tesfatsi AT iastate.edu
Syllabus for Econ 308
Basic Concepts
Key Issues
"The emerging perspective on embodied cognition may also offer the best hope so far for understanding central features of human thought and development."
"(The) intellectual alliance between development psychology and the other sciences of the embodied mind may prove to be one of the most exciting interdisciplinary ventures of the coming decade."
(Clark, p. 39): "In many cases, perception should not, it seems be viewed as a process in which environmental data are passively collected. Instead, perception may be geared, from the outset, to specific action routines."
Clark supports this view by discussing evidence that appears to demonstrate the context specificity of infant knowledge.
(Mitchel Resnick, 1994): "...people tend to look for the cause, the reason, the driving force, the deciding factor. When people observe patterns and structures in the world (for example, the flocking patterns of birds or the foraging pattern of ants), they often assume centralized causes where none exist. And when people try to create patterns or structures in the world (for example, new organizations or new machines), they often impose centralized control when none is needed."
(Clark, p. 40): "..a central message of our investigations -- a message that will recur again and again in this book: Complex phenomena exhibit a great deal of self-organization."
(Clark, p. 43): "Centralized control via detailed inner models or specifications seems, in general, to be inimical to ... fluid, contextual adaptation." Rather, systems "create actions from an `equal partners' approach in which the local environment plays a large role in selecting behaviors. In situations where a more classical, inner-model-driven solution would break down as as a result of the model's incapacity to reflect some novel environmental change, `equal partners' solutions often are able to cope because the environment itself helps to orchestrate the behavior."
Example: (Due to Pattie Maes, MIT, described by Clark on p. 43)
(Clark, p. 45): "In sum, the task is to learn how to soft-assemble adaptive behaviors in ways that respond to local context and exploit intrinsic dynamics. Mind, body, and world thus emerge as equal partners in the construction of robust, flexible behaviors."
(Clark, p. 46): "Biologists have tended to focus solely on the individual organism as the locus of adaptive structure. They have treated the organism as if it could be understood independent of its physical world. In this respect, biologists have resembled those cognitive scientists who have sought only inner-cause explanations of cognitive phenomena."
(Clark, p. 47): "In place of the intellectual engine cogitating in a realm of detailed inner models, we confront the embodied, embedded agent acting as an equal partner in adaptive responses which draw on the resources of mind, body, and world."
(Clark, p. 47): "This perspective leads to a rather profound shift in how we think about mind and cognition -- a shift I characterize as the transition from models of representation as mirroring or encoding to models of representation as control.... The idea here is that the brain should not be seen as primarily a locus of inner descriptions of external states of affairs; rather, it should be seen as a locus of inner structures that act as operators upon the world via their role in determining actions."
Example (Due to Maja Mataric, USC, described by Clark on pp. 47-49)
Clark rejects the traditionally perceived tripartite division of labor between perception, cognition, and intentional action. Instead, Clark argues (p. 51): "The internal representations the mind uses to guide actions may ... be best understood as action-and-control-specific control structures rather than as passive recapitulations of external reality."
Some Questions Arising During In-Class DiscussionRegarding Jen Diaz's comments on the approach to child development taken by development psychologist Jean Piaget, how does the approach of Piaget compare or contrast with the ideas espoused by Andy Clark in this chapter? In particular, does Piaget represent the kind of traditional tripartite perspective (division of mind into perception, cognition, and intended action) that Clark forcefully dismisses in Chapter 2 (Section 2.6)? Or is this view of Piaget's work too simplistic?
As A.J. Blair carefully discussed, Clark in chapter 2 bases his ideas about learning on learning in infants. A question raised by Will Rock is whether Clark's ideas about learning will hold up when extended to adults. We shall see in later chapters.
Regarding Will Rock's comments on "hard versus soft assembly" (Clark, p. 42), certain mental disorders in people seem to manifest themselves in a strong desire for rigid scheduling and repetition -- recall the autistic character Raymond played by Dustin Hoffman in the movie "Rainman." What does this imply about the organization of the brain and its potential for expressing the two types of behavioral modes?
Again regarding Will Rock's comments (and the comments of many others) regarding hard versus soft assembly, don't we as a matter of course build a lot of "hard assembly" rules and regulations into institutional structures (e.g., markets, highway systems, legislative processes, etc.) in an attempt to channel the behaviors of "soft assembled" human participants in socially desirable directions? For example, to what extent do current U.S. market institutions REQUIRE a large degree of flexible (adaptive) response on the part of human participants?
On the other hand, regarding Patrick Jordan's comments and questions on "data storage" intelligence versus "efficient flexible response" intelligence, what kind of intelligence is required to be economically SUCCESSFUL in the U.S. today? Many "New Economy" advocates are arguing that, as we move more towards an information-based economy, it is the latter kind of intelligence that is increasingly required for economic success, in the sense of being able to secure and sustain a relatively high income.
Regarding the comments on IQ by Chris Cook and Matt Neubauer (among others), how/when is "IQ" actually measured and what is it supposed to represent about intelligence?
Discussion among moderators, Kunil Daftari, Jordan Swanson, Andy Ruff, and others: Are robots capable of being truly innovative? For example, could they ever paint pictures in "new" (not just directly programmed) ways that are pleasing to humans? What if they paint pictures that are "pleasing" to other robots, even if not to humans? (Should robots be designed always and everywhere just to serve mankind?)
Regarding Matt Neubauer's interesting discussion of Patti Maes' rat model (pp. 47-49), when you have to drive somewhere, how many of you rely on "algorithmic" directions -- turn left at Main, go two blocks, then turn right.... -- and how many of you instead visualize and orient your direction path in the context of an encompassing spatial map? It seems that the Maes rat is proceeding to learn to navigate its environment in a purely algorithmic fashion.
What role does (or should) human moral judgement play in the construction of autonomous computational agents (in a computer) or situated agents (robots)? Might artificial agents evolve behaviors that humans would term "immoral" if the agents are trying to achieve goals with no moral restrictions on means? (Recall Hal of 2001 fame, and the issues along these lines raised in the movies Terminator 1, 2, and soon-to-be 3). See, for example, the ongoing controversial work by Peter Danielson (University of British Columbia) on Evolving Artificial Moral Ecologies.
Regarding the concerns expressed by Will Rock and David Gillingham about whether we really should be encouraging the development of artificial agents to the point where they equal or exceed humans with regard to intelligent capabilities, is there any way to stop this "march of science" development even if we wanted to do it?
An interesting discussion arose among Jen Diaz and many others in class regarding whether artificial agents will ever be able to evolve something akin to "emotions." Are Data (of StarTrek fame) and the little boy in the Speilberg-Kubrick movie AI (Artificial Intelligence) just far-out SF ideas, or could they become reality?
Patrick Jordan and Jeremy Spring raised a number of additional provocative questions for the moderators regarding possible future directions for artificial life forms. For example, what if we ever come to a point where self-replicating robots or other artificial agents evolve a morality that is good for robots but not necessarily for humans? What then? For example, suppose these robots decided that the earth would be better off without humans (and all other living things agreed with them)?
Finally, experiments are already being conducted in Japan and elsewhere in which biological life forms are combined with machinery (shades of the Borg in StarTrek:TNG). For example, researchers have modified the brains of biological cockroaches by insertion of computer chips so that their bodies are controllable by external programming instructions. And stunning advances in nanotechnology raise the possibility that, someday, people may routinely ingest nanorobotic swarms to carry out medical procedures. Will human life ever merge with machinery to the point that we have a "third" form of life between artificial and human?