Thursday, August 21, 2008

Method: Push the Metaphor Until it Breaks
(or, mocking "imbricated" yet again)

I've been involved in an interesting exchange of emails with a student about how to do Tolkien studies, and that discussion have evolved into a larger discussion about intellectual practice.  This student is non-traditional, someone who has come to academia from a physical-labor and highly skilled job.  He/she asked me for some trick about how to generate ideas for papers and arguments.  I came up with a few and thought I would share one here:  push the metaphor until it breaks, then look at the broken pieces and figure out why it broke. 

So, for example, if you hear Foucault's metaphor of the "prisonhouse of language," push the metaphor: who is the warden? what shape would that prisonhouse be? Do people get work release? Is there parole?  Do people in it have just one cell mate?  Communal showers? Exercise yard?   Etc., etc. 

If you can build the metaphor bigger and bigger, and figure out how all those pieces might fit in, then that metaphor might be robust.  In Daniel Dennett's terms, it's a good "intuition pump." But if the metaphor collapses when pushed, then you know that perhaps it wasn't a good one, that it wasn't carrying the things you wanted it to carry. 

Which brings me to "imbricated," a metaphor in theory-speak for discourses that overlap, but don't overlap completely. 

Here's the problem.  "Imbricatus," especially when applied, as it originally was, to fish and reptile taxonomy (to describing animals with scales), really means "overlapping like shingles on a roof." (The alternates for scale descriptions are often "tessellatus," tessellated, like floor tiles, and tuberculatus, having small patches of scales surrounded by a lump, a tubercle). 

But do discourses really overlap like shingles on a roof?  To push the metaphor, that means that the overlap on each shingle is exactly the same amount.  And furthermore, that each shingle is identical to each other shingle, so the "imbrication" is really the same everywhere.  It seems to me that the metaphor of "imbricated discourses" breaks right here, and it breaks because the metaphorical description is not a particularly good intuition pump.  When you think of discourses as overlapping like shingles, the abstraction doesn't really help you understand anything else about the discourses.  When you start to manipulate the metaphor in your mind, you don't really find anything that you didn't already know (the way that manipulating other metaphors, such as genes as "wanting" to assist their copies--though I'm not a great fan of this metaphor--does help to uncover new relationships). 

And this is why I think that using the phrase "imbricated discourses" is really just a bit of ossified jargon whose real purpose is to obscure, not to illuminate.  "Partially overlapping" or "networked" or "intertwined" might not seem so technical, but these are actually better intuition pumps.  Hence I take the phrase "imbricated discourses" as being equivalent to a cliche in a short story: I start to think that the author hasn't put as much thought as was needed into the argument and is lazily relying upon materials pre-fabricated (and not pre-fabricated very well) by others. 

So the method is this:  don't automatically avoid the metaphor, but see if, by going along with it to the extreme of reducto ad absurdum, you can get it to break.  Then do failure analysis. 

(and I'll note that a very extended metaphor, that doesn't easily break, was developed by King Alfred in the Preface to the translation of Augustine's Soliloquies.  Alfred himself pushes that metaphor a long, long way without breaking it). 

No comments: