Overview

Metaphor and Thought is a collection of essays about metaphor. The book is primarily concerned with linguistics, but several essays trace back to the larger problem of understanding the role of metaphor in cognition. The general questions about metaphor are: How do people create and understand metaphors? How do we understand metaphors as linguistic structures? What is the relationship between metaphorical and literal speech? Is there such a thing as literal speech? The last question raises interesting concerns about the affect and use of language in communication.

Directions

Andrew Ortony: Metaphor: A Multidimensional Problem

Literal understanding and conception of meaning reached a height with Russel 1956 and Wittgenstein 1921/1961 and the rise of logical positivism. This is opposed to relativism and “mental construction” that arose with E Sapir 1921 and Whorf 1956. The latter asserts that types and uses of language are interdependent. (p. 1) Ortony distills questions down to distinguishing metaphors from nonmetaphors? “What are metaphors?” as well as a more interesting question, “What are metaphors for?” Max Black asserts that criterion for metaphorhood is inherently flawed. Something new is created when a metaphor is understood. What is that? (p. 5) In Development, new words are understood: “open” for both window and mouth. Is this the origin of embodiment in meaning? (p. 7)

Several widely asserted theories of metaphor: interaction view, comparison view, substitution view. Each equally compatible, incompatible w actual metaphor. (p. 10)

Max Black: More about Metaphor

This paper supplements the interaction view of metaphor. There is a mysterious, conventional nature of metaphor. It is not rationalist or literal certainly. “Say one thing and mean another” but why? (p. 21) There is a reductionist standard that the literal as norm, but metaphor is playful. Linguistically, metaphor violates some rules, but some must still be met to be acceptable. Can be no dictionary of metaphors, but there can be themes. Metaphors would be HARD in NLU. There are two categories of metaphors, emphatic and resonant. Emphatic is the tone and force of the metaphor, resonance is the allusion to other broader characteristics and qualities. A strong metaphor is both resonant and emphatic. (p. 26)

Black begins deriving a vaguely mathematical model of metaphors. “Every implication complex supported by a metaphor’s secondary subject, I now think, is a model of the ascriptions imputed to the primary subject: Every metaphor is the tip of a submerged model.” (p. 31) Translation and metaphorical equivalence…

Jerrold M. Sadock: Figurative Speech and Linguistics

Problem of linguistics: Form vs meaning. Linguistics seeks to make explicit the relationship between them. This is not possible without understanding what sentences mean. (p. 52) Sadock introduces a mathematical notion of semantics. U_c(E)=>P is the utterance of E on some occasion c. Where P is a porposition that is implied or conveyed. Sadock introduces confidence in implication logic, forming some implicative logic system. (p. 59)

L. Jonathan Cohen: The Semantics of Metaphor

Discussion of metaphor and its relation to language. Is metaphor independent of language? What are the mechanics of metaphor? “analogical urge” Problem of langue vs parole. Cohen introduces a great many examples of metaphors and examines the possible meanings and complex relationships thereof. (p. 65) Cohen introduces a property attribution style of analysis. That is, some features are attributed, while others are canceled. (+update, +cylindircal, etc). Metaphor cancels some propositions. But enables, alludes to a larger set. This sounds like semantic networking. (p. 70)

David E. Rumelhart: Some Problems with the Notions of Literal Meanings

Rumelhart challenges the very notion of literal speech in this paper. Performing an analysis of Searle and Sadock, figurative language is integral to learning, it is intuitive and natural. The difference between figurative and “conventional” or literal language is very hazy, usually there is overlap of some kind. (p. 78) “The traditional program of semantic analysis (cf. Katz & Fodor) provides a set of meanings for the individual lexemes of the language and then provides a set of rules of composition whereby the individual meanings of the lexemes are combined to form the meaning of the sentence.” These combined lexemes form the literal meanings of sentences, and form the set of literal sentences. This approach is lacking due to the inflexibility of meaning. (p. 81) There are several approaches to addressing metaphor: 1) Reject the traditional program of semantics and forma new account of literal and conveyed meanings. 2) Retain traditional program, but assume metaphor and other cases of figurative are accountable by another theory altogether. 3) Assume traditional theory works fine for most cases, but additional processing is required for violations. 4) Modify traditional theory to make it work for metaphor as well as literal language. Sadock has opted for (3), and Cohen for (4), but Rumelhart asserts that neither are valid and whole theory must be torn down and reconstructed to account for metaphor. Dominant theories are suspect. Some interesting issues to follow: NLU and representation, psychological analysis of human understanding (accounting for figurative) (p. 82)

Metaphor depends on our background, real world understanding, does not rely on logical order of lexemes. Even literal sounding statements require context in order to be processed and understood correctly, since many details are ambiguous or possible to misinterpret. Spoken words are gnomon. (p. 84) A process oriented model of cognition: Evaluate, hypothesize, search. (Sounds like SOAR) (p. 87)

John R. Searle: Metaphor

George Miller’s mathematical/linguistic approach to metaphor. Uses first order logic formula of property attribution. (p. 110) Searle outlines 8 principles of metaphor using some of the propositional approach. (p. 116) Searle finally creates some helpful graphs of meanings, expressing specifically various types of metaphors and how they relate to sentence and utterance meanings; literal utterance, metaphorical utterance (simple), metaphorical utterance (open ended), ironical utterance, dead metaphor, and indirect speech act.

Samuel R. Levin: Standard Approaches and Literary Metaphors

Levin describes a mapping graph accounting for the metaphorical construction, between the linguistic and phenomenalistic constructural transformations. (p. 132)

Allan Paivio: Psychological Processes in the Comprehension of Metaphor

The question at the bottom here: Matter of why, not as workaround, but as integral portion of understanding. (p. 151) Skinner: generalization of stimulus-response; metaphor is related to “abstractive seeing” as visual imagery. (p. 156) Paivio examines dual coding of imagery and verbal associations, finding some sort approach to memory. Dual coding uses visual terms, but could be extended to embodiment in general. (p. 163)

George A. Miller: Images and Models, Similies and Metaphors

Miller starts by addressing comprehension issues, and addresses metaphor as an apperceptive problem. He claims that when we hear “x is y” then we must imagine a world in which x is y. This seems to be curious and flawed reasoning. (p. 213) Miller later extends Searle’s first order model of linguistics to one resembling lambda calculus. This resembles Cyc and other AI style modeling, using relational models. Several types of metaphors are examined using this approach, nominal metaphors, predicative metaphors, and sentenial metaphors. This forms some sort of peculiar replacement logic. (p. 231) Finally, Miller examines metaphor recognition as a functional interpretation of lexemes. This model may potentially have some helpful uses. (p. 239) This also uses a symbol-object view on literal terms and meanings. (p. 247)

Richard Boyd: Metaphor and Theory Change: What is “Metaphor” a Metaphor for?

Boyd here uses metaphor as an approach to examine theory. Different theories have their own languages, and employ metaphors in context and frame specific ways of expressing meaning. Metaphor applies to symbol representation in science, metaphors are constitutive of the theories that they express. Some examples, “thought is a kind of information processing, and the brain is a sort of computer”, or “certain motoric or cognitive processes are pre-programmed”, etc… (p. 360) Logical positivism claims that relevant qualities of objects are defined by sensation (phenomenological perception) rather than verbal description. (p. 366) Metaphor is bound up in the notion of reference, and it is necessary to define a theory of nondefinitional reference to deal with referentiality, ambiguity, and linguistic precision. (p. 377)

Boyd spends some time on how to approach a theory of reference: reference is an epistemological notion, since semantic theory is a branch of epistemology. “A causal theory of reference is true precisely because reference is an epistemological notion and causal theory of knowledge is true.” Reference must be thougth of as dynamic and dialectical (as opposed to synchronic, piecemeal, and nondialectical): Changes in language do not represent changes of reference. This becomes study of reference by means of “epistemic access”. (p. 381)