Linking Parser Development to Acquisition of Syntactic Knowledge

Abstract

Traditionally, acquisition of syntactic knowledge and the development of sentence comprehension behaviors have been treated as separate disciplines. This article reviews a growing body of work on the development of incremental sentence comprehension mechanisms and discusses how a better understanding of the developing parser can shed light on two linking problems that plague language acquisition research. The first linking problem is that children’s behavioral data that are observable to researchers do not provide a transparent window into the developing grammar, as children’s immature linguistic behaviors may reflect the immature parser. The second linking problem is that the input data that researchers investigate may not correspond veridically to the intake data that feed the language acquisition mechanisms, as the developing parser may misanalyze and incorrectly represent the input. Based on reviews of child language comprehension studies that shed light on these two linking problems, it is argued that further research is necessary to closely integrate parser development and acquisition of syntactic knowledge.

Notes

¹ Intake is defined here as linguistic representations in the mind that serve as data for the purpose of language development, but it is important to note that there are possibly two (or more) alternative definitions of intake. One commonly used definition is that intake is the data that are internalized through attention and perception (Corder 1967; Omaki 2010) and is available for further computation in the mind, including language acquisition. This type of intake can be dubbed perceptual intake to highlight the fact that the input signal may not be veridically represented in the learner’s mind. Another commonly used definition is that intake is data that the language acquisition mechanism selectively extracts out of the perceptual intake for the purpose of making inferences about the grammatical structure of the language (Carroll 1999; Gagliardi & Lidz 2014; Pearl & Lidz 2009; Viau & Lidz 2011). This type of intake can be dubbed acquisitional intake, as it is hypothesized to directly feed the language acquisition mechanism. It is possible that these two levels of intake representations are identical (Fodor 1998), but this question awaits further research. This article focuses on the role of perceptual mechanisms, and for this reason we refer to the perceptual intake when we use the word intake.

² The problem of inference-under-uncertainty should sound familiar to language acquisition researchers, as language acquisition presents the same problem: Oftentimes, the learner must use various sources of information to infer the target grammar, because input strings could be compatible with multiple grammars (Chomsky 1965; Fodor & Sakas 2004; Gibson & Wexler 1994; Perfors, Tenenbaum & Regier 2011; Yang 2002). For example, the SVO word order may be derived from English-like phrase structure rules, but it could also be derived from a verb-final word order with a Verb Second rule, as in German.

³ The Intermodal Preferential Looking Paradigm, which has been widely used to study language development since Hirsh-Pasek & Golinkoff (1996), is essentially the same methodology as visual world eye-tracking techniques. A major difference lies in the questions they ask and how the data are used: Studies in the preferential looking paradigm tradition have focused on children’s linguistic knowledge and examined the ultimate interpretation by averaging looking times across a longer period of time (several seconds). This is partly because these studies tend to focus on young infants whose behaviors are not stable enough to reveal effects on the scale of milliseconds. Studies in the visual world eye-tracking tradition tend to focus on the time course of language comprehension, so as to shed light on how the interpretation evolves as the linguistic stimuli unfold. These studies usually set up the linguistic and visual stimuli in such a way that fine time course analyses are feasible.

⁴ The gloss abbreviations are as follows: Acc = accusative case marker, Comp = complementizer, Dim = diminutive marker, Gen = genitive case marker, Q = question particle, Top = topic marker.

⁵ Note that there were 219 instances of long-distance wh-argument fronting (e.g., What do you think Robin is doing in school?), which provides clear evidence that long distance wh-movement from the embedded clause is possible. Here, we will continue to focus on the distribution of adjunct wh-questions for two reasons. First, the experiment in Omaki et al. (2014) only used adjunct wh-questions, and therefore the discussion of how the input distribution can be skewed due to the child parser must also be restricted to adjunct wh-questions. Second, there are syntactic reasons to think that generalizing properties of argument wh-questions to adjunct wh-questions is a risky move. For example, there are differences in constraints on argument and adjunct wh-movements (e.g., C.-T. J. Huang 1982), and it is also known that wh-scope marking for arguments and adjuncts can take a very different syntactic property (e.g., Bruening 2004). Third, the conclusion drawn from this case is easily extendable to other domains of syntactic development, and it thus serves as a useful exercise.