Time and Mind
- Published
A summary of section 1.8 of Supersizing the mind by Andy Clark.
1.8. Time and mind
This section introduces dynamical systems theory as one tool for describing how the elements of the extended mind discussed earlier interact and unfold over time. (Information self-structuring uses expressions such as “time-locked unfolding” and “temporally coupled unfolding” for this.)
According to the author, there has been a conflict between two approaches to mind and cognition:
- The traditional position, which favors functional, computational, information-processing, and information-theoretic approaches.
- The embodied-cognition position, which emphasizes embodiment, active perception, and synchronized unfolding.
Dynamical systems theory (DST) is a theoretical tool favored by the embodied cognition camp.
But the book does not argue that DST is superior. Rather, while DST has advantages, it alone has difficulty providing satisfactory explanations, especially of intelligent behavior (the subject of section 1.8). The two positions therefore need to be combined appropriately (the subject of section 1.9). In the author’s words:
(the purpose is) to reveal computational, representational, information-theoretic, and dynamical approaches as deeply complementary elements in a mature science of the mind.
The aim is to show that computational, representational, information-theoretic, and dynamical approaches complement one another deeply within a mature science of mind.
Before discussing DST, it helps to look at the problems it seeks to capture. The author identifies CRC, continuous reciprocal causation, as one such problem. CRC occurs when a system S continuously influences another system O while continuously being influenced by O. Such relationships can occur at several scales:
- Between brain regions linked by multiple feedback and feedforward neural pathways.
- Between a dancing couple. The woman’s neural state influences her movements, her movements influence the man’s movements, and he influences her in turn, repeatedly.
- Between jazz musicians who influence one another as they improvise.
It is finally time to explain DST itself, but this was becoming too long, so I separated that discussion into Introduction to dynamical systems theory. What follows assumes you have read that introduction.
Because the factors involved in a system influence one another in complex ways, DST emphasizes understanding the system’s total state, for example by plotting changes in state variables over time. This is its biggest difference from standard computational approaches, and it has the following advantages and disadvantages.
Its advantage is that it captures CRC well, as described above. That is straightforward, so I will not elaborate. Its disadvantage is that when the system being explained contains an intelligent agent, such as the human brain, the agent has to be simplified and treated like the other, nonintelligent, reactive, mechanical elements. That makes it difficult to understand the entire system containing that intelligent agent properly. The main point of this section is therefore that the two approaches need to be combined appropriately.