Form-meaning relations in acquisition: the case of polysemous constructions

This work was presented at the Societas Linguistica Europaea 52nd Annual Meeting, Leipzig, Germany.



Many constructions are associated with a variety of pragmatic meanings, posing a challenge in acquisition. How do children learn the meanings of these constructions, and what factors influence their production and comprehension of the various forms? Here we consider two construction types; modals, and causal/conditional adverbial sentences. Individual modals (e.g. can, must) can convey a variety of deontic (e.g. ability – He can reach the book, permission – He can go to the park now) and epistemic meanings (e.g. He can read French) (Papafragou, 1998), while causal and conditional adverbial sentences explain real-world causality (Content, e.g. You’re hungry because you didn’t eat); justify/restrict conclusions (Epistemic, e.g. She must not have eaten, if there is food left); or justify/restrict speech acts (Speech-Act, e.g. Eat your dinner, because you’re hungry) (Sweetser, 1990). We ask to what extent the frequency distribution of form-meaning mappings in the input children hear predicts (i) the form-meaning mappings they produce and (ii) their comprehension of these forms with different meanings. Specifically, do children start out by producing and comprehending those mappings which they hear most frequently?
To address (i), we conducted two corpus studies. Utterances containing modals or the adverbials because/if were extracted from dense naturalistic corpora of mothers interacting with their 3-5yr-old English-speaking children. These utterances were coded for their form and pragmatic meaning, and the distribution of form-meaning mappings in the input and in the children’s speech was compared. We found that for all speakers, different modal/causal/conditional forms were associated with a different distribution of form-meaning mappings (e.g. because was most frequently produced in Speech-Acts, while if most frequently encoded Content). Some modals (e.g. may) appeared with only one or two meanings, whereas others (e.g. can) took many different meanings. In general, the children’s patterns of usage mirrored those in their input (e.g. frequency of modal form-meaning pairings in the input predicted use in the children’s speech); children did not start out producing only the most frequently encountered form-meaning mapping for each form. On the other hand, there were also some differences (e.g. in the use of because and acquisition of epistemic modal meanings) which suggests that immediate discourse priorities and the children’s cognitive capabilities interact with distributional information to determine the patterns of acquisition.

To address (ii), we conducted two comprehension experiments to explore whether 3-5yr-old children’s interpretation of modal can/must and adverbial because/if constructions varies as a function of their frequency of exposure to different form-meaning mappings in the input. Both studies utilize a forced-choice picture selection task, gathering picture choice and reaction time data. Bayesian analysis suggests that for conditional constructions, in line with the corpus data where we observed frequency asymmetries in form-meaning associations, children’s performance was better for higher frequency mappings, although this pattern was not observed for causal constructions. For modal constructions, data analysis is underway.

Both sets of results will be discussed in the context of the role of input frequency in determining the comprehension and production of polysemous constructions across acquisition.


Papafragou, A. (1998). The Acquisition of Modality: Implications for Theories of Semantic Representation. Mind & Language, 13(3), 370-399.
Sweetser, E. (1990). From Etymology to Pragmatics: Metaphorical and Cultural Aspects of Semantic Structure (Cambridge Studies in Linguistics). Cambridge: Cambridge University Press.


Citation Information

Theakston, A., Bell, K., Lemen, H., Brandt, S., & Lieven, E. (2019). Form-meaning relations in acquisition: the case of polysemous constructions. Paper presented at the Societas Linguistica Europaea 52nd Annual Meeting, Leipzig, Germany.