Search Results

This chapter argue that wellformedness constraints are inappropriate computational devices for modeling grammar. It develops a purely derivational theory with minimal theoretical apparatus and no filters or wellformedness constraints. The conceptual arguments will be bolstered by reference to recent work developing alternative approaches to phonological computation from constraint-based ones. The author also makes concrete proposals concerning the nature of phonological acquisition. The idea is to constrain the acquisition task without recourse to innate constraints.

This chapter reanalyses noun classification in the Australian language Dyirbal. While earlier analyses have proposed intricate class assignment principles rooted in conceptual features, we argue that Dyirbal noun classification is sensitive to salient phonological cues and a small core of cross‐linguistically common semantic cues in keeping with other familiar noun classification systems.

This chapter reconsiders the role of the Subset Principle in language acquisition and change, arguing that consideration of true formal optionality enables one to define grammars generating languages that are in inclusion relations. This in turn facilitates an explanation of diachronic changes where absence of sufficiently robust PLD led acquirers to ‘default’ to ‘smaller language’‐generating grammars.

Discusses the requirement that an adequate theory of meaning give a constructive account of the meanings of sentences in natural languages, that is, an account of how the meanings of sentences and their complex parts are understood ultimately on the basis of a finite number of semantical primitives and a finite number of rules for their composition.

This chapter highlights the crucial role that context plays in interaction. It hastens to add that the interplay between context and language use is an empirical issue; that is, the investigation of the social functions of a linguistic phenomenon, include the phenomenon of attitude datives examined in this study, requires a close and systematic analysis of the context in which the phenomenon occurs; thus, the focus on Levantine Arabic. The chapter goes on to delineate the three types of context that need to be taken into account in any sociopragmatic analysis. These are the co-textual context, the situation context, and the sociocultural context. Finally, it underlines the significance of sociopragmatic analyses of the type put forth and employed in this study from a learnability perspective, including foreign language education.

Davidson asks what properties a language must have to be learnable. He criticizes a (then) popular response that models the order of language acquisition on the epistemological priority of the types of expressions learnt; he labels this position the ‘building‐block theory’ (see further Essay 16). He discusses Strawson's critique of Quine's elimination of singular terms and shows how it is likewise premissed on the questionable derivation of claims about language learning from purely a priori considerations. On the positive side, Davidson proposes that a language is learnable by a creature with finite means if the language's number of semantic primitives or undefinables is finite. Using this criterion, he demonstrates that various theories in the philosophy of language introduce an infinite number of semantic primitives into the language and thus make it unlearnable; theories he alleges of this error (1) model quotations on names of expressions (Tarski, Quine; cf Essay 6), (2) analyse belief attributions in terms of linguistic marks (Scheffler, Carnap) or distinct one‐place predicates for each attributed belief (Quine; cf Essay 7), or (3) postulate intensional entities into their overall semantic framework (Frege, Church).

A consideration of `philosophical’ evidence, the view that typical concepts are complex. Connections with questions about analyticity and learnability. The alternative view (`conceptual atomism’)‐that typical concepts are primitives‐is introduced and defended. The importance of distinguishing the semantic issues about conceptual atomism from the epistemological ones.

This chapter discusses the question of whether T-theories explain how it is possible to understand new sentences, or learn an infinite language, as Davidson claimed. It argues against some commentators who claim that for explanatory power we need not require that T-theories are implicitly known or mirror cognitive structures. It is noted, contra Davidson, that the recursive nature of T-theories is not sufficient for explanatory power, since humans can work out only what is computationally tractable, and recursiveness by itself allows for intractable computational complexity. Finally, it considers the complexity of T-theories, transformed into term rewriting systems, and finds that the complexity of such systems is indeed tractable. Therefore, Davidson’s claim stands, even though a further condition had to be met.

This chapter comments on the discussion in Chapter 9. It largely agrees with Fitch's proposals and explores the extent to which the learnability of communicative systems informs the emergence of language- and music-like systems. In effect, language and music is conceived as culturally transmitted complex systems of which the development can be explored and evaluated in terms of the concept of fitness; this operationalization renders the exploration of language (and music) evolution computationally tractable and conditions the chapter's conclusions.

A language must be learnable by new generations of speakers. Contributing to the learnability of language and its semantic and syntactic categories are the skewed frequencies of its elements. Categories whose members are equally frequent would be unlearnable in principle.

A completely new conceptualization of the indirect negative evidence business in language acquisition, especially in syntax. Instead of thinking about retreating from over-generalization, a derivative application of the Tolerance Principle ensures that the child is much more careful before generalizing. Shows how the learner may acquire that adjectives such as “asleep” do not allow attributive in NPs (“*the asleep cat”), and how to resolve Baker’s classic problem of dative construction acquisition (“*I donated the museum a painting”). A critique of previous proposals, including Bayesian models of inference, is also included.

In 1990, Y.-C. Chien and Kenneth Wexler (1990) proposed that acquisition delays in children occur with coreference rather than the binding theory in general, in agreement with the notion that variable binding and coreference are governed by different types of linguistic conditions. A 1983 study suggested that the condition on coreference belongs to pragmatics and involves an inference based on knowledge of grammar, meaning, and appropriateness to context. This chapter explores the nature of coreference in human language and the role of processing in language acquisition by using data from a research program that focuses on adult acceptability judgments on possible sentences as well as child language acquisition. It considers intrasentential pronominal anaphora and comments on Rosalind Thornton and Kenneth Wexler’s arguments against the processing account, especially with regards condition C. It also discusses the pragmatic and the processing accounts of delay in the acquisition of coreference in the context of learnability and chance performance.

Research is reviewed that addresses itself to human language learning by developing precise, mechanistic models that are capable in principle of acquiring languages on the basis of exposure to linguistic data. Such research includes theorems on language learnability from mathematical linguistics, computer models of language acquisition from cognitive simulation and artificial intelligence, and models of transformational grammar acquisition from theoretical linguistics. It is argued that such research bears strongly on major issues in developmental psycholinguistics, in particular, nativism and empiricism, the role of semantics and pragmatics in language learning, cognitive development, and the importance of the simplified speech addressed to children.

The Sign Theory of Language leads to new insights in our understanding of a substantial sample of key constructions in decades of linguistic analysis, such as structure dependence, binding theory, existential constructions, raising, and long-distance dependencies, as well as considerations of learnability of syntax. If children can learn the meanings of signs, sometimes quite infrequent words, they should be able to squeeze through easily enough for the ubiquitous combinatorial signs they encounter. The Primary Linguistic Data is saturated with information germane to acquiring the combinatorial signs of a given language. Given the actual richness of the Primary Linguistic Data about the relevant properties, nativist speculations about a genotypically specified Universal Grammar are patently otiose. The canalization of language acquisition is not done by contingent Universal Grammar constraints on the functioning of the formal apparatus, but is due to the substantial properties of the signs.

This chapter surveys the current debate on the morphome, drawing attention to underexplored theoretical possibilities and underexploited empirical tools. Three separate claims made by proponents of the morphome are outlined: that there exist morphological patterns mapping arbitrary sets of exponenda onto arbitrary sets of exponents; that such patterns do not suffer from a learnability disadvantage; and that all patterns of exponence are mediated by purely morphological categories belonging to an autonomous level of linguistic representation. The chapter reviews the problems caused by the lack of positive criteria for morphomicity and disagreements over the application of negative criteria. It presents arguments for a learning bias in favour of realization patterns involving natural classes, and for greater use of wug-tests and artificial grammar learning experiments. Competing morphological theories turn out to exhibit a surprising amount of empirical overlap; their implications for the learnability of morphomic patterns are less straightforward than usually assumed.

This chapter begins by presenting the axiom that a game should be “easy to learn and hard to master,” a concept which earned the title “Bushnell’s Law”, based on Atari founder Nolan Bushnell. This is a recurring notion that “holds” the key in making a powerful addictive game that people want to play over and over again. Pong, which resembles tennis, is “easy to learn” because it assumes the basic rules and function of a familiar cultural practice. Familiarity is the primary property of the game, not learnability. This example suggests that habituation build on prior convention. Games can also produce their own conventions, which become familiar enough to be adopted later in the same way that Pong adopts table tennis.

This chapter stresses that the central problem of parametric theory is Plato’s problem (how do humans learn a language?), whereas what we may call Greenberg’s problem (what is the shape and degree of linguistic variability?) is derivative. The chapter reviews several macro- and microparameters and shows that they do not satisfy certain learnability conditions, whereby they are not plausible learning paths or plausible elements of UG. It explores a parametric approach, which conjoins a mechanism of data analysis parametrically defined (the Chunking Procedure), arguably used by the LAD to attain a morphological analysis of its PLD, and bootstrapping mechanisms that use the attained morphological analysis to specify more abstract syntactic properties of the target language, the clustering properties which standard parameters range over.

This chapter reviews Fodor’s contribution to the epic Chomsky-Piaget Royaumont debate. The issue that was under discussion was a familiar one, namely, what psychological processes underlie concept learning. Piaget thought concept learning involved the formation and confirmation of hypotheses that a learner generates through the construction and organization of stimuli gathered from the environment, and modifying them when they proved to be inconsistent. However, Fodor pointed out a fundamental flaw in this theory: it is silent about the origin of the concepts used in generating the hypotheses. Fodor argued that in order for these hypotheses to be tested, let alone generated, they needed to have been readily available to the learner, suggesting that all primitive concepts are innate, and that concept acquisition relies on the process of triggering these concepts that are innately available to the learner, and not through construction by means of progressive guesses and trial-and-error.