Browse By Day
Browse By Time
Browse By Panel
Browse By Session Type
Browse By Topic Area
Corpus (Wells, 1979) and comprehension (Hirst & Weil, 1982; Ozturk & Papafragou, 2014) studies suggest that children have difficulty learning modal words (e.g., 'can', 'must'). This is standardly attributed to conceptual or pragmatic difficulties (Papafragou, 1998; Noveck, 2001). We use a sentence-repair task to show that children’s difficulty arises instead from the many-to-one mapping between modal meanings and forms.
The same modal words can be used to express possibility and necessity "forces" in different "flavors", creating a one-to-many mapping problem from word-form to meaning. For example, in (1), 'must' is used to express a goal-oriented (teleological) necessity "flavor"; in (2), it is used to express a knowledge-based (epistemic) necessity "flavor".
(1) Kat must go down the pink path (given her goal to get to the bakery)
(2) Nic must be hiding in the pink box (given knowledge that the other box is empty)
Furthermore, different modals can express the same flavor (e.g., 'must'/'have to'), adding to the difficult mapping problem children must eventually solve: children have to (a) map the same modal to different meanings (b) map the same meaning to different modals, and (c) identify which abstract concepts pair with which forms.
Corpus studies show that children use non-epistemic modals before epistemic ones (e.g., Wells, 1979). Comprehension studies show that children are likely to accept possibility modals in necessity contexts, and necessity modals in possibility contexts (Noveck, 2001; Ozturk & Papafragou, 2014). From these results, it's not clear what meanings children map to modal words. Using a sentence-repair task, building on Cournane, (2014), we elicit modal words in teleological and epistemic necessity and possibility contexts (Fig.1). Children are trained to repeat pre-recorded sentences to a shy snail puppet, filling in any missing words for his benefit. In test sentences, white noise occurs where an adult would use a modal (1-2), which allows children to supply their own modal to fit the controlled context.
We crossed three factors: force (necessity vs. possibility), "flavor" (teleological vs. epistemic) and age (child: n=15, ages=3;1-4;5; mean=3;11; adult: n=11, ages=18-28). Here, we report only the target modal responses, which make up about 49% and 97% of the total child and adult responses, respectively. We find that children use different modal words than adults do: in epistemic contexts, children prefer 'might' for possibilities and necessities, while adults prefer 'could' for possibilities and 'must' for necessities (Fig.2). In teleological contexts, children prefer 'can', while adults prefer 'could' for possibilities, and both 'could' and 'has to' for necessities. In the epistemic condition, children use necessity and possibility modals the same amount, regardless of the force condition, but in the teleological condition, children differentiate between possibility and necessity force. Together our results suggest that the conceptual distinctions may already be in place, but children lack robust linguistic representations for the modal words adults prefer to use in different contexts. Hence, children's difficulty in comprehension studies may be due to their lacking the adult-like form-meaning relations for the modals being tested.