Cognition in Practice: Conceptual Development and Disagreement in Cognitive Science
Robert Brandom (chair), Edouard Machery, Mazviita Chirimuuta, Mark Wilson, James Woodward.
Cognitive science has been beset for thirty years by foundational disputes about the nature and extension of cognition—e.g. whether cognitive processes extend outside the body, and whether plants have them. Previous work on this topic has aimed to settle these questions, but I ask instead how scientists could possibly disagree so much about such fundamental issues. I conclude that cognitive science is currently in a period of major conceptual change. I develop logical tools for characterizing concepts that, like cognition, are in transition, and argue for a novel view of cognitive science that has revisionary consequences for functionalism and our understanding of natural representation. While my dissertation focuses on issues in contemporary cognitive science, the tools I develop have promising applications in other scientific and social contexts.
Cognitive science is the study of a thing called “cognition,” but what do we learn about when we do cognitive science? What, in other words, is the conception of the mind that informs contemporary science? This question is at the heart of fierce disputes that have divided cognitive science since the 1980s—e.g. is cognition like computation? Can it happen outside of brains? Do plants and microbes have cognitive processes? Most philosophers have addressed these disputes by taking sides; I ask instead what structure the concept cognition could have, given that these disputes are so enduring and fundamental. I argue that the disputes are a symptom of conceptual change in cognitive science. As scientific understanding progresses, we revise our use of scientific concepts (like temperature and heredity), entertaining counter-intuitive claims that in the end may be adopted or discarded. Since cognition, the organizing theoretical concept of cognitive science, is in the midst of such a change, I provide a conceptual explication that provides new perspective on entrenched debates, and serves as a common court of appeal that is acceptable to partisans of different theoretical camps.
Drawing on tools from the philosophy of language, I develop new techniques for describing embattled theoretical concepts. Since expert judgments about the extension of cognition vary so much, I suggest that a faithful and illuminating explication should be ecumenical: that is, it should explain the variance in scientists’ judgments, rather than taking sides or treating the variance as noise. Thus, rather than deciding whether e.g. Venus flytraps have cognitive processes like humans do, an ecumenical explication should classify humans as straightforward cases of cognitive systems and flytraps as controversial cases. I do this by identifying parameters, or terms that can be assigned variable interpretations. For example, a parameterized explication might require that cognition be “representational,” and deem flytrap activities as cognitive or non-cognitive depending on how “representation” is interpreted. So the elements of a conceptual analysis function not as necessary or sufficient conditions, but as objects of ongoing controversy. Parameterization provides a unified treatment of an embattled concept by isolating topics of disagreement in a small number of parameters. Parameterization therefore provides clarity without sacrificing analytic rigor or treating open questions as already settled.
I incorporate these innovations into a novel account on which cognition is the “sensitive management of organismal behavior.” The sensitive management hypothesis incorporates three parameters: organism, sensitivity to circumstances, and behavior. I describe various ways these three terms can be interpreted, and argue that the account accurately judges which cases, like the flytrap, are controversial among cognitive scientists. The sensitive management hypothesis does not adjudicate these cases, but it provides the neutral court of appeal that I sought and indicates directions for further work. In particular, I argue that extant treatments of cognition have overemphasized questions about representation and underemphasized questions about the nature of cognitive organisms and behavior.
Nevertheless, the sensitive management hypothesis has several revisionary consequences for philosophy of cognitive science, as well as broader debates in philosophy of science and philosophy of mind. For example, I argue that functionalism owes its attractiveness to a flawed conception of science as seeking laws of nature, which misrepresents cognitive scientific practice. Functionalism is therefore an unreliable court of appeal. In addition, it is an unusual feature of my view that it does not make the notion of representation central for understanding cognition. However, the sensitive management hypothesis can explain several puzzling features of scientific representation-talk, such as the existence of action-oriented representations and the liberality with which scientists ascribe representations. Thus, contrary to the conventional wisdom, it is not the case that understanding representation allows us to understand cognition. Rather, it is cognition that illuminates representation.