This is the first of three chapters of Juarrero’s book about which I’ve written summaries/reviews. The first two (Chapters 9 and 13) really just open the doors to the last (Chapter 14), which I find interesting for my own work.
Juarrero, A. (1999). Constraints as causes: the intersection of information theory and complex systems dynamics (Chapter 9). Dynamics in Action: Intentional Behavior as a Complex System (pp. 131–150). Cambridge, MA: The MIT Press.
In Chapter 9 of Dynamics in Action, Juarrero sets out to analyze the “interlevel causality of dynamical systems—as the workings of constraint” (p. 131). One of her goals is to explain top-down causality, “how intentions can cause and flow into behavior” (p. 131). Juarrero claims previously to have shown that “systems are created from interacting components, which they then, in turn, control” (p. 131), but that this raises the question of “self-cause.”
A central concept in this chapter is that of constraints. Juarrero defines physical constraints as “relational properties that parts acquire in virtue of being unified—not just aggregated—into a systematic whole.” As an example from physics, she describes a sphere that rolls down a plank—the plank constrains the movement of the sphere, forces it in a sense to behave in a certain way, without acting on it in the way that the force of gravity does (p. 132). She also notes how the kneecap is connected to tissues around it in a system, where none of the parts can act or be acted on without some reference to the others (p. 133). Juarrero argues that constraints don’t just limit the options of systems but must also “open up” options (p. 133). In the information theoretical context, constraint addresses that fact that the system that is potentially capable of expressing the greatest number of messages is the one that expresses none. For any of the potential messages to become actual messages, the system must be constrained to “harness the randomness”: “Constraints thus turn the amorphous potential into the definite actual” (p. 134).
These constraints are context-free if they “alter the probability distribution of the available alternatives,” if they “make a system diverge from chance, randomness, or equiprobability” (p. 135). One example she offers is the distribution of letters in a language: zs is less likely in English than es. The frequency of these letters in a string communicates probabilities for whether there is a message and whether it is in English. She explains why context-free constraints are insufficient for encoding complex systems, before introducing context-sensitive constraints. These constraints take into account system-wide probabilities, but also contextual probabilities. For example, a context-free constraint describes the frequency of the word “the” in English; a context-sensitive constraint describes the frequency of the word “the” immediately following a verb in English. The appearance of the letter q in English more likely presages a u than another q (making qu a feature called an i-tuplet).
Such context-sensitive redundancies or context-sensitive constraints, when applied hierarchically, permit an infinite set of messages from a constrained system: “there are more words than i-tuplets, more sentences than words” (p. 138). “By correlating and coordinating previously aggregated parts into a more complex, differentiated, systematic whole, contextual constraints enlarge the variety of states the system as a whole can access” (p. 138).
On a thing’s external structure: “Once the probability that something will happen [B] depends on and is altered by the presence of something else [A], the two have become systematically and therefore internally related…. A is now a part of B’s external structure” (p. 139). Juarrero calls the fact that a system’s current state “is in part dependent on a prior one” feedback, and she claims this allows “part of the system’s external structure to run through its history” (p. 139).
First-order contextual constraints are those constraints imposed by entities operating at the same level of organization, for example, two slime-mold amoebas foraging in the same environment for food; second-order contextual constraints arise when the system containing entities in first-order relationship itself imposes constraints on all the entities in the system, for example, when the amoebas begin to form a spore-stalk with fruiting body, and the individual amoebae take on different roles (p. 142).
She defines selectionist constraints: these top-down, second-order constraints “reduce the number of ways in which the parts can be arranged” (p. 144) but at the same time are creative in that they impose upon their components a function.
Juarrero offers an argument for why her model does not result in causal circularity (the dealing of cards to card players). She then offers evidence from the fields of biology (including Goodwin’s conceptualization of organisms as fields); neurology (including Damasio’s “structured structuring structure”; p. 147); auditory perception; olfactory perception; and visual perception.