From variable to optimal grammar: evidence from language acquisition and language change

Ever since the advent of generative theory, there has been a dispute in linguistics between formalist and functionalist approaches. At stake in this controversy is the autonomy thesis: i.e. the claim that linguistic knowledge is independent from, and irreducible to, the facts of language use. Under the autonomist view, language is both arbitrary and self-contained. Currently, the radically autonomous vision of syntax that is being pursued in Chomskyan circles has lent new urgency to the question whether phonology and syntax can in fact possess autonomy to the same degree. This problem has a long history (e.g. Anderson 1981), but it has lately received new and radical answers, notably through the Representational Hypothesis advanced by Burton-Roberts (2000) and Carr (2000): in the Representational Hypothesis, the language faculty is assumed to be radically autonomous, whereas phonology is functionally based and stands outside UG. This dissertation is concerned with the intermediate grammars and the variability that emerge in the course of linguistic change. It is based on evidence from the acquisition of Hebrew phonology and morphology, and a case study of a historical change affecting the grammar of Modern Hebrew. I propose a unified formal model for intermediate grammars in both language acquisition and (historical) language change, from a synchronic point of view, using the framework of Optimality Theory (Prince and Smolensky 1993). The topics discussed in this work are based on two different studies: (i) the acquisition path of Hebrew prosodic structure; and (ii) the variation involved in stop-fricative alternation in Modern Hebrew, as a result of historical changes. For both topics I suggest an Optimality Theoretic account of the phenomena, their cause, and their consequences. The study of language acquisition is based on two interrelated acquisition paths: (a) of prosodic structure of words out of morphological context; and (b) of prosodic structure within paradigms with regular morphophonological alternation, where the acquisition of Hebrew verb inflectional paradigms are examined. The acquisition path of words out of morphological context shows a pattern of gradual increase in prosodic structure (e.g. the number of syllables), involving a gradual increase in prosodic contrasts (e.g. diverse stress patterns), up until the child’s production is phonologically identical to the target forms. The increase in structure and contrast indicates a gradual transition from unmarked structures with input-output disparities (indicating phonological dominance) to marked structures, lacking input-output disparities (indicating lexical dominance). In the acquisition path of alternating paradigms, not only the child-adult relations are examined, but also the relation between words, where the target language itself exhibits input-output disparities, namely words drawn from alternating paradigms. This acquisition path provides evidence for the child’s restructuring of a lexical representation, as well as evidence for the transition from child-adult relation to the child’s input-output relation. In addition, I show that the phonological account of the acquisition of inflectional paradigms provides evidence for morphological development. One of the interesting findings in this regard is the emergence of morphological knowledge before actual inflectional suffixes are produced by the children. This finding would not be available without examining the interaction between prosodic and morphological factors. Another crucial finding dealt with in my study of acquisition is the scope of variation in the course of acquisition. I found that variation in children’s production is limited to the acquisition of the lexical contrasts existing in the language, and is not found in the acquisition of alternation resulting from phonological restrictions. The model I suggest for change in the course of acquisition predicts and explains this finding, based on evidence that children do not violate universal restrictions respected in the target language. The study of language change deals with spirantization in Modern Hebrew, where the alternation between stops and fricatives involves a great deal of opacity and free variation. The analysis suggested is based on a set of paradigms co- existing in the language and contradicting each other. I show that variation is restricted only to certain types of paradigms, conditioned by certain phonological properties. However, I argue that the co-existence of different types of paradigms indicates a change in the language, where the variable grammar is an intermediate phase, and the invariable paradigms indicate the direction of change. Similar to the acquisition path of single forms, and unlike the acquisition path of alternation within a paradigm, the direction of change in the case of Modern Hebrew spirantization is towards a loss of phonological generalization (i.e. towards lexical dominance). Through the investigation of language acquisition and language change, I examine in this work central issues in phonological theory, such as morphophonemic alternation and prosodic structure, in conditions of change. The phenomena investigated here provide insight to the nature of linguistic change in the following aspects: (i) the distinction between phonological properties that undergo change vs. phonological properties that do not undergo change; (ii) the directionality of change; and (iii) the scope of variation in the course of change. I argue that both language acquisition and language change share similar patterns with respect to these issues. Moreover, the change, its directionality and the variation involved, are accounted for by the same theoretical terms that account for the linguistic knowledge of a single ideal native speaker. I show that Optimality Theory suggests an appropriate theoretical framework for a unified generative analysis of this type.Ph.D