As a linguist – and in particular, one working with older texts as data– a major benefit of cognitive framing for textual analysis is the awareness that I’m constrained by knowledge about human cognition. If I can’t read long-ago authors’ or redactors’ minds, I can at least propose textual meanings and readings that are plausible for a human, embodied brain.

For example, extensive work on linguistic and cognitive categorization suggests a model for how the mind works to categorize things like people, events, and concepts. First, it tells us that categories are prototype-structured (some members are more central than others) and frame-based (knowledge representation “clumped” according to experience). Thus, the more I know about culture specific as well as universal aspects of experiential frames (legal frames, social frames), the more I can feed that into my semantic analysis.

I also know that lexical and constructional polysemy (coexistence of multiple meanings) are “rampant” (as Ronald Langacker liked to say). So, I should not be surprised to find that a word refers to two different levels in a category structure. For example, if a word for Man refers to both Man and Person, this tells me that Man is the central sub-category of Person for that culture. Also, knowing that

• frames are basic to semantics,
• metaphoric mappings between frames (creation of new meaning by mapping similar elements of one frame onto another) are basic to cognition,
• and frame-metonymic mappings are a natural fallout from frame structure,

I should not be surprised to find pervasive lexical polysemy patterns which are metaphoric (high prices, close relationship) or metonymic (cut refers both to an action and to a resulting surface discontinuity).

Analytically, what I get from doing formal frame and metaphor analysis, and formal mappings of mental spaces (cognitive structures created “online” in discourse) is precision and transparency. Without these tools I could perform a good reading and correctly identify many metaphors in texts, but with them I can see the problems and complexities within such identifications. When I work through all the details of mappings entailed by other mappings, I can see which ones don’t work, which are mutually contradictory, and which are unpredicted by my analysis so far.

These tools also help with precise identification: is this really love is fire, or a more specific lust is fire; is it love is a connecting tie or a relationship (including love) is a connecting tie? Looking at the complex of frames called love (including Relationship, Affection and Lust) helps me notice which frames are being mapped in a particular usage. Students in my metaphor class feel as if a lightbulb goes on when they see how these methods let them rigorously test their insights – they can see not only what metaphoric connections are present, but also how they are structuring the reasoning of the text.

Mental spaces theory, originally developed by Gilles Fauconnier to solve reference problems, has turned out to be an invaluable tool for doing this kind of precise work on viewpoint. According to this theory, we construct meaning in texts by creating arrangements of mental spaces, noting everything the linguistic and constructional cues let us know, for example, that both a narrator’s and a character’s viewpoint are manifest in a text.

Cognitive tools also eventually brought me to the realization that I need to study language as a whole multimodal communicative activity, not just as a symbolic system in the linguistic channel. Of course, I can’t examine the gestures of characters in old texts, but I can examine instances in which bodily actions are described. And in general, I need to remember that these older texts, like modern ones, were written by people whose spoken communication was multimodal.  These written texts are iceberg-tips: to the authors and contemporary readers, they might hint at intonation (even with no punctuation) and evoke many aspects of full multimodal communication.

Like cognitive science, comparison of a wide range of unrelated languages allows scholars to get a picture of what the constraints on language structures are. The linguistics world in which I live is a strongly comparative one – my academic career came of age along with the subfield of linguistic typology. What’s the range of ways a language can represent time, or space? I’ve had the fun of working with cognitive scientist/anthropologist Rafael Núñez on the time-metaphor system of Aymara, an Andean language which views Time as a static landscape with the “visible, known” Past in front of Ego and the Future in front of Ego. You can’t know a culture without knowing its frames and metaphors. And once you see what the crosslinguistic repertoire of metaphors is (more is up is my top candidate for a universal) this can help you limit your analytic options for languages which no longer have native speakers around to work with.

Obviously, I don’t have to be a disciplinary linguist to use these tools; they are a basic part of the cognitive linguistic “tool kit” but they belong to cognitive science as well, and to cognitively oriented scholars in any field. In my case, I can’t help practicing cognitive linguistic analysis of poetic texts, even though I sometimes acknowledge that I’m “practicing literary criticism without a license.”[1]  Finding out what metaphor systems are present in a text is an old endeavor, long preceding Modernist analyses (cf. Northrop Frye or Tillyard’s Elizabethan World Picture). But it’s a basic one, and we have new tools for it.

In my ideal world, cognitive scientists would all fully recognize the obvious: the works of Shakespeare, Jane Austen and Homer are exceptional products of human cognition, as are the Psalms. They’d find great authors’ cognitive processes as worthy of study as the mental processes of great scientists, or great chess players. And literary scholars would all likewise consider their subject-matter, literature, as a cognitive/linguistic achievement.  Of course, it hasn’t turned out that way. In the 1980’s and 90’s, literary scholars (in rather natural reaction to previous waves of highly objectivist criticism, and to the extremely objectivist direction of linguistics) turned to postmodernism and have only somewhat swung back with new historicism. Literary scholars, linguists and cognitive scientists tend to view each other’s frameworks with skepticism and to use non-overlapping data. But some of us continue to refuse to live in our boxes.  We will just be literary scholars and cognitive scientists, or linguists and cognitive scientists with an interest in literary texts.

I’ve only gradually realized some of what I describe above; but for me, the journey started when I took the first cognitive linguistics course ever taught, as a first-year graduate student in Spring 1977 on the Berkeley campus. Thank you, Chuck Fillmore, George Lakoff, Paul Kay, Sue Ervin-Tripp, Robin Tolmach Lakoff, John Searle, John Gumperz, and Bert Dreyfuss. I have never been able to go back into the box of formal linguistic analysis done without the context of meaning, culture, history, and embodied cognition.

[1] Don H. Doyle, Faulkner’s County: The Historical Roots of Yoknapatawpha (Chapel Hill: University of North Carolina Press, 2001), p. 17.