A theory of analogical mapping between source and target analogs based upon Interacting structural, semantic, and pragmatic constraints is proposed here. The structural constraint of isomorphism encourages mappings that maximize the consistency of relational corresondences between the elements of the two analogs. The constraint of semantic similarity supports mapping hypotheses to the degree that mapped predicates have similar meanings. The constraint of prog-mafic central/! / favors mappings involving elements the analogist believes to be Important in order to achieve the purpose for which the analogy Is being used. The theory is implemented in a computer program called ACME (Analogical Constraint Mapping Engine), which represents constraints by means of a network of supporting and competing hypotheses regarding what elements to map. A coop-erative algorithm for parallel constraint satisfaction identifies mapping hypotheses that collectively represent the overall mapping that best fits the interacting constraints. ACME has been applied to a wide range of examples that include problem analogies, analogical arguments, explanatory analogies, story analogies, formal analogies, and metaphors. ACME is sensitive to semantic and pragmatic Information if it Is available,.and yet able to compute mappings between formally Isomorphic analogs without any similar or identical elements. The theory Is able to account for empirical findings regarding the impact of consistency and similarity on human processing of analogies.

Abstract not found

Hofstadter and his colleagues have criticized current accounts of analogy, claiming that such accounts do not accurately capture interactions between processes of representation construction and processes of mapping. They suggest instead that analogy should be viewed as a form of high level perception that encompasses both representation building and mapping as indivisible operations within a single model. They argue specifically against SME, our model of analogical matching, on the grounds that it is modular, and offer instead programs such as Mitchell and Hofstadter's Copycat as examples of the high level perception approach. In this paper we argue against this position on two grounds. First, we demonstrate that most of their specific arguments invo lving SME and Copycat are incorrect. Second, we argue that the claim that analogy is high-level perception, while in some ways an attractive metaphor, is too vague to be useful as a technical proposal. We focus on five issues: (1) how perception relates to analogy, (2) how flexibility arises in analogical processing, (3) whether analogy is a domain-general process, (4) how micro-worlds should be used in the study of analogy, and (5) how best to assess the psychological plausibility of a model of analogy. We illustrate our discussion with examples taken from computer models embodying both views.

This report is intended as a companion to another report, Asymmetry and Reverse Transfers in Metaphor, [Barnden et al., 2002]. In that report, it is argued that the transfer of information and other effects from the tenor or target domain to the vehicle or source domain plays a highly important, albeit subsidiary, role in metaphor understanding; the main transfer in metaphor understanding goes from source to target. Furthermore, it is claimed that the importance of this type of transfer has been underestimated or overlooked entirely, hidden from detailed study by a confusion between different notions of direction and asymmetry in metaphor that one finds in the literature. However, the latter report, whilst it refers to an AI program for metaphorical reasoning called ATT-Meta, which implements target to source (and source to target) transfers, does not report any results from a simulation. It does not give a detailed listing of the rules used in illustrative discourse examples exemplifying the various types of reverse transfer, and it does not give a step-by-step account of ATT-Meta’s reasoning as it processes the different examples. This report provides enough detail of the ATT-Meta system to understand its approach to metaphorical reasoning, and it lists the rules and shows the reasoning involved in processing the examples. 0

We present and evaluate an approach to the reasoning needed to handle a broad class of metaphorical ut-terances, and a computer program (ATT-Meta) partially implementing and further specifying that approach. The approach emanates from artificial intelligence but is offered also for consideration by cognitive scien-tists generally. The utterances of interest are ones that (a) rest on conceptual metaphors that are familiar to the understander but (b) transcend the mappings in the conceptual metaphors by using concepts not han-dled by the mappings. Our approach advocates possibly-extensive inferencing in the terms of the source (vehicle) domains of the conceptual metaphors, while avoiding as far as possible the extension of the map-pings to deal with the concepts they do not handle. The general approach is similar in flavor to those of a small number of other metaphor researchers, but we provide a more extensive analysis, additional principles and a more thorough-going implementation. The approach contains a number of “view-neutral mapping adjuncts, ” which are default mapping principles that enable important source-domain aspects to be mapped to the target domain, independently of which specific metaphorical views are in play. Many discussions of metaphor appear to assume that such mapping actions occur, but rarely address them systematically and explicitly. In addition, in the approach, a conceptual metaphor can consist not only of a between-domain mapping but also of special, ancillary assumptions that serve to enrich the source domain with specific de-tails needed by the metaphors. The implemented system supports ancillary assumptions but currently only has a preliminary handling of view-neutral mapping adjuncts. 2 1

If metaphor is to be viewed as a fundamental cognitive agency, as recent work suggests, what ramifications does this view have for a model of semantic memory? This paper argues that if various signature phenomena of metaphor are to be given adequate computational treatment, a parallel adaptive (learning) network model of memory is required. Such a model, the Sapper framework, is described; Sapper is a hybrid symbolic/ connectionist model which views the interpretation of novel metaphors as a process of connectionist bridgebuilding, which subsequently alters the activation dynamics between conceptual schemata. 1. Introduction Recent years have heralded the acceptance of metaphor by the natural language processing community as a deep cognitive and linguistic phenomenon (see for example Lakoff & Johnson 1980; Weiner 1984; Martin 1990; Way 1991; Veale & Keane 1992a,b). Metaphor is now recognised to serve a fundamental structural role in the organization of our conceptual systems, overturni...

. The distinction between literal and gurative language (metonymies, metaphors, etc.) is often not made formally explicit, or, if formal criteria exist, insucient. This poses problems for an adequate computational treatment of these phenomena. The basic criterion for delineating literal from gurative speech we propose is centered around the notion of categorization conicts that follow from the context of the utterance. In addition, we consider the problem of granularity, which is posed by the dependence of our approach on the underlying ontology. 1 Introduction Figurative speech comes in dierent varieties (e.g., the metonymy in example (2) and the metaphor in example (3) below), and is typically contrasted with literal language use (e.g., example (1)) on the basis of some notion of deviance. (1) \The man left without paying." (2) \The ham sandwich left without paying." (3) \The Internet is a gold mine." Currently, two approaches prevail, which spell out this distinction...

Using metaphor-annotated material that is sufficiently representative of the topical composition of a similar-length document in a large background corpus, we show that words expressing a discourse-wide topic of discussion are less likely to be metaphorical than other words in a document. Our results suggest that to harvest metaphors more effectively, one is advised to consider words that do not represent a discourse topic. Traditionally, metaphor detectors use the observation that a metaphorically used item creates a local incongruity because there is a violation of a selectional restriction, such as providing a non-vehicle object to the verb derail in Protesters derailed the conference. Current state of art in metaphor detection therefore tends to be “localistic ” – the distributional profile of the target word in its immediate grammatical or collocational context in a background corpus or a database like WordNet is used to determine metaphoricity

Children encounter metaphor in all aspects of their daily lives, including the spoken and written discourse of school, and metaphor can be one of the routes through which socio-cultural norms are appropriated. The use of metaphor seems to be a basic human skill, which develops in interaction with developing world knowledge and linguistic skills. Both use of metaphor and the development of metaphor capacity take place in situations in which language is used for personal and interpersonal goals. This paper explores how the discourse context of metaphorical language supports children’s understanding, and examination of classroom data shows how language, situation and interaction all offer assistance to the interpretation of metaphorically-used language. I further argue for a discourse approach to the study of children’s developing metaphor capacity, and identify some of the dimensions of metaphor capacity: knowing when metaphor use or metaphorical interpretation is appropriate, negotiating appropriately rich meanings of metaphor, accessing stored contextualised metaphorical language and meanings, and using metaphor to achieve cognition or affective goals. The implications of a discourse approach for researching metaphor, and for teaching and learning are discussed.

Choosing the right word is difficult. One reason is that the context affects the meaning expressed by a word in complex ways. In particular, when a word is used in a context that is not normal for the word, it may evoke a special meaning. This paper presents a lexical choice process that chooses the word from a set of near-synonyms that best produces the desired effects in the given context. It relies on a clustered representation of lexical knowledge that unites both a statistical model of word co-occurrence (for determining when a word use will be marked), and a knowledge-based model (for determining what specific effects will occur). In Proceedings of the First Student Conference in Computational Linguistics in Montreal, 1996, pages 80-87. Evoking Meaning by Choosing the Right Words Philip Edmonds University of Toronto 1. Context and lexical choice Choosing the right word is difficult for computers, let alone for people. Effective communication requires that one seek to express ...