Learning the Language of Locomotion: Do Children Use Biomechanical Structure to Constrain Hypotheses about Word Meaning?

ABSTRACT

Much has been said about children’s strategies for mapping elements of meaning to words in toddlerhood. However, children continue to refine word meanings and patterns of word use into middle childhood and beyond, even for common words appearing in early vocabulary. We address where children past toddlerhood diverge from adults and where they more closely approximate them, and why. In two studies, we examined naming of locomotion (walking, running, hopping, etc.) by children aged four to nine and compared their patterns of word use to adult patterns. We evaluated whether the children are sensitive to the biomechanical discontinuity between pendulum-type and impact-and-recoil-type actions that constrains adult word use. We also evaluated whether they appreciate this constraint by age four or only develop appreciation later. Children from four onward were sensitive to the biomechanical distinction in their word use. Perceived domain structure plays a role in explaining later lexical development.

Acknowledgments

Some of the data reported here were previously reported in Malt, B.C., Gennari, S., Imai, M., Ameel, E., Tsuda, N., & Majid, A. (2008). Talking about walking: Biomechanics and the language of locomotion. Psychological Science, 19, 232–240, and Malt, B.C., Gennari, S.P., Imai, M., Ameel, E., Saji, N., & Majid, A. (2014). Human locomotion in languages: Constraints on moving and meaning. Journal of Memory and Language, 74, 107–123.

We thank Dorien Cools for collecting the children’s data, their schools for permitting access, and Padraig O’Seaghdha for helpful discussion. Kristine Schuster developed the web page for collecting naming data, and Naoaki Tsuda developed the program for collecting physical similarity data.

Funding

This research was supported by Grant OT/10/024 from the Leuven University Research Council to Gert Storms. Anne White is a research assistant of the Research Foundation-Flanders (FWO-Vlaanderen).

Notes

¹ We use “overextension” and “underextension” in the classic sense based on production (e.g., Rescorla, 1980). Whereas Ameel et al. (2008) found overextension of some container and dishware terms in a production task and underextension of others for children ages 5–14, Ameel et al. (2014) found only overextension of terms in two receptive tasks with the same stimuli for children in a similar age range. Given the substantially younger children tested in the current study, it is an open question exactly how they would respond to the adult dominant locomotion terms in a receptive task. It is unlikely in any case that they would display adult-like performance.

² One small adjustment was made relative to Malt et al. (2008): Wandelt snel was counted as an instance of wandelen (plus modifier), not as the conventional compound snelwandelen, based on information from native speakers that the simple present of snelwandelen would be snelwandelt (not wandelt snel). This change was relevant to only a few responses and had little impact on final tallies.

³ We do not report exact word frequencies because their interpretation is clouded by the substantial dialect differences in use of locomotion words between Netherlands and Belgian Dutch (Majid & Malt, 2012) and by lack of information as to which dialect(s) are represented in the frequency norms. In addition, lopen is highly polysemous (similar to English run: run for office, run a business, water runs down the window, etc.), making it impossible to know how frequently it is used with reference to locomotion.

⁴ Again, both tallies had the small modification noted in Footnote 2 applied.

⁵ Again, we do not specify word frequencies due to the complications related to dialect variations and polysemy.

⁶ In Malt et al. (2014), the walk_backwards clip is shown in the third-level cluster with the clips labeled noh, pigeon toed, sneak, and tiptoe. This appears to have been an error in interpreting the program output. No study conclusions are affected by the difference.

Additional information

Funding

This research was supported by Grant OT/10/024 from the Leuven University Research Council to Gert Storms. Anne White is a research assistant of the Research Foundation-Flanders (FWO-Vlaanderen).