Showing papers in "Language in 2010"

Comparative concepts and descriptive categories in crosslinguistic studies

Abstract: In this paper I argue that cross-linguistic grammatical comparison cannot be based on grammatical categories, because these are language-specific. Instead, typology must be (and usually is) based on a special set of comparative concepts that are specifically created by typologists for the purposes of comparison. Descriptive formal categories cannot be equated across languages because the criteria for category-assignment are different from language to language. This old structuralist insight (called categorial particularism) has recently been emphasized again by several linguists, but the idea that typologists need to identify "crosslinguistic categories" before they can compare languages is still widespread. Instead, what they have to do (and normally do in practice) is to create comparative concepts that help them to identify comparable phenomena across languages and to formulate cross-linguistic generalizations. Comparative concepts have to be universally applicable, so they can only be based on other universally applicable concepts: conceptual-semantic concepts, formal concepts, general concepts, and other comparative concepts. If, by contrast, one espouses categorial universalism and assumes crosslinguistic categories, as many generative linguists do, typology works by equating comparable categories in different languages, which are said to "instantiate" a cross-linguistic category. But in typological practice, all that is required is that a language-specific category matches a comparative concept. For example, the Russian Dative, the Turkish Dative and the Finnish Allative all match the comparative concept 'dative case', but they are very different distributionally and semantically and therefore cannot be equated and cannot instantiate a cross-linguistic category 'dative'. Comparative concepts are not always purely semantically-based concepts, but outside of phonology they usually contain some semantic components. If one is not confident about the universality of meanings, one can substitute extralinguistic contexts for universal meanings. The view that descriptive categories are different across languages and different from comparative concepts leads to terminological problems, which are also discussed here. Finally, I observe that the adoption of categorial universalism has actually impeded, not facilitated, cross-linguistic research.

Predicting syntax: Processing dative constructions in American and Australian varieties of English

Abstract: The present study uses probabilistic models of corpus data in a novel way, to measure and compare the syntactic predictive capacities of speakers of different varieties of the same language. The study finds that speakers' knowledge of probabilistic grammatical choices can vary across different varieties of the same language and can be detected psycholinguistically in the individual. In three pairs of experiments, Australians and Americans responded reliably to corpus model probabilities in rating the naturalness of alternative dative constructions, their lexical-decision latencies during reading varied inversely with the syntactic probabilities of the construction, and they showed subtle covariation in these tasks, which is in line with quantitative differences in the choices of datives produced in the same contexts.

Cognitive constraints and island effects

Abstract: Competence-based theories of island effects play a central role in generative grammar, yet the graded nature of many syntactic islands has never been properly accounted for. Categorical syntactic accounts of island effects have persisted in spite of a wealth of data suggesting that island effects are not categorical in nature and that non-structural manipulations that leave island structures intact can radically alter judgments of island violations. We argue here, building on work by Deane, Kluender, and others, that processing factors have the potential to account for this otherwise unexplained variation in acceptability judgments.We report the results of self-paced reading experiments and controlled acceptability studies which explore the relationship between processing costs and judgments of acceptability. In each of the three self-paced reading studies, the data indicate that the processing cost of different types of island violations can be significantly reduced to a degree comparable to that of non-island filler-gap constructions by manipulating a single non-structural factor. Moreover, this reduction in processing cost is accompanied by significant improvements in acceptability. This evidence favors the hypothesis that island-violating constructions involve numerous processing pressures that aggregate to drive processing difficulty above a threshold so that a perception of unacceptability ensues. We examine the implications of these findings for the grammar of filler-gap dependencies.

English filler-gap constructions

Abstract: This article delineates and analyzes the syntactic and semantic parameters of variation exhibited by English filler -gap constructions . It demonstrates that a detailed, fully explicit account of the observed variation is available within a framework embracing the notion ‘grammatical construction’. This account, which explicates similarities and differences among topicalization, interrogatives, relatives, exclamatives, and comparative correlatives in terms of linguistic types and hierarchical constraint inheritance, is articulated in detail within the framework of sign -based construction grammar (SBCG), a version of head -driven phrase structure grammar (HPSG) integrating key insights from Berkeley construction grammar . The results presented here stand as a challenge to any analysis incorporating transformational operations, especially proposals couched within Chomsky’s ‘Minimalist program’.

The centrality of metrical structure in signaling information structure: A probabilistic perspective

Abstract: This article introduces a new way to explain how information structure is signaled prosodically in English. I claim that METRICAL STRUCTURE plays a central role (Ladd 2008, Truckenbrodt 1995). Information structure (defined as in Steedman 1991 and Vallduvi & Vilkuna 1998) places strong constraints on the PROBABIILISTIC mapping of words onto metrical prosodic structure—that is, foci usually align with nuclear accents and theme/rheme units with prosodic phrases, and themes are less metrically prominent than rhemes. It is shown that focus position, scope, and pragmatic interpretation are then derived by manipulating EXPECTED PROMINENCE within metrical structure. Broadly, the more prominent a word than expected, the more likely a contrastive reading; the less prominent, the more likely a givenness reading. Both constructed and naturally occurring examples from the Switchboard corpus are used.

Convergent evidence for categorial change in French: From subject clitic to agreement marker

Abstract: The status of subject clitics in French has been heavily debated (Kayne 1975, Rizzi 1986, Roberge 1990, Auger 1994b, Miller & Sag 1997, De Cat 2007b, and many others). Distributional properties of French subject clitics have led Kayne (1975), Rizzi (1986), and others to analyze them as argument-bearing elements occupying canonical subject position, cliticizing to the verb only at the level of the phonology. While this hypothesis enjoys a wide following, a growing body of evidence suggests that it fails to capture patterns of subject-clitic use in colloquial French dialects/registers (Roberge 1990, Auger 1994b, Zribi-Hertz 1994, Miller & Sag 1997). Using new evidence from prosodic and corpus analyses, speaker judgments, and crosslinguistic typology, this article argues that (i) European Colloquial French exhibits differences from Standard French that impact how subject clitics are best analyzed, and more specifically (ii) subject clitics in European Colloquial French are affixal agreement markers, not phonological clitic arguments.

Developing Noun Phrase Complexity at School Age: A Text-Embedded Cross-Linguistic Analysis

Abstract: Development of noun phrase structure and use is analyzed as an important facet of syntactic acquisition from middle childhood to adolescence. Noun phrases occurring in narrative and expository texts produced in both speech and writing by 96 native speakers of English and Hebrew were identified and examined by a set of specially devised criteria including length in words, syntactic depth, abstractness of head nouns, and nature of modifiers. Results reveal a clear and consistent developmental increment in NP complexity from age 9 to 12, and particularly from age 16 years; written expository texts emerge as a favored site for use of syntactically complex constructions; and nominal elements play a more central role in the discursive syntax of Hebrew than English. Findings are discussed in terms of the interplay between psycholinguistic factors of cognitive processing constraints and the impact of increased literacy in later language development.

Cognitive and linguistic factors affecting subject/object asymmetry: An eye-tracking study of prenominal relative clauses in Korean

Abstract: Object relatives (ORs) have been reported to cause heavier processing loads than subject relatives (SRs) in both pre- and postnominal position (prenominal relatives: Miyamoto & Nakamura 2003, Kwon 2008, Ueno & Garnsey 2008; postnominal relatives: King & Just 1991, King & Kutas 1995, Traxler et al. 2002). In this article, we report the results of two eye-tracking studies of Korean prenominal relative clauses that confirm a processing advantage for subject relatives both with and without supporting context. These results are shown to be compatible with accounts involving the accessibility hierarchy (Keenan & Comrie 1977), phrase-structural complexity (O’Grady 1997), and probabilistic structural disambiguation (Mitchell et al. 1995, Hale 2006), partially compatible with similarity-based interference (Gordon et al. 2001), but incompatible with linear/temporal analyses of filler-gap dependencies (Gibson 1998, 2000, Lewis & Vasishth 2005, Lewis et al. 2006).

Iconicity and metaphor: Constraints on metaphorical extension of iconic forms

Abstract: Some conceptual metaphors common in spoken languages are infelicitous in sign languages. The explanation suggested in this article is that the iconicity of these signs clashes with the shifts in meaning that take place in these metaphorical extensions. Both iconicity and metaphors are built on mappings of two domains: form and meaning in iconicity, source domain and target domain in metaphors. Iconic signs that undergo metaphoric extension are therefore subject to both mappings (Taub 2001). When the two mappings do not preserve the same structural correspondence, the metaphorical extension is blocked. This restriction is formulated as the double-mapping constraint , which requires multiple mappings to be structure-preserving. The effects of this constraint go beyond explaining possible and impossible metaphors in sign languages. Because of the central role of metaphors in various linguistic processes, constraints on their occurrence may affect other linguistic structures and processes that are built on these metaphors in both sign and spoken languages.

Context effects on young children’s language use: The influence of conversational setting and partner:

Abstract: This article reports on two studies investigating the effect of contextual variables on young children’s language use in conversation. In Study 1, 20 children between age 1;5 and 2;2 were recorded ...

How New Technology Influences Parent—child Interaction: The Case of e-book Reading:

Abstract: This article reports on a study focusing on mother—child interactions during e-book reading compared to print book reading. Two different types of e-books were used, commercial and educational. Forty-eight kindergarten children and their mothers were assigned randomly to one of four groups, reading: (1) the printed book Just grandma and me; (2) the electronic commercial book Just grandma and me; (3) the printed book The tractor in the sandbox; and (4) the electronic-educational book The tractor in the sandbox. Compared to the printed book reading, e-book reading yielded more discourse initiated by the child and more responsiveness to maternal initiations. Printed book reading yielded more initiations and responses of mothers. Discourse during printed book reading compared to the digital context showed more expanding talk. Educational e-book reading showed more word meaning than reading the commercial e-book. The study concludes that different reading contexts influence adult—child interactions, and this m...

What 'you' and 'I' mean to each other: Person indexicals, self-ascription, and theory of mind

Abstract: This article offers a DE SE THEORY of person indexicals, wherein first- and second-person indexical pronouns indicate REFERENCE DE SE (also called SELF-ASCRIPTION ). Long observed for first-person pronouns (Castaneda 1977, Kaplan 1977, Perry 1979, inter alia), self-ascription is extended here to second person as well. The person feature of a pronoun specifies the speech-act roles that must be played by the self-ascribers: the speakers (uttering a first-person pronoun), the addressees (interpreting a second-person pronoun), or both (for first-person inclusive). Other agents who are not among the designated self-ascribers for a given pronoun interpret the pronoun indirectly by inferring the self-ascriber's interpretation, a process requiring THEORY OF MIND , that is, the cognitive ability to impute mental states to others (Premack & Woodruff 1978). This de se theory is supported by convergent evidence from multiple domains: (i) It explains a typological universal: first- and second-person plurals always allow associative semantics ('speaker(s) plus others', 'addressee(s) plus others') rather than requiring regular plural semantics ('speakers only', 'addressees only') (Greenberg 1988, Noyer 1992, Cysouw 2003, Bobaljik 2008). (ii) It belongs to a family of approaches that solve the problem of the essential indexical (Perry 1979). (iii) It correctly predicts observed patterns of indexical pronoun production and comprehension by two populations lacking a fully developed theory of mind: typically developing children in the stage before theory of mind has developed, and children with autism. (iv) It correctly predicts the interpretation of second-person pronouns in utterances with multiple addressees.

What is a perfect state

Abstract: Although many previous studies have tried to explain the English perfect’s various readings, none of them have been entirely successful. In this article, we argue that the perfect is pragmatically, rather than semantically, ambiguous. The meaning of the perfect introduces a base eventuality and a perfect state whose category is underspecified semantically. Neo-Gricean reasoning leads hearers to appropriately fill in the value of that variable. We present the results of a corpus study of over 600 present perfect examples. The results of this study suggest that (i) most English present perfects receive entailed resultative or continuative readings, (ii) the English perfect need not elaborate on a preexisting topic, and (iii) the English perfect plays a role in establishing discourse coherence by helping hearers establish discourse relations between discourse segments.

Hebrew Adjectives in Later Language Text Production

Abstract: The study investigates the distribution and use of adjectives in 252 texts produced by 63 Hebrew-speaking children, adolescents, and adults who were asked to tell and write a story about a personal...

Development of word order in German complement-clause constructions: Effects of input frequencies, lexical items, and discourse function

Abstract: We investigate the development of word order in German children’s spontaneous production of complement clauses. From soon after their second birthday, young German children use both verb-final complements with complementizers and verb-second complements without complementizers. By their third birthday they use both kinds of complement clauses with a variety of complement-taking verbs. Early in development, however, verb-final complements and verb-second complements are used with separate sets of complement-taking verbs, and they are used with separate sets of item-specific main-clause phrases. For example, initially phrases such as ‘I want to see’ were used exclusively with verb-final complements, whereas phrases such as ‘do you see’ and ‘you have to say’ were used exclusively with verb-second complements. Only later in development— when specific complement-taking verbs were used with both verb-second and verb-final complements, with a greater variety of main-clause phrases, and when specific main-clause phrases were used with both verb-second and verb-final complements—was there evidence for structural links between these various, item-based, complement-clause constructions.

Adult offer, word-class, and child uptake in early lexical acquisition

Abstract: How do adults offer new words from different parts of speech? This study examined the offers in book-reading interactions for 48 dyads (parents and children aged 2- to 5-years-old). The parents relied on fixed syntactic frames, final position, and emphatic stress to highlight unfamiliar words. As they talked to their children about the referent objects, events, or scenes, they also linked new words to other terms in the pertinent semantic domain, thereby presenting further information about possible meanings. Children attended to new words, often repeating them in the next turn, and, as they got older, they too related new words to familiar terms as they talked about their referents with their parents. These data add further evidence that interaction in conversation supports the process of language acquisition.

Exploring the Feasibility of a Pedagogy of Multiliteracies in Introductory Foreign Language Courses

Abstract: The 2007 MLA Report calls for large-scale reform in university foreign language (FL) departments to integrate the study of language, literature, and culture and move beyond the the language-content dichotomy that has characterized the undergraduate curriculum for decades. This article explores the implications of these recommendations for introductory FL courses, arguing in favor of a pedagogy of multiliteracies (New London Group, 1996; Kern, 2000) as one pathway toward curricular reform. The adoption of a multiliteracies framework in response to calls for curricular change is not entirely novel, yet most scholarship to date has focused on the need for more explicit attention to students' linguistic development in advanced-level content courses rather than on pedagogical models for integrating textual content into introductory language courses. To support our position, three challenges to realizing curricular change and fostering literacy in introductory FL courses are discussed – pedagogy, course content, and departmental buy-in – and strategies to address each challenge are proposed. We conclude that in light of the changing landscape in U.S. higher education today, a pedagogy of multiliteracies represents a means of keeping the introductory FL curriculum relevant to students as well as the broader intellectual mission of the university.

The structure of lexical meaning: Why semantics really matters

Abstract: This article explores the architecture of the interface between morphosyntax and lexical semantics, in particular the semantic underpinnings of argument realization. Many theories of lexical meaning assume that argument realization is derived from underlying event structure: the relative prominence of coarguments in a clause follows from their relative semantic prominence in how the event unfolds. I show that event structure is not sufficient to capture certain generalizations about argument realization, however, focusing on arguments that alternate between direct and oblique realization. I show that for these alternations the relevant semantic contrast is in strength of truth conditions: direct realization encodes a monotonically stronger set of truth conditions associated with the alternating argument than oblique realization. This, I suggest, follows if word meanings are built from basic units that are related to one another implicationally, and the relative implicational strength of such components figures into argument realization. I use as a case study English locative and conative alternations, which, I argue, reflect stronger and weaker degrees of affectedness along an independently motivated affectedness hierarchy . I also show that similar contrasts are found with other alternations on other hierarchies. I conclude by suggesting that a theory of weakening truth conditions is not incompatible with event-structural analyses of verb meaning, and in fact the two augment one another.

The acquisition of early verbs in French: Assessing the role of conversation and of child-directed speech

Abstract: This article examines the production of early verbs by two children acquiring French as their first language. The study focuses on the developmental period during which verbs are produced in one form only. Child-directed speech (CDS) and conversational contingencies (CC) occurring around these verbal forms were analysed up to the moment when some verbs are produced in two different forms. Results show that children’s use of a single form per verb can also be found in CDS by adults where the majority of verbs are used in one morphophonological form only. Moreover, the particular form children use for a given verb corresponds to the one adults predominantly use in CDS. At the same time, child-produced verb forms are reinforced in the CC occurring in adult—child exchanges. When trying to separate the role of CCs from that of more general CDS, for both children the study found that for about half of the verbal forms CDS and CC provide the same congruent information. Of the remaining verb types, three-quarters...

Foreign Language Teachers’ Struggle to Learn Content-Based Instruction

Abstract: Research has shown content-based instruction (CBI) to be effective in various language settings, yet this promising curricular approach remains rarely implemented in mainstream foreign language educational contexts. While the existing body of research has identified important barriers to the implementation of CBI, it has neglected the problem of meaning which is essential to understanding educational reforms. This phenomenological study explores the meaning that the experience of learning CBI had for in-service foreign language teachers in traditional teaching contexts who were once enrolled in a year-long professional development program specifically designed to help them become familiar with CBI core principles and create CBI curricular materials. Findings suggest that teachers struggle mainly with the idea of teaching language through content, a concept they have difficulty grasping or even accepting as a possibility. Professional development programs must be designed to respond to this specific challenge if they are to help teachers explore new instructional possibilities.

Validity of a Parent Report Instrument for Irish-speaking Toddlers

Abstract: Previous research has demonstrated the validity of parent report as an assessment of early vocabulary and grammatical development across a range of languagesThis study adapted a well-known tool, the MacArthur—Bates Communicative Development Inventories (MCDI), to Irish (ICDI), in order to measure vocabulary and grammatical development longitudinally across a sample of 21 children Results from the parent checklists were validated against spontaneous language samples Correlations demonstrated high concurrent validity for ICDI reported vocabulary with measures of lexical diversity derived from the spontaneous samples In addition, ICDI measures of grammar also correlated highly with various indices of grammar from the spontaneous samples Although both ICDI and spontaneous vocabulary measures were highly correlated, the ICDI checklist includes a broader range of language skills and so would seem to capture the range of language ability more comprehensively than direct observation The findings have implic

Variable affix order: Grammar and learning

Abstract: While affix ordering often reflects general syntactic or semantic principles, it can also be arbitrary or variable. This article develops a theory of morpheme ordering based on local morphotactic restrictions encoded as weighted bigram constraints. I examine the formal properties of morphotactic systems, including arbitrariness, nontransitivity, context-sensitivity, analogy, and variation. Several variable systems are surveyed before turning to a detailed corpus study of a variable affix in Tagalog. Bigram morphotactics is shown to cover Tagalog and the typology, while other formalisms, such as alignment, precedence, and position classes, undergenerate. Moreover, learning simulations reveal that affix ordering under bigram morphotactics is subject to analogical pressures, providing a learning-theoretic motivation for the specific patterns of variation observed in Tagalog. I raise a different set of objections to rule-based approaches invoking affix movement. Finally, I demonstrate that bigram morphotactics is restrictive, being unable to generate unattested scenarios such as nonlocal contingency in ordering.

Explanations in naturally occurring peer talk: Conversational emergence and function, thematic scope, and contribution to the development of discursive skills:

Abstract: This study examines explanations in children’s natural peer talk, as they emerge in collaborative interaction among children, in two cohorts: preschoolers (aged 4—5) and preadolescents (aged 9—10). The study examined 322 explanations for the diversity of their content-components, their modes of emergence, and social functions. Quantitative analysis revealed a wide range of explanation topics in both cohorts, with a shift with age from the immediate to the more distant. Discourse analysis of explanatory sequences demonstrated a high sensitivity to conversational notions of expectedness and an effective use of explanations for a rich array of pragmatic and social functions, as well as the affordances of peer talk explanations as a potential site for learning and developing the discursive skills of decontextualized discourse.

Interpretation and Recall of Proverbs in Three School-age Populations

Abstract: The study investigates schoolchildren’s command of proverbs as a facet of figurative language, testing their ability to go beyond the referential content of the linguistic message and their familiarity with established non-literal sayings as indicative of lexical development. The tasks involved (1) interpretation of unfamiliar proverbial sayings that are non-conventionalized in Hebrew — in context-free and contextualized conditions — and (2) recall of established traditional Hebrew proverbs. Participants were 4th- and 8th-graders from three populations: typically developing children of high and low SES backgrounds respectively and a group of high SES language-impaired children. Results show a clear rise in performance with age and schooling on both tasks, with greater success in interpreting novel sayings than in recalling traditional proverbs. The language-impaired group scored lowest on all tasks, with the low SES children doing less well than their high SES peers on interpretation but better on recall.

Dialogical factors in toddlers’ use of clitic pronouns:

Abstract: Young (1;9—2;4) children’s use of third person clitic subject pronouns in natural dialogues was examined in both longitudinal and cross-sectional data. Considering that young children mainly use pronouns in the context of referential continuity, this study aims at identifying some of the factors that affect this use. Two possible dialogical factors are examined: (1) the use of clitic pronouns can be interpreted as a reproduction of the adult’s discourse, either by taking up whole utterances containing a pronoun or by taking up only the clitic pronouns without reproducing the adult’s utterance. (2) The use of pronouns could be driven by pragmatic-discursive factors. In order to assess this hypothesis the use of clitic pronouns was observed in the context of dialogical continuity. Three kinds of links were considered: children repeat or reformulate the adult’s utterances, add a new predication on the same topic, or establish a contrast. The results suggest that the reproduction of the adult’s utterance does...

Language-particular syntactic rules and constraints: English locative inversion and do-support

Abstract: Locative inversion in English ( under the bridge lived a troll ) is ungrammatical in all of the contexts where do -support applies: subject-auxiliary inversion, sentential negation, emphasis or verum focus, VP ellipsis, and VP displacement. Importantly, it is ungrammatical in these contexts whether do -support applies or not: it is ungrammatical with other auxiliaries, and it is also ungrammatical in nonfinite clauses of these types, where do -support never actually applies. This indicates that all of these contexts have something in common, and that cannot be disruption of adjacency between tense/agreement and the verb because there is no such disruption with other auxiliaries or in nonfinite contexts. These facts therefore argue against the standard last-resort theory of do -support, which holds that it is inserted to save a stranded tense/agreement affix, and for a theory like that of Baker 1991. In this theory, VPs have corresponding SPECIAL PURPOSE ([SP]) VPs, and do heads a [SP] VP. All of the contexts for do -support have in common the featural specification [SP]. Locative inversion involves a null expletive subject, the licensing of which is blocked by a non-[SP] context. All of this argues for a view of syntax with language-particular licensing constraints, features, and rules, within a range of variation proscribed by universal grammar.

The interplay between comparative concepts and descriptive categories (Reply to Newmeyer)

Abstract: In his reply to my paper 'Comparative concepts and descriptive categories in cross-linguistic studies' (Haspelmath 2010, henceforth H10), Newmeyer (henceforth N) defends the widespread old idea that there are cross-linguistic categories and that the same kinds of concepts are and should be used both in cross-linguistic studies and in language-particular description. However, his critique does not amount to a full-blown endorsement of the categorial universalist position, which in its prototypical generative form assumes a single set of categories or features for both purposes. For instance, when N says that 'it is only by means of working out the interplay between the language-particular and the language-independent that we can hope to understand either' (p. 2-3), he implies that the two are not identical, and that the relationship between them is non-trivial, much as I have emphasized. The bulk of N's reply is devoted to listing various problems in H10's examples that he sees as undermining its general claims, especially with respect to the notions of subject, thematic roles, adjectives, clause, and word. In this brief reply, I will try to show that none of them are real, and that a closer consideration of these cases actually strengthens my position. N first observes that despite my criticism of the equation of languageparticular 'subject' categories across languages, I do use 'subject' as a comparative concept in some of the generalizations and definitions (in particular, in (3) and (13)). I did not give the relevant definition in H10 for lack of space, hoping that it could be inferred, but here it is: 'the agent of a simple transitive clause' (the same definition is assumed by Greenberg 1963). This is of course a very different concept from 'Subject in Tagalog' (defined in terms of case-marking by ang, very often different from the agent NP) and from 'Subject in English' (perhaps defined as in N's (6), comprising expletive there, which surely is not an agent). Thus, this underscores rather than undermines my central point. N's own concrete proposal for defining 'subject' ('[an entity that manifests] more of the core properties of subjects than of constructs that contrast with subjects', p. 12) is also a possible comparative concept, provided that the 'core properties' can be defined in a universally applicable way. However, as I noted in H10 (§7.3) in the discussion of Keenan's (1976) definition of 'subject', it is unclear how one can draw up such a list of core properties on a principled basis. And most crucially, it is unlikely that such a definition will be found useful in language-particular studies. Descriptive categories are normally defined in terms of specific conditions, not in vague quantitative terms ('more properties than'). N also criticizes me for using semantic roles such as 'recipient' (in 2) and 'agent' (in 16) (even though he uses the latter concept himself in (6)). But he overlooks that the problems with such semantic roles have never arisen in a typological context. Nobody has ever seen problems with defining the comparative concept 'ergative' in terms of 'agent'. It is only when one tries to describe particular languages with a single universal set of semantic roles that one runs into problems. But we can describe I play the sonata in purely grammatical terms, without reference to general semantic roles. Generative