Harder Words: Learning Abstract Verbs with Opaque Syntax

Abstract

Highly abstract predicates (e.g. think) present a number of difficulties for language learners (Gleitman et al., 2005). A partial solution to learning these verbs is that learners exploit regularities in the syntactic frames in which these verbs occur. While agreeing with this general approach to learning verbs, we caution that this strategy is not sufficient for learning another class of abstract verbs known as "raising" verbs (seem) since their argument structure frames cannot always be directly read off of the surface syntax. We conducted a novel verb learning study to test the role played by subject (in)animacy in determining the syntactic frame of novel verbs that could be categorized as raising verbs. Animacy is useful since raising but not control verbs admit inanimate subjects (The rock seems/#claims to be heavy). We outline a two-step process of probabilistic verb categorization that relies on subject animacy in addition to more traditional Syntactic Bootstrapping assumptions. We conclude by suggesting more general applications of this process to learning other kinds of opaque abstract predicates. HARD WORDS AND HARDER WORDSThe well-known Syntactic Bootstrapping hypothesis was born in an attempt to explain how blind children are able to acquire verbs like look and see. As Landau and Gleitman (1985) showed, blind children are able to understand these verbs and to use them appropriately even though the action or state denoted by the verb is beyond the experiential realm of these children. They have never directly experienced "looking" or "seeing," yet they form accurate lexical entries for these verbs and make remarkably few errors in their use. As Gleitman (1990) explains, while the extralinguistic contextual support for verb meanings is irregular and unreliable, for both sighted and nonseeing learners, the linguistic context, as indexed by the argument structure frames that verbs occur in, is highly regular and systematic.Moreover, the set of argument structure frames of verbs overlaps to a large degree with verb meanings. That is, verbs that have the same argument structure frames often have quite similar Correspondence should be addressed to

Comparing solutions to the linking problem using an integrated quantitative framework of language acquisition

“…Second, animacy for each NP argument was annotated by hand. We included animacy because a number of acquisition studies have demonstrated that animacy is a useful cue for learning verb classes (Scott & Fisher, 2009;Becker, 2009;Kirby, 2009aKirby, , 2010Becker & Estigarribia, 2013;Becker, 2014Becker, , 2015Hartshorne et al, 2015). Third, thematic roles for the arguments of each verb (except the copula be) were annotated by hand using 13 thematic role labels that are common in the literature (again, see the readme file mentioned above for details).…”

“…Animacy is something young children are known to both be sensitive to as a general property and also use as a cue in experimental studies to predict how verbs will behave (Scott & Fisher, 2009;Becker, 2009;Kirby, 2009aKirby, , 2010Becker, 2014Becker, , 2015Hartshorne et al, 2015). Moreover, if children are able to harness animacy effectively in their input, it's possible to use the animacy of a verb's arguments (in particular, whether the argument is inanimate) to distinguish verb behaviors such as those associated with subject-raising, subject-control, objectraising, and object-control (Kirby, 2009a(Kirby, , 2010Becker & Estigarribia, 2013;Becker, 2014). One way to implement this conceptual information is for the verb's NP arguments to be labeled as +/-animate, as in (6b).…”

How do children learn novel emotion words? A study of emotion concept acquisition in preschoolers.

“…Children are then asked to point to, or their eye gaze is tracked toward a visual image depicting the relevant action (Fisher, 2002; Hirsh-Pasek & Golinkoff, 1996; Naigles, 1990; Yuan & Fisher, 2009). The methodology has been adapted to test learners’ interpretations of not only verbs denoting observable actions but also relatively more abstract kinds of verbs, such as those that could have a meaning like think or seem , or to abstract adjectives such as easy (Becker, 2006, 2014, 2015; Becker & Estigarribia, 2013; Papafragou, Cassidy, & Gleitman, 2007). In some cases, such as for abstract verbs such as think , the sentence frame is even more informative than extragrammatical situational context (what is going on in the world when a particular verb is uttered) for cueing children to the predicate’s meaning (Gleitman et al, 2005; Papafragou et al, 2007).…”

Innate Mechanisms for Acquiring Syntactic Displacement

“…Evidence for the bias to map animate or agentive NPs to the most prominent syntactic position comes not only from psycholinguistic studies with children and adults (Clark & Begun 1971, Trueswell & Tanenhaus 1994, Mak et al 2002, Traxler et al 2005, Becker 2005, Becker & Estigarribia 2013, but also from argument hierarchies in language that are based on broad typological patterns. Well-known hierarchies such as the Animacy Hierarchy (Silverstein 1976, Dixon 1979) and the Thematic Hierarchy (Jackendoff 1972), as well as the Promotion to Subject Hierarchy (Keenan 1976) show how widespread the preference is in human language to associate animacy and agency with syntactic prominence, and inanimacy with lower positions.…”

“…In my previous work I have argued that hearing an inanimate subject in sentences like (1) and ( 2) can provide such a trigger by signaling that the subject and main predicate are not thematically related (i.e. not in an argument structure relationship), and therefore that the subject is displaced from elsewhere in the sentence (Becker 2005, 2006, 2014, 2015, Mitchener & Becker 2011, Becker & Estigarribia 2013. Inanimate subjects provide a helpful clue because inanimate NPs cannot be agents or experiencers and therefore can't be the subject argument of a control predicate (the ones found in (3) and ( 4)).…”