Search

Human language is often assumed to make "infinite use of finite means" - that is, to generate an infinite number of possible utterances from a finite number of building blocks. From an acquisition perspective, this assumed property of language is interesting because learners must acquire their languages from a finite number of examples. To acquire an infinite language, learners must therefore generalize beyond the finite bounds of the linguistic data they have observed. In this work, we use an artificial language learning experiment to investigate whether people generalize in this way. We train participants on sequences from a simple grammar featuring center embedding, where the training sequences have at most two levels of embedding, and then evaluate whether participants accept sequences of a greater depth of embedding. We find that, when participants learn the pattern for sequences of the sizes they have observed, they also extrapolate it to sequences with a greater depth of embedding. These results support the hypothesis that the learning biases of humans favor languages with an infinite generative capacity.

We present in this article corpus analyses, two experiments, and a preliminary English-French comparison on children's acquisition of "wh"-in-situ. Our examination of 10,000 "wh"-questions from CHILDES reveals that the reported empirical picture of "wh"-question acquisition in English is incomplete: A type of "wh"-in-situ, probe questions (PQs), has been left out from most discussions despite its presence in child-directed speech. Unlike "wh"-in-situ echo questions (EQs), PQs are used to request new information, and parents frequently use PQs and fronted information-seeking questions in alternation. The fact that PQs share the pragmatic space with fronted "wh"-questions while involving fewer syntactic operations and exhibiting lower input frequency allows us to test both structure-based and frequency-based theories of syntax acquisition. Our comprehension task with 3;06-5;06-year-olds confirms that children accept and understand PQs as information seeking. On the other hand, results from a production task show a strong avoidance of "wh"-in-situ, which is in line with reported elicited data from French-speaking children. We reason that a structural economy-based approach alone is not sufficient to account for children's disfavor of "wh"-in-situ. Depending on the input frequency and consistency, as well as the number of variants licensed by the grammar of a given language, children may treat part of the input as uninformative and initially only learn from higher-frequent, more regularized input. Their intake is thus selective.

A surprising comprehension-production asymmetry in subject-verb (SV) agreement acquisition has been suggested in the literature, and recent research indicates that task-specific as well as language-specific features may contribute to this apparent asymmetry across languages. The present study investigates when during development children acquiring Mexican Spanish gain competence with 3rd-person SV agreement, testing production as well as comprehension in the same children aged between 3;6 and 5;7 years, and whether comprehension of SV agreement is modulated by the sentential position of the verb (i.e., medial vs. final position). Accuracy and sensitivity analyses show that comprehension performance correlates with SV agreement production abilities, and that comprehension of singular and plural third-person forms is not influenced by the sentential position of the agreement morpheme. Issues of the appropriate outcome measure and the role of structural familiarity in the development of abstract representations are discussed.

We have developed principles integrating connectionist and symbolic computation by establishing mathematical relationships between two levels of description of a single computational system: at the lower level, the system is formallydescribed in terms of highly distributed patternsof activity over connectionist units, and the dynamics of these units; at the higher level, thesame system is formally described by symbolicstructures and symbol manipulation. In this pa-per, w e propose a specific treatment of recur-sion where complex symbolic operations on re-cursive structures are mapped to massively par-allel manipulation of distributed representationsin a connectionist network.

We describe harmonic grammar, a connectionist-based approach to formal theories of linguistic well-formedness. The general approach can be applied to various kinds of linguistic well-formedness, e.g., phonological and syntactic. Here, we address a syntactic problem: unaccusativity. Harmonic grammar is a two-lcvcl theory, involving a distributed, lower level connectionist network whose relevant aggregate computational behavior is described by a local, higher level network. The central hypothesis is that the connectionist well-formedness measure called "harmony"^ can be used to model linguistic well-formedness; what is crucial about the relation between the lower and higher level networks is that there is a harmony-preserving mapping between them: they are isoharmonic (at least approximately). A companion paper (Legendre, Miyata, & Smolensky, 1990; henceforth "LMS2") describes the theoretical basis for the two level approach, starting from general connectionist principles. In this paper, we discuss the problem of unaccusativity, give a high level characterization of harmonic syntax, and present a higher level network to account for unaccusativity data in French. W e interpret this network as a fragment of the grammar and lexicon of French expressed in "soft rules." Of the 760 sentence types represented in our data, the network correctly predicts the acceptability in all but two cases. This coverage of real, problematic syntactic data greatly exceeds that of any other formal account of unaccusativity of which we are aware.

In this paper, we derive the formalism of harmonic grammar, a connectionist-based theory of linguistic well formedness. Harmonic grammar is a two-level theory, involving a low level connectionist network using a particular kind of distributed representation, and a second, higher level network that uses local representations and which approximately and incompletely describes the aggregate computational behavior of the lower level network. The central hypothesis is that the connectionist well-formedness measure Harmon)^ can be used to model linguistic well-formedness; what is crucial about the relation between the lower and higher level networks is that there is a harmony-preserving mapping between them: they are isoharmonic (at least approximately). In a companion paper (Legendre, Miyata, & Smolensky, 1990; henceforth, "LMSi"), we apply harmonic grammar to a syntactic problem, unaccusativity, and show that the resulting network is capable of a degree of coverage of difficult data that is unparallelled by symbolic approaches of which we are aware: of the 760 sentence types represented in our data, the network correctly predicts the acceptability in all but two cases. In the present paper, we describe the theoretical basis for the two level approach, illustrating the general theory through the derivation from first principles of the unaccusativity network of LMSj.

How recurrent typological patterns, or universals, emerge from the extensive diversity found across the world's languages constitutes a central question for linguistics and cognitive science. Recent challenges to a fundamental assumption of generative linguistics--that universal properties of the human language acquisition faculty constrain the types of grammatical systems which can occur--suggest the need for new types of empirical evidence connecting typology to biases of learners. Using an artificial language learning paradigm in which adult subjects are exposed to a mix of grammatical systems (similar to a period of linguistic change), we show that learners' biases mirror a word-order universal, first proposed by Joseph Greenberg, which constrains typological patterns of adjective, numeral, and noun ordering. We briefly summarize the results of a probabilistic model of the hypothesized biases and their effect on learning, and discuss the broader implications of the results for current theories of the origins of cross-linguistic word-order preferences. (Contains 7 figures and 8 tables.)

Two comprehension experiments were conducted to investigate whether young French-learning children (N = 76) are able to use a single number cue in subject-verb agreement contexts and match a visually dynamic scene with a corresponding verbal stimulus. Results from both preferential looking and pointing demonstrated significant comprehension in 30-month-olds with no preference for either singular or plural. These results challenge previous claims made on the basis of English and Spanish that comprehension of subject-verb agreement expressed as a bound morpheme is late, around 5 years of age (V. A. Johnson, J. G. de Villiers, & H. N. Seymour, 2005; A.-T. Perez-Leroux, 2005). Properties of the adult input were also analyzed. Possible implications for theories of syntactic acquisition are discussed.

This article reports on a series of 5 analyses of spontaneous production of verbal inflection (tense and person-number agreement) by 2-year-olds acquiring French as a native language. A formal analysis of the qualitative and quantitative results is developed using the unique resources of Optimality Theory (OT; Prince & Smolensky, 2004). It is argued that acquisition of morphosyntax proceeds via overlapping grammars (rather than through abrupt changes), which OT formalizes in terms of partial rather than total constraint rankings. Initially, economy of structure constraints take priority over faithfulness constraints that demand faithful expression of a speaker's intent, resulting in child production of tense that is comparable in level to that of child-directed speech. Using the independent Predominant Length of Utterance measure of syntactic development proposed in Vainikka, Legendre, and Todorova (1999), production of agreement is shown first to lag behind tense then to compete with tense at an intermediate stage of development. As the child's development progresses, faithfulness constraints become more dominant, and the overall production of tense and agreement becomes adult-like.

Reanalyzes production data from three French children to make two basic points. Shows that tense and agreement inflection follow independent courses of acquisition (in child French). Using a mechanism of grammatical development based on partial rankings of constraints, analysis successfully models over three stages the frequency with which children use tensed, agreeing, and nonfinite verbs. ( Author/VWL)

A surprising comprehension-production asymmetry in subject-verb (SV) agreement acquisition has been suggested in the literature, and recent research indicates that task-specific as well as language-specific features may contribute to this apparent asymmetry across languages. The present study investigates when during development children acquiring Mexican Spanish gain competence with 3rd-person SV agreement, testing production as well as comprehension in the same children aged between 3;6 and 5;7 years, and whether comprehension of SV agreement is modulated by the sentential position of the verb (i.e., medial vs. final position). Accuracy and sensitivity analyses show that comprehension performance correlates with SV agreement production abilities, and that comprehension of singular and plural third-person forms is not influenced by the sentential position of the agreement morpheme. Issues of the appropriate outcome measure and the role of structural familiarity in the development of abstract representations are discussed. [ABSTRACT FROM AUTHOR]

We present elicited production data reflecting cross-linguistic interference effects in the English wh-questions of Spanish-English bilingual children to provide a proof-of-concept for a proposed new formal analysis of such effects across cross-linguistic influence phenomena. The observed interference effects are interpreted as evidence for the Strong Integration hypothesis of bilingual grammatical architecture, in the context of independently documented facilitation and interference effects in a range of bilingual acquisition contexts. Building on an existing Optimality-Theoretic (OT) model of monolingual acquisition and a specific account of the adult grammar of wh-structures across dialects of Spanish, we propose that the individual patterns documented, in particular the sensitivity in child English to distinctions made in Spanish dialects on the basis of an argument/adjunct contrast, find a straightforward explanation in the OT model of acquisition as adapted to bilingual situations. The generalizability of the model as well as effects of exposure and dominance are discussed. [ABSTRACT FROM AUTHOR]