Skip to main content
Social Sci LibreTexts

5.2: Abandoning Methodological Solipsism

  • Page ID
    21229
  • The goal of Cartesian philosophy was to provide a core of incontestable truths to serve as an anchor for knowledge (Descartes, 1960, 1996). Descartes believed that he had achieved this goal. However, the cost of this accomplishment was a fundamental separation between mind and body. Cartesian dualism disembodied the mind, because Descartes held that the mind’s existence was independent of the existence of the body.

    I am not that structure of limbs which is called a human body, I am not even some thin vapor which permeates the limbs—a wind, fire, air, breath, or whatever I depict in my imagination, for these are things which I have supposed to be nothing. (Descartes, 1996, p. 18)

    Cartesian dualism permeates a great deal of theorizing about the nature of mind and self, particularly in our current age of information technology. One such theory is posthumanism (Dewdney, 1998; Hayles, 1999). Posthumanism results when the content of information is more important than the physical medium in which it is represented, when consciousness is considered to be epiphenomenal, and when the human body is simply a prosthetic. Posthumanism is rooted in the pioneering work of cybernetics (Ashby, 1956, 1960; MacKay, 1969; Wiener, 1948), and is sympathetic to such futuristic views as uploading our minds into silicon bodies (Kurzweil, 1999, 2005; Moravec, 1988, 1999), because, in this view, the nature of the body is irrelevant to the nature of the mind. Hayles uncomfortably notes that a major implication of posthumanism is its “systematic devaluation of materiality and embodiment” (Hayles, 1999, p. 48); “because we are essentially information, we can do away with the body” (Hayles, 1999, p. 12).

    Some would argue that similar ideas pervade classical cognitive science. American psychologist Sylvia Scribner wrote that cognitive science “is haunted by a metaphysical spectre. The spectre goes by the familiar name of Cartesian dualism, which, in spite of its age, continues to cast a shadow over inquiries into the nature of human nature” (Scribner & Tobach, 1997, p. 308).

    In Chapter 3 we observed that classical cognitive science departed from the Cartesian approach by seeking materialist explanations of cognition. Why then should it be haunted by dualism?

    To answer this question, we examine how classical cognitive science explains, for instance, how a single agent produces different behaviors. Because classical cognitive science appeals to the representational theory of mind (Pylyshyn, 1984), it must claim that different behaviors must ultimately be rooted in different mental representations.

    If different behaviors are caused by differences between representations, then classical cognitive science must be able to distinguish or individuate representational states. How is this done? The typical position adopted by classical cognitive science is called methodological solipsism (Fodor, 1980). Methodological solipsism individuates representational states only in terms of their relations to other representational states. Relations of the states to the external world—the agent’s environment—are not considered. “Methodological solipsism in psychology is the view that psychological states should be construed without reference to anything beyond the boundary of the individual who has those states” (Wilson, 2004, p. 77).

    The methodological solipsism that accompanies the representational theory of mind is an example of the classical sandwich (Hurley, 2001). The classical sandwich is the view that links between a cognitive agent’s perceptions and a cognitive agent’s actions must be mediated by internal thinking or planning. In the classical sandwich, models of cognition take the form of sense-think-act cycles (Brooks, 1999; Clark, 1997; Pfeifer & Scheier, 1999). Furthermore, these theories tend to place a strong emphasis on the purely mental part of cognition—the thinking—and at the same time strongly de-emphasize the physical—the action. In the classical sandwich, perception, thinking, and action are separate and unequal.

    On this traditional view, the mind passively receives sensory input from its environment, structures that input in cognition, and then marries the products of cognition to action in a peculiar sort of shotgun wedding. Action is a by-product of genuinely mental activity. (Hurley, 2001, p. 11)

    Although connectionist cognitive science is a reaction against classical cognitivism, this reaction does not include a rejection of the separation of perception and action via internal representation. Artificial neural networks typically have undeveloped models of perception (i.e., input unit encodings) and action (i.e., output unit encodings), and in modern networks communication between the two must be moderated by representational layers of hidden units.

    Highly artificial choices of input and output representations and poor choices of problem domains have, I believe, robbed the neural network revolution of some of its initial momentum. . . . The worry is, in essence, that a good deal of the research on artificial neural networks leaned too heavily on a rather classical conception of the nature of the problems. (Clark, 1997, p. 58)

    The purpose of this chapter is to introduce embodied cognitive science, a fairly modern reaction against classical cognitive science. This approach is an explicit rejection of methodological solipsism. Embodied cognitive scientists argue that a cognitive theory must include an agent’s environment as well as the agent’s experience of that environment (Agre, 1997; Chemero, 2009; Clancey, 1997; Clark, 1997; Dawson, Dupuis, & Wilson, 2010; Dourish, 2001; Gibbs, 2006; Johnson, 2007; Menary, 2008; Pfeifer & Scheier, 1999; Shapiro, 2011; Varela, Thompson, & Rosch, 1991). They recognize that this experience depends on how the environment is sensed, which is situation; that an agent’s situation depends upon its physical nature, which is embodiment; and that an embodied agent can act upon and change its environment (Webb & Consi, 2001). The embodied approach replaces the notion that cognition is representation with the notion that cognition is the control of actions upon the environment. As such, it can also be viewed as a reaction against a great deal of connectionist cognitive science.

    In embodied cognitive science, the environment contributes in such a significant way to cognitive processing that some would argue that an agent’s mind has leaked into the world (Clark, 1997; Hutchins, 1995; Menary, 2008, 2010; Noë, 2009; Wilson, 2004). For example, research in behaviour-based robotics eliminates resource-consuming representations of the world by letting the world serve as its own representation, one that can be accessed by a situated agent (Brooks, 1999). This robotics tradition has also shown that nonlinear interactions between an embodied agent and its environment can produce surprisingly complex behavior, even when the internal components of an agent are exceedingly simple (Braitenberg, 1984; Grey Walter, 1950a, 1950b, 1951, 1963; Webb & Consi, 2001).

    In short, embodied cognitive scientists argue that classical cognitive science’s reliance on methodological solipsism—its Cartesian view of the disembodied mind—is a deep-seated error. “Classical rule-and-symbol-based AI may have made a fundamental error, mistaking the cognitive profile of the agent plus the environment for the cognitive profile of the naked brain” (Clark, 1997, p. 61).

    In reacting against classical cognitive science, the embodied approach takes seriously the idea that Simon’s (1969) parable of the ant might also be applicable to human cognition: “A man, viewed as a behaving system, is quite simple. The apparent complexity of his behavior over time is largely a reflection of the complexity of the environment in which he finds himself” (p. 25). However, when it comes to specifics about applying such insight, embodied cognitive science is frustratingly fractured. “Embodied cognition, at this stage in its very brief history, is better considered a research program than a well-defined theory” (Shapiro, 2011, p. 2). Shapiro (2011) went on to note that this is because embodied cognitive science “exhibits much greater latitude in its subject matter, ontological commitment, and methodology than does standard cognitive science” (p. 2).

    Shapiro (2011) distinguished three key themes that are present, often to differing degrees, in a variety of theories that belong to embodied cognitive science. The first of Shapiro’s themes is conceptualization. According to this theme, the concepts that an agent requires to interact with its environment depend on the form of the agent’s body. If different agents have different bodies, then their understanding or engagement with the world will differ as well. We explore the theme of conceptualization later in this chapter, in the discussion of concepts such as umwelten, affordances, and enactive perception.

    Shapiro’s (2011) second theme of embodied cognitive science is replacement: “An organism’s body in interaction with its environment replaces the need for representational processes thought to have been at the core of cognition” (p. 4). The theme of replacement is central to the idea of cognitive scaffolding, in which agents exploit environmental resources for problem representation and solution.

    The biological brain takes all the help it can get. This help includes the use of external physical structures (both natural and artifactual), the use of language and cultural institutions, and the extensive use of other agents. (Clark, 1997, p. 80)

    Shapiro’s (2011) third theme of embodied cognitive science is constitution. According to this theme, the body or the world has more than a causal role in cognition—they are literally constituents of cognitive processing. The constitution hypothesis leads to one of the more interesting and radical proposals from embodied cognitive science, the extended mind. According to this hypothesis, which flies in the face of the Cartesian mind, the boundary of the mind is not the skin or the skull (Clark, 1997, p. 53): “Mind is a leaky organ, forever escaping its ‘natural’ confines and mingling shamelessly with body and with world.”

    One reason that Shapiro (2011) argued that embodied cognitive science is not a well-defined theory, but is instead a more ambiguous research program, is because these different themes are endorsed to different degrees by different embodied cognitive scientists. For example, consider the replacement hypothesis. On the one hand, some researchers, such as behaviour-based roboticists (Brooks, 1999) or radical embodied cognitive scientists (Chemero, 2009), are strongly anti-representational; their aim is to use embodied insights to expunge representational issues from cognitive science. On the other hand, some other researchers, such as philosopher Andy Clark (1997), have a more moderate view in which both representational and non-representational forms of cognition might be present in the same agent.

    Shapiro’s (2011) three themes of conceptualization, replacement, and constitution characterize important principles that are the concern of the embodied approach. These principles also have important effects on the practice of embodied cognitive science. Because of their concern with environmental contributions to behavioral complexity, embodied cognitive scientists are much more likely to practice forward engineering or synthetic psychology (Braitenberg, 1984; Dawson, 2004; Dawson, Dupuis, & Wilson, 2010; Pfeifer & Scheier, 1999). In this approach, devices are first constructed and placed in an environment, to examine what complicated or surprising behaviors might emerge. Thus while in reverse engineering behavioral observations are the source of models, in forward engineering models are the source of behavior to observe. Because of their concern about how engagement with the world is dependent upon the physical nature and abilities of agents, embodied cognitive scientists actively explore the role that embodiment plays in cognition. For instance, their growing interest in humanoid robots is motivated by the realization that human intelligence and development require human form (Breazeal, 2002; Brooks et al., 1999).

    In the current chapter we introduce some of the key elements that characterize embodied cognitive science. These ideas are presented in the context of reactions against classical cognitive science in order to highlight their innovative nature. However, it is important to keep potential similarities between embodied cognitive science and the other two approaches in mind; while they are not emphasized here, the possibility of such similarities is a central theme of Part II of this book.

    • Was this article helpful?