Recursive Metaphors in Naturalistic Explanation:
Directionality and feedback loops between conceptual domains

by Tim Rohrer

Dept of Philosophy, University of Oregon
e-mail: rohrer@darkwing.uoregon.edu
to be presented on Thursday, July 17, 1997 at 3:45-4:15 pm in Room 2A-00 during the ICLC 97 in Amsterdam

Most of us are (rightly) suspicious of explanations that purport to explain everything, including themselves. In Darwin's Dangerous Idea the cognitive scientist and philosopher Daniel Dennett recounts a paradox about a fictional 'universal' acid which by definition dissolves everything with which comes in contact, including its own container. Dennett argues that the idea of evolution raises precisely the sort of paradox posed by universal acid: it is an explanation which purports to explain everything, including itself. As Dennett writes
"Like universal acid, Darwin's idea began to quickly eat its way out of the original container. If the redesign of organisms could be a mindless, algorithmic process of evolution, why could that process itself not be the product of evolution, and so forth ... ?"
Such an explanation of evolution would be recursive in its defining things in terms of itself. In Godel, Escher and Bach, Douglas Hofstadter argued that a recursive definition is a legitimate explanation as long as it defines itself in terms of simpler versions of itself. That is, as long the directionality of explanation is constrained as recursing from the more complex to the simpler, the explanation is legitimate. On Denett's terms, evolution is an explanation which can legitimately explain itself.

My interest in Dennett lies less in his assessing the merits or failings of his particular theory of evolution than as an example of how conceptual metaphors are used recursively. For Dennett, an avowed philosophical mechanist, evolution is an algorithmic and mechanical process. He uses the conceptual metaphor of EVOLUTION AS BUILDING to explain this mechanism, drawing on the fact that a large crane is often set up by a series of smaller cranes. The process of natural selection among organisms is itself explained as simply the result of smaller organismic machines--"cranes"--which have built organisms complete with larger cranes, one of which was the crane of natural selection. But the process of natural selection now itself operates as a crane to build, with greater speed and efficiency, more complex organisms which are complete with even larger cranes capable of building even more complex ... and so forth. We might schematize this recursive use of metaphor in the Lakoff-Johnson source-target notation for conceptual metaphors as:

( ( ( building --> evolution ) --> evolution ) --> evolution ... )

Recursive metaphors have interesting metaphorical entailments. Given his use of recursive metaphors, Dennett unsurprisingly favors Richard Dawkins' recent proposal of "cultural evolution" via 'memes,' or ideas which replicate themselves as genes do. Human culture, Dennett writes, is "a crane-making crane." Dennett regards an evolutionary explanation at the cultural level as simply the next logical step outward in recursing the EVOLUTION AS BUILDING metaphor. Like universal acid, Dennett's dangerous idea (his theory of evolution) has, via recursion, eaten "its way out of its its own container"--biology--and entailed a new anthropological theory.

Striking though recursive metaphors like Dennett's might be, they are not altogether unique. In this talk I shall give several examples of the use of recursive metaphors in naturalistic explanation ranging from Dennett to Hobbes and Aristotle. The theoretical focus of the talk is on assessing the implications of this new set of phenomena with respect to two problematic points in the Lakoff-Johnson two-domain model of conceptual metaphor: the strong directionality constraint from source to target and the need to further explain how domains change on the basis of feedback effects which violate the directionality constraint. I will also discuss the implications of recursive metaphoricity with respect to the attempt that Fauconnier and Turner have made to address these weaknesses in their many-space model of conceptual blending.

Keywords: metaphor theory; recursion; conceptual blending; evolution

Call for Examples:

While I have collected numerous examples of recursive metaphors, I am always eager to hear about more. Additionally I am currently collecting any and all examples which challenge the strong directionality constraint (from the bodily to the more 'abstract') of the Lakoff-Johnson two-domain theory of metaphor. I'm not only interested recursions but in cases which appear to either reverse this directionality or ones which exhibit feedback between the two domains. If you have any examples, please send them to the email address above.

Back to the metaphor center online