Here are some comments about the diagrams I drew on the chalk board in lecture on November 2. In these diagrams, the boundaries of the individual cognitive system (early views) or the person (later views) are indicated by a bounding circle or line. Outside the line lies the environment for thinking. Notice that for the later views it becomes important to specify in some detail what lies in the environment, setting, habitat of thinking.
The traditional unit of analysis has been a Physical Symbol System (PSS) residing somewhere inside a person. In the diagram, the heavy circle depicts the boundaries of the cognitive system. Notice that the boundary of the system is not the skin or the skull of a person. It lies somewhere inside the person, presummably deep in the brain. The system functions by transforming strings of symbols into other strings of symbols (shown by the arrows inside the heavy circle. In this view, perception and action are seen as peripheral systems, not part of the core cognitive system. Perception transduces or translates the messy world into symbolic representations suitable for processing by the inner symbol system. Action translates computed symbolic descriptions into movements of the body.
It is possible to describe learning in a traditional cognitive system. Continued exposure to strings of symbols, for example, could lead to the gradual formation of internal strings of symbols that are near copies of the external strings of symbols.
In this example, the strings of symbols containing P and Q are meant to express a logical relationship of implication. The premise is that P implies Q. The notation shown indicates that if P is true then it can be inferred from the premise that Q is also true. The circles represent the cognitive system through different moments in time. Initially, the internal system does not represent the relationship internally, later it does.
This traditional view of the cognitive system now seems unlikely as a description of what actually happens in human brains. It is interesting to note that the PSS description works quite well for systems of Socially Distributed Cognition.
The PSS operates by creating and transforming physical representations (which may be strings of symbols or other physical structures). We describe the propagation of representational state across a variety of representational media in the system.
A more direct alternative to the traditional approach is provided by the Situated Cognition approach. This view emphasizes the role of the organied environment in the organization of cognition and action. Where the traditional approach focused on the internal processes, possibly neglecting processes that are external to the central PSS, the situated cognition view focuses on the activity of a person in a personally edited and engaged world, also known as a setting. The situated view has been criticized (perhaps justly) for neglecting internal processes. Much of the work in situated cognition remains disembodied, because the focus is outside the person.
The situated cognition approach introduced the notion of resouces and the idea that cognitive systems can be built up by getting resources into coordination. There is a dialectic relation between activity and setting. This means that one cannot understand changes in the organization of either activity or setting without understanding the changes in the other. The famous cottage cheese example is taken as a showcase instance of situated cognition. A weight-watcher computed 3/4 of 2/3 of a cup of cottage cheese by filling a measuring cup to the 2/3 mark, dumping the lump of cottage cheese on the cutting board, cutting a centered cross into the lump, and then taking three of the four pie slices produced. The challenge is this: This computation produces a cognitive outcome. Where was this computation performed?
The embedded view adds dynamics to situated cognition. It provides a more modern and sometimes neurally-informed account of systems composed of coordinated internal and external resources.
One version of embedded cogition is a modern descendant of Ecological Psychology, a field pionered by J.J. Gibson. Rather than seeing input and output as peripheral, perception and action are central parts of the embedded cognitive system. The relations between internal and external processes are not seen in terms of translation or transduction, as they were in the traditional view, but are now understood in terms of coupling and entrainment and mutual dependence.
I like to think that the sequence learning system that I described in chapter 7 of Cognition in the Wild is a good example of an embedded cognitive system.
The diagram I drew in lecture does not show much. See Ch 7 for an explanation of how, in this framework, learning appears as the propagation of organization from one part of a complex system to other parts. While this example did acknowledge the importance of multiple sensory and motor modes, it was actually still disembodied.
Embodied cognition is a rich framework that emphasizes the role in thinking of the resources that develop as a conseqence of the fact that human experience is shaped by the kinds of bodies we have.
The brain is in the body, which is in the world. Perception and action are linked to each other - perception is not something that happens to you, it is something you do. One of the most important kinds of knowledge we have is implicit knowledge about how our actions change our sensations. The regularities in these relations are called sensorimotor contingencies. A radical claim in this view is that what characterizes a sensorimotor experience is not where it happens in the nervous system, but the nature of the sensorimotor contingencies. The Tactile Visual Substitution System provides a strong argument in favor of this idea. The flower petals radiating out from the blob in the middle of the shape are supposed to depict sensorimotor loops that connect the brain to the world via the body. The dotted line depicts an approximate location for the traditional boundary of the cognitive system.
The Extended Cognition framework explores the possibilty that structures and processes outside the person could, under certain circumstances, play functional roles in cognitive processes that are indistinguishable from the functional roles played by internal resources.
This diagram is supposed to represent the situation described in the appendix to Supersizing the Mind. Otto's notebook and Inga's internal memory play very similar roles in answering the question, "Where is the Museum of Modern Art?" The announcement that there is an exhibit at MOMA is part of the world inhabited by Otto and Inga. Inga's internal memory is intact and she simply recalls the location of MOMA. Otto's internal memory is damaged, but he compensates by writing down useful bits of information, such as the location of MOMA. This allows him to "look up" the location of the museum. I do not find this example very compelling. Clark's analysis obscures our view of the important differences between the functional system used by Inga and the functional system used by Otto.