Bootstrapping into Filler-Gap: An Acquisition Story

Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, pages 1084–1093, Baltimore, Maryland, USA, June 23-25 2014. c⃝2014 Association for Computational Linguistics Bootstrapping into Filler-Gap: An Acquisition Story Marten van Schijndel Micha Elsner The Ohio State University {vanschm,melsner}@ling.ohio-state.edu Abstract Analyses of filler-gap dependencies usu- ally involve complex syntactic rules or heuristics; however recent results suggest that filler-gap comprehension begins ear- lier than seemingly simpler constructions such as ditransitives or passives. Therefore, this work models filler-gap acquisition as a byproduct of learning word orderings (e.g. SVO vs OSV), which must be done at a very young age anyway in order to extract meaning from language. Specifically, this model, trained on part-of-speech tags, rep- resents the preferred locations of semantic roles relative to a verb as Gaussian mix- tures over real numbers. This approach learns role assignment in filler-gap constructions in a manner con- sistent with current developmental findings and is extremely robust to initialization variance. Additionally, this model is shown to be able to account for a characteristic er- ror made by learners during this period (A and B gorped interpreted as A gorped B). 1 Introduction The phenomenon of filler-gap, where the argument of a predicate appears outside its canonical posi- tion in the phrase structure (e.g. [the apple]i that the boy ate ti or [what]i did the boy eat ti), has long been an object of study for syntacticians (Ross, 1967) due to its apparent processing complexity. Such complexity is due, in part, to the arbitrary length of the dependency between a filler and its gap (e.g. [the apple]i that Mary said the boy ate ti). Recent studies indicate that comprehension of filler-gap constructions begins around 15 months (Seidl et al., 2003; Gagliardi et al., 2014). This finding raises the question of how such a complex phenomenon could be acquired so early since chil- dren at that age do not yet have a very advanced grasp of language (e.g. ditransitives do not seem to be generalized until at least 31 months; Gold- berg et al. 2004, Bello 2012). This work shows that filler-gap comprehension in English may be Age Wh-S Wh-O 1-1 13mo No No 15mo Yes (Yes) 20mo Yes Yes Yes 25mo Yes Yes No Figure 1: The developmental timeline of subject (Wh-S) and object (Wh-O) wh-clause extraction comprehension suggested by experimental results (Seidl et al., 2003; Gagliardi et al., 2014). Paren- theses indicate weak comprehension. The final row shows the timeline of 1-1 role bias errors (Naigles, 1990; Gertner and Fisher, 2012). Missing nodes de- note a lack of studies. acquired through learning word orderings rather than relying on hierarchical syntactic knowledge. This work describes a cognitive model of the de- velopmental timecourse of filler-gap comprehension with the goal of setting a lower bound on the mod- eling assumptions necessary for an ideal learner to display filler-gap comprehension. In particular, the model described in this paper takes chunked child-directed speech as input and learns orderings over semantic roles. These orderings then permit the model …

Content truncated to 500 words