Consciousness and Cognition

Consciousness and Cognition 16 (2007) 583599 www.elsevier.com/locate/concog

Pre-reective self-as-subject from experiential and empirical perspectives

Dorothee Legrand
*
CREA CNRS, 1 rue Descartes, 75005 Paris, France Received 8 January 2007 Available online 29 May 2007

Abstract In the rst part of this paper I characterize a minimal form of self-consciousness, namely pre-reective self-consciousness. It is a constant structural feature of conscious experience, and corresponds to the consciousness of the self-as-subject that is not taken as an intentional object. In the second part, I argue that contemporary cognitive neuroscience has by and large missed this fundamental form of self-consciousness in its investigation of various forms of self-experience. In the third part, I exemplify how the notion of pre-reective self-awareness can be of relevance for empirical research. In particular, I propose to interpret processes of sensorimotor integration in light of the phenomenological approach that allows the denition of pre-reective self-consciousness. 2007 Elsevier Inc. All rights reserved.
Keywords: Subjectivity; Pre-reectivity; Consciousness; Self-recognition; Neural correlates; Sensori-motor integration; Self-related processes; Phenomenology; Cognitive neurosciences

1. Pre-reective consciousness of the self-as-subject1 1.1. Self-consciousnesses Self-recognition is by and large considered as an important criterion for self-consciousness. In developmental and comparative psychology, the so-called mirror-recognition task has occasionally been heralded as the decisive test for self-consciousness (TaylorParker, Mitchell, & Boccia, 2006). Children would be self-conscious only when capable of recognizing themselves in the mirror (cf. Lewis, 2003, pp. 281282). The question whether chimpanzees (Gallup & Povinelli, 1998; Gallup, 1970; Povinelli et al., 1997), non human primates (Itakura, 2001), birds (Epstein, Lanza, & Skinner, 1981), dolphins (Reiss & Marino, 2001) or elephants (Plotnik, de Waal, & Reiss, 2006) are self-conscious is considered according to their ability or inability to behave in front of a mirror as if in front of themselves vs. in front of a conspecic.
*

Fax: +33 1 55 55 90 40. E-mail address: legrand@shs.polytechnique.fr A good deal of the content and structure of this rst part is indebted to published as well as still unpublished writings of Dan Zahavi.

1053-8100/$ - see front matter 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.concog.2007.04.002

584

D. Legrand / Consciousness and Cognition 16 (2007) 583599

However, self-consciousness comes in many forms and degrees and does not only emerge the moment one recognizes ones own mirror image or scrutinizes ones experiences attentively. Rather, a more minimalist account defends the view that the most primitive form of self-consciousness simply corresponds to the subjective dimension of experience (cf. Zahavi, 1999, 2005). In what follows, I designate this primary form of selfconsciousness as a pre-reective consciousness of the self-as-subject. The rst section intends to clarify this notion, by comparing it with other conceptions of self-consciousness. 1.2. Primary self-consciousness Pre-reective self-consciousness specically corresponds to consciousness of the self as it is the subject of any given experience. Compare two dierent experiences: the smelling of fresh coee and the seeing of midnight sun. These experiences dier in their phenomenality, i.e. in what it feels like to undergo them. More precisely, the experiences dier from each other both in terms of content (coee vs. sun) and mode of presentation of these contents (smelling vs. seeing). However, these experiences do not dier in every aspect. They share a specic dimension in the fact that they are all given from the rst-person perspective, they are given (at least tacitly) as my experiences, as experiences I am undergoing: they feel like something for me. This quality of mineness or for-me-ness is what the notion of pre-reective consciousness of the self-as-subject designates. It is important to note that the specicity of pre-reective consciousness of the self-as-subject is not fully captured by the notion of phenomenal consciousness (Block, 1997; Zahavi, 2005, p. 224). Indeed, as the previous examples illustrate, pre-reective self-consciousness remains constantly present even when other aspects of phenomenal consciousness vary, implying that these dimensions of experience cannot be reduced to each other. The mineness or subjectivity in question is not a quality as bitter or bright, black or orange that varies with the intentional object experienced. Rather, it refers to the fact that every experience is characterized by a subjective mode of givenness in the sense that it feels like something for the experiencing subject. For this reason, pre-reective self-consciousness does not vary either with the modulation of phenomenal experience of the self taken as an intentional object (just like it does not vary with the modulation of phenomenal experience of any other intentional object). For example, when I experience myself as being hungry, I experience myself in a double manner: I experience an intentional aspect (hunger), but the latter does not suce to capture the subjective aspect at stake here: pre-reective consciousness of the self-as-subject (I). The point here is that pre-reective self-consciousness is not reducible to phenomenal consciousness, as evidenced by the fact that the latter can change without involving any modulation of the former. Note that this does not imply that pre-reective self-consciousness always remains constant (it can be more or less recessive). It only implies that its potential modulations are not directly and systematically due to changes of the intentional aspects of experience. 1.3. Pre-reective self-consciousness versus anonymity One reason why this dimension of experience can be adequately described as a primary form of self-consciousness is because it corresponds to a quality or dimension of mineness. However, according to what might be termed the anonymity objection, there would be no experience of self at the pre-reective level. Rather, experiences would be characterized by a certain anonymity or neutrality. To claim that every experience has a quality or dimension of mineness would consequently be a post-hoc fabrication. In reply, it is crucial to understand that the notion of pre-reective self-consciousness does not suppose that the self would be experienced as standing opposed to the stream of consciousness. Rather, at the pre-reective level, it is an integral part of conscious experience. I do not rst experience a neutral or anonymous toothache or intention to act, then ask the question Whose experience is this actually? to nally nd myself as the owner of these experiences (Legrand, in press; Shoemaker, 1968). Rather, any experience is pre-reectively experienced as intrinsically subjective in the sense that it is experienced from the perspective of the experiencing subject (Zahavi, 2005). The latter is a rst-person perspective; it is tied to a self in the sense of being tied to the point of view of the experiencing, perceiving, acting subject. It could consequently be claimed that

D. Legrand / Consciousness and Cognition 16 (2007) 583599

585

anybody who denies the subjectivity of experiences and considers that experiences can be given in a neutral way simply fails to recognize an essential aspect of what it feels like to undergo an experience. 1.4. Pre-reective self-consciousness versus transparency As just proposed, one way to cash out the notion of pre-reective consciousness of the self-as-subject is by defending that phenomenal experiences are intrinsically subjective: pre-reective self-consciousness is an intrinsic aspect of phenomenal experience. Insofar as conscious experiences are characterized by a subjective feel, i.e., a certain what it is like or what it feels like to have them, they also come together with a minimal form of self-consciousness. This description of pre-reective self-consciousness can appear inconsistent with descriptions of experience as transparent defended by both externalists and representationalists: When we try to introspect the sensation of blue, all we can see is the blue: the other element is as if it were diaphanous (Moore, 1903, p. 25). The darkness of a coee and the brightness of the midnight sun are not qualities of experiences, they are qualities of the things represented, the coee and the sun. In this view, what it is like to have a certain experience is reducible to the quality of what is being intentionally represented. The latter is not an intrinsic and non-intentional quality of experiences themselves but consists entirely in the qualitative properties of the experienced objects (Dretske, 1995, p. 1). These views reduce intentional experiences to world-presenting components. Therefore, they give no place to the primary form of self-consciousness that has been described above as being intrinsic to intentional experiences: introspection of your perceptual experiences seems to reveal only aspects of what you experience, further aspects of the scenes, as represented. Why? The answer, I suggest, is that your perceptual experiences have no introspectible features over and above those implicated in their intentional contents. So the phenomenal character of such experiences . . . is identical with, or contained within, their intentional contents (Tye, 1995, p. 136). The externalist attempt to locate the intentional dimension of experience outside is in fact not incompatible with the notion of pre-reective self-consciousness. However, considering pre-reective self-consciousness in its specicity implies to refute the reduction of phenomenal experience to its intentional dimension (Legrand, 2005). In other terms, what might be called the transparency objection operates with a conception of phenomenality that is too impoverished to address specically the subjective dimension of experience that the notion of pre-reective consciousness of the self-as-subject intends to capture. What it is like to be dazzled by the sun diers from what it is like to smell fresh coee, and what it is like to perceive diers from what it is like to remember or imagine. But what is crucial to account for pre-reective self-consciousness is to underline that all these dierent phenomenal experiences are necessarily experienced by a given subject. These distinct experiences bring me into the presence of dierent intentional objects. Not only am I phenomenally acquainted with various properties of these objects, but also these objects are there for me. Given that this for me quality remains constant whatever the intentional object, it makes little sense to suggest that this mineness can be reduced to a qualitative feature of the object experienced. Phenomenality is world-presenting but it is also self-involving. To put it dierently, it has an intentional and a subjective component, both being irreducible to each other. The notion of transparency of experience as used by externalist conceptions of consciousness only concerns the former aspect and does not consider specically (nor threaten) the notion of pre-reective self-consciousness. 1.5. Pre-reective self-consciousness versus objectifying consciousness Considering pre-reective self-consciousness in its specicity implies to refute the equation of consciousness with object-consciousness. For example, Searle (2005) proposed that the conscious eld should not be conceived of as a eld constituted only by its contents and their arrangements. The contents require a principle of unity, a self that is not a separate entity distinct from the eld. So far, this claim coheres with the present account of pre-reective self-consciousness. However, Searle further argues that the postulation of a self is like the postulation of a point of view in visual perception. Just like we cannot make sense of our perceptions unless we suppose that they occur from a point of view, even though the point of view is not itself perceived,

586

D. Legrand / Consciousness and Cognition 16 (2007) 583599

we cannot, according to Searle, make sense of our conscious experiences unless we suppose that they occur to a self, even though the self is not consciously experienced. All (nonpathological) consciousness has to be possessed by a self. This self is not the object of consciousness, nor is it part of the content of consciousness. In Searles view this implies that there is no experience of the self, even though we have to infer that there is a self. The problem is that Searle equates consciousness with object-consciousness. Failing to consider subjectconsciousness, he concludes that the self must either be given as an object or not be given at all. But as Evans put it: [F]rom the fact that the self is not an object of experience it does not follow that it is non-experiential (Evans, 1970, p. 145). The present view proposes that a crucial aspect of pre-reective self-consciousness is that it is thoroughly non-observational and non-objectifying in the sense that it does not correspond to any form of consciousness of some object towards which I would occupy the position or perspective of a spectator or in(tro)spector. Prereective self-consciousness sensu stricto is an awareness of oneself as subject. Therefore, it cannot be a question of the self taking itself as an object of experience. More in detail, to understand the specicity of pre-reective self-consciousness as a non-objectifying form of consciousness of the self-as-subject, it is crucial to underline that subject-consciousness and object-consciousness are fundamentally dierent modes of consciousness. This distinction between subject and object is not ontological but phenomenological: object here means object of intentional consciousness and subject means subject of intentional consciousness. For x to be given as an object of experience is for x to be given as diering from the subjective experience that takes it as an object. In short, subject and object of experience are necessarily dierent from each other: an object of experience is something that stands in opposition to the subject of experience. Even if x can be identical to y and be an object for y (Williford, 2006), the subject cannot stand in a subjectobject relation to itself at the pre-reective level. Therefore pre-reective selfconsciousness cannot be reduced to the experience of a particular object: Object-consciousness necessarily entails a dierence between the subject and the object of experience, and it is this very feature that makes object-consciousness singularly unsuited as a model for self-consciousness. This conception of the subjectivity of experience contrasts with the description of self-consciousness in terms of a peripheral or marginal object-consciousness as well as with the view espoused by Brentano according to which a mental state becomes conscious by taking itself as an object (Zahavi, 2004, 2006). For phenomenologists, experience is conscious without being itself an intentional object (Husserl, 1984a, 1984b, p. 399; Sartre, 1936, pp. 2829). Of course, we can direct our attention towards our experiences and thereby take them as objects, but this only occurs the moment we reect upon them. In everyday life, I rather enjoy a continuous and pre-reectively rst-personal access to myself as the subject of the experiences I undergo. 1.6. Pre-reective self-consciousness grounds consciousness of the self-as-object These considerations raise the question of the link between consciousness of the self-as-subject and consciousness of the self-as-object, i.e. between the subjective and the intentional aspects of phenomenal consciousness. This section will argue that pre-reective self-consciousness is the necessary ground to which other forms of self-consciousness are anchored, i.e. the subjective rst-person perspective anchors the experience of intentional objects (the self-as-intentional-object included). Consider again self-recognition in a mirror. This is a case of consciousness of the self-as-object in the sense that the self is taken as an object of visual observation. Such self-recognition is often taken as paradigmatic of self-consciousness even though it let unexplored pre-reective consciousness of the self-as-subject. It is crucial to consider the latter not only because it is an essential constituent of consciousness and a primary form of self-consciousness but also because it provides the necessary ground anchoring conscious self-recognition. Recalling a well-known story described by John Perry (1993) will allow a better understanding of this point. Ernst Mach enters a bus, he sees a person in front of him, and judges that this person is a shabby pedagogue. Only latter on, he realizes that this shabby pedagogue is nobody else but himself who he was watching in a mirror. This story provides the occasion to describe a case of self-observation without self-consciousness and illustrates that observing the self-as-object (e.g. in a mirror) is insucient for self-consciousness. More in detail, self-observation implies that the same self is both the subject and the object of the observation. In addition,

D. Legrand / Consciousness and Cognition 16 (2007) 583599

587

self-consciousness implies that the subject recognizes that he is both the subject and the object of observation. However, this cannot be given by observation itself since the latter implies a dissociation of the self which is cut into an observing subject and an observed object. Therefore, self-observation itself cannot ground self-consciousness. By analogy, Machs mirror reection is not enough for him to consciously recognize himself as himself. One may think that self-recognition could be possible thanks to a second act of consciousness which would take the rst one as its object. Thanks to this additional act of consciousness, one would be able to observe the subject and the object of the act of consciousness, and would be able to identify them as identical. However, observing that the subject and the object of an act of consciousness are identical does not imply that I recognize myself as being the subject. Quite the contrary, the self previously dissociated in a subject and an object of observation is now dissociated in a meta-observer, an observer and an observed object. This conception thus seems particularly unsuited to account for self-consciousness. Shoemaker has provided a classical argument purporting to show why it is impossible to account for selfconsciousness only in terms of a successful object-identication. In order to identify someone as oneself, one has to identify certain features and already know that these features characterize oneself. In some cases, this self-knowledge might be grounded in some further identication, but, in turn, the latter also needs to rely on some further self-knowledge. The supposition that every item of self-knowledge rests on identication thus leads to an innite regress (Shoemaker, 1968, p. 561). This holds even for self-identication of mental event obtained through introspection. If no other self but me could possibly have the private and exclusive object of exactly my introspection, it would still be insucient to make this introspection immediately identied as being held by me. Indeed, I would be unable to identify an introspected self as myself by the fact that it is introspectively observed by me, unless I know it is the object of my introspection, i.e., unless I know that it is in fact me who undertakes this introspection. Importantly, this knowledge cannot itself be based on identication, on pain of innite regress (Shoemaker, 1968, pp. 562563). The point here is that identication of objects cannot ground consciousness of the self-as-subject: intentional aspects of consciousness cannot ground its subjective aspects. Therefore, a non-objectifying form of self-consciousness where the self is not an intentional object of consciousness detached from the subject of consciousness is a necessary condition to be conscious of oneself as such. 1.7. Pre-reective self-consciousness as an intrinsic aspect of consciousness Previous sections argue that consciousness of the self-as-subject grounds consciousness of the self-as-object. More generally, it is crucial to understand that even though subjectivity and intentionality are fundamentally dierent dimensions of experience, they are closely intertwined: subjectivity is an intrinsic, constitutive feature of intentional consciousness. Without self-consciousness, there would be nothing it is like to undergo a mental episode, which therefore could not be a phenomenally conscious process (Zahavi, 1999, 2005). This view has been forcefully defended by numerous phenomenologists, and by a growing number of contemporary analytical philosophers and cognitive neuroscientists. As a case in point, consider the following quotes: This self-consciousness we ought to consider not as a new consciousness, but as the only mode of existence which is possible for a consciousness of something (Sartre, 1943, p. 20). What would it be like to be conscious of something without being aware of this consciousness? It would mean having an experience with no awareness whatever of its occurrence. This would be, precisely, a case of unconscious experience. It appears, then, that being conscious is identical with being self-conscious. Consciousness is self-consciousness (Frankfurt, 1988, p. 162). If self-consciousness is taken to mean consciousness with a sense of self, then all human consciousness is necessarily covered by the termthere is just no other kind of consciousness (Damasio, 1999, p. 19). However, one might share the view that there is a close link between phenomenal consciousness and selfconsciousness and still disagree about the nature of the link. The specicity of self-consciousness at the prereective level is that it is not present only when I entertain some kind of reection or introspection. Rather, the primary form of self-consciousness is pre-reective in the sense that it does not involve an additional

588

D. Legrand / Consciousness and Cognition 16 (2007) 583599

mental state. It is not a quality added secondarily to the experience (Zahavi, 1999, 2005). Therefore, the specicity of this primary form of self-consciousness cannot be adequately accounted for in terms of higher-order monitoring as suggested by higher-order theories (Armstrong, 1968; Carruthers, 1996; Lycan, 1987; Rosenthal, 1986). Pre-reective self-consciousness is rather to be understood as an intrinsic feature of experience. As claried above, the intentional and the subjective dimensions of experience cannot be reduced to each other but intentionality and subjectivity are nonetheless intertwined with each other and are both intrinsic dimensions of experience. Consider Wittgensteins examples of the I-as-subject: I see so-and-so, I try to lift my arm, I think it will rain, I have a toothache (1958, pp. 6667). The I-as-object is rather of the form: I have grown 6 inches , I have a bump on my forehead. As these examples illustrate, in the case of an I-as-object, I am aware of a specic object me and I identify specic properties of it, as its shape and size. Conversely, in the case of an I-as-subject, the object of consciousness is not me. Rather, it is, e.g. a visual percept (I see so-andso), an attempt to move (I try to lift my arm), a thought (I think it will rain), a pain (I have a toothache). I am conscious of myself only insofar as I experience these contents of consciousness as experienced by me: I am the subject of the visual perception, of the trying to move, of the thought and of the pain, and I experience myself as such at the pre-reective level. To put it dierently, consciousness of the world and pre-reective self-consciousness are strictly contemporary. There is a world for me because I am not unaware of myself; and I am not concealed from myself because I have a world (Merleau-Ponty, 1962, p. 298). The self-as-subject is neither an external object (for example, it is not my body that I can observe in the mirror) nor an internal object: when I am conscious of myself as the subject of an experience, I am not scrutinizing an internal self looking at the external world. I am simply looking outside at the external world, and within this single act of consciousness I pre-reectively experience myself-as-subject. Interestingly, the framework defended here sheds light on what is probably the most quoted description of self experience: For my part, when I enter most intimately into what I call myself, I always stumble on some particular perception or other, of heat or cold, light or shade, love or hatred, pain or pleasure. I never can catch myself at any time without a perception, and never can observe any thing but the perception (Hume, 1739). This position is most often considered as dismissing the notion of self all together. However, in the current framework, this description of the experience of oneself interestingly coheres with an anchoring of the self to experiences of non-self intentional objects, to some particular perception or other. Self-consciousness primarily corresponds to pre-reective consciousness of the self-as-subject, which corresponds to the subjective dimension of experience, intertwined with its intentional dimension. At this pre-reective level at least, there is no such thing as a world-less self-experience, as well as there is no such thing as a self-less world-experience. 2. Cognitive neuroscience misses the investigation of pre-reective consciousness of the self-as-subject In the rst part of the present article, I proposed a shift of focus from consciousness of the self-as-object to pre-reective consciousness of the self-as-subject. In what follows, I will complement the phenomenological perspective developed above with an empirical perspective. In the following sections, I will review the major approaches in cognitive neuroscience that aim to investigate self-consciousness and will consider whether they explore pre-reective consciousness of the self-as-subject. 2.1. The content-approach The most recurrent approach in the search for the neuronal correlates of the self records brain activations during tasks involving the representation of self-specic contents (e.g. my name, my face, my body, my action vs. your name, your face, your body, your action). A recent extensive review of such results concludes that most brain activations reported to be activated during self-recognition did in fact lack self-specicity, e.g. they overlapped with cerebral activation involved in the

D. Legrand / Consciousness and Cognition 16 (2007) 583599

589

representation of other-related stimuli (Gillihan & Farah, 2005). But it is noteworthy that this review focused on investigations of representational contents that are recognized as self-related (e.g. the recognition of my face, the visual observation of my action). In other terms, the reviewed results shared two common features: (1) they failed to involve self-specic brain activations; and (2) they only involved the self-as-object (as an object of perception, attention), thereby ignoring the self-as-subject (as a subject of perception, representation, action). To clarify how these two features are linked to each other (see also Ruby & Legrand, in press), let me recall the distinction between the self-as-object and the self-as-subject. The term object is not meant to refer exclusively to physical objects in the external world. Thus, the contrast between the self-as-object and the selfas-subject does not contrast a physical object and a non-physical subject. The notions of self-as-subject and self-as-object are neutral terms at the ontological level. The notion of self-as-object simply refers to the self as an intentional object of consciousness, perception, attention, representation. In the same vein, the notion of self-as-subject refers to the self as the subject of experience, i.e. the subject who consciously intends objects (including potentially himself-as-object). For example, when I experience my body, I entertain a body image, thereby having a double experience of myself characterized by intentional aspects of the self-as-object (the body image) and subjective aspects of the self-as-subject (I). The content-approach to the self seeks to record those brain activations that occur whenever a content related to the self-as-object is processed. However, in such a way, it simply collects a variety of heterogeneous brain activations correlated with such self-specic contents. This approach fails to identify a core brain activation that would be common across the processing of dierent contents attributed to the self. It reduces the self to a collection of self-related traits characterizing a particular object, thereby failing to investigate the selfas-subject, i.e., the self that is involved in the conscious processing of any content whatsoever. Because it focuses on the former and excludes the latter, the content-approach is limited and misleading. Moreover, the content-approach focuses on those features that allow me to be identied as myself from a third-person perspective. I can refer to an object by way of a proper name, a demonstrative, or a denite description, and occasionally this object is myself. When I refer to myself in this way, I am referring to myself in exactly the same way that I can refer to others, and others can refer to me, the only dierence being that I am the one doing it, thus making the reference a self-reference. Apart from being contingent, this kind of selfreference is also insucient to account for self-consciousness, since it can occur without my knowledge of it, that is, I can refer to myself from the third-person perspective without realizing that I myself am the referent (as illustrated by E. Mach story recalled above). Because it focuses on this third-personal access to oneself and excludes a specic consideration of the rst-person perspective, the content-approach is of limited value. What needs to be emphasized is that the question of self-awareness is not primarily a question of a specic what, but of a unique how. It does not concern the specic content of an experience, but its unique mode of givenness (Zahavi, 2005, p. 204). 2.2. The baseline-approach In the current framework, the content-approach is problematic because of its identication of the self with some kind of self-specic content. It might therefore seem that a better alternative approach should detach the self from the processing of any type of content. However, this approach is problematic as well. More in detail, according to a recent suggestion, one crucial neuronal signature of the self could be the baseline cerebral activity that increases whenever the subject does not focus on a specic external content (rest condition), and decreases whenever the subject is engaged in some task. It has been hypothesized that the resting state can be considered as an ultimate state of inspection of the self (Wicker et al., 2003, p. 229), and that being self-conscious corresponds to a default mode that is ongoing unless attenuated during the performance of an attentionally demanding task (Gusnard, Akbudak, Shulman, & Raichle, 2001, p. 4262). Self-related activity could consequently be characterized by stimulus independent (self-related) thought (Gusnard, 2005, p. 680), also referred to as uncontrolled self-referential or introspectively oriented mental activity (Gusnard et al., 2001, p. 4263). The baseline-approach may seem related to the present consideration of pre-reective self-consciousness in that it conceives of the self at a basic level. However, it diers from it in two crucial ways. First, this approach

590

D. Legrand / Consciousness and Cognition 16 (2007) 583599

invokes a conception of the self that makes it strongly linked to an introspective perspective that can be either spontaneous or explicitly triggered by an attentional task. It is argued that, by contrast with tasks where the subjects direct their attention towards external stimuli, the rest state can . . . be considered as a task where subjects had to direct their attention towards internal processes (Wicker et al., 2003, p. 229). In other terms, during introspection, the subject takes his own states as an object of conscious attention, thereby focusing on the self-as-object. On this point, the baseline-approach is in line with the content-approach, the major dierence being that the former recognizes the importance of self-directed cognitive processes during the so-called resting state. Just like the content-approach, the baseline-approach consequently leaves the very subjectivity of phenomenal consciousness unexplored. Second, the baseline-approach operates with a rather peculiar resting self. Any time it represents the world, the resources allocated to its experience of itself would decrease. Its experience of itself would thus compete with and contrast with its experience of the world. According to this view, self-awareness requires the ability to escape from ambient stimulus control so that attention may be disengaged from the salient stimulus eld and shifted internally to the processing of representations about the self, its attributes and experiences (Gusnard, 2005, p. 689). Surely, while I experience objects in the world, I do not need to reectively experience myself as a particular object. However, it would be a mistake to conclude from this observation that the self is not experienced at all when we experience the world, and that self-consciousness corresponds to consciousness of a self withdrawn from the experienced world. Rather, and as developed earlier, the phenomenology of conscious experience emphasizes the unity of world-awareness and self-awareness. When perceiving an object, one is aware of the object as appearing in a determinate manner to oneself. Furthermore, consciousness cannot be equated with object-consciousness. It is thus misleading to consider that there is no self-consciousness when the subject does not thematically experience himself but only the external world. Rather, our experience of the world includes a pre-reective consciousness of the self-as-subject. The latter does not compete nor contrast with the former, since the mineness or for-me-ness of experience simply amounts to the rst-personal givenness of its objects, i.e. to the rstpersonal way of processing its content: Experiences are both and at the same time subjective and intentional. In short, the baseline-approach fails to recognize that the self-as-subject is being-in-the-world. It disregards the fact that self-experience and world-experience are two sides of the same coin. We experience ourselves pre-reectively precisely when we consciously experience some intentional object. The characterization of the resting-self as being somehow in competition with the world is phenomenologically unsound and thus inadequate for the present purpose. 2.3. The perspective-approach The content-approach mistakenly considers only the self-as-object and the baseline-approach fails to do justice to the fact that the self-as-subject is an intentional subject (a world-directed subject). To overcome these problems, a promising approach would be one that sought to clarify self-experience by focusing on the rstperson perspective. Indeed, the latter involves both a subject holding a perspective, and a world on which this perspective is held. Thus, this approach is relevant to consider at one and the same time the self as a subject and its involvement into the world. Such an account has recently been empirically implemented using neuroimagery (David et al., 2006; Ruby & Decety, 2001, 2003, 2004). In contrast with the content approach, the distinguishing characteristic of the perspective-approach is that the purpose of the dierent experimental conditions is not to contrast dierent contents but rather to focus on situations where the same stimuli is represented either from the subjects own perspective or from a third-person perspective (e.g. the subject is asked either what can you see? or what can he see?). It is worth noticing, however, that even in the case where the subject adopts the so-called third-person perspective, he accomplishes the task from a perspective that remains his own (Gallagher, 2001, note 18). The experiments involve a contrast between a rst-person perspective anchored in the subjects current position and a rst-person perspective moved towards the other (third) persons position. Thus, strictly speaking, they do not involve a contrast between a rst-person perspective and a perspective that would not be rstpersonal.

D. Legrand / Consciousness and Cognition 16 (2007) 583599

591

Moreover, it should be underlined that the perspective-approach has been articulated in recent cognitive neuroscience in a way that is particularly inadequate when it comes to an investigation of pre-reective self-consciousness. More specically, despite its claim that the ability to take a rst-person-perspective in contrast to a third-person-perspective is an essential constituent of human selfconsciousness (Newen & Vogeley, 2003, p. 530), this view in fact conceives of the specicity of self-consciousness not in terms of a rst-person perspective (1PP) per se, but in terms of a self-ascription or reference to 1PP (Newen & Vogeley, 2003, p. 540). In this view, self-consciousness would imply the attribution to oneself of a perspective (hence a rst-person perspective) and this process is thought to require the ability to refer to our body schema representation in the brain (Vogeley & Fink, 2003, p. 40). Possessing a rst-person perspective would consequently entail having a literally spatial model of ones own body, upon which the experiential space is centered (Newen & Vogeley, 2003, p. 538; Vogeley & Fink, 2003, p. 40). This approach consequently conceives of the rst-person perspective as involving the representation of ones own (bodily) states in relation to some world model: The specic subjective perspectivalness in the rst-person-account is realized by the integration of both the subject and the world model as the two main constituents of the internal representation framework in our nervous system (Newen & Vogeley, 2003, p. 541). These remarks clarify that this version of the perspective-approach defends a representationalist conception of self-consciousness. Such representationalism relies on the content-approach discussed above in that it considers that the rst-person perspective itself has to be part of the content of the representations of what is experienced from this rst-person perspective. Therefore, for reasons already mentioned above, this account of the rst-person perspective contrasts with the view defended here. When we are pre-reectively self-conscious, we are aware of the world from our own perspective, we are not reectively conscious of our own perspective on the world (Legrand, 2003). What is crucial to better understand pre-reective self-consciousness is that even when we are intentionally directed at the world and not at ourselves, we remain pre-reectively self-conscious. This point also coheres with Evans following description of an immediate awareness of oneself as an information-gainer and experiencing subject: It is of the utmost importance to appreciate that in order to understand the self-ascription of experience we need to postulate no special faculty of inner sense or internal self-scanningldots For what we are aware of, when we know that we see a tree, is nothing but a tree. In fact, we only have to be aware of some state of the world in order to be in a position to make an assertion about ourselves (1982, 2301). From a phenomenological perspective, it is of the utmost importance to realize that pre-reective selfconsciousness does not necessarily involve the processing of self-specic contents; it might be nothing but the rst-personal processing of non self-specic contents (Legrand, 2006; Ruby & Legrand, in press). To consciously perceive a tree is already to be pre-reectively self-conscious: it involves self-consciousness in the sense that the tree is perceptually given to me, there is a conscious rst-person-perspective, a dimension of mineness. This self-consciousness is pre-reective since the intentional object of the experience is the tree rather than oneself. In its representationalist formulation, the perspective-approach does not consider the possibility of such self-consciousness without self-representation. Therefore, it fails to do justice to pre-reective self-consciousness. 2.4. The core-self In the current framework, Damasios position is worth considering in its specicity since it is both very close and crucially dierent from the present proposal. Our respective approaches to subjectivity cohere on several points. First, as a counter-position to attempts to eliminate the self, Rudrauf and Damasio (2006) recently claim that the idea of a perceiving or feeling subject interacting with an object, which partly constitutes the content of its experience, has some phenomenological validity, and should not consequently be rejected as a description of what it is like to perceive and feel. As a phenomenon, it constitutes an object of interest for science and has to be explained, not dismissed (431).

592

D. Legrand / Consciousness and Cognition 16 (2007) 583599

Second, and more specically, Damasio points out that what we must explain if we are to address the issue of consciousness is the generation of a sense of self and the generation of the sense that such a self is involved in the process of perceiving the stimulus (Damasio, 1998, 1880). In this sense, his position differs markedly from the baseline-approach, since it explicitly conceives of the self as a world-experiencing subject. Nevertheless, our respective positions dier on other crucial points and most notably on the conception of what subjectivity involves specically. According to Rudrauf & Damasio, feeling and the core consciousness are one and the same. (In the following we will often use core consciousness, consciousness, self, sentience, and subjectivity interchangeably.) (Rudrauf & Damasio, 2006, p. 423). Subjectivity is feeling and feeling is related to the sensing of body changes and constraints (429). In other terms, what grounds subjectivity is the feeling of ones own (bodily) states. More specically, even if they argue that the biological substrate of feeling is . . . certainly not conned to seemingly passive brain representations of body states (430), they nonetheless maintain that an essential constituent of the phenomenology of subjectivity is the experience of a large range of body state changes and body state constraints, such as those occurring during the unfolding of an emotion, and, more generally, pain- and pleasure-related behaviours (429). Therefore, even if they avoid the problematic baseline-approach and representationalist perspectiveapproach, they nevertheless consider that self-consciousness entails the processing of self-specic information. Just like the content-approach identies self-consciousness with an awareness of the self-as-object, Damasio claims that the core-self emerges thanks to the dynamical processing of states of oneself (Damasio, 1999). In fact, Damasio even claims that living creatures such as ourselves produce core consciousness when our organisms construct images of a part of themselves forming images of something else (Damasio, 1998, 1881, emphasis added). Surely, this approach does not isolate a self-as-object from the rest of the world and one should not deny the importance of such an attempt to pin down the biological basis of this form of self-consciousness. However, Damasios account suggests that core consciousness relies on a (implicit) body image (self-as-intentionalobject), thereby leaving unexplored the form of self-consciousness that involves a rst-personal perspective on non self-specic contents (self-as-subject): his investigation misses the specically pre-reective self-as-subject that I am focusing on here. 3. Empirical relevance of the proposed approach The preceding set of remarks on the major neuroscientic approaches to self-consciousness leads to the following conclusion: their common point is that they conceive of the processing of self-specic contents as a necessary ingredient of self-consciousness. I have underlined above several reasons why this position is problematic, the main reason being that it focuses on consciousness of the self-as-object thereby failing to account for the primary form of self-as-subject experienced at a pre-reective level. At this point, sceptics might argue that the present approach faces an insurmountable diculty in that the very notion of self and self-consciousness at a pre-reective level does not seem to be a scientically addressable question. However, in the next sections, I propose that the consideration of pre-reective self-consciousness in the empirical domain requires only a change of perspective on already available methodologies, in order to ask phenomenologically sound questions, and a change of perspective on already available results, in order to propose phenomenologically sound interpretations. In what follows, I rst propose that some classical experiments developed in the standard framework are not only limited in scope but also importantly misleading. Specically, they cannot be fully understood without a consideration of subjectivity at the pre-reective level: self-recognition (that is the central component of self-consciousness according to standard approaches) is grounded on pre-reective self-consciousness. These considerations will underline the relevance of a ne-grained phenomenological approach to the empirical investigation of self-consciousness, by clarifying their explanandum (what needs to be explained). Moreover, I will also consider the relevance of the present approach at the level of the explanans (what intends to provide an explanation) by showing that the shift away from a narrow focus on the self-as-object can be interestingly implemented at the sub-personal level.

D. Legrand / Consciousness and Cognition 16 (2007) 583599

593

3.1. Self-recognition implies pre-reective self-consciousness Following standard empirical approaches, one way to investigate the self is to consider the process of self-recognition, e.g., the ability to attribute ones action to oneself. In a classical setup, the subject is asked to perform a simple action (e.g. lift his nger). She then receives an indirect visual feedback from a computer monitor or a set of mirrors. This visual feedback can be biased so that in some condition, the experimenters action is presented in place of the subjects action, or the subjects own action is presented with a temporal or angular distortion. The subject is then asked to report if the action seen is her own or not, and as a result of the manipulation she either attributes the seen action to herself or to the experimenter (e.g. Daprati et al., 1997; Farrer et al., 2003). These experiments ask the subject to focus her attention on an action and attribute it to an agent (herself or somebody else). In other terms, this paradigm explicitly conceives of self-consciousness in terms of identication and attribution to oneself of self-specic contents (This is my action). In such experimental conditions (Daprati et al., 1997), normal subjects were able to correctly attribute the action to its proper agent regardless of whether they were observing their own hand or the experimenters hand as long as the two were performing dierent actions. However, they made erroneous attribution of the experimenters hand to themselves in 30% of the cases when the actions performed by the two hands were similar. In order to explain these results, Daprati et al., 1997 dierentiate between the processes that are involved in generating an action and the processes that are involved in generating a judgement about the action; or to put it dierently, they distinguished simply being an agent from reporting that one is an agent. In accordance with this distinction, they then argued that consciousness of ones own actions requires the ability to consciously report being the agent of the action, that is, it involves processes necessary for the generation of a perceptual judgement regarding the action, over and above the processes necessary for simply generating the action itself. To repeat in this particular case what has been said generally above, one important problem of these studies is that their focus on observational consciousness of the self-as-object (object of identication and attribution) does not allow the investigation of the primary form of self-consciousness (consciousness of oneself-as-subject, agent). Now, to fully appreciate the importance of this remark, it must be claried that even if one restricts the scope to consciousness of the self-as-object, the content approach cannot avoid the consideration of the notion of pre-reective self-consciousness. To clarify this point, consider the following question: How can the subject achieve the task? What is self-specic in this task which allows the subject to accurately answer the question is this my action? First, the subject cannot succeed in this task by simply relying on the perceptual appearance of the hand she observes since the distinguishing visual features of the hand have been empirically neutralized: For example, the subject and the experimenter typically wear identical gloves so that visual cues cannot be used for self-recognition. Moreover, the task cannot be solved on the sole basis of proprioceptive information either, since the subjectwho is acting in either conditionwill receive the same proprioceptive information regardless of whether she attributes the action to herself or not. Second, it is important to underline that the appeal to visual observation and cognitive processes of identication and attribution is insucient too. These processes are not self-specic since the subject employs the very same kind of processes both for self-attribution as for other-attribution, i.e. regardless of whether she attributes a visually observed action to herself or to somebody else. At this level, there is no fundamental dierence between seeing ones own hand and attributing it to an agent and seeing the hand of another and attributing it to an agent, so the crucial dierence between self-attribution and other-attribution cannot be based on processes of observation and attribution. Coherently with this observation, the neuronal resources involved in the attribution of features to oneself would overlap with those activated during attribution of features to others, and would also be shared with general evaluative cognitive processing. This suggests that such high-level cognitive processes are not specically involved in self-related tasks (Ruby & Legrand, in press). In short, the crucial dierence between self-attribution and other-attribution cannot be explained simply with reference to the available perceptual information nor to any judgements based on it. Therefore, an account of how we attribute action to ourselves cannot explain self-recognition if it reduces self-consciousness to a judgement on the self-as-(perceptual)-object.

594

D. Legrand / Consciousness and Cognition 16 (2007) 583599

What such an account fails to realize is that ones action and others action do not primarily dier at the level of perceptual judgement but at the pre-reective level, where we are pre-reectively conscious of ourselves as agents when acting, independently of any judgement about the action. At this level, the subject is simply conscious of herself as acting (self-as-subject). According to the view developed above, this corresponds to a pre-reective consciousness of the self-as-subject that does not rely on the perceptual judgement grounding consciousness of the self-as-object (Hanna & Thompson, 2003; Legrand, 2006; Thompson, 2005). Importantly, it only makes sense to ask the subject whether she recognizes her action as her own if it is presupposed that she already experiences herself as acting. If that is correct, it implies that self-recognition presupposes pre-reective consciousness of the self-as-subject. Therefore, standard approaches do not only have too narrow a focus on the self-as-object, thereby failing to account for the self-as-subject. More importantly the point here is that even accounts of reective self-recognition need to consider pre-reective self-consciousness since the former is grounded on the latter. 3.2. From representations of the self-as-object to self-related processes So far, I have considered the personal level of description and illustrated the relevance of a ne-grained phenomenological account of the explanandum. In what follows, I will consider to what extent these considerations are also applicable to the explanans, at the sub-personal level. Classical empirical investigations describe mental states (e.g. reective consciousness of the self) as involving mental representations (e.g. representation of the self-as-object) and correlates the latter with neuronal representations, themselves corresponding to neuronal activations (e.g. the prefrontal cortex). This methodology is obviously unsuited for investigating pre-reective consciousness of the self-as-subject since the latter does not involve any representation of the self-as-object. However, there are other ways to approach the pre-reective dimension from a neuroscientic perspective. To clarify this point, let me recall the distinction between re-aerent and ex-aerent signals (Ruby & Legrand, in press). The latter amounts to signals coming from the external world while the former amounts to signals issuing from the perceiving subjects own action. The relevant point for the issue at stake here is that a given perceptual signal will be processed dierently depending on whether it is re-aerent or ex-aerent, i.e., whether or not it is directly related to oneself as a perceiving agent/active perceiver. The interesting fact is that there is no way to dene what a re-aerence is (by contrast with an ex-aerence) without mentioning the fact that it is related to the perceiving subjects own action. In other words, there is nothing like a non self-related re-aerence. At the sub-personal level, the notion of self-relatedness captures the fact that ones perception is related to ones own (hence the use of the term self) action. In what follows, I will use the notion of self-relatedness only to refer to processes at the sub-personal level of description. Moreover, it should be clear that the mere use of the term self at this level does not suce to conclude that pre-reective self-consciousness emerges from the processing of aerent signals in a self-related manner. Rather, what I want to underline is that the present approach proves to have repercussions both at the personal level (allowing a shift of focus from consciousness of the self-as-object to consciousness of the self-as-subject) and at the sub-personal level (allowing a shift of focus from representation of the self-as-object to self-related processes that are not selfrepresentational). To see the neuroscientic relevance of the notion of self-related processes, let me now return to the experiments on self-attribution of actions described above. When I perceive my own action and when the feedback is not biased, the aerent information is not a mere ex-aerence but a re-aerence, i.e., it is a consequence of the execution of my action. By contrast, when I perceive anothers action or when the feedback from my own action is biased, the aerent information is not re-aerent but ex-aerent. The proposal here is that the subject accurately perceives his hand as his own if he can accurately dierentiate between self-induced (re-aerent) and externally-induced (ex-aerent) perceptual information, i.e. if the feedback is accurately processed as a perceptual consequence of his action. If this hypothesis is correct, an experimental manipulation of the aerent feedback that disrupted the link between perception and action thereby transforming a re-aerence into an ex-aerence would also impede the subjects recognition of her action as her own. This is exactly what the experiments on the self-attribution of

D. Legrand / Consciousness and Cognition 16 (2007) 583599

595

actions show. When the experimenter wants to inuence the accuracy of the subjects reply to the experimental question (is this your action; are you observing your own action?) he in fact does not manipulate self-specic content. Rather, he specically manipulates the relation between the subjects action and her perception (e.g. he introduces a temporal or spatial bias between the action and the perceptual feedback). When the perceptual feedback is not biased, it is processed as a re-aerence and leads to self-recognition. When the perceptual feedback is biased, it is processed as an ex-aerence and leads to the attribution of the observed action to another agent. In other words, the subjects answer to the experimental task specically depends on the selfrelatedness of his perceptual experience. This consideration implies that the classical experiments on the self-attribution of action are importantly ambiguous: they interpret their own results as related to the representation and self-ascription of self-specic features, thereby conning their investigation to the framework of consciousness of the self-as-object. However, they in fact manipulate non self-representational processes, specically the self-relatedness of ones perception and action. Of course, this ambiguity is problematic in that it precludes any clear-cut interpretation of the results. But interestingly, it also coheres with the hypothesis that in these experiments, what distinguishes self from others (ones action from others) is not only a self-specic content, but also and crucially a self-related way of processing this content: ones own actions only, and not others actions, specically generate re-aerent information. Therefore, I propose that the ability to report that an observed action is ones own is grounded on the ability to process the content of ones perception in a self-related manner (in this particular case, related to ones own action). This process grounds self-recognition on a process that does not involve representation of the self-as-object but rather involves self-related processing of contents, while these contents are not intrinsically self-specic. The present approach proposes to complement the consideration of the self-as-object with a consideration of the self-as-subject. At the personal level, it allows me to describe the specic state of pre-reective self-consciousness. At the sub-personal level, it allows me to consider self-related processes that are non self-representational. So far, this cannot justify any reduction of one level of investigation to the other. However, note that from the two previous sections, it appears that these two levels are correlated with each other: at the personal level, self-recognition is grounded on the pre-reective experience of the self-as-subject, which is correlated with non self-representational self-related processing at the sub-personal level. This conception of pre-reective self-consciousness (at the personal level) and of self-related processes (at the sub-personal level) does not only allow the reinterpretation of these particular experiments on explicit selfrecognition. Most importantly, two additional points needs to be underlined, that I will develop in the next sections. First, implicit self-recognition (modulating the subjects behaviour without explicit reports) corresponds to cases of recognition of the self-as-object just like explicit (verbally reported) self-recognition. As such, it diers from but is grounded on pre-reective self-consciousness and self-related processes. Second, pre-reective self-consciousness and non self-representational self-related processes are not only involved in explicit and implicit self-recognition but are also found at a more primary level that does not rely on any self-recognition (i.e. on any form of consciousness of the self-as-object). 3.3. The pre-reective self in developmental psychology The relevance of the present approach is not restricted to cognitive neurosciences. Due to space constraints, I will only take one example in another eld, namely developmental psychology, and show that the same approach holds. This will allow the consideration of more basic forms of self-consciousness. All healthy infants have an innate rooting response. When the corner of the infants mouth is touched, the infant turns her head and opens her mouth toward the stimulation. Rochat and Hespos (1997) have observed this response in two contrasted conditions: a self-stimulation (a double touch where hand and cheek touch each other) and an external stimulation (a simple touch where only the cheek is touched by an external stimulus). Newborns (24-h-old) showed the rooting reex three times more frequently in response to an external stimulation than in response to a self-stimulation. For the authors, such discrimination [between self-stimulation and external stimulation] is fundamentally self-specifying as it involves proprioception, a perceptual system that conveys rst and foremost information about the body and its situation in the environment. Proprioception in conjunction with other perceptual systems, is indeed the modality of the self par excellence

596

D. Legrand / Consciousness and Cognition 16 (2007) 583599

(Rochat & Striano, 2000, pp. 5167). Their hypothesis is that the sense of self would rely on a multi-sensorial integration characterized by an invariant correlation of exteroceptive information (touch in the case reported above) with proprioception (due to the touching movement performed only in the case of the double-touch). This invariant multi-sensory correlation is specic to the self, and would thus be self-specifying. Infants would be able to pick up these intermodal invariants that specify self- versus nonself-stimulation, and thereby possess the ability to develop an early sense of self (Rochat, 2001, pp. 4041). The problem with this account is that this multi-sensory hypothesis leaves open the following questions: how is the redundancy of sensory information recognized as such, and how is it recognized as self-specic? Recording and explaining that the child behaviourally reacts in distinctive ways to simple-touch and double-touch does not suce to account for the childs experience of the latter as self-related. One reason why this question remains unexplored is that this account (among others, like Sterns and Neissers) conceives of the embodied self as an object, and of embodied self-experience as a kind of object-awareness (Zahavi, 2005, p. 204). Indeed, their interpretation focuses exclusively on the determination of self-specic contents, which they characterize by the redundancy of sensory information about the same thing. In contrast, phenomenology has insisted that rst-personal experience presents me with an immediate, nonobjectifying and nonobservational access to myself (Zahavi, 2005, p. 204). While Rochat dierentiates the experience of self from the both the experience of objects and of other people (Rochat, 2001, p. 27), he nonetheless describes the body as an object of exploration and self-perception as a question of dierentiating ones own body from other objects in the environment (Rochat, 2001, pp. 34, 37). This reduces self-experience to the ability to discriminate correctly between two dierent objects. However, this mistakenly equates self-experience with object-identication and to recapitulate a point made earlier, it is importantly mistaken to identify pre-reective consciousness of the self-as-subject with some kind of peripheral, marginal or merely implicit object-consciousness. Now considering the sub-personal level of investigation, can the notion of self-related processes described above allow for a better understanding of the mechanisms involved in recognizing certain features as ones own? Let me propose a positive answer to this question by considering the following experimental result: Rochat (1998) showed that 3-month-old infants are able to dierentiate between their own moving legs and the moving legs of other infants (both were presented to them on a TV monitor), even when the legs look alike. Rochat explained this result by proposing that only the perception of ones own legs involves a visuo-proprioceptive coherence. But again the question remains: what does allow the infant to experience this coherence as self-specic? In line with the hypothesis presented above, I propose that the reason why vision and proprioception are integrated when one perceives ones leg moving, and not when one perceives anothers leg moving, is because the visual and the proprioceptive contents are both generated by the subjects own action only in the case of self-perception: in these experiments, what makes some contents (rather than others) recognized as self-related is that they are self-generated rather than externally-generated. In fact, the crucial dierence between perceiving oneself and perceiving others is not purely sensory but sensory-motor: what is self-specic is not a multisensory redundancy but a sensori-motor coherence. In other terms, an aerent information processed as re-aerent rather than as ex-aerent does not involve representation of the self at the sub-personal level but it is nonetheless processed in a self-related manner, and it is correlated with self-recognition, i.e. with a representation of oneself as such, at the personal level. I consequently suggest that the empirical data reported here support the view that reexive self-recognition relies on pre-reective self-consciousness and hypothesize that this is matched at the sub-personal level by self-related processes that do not involve the representation of self-specic contents. 3.4. Experiencing the world in a self-related manner The preceding sections proposed that the representation of the self-as-object is not sucient to determine self-specicity: pre-reective self-consciousness and self-related processes are necessary at the personal and sub-personal level, respectively. Now I would like to defend that the representation of the self-as-object is not necessary either: pre-reective self-consciousness and self-related processes are involved in situations where no representation of the self-as-object occurs.

D. Legrand / Consciousness and Cognition 16 (2007) 583599

597

The hypothesis proposed above is that one way a perceptual content becomes self-specic and allows selfrecognition is thanks to self-related processes that link the content of ones perception to ones action. There is no reason to think that such processes are conned to features characterizing oneself (ones action, ones legs, ones face. . .). Rather, the perception of non-self stimuli can be processed in a self-related manner as well. For example, it has been shown (Rochat & Striano, 1999) that two-month-old children react dierently depending on the perceptual consequence of their actions. In an experiment, children would suck on a dummy pacier and each time they did so, they would hear dierent sounds. Depending on the experimental conditions, the auditory feedbacks were either analogue or non-analogue to the sucking. Results show that children modulate their sucking dierently in the analogue condition, i.e., when the perceptual content was self-initiated. In the same vein, it has been shown with adult subjects that the temporal structure of ones auditory perception depends on whether the sounds are produced as an outcome of ones intentional action or not (Haggard, Clark, & Kalogeras, 2002). This eect is called intentional binding since it does not occur with passive movement but only with intentional actions, and since it binds together in time the intentional action and its sensory outcome. At the sub-personal level, these results imply that perceptual contents are processed dierently when the subject is involved in their formation. However, no representation of the self-as-object is involved. In parallel, at the personal level, the fact that the experience of the perceptual content is modulated when it is inuenced by ones own action suggests the presence of a minimal form of self-consciousness, namely a pre-reective sense of oneself as an active perceiver. This form of self-consciousness is non-objectifying, since the represented content is not self-specic: here the object of attention is the sound rather than the subject herself. The framework developed here thus allows for an investigation of one of the crucial features of the pre-reective self: without taking herself as an intentional object, the subject remains selfconscious when intentionally experiencing the world: As defended above, subjectivity and intentionality are closely intertwined. These considerations open an interesting question: What allows the distinction between self-related processing correlated to self-recognition (of ones body, ones action, etc.), on the one hand, and self-related processing correlated to the rst-personal experience of non-self stimuli (e.g. sounds), on the other? One possibility, coherent with the empirical results recalled above, is that, at this level, proprioception would play a crucial role. If this hypothesis were correct, it would imply that proprioception, by itself or in conjunction with other sensory modalities, might not be sucient but might nonetheless be necessary to allow self-recognition. 4. Conclusion Neither correlation nor isomorphism between the personal and the sub-personal levels allow any easy reduction or even explanation of the former by the latter. Nevertheless, the present investigation illustrates the experiential and empirical relevance of the phenomenological approach (see also Gallagher & Sorensen, 2005). The notion of pre-reective self-consciousness seems crucial when it comes to characterizing the experiential states that ground the very possibility of self-recognition. Moreover, at the sub-personal level this approach leads to draw a fundamental distinction between representations of the self-as-object and non self-representational self-related processes. This approach intends to provide a framework that allows the development of a more complete theory of self-consciousness that anchors subjectivity to intentionality and that is applicable both at the personal and the sub-personal levels, thereby meeting both with phenomenological and naturalistic constraints. Acknowledgments I express my gratitude to Dan Zahavi for his patient assistance and his critical appraisal. Thanks to Samantha Perez for her linguistic help. References
Armstrong, D. M. (1968). A materialist theory of the mind. London: Routledge and Kegan Paul.

598

D. Legrand / Consciousness and Cognition 16 (2007) 583599

Block, N. (1997). On a confusion about a function of consciousness. In N. Block, O. Flanagan, & G. Guzeldere (Eds.), The nature of consciousness (pp. 375415). Cambridge, Mass: MIT Press. Carruthers, P. (1996). Language thought and consciousness. Cambridge: Cambridge University Press. Damasio, A. R. (1998). Investigating the biology of consciousness. Transactions of the Royal Society (London), 353, 18791882. Damasio, A. R. (1999). The feeling of what happens: Body and emotion in the making of consciousness. New York: Harcourt. Daprati, E., Franck, N., Georgie, N., Proust, J., Pacherie, E., Dalery, J., et al. (1997). Looking for the agent: An investigation into consciousness of action and self consciousness in schizophrenic patients. Cognition, 65, 7186. David, N., Bewernick, B. H., Cohen, M. X., Newen, A., Lux, S., Fink, G. R., et al. (2006). Neural representations of self versus other: Visual-spatial perspective taking and agency in a virtual ball-tossing game. Journal of Cognitive Neuroscience, 898910. Dretske, F. (1995). Naturalizing the mind. Cambridge, Mass: MIT Press. Epstein, R., Lanza, R. P., & Skinner, B. F. (1981). Self-awareness in the pigeon. Science, 212, 695696. Evans, C. O. (1970). The subject of consciousness. London: George Allen & Unwin. Evans, G. (1982). Varieties of reference. Oxford: Oxford University Press. Farrer, C., Franck, N., Georgie, N., Frith, C. D., Decety, J., & Jeannerod, M. (2003). Modulating the experience of agency: A positron emission tomography study. Neuroimage, 18, 324333. Frankfurt, H. G. (1988). The importance of what we care about. Cambridge: Cambridge University Press. Gallagher, S. (2001). The practice of mind: Theory, simulation or interaction? Journal of Consciousness Studies, 8(57), 83108. Gallagher, S., & Sorensen, J. (2005). Experimenting with phenomenology. Consciousness and Cognition, 15(1), 119134. Gallup, G. G. (1970). Chimpanzees: Self-recognition. Science, 167, 8687. Gallup, G., & Povinelli, D. J. (1998). Can animals empathize? Scientic American Presents, 9, 6676. Gillihan, S. J., & Farah, M. J. (2005). Is self special? A critical review of evidence from experimental psychology and cognitive neuroscience. Psychological Bulletin, 131, 7697. Gusnard, D. A. (2005). Being a self: Considerations from functional imaging. Consciousness and Cognition, 14, 679697. Gusnard, D. A., Akbudak, E., Shulman, G. L., & Raichle, M. E. (2001). Medial prefrontal cortex and self-referential mental activity: Relation to a default mode of brain function. The Proceedings of the National Academy of Sciences Online (US), 98, 42594264. Haggard, P., Clark, S., & Kalogeras, J. (2002). Voluntary action and conscious awareness. Nature Neuroscience, 5(4), 382385. Hanna, R., & Thompson, E. (2003). The mindbodybody problem. Theoria et Historia Scientiarum, 7, 2444. Hume, D. (1739). Teatrise of human nature. Oxford: Wiley. Husserl, E. (1984a). Logische Untersuchungen II, Husserliana XIX/1-2. Den Haag: Martinus Nijho; Logical Investigations I-II. Trans. J.N. Findlay. London: Routledge, 2001. Husserl, E. (1984b). Einleitung in die Logik und Erkenntnistheorie. Vorlesungen 1906/07. Dordrecht: Martinus Nijho. Itakura, S. (2001). Level of self-knowledge in nonhuman primates. In T. Matsuzawa (Ed.), Primate foundation of human behavior. Springer. Legrand, D. (2003). How not to nd the neural signature of self-consciousness (Commentary on Newen & Vogeley). Consciousness and Cognition, 12, 544546. Legrand, D. (2005). Transparenty oneself. A commentary on Metzinger, being no one. Psyche, 11(5). Legrand, D. (2006). The bodily self. The sensori-motor roots of pre-reexive self consciousness. Phenomenology and the Cognitive Sciences(5), 89118. Legrand, D. (in press). Naturalizing the acting self: Subjective vs. anonymous agency. Philosophical Psychology. Lewis, M. (2003). The development of self-consciousness. In J. Roessler & N. Eilan (Eds.), Agency and self-awareness (pp. 275295). Oxford: Oxford University Press. Lycan, W. G. (1987). Consciousness. CambridgeMass: MIT Press. Merleau-Ponty, M. (1962). Phenomenology of perception. London, New-York: Routledge, Trans. Colin Smith. Moore, G. E. (1903). The refutation of idealism. Mind, 12, 433453. Newen, A., & Vogeley, K. (2003). Self-representation: searching for a neural signature of self-consciousness. Consciousness and Cognition, 12, 529543. Perry, J. (1993). The problem of the essential indexical and other essays. New-York: Oxford University Press. Plotnik, JM., de Waal, FB., & Reiss, D. (2006). Self-recognition in an Asian elephant. Proceedings of the National Academy of Science of the United States of America, 103(45), 1705317057. Povinelli, D. J., Gallup, G. G., Eddy, T. J., Bierschwale, D. T., Engstrom, M. C., Perilloux, H. K., et al. (1997). Chimpanzees recognize themselves in mirrors. Animal behavior, 53, 10831088. Reiss, D., & Marino, L. (2001). Mirror self-recognition in the bottlenose dolphin: A case of cognitive convergence. PNAS, 98(10), 59375942. Rochat, P. (1998). Self-perception and action in infancy. Experimental Brain Research, 123, 102109. Rochat, P. (2001). The infants world. Cambridge: Harvard University Press. Rochat, P., & Hespos, S. J. (1997). Dierential rooting response by neonates: Evidence for an early sense of self. Early Development and Parenting, 6(2), 150.1150.8. Rochat, P., & Striano, T. (1999). Emerging self-exploration by 2 month-old infants. Developmental Science, 2(2), 206218. Rochat, P., & Striano, T. (2000). Perceived self in infancy. Infant Behavior and Development, 23, 513530. Rosenthal, D. (1986). Two concepts of consciousness. Philosophical Studies(94/3), 329359. Ruby, P., & Decety, J. (2001). Eect of subjective perspective taking during simulation of action: A PET investigation of agency. Nature Neuroscience, 4, 546550.

D. Legrand / Consciousness and Cognition 16 (2007) 583599

599

Ruby, P., & Decety, J. (2003). What you believe versus what you think they believe: A neuroimaging study of conceptual perspectivetaking. European Journal of Neuroscience, 17, 24752480. Ruby, P., & Decety, J. (2004). How would you feel versus how do you think she would feel? A neuroimaging study of perspective-taking with social emotions. Journal of Cognitive Neuroscience, 16, 988999. Ruby, P., & Legrand, D. (in press). Neuroimaging the self? In Y. Rossetti, P. Haggard & M. Kawato (Eds.), Sensorimotor foundations of higher cognition (22nd Attention & Performance Meeting). Oxford University Press. Rudrauf, D., & Damasio, A. R. (2006). The biological basis of subjectivity: A hypothesis. In U. Kriegel & K. Williford (Eds.), Selfrepresentational approaches to consciousness (pp. 423464). Cambridge MA: The MIT Press. Sartre, J.-P. (1936). La transcendance de lego. Paris: Vrin. Sartre, J.-P. (1943). Letre et le neant. Paris: Gallimard. Searle, J. (2005). The self as a problem in philosophy and neurobiology. In T. E. Feinberg & J. P. Keenan (Eds.), The lost self: Pathologies of the brain and identity. Oxford: Oxford University Press. Shoemaker, S. (1968). Self-reference and self-awareness. The Journal of Philosophy LXV, 556579. TaylorParker, S., Mitchell, R. W., & Boccia, M. L. (Eds.). (2006). Self-awareness in animals and humans: Developmental perspectives. Cambridge: Cambridge University Press. Thompson, E. (2005). Sensorimotor subjectivity and the enactive approach to experience. Phenomenology and the Cognitive Sciences, 4, 407427. Tye, M. (1995). Ten problems of consciousness. Cambridge MA: The MIT Press. Vogeley, K., & Fink, G. (2003). Neural correlates of rst-person-perspective. Trends in Cognitive Science, 7, 3842. Wicker, B., Keysers, C., Plailly, J., Royet, J. P., Gallese, V., & Rizzolatti, G. (2003). Both of us disgusted in My insula: The common neural basis of seeing and feeling disgust. Neuron, 40, 655664. Williford, K. (2006). The self-representational structure of consciousness. In U. Kriegel & K. Williford (Eds.), Self-representational approaches to consciousness (pp. 111142). Cambridge MA: The MIT Press. Wittgenstein, L. (1958). Preliminary studies for the Philosophical Investigations. Blue and Brown Books. Gallimard (Ed. 1996). Zahavi, D. (1999). Self-awareness and alterity: A phenomenological investigation. Evanston: Northwestern University Press. Zahavi, D. (2004). Back to Brentano? Journal of Consciousness Studies, 6687. Zahavi, D. (2005). Subjectivity and selfhood: Investigating the rst-person perspective. Cambridge MA: The MIT Press. Zahavi, D. (2006). Two takes on a one-level account of consciousness. Psyche(12/2), 19.