Search

The typological classification of Mandarin Chinese as an isolating language lacking inflection brings about a theoretical consideration that word order in Mandarin serves as an interface between syntax and pragmatics. This paper provides from typological perspective a unified framework to examine the role that word order in Mandarin plays in marking case as well as topics. The study finds that Mandarin exhibits a mixed case-marking pattern through word order, namely that nominative-accusative pattern coexists with ergative-absolutive pattern. Both of the two case-marking patterns allow intra-clausal topics and extra-clausal topics, and it is the intra-clausal topics rather than the extra-clausal ones that are associated with case-marking patterns. With respect to the intra-clausal topics, in nominative-accusative pattern, topic assignment in unmarked constructions is A/S oriented, whereas in ergative-absolutive pattern, topic assignment in unmarked two/three-argument constructions is post-ba P oriented, and topic assignment in unmarked one-argument constructions is preverbal-S oriented. In light of the two case-marking patterns, context-independent topics in various constructions can be identified, and thus the controversial issue of distinguishing topics from other arguments can be properly addressed.

Researchers have long noticed the phenomenon of inflection decay in the evolution of English, which may be driven by the Principle of Least Effort ( Millward and Hayes, 2010 , Zipf, 1949 ). However, most studies in this line of research were based on exemplar-oriented observations while few explored the phenomenon from a quantitative perspective. In addition, it remains unclear whether the trend of inflection decay has continued in Modern English. In the present study, an index named Inflectional Entropy was developed to measure the diachronic change of the inflectional trend of Modern English in the 20th century. A trend of inflection decay in Modern English was found based on experiments on several large-scale corpora. The findings provide quantitative evidence to linguistic theories such as Principle of Least Effort, Language Niche Hypothesis ( Lupyan and Dale, 2010 ), and Principle of Entropy Minimization ( Ferrer-i-Cancho, 2018 ).

In spite of earlier implantation, bilateral implants and advances in the care of deaf children with cochlear implants (CI) within-group variation in language skills is large. We use three tasks with predictive value of language impairment in Swedish to identify children with CI most at risk for persistent difficulties in language development, in need of language intervention. The clinical markers investigated are nonword repetition and past tense inflection. We also assessed language comprehension, a predictor of severe language impairment associated with poorer prognosis. Fifteen Swedish-speaking deaf children with CI and 15 controls aged 5–8 years participated. Most children with CI had bilateral implants and had been fitted with CIs before 12 months. At least 70% in the group with CI performed >1.25 and 47% >2 SD below controls on more than one measure, showing risk for persistent language impairment. Speech perception was more crucial in the nonword repetition and language comprehension tasks. No time factor was significantly related to outcome. We conclude that it is important to allocate resources for continuous follow-up and language intervention. Research and care of children with CI will profit from better integration of knowledge from the fields of audiology, speech-language pathology and linguistics.

This article shows that in Jumjum, a little-described language belonging to the Southern Burun subbranch of Western Nilotic, subjects and objects are obligatorily cross-referenced in the verb. Only definite patients are coded as objects, while indefinite patients are coded as non-core arguments of antipassive verbs. Jumjum and its sister language Mabaan are morphosyntactically unique among the Western Nilotic languages in three respects. Firstly, the object of a transitive verb always precedes the verb (OV order). Secondly, subjects and objects are always expressed by the inflection of the verb, which thus cross-references subject and object noun phrases. And thirdly, some subject suffixes are deleted before some object suffixes, which gives rise to many cases of segmentally homonymous forms, but these forms are kept phonologically distinct by their tone patterns. The article documents these features by also taking into account the effect of derived verb stems and of tense inflection on the expression of cross-reference. To some extent, the derivational morphemes are expressed by suffixation, but in addition, root-internal alternations play a major role, including alternations in vowel quality, vowel length, and the root-final consonant.

Construction Grammar is a linguistic framework which postulates that all grammar is based on a single, relatively simple entity, the construction, defined as a conventionalized mapping between a form and a meaning. Construction Morphology is a sub-theory of Construction Grammar which deals with the word-level domain. This article shows how Construction Morphology can account for the full range of inflectional types. It argues that Construction Morphology falls into the type of morphological theory which Stump (2001) has designated ‘inferential—realizational’, and which he contends is the only kind of theory that accounts for all the known facts of morphology.

The present study demonstrates the existence in Wandala (Central Chadic) of two inflectional markers which, although they change the form of the word, are not word-formation means but rather sentence-formation means. While neither of the forms indicates any specific semantic relation, they both enable the listener to parse the sentence into units that can undergo further analysis and allow for inferences about a large number of semantic relations within the clause and within the sentence. The types of inflectional marking described in this study are unrelated to the dichotomy of head marking and dependent marking. Although this study is based on data from a single language, it has implications for the typology of coding means and for theories of clausal structures.

Verbal accent in Middle Korean (15–16th c., MK) is not 100% predictable but very strongly correlates with the segmental shape of the stem to varying degrees, leading Ramsey (1978, 1986, 1991, 2001) and Whitman (1994) to reconstruct predictable accent for both stems and suffixes in the protolanguage. In a traditional rule-based approach or in optimality theory, the only possible analysis of the data of MK verbal accent is to list the stems with exceptional accent, and it is difficult to explain the gradient lexical frequency patterns. However, it is now possible to capture these gradient patterns and construct a grammatical model in terms of weighted constraints. This paper first attempts to formalize the tone-segment correlation of MK verbal stems noticed in previous research with a weighted constraint OT grammar (Hayes, 2009). Then, we compare the MK grammar with a corresponding investigation of the accentuation of monosyllabic verbal stems in contemporary Yanbian Korean (north-eastern China). An important fact with regard to the two stages of the language is that due to sound changes, some segmental contrasts that were crucial in classifying various accent sub-categories in the verbal inflection have merged in Yanbian, and as a result the original association between accent class and segmental shape seen in MK became partially obscured. In the face of these changes, Yanbian speakers have restructured the correlation between accent class and segmental shape. They either retained the original accent in spite of the segmental change (conservative accent) or have shifted the accent class of a given verbal stem based on the output of the segmental change (innovative accent). The innovative accent attests to the productivity of these tone-segment associations. This finding is modeled as a diachronic restructuring of the weight hierarchy of relevant OT constraints based on the updated lexical statistics, which led to a shift in the accent.

The present paper investigates null arbitrary subjects in Greek that trigger either a generic or an existential interpretation. Null arbitrary subjects in Greek are permitted with 2sg , 1pl and 3pl verbal agreement. The synergy between imperfective marking alongside a clause-initial XP (overt or null) results in genericity, whereas perfective marking alongside an overt or covert clause-initial XP corroborates an existential reading. Moreover, the difference between a 2sg generic reading vs. a 2sg referential one amounts to the absence vs. the presence of a D-feature (=[ definite ]) on the verbal agreement morpheme/inflection (i.e. empty [ d ] vs. specified [ d ]-marked verbal agreement). Contra Alexiadou and Anagnostopoulou (1998) , V-T movement is unable to delete the D-feature of the EPP on T in these arbitrary constructions. Instead, both generic and existential constructions satisfy the EPP in a uniform manner; an XP (overt or covert) merges on SpecTP and satisfies the D-feature of the EPP, whereas movement of the verb to T satisfies the remaining agreement feature bundle of T. It is further postulated that the defective verbal morphology of these constructions also satisfies an Arb-gen or an Arb-ex feature through movement to AspP.

This comparative study examines the morphosyntactic parameters governing the distribution of ‘special inflection’ in constituent interrogatives and focus constructions in Hausa and Coptic Egyptian. In these languages, ‘relative’ tense-aspect-mood (TAM) markers occur in relative clauses, constituent interrogatives and declarative focus constructions. However, special inflection is not a clause-typing device but is governed by syntactic conditions, since both languages also have focus/wh-constructions lacking relative TAMs: both languages allow in situ and ex situ focus/wh-constructions, but while Hausa special inflection occurs in ex situ constructions, Coptic special inflection occurs in in situ constructions. A transformational copy theory analysis reduces the parametric differences between Hausa and Coptic to different pronunciation sites of the displaced focus/wh-phrase: either the ‘top copy’ in its displaced position in the specifier of Focus Phrase (FP), or the ‘lower copy’ in its thematic position. An additional parameter reminiscent of the Doubly-filled Comp Filter is set for Hausa to allow both specifier and head of FP to be spelt out (pronounced) at once, resulting in the coincidence of fronting and special inflection, while the same parameter in Coptic is set to prohibit both the head and the specifier of FP spelling out at once, ruling out the coincidence of special inflection and fronting.

International audience; In this article, we examined how complex words are recognized as being mediated by their morphological operations and structure. French verbal inflection is a system where the stems provide the lexical meaning and the inflectional suffixes activate the functional information by the morphosyntactic features. We investigated the morphological decomposition and inflectional suffixes processing through visual lexical decision tasks. Experiment 1 accessed general differences in the number of morphological operations regarding low and high frequencies, and regular and irregular verbal forms (e.g., jou-ent/jou-ai-ent ‘they play/played’, prend-s/pren-ai-s ‘yousg take/took’). Experiment 2 tested specific differences in the tense and agreement inflectional suffixes (e.g., jou-ons/jou-i-ons/jou-ez/jou-i-ez ‘we/youpl play/played’). Our hypothesis is that words are automatically decomposed early for morphological processing and that morphemes are later hierarchically recombined for word recognition. We found significant differences between the number of morphological operations in regular and irregular verbs in low and high frequencies; we also found significant differences in tense and agreement suffix processing with longer responses for the past tense and first plural agreement verbal forms, suggesting additive effects. Our results are supported by single-mechanism pre-lexical decompositional models; we propose a model where stems and inflectional suffixes are processed differently for lexical access and word recognition.

This paper examines children's use of verbal agreement in comprehension in two varieties of Spanish with the goal of determining how the input affects their behavior. Both dialects are identical in allowing null subjects and having rich verbal morphology in contrast to English. The dialects differ however in that in Chilean Spanish a phonological lenition process affects the realization of second person singular /-s/. In Mexican Spanish, no such variation is found for the second singular. The results indicate that children acquiring languages with rich verbal morphology use agreement in comprehension in much the same way as children acquiring a language with less robust agreement. Moreover, we find that variable input for the second singular marker affects acquisition.

We review and discuss some issues to do with the relation between morphology and syntax which have played a prominent part in generative linguistic research in the past three decades. Focusing on verbal inflection, we first discuss the relation between inflection and verb placement, with special attention given to verb-initial languages. We then discuss the relation between pro-drop and agreement, where we articulate a partly new understanding of Huang's (1989) generalization that pro-drop is characteristic of languages with rich agreement and languages with no agreement, but not languages that are in-between. We then present and discuss the Mirror Principle, one of the most significant findings in recent linguistic research. We pay special attention to the Mirror Principle as it applies in head-final languages, in the context of a model adopting Kayne's (1994) Linear Correspondence Axiom. The idea is to show how fairly complex aspects of clausal syntax, including word order and the possibility of phonetically silent arguments of a predicate, may be correlated with readily observable and, in themselves, rather simple properties of verbal inflection, and to show how, given a restrictive theory of Universal Grammar, this follows from the fact that inflections are syntactic categories, albeit realized as parts of words.

Breton, a language rich in compounds with both initial and final heads, provides a unique opportunity to study the relation between the head of a compound and the meaning of the compound. A novel approach is proposed. The relation between the head and the meaning of a compound is captured by the distinction between grammatical and logical head. The grammatical head determines the syntax of the compound. It gives the grammatical category of the compound and is the locus of inflection. The logical head is determined by the logical relation between the two terms of the compound. When the grammatical head corresponds to the logical head, we have endocentricity. When the grammatical head does not correspond to the logical head, we have exocentricity. In the latter case the head is grammatical or logical; hence there are no headless compounds. The concept of grammatical and logical head can also be applied to derivation.

This study investigates the ability of two Greek-speaking agrammatic individuals to produce subject–verb Agreement, Tense, and Aspect in light of recent hypotheses formulated to explain the patterns of performance observed in agrammatic aphasia. The role of frequency and (morphological) regularity of verbs in agrammatism is also explored. For these investigations a sentence completion task was developed. Overall, Aspect was found significantly more impaired than Tense, and Tense significantly more impaired than Agreement. No dissociation was observed between past and future Tense, or between perfective and imperfective Aspect, for any of the two agrammatic participants. No verb regularity effect was detected and only a verb frequency effect in the Tense condition was observed for one of the participants. It is suggested that the two agrammatic participants have a morphosyntactic deficit, while one of them has an additional lexical deficit. Moreover, it is argued that the increased processing demands of Tense and Aspect, which carry interpretable features, render them more vulnerable compared to Agreement, which bears an uninterpretable feature and is a local, strictly grammatical operation. The significantly worse performance of the two agrammatic participants on Aspect than on Tense is attributed to the more “subjective” nature of the former category.

Children acquiring Basque begin producing absolutive agreement earlier in development than ergative inflection (Barrena, 1995; Ezeizabarrena, 1996). In this paper, I compare the acquisition of absolutive and ergative to dative agreement, which is also inflected on the auxiliary in Basque, in order to evaluate factors which may influence the order of emergence of inflectional morphemes, including input frequency, the position of the affix on the verb and feature complexity. Data from 20 bilingual and 11 monolingual children between 2;00 and 3;06 years old showed a consistent pattern; children acquired absolutive agreement first, then ergative, and dative inflection last of all. This sequence replicates the implicational hierarchy of cases found in many languages (Blake, 2001). I argue that the complexity of inflectional features of the agreement morphemes best predicts this developmental order, along with the assumption that children acquire the inflectional forms earliest which encode the fewest positive values for features.

In this paper, I analyze null subjects in absolute clauses (AC) in Spanish, comparing them to finite-clause null subjects and infinitival subjects. AC null subjects mostly pattern like finite-clause null subjects, but they cannot be identified by a rich inflection and they require a grammatically encoded antecedent (like some infinitival subjects). I propose that AC null subjects must be identified by a topic operator through the operation of AGREE, a matching operation between two categories in which one of them copies its values to the other. AC null subjects remain unvalued in their own clause, seeking an antecedent in the preceding clause through a mediating null topic in the absolute clause’s periphery. Finite-clause null subjects, on the other hand, are identified within their own clause by verbal morphology through the same AGREE mechanism. Two distinct ways of relating null subjects to discourse-antecedents are proposed, one for finite-clause subjects, one for absolute-clause subjects. By proposing a distinct discourse-identification process not mediated by rich agreement, this proposal opens the door to a unified theory of null referential categories, in which discourse-topic identification plays a central role.

Mesoclisis in the imperative and parasitic plurals in Spanish are currently accounted for either at Morphological Structure ( Halle and Marantz, 1994 ) or at the PF interface ( Harris and Halle, 2005 ). In previous work ( Manzini and Savoia, 1999 , Manzini and Savoia, 2004a , Manzini and Savoia, 2004b , Manzini and Savoia, 2005 , Manzini and Savoia, 2007 , Manzini and Savoia, 2008a , Manzini and Savoia, 2008b ) we argued that these and similar phenomena in both Romance and Albanian are best accounted for at the syntactic level. Since sub-word constituency is involved, this amounts to saying that syntax subsumes morphology. Here we defend the conclusions of our previous work against available alternatives, on the basis of a strictly lexicalist stance on the projection of morphosyntactic structures from the lexicon.

Japanese has been conventionally considered to show two distinct distributional regularities holding in morpheme-final position: at the lexical level, verb stems can end with either a vowel or a consonant, whereas other morpheme types and derived/inflected forms must end with either a vowel or the placeless nasal n . Mono-stratal models of phonology call into question the validity of the traditional underlying-surface distinction, and respond by allowing only a single type of static pattern. Unlike the standard derivational approach, however, the mono-stratal approach cannot express the disparity between lexical and non-lexical patterns ( Scobbie et al., 1996; Harris, 1997, 2004 ). This paper adopts a mono-stratal approach to investigate regularities in the pattern of morpheme/word-final segments in Japanese, and challenges the widespread view that Japanese employs consonant-final verb stems (e.g. jom ‘read’). Instead, it claims that the actual shape of apparently consonant-final stems is the same as that of non-past forms which end with ɯ (e.g. jomɯ ), the neutral vowel of Japanese. This has the result of excluding consonant-final stems from the Japanese system — a system which prefers vowel-final forms. In terms of phonology-phonetics mapping, the stem-final ɯ is the phonetic manifestation of a melodically empty nucleus. This analysis has the benefit of allowing us to eliminate operations such as ɯ -deletion and vowel alternation. By positing a morphologically driven process which avoids empty positions at the morphological boundary, the empty position is simply filled by the initial vowel of the negative suffix (e.g. jo.m ∅ [jomɯ] + - a . na .i → jo.ma.na.i [jomanai]).

Two recent studies ( Johnson et al., 2005 , Perez-Leroux, 2005 ) found that English- and Spanish-learning children do not show the ability to use verb inflection as a cue to subject number before the age of 5 to 6 years. These findings suggest an asymmetric development as verb inflections are usually correctly produced before this age. In the present study we investigated whether German 3- to 4-year-olds take advantage of the information provided by the verb inflection in sentence comprehension. In a first study, children's looking behavior at two pictures was measured after presentation of a sentence in which the subject number was coded only by the verb inflection. The results from this study suggest that children's looks reflect correct interpretation of the sentences and thus show their ability to make use of verb inflection. In a second experiment, preferential looking was combined with an additional task in which the children had to point to the matching picture. In this case children did not perform above chance level. Our results underline the relevance that specific task demands have on the performance of children in comprehension testing. These have to be accounted for when interpreting findings on production and comprehension asymmetries in language acquisition.

The ergative inflection of Kuuk Thaayorre nouns is remarkably irregular. More than 15 distinct ergative forms are attested, none of which is phonologically predictable. This ergative inflection is also remarkable for its optional use. In spite of its primarily syntactic function, it may also be used to mark pragmatically ‘unexpected’ intransitive subjects, or be omitted where the referent of a transitive subject is easily retrieved from the discourse context or world knowledge. This article proposes that both the formal irregularity of the Kuuk Thaayorre ergative morpheme and its optionality have their origins in a language-wide historical process of phonological erosion. The suite of diachronic processes this erosion triggered a bidirectional grammaticalisation pathway, whereby grammatical (ergative) morphs were imbued with pragmatic functions even as originally pragmatic (focal) morphs were reanalysed as grammatical.

Recent research has shown that there is more variation in the placement of verbs and adverbs in Scandinavian than previously thought. This paper argues that the theory of verb movement originally proposed by Bobaljik and Thrainsson [Bobaljik, J.D., Thrainsson, H., 1998. Two heads aren’t always better than one. Syntax 1, 37–71] does in fact predict much of the observed variation, despite recent claims to the contrary. According to this theory, this variation is intimately related to morphological differences between the languages (or dialects): Languages with clearly separable agreement and tense morphology (Icelandic, for instance) have separate agreement and tense projections and this makes verb movement (of the V-to-I type) obligatory because of the nature of checking operations. This does not mean, however, that V-to-I can only occur in languages with rich verbal morphology. The theory also predicts that a certain kind of adverbial modification can trigger V-to-I, but this has typically been overlooked in the literature: If adverbs are specifiers of separate functional projections, then V-to-I is necessary, whereas it is not if adverbs are adjuncts. This means that adverbial adjunction must be the rule in Scandinavian and adverbs in separate functional projections an exception (optionally available for some sentence adverbs but not all in Regional Northern Norwegian, for instance).

This paper reports four experiments that test whether persistent problems of second-language (L2) learners with L2 inflection, such as case or subject–verb agreement, are the consequence of age-related grammatical impairment of L2 morphosyntax or differences in processing efficiency between natives and non-natives. Fifty-nine L1 English, Dutch and Russian advanced to near-native L2 speakers of German were tested on their ultimate attainment of case marking, subject–verb agreement and gender concord. Different off-line and on-line tasks were employed. Cross-linguistic and cross-experiment comparisons of native and non-native performance show that native-like ultimate attainment of L2 inflection is possible for postpubescent learners in L2 grammar and L2 processing. Non-target-like L2 inflection is systematically related to L1 transfer and limitations in L2 processing efficiency. In conjunction, these findings argue against a critical period for morphosyntax in L2 acquisition; rather, they suggest that non-native and native grammars and processing systems are fundamentally identical, with L2 systems being computationally less efficient due to L1 influence.

This paper proposes a new analysis of Spanish verb forms. It works from a general assumption in the Spanish literature that the Spanish verb consists of a root followed by a verb theme marker; these are followed by tense/aspect/mood and finally by person/number. While this structure is often transparent, there are cases where it is unclear, leading to a variety of previous analyses. Problematic cases are accounted for here without violating this order of morphemes. The proposal is compared with the classical analysis by Harris and with a recent one by Arregi.

Jouitteau's 2005 generalization that absolute V1 is not allowed introduces an interesting perspective on the possible structural relatedness between V1 and V2 grammatical systems. This article examines whether Celtic and Germanic display additional convergences beyond those that involve the left periphery. The verb–subject order in finite declarative clauses is one of them. It is shown that, once certain assumptions are made concerning the syntax of formal features (they “fission” when unsatisfied) and the featural endowment of Tense (the phi-requirement should carefully be distinguished from the EPP-requirement), some of the characteristics of the Celtic and Germanic peripheries straightforwardly follow from the properties of the inflectional domains. As far as the latter are concerned, three dimensions of variation are isolated, which decisively contribute to shaping clausal structure and determining word order: the inflectional richness of verbal forms, whether tense has a featural/morphemic representation within vP, and the presence/absence/optionality of T's [EPP].

The aim of this paper is to provide an account of na-clauses in Modern Greek, by focusing on the properties of the ‘mood’ particle na and its relation to the subject it embeds. It is argued that na has the effect of subsuming the lexical realization of the subject, mediating (along with verbal agreement) its interpretation. In control contexts, clause-union takes place, leading to the identification of the embedded (unrealized) subject with one of the matrix arguments. In non-control contexts, where no clause-union is at stake, free reference is the available option. It is also argued that the same approach can extend to to-complements in English. Any differences between the two grammars are further related to the presence vs. absence of verbal agreement.

In this paper, we propose a lexicalist account of constructions which initially appear to be nominalisations of phrasal structures. We argue that lexical nominalisation of verbs preserves their form-selectional properties, creating nouns which select for dependents in exactly the way that the source verbs do. These properties hold in the ‘sub-phrasal’ syntax – essentially a domain of complex predicate formation – but do not hold in the regular phrasal syntax, where the form-selectional properties of verbs and nouns are disjoint. The overall view of grammar is one in which all morphological derivation and inflection is in the lexicon, interacting with a syntax in which there is a crucial distinction between words – items from the lexicon which do not project phrases, but create sub-phrasal constituents – and phrases, which are standardly projected from words.

This paper investigates the theoretical implications of the negative inflected forms of verbs in Japanese. Japanese verbs and adjectives overlap considerably in their morphology and syntax, but there is strong evidence that they constitute distinct morpholexical classes. Verbs and adjectives realize negative polarity, but in verbs this is realized by synthetic suffixal forms (inflections), while in adjectives it is achieved by a periphrastic construction with the negative form of the copula. However, the negative form of the plain style verb (as opposed to the polite style verb in -masu) inflects in the same way as the positive polarity adjective. Thus, in the negative conjugation the verb is morphologically an adjective. This is not a case of heteroclisis or deponency because it affects all verbs and all plain negative forms. It is also not an instance of a participle, since the primary function of the negative forms is as the head of the predication as with any other verb form, while a participle is canonically a verb transposed to an adjective so that it can function as an attributive modifier to a noun. We therefore have an instance in which a word of one morphological category shifts to another category for a large portion of its paradigm, without shifting in syntactic category. Such cases represent an important, though hardly studied, instance of morphosyntactic mismatch.

In this paper, we argue against the emergentist position that functional morphology can be acquired by induction from properties of the target-language surface phonetic string. Two experiments are presented, which examine oral production data from Mandarin speakers at intermediate levels of proficiency in English. The first experiment is on the acquisition of English past tense morphology, the second on the acquisition of English articles. In both cases, learners are quite successful at supplying the appropriate morphology. A detailed examination of the phonetics of learners’ outputs shows that many of their productions cannot be derived from the surface strings available in the ambient input. In particular, the L2 learners produce tense and articles in ways which are not found in either the L1 or the L2, namely, with fortis release on tense inflection and stress on articles. We maintain that options made available by Universal Grammar, together with transfer of prosodic representations from the L1, best account for such behaviour.

In Modern Greek nominals, the accent is sometimes faithful to the properties of stems, and sometimes to those of affixes. Several facts challenge solutions based on stipulating the ranking of faithfulness constraints in OT, however. These include the fact that accentual dominance of affixes correlates with their semantic markedness (marked Case; marked number); the fact that neighboring cells within a paradigm tend to be accentually uniform; and the fact that control of accent shifts from affix to stem when stems are followed by Class Markers. These facts are argued to motivate an approach in which faithfulness effects and their relative magnitudes are not primitive notions, but are rather derived in certain specific ways that take account of both the internal structure of representations and their overall distance in space.

In spite of the substantial literature dedicated to it, the status of French subject clitics is still an unresolved issue within morpho-syntactic theory. Two main analyses have been proposed and defended over the past three decades: one advocating that French subject clitics are syntactic arguments bearing a theta-role and the other viewing such clitics as inflectional morphemes on the verb. This paper demonstrates that the empirical basis motivating the morphological analysis of French subject clitics is much narrower than has been assumed in the literature and shows that the implementation of such an analysis faces numerous theoretical and empirical difficulties. It concludes that the limited similarities between the behaviour of French subject clitics and that of morphemes should be treated as accidental rather than as decisive factors in favour of a morphological analysis. Under a derivational approach to grammar, the syntactic analysis appears to be the only viable one.

Adjectives modifying singular human nouns in the Ganja dialect of Balanta bear one of two prefixes: u- and a- . It is argued that these prefixes represent two different types of inflection, agreement ( u- ) and adverbial marking ( a- ). Alternative analyses based on stage-level vs. individual-level interpretations of adjectives and the individual-event distinction are shown to be insufficient. Finally, the set of adjectival agreement prefixes occurring in other Balanta dialects is considered, and a diachronic explanation based on Kurylowicz's laws of analogical change and Lass’ theory of linguistic exaptation is proposed [J Linguistics 26 (1990) 79–102].

Starting out from observational gaps in the Norwegian verb inflection, a constraint is suggested on non-affixal inflection. A previous suggestion is shown not to suffice. A revised version, the Stem Alternation Constraint, is suggested. This constraint seems empirically adequate when confronted with examples from German, Italian and Polish. Theoretical implications are considered. The constraint indicates that paradigms are organised, that Bybee's notion of autonomy is useful, and that non-affixal inflection is not a ragbag of isolated facts without theoretical interest.

This paper examines the object fronting construction of American sign language (Liddell, S., 1980. American Sign Language Syntax. Mouton, The Hague; Matsuoka, K., 1997. Verb missing in American sign language. Lingua 103, 127–149), and its relation to verb raising and aspectual inflection. In this construction, word order, canonically S–V–O, is modified to O–S–V. Matsuoka argues persuasively for a relationship between object fronting and overt verb raising, driven by the presence of an affixal verbal inflection [Lasnik, H. 1995b. Verbal morphology: syntactic structures meets the minimalist program. In: Campos, H. and Kempchinsky, P. (Eds.), Evolution and Revolution in Linguistic Theory. Georgetown University Press, Washington, DC, pp. 251–275]. Thus, her analysis reduces ASL object fronting to an instance of ‘object shift’ as occurs in the Scandinavian language family. However, evidence is provided here, from the distributions of adverbs and modals, and from the interaction of object fronting with question formation and raising contexts, which indicates that ASL object fronting targets a position high in the clause. A new model of ASL object fronting is developed against the backdrop of Matsuoka and the Lasnik theory of verbal morphology. This new framework has much in common with Matsuoka's account, while attaining greater empirical coverage. In this framework, as in Matsuoka, overt verb raising is triggered by the presence of an aspectual head (Asp). However, I argue that AspP is part of an articulated COMP layer [Rizzi, L. 1997. The fine structure of the left periphery. In: Haegeman, L. (Ed.), Elements of Grammar: Handbook in Generative Syntax. Kluwer, Dordrecht. pp. 281–337]. Object fronting targets spec-AspP, an A′-position. This model is shown to provide a better account of the distribution of adverbs and modals in object-fronted clauses, as well as explanations of the behavior of object fronted clauses in questions and in embedded contexts. Lasnik's model of verbal morphology is extended such that verbs in (some varieties of) ASL allow both derivational and lexical options for aspectual inflection. This mechanism provides for an explanation of the infelicity of object fronted clauses in embedded contexts. The analysis crucially relies on the categorial status of object-fronted clauses and the mechanism of c-selection.

Subject—object distinctions in pronominal systems of languages like Dutch and English are not similar to nominative—accusative oppositions in languages with morphological case, since pronouns do not show the syntactic effects of morphological case. This does not mean that these pronominal distinctions are only relics of earlier stages with a richer inflection. In fact, they do show a fundamental distinction between what is sometimes called head marking (here: agreement) and dependency marking (here: case marking). Consequently, subjects are DPs and objects are extended with a Case Phrase. However, in languages like Dutch and English dependency marking is not morphologically specified, i.e. the head of the Case Phrase is empty. The special property of pronouns is that they are not just nouns, since they only contain functional information. They are organized in a paradigm and correspond to (or spell out) some higher, extended nominal projection. More specifically, Dutch and English object pronouns spell out the Case Phrase, whereas the subject pronouns in these languages correspond to a DP (licensed by agreement). As a result, object pronouns differ in form from subject pronouns. Ordinary nouns, containing lexical information, correspond to N. Since N can be present in subject as well as in object position, ordinary nouns can appear in both types of argument positions. Several peculiar characteristics of Dutch and English pronouns follow from this theory.

We have examined potential dissociations between lexical and grammatical knowledge in language impairment by investigating noun plurals and plural formation inside compounds in Williams Syndrome (WS) subjects. For comparison, we also report results from studies investigating the same linguistic phenomena in children with Specific Language Impairment (SLI). We found that nouns taking regular plurals were always correctly inflected by the WS subjects, whereas they performed much worse on nouns that take irregular plurals. The WS subjects heavily overregularized the regular − s plural, and they even used the plural − s as non-head elements inside compounds, i.e. in circumstances in which unimpaired children (and adults) would typically not use regular plurals. We argue that the excessive use of the regular − s plural in WS results from an impairment of the lexical system and/or its access mechanisms. The dissociation between lexical and grammatical phenomena in WS supports the theoretical distinction between listed lexical entries and a (rule-based) computational system for language.

An attempt is made to gain a general understanding of how referential hierarchy phenomena work morphosyntactically. Languages with and without morphological case turn out to adopt rather different strategies for dealing with the demands that referential hierarchies place on their clause structure. An investigation of Kashmiri (with morphological case) and Northern Tiwa (without case) indicates that at least two factors influence the morphosyntactic manifestation of hierarchy phenomena: a structural association that appears to exist between referential features of clitics and agreement and inflectional structure on the one hand, and the competing demands on inflectional structure from Nominative agreement vs. high ranking referential features on the other. A consideration of these two factors contributes to an understanding of why certain hierarchy phenomena take the form they do, why different strategies are used in different types of languages to avoid violations of the hierarchy, and the role that morphological case marking can play in hierarchy phenomena.

This paper argues that predication theory is instrumental in capturing the distribution of expletives. (i) In interaction with well-known verb movement parameters, predication theory explains why not all languages have transitive expletive constructions. Conditions on the assignment of external thematic roles have the effect that the functional domain of a clause is only large enough to host both an expletive and a subject if the verb undergoes verb second as well as V to I. (ii) On the assumption that sufficiently rich inflection is argumental, predication theory also provides a trigger for V to I. It is argued that, after verb movement, I can be interpreted as VP's subject. This explains why rich inflection blocks expletive insertion within the verb's predicational domain. (iii) Finally, predication theory forces an analysis of NP raising as predicate formation. As a result, NP raising is optional in many languages, even if no expletive is inserted.

Prasada and Pinker's (1993) subjects provided past tense forms of nonce verbs. The subject's willingness to provide irregular past tense forms correlated with the verb's phonological similarity to existing irregular English verbs. However, there was no correlation between the number of nonce verbs assigned regular inflection, and the verb's similarity to existing regular verbs. According to the dual-route model, this is expected since irregular items are stored in associative memory, while regular items take an allomorph of -ed by rule. A singleroute connectionist simulation failed to duplicate the subject's behavior on regular verbs. Two instance-based models were applied to the data: Analogical Modeling of Language and the Tilburg Memory Based Learner. Each model employed a similarity algorithm to determine the behavior of all regular and irregular items. Both models successfully mirrored the subject's responses. Therefore, the data are consistent with an instance-based single-route model of morphology.

The treatment of morphology in Word Grammar has received less attention than syntax and semantics, but the general principles are equally applicable to morphology. We outline a theory of inflectional morphology which uses ideas familiar from the Word-and-Paradigm tradition — Inflection, Lexeme and Stem — in combination with the logic of default inheritance. We apply this theory to a range of different morphological data: agglutinative (Swahili) and fusional, with and without syncretism (English, Welsh). We show that it is possible to analyse each of these types in a natural way without forcing it into an unsuitable mould, and in a discussion of Cree we show how dialect variation can be accommodated. We compare the WG theory with other approaches which are currently popular, especially a-morphous morphology, distributed morphology and network morphology.

The vast expressive power of language is made possible by two principles: the arbitrary sound-meaning pairing underlying words, and the discrete combinatorial system underlying grammar. These principles implicate distinct cognitive mechanisms: associative memory and symbol-manipulating rules. The distinction may be seen in the difference between regular inflection (e.g., walk-walked ), which is productive and open-ended and hence implicates a rule, and irregular inflection (e.g., come-came , which is idiosyncratic and closed and hence implicates individually memorized words. Nonetheless, two very different theories have attempted to collapse the distinction; generative phonology invokes minor rules to generate irregular as well as regular forms, and connectionism invokes a pattern associator memory to store and retrieve regular as well as irregular forms. I present evidence from three disciplines that supports the traditional word/rule distinction, though with an enriched conception of lexical memory with some of the properties of a pattern-associator. Rules, nonetheless, are distinct from pattern-association, because a rule concatenates a suffix to a symbol for verbs, so it does not require access to memorized verbs or their sound patterns, but applies as the ‘default’, whenever memory access fails. I present a dozen such circumstances, including novel, unusual-sounding, and rootless and headless derived words, in which people inflect the words regularly (explaining quirks like flied out, low-lifes , and Walkmans ). A comparison of English to other languages shows that contrary to the connectionist account, default suffixation is not due to numerous regular words reinforcing a pattern in associative memory, but to a memory-independent, symbol-concatenating mental operation.

A number of scholars in a variety of sign languages (Smith, 1990: Taiwan Sign Language; Bos, 1994: Sign Language of the Netherlands; Engberg-Pedersen, 1993: Danish Sign Language) have recently found auxiliary or auxiliary-like structures which serve to carry agreement, usually when the main verb does not. I have found a similar phenomenon in Japanese Sign Language (JSL). Other papers in this issue address the linguistic status of agreement and the auxiliaries to which it attaches. I would like to approach the problem from a somewhat orthogonal perspective. Although I have suggested elsewhere that verb agreement is derived historically from clitic pronouns, I assume the conclusion of the other papers that agreement is indeed an inflection and not a clitic pronoun. I turn the question around and address the issue of whether some indexical pronouns in JSL are in fact manifestations of agreement as argued by Torigoe (1994) (I argue in the negative), suggest the beginnings of a theoretically coherent account of the auxiliaries in JSL, and discuss questions of universality in both language and gesture.