Rules vs analogy in English past tenses a computational

Rules Vs Analogy In English Past Tenses A Computational-PDF Download

  • Date:29 Jul 2020
  • Views:1
  • Downloads:0
  • Pages:43
  • Size:516.37 KB

Share Pdf : Rules Vs Analogy In English Past Tenses A Computational

Download and Preview : Rules Vs Analogy In English Past Tenses A Computational


Report CopyRight/DMCA Form For : Rules Vs Analogy In English Past Tenses A Computational


Transcription:

120 A Albright B Hayes Cognition 90 2003 119 161, 1 Introduction rules in regular and irregular morphology. What is the mental mechanism that underlies a native speaker s capacity to produce. novel words and sentences Researchers working within generative linguistics have. commonly assumed that speakers acquire abstract knowledge about possible structures of. their language and represent it mentally as rules An alternative view however is that new. forms are generated solely by analogy and that the clean categorical effects described by. rules are an illusion which vanishes under a more fine grained gradient approach to the. data Bybee 1985 2001 Rumelhart McClelland 1986 Skousen 1989. The debate over rules and analogy has been most intense in the domain of inflectional. morphology In this area a compromise position has emerged the dual mechanism. approach see e g Clahsen 1999 Pinker 1999a Pinker Prince 1988 1994 adopts a. limited set of rules to handle regular forms in most cases just one extremely general. default rule while employing an analogical mechanism to handle irregular forms. There are two motivating assumptions behind this approach 1 that regular default. processes are clean and categorical while irregular processes exhibit gradience and are. sensitive to similarity and 2 that categorical processes are a diagnostic for rules. while gradient processes must be modeled only by analogy. Our goal in this paper is to challenge both of these assumptions and to argue instead for. a model of morphology that makes use of multiple stochastic rules We present data from. two new experiments on English past tense formation showing that regular processes are. no more clean and categorical than irregular processes These results run contrary to a. number of previous findings in the literature e g Prasada Pinker 1993 and are. incompatible with the claim that regular and irregular processes are handled by. qualitatively different mechanisms We then consider what the best account of these. results might be We contrast the predictions of a purely analogical model against those of. a model that employs many rules including multiple rules for the same morphological. process and that includes detailed probabilistic knowledge about the reliability of rules in. different phonological environments We find that in almost every respect the rule based. model is a more accurate account of how novel words are inflected. Our strategy in testing the multiple rule approach is inspired by a variety of previous. efforts in this area We begin in Section 2 by presenting a computational implementation. of our model For purposes of comparison we also describe an implemented analogical. model based on Nosofsky 1990 and Nakisa Plunkett and Hahn 2001 Our use of. implemented systems follows a view brought to the debate by connectionists namely. that simulations are the most stringent test of a model s predictions Daugherty. Seidenberg 1994 MacWhinney Leinbach 1991 Rumelhart McClelland 1986. We then present data in Section 3 from two new nonce probe wug test Berko 1958. experiments on English past tenses allowing us to test directly as Prasada and Pinker. 1993 did whether the models can generalize to new items in the same way as humans. Finally in Section 4 we compare the performance of the rule based and analogical models. in capturing various aspects of the experimental data under the view that comparing. differences in how competing models perform on the same task can be a revealing. diagnostic of larger conceptual problems Ling Marinov 1993 Nakisa et al 2001. A Albright B Hayes Cognition 90 2003 119 161 121,2 1 Rules and analogy. To begin we lay out what we consider the essential properties of a rule based or. analogical approach The use of these terms varies a great deal and the discussion that. follows depends on having a clear interpretation of these concepts. Consider a simple example In three wug testing experiments Bybee Moder 1983. Prasada Pinker 1993 and the present study participants have found splung spl fairly. acceptable as a past tense for spling splI This is undoubtedly related to the fact that English. has a number of existing verbs whose past tenses are formed in the same way swing string. wring sting sling fling and cling In an analogical approach these words play a direct role in. determining behavior on novel items splung is acceptable because spling is phonologically. similar to many of the members of this set cf Nakisa et al 2001 p 201 In the present case. the similarity apparently involves ending with the sequence I and perhaps also in. containing a preceding liquid s consonant cluster and so on Bybee Moder 1983. Under a rule based approach on the other hand the influence of existing words is. mediated by rules that are generalized over the data in order to locate a phonological. context in which the I change is required or at least appropriate For example. one might posit an I rule restricted to the context of a final as in 1. At first blush the analogical and rule based approaches seem to be different ways of saying. the same thing the context past in rule 1 forces the change to occur only in words. that are similar to fling sting etc But there is a critical difference The rule based approach. requires that fling sting etc be similar to spling in exactly the same way namely by ending in. I The structural description of the rule provides the necessary and sufficient conditions that. a form must meet in order for the rule to apply When similarity of a form to a set of model. forms is based on a uniform structural description as in 1 we will refer to this as structured. similarity A rule based system can relate a set of forms only if they possess structured. similarity since rules are defined by their structural descriptions. In contrast there is nothing inherent in an analogical approach that requires similarity. to be structured each analogical form could be similar to spling in its own way Thus if. English hypothetically had verbs like plip plup and sliff sluff in a purely analogical. model these verbs could gang up with fling sting etc as support for spling splung as. shown in 2 When a form is similar in different ways to the various comparison forms. we will use the term variegated similarity,122 A Albright B Hayes Cognition 90 2003 119 161. Since analogical approaches rely on a more general possibly variegated notion of. similarity they are potentially able to capture effects beyond the reach of structured. similarity and hence of rules If we could find evidence that speakers are influenced by. variegated similarity then we would have good reason to think that at least some of the. morphological system is driven by analogy In what follows we attempt to search for such. cases and find that the evidence is less than compelling We conclude that a model using. pure analogy i e pure enough to employ variegated similarity is not restrictive. enough as a model of morphology, It is worth acknowledging at this point that conceptions of analogy are often more. sophisticated than this permitting analogy to zero in on particular aspects of the. phonological structure of words in a way that is tailored to the task at hand We are. certainly not claiming that all analogical models are susceptible to the same failings that. we find in the model presented here However when an analogical model is biased or. restricted to pay attention to the same things that could be referred to in the corresponding. rules it becomes difficult to distinguish the model empirically from a rule based model. Chater Hahn 1998 Our interest is in testing the claim of Pinker and others that some. morphological processes cannot be adequately described without the full formal power of. analogy i e beyond what can be captured by rules Thus we adopt here a more powerful. if more na ve model of analogy which makes maximally distinct predictions by. employing the full range of possible similarity relations. 2 2 Criteria for models, Our modeling work takes place in the context of a flourishing research program in.
algorithmic learning of morphology and phonology Some models that take on similar. tasks to our own include connectionist models Daugherty Seidenberg 1994. MacWhinney Leinbach 1991 Nakisa et al 2001 Plunkett Juola 1999 Plunkett. Marchman 1993 Rumelhart McClelland 1986 Westermann 1997 symbolic. analogical models such as the Tilburg Memory Based Learner TiMBL Daelemans. Zavrel van der Sloot van den Bosch 2002 Analogical Modeling of Language AML. Eddington 2000 Skousen 1989 the Generalized Context Model Nakisa et al 2001. Nosofsky 1990 and the decision tree based model of Ling and Marinov 1993. In comparing the range of currently available theories and models we found that they. generally did not possess all the features needed to fully evaluate their predictions and. performance Thus it is useful to start with a list of the minimum basic properties we think. are necessary to provide a testable model of the generative capabilities of native speakers. First a model should be fully explicit to the point of being machine implemented It is. true that important work in this area has been carried out at the conceptual level for. example Bybee 1985 Pinker Prince 1988 but an implemented model has the. advantage that it can be compared precisely with experimental data. Second even implemented models differ in explicitness some models do not actually. generate outputs but merely classify the input forms into broad categories such as. regular irregular or vowel change As we will see below the use of such broad. categories is perilous because it can conceal grave defects in a model For this reason a. model must fully specify its intended outputs,A Albright B Hayes Cognition 90 2003 119 161 123. Third where appropriate models should generate multiple outputs for any given input. and they should rate each output on a well formedness scale Ambivalence between. different choices with gradient preferences is characteristic of human judgments in. morphology including the experimental data we report below. Fourth models should be able to discover the crucial phonological generalizations. on their own without human assistance This means that models should not. require that the analyst select in advance a particular group of phonological properties. for the model to attend to 1 Models that satisfy this criterion are more realistic. and also produce clearer comparative results since their performance does not. depend on the ability of the analyst in picking out the right learning variables in. Finally for present purposes we need a pair of models that embody a. maximally clear distinction between rules and analogy following the criterion of. structured vs variegated similarity laid out in the previous section From this point of. view a number of existing models could be described as hybrid rule analogy models. While such models are well worth exploring on their own merits 2 they are less. helpful in exploring the theoretical predictions of analogical vs rule based. approaches, Below we describe two implemented models that satisfy all of the above criteria. 2 3 A rule based model, 2 3 1 Finding rules through minimal generalization. Our rule based model builds on ideas from Pinker and Prince 1988 pp 130 136. The basic principle is that rules can be gradually built up from the lexicon through. iterative generalization over pairs of forms The starting point is to take each learning. pair here a verb stem and its past and construe it as a word specific rule thus for. example the pair shine shined3 SaI n SaI nd is interpreted as SaI n becomes. Some examples Plunkett and Juola 1999 fitted input verbs all monosyllabic into templates of the form. CCCVVCCC They used right alignment so that final consonants were always placed in the final C slot whereas. initial consonants would be placed in any of the first three slots depending on the initial cluster length In. Eddington s 2000 analysis of English past tenses using AML and TiMBL verbs were coded with a predefined. set of variables that included the final phoneme an indication of whether the final syllable was stressed and a. right aligned representation of the last two syllables In both cases the choice was highly apt for learning English. past tenses but would not have been if some quite different morphological process such as prefixation had. happened to be present in the learning data, In contrast the actual input data to children consist of whole words composed of dozens or even hundreds of. frequently correlated feature values Furthermore phonological environments are often formed from. conjunctions of two or more features e g d is selected when the final segment is both alveolar and a. stop and different features are relevant for different classes cf t when the final segment is voiceless Recent. work in the area of instance based learning has made headway on the task of finding the relevant features from. among a larger set see Daelemans et al 2002 Howe Cardie 1997 Wettschereck Aha Mohri 1997 Zavrel. Daelemans 1997 however we are not aware of any feature selection technique that would allow the learner. on the basis of limited data to isolate all the different combinations of features that we find to be relevant below. To encourage such exploration we have posted our learning sets features and experimental data on the. Internet http www linguistics ucla edu people hayes rulesvsanalogy. wug test data on English past tenses which show that participant ratings of novel pasts depend on the phonological shape of the stem both for irregulars and surprisingly also for regulars The latter observation cannot be explained under the dual mechanism approach which derives all regulars with a single rule To evaluate the

Related Books