MetaCreativity

May 13, 2004 6:45 AM

What is the modus operandi of creativity? According to two cognitive scientists, Gilles Fauconnier and Mark Turner, the subconscious operation of conceptual blending. The formal theory, known as the Network Model of Conceptual Integration (CI), seeks to explain how creative insights are derived from pre-existing knowledge and understanding.
[More Inside]
posted by Gyan (10 comments total)

In Conceptual Blending theory(CI), one works with mental spaces(pdf), which are schematical representations of knowledge and relationships i.e. an organizing frame. So, the mental space of driving, would contain knowledge about vehicles, their components, the functional relationships between human actions and resultant activity (steering, braking,...). One can take diverse mental spaces dealing with different subjects and "blend" them. Take the example of the computer desktop UI. You deal with folders, trash bin, files, and with activities like copy, move...etc. Here the source mental spaces are that o fthe regular office and the computer display. The blended output is the metaphorical organization of computer data in terms of concepts derived from the office. In general terms, one can incorporate any number of source mental spaces, link their functional counterparts, and derive a target blend. Creativity is said to result, when the blended space contains emergent structure (organization) not present in the source mental spaces.

There are many ways to perform blending. According to the authors, one way in particular, double-scope blending is responsible for the uniqueness of human thought and the emergence of language. The governing principles of CI are presented here(pdf).

The theory of Conceptual Blending is in many ways a fleshing out of George Lakoff's Conceptual Metaphor theory. It is comprehensively explained in the book The Way We Think. Fortunately, the internet also has a lot of resources.
posted by Gyan at 6:46 AM on May 13, 2004

Don't forget that the brain is organic system, however - the neural substrate and the semantic are not necessarily closely linked. Secondarily, and perhaps more relevant to the topic at hand - bear in mind that neurons fire after achieving a certain threshold of signal - so sometimes just surrounding neurons in the neural substrate can 'accidentally' set off random neighboring neurons, which may represent something completely unrelated within the semantic substrate.

It would be more accurate to say that creativity arises from both conceptual blending and the occasional misfire within one's neural network.
posted by Ryvar at 7:03 AM on May 13, 2004

Ryvar: It would be more accurate to say that creativity arises from both conceptual blending and the occasional misfire within one's neural network.

I don't think that CI is really meant to be an accurate representation of neural activity. I look at it as a schema that approximately conveys the essence of creativity.
posted by Gyan at 7:07 AM on May 13, 2004

More data please. I can't seem to distinguish conceptual blending from any of the other analogical reasoning theories from the links provided. I can't even find anything beyond a very abstract discussion of probable mechanisms that mediate these processes from the list of links on their home pages.
Given that I can't nail down many of their claims I can't say much more about their theory.
posted by snarfodox at 8:39 AM on May 13, 2004

snarfodox, blending is an extension of analogy. Analogy and conceptual metaphor work with two spaces and has constraints on directionality. Blending can work with n-spaces and directionality and operations are more flexible.
posted by Gyan at 9:03 AM on May 13, 2004

I hate to sound derisive and snarky, so please take this with a grain: Has this genius just figured out that innovation, creativity and imagination are in large part processes involving associating (or creating "cognitive connections") and then comparing and contrasting different concepts?

Or is my explanation an oversimplification? Or is his dissertation an overly complex aggregation of $10 words?

Hmm. I will conceptualize these two ideas in what passes for my mind and blend them.
posted by Shane at 10:25 AM on May 13, 2004

Shane, who's "this genius"? There are two authors behind the theory. Their theory deals with how "cognitive connections" are made and compared. It's a formal theory.
posted by Gyan at 10:34 AM on May 13, 2004

Very sorry. Genii.
posted by Shane at 10:36 AM on May 13, 2004

I'm also having a hard time understanding what this theory is actually claiming about the way the mind works. I'm far from being an expert in any of these areas, but my initial impression is that this theory builds on the basic ideas of "Classical" AI, and repeats its fundamental error of postulating that the mind can be modeled as symbol manipulation, where the symbols correspond directly to semantic concepts of human language and thought.

Can anyone point to either a human experiment that demonstrates a concrete phenomenon explained by this theory, a computer model based on this theory that successfully replicates some non-trivial aspect of human behavior, or a theory that explains how this model maps to known neurological processes? Am I wrong in my assumption that no Classical AI theory has ever managed to achieve any of these goals?

To put it differently, how would we know whether this theory is true or not? What does it predict?
posted by fuzz at 11:44 AM on May 13, 2004

On a tangential note, here's Bruce Mau's Incomplete Manifesto which I interpret as a set of loose principles for improving your own creativity.
posted by euphorb at 1:23 PM on May 13, 2004