This study reconsiders the acquisition of relative clauses based on data from two sentence-repetition tasks. Using materials modeled on the relative constructions of spontaneous child speech, we asked four-year-old English- and German-speaking children to repeat six different types of relative clauses. Although English and German relative clauses are structurally very different, the results were similar across studies: intransitive subject relatives caused fewer errors than transitive subject relatives and direct object relatives, which in turn caused fewer errors than indirect object relatives and oblique relatives; finally, genitive relatives caused by far the most problems. Challenging previous analyses in which the acquisition of relative clauses has been explained by the varying distance between filler and gap, we propose a multifactorial analysis in which the acquisition process is determined primarily by the similarity between the various types of relative clauses and their relationship to simple sentences.

Noun incorporation (NI) in Mapudungun is different from NI in better-studied languages like Mohawk in three ways: the incorporated noun is invisible to verbal agreement, incorporation into unaccusative verbs is impossible unless a possessor is stranded, and possessors are the only modifiers that can be stranded. These differences can be explained by saying that the trace of NI retains its person, number, and gender features in Mohawk but not in Mapudungun. Those aspects of grammar that do not involve these features treat NI in the two languages the same; thus, NI has the same gross distribution and anaphoric possibilities in both languages. We extend these results to Nahuatl, Chukchee, Ainu, Southern Tiwa, Mayali, and Wichita, showing that our theory accounts for Mithun’s (1984) distinction between Type III and Type IV noun incorporation in a general way.

It is widely believed that the difference between regular and irregular verbs is restricted to form. This study questions that belief. We report a series of lexical statistics showing that irregular verbs cluster in denser regions in semantic space. Compared to regular verbs, irregular verbs tend to have more semantic neighbors that in turn have relatively many other semantic neighbors that are morphologically irregular. We show that this greater semantic density for irregulars is reflected in association norms, familiarity ratings, visual lexical-decision latencies, and word-naming latencies. Meta-analyses of the materials of two neuroimaging studies show that in these studies, regularity is confounded with differences in semantic density. Our results challenge the hypothesis of the supposed formal encapsulation of rules of inflection and support lines of research in which sensitivity to probability is recognized as intrinsic to human language.

Segmental features of child-directed speech (CDS) were studied in a corpus drawn from thirty-nine mothers living in Tyneside, England. Focus was on the phonetic variants used for (t) in word-medial and word-final prevocalic contexts since it is known that these variants display clear sociolinguistic patterning in the adult community. Variant usage in CDS was found to differ markedly from that in interadult speech. Effects were also found with respect to the age and gender of the children being addressed. Speech to girls generally contained more standard variants than speech to boys, which, by contrast, contained higher rates of vernacular variants. The differentiation by gender was most apparent for the youngest children. The findings are assessed in comparison to other studies of CDS. It has previously been claimed that modifications made in the CDS register help children to learn linguistic structures and also to learn that speech is a social activity. Our findings suggest that CDS may play an additional role, providing boys and girls as young as 2;0 with differential opportunities to learn the social-indexical values of sociolinguistic variables.

This article describes the syntactic distribution and semantic interpretation of determiners in Tongan (Polynesian). This determiner system is unusual because it contains, in addition to the familiar definite and indefinite, a third contrasting morpheme, labeled ‘semi-definite’ in traditional descriptions. A description of the behavior of these three determiners is offered that makes use of some conceptual tools provided by recent crosslinguistic work in formal semantics.

Chomsky and Halle (1968) and many formal linguists rely on the notion of a universally available phonetic space defined in discrete time. This assumption plays a central role in phonological theory. Discreteness at the phonetic level guarantees the discreteness of all other levels of language. But decades of phonetics research demonstrate that there exists no universal inventory of phonetic objects. We discuss three kinds of evidence: first, phonologies differ incommensurably. Second, some phonetic characteristics of languages depend on intrinsically temporal patterns, and, third, some linguistic sound categories within a language are different from each other despite a high degree of overlap that precludes distinctness. Linguistics has mistakenly presumed that speech can always be spelled with letter-like tokens. A variety of implications of these conclusions for research in phonology are discussed.

In an article in a recent issue of Language (Newmeyer 2003), Frederick Newmeyer argues for a clear separation between what he terms ‘knowledge of language [i.e. grammar] and use of language [i.e. pragmatics]’ (682). In developing his argument, N makes frequent reference to linguistic analyses that are corpus-based, that is, centered on information taken from large datasets of actual speech or writing (e.g. transcriptions of conversations, newspaper articles, novels). In N’s view, grammar is distinct from pragmatics, and because a corpus contains examples of actual language use drawn from a large community of speakers and writers, it can yield only performance data. As a result, N argues, ‘there is no way that one can draw conclusions about the grammar of an individual from usage facts about communities, particularly communities from which the individual receives no speech input’ (696).