Avoiding dative overgeneralisation errors: semantics, statistics or both?

Abstract

How do children eventually come to avoid the production of overgeneralisation errors, in particular, those involving the dative (e.g., *I said her ''no'')? The present study addressed this question by obtaining from adults and children (5Á6, 9Á10 years) judgements of well-formed and over-general datives with 301 different verbs (44 for children). A significant effect of pre-emption*whereby the use of a verb in the prepositional-object (PO)-dative construction constitutes evidence that double-object (DO)-dative uses are not permitted*was observed for every age group. A significant effect of entrenchment*whereby the use of a verb in any construction constitutes evidence that unattested dative uses are not permitted*was also observed for every age group, with both predictors also accounting for developmental change between ages 5Á6 and 9Á10 years. Adults demonstrated knowledge of a morphophonological constraint that prohibits Latinate verbs from appearing in the DO-dative construction (e.g., *I suggested her the trip). Verbs' semantic properties (supplied by independent adult raters) explained additional variance for all groups and developmentally, with the relative influence of narrow-vs broad-range semantic properties increasing with age. We conclude by outlining an account of the formation and restriction of argument-structure generalisations designed to accommodate these findings.

Generalizing beyond the input: The functions of the constructions matter

“…This is a question that has bedeviled researchers of argument structure for quite a long time (e.g., Ambridge, Pine, Rowland, Freudenthal, & Chang, 2014;Baker, 1979;Bowerman, 1988;Braine, 1971;Goldberg, 1995;Lakoff, 1970;Pinker, 1989). Recent work on artificial grammar learning has demonstrated that the statistics in the input can play a key role.…”

A Constructivist Account of Child Language Acquisition

“…This weighting explains the finding that a slot that is highly skewed towards one particular filler largely takes on the properties of that filler (Casenhiser and Goldberg 2005;Goldberg, Casenhiser, and Sethuraman, 2004;Goldberg, Casenhiser, and White, 2007). For example, the properties of the "VERB" slot in the double-object dative construction (not just semantic, but also morphophonological) are largely those of give, which accounts for the lion's share of all occurrences of this construction (Ambridge, Pine, Rowland, Freudenthal, and Chang, 2014).…”

“…Consequently, the likelihood of children's using (e.g., Brooks, Tomasello, Dodson, and Lewis, 1999), accepting (Ambridge, Pine, Rowland, and Young, 2008;Ambridge, Pine, Rowland, Jones, and Clark, 2009;, and comprehending (Dittmar, Abbot-Smith, Lieven, and Tomasello, 2013) verbs in non-attested constructions decreases with increasing verb frequency. 10 With regard to semantics, children learn the fine-grained semantic properties of particular construction slots (e.g., that the [Z] slot in the [X] [Y][Z] construction denotes direct causation), meaning that verbs that are not compatible with this meaning (e.g., giggle) are a better fit for slots in other, competing constructions such as the intransitive and periphrastic causative (Ambridge, Pine, Rowland, Jones, and Clark, 2009;Ambridge, Pine, Rowland, and Chang, 2012;Ambridge, Pine, Rowland, Freudenthal, and Chang, 2014;Ambridge, Pine, and Rowland, 2011).…”

“…I suggested the idea to him) (Pinker, 1989). Ambridge, Pine, Rowland, Freudenthal, and Chang (2014) and Ambridge, Pine, Rowland, and Chang (2012) showed that independently obtained ratings of verbs' semantic and phonological properties were able to predict their relative acceptability in PO-and DO-dative constructions.…”

Statistics and semantics in the acquisition of Spanish word order: Testing two accounts of the retreat from locative overgeneralization errors

“…With regard to the semantic verb class hypothesis, both production and judgment studies have shown that if adults and children are taught novel verbs, they use their notional semantic class membership to determine the constructions in which they can and cannot appear (Gropen et al 1989;Gropen et al 1991a;Gropen et al 1991b;Ambridge et al 2008, Ambridge et al 2009Ambridge et al 2011;Ambridge et al 2012). With regard to the statistical learning hypothesis, many studies have observed the predicted negative correlation between verb frequency and the rated acceptability/ production probability of overgeneralization errors, in judgment and production tasks respectively Brooks and Zizak 2002;Theakston 2004;Ambridge et al 2008;Stefanowitsch 2008;Wonnacott et al 2008;Ambridge et al 2009;Goldberg 2011;Boyd and Goldberg 2011;Wonnacott 2011;Ambridge et al 2012a;Ambridge et al, 2012b;Ambridge and Brandt 2013;Ambridge 2013;Ambridge et al 2014).…”