We, like many animals, are conscious of our environment, both physical and social. We think about and experience the objects and people that the world contains. But at the same time, those very thoughts and experiences can themselves become the objects of conscious thought and experience, so that we can think critically about them. When this happens, we are often thinking about ourselves—holding ourselves to account by evaluating our own reasoning, for example. According to a long philosophical tradition, rationality and self-consciousness go hand in hand. A rational thinker is one who constantly monitors her own thoughts and updates her beliefs in the light of changing evidence in order to ensure consistency and to increase the likelihood that they will track the truth.
Moreover, it is because we are aware of ourselves as temporally extended beings with a past and a future that we are able to develop a narrative understanding of our lives. The narrative we tell of ourselves can incorporate an interpretation of our personal history. This personal history, in turn, can inform a forward-looking sense of the kind of person we strive to be. That forward-looking sense of how we want to turn out sets the agenda for our plans and projects. It also sets the agenda for how we navigate the social world, for how we engage with other people and with institutions.
The fundamental source for these (and many other) abilities and achievements is our capacity for self-consciousness, or self-awareness. This capacity is at the heart of much of what is typically taken to be distinctively human. Some aspects of self-consciousness have been studied extensively. Every student of philosophy will be familiar, for example, with Descartes’s cogito ergo sum. Most will have encountered Kant’s guiding idea, in the Critique of Pure Reason, that self-consciousness and consciousness of the physical world are inextricably linked.
For Descartes, Kant, and many other philosophers, self-consciousness is a phenomenon closely tied to language. Kant famously wrote of “the ‘I think’ that accompanies all my representations.” To be self-conscious for Kant (and again, for many others) is to be capable of thinking about oneself in a special way. He took that special way of thinking of oneself to be coeval with the ability to refer to oneself using the first person pronoun “I” (or its equivalent in other languages). And this is not surprising. The ability to think about oneself and the ability to refer to oneself seem inextricably linked. After all, how can one think about oneself without referring to oneself in thought? And, many have thought, referring to oneself in thought is really just an internalized form of how one refers to oneself in public language. It is no accident that self-conscious thoughts are often referred to as “I”-thoughts.
The close connections between self-conscious thought and linguistic self-reference raise many fascinating and important questions. I explore some of them in my recent book Understanding “I”: Language and Thought (Bermúdez 2017). However, they are not the focus of the essays in this volume. In one form or another, all of the essays collected here explore different dimensions of a single basic idea. This guiding idea is that the rich and sophisticated forms of self-consciousness with which we are most familiar (not just as philosophers or psychologists, but also as ordinary, reflective individuals) rest on a complex underpinning that has largely been invisible to students of the self and of self-consciousness.
Full-fledged linguistic self-consciousness emerges from multiple layers of more primitive forms of self-consciousness, and, even when linguistic self-consciousness is fully operational, these more primitive forms of self-consciousness persist in ways that structure and frame self-conscious language and thought. Moreover, and unlike linguistic self-consciousness, these primitive forms of self-consciousness extend widely throughout the animal kingdom. Some are present in human infants from the earliest moments outside the womb.
The essays in this volume focus on three primitive forms of self-consciousness in particular:
modes of awareness of one’s spatial orientation and trajectory built into outward-directed perception,
modes of awareness of how one’s body is disposed that are derived from awareness of our own bodies, and
the complex spatial self-awareness implicated in navigating through the environment.
The object of these forms of self-awareness is primarily the embodied self. Hence the title of this volume: The Bodily Self.
The distinction between conceptual and nonconceptual content offers a way of understanding some of the significant differences between full-fledged linguistic self-consciousness, on the one hand, and these primitive forms of self-consciousness, on the other. The issue of content is important because the three primitive forms of self-consciousness all yield representations of the bodily self (in the context, usually, of representations of the physical/social environment). Whereas the content of linguistic self-consciousness is conceptual, that of primitive self-consciousness is nonconceptual. Precisely characterizing the differences between conceptual and nonconceptual content is a complicated undertaking (see Bermúdez and Cahen 2015 for an overview of relevant debates), but a rough characterization will be adequate here.
A representational state has conceptual content just if it can only be truly attributed to a thinker who possesses the concepts required to specify the relevant content. So, for example, my belief that Bishkek is the capital of Kyrgyzstan has conceptual content because my ability to entertain it requires me to possess all the relevant concepts (Bishkek, Kyrghyzstan, and —— is the capital of ——). Someone lacking even one of those concepts would not be able to have a belief with that content. Representational states have nonconceptual content, on the other hand, when that requirement of concept possession does not hold. So, to be in a state with nonconceptual content is to be in a state that one need not be in a position to conceptualize or articulate—because one lacks some or all of the relevant concepts. In fact, as I understand the notion of nonconceptual content, it is perfectly possible for a creature possessing no concepts whatsoever to be in states that have nonconceptual content.1
The essays that follow explore three different, but of course interrelated, dimensions of these forms of nonconceptual self-consciousness.
First, I make the case that it is indeed correct to describe them as forms of self-consciousness. My principal strategy for doing this is to argue that they share certain structural and epistemological features with full-fledged linguistic self-consciousness. Several of the essays discuss, for example, how forms of nonconceptual self-consciousness give rise to judgments that have the important property of being immune to error through misidentification relative to the first-person pronoun.
Second, I try to give accounts of the content of certain important classes of states of nonconceptual self-consciousness that elucidate how they represent the self. In this context I focus primarily on the self-specifying dimension of visual perception and on the content of bodily awareness.
Third, I explore the role of nonconceptual self-consciousness more generally in our cognitive and affective lives. One topic here is the role nonconceptual self-consciousness plays in epistemically grounding fully conceptual first-person judgments. A theme that recurs in several essays is the relation between nonconceptual awareness of our bodies and what many theorists call our “sense of ownership” for our own bodies.
The remainder of this introduction explains how these themes are developed in the individual essays.
1 The Paradox of Self-Consciousness
The intellectual framework for these essays is set in part by my first book The Paradox of Self-Consciousness (Bermúdez 1998). The paradox referenced in the title comes from taking self-consciousness and self-reference to be interdependent. It is natural (and common) to think of self-conscious thoughts as thoughts that are typically expressed through the first-person pronoun “I.” What is distinctive about self-conscious thought is often expressed by contrasting how one might talk about oneself using a linguistic device of intentional self-reference, such as “I,” as opposed to a third-person linguistic device, such as one’s name. Philosophers such as Héctor-Neri Castañeda and John Perry have emphasized that third-person self-reference always leaves open the possibility of errors of self-identification. It seems to be a defining feature of self-conscious thought that it knowingly and intentionally be about oneself. And yet, one might not realize that one is in fact referring to oneself when one uses a device other than the first-person pronoun or its equivalent—a dramatic example being an amnesiac who has forgotten his own name. So it is natural to think that the capacity for self-conscious thought depends on the capacity for self-reference using “I.”
At the same time, however, as Elizabeth Anscombe observed in her famous article “The First Person” (Anscombe 1975), it seems impossible to characterize the reference rule for the first person without building into it the requirement that the self-referrer be self-consciously thinking of herself. After all, it is not enough to say that “I” is the pronoun that a speaker uses to refer to herself, because that leaves open the possibility of an error of self-identification (not knowing that one is actually referring to oneself). To rule out the possibility of an error of misidentification, “I” must be used knowingly and intentionally as a device of self-reference. But to intend to use the pronoun “I” to refer to oneself is of course itself a form of self-conscious thought.
In The Paradox of Self-Consciousness, I suggested that this interdependence between self-conscious thought and self-reference creates two forms of circularity. The first type (explanatory circularity) arises because neither capacity can be explained in terms of the other. The second type of circularity (capacity circularity) arises because this explanatory circularity seems to rule out the possibility of explaining how either the capacity for self-conscious thought or linguistic mastery of the first-person pronoun arises in the normal course of human development. It does not seem possible to meet the following constraint:
TheAcquisition Constraint If a given psychological capacity is psychologically real, then there must be an explanation of how it is possible for an individual in the normal course of human development to acquire that capacity.
Neither self-conscious thought nor linguistic mastery of the first-person pronoun is innate, and yet each presupposes the other in a way that seems to imply that neither can be acquired unless the other capacity is already in place. I termed these two types of circularity collectively the paradox of self-consciousness.
The resolution I proposed to the paradox of self-consciousness is sketched out in chapter 1, “Nonconceptual Self-Consciousness and Cognitive Science,” originally published in 2001 in the journal Synthese. Chapter 1 summarizes the overall argument of The Paradox of Self-Consciousness and identifies key points of contact with research programs in scientific psychology and cognitive science. As it makes clear, the basic strategy for defusing both the explanatory and capacity circularities is to show how full-fledged linguistic self-consciousness is grounded in forms of nonconceptual self-consciousness that do not presuppose linguistic mastery of the first-person pronoun.
Full-fledged self-consciousness is conceptual. It presupposes mastery of the first-person pronoun and, correlatively, of the first-person concept. But neither is presupposed by the nonconceptual forms of self-consciousness. That is the key to escaping the paradox of self-consciousness.
These forms of nonconceptual self-consciousness are more primitive than linguistic self-consciousness in three ways—theoretically, ontogenetically, and phylogenetically. The first of these opens up the possibility of analyzing linguistic self-consciousness in terms of the various forms of nonconceptual self-consciousness. The second and third open up the possibility of explaining the emergence of linguistic self-consciousness out of nonconceptual self-consciousness, both (ontogenetically) in the normal course of human development and (phylogenetically) in evolutionary development.
In addition to the three forms of nonconceptual self-consciousness identified earlier, The Paradox of Self-Consciousness discussed the psychological self-awareness that emerges from joint attention and other primitive forms of interpersonal interaction. As outlined in chapter 1, this psychological self-awareness is important in explaining mastery of the first-person pronoun. Circularity is averted, I claim, because the communicative intentions that make it the case that “I” is used knowingly and intentionally as a device of self-reference implicate nonconceptual psychological self-awareness, rather than any form of conceptual self-awareness. The remaining essays in this volume do not develop this analysis further, focusing instead on forms of self-awareness implicated in visual proprioception, self–world dualism, and bodily awareness.
2 Nonconceptual Self-Consciousness in Visual Perception
The starting point for chapters 2 through 4 is the ecological theory of perception proposed by the psychologist J. J. Gibson, particularly in his two books The Senses Considered as Perceptual Systems (Gibson 1966) and The Ecological Approach to Visual Perception (Gibson 1979). Gibson’s ecological approach to visual perception is almost completely unknown to philosophers, and has had less influence than I believe it deserves within scientific psychology and cognitive science. Yet, as I illustrate in these three essays, Gibson’s work points to one of the most basic forms of nonconceptual self-consciousness.
Part of the reason for Gibson’s relative neglect is that he very polemically set himself against the dominant information-processing paradigm at the heart of contemporary cognitive science (preferring instead to speak of perceptual systems “resonating” to the environment and picking up information directly). This aspect of his thinking, however, is in many ways orthogonal to his insights into the nature of visual perception (and indeed the other sensory modalities). In particular, it seems independent of the aspect of his thinking that is most important for thinking about self-consciousness, namely, his insight that all forms of perception contain both propriospecific information about the self and exterospecific information about the environment.
For Gibson, the starting point for thinking about visual perception (which was the modality on which he primarily focused) is the flow of movement in the ambient optic array—what he termed optic flow. The basic constituents of the ambient optic array are visual solid angles—angles with their apex at the eye and their base at a perceived surface. Optic flow is, in essence, the changing patterns of visual solid angles, as surfaces move relative to each other, occlude each other, and so forth. The optic flow contains both variant and invariant features, and is constantly changing as the perceiver moves through the environment (which can itself be moving). Because of this, as Gibson himself puts it, “information about the self accompanies information about the environment, and the two are inseparable” (Gibson 1979, 116).
Gibson identifies three principal categories of propriospecific (or: self-specifying) information in the optic flow. These include:
Self-specifying structural invariants
These are features of the bodily self that directly structure the optical array (such as the nose, which has the highest value of motion parallax of all directly perceived objects) as well as bodily extremities that have distinctive visual properties (such as subtending visual solid angles that cannot be reduced below a certain minimum).
Visual kinesthesis
This refers to information about the perceiver’s relative motion that is available in the optic array. The flow of the ambient array originates from the aiming point of locomotion, for example.
Affordances
According to the ecological approach, the perception of objects is not neutral. Instead, the optical array carries information about the possibilities that surfaces and objects afford the organism. A surface might be seen as affording support, for example, or an object as being manipulable or edible. Affordances carry information about the behavioral possibilities open to an organism.
In these respects, then, Gibson’s concept of ecological perception reveals a basic awareness of the bodily self that can serve as the core of a comprehensive account of full-fledged self-consciousness in thought and action. On the ecological understanding of perception, sensitivity to self-specifying information is built into the very structure of perception in such a way that, as Gibson famously put it, all perception involves coperception of the (bodily) self and the environment. Nonetheless, his suggestive analysis opens up (at least) two obvious questions. How do we scale up from this type of self-perception (or rather, these types of self-specifying information) to full-fledged self-consciousness? What role, or roles, do they play in conceptual thought about ourselves and about the world? The first of these questions is addressed in chapter 2, and the second in chapters 3 and 4.
Chapter 2, “Ecological Perception and the Notion of a Nonconceptual Point of View,” was first published in 1995 in The Body and the Self, which I coedited with Anthony J. Marcel and Naomi Eilan. This essay offers a way of bridging the gap between the ecological coperception of self and environment, on the one hand, and full-fledged self-consciousness, on the other. The link between the two follows from the notion of a point of view on the world, an idea initially developed by Peter Strawson in The Bounds of Sense (Strawson 1975), his commentary on Kant’s Critique of Pure Reason. One of the guiding ideas of Kant’s Critique is that self-consciousness and consciousness of an external world are inextricably linked. Strawson’s distinctive gloss on Kant’s interdependence thesis depends crucially on the notion of a point of view, which he sees as a necessary (but not sufficient) element in full-fledged self-consciousness. For Strawson, a creature can count as a subject of experiences only if it can draw certain basic distinctions—most fundamentally, a distinction between its experiences and what those experiences are of. A genuine subject of experience must, moreover, be able to draw this distinction not just at a time, but also over time. Being able to do this is, in Strawson’s phrase, possessing a point of view on the world.
For Strawson and Kant, having a point of view is a highly sophisticated cognitive achievement, requiring the ability to ascribe experiences to oneself as well as the ability to conceptualize the basic spatiotemporal and causal structure of the world. Chapter 2 develops a more pared down notion of a nonconceptual point of view. Key here is the idea that an awareness of the distinction between self and nonself can emerge from the ability to engage in spatial reasoning. In particular, spatial reasoning depends on the interplay of two cognitive capacities, one self-directed and one outward-directed. The self-directed capacity is the awareness of one’s own agency—of one’s ability to bring about changes in the world. The outward-directed capacity is the ability to recognize and reidentify places over time. Both of these capacities can exist in a nonconceptual form, I argue, and possessing both of them yields a nonconceptual point of view on the world.
The general theme of the relation between nonconceptual self-consciousness and full-fledged linguistic self-consciousness is pursued further in chapter 3, “Sources of Self-Consciousness,” first published in the Proceedings of the Aristotelian Society in 2002. The starting point for this essay is that we can think about the sources of linguistic self-consciousness in two different ways. On the one hand, we can take a genetic perspective. That is, we can think about the origins of the capacity for thinking and articulating self-conscious thoughts—about how this capacity emerges in normal human development, or how it emerged in the course of human evolution. Alternatively, we can take an epistemic perspective on the sources of self-consciousness. To take an epistemic perspective is to ask not where those judgments come from, but rather what the warrant for them is, or how they are justified.
Within philosophy, these two types of question are typically kept completely separate, with only questions of the second, epistemic variety held to count as genuinely philosophical, on pain of committing some version of the genetic fallacy, or fallacy of origins. A key claim of chapter 3 (and, in fact, of my work more generally) is that this approach can be, and often is, a mistake. In that spirit, the essay explores the role that the genetic dimension of self-consciousness plays in understanding the epistemology of self-consciousness. I take the representation of the bodily self in “ordinary” visual perception as a paradigm example of a genetic source of (full-fledged) self-consciousness, since it seems to be primitive and foundational from the perspective both of ontogeny and of phylogeny. I claim that these primitive foundations from which self-consciousness emerges in the course of cognitive development are also the foundation for the epistemic status of full-fledged self-conscious thoughts.
My foil in the essay is Christopher Peacocke’s account, in his book Being Known (Peacocke 1999), of the epistemic dimension of a particular type of self-conscious judgment. Peacocke distinguishes two types of self-conscious judgment—those that are, in his terminology, representation dependent and those that are representation independent. A representation-dependent judgment is one that involves taking a first-person content at face value. Perceptual states can have first-person contents. I might see that I am in front of the football stadium, for example. If I take that visual content at face value and on that basis judge that I am in front of the football stadium, then my judgment would count as representation dependent. For self-conscious judgments that are representation independent, however, there are no such corresponding first-person contents.
The distinction is a subtle one. To go back to the original example, I might see that I am in front of the football stadium and, by taking that content at face value, judge that I am in front of the football stadium. The corresponding first-person content is my seeing that I am in front of the football stadium. Here we have a case of representation-dependence. At the same time, though, I might judge that I am seeing the football stadium. Here the basis for my judgment is the same—namely, my knowing that I am in front of the football stadium and looking at it. But the judgment does not count as representation dependent, because it does not involve taking my visual state at face value. The fact that I am in front of the stadium is part of what I see, which is why it counts as a first-person content. But the fact that I am seeing the football stadium is not part of what I see, and so is not available to be taken at face value.
Clearly, an important question here is what “taking at face value” means. At one extreme, if a perceptual belief takes a perceptual content at face value, then that requires the content of the belief to be exactly the same as the content of the perception. Since belief contents are standardly taken to be conceptual, perceptual states would also have to be conceptual.2 Peacocke himself is neutral on this question.3 But (as one of the principal theorists of nonconceptual content) he certainly wants to leave open the possibility that first-person judgments can involve taking nonconceptual first-person contents at face value. Chapter 3 develops this possibility, arguing that properly understanding both the first-person content and the content of perception requires a nonconceptual model of how the self is represented in perception. The self cannot be represented conceptually in perception because the first-person concept lacks the perspectival dimension built into perceptual self-representation. Instead, I argue, a broadly Gibsonian account of how the bodily self is represented in visual perception better captures the epistemic dimension of representation-dependent first-person judgments.
Both representation-dependent and representation-independent judgments can have the important property of being immune to error through misidentification relative to the first-person pronoun (henceforth: the IEM property). A judgment “I am F” has the IEM property just in case it is not possible to be mistaken about whom it is one is judging to be F. To continue with the earlier example, if I judge that I am in front of the football stadium because I see the football stadium in front of me, then I cannot be mistaken about whom it is that I am taking to be in front of the football stadium. I can of course be mistaken about every other aspect of the judgment. I might be behind the football stadium, not in front of it, for example, or it might not even be a football stadium at all. But it would make no sense for me to ask: “Someone is in front of the football stadium, but is it me?”
Judgments either have or lack the IEM property relative to the grounds on which they are made. My judgment about being in front of the football stadium could well lack the IEM property if it were made on different grounds. Suppose, for example, that I am trying to locate myself on a GPS map. I see a blue dot in front of the football stadium and mistakenly think that it corresponds to my GPS tracker, while actually it is someone else’s tracker and mine is malfunctioning. Then I have misidentified the person who is in front of the football stadium. The grounds on which I made my judgment do not confer the IEM property upon it.
So, can we say anything specific about the types of ground that do confer the IEM property upon self-conscious judgments? One thing is clear. A judgment with the IEM property cannot involve any identification of the self. So, it must be derived from sources of information that can only provide information about the self (so that no identification is required). In other words, judgments with the IEM property must ultimately be based on what might be termed identification-free self-awareness. The forms of nonconceptual self-consciousness that we have been discussing all fit this description.4 Self-specifying information in visual perception can only be about the perceiving self. Likewise for bodily self-awareness and the self-awareness implicated in spatial reasoning.
The relations between nonconceptual self-consciousness and the immunity property are the focus of chapter 4, “The Elusiveness Thesis, Immunity to Error through Misidentification, and Privileged Access,” originally published in 2003 in a volume entitled Privileged Access: Philosophical Accounts of Self-Knowledge edited by Brie Gertler. The essay explores two ideas that have played a prominent role in philosophical discussions of self-knowledge. The first is the idea that we enjoy introspective ways of finding out about ourselves that are fundamentally different from our ways of finding out about ordinary physical objects and other psychological subjects. These ways of finding out about ourselves yield a certain type of privileged access to our own properties and states that we do not have to the properties and states of other people (and nor do they have it to our own properties and states). The second is an idea most often associated with David Hume. According to the so-called elusiveness thesis, when we find out about our own properties through introspection we are not acquainted with any object whose properties they are. It is natural to think that these two ideas are related—and, in particular, that it is (at least partly) because we do not encounter the self as an object in introspection that the knowledge of the self gained through introspection is epistemically privileged. In other words, the elusiveness thesis explains the IEM property. This idea has been emphasized by Sydney Shoemaker, who bears primary responsibility for bringing the IEM property to the center of philosophical discussions of self-awareness (see the essays reprinted in Shoemaker 1996).
The forms of nonconceptual self-consciousness that we have been considering, however, show the limitations of the elusiveness thesis as an explanation of identification-free self-awareness. Simply put, the bodily self just is directly encountered in nonconceptual self-consciousness. This means that we need to provide an alternative account of the warrant for self-conscious judgments with the IEM. This account will have two components. First, we need an explanation of how and why these self-conscious judgments are identification free (and hence of why they have the IEM property). At the same time, though, we also need to explain how the thinker is justified in trading on identification-free information sources. Where φ is some property whose presence is revealed by a form of nonconceptual self-consciousness, what justifies the immediate transition from one’s awareness of φ to the self-conscious judgment that one is oneself φ? Chapter 4 offers an answer to that question for two forms of nonconceptual self-consciousness—bodily awareness, and the self-specifying dimension of ordinary visual perception.
3 Nonconceptual Self-Consciousness in Bodily Awareness
Whereas chapters 2–4 focus primarily on nonconceptual self-consciousness in visual perception, chapters 5–9 turn to the complex phenomenon of bodily awareness.
In a sense, of course, visual proprioception is a form of bodily awareness, since it incorporates awareness of the bodily self. The principal emphasis in these chapters, however, is on what is often termed awareness of one’s body “from the inside.” This awareness of the body from the inside comes from multiple information sources. Some of these information sources are nonconscious. Others are conscious. Of the conscious varieties, some are conceptual and others nonconceptual. Figure 6.2 (reproduced here) offers an overview of the typology of bodily awareness. Bodily awareness from the inside corresponds to the branch labeled “First-person.”
It is hard, and not particularly profitable, to try to disentangle the philosophical study of bodily awareness from the scientific study in psychology and physiology of how the mechanisms of bodily awareness function, or from neuropsychological inquiries into what happens in disorders of bodily awareness. In part, this is because bodily awareness is a somewhat concealed phenomenon. We depend on it constantly, but rarely attend to it. In part, it is because (unlike vision, touch, hearing, taste, and smell) it does not derive from a single dedicated sensory modality. And, in part, it is because bodily awareness plays an important role structuring and framing our general perceptual experience of the world (and of ourselves) so that it is not always easy to identify which elements in our multimodal experience of the world are contributed by bodily awareness.
Both the significance of bodily awareness and the importance of studying it in an empirically informed and multidisciplinary way were recognized in the phenomenological tradition long before philosophers in the analytic tradition started to take notice. Maurice Merleau-Ponty’s The Phenomenology of Perception, originally published in French in 1945 but not translated into English for another two decades (Merleau-Ponty 1962), is undoubtedly the most important work here. Chapter 5, “The Phenomenology of Bodily Awareness,” first published in a volume entitled Phenomenology and the Cognitive Sciences edited by Daviod Woodruff Smith and Amie L. Thomasson, is framed by Merleau-Ponty’s discussion, exploring how phenomenological insights into bodily awareness and its role in agency can be developed and illuminated by research into somatic proprioception and motor control.
From my perspective (as a sympathetic reader from a very different philosophical tradition, rather than as a Merleau-Ponty scholar), Merleau-Ponty’s guiding aim was to elucidate the distinctiveness of the experience of embodiment—the distinctiveness of how we find out about, and act through, our physical bodies. In doing this, he was deeply informed by the scientific study of the body. At the same time, though, he drew a sharp distinction between how we experience the body from a first-person perspective (as something that structures and gives meaning to our engagements with the physical world) and how we study the body from a third-person perspective as one physical object among others, as a complex organization of muscles, bones, and nerves. He formulated this distinction through the contrast between the for-itself and the in-itself, terminology that goes back at least as far as Hegel but may be unfamiliar to many contemporary philosophers and cognitive scientists. Still, the basic distinction he draws between these two ways of thinking about and studying the body seems incontrovertible.
However, Merleau-Ponty seems (at least to this reader) to draw significant, though ultimately untenable, metaphysical conclusions from his basic starting point. He distinguishes, for example, between objective space and what he terms “the natural system of one’s body.” He makes the evocative remark that “the frontier of my body is a line that ordinary spatial relations do not cross” (Merleau-Ponty 1962, 68) and draws a sharp distinction between the phenomenal body and the objective body (as a special case of his more general distinction between the phenomenal world and the objective world). It is not always easy to see how exactly these remarks are supposed to be read, but if one takes them at face value they appear significantly to constrain the explanatory power of the scientific study of the body.
So, is there a way to do justice to the phenomenological distinctiveness of our experience of our own bodies that Merleau-Ponty so eloquently characterizes, while still treating the body (from an ontological perspective) as one object among others? Chapter 5 sets out to do this, proposing a way to understand the phenomenological distinctiveness of somatic experience at the level of sense, not at the level of reference (to borrow Frege’s terminology). The strategy is to emphasize distinctive features of how the space of the body is represented within bodily experience, contrasting it with how extrabodily space is experienced.
It is standardly assumed within scientific psychology and cognitive science that extrabodily space is represented in multiple frames of reference. Some of these frames of reference are allocentric (centered on a nonbodily object). Others are egocentric (centered on the subject’s body). The egocentric frames of reference differ according to which body part they take as their origin, which in turn can depend on the relevant sensory modality. So, for example, visual information is standardly represented in retina-centered coordinates (coordinates centered on the retina), while information from smell and hearing is coded in head-centered coordinates and tactile information in hand-centered coordinates.
Despite these differences, all of these frames of reference are in essence Cartesian. That is, they allow spatial locations to be identified relative to axes centered on an origin. The experienced space of the body, however, seems fundamentally non-Cartesian. The simple reason for this is that there is no single part of the body that can serve as the origin of a Cartesian frame of reference, in the way that the eye (for example) is the obvious origin for a visual frame of reference. Whereas it makes perfect sense to ask, of any two visual perceived objects, whether they are in the same direction from the perceiver, or which is further way, neither question makes sense for bodily events apprehended through bodily awareness. These questions about direction and distance incorporate an implicit self-reference (further away from me, for example) and there is no privileged part of the body that can count as me for this purpose.
To accommodate these facts about the phenomenology of bodily experience, I proposed in The Paradox of Self-Consciousness an alternative model of how we experience the space of the body. The model starts from a distinction between two ways of thinking about bodily location. We can think about the location of a bodily event without taking into account the body as a whole. This would be to identify a bodily location relative to an abstract and relatively unchanging bodily map. Alternatively, we can think about the bodily event in a way that takes into account what the rest of the body is doing. I term these A-location and B-location, respectively.
Developing these ideas further exploits the fact that the human body can be viewed as a relatively immoveable torso to which are connected moveable limbs, each of which is further articulated into multiple, independently moving body parts connected by joints. The basic idea is that A-location specifies a bodily event within a particular body-part, while B-location fixes the location of that body part in terms of the angles of the joints that lie between it and the immoveable torso. So, the B-location of the pain in the ball of my foot is its A-location within my foot, supplemented by specifying the angles of the foot relative to the lower leg, and the lower leg relative to the upper leg.
Chapter 5 suggests that this fundamentally non-Cartesian model of how we experience the space of the body offers a metaphysically neutral way of doing justice to Merleau-Ponty’s insights about the distinctiveness of our experience of our own bodies. To illustrate how the model can be integrated with the scientific study of the body, I also show how it is consistent with important work in motor control and action planning.
Chapter 6, “Bodily Awareness and Self-Consciousness,” originally published in the Oxford Handbook of the Self edited by Shaun Gallagher, gives an overview of how bodily awareness functions as a form of self-consciousness. Emphasizing the multiple and complex sources of bodily awareness, the essay addresses why bodily awareness counts as a (nonconceptual) form of self-consciousness. This involves exploring topics continuous with those discussed in the context of visual proprioception, such as the significance of the IEM property for self-consciousness, as well as considering (and rejecting) some of the objections raised in the literature to characterizing bodily awareness as a form of self-consciousness.
Chapter 6 also introduces the topic of ownership, which is the focus of chapters 7 through 9. It has become common both in philosophy and in the cognitive sciences to talk about the sense of ownership that subjects have for parts of their body and, indeed, for their bodies as a whole. An important part of how each of us experiences our own body in a distinctive way is that we each experience our body and our limbs as our own. This seems to be a ground-level fact about bodily experience, which is often described by saying that we each have a sense of ownership for our own bodies and body parts. This raises a number of questions about how exactly we should understand that sense of ownership and how it is grounded. This is the problem of bodily ownership.
The concept of ownership seems to mean different things to different people, and, in an attempt to clarify the structure of ongoing debates, I make a distinction in chapter 6 between inflationary and deflationary conceptions of the sense of ownership. According to the inflationary conception, our awareness of our own bodies incorporates a specific feeling of “mineness” or “myness,” and it is because we have this feeling that we experience our bodies and body parts as our own. For the inflationary theorist, therefore, the sense of ownership is grounded in what might be termed a phenomenologically salient quale of ownership. According to deflationary theorists, on the other hand, the sense of ownership is grounded in other, more fundamental, aspects of bodily experience—with different theorists looking to different aspects of bodily experience as potential grounds. To be clear, both deflationary and inflationary theorists hold that the sense of ownership is phenomenologically salient, where that means that we experience our bodies and body parts as our own.5 Theorists differ on how they explicate that phenomenological salience.
Chapter 6 offers an argument against the inflationary conception of ownership (an argument that is developed further in chapter 7). The argument adapts and develops some important themes from Elizabeth Anscombe’s discussion of knowledge without observation in her classic paper “Sensations of Position” (Anscombe 1962) and in her book Intention (Anscombe 1957). In brief, Anscombe poses a dilemma for the thesis that our knowledge of how are limbs are disposed is derived from bodily sensations. The dilemma emerges when we ask about the content of bodily sensations and how that content is supposed to ground our knowledge. On the one hand, we can think of the content of sensations purely internally (in terms of feelings of pressure, for example, or strain in muscles, or inchoate tinglings). On the other hand, we can think of their content externally (in what we would now call representational terms).
According to Anscombe, neither way will allow sensations to serve as a source of knowledge. When construed internally, the content of sensations is too vague and general to ground fine-grained knowledge of how limbs are distributed. But when construed externally, the content of the sensations is effectively the same as the content of the knowledge it is supposed to be grounding, and hence incapable of providing independent support for it. It is straightforward to extend this argument to the putative quale of ownership postulated by inflationary theorists, who typically fall foul of the external horn of the dilemma. The “feeling of myness” is invariably presented as precisely that—namely, as a feeling whose (external) content is that the body part within which the feeling is located is mine. So, they are susceptible to the charge that the feeling of ownership simply recapitulates what it is supposed to be independently justifying.
One of the targets of my argument in chapter 6 was Frédérique de Vignemont, who subsequently published an essay in the journal Analysis objecting to my argument and offering a new line of argument appealing to the rubber hand illusion in support of the inflationary conception. Chapter 7, “Bodily Ownership, Bodily Awareness, and Knowledge without Observation,” a much-expanded version of an article first published in Analysis, continues the discussion. I remain unconvinced by her arguments and objections, but welcome the opportunity to clarify further my own thinking about ownership and bodily awareness.
In particular, chapter 7 addresses an important contrast between the content of visual perception and the content of bodily awareness that emerges from the discussion of Anscombe’s argument and the rubber hand illusion. Visual perception stands in justificatory relations to perceptual judgments. This was explored in chapter 4 in the context of a very specific class of first-person judgments, but of course it holds more generally. Visual perception is a paradigm example of knowledge through observation and hence that to which bodily sensations and the sense of ownership are being contrasted. At the same time, though, it would be reasonable to ask why the content of visual perception does not fall foul of Anscombe’s dilemma—which surely would be a reductio both of her argument and of my use of it.
We know from the demise of sense datum theory that there is no prospect of characterizing the content of perception in purely sensational terms (at least, not if one wants perception to justify belief). But if we think about the content of visual perception externally or representationally, then in what sense is it sufficiently independent to ground beliefs and judgments that are based on it? And, if it is sufficiently independent, then why can the same not be said for the content of bodily awareness?
This challenge raises subtle questions that go back to the discussion in chapter 3 of what it is to take a perception at face value. The answer to the challenge that I propose in chapter 7 is, in brief, that the nonconceptual content of perception can be specified in ways that are completely independent of any of the concepts featuring in those perceptual judgments that they might be called on to justify. The theory of scenario content developed by Christopher Peacocke is a good example of how this might be done, and in fact Peacocke has used his model of scenario content to elucidate how perceptual states with nonconceptual content can justify belief states with conceptual content (see particularly Peacocke 1992). The same cannot be said, however, for the putative “feeling of ownership.” The type of feeling proposed by the inflationary theorist cannot be specified in the type of “ownership-neutral” manner that would be required to escape an Anscombe-type argument. This is because the feeling of ownership itself counts as a bodily sensation and hence will have the same general properties as any other bodily sensation. One such general property is that facts about ownership are an integral part of the content of bodily sensations (and other forms of bodily awareness). So, as a bodily sensation, the feeling of ownership proposed by the inflationary theorist is insufficiently independent of the phenomenon that it is being called on to explain.
This basic point about bodily sensation and bodily awareness in general is developed in much more detail in chapter 8, “Ownership and the Space of the Body,” published in a volume entitled The Subject’s Matter edited by Frédérique de Vignemont and Adrian Alsmith. In that essay, I take the following two principles as starting points for discussing the content of bodily awareness and how it supports the phenomenology of ownership.
Boundedness
Bodily events are experienced within the experienced body (a circumscribed body-shaped volume whose boundaries define the limits of the self).
Connectedness
The spatial location of a bodily event is experienced relative to the disposition of the body as a whole.
Chapter 8 does not directly discuss the inflationary view of ownership, but it offers a natural way of extending the line of argument from chapter 7, since a quale of ownership that satisfies Boundedness and Connectedness will plainly fail Anscombe’s independence constraint.
The main focus of chapter 8 is providing a deflationary account of ownership—that is to say, an account that grounds the phenomenology of ownership in other, more fundamental aspects of bodily experience. My proposal is that judgments of ownership are ultimately grounded in the distinctive way we experience the space of the body. To develop this idea, chapter 8 extends the model of A-location and B-location proposed in chapter 5 (and earlier in The Paradox of Self-Consciousness). The A-location of a bodily event is its location in a specific body part relative to an abstract map of the body. A-location does not take into account the current position of the body. That is the main difference from B-location, which is a bodily event’s location relative to the current disposition of body parts. In The Paradox of Self-Consciousness and chapter 5, I discuss A-location and B-location very generally, in terms of a model of the body as an articulated structure in which a relatively immoveable torso connected by joints to moveable body parts, each of which is further articulated into smaller body parts and joints. Chapter 8 develops this general picture further in light of two very different (but complementary) approaches to modeling the body.
The first approach is most prominent in kinesiology and robotics, where the body is typically modeled as a system of rigid links connected by mechanical joints. This way of modeling the body typically represents both bodily position at a time and movement over time in terms of the angles of the relevant joints. The second derives from the well-known model of object recognition proposed by Marr and Nishihara (1978). In that model, objects are represented schematically as complexes of generalized cones (surfaces generated by moving a cross-section along an axis, maintaining its shape but possibly varying its size). The human body, in particular, is represented as a hierarchy of generalized cones. Chapter 8 shows how these two approaches can be combined to yield a model of the body as a hierarchy of generalized cones linked by mechanical joints. I argue, first, that this model captures how we experience the space of our bodies and, second, that our experiencing the space of our bodies in this distinctive way is what ultimately grounds both the phenomenology of ownership and our judgments of ownership.
Chapter 9, “Bodily Ownership, Psychological Ownership, and Psychopathology,” first published in a special issue of the Review of Philosophy and Psychology edited by Alexandre Billon, explores the relation between this general model of bodily ownership and discussions of psychological ownership—the phenomenon of taking one’s thoughts, emotions, and other psychological states as one’s own. In chapter 9 I term these φ-ownership (bodily) and ψ-ownership (psychological), respectively. The particular model of ψ-ownership I consider is that proposed by John Campbell in a number of papers. Reflecting on cases of schizophrenia (and other psychopathologies), Campbell proposes fractionating ψ-ownership into two components: a self-ascriptive component, on the one hand, and a causal component, on the other. One can think of those two dimensions of ownership as corresponding to the following two different types of judgment of ownership:
(1) “I am thinking this thought”
(2) “I am the producer of this thought”
Campbell’s point is that schizophrenic patients suffering from delusions of thought insertion sometimes make judgments of the first type while rejecting judgments of the second type—they can think of themselves as thinking thoughts that they did not themselves produce.
On the view laid out in chapter 9, φ-ownership and ψ-ownership are distinct phenomena for which different explanations need to be offered. They are closely related in various ways, however. As I bring out, each presupposes the other. Our awareness of our own bodies and our awareness of our ongoing thoughts, emotions, and feelings are complementary elements of a single, embodied perspective on the world. Both φ-ownership and ψ-ownership are parts of what it is to be a bodily self, and for that reason one would expect there to be a deep commonality between them. As chapter 9 brings out, this commonality emerges from the fundamental role of agency in the two types of ownership.
Agency is at the heart of the model of φ-ownership that I have developed because it derives ownership from the spatial content of bodily awareness, and the spatial content of bodily awareness brings with it awareness of the agent’s agential abilities. It is not just that the limits of the body are the limits of what is directly responsive to the will; the key variables for specifying both A-location and B-location are also given in terms of the agent’s capacities for movement (since limbs and other effectors are manipulated through joints). Likewise, for the reasons I brought out earlier, agency is central to Campbell’s model of ψ-ownership, which incorporates the idea that (in the normal case) ownership of one’s thoughts goes beyond simple self-ascription and incorporates taking oneself to be (causally) responsible for the thought. Taking both φ-ownership and ψ-ownership into account, therefore, to experience oneself as a bodily self is to experience oneself as an agent.
The final essay in this volume takes a broader perspective on the issue of agency. Chapter 10, “The Bodily Self, Commonsense Psychology, and the Springs of Action,” appears here for the first time, although it draws on elements of two previously published papers. It programmatically suggests how the general picture of the bodily self that has emerged from the essays reprinted here might be extended to modeling action and how it is explained. Let me explain the analogy and the suggestion.
The earlier essays in this volume have all, in one way or another, explored how full-fledged, conceptual self-consciousness rests on a multilayered complex of forms of nonconceptual self-consciousness. These forms of nonconceptual self-consciousness are so omnipresent and pervasive that they can easily seem invisible. But the moral of the discussion is that our sophisticated abilities to think and speak about ourselves rest on a bed of much more primitive self-specifying representations. One way of framing this idea would be to say that the domain of conceptual self-consciousness turns out to be much narrower than it has typically been taken to be. It is not just that we can represent ourselves wordlessly and without concepts. It turns out that these representations guide and structure much (if not most) of our activity as we navigate the physical world. The bodily self, one might say, is primary.
So now, with this general picture in mind, consider standard models of the springs of action. Within philosophy and cognitive science, action is often understood in a highly conceptual way. That is to say, the springs of action are taken to be beliefs, desires, and other propositional attitudes, with the general idea being that action is caused by desires (or comparable pro-attitudes) and guided by belief (or comparable information states). Within philosophy, this way of thinking about action has a long and distinguished history, going back at least as far as Aristotle’s discussion of the practical syllogism. More recently, it finds a very clear expression in the distinction between modular and central processing that Jerry Fodor, among others, sees as key to contemporary cognitive science. Central processing is the domain of the propositional attitudes, of practical reasoning and decision making. Intentional actions (that is to say, those not governed by reflex or habit) emerge from central processing. Since it is standard to think of beliefs and other propositional attitudes as having conceptual contents, this picture is one on which the springs of action fall within the domain of the conceptual.
This dominant way of thinking about the springs of action goes hand in hand with a model of social understanding and social coordination. According to this model, we navigate the social world by tacitly applying a conceptual understanding of how beliefs, desires, and other propositional attitudes work. The model can be developed in various ways. So, the conceptual understanding might take the form of an implicit theory (as proposed by so-called theory theorists) or it might be built into simulation mechanisms. But the common thread is that the conceptual framework of propositional attitude psychology (also known as folk psychology, or commonsense psychology) enables and structures our social interactions.
It is unsurprising that the two models go hand in hand, for it is natural to try to explain things in terms of what one thinks causes them. And, for that very reason, it seems unpromising to take issue with one without taking issue with the other. Accordingly, in chapter 10 I tackle both. The discussion is programmatic, designed to prime the pump, as it were, by pointing to a range of phenomena that fall outside the domain of commonsense psychology and propositional attitudes (from both an etiological point of view and an explanatory one).
First, I point toward ways of thinking about the springs of action that do not engage the propositional attitudes. As far as the etiology of action is concerned, human organisms act within the world in virtue of how they represent the world. But there often seems to be a fundamental lack of fit between our best models of the representations that generate behavior and the model of representation built into propositional attitude psychology. Evidence for this lack of fit comes from range of sources, from neural network modeling to detailed studies of action control and perceptual processing. Second, and correlatively, I explore tools for achieving social understanding and social coordination that bypass the propositional attitudes. These range from direct perception of emotions to game-theoretic heuristics (such as tit-for-tat in the iterated prisoner’s dilemma) and what AI theorists term scripts and frames.
This introduction has brought out some of the common themes that run through all the chapters. In contrast, the book ends with an afterword that seeks to identify some of the principal avenues for further inquiry. If there is one unifying theme for this volume, it is that the fully conceptual understanding that we have of ourselves and others rests upon the complex underpinnings of many types and levels of nonconceptual (self-)awareness. Thinking through the implications of this will require reconfiguring a number of fundamental concepts, most obviously perhaps the concept of rationality. Our notion of rationality is closely tied to paradigms of highly reflective and self-aware thinking. As we begin to see, though, that the scope of such self-aware reflection is much narrower than standardly thought, it becomes pressing to ask how, if at all, standards of rationality can be applied to the actions and reactions of the bodily self. In an important sense, therefore, the message of this book is a challenge and a question: Now that the scope and importance of the nonconceptual bodily self is (I hope) clearly in view, how do we need to change the basic evaluative framework that we apply to human thinking and human behavior?
Notes
References
Anscombe, G. E. M. 1957. Intention. Oxford: Blackwell.
Anscombe, G. E. M. 1962. On sensations of position. Analysis 22:55–58.
Anscombe, G. E. M. 1975. The first person. In Mind and Language, ed. S. Guttenplan. Oxford: Clarendon Press.
Bermúdez, J. L. 1994. Peacocke’s argument against the autonomy of nonconceptual representational content. Mind & Language 9:402–418.
Bermúdez, J. L. 1998. The Paradox of Self-Consciousness. Cambridge, MA: MIT Press.
Bermúdez, J. L. 2017. Understanding “I”: Language and Thought. Oxford: Oxford University Press.
Bermúdez, J. L., and A. Cahen. 2015. Mental content, nonconceptual. Stanford Encyclopedia of Philosophy. http://plato.stanford.edu.
Bermúdez, J. L., A. J. Marcel, and N. Eilan, eds. 1995. The Body and the Self. Cambridge, MA: MIT Press.
Davidson, D. 1986. A coherence theory of truth and knowledge. In Truth and Interpretation: Perspectives on the Philosophy of Donald Davidson, ed. E. Lepore, 307–319. Oxford: Blackwell.
Gibson, J. J. 1966. The Senses Considered As Perceptual Systems. Boston: Houghton Mifflin.
Gibson, J. J. 1979. The Ecological Approach to Visual Perception. Boston: Houghton Mifflin.
Marr, D., and H. K. Nishihara. 1978. Representation and recognition of the spatial organization of three-dimensional shapes. Proceedings of the Royal Society of London 200:269–294.
McDowell, J. 1994. Mind and World. Cambridge, MA: Harvard University Press.
Merleau-Ponty, M. 1962. The Phenomenology of Perception. London: Routledge.
Peacocke, C. 1992. A Study of Concepts. Cambridge, MA: MIT Press.
Peacocke, C. 1994. Nonconceptual content: Kinds, rationales, and relations. Mind & Language 9:419–429.
Peacocke, C. 1999. Being Known. Oxford: Oxford University Press.
Peacocke, C. 2002. The relations between conceptual and nonconceptual content (Postscript to reprint of Peacocke 1994). In Essays on Nonconceptual Content, ed. Y. Guenther. Cambridge, MA: MIT Press.
Shoemaker, S. 1996. The First-Person Perspective and Other Essays. Cambridge University Press.
Strawson, P. F. 1975. The Bounds of Sense: An Essay on Kant’s “Critique of Pure Reason,” vol. 18. New York: Routledge.
