S. Gallagher (forthcoming). Phenomenological contributions to a theory of social cognition. Husserl Studies. The version available here is the presented text of the Gurwitsch Lecture, without full references.
Phenomenological contributions to a theory of social cognition
The Aron Gurwitsch Memorial Lecture
Society for Phenomenology and Existential Philosophy
Boston, November 7, 2003
Philosophy and Cognitive Sciences
University of Central Florida
Hidden away in the remote corners of one of the largest parts of Husserl's Körper, if we can use that word to translate Corpus, there is ein Leib, an animate body of text that reverberates not only with some of Husserl's other little known texts, but also with some of the most recent discoveries in neuroscience. These texts suggest a theory of intersubjectivity, or what psychologists term social cognition. Let me start with a proviso: whether Husserl ever fully settled on this theory is completely open to interpretation. Accordingly, I will leave it to Husserlian scholars to determine whether this is a position that Husserl actually or ultimately embraces, either in the years 1906 through 1913 when he writes some of these texts, or later when he writes other related texts. The texts in question are from Husserl's early years and they are rather tentative and reactive in nature, since he was just beginning to think of these problems, and he was responding to proposals put forward by Lipps and Meinung. That's also why I claim that these texts suggest a theory of social cognition rather than anything like a pure phenomenology of intersubjectivity. Husserl was trying to work out ways to counter the theories of Lipps and Meinung, and his reflections, I think, should count as phenomenologically informed theory.
The texts that I want to call to your attention are found in Husserl's volumes on intersubjectivity and I also want to make reference to the volume on Ding und Raum, and Ideas II. But my title is not "Husserl's contributions"; it's rather "Phenomenological contributions to a theory of social cognition," and I will also be pointing to texts by Max Scheler and Aron Gurwitsch. Here is what I want to claim.
(1) When Husserl wrote the texts in question, he was suggesting a theory that was not identical to, but also not very far from the one that Scheler defended in 1913 and 1923.
(2) Husserl's theory is preferable in its details.
(3) The insights developed by Husserl constitute an important but only partial explanation of social cognition.
(4) Aron Gurwitsch develops another complementary part of the explanation, equally important, but only at the expense of giving up the insights of Husserl and Scheler.
(5) Only when taken together, these two parts do constitute the best explanation of the most basic processes of social cognition and provide an account that is superior to any of the standard accounts in circulation today.
(6) And finally, recent discoveries in the neurosciences corroborate and support these phenomenological accounts.
In the first of the Intersubjectivity volumes Husserl, in texts from 1906-13 (Text 2), suggests that our understanding of others involves processes that happen on the level of bodily sensations, and that this provides access to others that predates or prefigures anything that would involve inference or analogy - and here he is targeting Lipp's (1905) account of empathy which depends on the argument from analogy (Husserl, p. 70). For Husserl, it is not a matter of intellectual inference but a matter of sensations that are unified in or by the animate organism or lived body. But Husserl's thought here is offered more as a question than as a confirmed view. He asks: "Can what effects the unitary lived embodiment [Leiblichkeit] extend itself to the separate and movable bodies
in the spatial world? (p. 33). And here he suggests that the perception of a body as object [Körperwahrnehmung] is in some way different from the perception of a lived body [Leibwahrnehmung]. These suggestions are reminiscent of Husserl's writings in Ding und Raum. There he talks about kinaesthetic sensations that are activated in perception. When I perceive something, the generated hyletic data, or what he calls the aspect data reflective of the particular object, are joined by a corresponding activation of kinaesthetic sensations in my lived body. If Husserl were Merleau-Ponty he would have put it in this way; my body reverberates with the things of the world. What Husserl actually said was:
The system of [hyletic] aspect data [which refer to the constitution of objectivities] is consciously referred to the system of kinaesthetic movements and in the manner of an ontical motivation. (Ms. D 13, I, p. 15) .
[For example], a sensation of eye-posture corresponds to each place in the field of sight (Hua XVI, 170).
Kinaesthesia is the sensory experience of one's own movement. Quite simply, when I move, I have a pre-reflective and recessed sense of moving. Husserl's claim, however, is not simply that we have kinaesthetic sensation when we move, but that we have kinaesthetic sensation when we perceive something - the something that we perceive registers in a certain way within our kinaesthetic system. Let's make this more specific, in line with his remarks on intersubjectivity. When we see someone else act in a certain way, our own kinaesthetic system is activated. This, in part, is what allows us to understand the other person. Moreover, and importantly, this kinaesthetic activation is part of the perceptual process - part of the hyletic processes that underpin the noetic aspect of perception. This is a point that I want to bring into focus. We are talking about transformations that inform the noetic side of experience, although, as we know, this is tightly correlated with noematic content.
The emphasis on kinaesthetic sensation is a theme that we can find in the early texts, But Husserl also comes back to it in his later texts, in the late 20's and the 1930s (Intersubjectivit,y texts 33-36 and 18 repectively). And he develops a certain aspect of this idea in the Fifth Cartesian Meditation, namely, that my own body is given to me kinaesthetically as the center of my egocentric spatial framework, and this contrasts to the givenness, also in the kinaesthtic register, of the other's body - which can move around without re-ordering my own egocentric perspective.
This seems to me to be an excellent and much more detailed account of what Scheler later proposed as a perceptual theory of intersubjectivity, also developed in opposition to Lipp's argument from analogy. Scheler suggested that rather than a process of inference based on an analogy from our own mental and emotional processes to the mental and emotional processes of the other, we actually perceive the mental and emotional states of others.
For we certainly believe ourselves to be directly acquainted with another person's joy in his laughter, with his sorrow and pain in his tears, with his shame in his blushing, with his entreaty in his outstretched hands, with his love in his look of affection, with his rage in the gnashing of his teeth, with his threats in the clenching of his fist, and with the tenor of his thoughts in the sound of his words. (1954, p. 254).
What Scheler leaves relatively undeveloped, and what differentiates his account from the Husserlian one I have outlined, is his conception of the perceptual process involved. We have details of what we perceive - the gestures, behaviors, movements, postures, etc. of others from which we can directly read off their emotions and thoughts. But in Scheler, the status of the perceptual act involves a level of ideation that makes it somewhat mysterious. For Scheler, the other person's body is first and primarily a field of expression (see Schutz, I, p. 177). What Husserl proposes is that the visual perception of others includes not only the visual sensations, but the activation of the kinaesthetic modality, and this activation keys us into what the other is experiencing. So, on this Husserlian view, an intermodal view of perception that is supported by scientific evidence today, the other person's body is first and primarily not a field of expression to thematize visually, but something that interacts with my body. The problem is not solved on the level of the cognitive aspects of visual perception, but on the level of the sensory-motor body.
I want to take these phenomenologically informed insights into some current debates about social cognition or what some philosophers and psychologists call "theory of mind." The standard alternative in the contemporary theory of mind debate is between an approach called "theory theory" and another called simulation theory. Theory theory is so called because it claims that our understanding of others relies on taking a theoretical stance toward them, specifically appealing to a particular theory, namely, folk psychology for our understanding of others. Simulation theory, in contrast, is the grandchild of the argument from analogy (Gordon and Cruz, 2003). It claims that our understanding of the other is based on self-simulating their beliefs, desires, or emotions. I put myself in their place and ask what I would be thinking or feeling, and then I project the results onto them.
I want to focus briefly on simulation theory - elsewhere I've published phenomenological critiques of theory theory (Gallagher 2001, 2004). Simulation theory is relevant here because it appeals to some scientific evidence which can also be enlisted to support the phenomenological insights offered by Husserl and Scheler. Jean-Luc Petit has already linked Husserl's remarks on kinaesthetic reverberations to the recent discovery of mirror neurons in the motor systems of macaque monkeys and humans. Mirror neurons in the pre-motor cortex are activated under two different conditions: when the subject engages in intentional action, and when the subject observes another person engaging in that action. The evidence for this in macaque monkeys is very strong, involving recording from single neurons. In humans the evidence is more indirect, but significantly so: in humans, in addition to PET and fMRI results, one finds higher pre-activation processes in just those muscles that would be engaged in the action at stake. One would get here, at least, a subliminal activation of proprioceptive and kinaesthetic systems.
An example might make this clear. If, watching my daughter play soccer, I see that she is about to make a kick, I may feel my own leg being readied to kick. Indeed, if the ordinary motor inhibition does not "kick in," I may even find myself kicking as if I were in fact kicking the ball that my daughter just missed (or perhaps with which she just scored).
Simulation theorists have cited this as
evidence for a simulation process that takes place non-consciously. That is, they interpret this evidence
as involving three steps: (1) perception - I perceive the other person doing X;
(2) simulation: my own motor system (the mirror neurons) are activated and
contribute to a simulation of what it would be like for me to do what I see the
other person do. (3) By way of
analogy, I then use that simulation to project certain intentions onto the
other person. The question,
however, is where precisely do we draw the line between perception and anything
else that happens in the cognitive system. If, following Husserl, we understand perception to be a non-momentary
act of consciousness, then certainly we can think of an act of perception as
extended over time. Setting aside
questions about how one objectively measures the living or specious present,
almost everyone agrees that the living now of perception is significantly more
than 100 msecs, and perhaps on the order of seconds rather than msecs. But it does not take more than 100
msecs for the mirror neurons to fire subsequent to seeing the action of another
person. Where one draws the line
between perception and non-perceptual mental processes is more complex than
this, but I want to suggest that mirror neuron activation, and more generally,
the activation of a more distributed set of shared neural representations
activated when one acts, or perceives another's action, or imaginatively enacts
another person's action (see Jeannerod, Decety, etc.), is part of the
perceptual process, and contributes to the noetic structure of this kind of
perception. This is quite
consistent with recent philosophical arguments, based on psychological
evidence, that perception and action are not fully separate processes (e.g.,
Susan Hurley). And of course it is quite consistent with Husserl's analysis of kinaesthetic resonance with perceived objects.
In regard to this last point, however, one still needs to ask whether the perception of objects is equivalent to the perception of other persons. Husserl and Scheler suggest that it is not. And if we understand kinaesthetic-motor activation to be part of the perceptual process we have good reason to argue that there is an eidetic difference between the perception of objects and the perception of other persons. Still, to claim that person-perception is different from object-perception is somewhat controversial in the neuropsychological literature. What is well established there, at least in terms of vision, is that visual perception of an object for purposes of recognition (being able to say what it is that you see) is different from visual perception for action (locating and pragmatically relating to something in order to pick it up, for example). To put this into familiar phenomenological terms, to see a cup as something Vorhanden (e.g., as something that you lost and now found) is different from seeing the cup in the context of reaching for it to take a drink - that is, as something Zuhanden. The relevant neuropsychological literature distinguishes between the Ventral visual pathway for recognition and the Dorsal visual pathway for motor action (Milner and Goodale).
Beyond this distinction, however, it would be controversial to suggest that there is a third type of vision, namely, visual perception of other subjects who are neither merely Vorhanden nor merely Zuhanden. Yet there is good evidence in neuroscience that this is the case, for both perception and for action. It is a well-established fact that face recognition depends on the activation of a particular part of the brain (the fusiform gyrus). At the very least we can say that the visual perception of a face is quite a different thing than the visual perception of a tree or a hammer - both neurologically and phenomenologically. Add to this on the neurological side, the research on mirror neurons which are activated only in very specific circumstances when we perceive another animate body (human or in some cases animal) doing an intentional action - and they are not activated if we see that action performed by a tool or machine, or if I simply see an object. Add further evidence developed in neuropsychological studies of movement in cases of apraxia (see Gallagher and Marcel, 1999) and studies of gesture in cases of deafferentation (see Cole, Gallagher, and McNeill, 2001) that suggest that there are qualitative differences in social and communicative actions versus instrumental actions.
Phenomenologically speaking, the claim is that the act of perception, when it is the perception of another person, is different, not only noematically but also noetically (and this is a point I want to emphasize); it is noetically different from when we perceive a tree or a hammer. Kinaesthetically, our experience is different - and our perceptions are shaped differently and have a different feel - when we perceive a tree as something Vorhanden, a hammer as something Zuhanden, and a person as a person. Just as what it is like to perceive an object is qualitatively different from what it is like to perceive an instrument that I am about to use, what it is like to perceive a face is qualitatively and noetically different from what it is like to perceive an object or an instrument. One could appropriate Husserl's notion of "pairing" here, and think of it as a natural paring that takes place from body to body, at the kinaesthetic level, in a way that shapes our perceptual experience. I think we can find resources in Heidegger, Sartre, Levinas, and others, to bear this out.
Even if all of this is right, it is not the complete story of intersubjectivity. To supplement this story in the right way we can turn to the work of Gurwitsch, and specifically his 1931 habilitation thesis, translated into English as Encounters in the Social Environment.
Gurwitsch, like Husserl and Scheler, offers a critique of the argument of inference from analogy. In working out his phenomenological critique, Gurwitsch borrows from developmental psychology and from scientific studies of neuropathology. He also borrows heavily from Heidegger. But he starts by following and expanding on the critique of the argument of inference from analogy offered by Scheler in 1913. One point of this critique shared by Gurwitsch and Scheler, and many other phenomenologists who have considered it, including Schutz and most recently, Dan Zahavi is, I think off-base in an important way. The criticism is as follows: In the argument from analogy, the analogy depends on our ability to compare our own body and bodily behaviors with those of others. Scheler notes that the way my body is experienced by me (e.g., more through kinesthetic-proprioception than through exteroceptive or visual perception) is quite different from the way the body of another person is perceived by me (exclusively through exteroception). If this objection holds, it would challenge the earlier mentioned insights of Husserl. But this objection is questionable. Developmental studies have shown that intermodal perception is operative from the very beginning of life (e.g., Meltzoff and Moore, 1997). What I see is automatically registered in a code that is common to other sense modalities, including proprioception; and in the case of seeing biological movement, perception includes motoric, kinaesthetic activation. So when I see the other's body moving in a certain way, I have a kinesthetic-proprioceptive sense of what that is like.
One thing we can say against the argument from analogy is that if such intermodal links are made at the level of embodied enaction, there is no requirement to reduplicate that link at the level of cognitive inference. That is, to have some sense of what the other is thinking or feeling, we do not have to consciously compare the other with myself, or run through a simulation of the other person's action in our own cognitive system.
We certainly do not encounter an absolutely other person, as it were, a second example of the genus to which I belong and which I interpret by analogy or empathy and only make comprehensible by the 'transfer of my own livingness' and its structure, of my own mental processes and their concatenations such that I conceive it in an entirely determinate way as alter ego. (1931, p. 108).
Rather, our primary encounters are framed in pragmatic contexts. "Prior to all specific cognition, and independent of it, we are concerned with other people in our 'natural living' of daily life; we encounter them in the world in which our daily life occurs" (1931, p. 35). The nature of our concern with others, as this word signifies, involves mixing with others in relations that bear on practical matters and interests. It follows that there is not one uniform way in which we relate to others, but that our relations are mediated through various pragmatic circumstances.
Acknowledging the importance of this insight, I still want to ask whether this gives us the full picture of intersubjectivity. Are we, in fact, caught up in such pragmatic circumstances from the very beginning? One might suggest the example the infant's dependency on others for nourishment - but it is not clear to me that this situation is entirely or primarily pragmatic. The emphasis on the pragmatic is both a positive and a negative for Gurwitsch's analysis. Its positive significance is that it brings out the importance of context and situation for more sophisticated understandings of meaning and the significance of others's behaviors. The upshot of this account is that in our ordinary everyday encounters, we never come upon others as absolutely nondescript others who conceal a hidden realm of beliefs, desires, or intentions that we need to theorize about or try to simulate. How we understand others is implicitly revealed in the pragmatic circumstances that define their specific roles. We understand the intentions of others primarily in what they do in the practical contexts that define our relations to them.
Its negative significance, however, is that it precludes any appeal to intersubjectivity as a primary element in an account of the meaning-consititution of the surrounding world. To put it most succinctly, intersubjectivity is portrayed as emerging out of pragmatic contexts. In the Heideggerian framework, there is a clear and justified distinction between the Vorhandenheit and the Zuhandenheit - between mere objects and instrumental beings.
The "being encountered first of all," under which all that is comprehended with which we deal in our daily living, of which we make use, with which we go about our business and manipulate, is immediately delimited by Heidegger in a radical way from physical things as physical things 'on hand' [present-at-hand] that belong to Nature with the material properties and determinations accruing to them. (Gurwitsch, 1931, p. 66).
But the Mitsein is framed within our Zuhanden encounters. It is through such scenes of instrumental dealings with the surrounding world that we discover an implicit contact with others. The practical context operates as an implicit scaffolding of meaning that rises around us, and shapes our individual encounters. Only when this scaffolding collapses do we find ourselves taking a cognitive and abstract attitude toward the other. "It is in this milieu that we also encounter our fellow human beings. We encounter them because we live in determinate situations of the surrounding world of utensils" (p. 95). Specifically, Gurwitsch suggests that other people are always implicitly on the horizon of our practical dealings with the world. It is the constant and implicit references to others in our pragmatic affairs that generate the inescapable conviction that we live in a human world. This network of implicit references to others is the basis for our experience of being-with, or being together with others. "In these horizonal situations the 'co-included' others appear. That they come to light in this situation, and are not 'near by' or 'merely beside' it, signifies that they appear as belonging to the situation in their specific roles and functions" (p. 97). Here Gurwitsch suggests that our understanding of others is from the beginning in terms of the roles that they play in relation to our projects. "But it is always a matter of a person in his role. Understanding is yielded here by virtue of the situation and is, therefore, limited to what is inherent in it" (p. 114). For Gurwitsch, there are no perceptual or cognitive acts that have special empathic insight -- the explanation is not to be found in "some magical psychical mechanism" (perhaps a way to characterize Scheler's notion of direct perception). As such, however, this theory does not address the question of how our pragmatic understandings of others get off the ground - and specifically, why we are able to discern others as other persons rather than just other pragmatic elements of the context - rather than mere role players. As Gurwitsch himself puts it,
[...] this orientation characterizes this understanding not primarily because it involves other people but because it represents a moment in the knowledge about the immediately actual situation-totality. This orientation grips fellow human beings only insofar as it is a component of the situation. In other words, we are dealing with a functional understanding. (Gurwitsch, 1931, p. 112).
What is left unexplained, however, even in his detailed explication of the appearance of the other on the horizon of our pragmatic projects, the supposed basis for our strong conviction, is precisely how, in the first place, the other appears as another person rather than simply another piece of equipment. It is also not clear how just such pragmatic encounters are necessarily the basis for encounters that are not just pragmatic - emotional and ethical encounters, for example.
I could carry this critique further if time allowed, and I have done so in other papers (Gallagher, 2003, 2004). Like Gurwitsch himself, I could cite evidence from developmental psychology to show in opposition to Gurwitsch, that the solution framed in terms of pragmatic encounters is inadequate. In the developmental literature reference is made to both primary intersubjectivity, which correlates directly to the embodied perceptual, sensory-motor level of analysis that we find in Husserl, and secondary intersubjectivity, which has a direct correlation to the emphasis on pragmatic contexts that we find in Gurwitsch. Secondary intersubjectivity, however, can be shown to depend on an innate primary intersubjectivity which is built into our sensory-motor embodiment, and without the latter, the pragmatic framework of secondary intersubjectivity simply does not get off the ground. Thus, for a full phenomenological theory of intersubjectivity, we need accounts of both primary and secondary intersubjectivity - we need both Husserl-Scheler and Heidegger-Gurwitsch.
I can also bring to bear on this issue studies of pathologies, and specifically studies of recovery from apraxia (Marcel, 199X; Gallagher and Marcel, 1999) and of gesture (Cole, Gallagher, and McNeill, 2001) which show that social comportment is irreducible to pragmatic movement, and which suggest that there is something unique and quite basic to social cognition that cannot be captured in purely instrumental relations.
Since time does not permit further explication along these lines, let me conclude with a summary of what I have argued.
(1) Husserl, in a number of unsystematically developed texts suggested a theory of intersubjectivity or social cognition that was based on an embodied, intermodal perception. This perceptual experience of the other person involved a kinaesthetic reverberation of the perceived action of the other in our own sensory-motor system.
(2) Husserl's analysis explicated certain aspects of a theory that on the surface was close to Scheler's, in details that not only went beyond Scheler's demotion of the body, but find support in contemporary scientific evidence.
(3) The insights developed by Husserl constitute an important but only partial explanation of social cognition - specifically the part that directly contributes to what developmental psychologists term primary intersubjectivity.
(4) Gurwitsch develops, within a Heideggerian framework, another important aspect of what would be a fuller explanation - specifically he emphasizes the contribution of our involvement in pragmatic contexts to gain more sophisticated understandings of others. Unfortunately he does this at the expense of giving up Husserl's insights into the perceptual mechanisms that allow us to key into others in the first place - something, however, that Gurwitsch did not have give up.
(5) When taken together, these two analyses go a long way to constitute the best explanation of the most basic processes of social cognition and provide an account that is superior to any of the standard accounts in circulation today, for example theory theory or simulation theory.
(6) And finally, as mentioned, these phenomenological accounts are fully supported by recent discoveries in the neurosciences that show the importance of mirror neurons, and neural representations that are activated both when I act and I observe someone else act.
There is much more to say, much more evidence to discuss, more phenomenology to do. Let me close, however, by indicating what I suggested was a controversial claim, namely that there is a special kind of perception that is a perception of others and that contrasts with a perception of things and of instruments. This suggestion is controversial in the neurosciences, although there is good evidence for the distinction between a visual perception of things and a visual perception of instruments. If I am right to foreground Husserl's insights into kinaesthetic reverberations, however, I think this is a less controversial claim in phenomenology. I would suggest that there are more phenomenological resources to explore (I'm thinking of Sartre, Merleau-Ponty, Levinas) to make secure the idea that the noetic structure, and not just the noematic structure, of this kind of perception is quite different from the noetic structure of object perception or instrumental perception.
 Gurwitsch does not put phenomenology into opposition with science. Rather, with respect to psychology he describes part of his own project as follows: to endeavor "to interpret phenomenologically certain psychological theories as well as to use them for the advancement of phenomenological problems. Here, psychology is not looked at from without; rather, the psychological theories are received into the philosophical context and are made instruments, not objects, of analysis" (Gurwitsch, 1966, xxi).
 Gurwitsch supports his account with references to developmental psychology and studies of pathology. Specifically, the development of the naming function in the child's acquisition of language (as explicated by C. and W. Stern ) reflects the fact that children are acquainted not with things in their present-at-hand objectivity, but with how one can interact with them and what one can do with them. Similar evidence can be found in cases of aphasia, where the loss of the ability to use proper names for things may lead to the use of functional descriptions to signify (e.g., instead of 'pen' the patient says 'for writing') (Gurwitsch, 1931, pp. 68-70.
 Gurwitsch offers an account of the difference between a tool or instrument and a fellow human being, but only in terms that come too late to count as an answer to how "in the first place" we come to recognize the distinction. The other person is represented as having a freedom to be outside of the situation, whereas the tool has no such possibility (see, 1931, pp. 116-117). This distinction would seem to be something extra or secondary to what we already accomplish in our situational understanding of the other. In other words, to recognize the freedom of the other person (as a set of possibilities outside of the present pragmatic situation) already implies that we have recognized the other person as a person rather than a tool. Inside of the pragmatic situation where we have our primary encounters with others, as Gurwitsch suggests, "the determination of the role [of the other] and the imprinting of its sense by the whole situation is not essentially different from the way in which a utensil maintains the concrete function accruing to it hic et nunc" (p. 117).