Wednesday, September 12, 2018

Reading :: Supersizing the Mind

Supersizing the Mind: Embodiment, Action, and Cognitive Extension
By Andy Clark

I'm pretty sure I picked this 2008 book up several years ago with the intention of reading and reviewing it, but somewhere along the way, I put it down again. Now I can see why: It's really best read as a lengthy reply to critics of the original Clark and Chalmers (1998) article, which proposed the extended mind concept 20 years earlier. As such, it's an intimidating book for an outsider to pick up and read: the author discusses a broad range of psychological phenomena (perception, processing, sensorimotor dependencies) and objections involving each, including alternate readings. For someone like me, there's a lot of inside baseball.

However, I stuck with it this time and found some parts to be general enough to be useful to me. Among other things, it helped me to better and more systematically understand the principled objections to the extended mind thesis—a thesis with which I have felt more or less comfortable ever since Hutchins' Cognition in the Wild (1995).

In the Introduction, Clark contrasts two views of the mind: "BRAINBOUND" and "EXTENDED." To kick off the discussion, Clark quotes physicist Richard Feynman discussing some of his papers with historian Charles Weiner. Weiner characterizes the papers as a "record" of the work done in Feynman's head, and Feynman objects: no, the paper is the work (p.xxv). Clark claims:
Such considerations of parity [between in-head and out-of-head computation], once we put our bioprejudices aside, reveal the outward loop as a functional part of an extended cognitive machine. Such body- and world-involving cycles are best understood, or so I shall argue, are quite literally extending the machinery of mind out into the world—as building extended cognitive circuits that are themselves the minimal material bases for important aspects of human thought and reason. Such cycles supersize the mind. (p.xxvi)
That is, embodied cognition arises from bodily interaction with the world (p.xxvi).

Clark characterizes two extremes to understandings of the mind: (1) BRAINBOUND, in which "all human cognition depends on neural activity alone," the brain is the mind, and the rest of the body provides merely sensory input (p.xxvii); and (2) EXTENDED, in which "the actual local operations that realize certain forms of human cognizing include inextricable tangles of feedback, feed-forward, and feed-around loops: loops that promiscuously criss-cross the boundaries of brain, body, and world. The local mechanisms of mind, if this is correct, are not all in the head. Cognition leaks out into body or world" (p.xxvii).

With this succinct characterization of the two views, let's get into the body of the book.

Ch.1, "The Active Body," explores the consequences of being embodied. According to the "Principle of Ecological Assembly (PEA)," "the canny cognizer tends to recruit, on the spot, whatever mix of problem-solving resources will yield an acceptable result with a minimum of effort" (p.13, his emphasis). The explanatory strategy of "distributed functional decomposition (DTD)" (p.13) states that "it is the roles played by various elements, and not the way those elements are realized, that do the explanatory work" (p.14)—that is, when looking at cognition, we should identify chunks of computation, no matter where they take place (e.g., whether someone holds a figure in memory or writes it on paper). And finally, Clark invites us to think of sensing as coupling: people often "use the sensor as an open conduit allowing environmental magnitudes to exert a constant influence on behavior" (p.16, his emphasis). That is, "The embodied agent is empowered to use active sensing and perceptual coupling in ways that simplify neural problem solving by making the most of environmental opportunities and information freely available in the optic array" (p.17, still his emphasis).

With these principles in mind, Ch.2, "The Negotiable Body," explores the author's assertion that "human minds and bodies are essentially open to episodes of deep and transformative restructuring in which new equipment (both physical and 'mental') can become quite literally incorporated into the thinking and acting systems that we identify as our minds and bodies" (pp.30-31). He of course brings up the illustration of the blind man and the stick, an illustration that has been used by Merleau-Ponty, Gibson, and many others (p.31). Clark acknowledges the obvious criticism (is the tool used as a tool or actually incorporated?) (p.37) and uses research into mirror neurons to argue that, functionally, it is the latter (p.38). He concludes that "our own embodied activity enacts or brings forth new systemic wholes" (p.39, his emphasis, of course).

In Ch.3, "Material Symbols," Clark explores language as cognitive scaffolding with three effects:

  • "First, the simple act of labeling the world opens up a variety of new computational opportunities and supports the discovery of increasingly abstract patterns in nature."
  • "Second, encountering or recalling structured sentences supports the development of otherwise unattainable kinds of expertise."
  • "third, linguistic structures contribute to ... our ability to reflect on our own thoughts and characters and our limited but genuine capacity to control and guide the shape and contents of our own thinking." (p.44)
Labeling, he says, allows us to "cheaply and open-endedly project new groupings and structures onto a perceived scene" (p.46). 

Structured language, he says, has been explored in many literatures, including the Vygotskian (p.47). He's particularly interested in "the role of linguistic rehearsal in expert performance itself," i.e., how it "enables experts to temporarily alter their own focus of attention, thus fine-tuning the patterns of inputs that are to be processed by fast, fluent, highly trained subpersonal resources" (p.48). 

In Ch.4, he examines "niche construction," in which an organism alters its environment "in ways that may sometimes change the fitness landscape of the animal itself" (p.61). Niche construction can include "the cultural transmission of knowledge and practices resulting from individual lifetime learning, when combined with the physical persistence of artifacts" (p.62; compare Cole and Gajdamaschko's chapter "Vygotsky and Culture" for a Vygotskian understanding of this view). Ultimately, what matters is "the way niche-construction activity leads to new feedback cycles" (p.62). Niche construction involves space (p.64; think of how people do paperwork by arranging stacks, for instance). Clark notes that Sterelny (2003) dubs humans as "epistemic engineers" who "engineer their own habitats, and ... these are transmitted to the next generation, who further modify the habitat," including the "epistemic environment," modifications that "affect the informational structures and opportunities presented to each subsequent generation" (p.66). Clark compares epistemic and pragmatic actions: Pragmatic actions advance toward a goal, while epistemic actions extract or uncover information (p.71). Humans do both: "the human agent, one might say, is nature's expert at becoming expert" (p.75).

The machinery of mind, then, is not just biomachinery: external traces are among the physical vehicles of beliefs (p.76). Clark and Chalmers (1998) offered the Parity Principle: rather than starting with the assumption that cognition is in the head, they identify functions that, if they were to be implemented in the head, would count as cognitive functions. If those functions happen outside the head, Clark and Chalmers say, they should still be counted as cognitive (p.77). 

So ends Section I. In Section II, Clark addresses criticisms of the argument.

Ch.5, "Mind Re-bound?", acknowledges a criticism by Adams and Aizawa: "The fallacy is to move from the causal coupling of some object or process to some cognitive agent to the conclusion that the object or process is part of the cognitive agent or part of the agent's cognitive processing" (p.86—Clark's summary of the argument). 

Ch.6, "The Cure for Cognitive Hiccups," acknowledges Rupert's (2004, 2006) criticism that "cognitive processes lean heavily on environmental structures and scaffoldings but do not thereby include those structures and scaffoldings themselves" (p.111). 

Ch.7, "Rediscovering the Brain," acknowledges a criticism from a variety of quarters: that of "the asymmetric relationship that obtains between the organism and its props and aids. Subtract the props and aids, they argue, and the organism may create replacements. But subtract the organism, and all cognitive activity ceases" (p.162). Clark counters: "it is true that the key microlocus of plasticity is the individual human brain. It is the brain's great plasticity and thirst for cheap, outsourced labor that drives the distributed engines of sociotechnological adaptation and change. It is true, too, that by subtracting [the brain], the whole process grinds to a standstill. ... But it by no means follows, from the fact that [brains] are in that way lopsidedly essential to all this, that the rest of the hybrid, distributed circuitry is not part of the mechanistic base for specific episodes of cognitive processing" (p.162). 

In Ch.9, Clark summarizes his argument. He argues that embodiment matters to cognition in three ways:
  • "spreading the load"
  • "self-structuring of information"
  • "supporting extended cognition" (p.196). 
He goes on to sketch out theses for supporting a science of the extended mind. 

In all, the book grapples with the objections to, and implications of, understanding cognition as extending beyond the brain. Yet, as I found on my first attempt, these discussions are really focused on developing the above-mentioned science of the extended mind—that is, they go deeply into the philosophy of psychology. Since I'm neither a philosopher nor a psychologist, I found these aspects of the book to be less interesting (although I recognize their value). 

If you have become interested in extended cognition or others in this area (situated cognition, distributed cognition), this book is a great way to think more deeply about the issues involved and the objections leveled against such views. Notice that it takes as given the notion that cognition is computation, something that Hutchins (1995) also takes as his starting point, but that situated cognitionists such as Lave (I think) would not concede. Check it out.

No comments: