Past evidence suggests that adult Mandarin speakers rely on shape more heavily than English speakers when categorizing solid objects (Kuo and Sera 2009). In this experiment, we began to examine that effect developmentally by investigating the acquisition of the three most common Mandarin Chinese classifiers for solid objects (i.e. ge, zhi and tiao) in relation to development in shape-based categorization by native speakers of Mandarin and English from 3 years of age to adulthood. We found that 3-year-old Mandarin speakers were above chance in their classifier knowledge, but this knowledge continued to develop through 7 years of age. We also found that Mandarin speakers relied more heavily on shape than English speakers, and that shape-based categorization among English speakers tended to decline with age on the trials in which shape choices matched the Mandarin classifiers. The findings suggest that classifiers initially augment Mandarin speakers' attention to the shape of solid objects, and then maintain this early stronger shape bias after they are fully learned. The work highlights how categorization and word learning are graded and intertwined.

As speakers, we are frequently called upon to specify the locations of objects and landmarks in our environment. However, despite the considerable body of work on spatial cognition and semantics (i.e. Levinson 1992, 1996, 2003; Tyler and Evans 2003; Herskovits 1986; Vandeloise 1986), there has been almost no research on the expression of location from a multimodal viewpoint: that is, how do speakers use gesture, as well as speech, to express object location? This paper reports on a filmed study of 10 English speakers and 10 French speakers who were asked to express the locations of 28 objects in two spatial scenes. On the basis of our results we argue that a functional, as opposed to a grammatical, approach to the segmentation and analysis of these sorts of expressions is crucial. Such an approach reflects the fact that different Figures (i.e. objects to be located) can exist simultaneously across speech and gesture. Using filmed examples from our data set, we propose a new definition of static locative expressions and outline a multimodal approach for their analysis in oral discourse.

This paper examines the relationship between the concepts of ‘seeing’ and ‘attempting/trying’ in various languages. These concepts have so far been found to be co-lexified in languages spoken in Eurasia, Papua New Guinea, India and West Africa, with an added implicature of politeness present in some languages when this lexical item is used in directives. After establishing a cross-linguistic sample, the paper proposes a specific grammaticalization mechanism as responsible for producing this semantic relationship. The explanation centers on a process involving metaphorical transfer, the loss of semantic features, generalization, and a specific syntactic context conducive to this meaning shift. First, the Mind-as-Body metaphor is applied to the mind-related notion of ‘seeing an object’ to derive the body-related notion of ‘controlling an object’, as has previously been demonstrated to be the case in the history of certain Indo-European languages. Second, semantic bleaching causes the meaning component of physical sight to be lost from the overall meaning of the morpheme, and semantic generalization allows attempted actions to be mentally treated the same as physical objects that are manipulated. Finally, the context in which this meaning shift occurs is posited as constructions involving multiverbs, such as serial verbs or converbs.

Objects can almost always be called by more than one name, and the name chosen depends on the context. However, studies of naming in some traditions elicit names by showing objects in isolation and asking merely What is it? If the names produced are not the same as those that would be given in more fully specified communicative contexts, the value of the tasks for understanding object naming is in doubt. The current study examined the effects of different communicative demands on naming of 60 household containers. A standard “free naming” task was contrasted with two other versions. In these two, participants had the goal of getting an addressee to find a target object among others. In one, each object needed to be distinguished from a small set of dissimilar objects, visible to the addressee. In the other, each needed to be distinguished from a large set of similar objects, not visible to the addressee. Responses were sensitive to context in the number of modifiers produced. However, they were insensitive in the head nouns produced. Simple naming tasks such as What is it? can produce results for head nouns equivalent to those from tasks providing more fully specified communicative contexts.

The role of numeric language in basic numerical cognition is explored via the consideration of results obtained in two recent independent studies, one with Nicaraguan homesigners and one with speakers of Pirahã. Attention is drawn to remarkable parallels between the relevant findings, parallels that provide compelling evidence that adults without access to numeric language face difficulties when simply attempting to differentiate quantities greater than three.