2018 Alternate 6371
2018 Alternates | Saturday, November 3, 2018 | Poster Session II, Metcalf Small | 3:15pm
Early word meanings are structured around similarity: evidence from lexical processing
S. Floyd, C. Lew-Williams, A. Goldberg
While a large proportion of frequent words are ambiguous in that they are associated with more than a single meaning [1], developmental theory has assumed a one-to-one mapping between form and meaning, making no predictions for the structure of multiple meanings [2].
Additionally, ambiguous words differ in how related their meanings are: homonyms like bat have unrelated meanings (“baseball”, “mammal”), while most words with multiple meanings are polysemous, sharing featural or functional similarity (e.g., baseball cap and pen cap) [3,4]. In fact, 40% of commonly used words in English are polysemous [4]. Yet, no study has examined how children process words with multiple meanings. Relations between meanings could lead to greater co-activation, yielding faster processing when identifying additional meanings in the case of polysemy compared with homonymy. Alternatively, children could start by representing polysemous meanings independently, yielding equal cross-meaning priming for polysemy and ambiguity.
In a touchscreen picture identification task with 4-7-year-old children (n=48), we compared cross-meaning priming of homonyms and polysemes. On each trial, participants heard a familiar word and selected the target meaning from a distractor, followed by a second trial in which the target was another meaning of the same word – either polysemous or ambiguous (Figure 1). Additionally, in a baseline condition, children were tested on sequential depictions of the same meaning (e.g., treehouse). We used “irregular” polysemes (e.g. step of stairs, step-counter), so that children could not recruit a productive extension rule to speed up activation of the additional sense, instead relying on word knowledge [5, 6]. Frequency was matched within pairs, and word length was matched across conditions. 48 adults were tested in order to validate the paradigm and assess developmental change. We hypothesized that children’s representations of polysemous words are sensitive to relatedness across meanings, and therefore predicted faster reaction times to the second-encountered meaning of a polysemous word compared to an ambiguous word.
With log-transformed data, we fit a multilevel model with Condition (baseline vs. polysemy vs. ambiguity) and Meaning number (1st– or 2nd-meaning encountered) and their interaction as fixed effects (maximal random structure). As predicted, children were primed in the case of polysemes but not homophones, i.e., we found a significant interaction between Condition (polysemy) and Meaning number (2nd): ß=-0.12010, p=0.00591, without a main effect of word type or sense number. Replicating previous findings [7, cf.8], adults were primed across meanings in all trials: main effect of Meaning Number (2nd) (ß=-0.15515, p=0.00129), no interaction with Condition (ß=0.04156, p=0.32453).
Despite the ubiquity of polysemy in language, our work is the first to show that children’s processing reveals organization of such representations. We conclude that children rely on meaning relationships when activating ambiguous words, suggesting that activation of multiple unrelated meanings of a word develops later. Further directions will explore the development of homonymous priming as a function of literacy, and explore whether adults also exhibit a polysemy-specific priming boost when additional meanings are infrequent as evidence for a common mechanism across development.
References
[1] Zipf, GK (1949) Human Behavior and the Principle of Least Effort: An Introduction to Human Ecology (Addison-Wesley, Cambridge, MA). [2] Trueswell, JC, Medina, TN, Hafri, A., & Gleitman, LR (2013). Propose but verify: Fast mapping meets cross-situational word learning. Cognitive Psychology, 66(1), 126–156. [3] McCarthy, M. (Ed.) (1997). Vocabulary: Description, Acquisition and Pedagogy. Cambridge University Press. [4] Durkin, K & Manning, J (1989). Polysemy and the Subjective Lexicon: Semantic Relatedness and the Salience of Intraword Meanings. Journal of Psycholinguistic Research, 18, 577-612. [5] Brocher, A, Koenig, JP., Mauner, G & Foraker, S. (2017) About sharing and commitment: the retrieval of biased and balanced irregular polysemes. Language, Cognition and Neuroscience, 1-24. [6] Wittgenstein, 18. (1953). Philosophical investigations. John Wiley & Sons. [7] Swinney, DA. (1979) Lexical access during sentence comprehension: (re)consideration of context effects. Journal of Verbal Learning and Verbal Behavior 18.6:645–660. [8] Sereno, SC, Brewer, CC, & O’Donnell, PJ. (2003) Context effects in word recognition: Evidence for early interactive processing. Psychological Science, 14(4), 328-333.