Does knowledge of language consist of mentally-represented rules? Rumelhart and McClelland have described a connectionist (parallel distributed processing) model of the acquisition of the past tense in English which successfully maps many stems onto their past tense forms, both regular (walk/walked) and irregular (go/went), and which mimics some of the errors and sequences of development of children. Yet the model contains no explicit rules, only a set of neuron-style units which stand for trigrams of phonetic features of the stem, a set of units which stand for trigrams of phonetic features of the past form, and an array of connections between the two sets of units whose strengths are modified during learning. Rumelhart and McClelland conclude that linguistic rules may be merely convenient approximate fictions and that the real causal processes in language use and acquisition must be characterized as the transfer of activation levels among units and the modification of the weights of their connections. We analyze both the linguistic and the developmental assumptions of the model in detail and discover that (1) it cannot represent certain words, (2) it cannot learn many rules, (3) it can learn rules found in no human language, (4) it cannot explain morphological and phonological regularities, (5) it cannot explain the differences between irregular and regular forms, (6) it fails at its assigned task of mastering the past tense of English, (7) it gives an incorrect explanation for two developmental phenomena: stages of overregularization of irregular forms such as bringed, and the appearance of doubly-marked forms such as ated, and (8) it gives accounts of two others (infrequent overregularization of verbs ending in t/d, and the order of acquisition of different irregula...

late 1950's. It's basic premises are that phonological structure reflects the linguistic competence of the individual native speaker to compute a phonetic representation for the potentially infinite number of sentences generated by the syntactic component of the grammar and that this competence can be investigated in a serious scientific fashion. The generative point of view has become dominant in the field of linguistics and has had varying degrees of influence on other cognitive sciences. This entry surveys the development of the generative approach over three fifteen-year segments and concludes with current research trajectories. 1. SPE: 1960- 1975 The early work of Chomsky and Halle both embraces and rejects various aspects of the two major schools of American Structural Linguistics inaugurated by Edward Sapir 1 (1884-1942) and Leonard Bloomfield (1887-1949). Sapir's "Item and Process " model posits an abstract Phonological Representation that is converted to a Phonetic Representation by processes that delete, add, and change sounds. Sapir stressed the

This thesis is a qualitative and quantitative study of the changes that occurred in the phrase structure and system of pronominal clitics in medieval and renaissance Spanish, with the goal of explaining the basic differences between the syntactic properties of clitics in Old Spanish and their counterparts in the various dialects of modern Spanish. Specifically, I argue that these differences are explainable if we classify OSp clitics as Second Position (2P) clitics, in contrast to their modern counterparts. 2P clitics are treated here as prosodically deficient phrasal constituents that appear displaced from their canonical positions as internal arguments of the verb and are adjoined to a phrasal projection at the left edge of the clause (IP). The elements encompassed under the pre-theoretical notion clitic in modern Spanish, however, are not linked to an argument ...

We argue for a surficial pronunciation model: a model without underlying forms. The surficial model outperforms a traditional generative model by a significant margin on conversational speech (Switchboard) as well as on read speech (TIMIT). Our results suggest that the true mapping from underlying forms to surface forms is too complex to be accurately modeled using current techniques, and that we would be best served to model the surface forms directly.

This article presents detailed complexity analysis of the encoding and decoding problems for the segmental phonology. The segmental model is a generative theory of phonological dependencies. As such, it poses two computational problems. The first computational problem is the problem of actually encoding the dependencies of a given phonology into a given representation of phonological knowledge (the encoding problem). The second computational problem is the problem of deciding whether a given phonological representation encodes the dependencies of a given phonology (the decoding problem). We begin by proving that the encoding and decoding problems are both undecidable in the segmental model. Next, we motivate a simplified segmental model, that more accurately models the phonological dependencies actually proposed in the phonological literature. The simplified segmental model is a more restricted and more natural reformalization of the segmental model. To conclude, we prove that the enco...

. Our aim is to explain a general theory of inductive reasoning which is close enough to the concerns of language studies. In this set-up, the optimal prediction rate is assigned to the hypothesis considered most likely by a prior-free form of Bayesian inference. In terms of practical applications a most attractive form of this approach is embodied by the so-called minimum description length (MDL) principle. There, the most likely hypothesis is the one which minimizes the sum of the length of the description of the hypothesis and the length of the description of the data relative to the hypothesis. This theory is solidly based on a provably ideal method of inference using Kolmogorov complexity. We give references to several applications. Similar approaches should work for computational learning of features of language. The genesis of this work is not rooted in traditional approaches to artificial intelligence (AI), but rather in new exciting general learning theories which have develop...

Phonological notations need major overhaul. I would like to suggest how this can be succesfully done. My argument has four parts. First, I would like to make a simple distinction — well known in semiotics — between what I will call notations as symbols and notations as models. Notations which merely symbolize the thing they stand for are in all respects arbitrary in their form except insofar as they must be differentiated from other symbols used in the same context. Notations which model the thing they stand for are, in at least one respect, nonarbitrary in their form, that is. in some way they are or purport to be isomorphic with the thing they represent. A few simple examples of these two types of notations are given in Figure 1. In the case of the notation under the heading ‘symbols’, there is no recognizable part of the symbols that is isomorphic with any part of the entities they stand for. Only in Playboy cartoons and for humorous purposes is any ' isomorphism suggested to exist between the scientific symbols for 'male ' and 'female ' and actual males and females. Likewise the graphic symbol '7 ' does not have seven distinct parts, although the tally mark for '7', on the other side, does. Words, of course, are well known as arbitrary symbols for the things or concepts they stand for. Thus, we do not find it inappropriate that the word 'big ' is so small nor that the word 'microscopic ' is relatively large. Most mathematical notations are symbols in this sense. However, as for the notations under the heading 'models', there are obvious respects in which they are isomorphic with the entities they represent. Architects ' floor plans are isomorphic with the geometry and proportions of the building they represent. The electrical circuit diagram is isomorphic with the topological arrangement and connectedness of the various components. And so on.

revisions which make it slightly different from the version

Abstract not found

The effects of frequency and neighbourhood density on adult speakers ’ productivity with Polish case inflections: An empirical test of usage-based approaches to morphology