Uncovering the Living Body: Bodies and Agents in the Cognitive Sciences

Fred Keijzer

As long as the concept “embodiment” remains tied to the widely applicable notion of “agent,” it will remain insufficiently constrained and kept away from the empirical foundation that living bodies provide for the cognitive sciences.

Usually, we keep our bodies covered. Physically, we wrap them in textiles, and conceptually we hide them under layers of interpretation and meaning. In contrast to the physical ones, the conceptual wrappings are not so easily removed or changed. They consist of enduring encrustations that shape our perceptions and thoughts about our bodies - and ourselves - in more permanent ways. The case that I will address here concerns the concept of “agent” and the way it influences our view of the body within the cognitive sciences.

Starting about 30 years ago - as signaled by Varela, Thompson, and Rosch’s 1991 book The Embodied Mind, and Rodney Brooks’ seminal work on robotics - the cognitive sciences have taken on the notion of embodiment as a central concept: To explain human behavior and cognition, we should not only look at neural processes but also to the body and its interactions with the environment. Given such a general commitment to embodiment, it is important to specify what this commitment entails. The notion of embodied agents is central here.

Embodied Agents

Within the cognitive sciences, embodiment is tied to agents, which in turn are characterized as entities having knowledge and goals in relation to a situation or world in which they can act. The general structure of agent-based explanations incorporates four central ideas. First, such explanations differentiate between a world and an agent that interacts with this world. Second, we have some description of this world, providing us with an ontology. Third, we can reuse this ontology to describe the agent itself in terms of intentional or goal-oriented states that incorporate these descriptions, such as in propositional attitudes. Fourth, we assume the agent applies some degree of rational choice with respect to these intentional states and by doing the same, we can predict the actions of the agent. The resulting explanations take familiar forms, such as “When the woman sees the red light, she will stop her car” and “He said he would make lasagna so he must have gone to the supermarket.” This basic explanatory structure applies to both simple agents with a basic goal structure as well as complex ones with mental states like beliefs and desires.

In such an account, being an agent is central, while embodiment is an additional constraint. Embodiment is tied to an agent rather than being a substantial entity in its own right. Bodies are important because they instantiate an agent, not the other way round. After all, the standard phrase is “embodied agents,” not “agentive bodies.” Agents can be conceived and constituted in many different ways, and embodiment can refer to a human body, a simple robot, or even to the interface of a software agent in a simulated environment. There are no evident material criteria involved except for functional requirements that make the entity an agent.

Without clear material criteria, establishing whether some entity is an agent remains a matter of interpretation and ascription. Daniel Dennett’s famous notion of an intentional stance provides a very influential example of how agent-based explanations work. However, using such an intentional stance and ascribing agency to an entity does not merely come naturally to us. It is actually difficult to avoid and therefore can be thought of as an intentional urge rather than an optional theoretical position that we might or might not apply.

Simple movements often suffice to ascribe agency. Humans get upset when they see a robotic device being kicked, and overall humans have a “tendency to innately, automatically, and spontaneously view a broad variety of different targets as holding goals and mental states” (pp. 117-118) as Mar and Macrae phrased it in their 2007 paper Triggering the Intentional Stance. One can give evolutionary reasons for the ease with which we discern agents, as agents - predators, group members - which tend to be highly significant and should not remain undetected. Susan Carey, in her 2009 book The Origin of Concepts, names agency as one of the

“core cognitions” that constitute innate input analyzers that also “ground the deepest ontological commitments and the most general explanatory principles in terms of which we understand our world” (p. 22). Finally, child psychologists Gergely and Csibra, who study the evolutionary and ontogenetic origins of how we interpret others’ actions in terms of intentional mental states, report that infants of only 12 months (a) can already interpret others’ actions as goal-directed; (b) can evaluate which one of the alternative actions available within the constraints of the situation is the most efficient means to the goal; and (c) expect the agent to perform by the most efficient means available. Taking this all in, ascribing agency is fundamental to our way of looking at the world itself.

Given our tendency to ascribe agency so easily and widely, embodied agents can take many different forms. Embodiment itself need not be specific as long as it provides a material contraption that enables the agent to function. From this perspective, we can -or even should - resist tying human agency too closely to the living human body as this could lead to a form of biochauvinism that excludes the possibility of non-biological agents, such as robots. While our bodies enable us to act as agents, these bodies themselves are only relevant for understanding mind and cognition as far as they subserve this agency. Under this interpretation, the many details of our bodies - such as the presence of livers and muscles - are not considered important for the cognitive sciences. Even the relevance of the biological characteristics of the brain can be questioned, as exemplified by science fiction scenarios where minds can be uploaded to a computer leaving the brain behind. Thus, accepting the importance of embodiment, but interpreting this in terms of embodied agency, keeps out the living human body itself as a starting point for the cognitive sciences.

Ascribing agency and intentionality is a basic feature of human minds and comes with a cluster of ideas concerning the limited relevance of our living bodies. However, given our susceptibility to the intentional urge and the intuitive force of the conceptualizations that stem from it, we must consider the possibility that the very salience of agentive interpretations keeps the potentially much wider cognitive relevance of living bodies in the background.

Living Bodies

One reason to be wary about the impact of our intentional urge comes from the rapidly expanding study of cognitively relevant phenomena in a very broad range of living systems. This expanded scope starts with bacteria, but also involves plants, fungi, and both invertebrate and vertebrate animals. Excluding a few exceptions such as humans, these organisms do not have minds nor do they constitute smart agents in any obvious way. Using our intentional urge, this easily disqualifies these cases as interesting targets for the cognitive sciences. However, when studied in more detail, all organisms exhibit behavioral and growth-related decision-making and fit cognitive criteria such as perception, action, memory, and valuing (see Lyon, 2015, for bacteria and Baluska & Levin, 2016, for the general case). Whether we should use the word “cognition” for this broad set of widely dispersed phenomena is a matter of debate, but it is evident that these atypical cases are highly relevant for the cognitive sciences as an explanatory target. In this context, the concept of “agent” becomes less important as a guideline forjudging cognitive relevance. Being a living system provides a more certain criterion when these atypical cases are taken into account.

The need to look at living systems more widely provides further reasons for taking our own living bodies seriously as an important target for the cognitive sciences, and even as an empirical condition for agency in the first place. Basic life forms consist of single cells, and are wonders of natural nanotechnology, which enables unicellulars to construct themselves, metabolize, multiply, adapt, change shape, and do all the things that enable them to thrive. Importantly, unicellulars are small. They are usually not visible to the naked eye and live in microscopic worlds. As our own ancestors started out as such unicellulars, it follows that our current perceptual access to a macroscopic world is not a basic situation but the result of major evolutionary changes that took place when our ancestors became multicellular animals.

Becoming multicellular was a major change where (eventually) billions of separate cells came to sense and act as a macroscopic individual. These cells need to be coordinated in many ways, such as for body organization, physiology, and behavior. The evolution of nervous systems must have played a major role in the transition towards multicellular organizations capable of sensing and acting as a collective and, in this way, for the first time becoming capable of accessing our own familiar macroscopic world - the world of extended surfaces and things in it as James Gibson nicely described it in his 1979 book The Ecological Approach to Visual Perception. How this transition took place and how it bears on our understanding of nervous systems are key questions for the cognitive sciences, as they turn on explaining how animals turned into the agents that we are so familiar with.

One intriguing idea is that nervous systems originally were mostly involved in controlling the body itself, enabling it to move in a coordinated way. The resulting motile body could then have acted as a multicellular unit capable of sensing in a way that was dispersed across many cells and so enabling perceptual access of the macroscopic world (see Keijzer, 2015, for more details and references). If this is on the right track, it ties the ontology of the perceptual and behavioral world to the makeup of a living body. How we should interpret this option remains future work, but we will need to reassess many assumptions that seem clear and fundamental from our own human perspective, which is itself tied to such a multicellular organization.

Our living animal bodies incorporate many characteristics that may be crucial for understanding our minds and cognition. The concept of “agent” provides a skewed perspective on our own living bodies that backgrounds much that is relevant. Using the concept of (embodied) agents as a starting point for the cognitive sciences ignores not only a wide array of fundamental questions about the origins and nature of such agents, but also keeps these questions invisible by downplaying the crucial role of the living body.

Lived Bodies

One final issue must be addressed. Within phenomenology and en-activism (two related philosophical approaches that both focus on conscious experience), the concept of a living body is often used as a contrast to the lived body (see, for example, Evan Thompson’s 2007 book, Mind in Life). The lived body refers here to our first-person experience of our own bodies and aims to describe it from this perspective. Advocates for this perspective argue that study of the living body is limited to a third-person scientific perspective on the body that is associated with a reductive view that neglects experience.

There are various reasons for being unhappy with this restricted view of the living body that sees it as disconnected from first-person experience, which only surfaces in the “lived body.” Foremost, highlighting the experiential lived body brings us back to high-level agentive conceptualizations of the body that fit in with an agentive interpretation. Such a view easily remains oblivious to the underlying complexity and organization of our own bodies that can only be accessed by systematic, empirical, and socially shared scientific work. Another reason concerns the critique that scientific work on the living body involves a reductive view that is antithetical to conscious experience. Although such a reductive tendency has indeed been present in a lot of scientific work, by now the need to explain consciousness as we all experience it firsthand has become part of the scientific enterprise (see Ginsburg & Jablonka, 2019, for an overview). When the cognitive sciences start to pay attention to the living body in earnest, this inquiry must also involve the question of how a first-person perspective and the experience associated with it comes into being in the natural world. We do not need a dualism between living and lived bodies.

For Sources and Further Reading

Baluska, F., & Levin, M. (2016). On having no head: Cognition throughout biological systems. Frontiers in Psychology, 7,902.

Csibra, G., & Gergely, G. (2013). Teleological understanding of actions. In M. R. Banaji & S. A. Gelman (Eds.), Navigating the social world: What infants, children, and other species can teach us (pp. 38-43). New York: Oxford University Press.

Ginsburg, S., & Jablonka, E. (2019). The evolution of the sensitive soul: Learning and the origins of consciousness. Cambridge: MIT Press.

Keijzer, F. A. (2015). Moving and sensing without input and output: Early nervous systems and the origins of the animal sensorimotor organization. Biology & Philosophy, 30(3). 311-331.

Lyon, P. (2015). The cognitive cell: Bacterial behavior reconsidered. Frontiers in Microbiology, 6, 264.