Acquisition and island constraints My main research focus is the role that syntactic constraints (e.g., island constraints) play in the acquisition of grammar. Are such constraints intrinsic to the minds of children, or do children actually infer these constraints from their representations of the speech that they hear? My own empirical work (with Jeffrey Lidz) suggests that 3 year-old children have acquired knowledge of locality constraints on wh-movement — which isn’t long after they’ve acquired the syntax of wh-movement. This empirical work places tangible limitations on learning models of island constraints. [Dissertation Chapter]

Acquisition of modal words Around 2-3 years-old, children begin to use modal words in their own speech (e.g., can, must, might, have to), but it’s not clear that they have acquired the complex meanings of these words. In American English, modals express different “flavors” of possibility or necessity. For example, can, might, and could express possibility, but not necessity, while must, have to, and should express necessity, but not possibility. The modal flavor is determined by the context. The sentence Anouk might paint the walls could express a means to achieve her new home makeover (goal oriented flavor), or express what Anouk has planned for the room, as far as the speaker knows (knowledge-based flavor). In empirical work with Ailis Cournane and Valentine Hacquard, we prompted children to use modal words in carefully controlled contexts to study their early meaning postulates. We find that children use different modal words to express different “flavors” of modality (like adults), but use the same modal words to express possibility and necessity (unlike adults).

Acquisition and sentence processing Infants and children speak slowly, but they acquire a grammar from listening to adult speakers and older siblings, who use 4-6 words a second! From a young age, infants use incremental language processing strategies to quickly form linguistic representations of the speech they hear. They make predictions about what speakers will say, before they say it, so that they can keep up and learn a grammar. In my work with Laurel Perkins and Jeffrey Lidz, I study how (partial) knowledge of grammar feeds language processing strategies in 19 month-old infants. Specifically, we find that 19 month-olds process wh-questions incrementally, and use their knowledge of verb transitivity to interpret fronted wh-phrases — all before the sentence ends.

Acquisition and argument structure Infants manage to learn verbs when they’re only 1 years-old. Research suggests that they learn verb meanings by using their knowledge of syntax to zero in on what people are talking about (syntactic bootstrapping). But how exactly do infants link linguistic relations and event concepts to acquire verb meanings? In on-going work with Tyler Knowlton, Jeffrey Lidz and Alexander Williams, we show infants events that we already know they conceptualize as having either 2 or 3 participants. Each event is described by a made-up verb with only 2 syntactic arguments (e.g., She florped the cup!). If infants use linguistic relations like agent and patient (as opposed to matching arguments and participants one-to-one), then they will encode a 3 participant event concept for florp, even though the sentence only had 2 arguments. This more abstract strategy would grant them the flexibility to learn actual verbs like rob and jimmy, which can express 3 participant events in transitive frames, like She robbed the vault (of its precious jewels) and She jimmied the lock (with the wire hanger).