found in historical analyses (e.g., Habibi et al, 2020 ; Yu et al, 2020 ) plays a role in dynamic adaptation of language. This is an important point, because one-off usages, rather than being statistical noise, serve as informative signals to people of legitimate creativity.…”

Chaining and the temporal dynamics of scientists’ publishing behaviour

“…We consider five models of chaining following the existing literature [ 22 – 24 ]. Each model postulates a plausible yet different cognitive mechanism for how future scientific outputs might relate to or stem from existing ones and constructs a continuous temporal path among the scientific publications of a scientist.…”

“…Our proposal is grounded in the cognitive theory of chaining [ 19 – 22 ], and we use the existing computational cognitive algorithms that have been shown to capture properties of chaining in the context of word meaning extension [ 22 – 25 ]. More specifically, chaining refers to a continuous and incremental process of meaning growth whereby novel ideas connect to existing ideas that are semantically similar, hence forming chain-like structures over time.…”

“…More specifically, chaining refers to a continuous and incremental process of meaning growth whereby novel ideas connect to existing ideas that are semantically similar, hence forming chain-like structures over time. Existing work has formulated chaining as probabilistic algorithms or models of categorization and applied these formal models to predict the historical growth of word meaning in a variety of domains including container names, numeral classifiers, adjectives, and verbs [ 22 – 27 ].…”

“…We believe that the general process of scientific progress might mirror the process of word meaning extension, whereby a scientist conceives of a new idea by linking to existing ideas that are closely related. Importantly, previous work has also suggested that certain forms of chaining are preferred or more cognitively natural than other forms, and recent findings suggest that chaining can be best understood as an exemplar-based process [ 22 , 24 , 25 ]. The exemplar model of chaining is rooted in the exemplar theory of categorization, also known as the General Context Model [ 28 ].…”