Institutional Agency: Department of Philosophy, University of Edinburgh (UK) [website of the Department]

Principal Investigator: Prof. Andy Clark [homepage]

Team members:

• Dr. J. Kallestrup
• Dr. J. Kiverstein
• Dr. M. Nudds
• Dr. T. Vierkant
• Miss Mog Stapleton

External collaborators:

• Prof. P. Calvo Garzon (Murcia, ES)
• Prof. A. Noë (UC Berkeley, USA)
• Prof. M. Rowlands (Hertfordshire, UK)
• Dr. M. Wheeler (Stirling, Scotland)

National Funding Agency: Arts and Humanities Research Council [website of the AHRC]

Weak Interactionism and Active Consciousness

A key hypothesis to be explored by our CRP is that the qualitative character of phenomenal experience can be explained by the skillful capacities an animal draws upon in the sensory exploration of its environment ([31], [33], [38], [41]). Phenomenal consciousness is, we claim, "active" because phenomenal conscious experience is a skilled activity in which the agent has implicit knowledge of the sensory effects of movement. We will call the proposal to treat phenomenal consciousness as active consciousness "what-interactionism" since it purports to explain what we experience - the qualities of experience - by appeal to the dynamic practical interaction of a perceiver and her environment.
IP-02 proposes a variety of interactionism about active consciousness according to which the physical vehicles of conscious experience may support channels of sensorimotor feedback which develop through time in characteristic dynamic patterns ([19], [21], [32], [39)]. This hypothesis complements our (IP03) proposal to treat phenomenal conscious experience as a skillful activity which is mediated by knowledge of the sensory consequences of movement. We will call this proposal "how-interactionism", so-called because it offers an explanation of how phenomenal conscious experience might be physically realised which allows the physical vehicles of consciousness to loop through the environment.
IP-02 proposes to investigate the strongest possible version of interactionism which we will dub "strong interactionism" by combining how with what- interactionism. Our IP seeks to investigate a weaker form of interactionism which endorses what-interactionism whilst denying how-interactionism.

We deny how-interactionism because we believe that there may be neural mechanisms which have been recruited and tuned through agent-environment interaction but whose activation is now sufficient for phenomenal conscious experience. We will argue for how-internalism which claims that once neural routines have been recruited through agent-environment interaction, that routine is then always sufficient for enabling phenomenal consciousness [28]. IP-02 denies that such neural routines are always the best explanation of how phenomenal consciousness is enabled, allowing that sometimes the physical vehicles of a conscious experience may loop into the environment.

The Role of Conscious Perception in the Guidance of Action

The two-visual systems hypothesis ([5], [15], [22], [30]) presents a challenge to what-externalism. Proponents of the two-visual systems hypothesis [Milner and Goodale, 1995] claim that activity in the dorsal stream activity is non-conscious, and that visual phenomenology ‘can arise only from processing in the ventral stream.’([30, p.200]). This suggests a role for conscious perception solely in recognition and the making of perceptual judgements, not in dynamic sensorimotor interactions as what-interactionism claims.
While there are reasons for seriously qualifying the latter hypothesis [see 22] there is nevertheless a serious challenge here for any proposal that seeks to identify qualitative experience with dynamic sensorimotor interactions. Clark [5] argues that conscious perception only plays a role in the reason and memory-based selection of appropriate action, not in the fine-tuned control of action. Clark hypothesises that motor signals work in conjunction with attentional mechanisms to direct sensory systems to the parts of the environment salient to the organism’s needs at the time. However this distinction implies that the bonds between conscious perception and action are a good deal weaker than our hypothesis that phenomenal consciousness is active consciousness would have us believe. Further investigation in collaboration with IP-02 will reveal the extent to which such qualification is required.

The Extended Mind, Consciousness and Cognition

IP-03 endorses how-interactionism about cognition. We subscribe to the extended-mind hypothesis ([4], [6], [7]) which claims that bits of the world are on a par with resources internal to skin and skull when it comes to explaining how humans solve problems.
We allow then that the physical vehicles which enable cognition can cross the boundary of skin and skull to include artefacts. Yet we stop short of endorsing how-interactionism of this kind for consciousness. Why treat consciousness as a different case from cognition? Behind these questions lies a deeper issue of concern to all IP’s in this project: namely the nature of the relation between consciousness and cognition. IP-02 proposes to develop the shared circuits model which argues that the very same simulations employed in off-line cognition may also play a role in sensorimotor interaction. IP-01 describes the developmental route from simulating the actions of others to acquiring a full-fledged theory of other minds. If this process begins with simulations employed in active consciousness, it will follow that there is a dependence of cognition on consciousness (and indeed of interactive consciousness on cognition as described by IP-01). Thus IP-01 and IP-02 put significant pressure on our attempt to isolate consciousness from cognition.

Aim 2: To understand the relation between the bodily self (active consciousness) and the narrative self (interactive consciousness).
Objectives: To write papers on the following topics.

The Bodily Self and Action

IP-01 propose to investigate the origins of our sense of self where the self is conceived of as a coherent narrative. IP-03 proposes to add to this an account of the self as it exists in its interactions with the natural world. We will call this aspect of the self, the bodily self. We believe that active consciousness is a kind of primitive or non-conceptual self-consciousness ([21], [1]). It is the bodily self that has knowledge of the sensory consequences of movement: it understands how its experiences will change with movement, or with changes in background conditions.
In the case of fast and automatic actions it is natural to think of the link between action-guiding perception and bodily movements as direct and as not proceeding via consciousness. However we reject the implication that the information used in the guidance of perception is always non-conscious. One question we wish to pursue concerns the extent to which this information might be employed by the bodily self to give a subject the feeling of an object’s presence. This feeling of the presence of an object is an aspect of the phenomenology of experience. We propose to investigate how the neural mechanisms which enable the perceptual guidance of action may play a role in enabling this aspect of an experience’s phenomenology.

The Bodily Self and the Narrative Self

IP-01 supplies an account of what they call “the narrative self”. What is the nature of the relation between the bodily self and the narrative self? The shared-circuits hypothesis proposed by IP-03 spells out a possible way in which the narrative self might develop out of the bodily self. IP-01 proposes to describe how our sense of self as a coherent narrative has its origins in the internalisation of primarily social cognitive competences. They also describe how these cognitive competences draw upon capacities for simulating in particular, the actions of others. According to the shared-circuits hypothesis the very same simulations are used in the perceptual guidance of action. We have just outlined how the bodily self might be involved in the perceptual guidance of action. It would seem to follow then that together with IP-01 and IP-02 we have the resources for explaining how the narrative self might develop out of the sense we have of ourselves as embodied.
A rival view has the narrative self arising as a solution to the attribution problem of lingual mental states. According to this theory championed by amongst others ([10], [36], [24]), the narrative self came into existence, because early lingual humans needed a locus for their communications in order to make long term planning effective. This locus is the narrative self. Thus we have two competing stories about the relation between the narrative and bodily aspects of the self. One locates the origins of the narrative self in language and thereby stresses a discontinuity between the bodily and narrative aspects of the self. IP-01 and IP-02 by contrast propose an account of the narrative self which stresses its continuity with the bodily self. In dialogue with IP-01 and IP-02 we will attempt to decide between these two accounts.

The Bodily Self and Emotion Experience

IP-04 make a distinction between irreflexive or first-order emotion experience and second-order or reflexive emotion experience. We are interested in the first-order variant of emotion experience and propose to explore the bodily dimension to this variety of emotion experience. According to one influential theory ([23], [9], [34]) emotions are embodied because they are perceptions of various types of bodily change (a feeling of elation, for instance, might be identified with perception of change in heart rate.) We believe that this approach to emotion experience neglects the active and interactive character of emotion experience. Emotions are not simply perceptions of the body (pace James and Damasio [23], [9]): what is experienced is a body’s changing relation to the natural and social world. We propose to characterise emotion experience as forms of skillful engagement with the world. Subjects of emotion experience are dynamically coupled to a world which both influences and is influenced by an unfolding emotion experience. Understanding the dynamic coupling between such a subject and its environment amounts to understanding how the unfolding of an emotion episode may affect the behaviour of other creatures and is in turn shaped by their behaviour.
Emotions have a temporal course of development and can involve an exchange of signals between subjects (e.g. facial expressions, tones of voice etc.) In appraising a situation a subject doesn’t simply receive inputs from the world and then use this information to appraise the situation they confront. Emotions we will argue are action-oriented representations ([7], [17]): the environment is represented in terms of what it affords the creature – what the creature may do or what the situation offers to the creature, or what the creature must do in order to cope with the situation ([11], [12]). IP-04 treat emotions as perceptions of events with emotional meaning. IP-03 suggest that it is for the bodily self that these events have meaning.

Aim 3: To explore the connection between consciousness and the will.
Objectives: To write papers on the following topics.

Two Kinds of Volition

As our empirical understanding of the human mind progresses, the need becomes clearer for a still better understanding of ourselves as conscious, rational, responsible agents. However, these concepts need to be understood in the context of the different levels on which human cognition works, many of which are arguably not subject to direct conscious control (see the two visual systems hypothesis). Our first paper will try to map two very different concepts of volition, which are constantly run together in the current interdisciplinary discussion on ‘willing’ onto the distinction between the narrative and bodily aspects of the self. The first sense of volition, which could be defined as spontaneity, is what is most prominent in the classical experiments by Libet [27], as well as in recent studies by Lau, Rushworth, Prinz and others ([26], [35], [37]). In all these experiments, willing is defined as a capacity to act without external stimuli guiding the action. An action is willed, according to this definition, if the individual has to select the action (or as in the Libet paradigm: the timing of the action) internally without any prompting from the individual’s environment. We will argue that all these actions are triggered in the same way as skillfull behaviours by fast, automatic systems. This fact explains why such actions are felt by the conscious subject to be spontaneous in the sense that the subject could not give a reason why it performed that action at that specific moment. We will examine whether fast, automatic actions of this kind might be said to bypass entirely any involvement from the bodily self. This first sense of volition is often conflated with an understanding of volition as the conscious rational control of one’s actions: it is commonly assumed that an action is under an agent’s conscious rational control iff it is internally generated. We will distinguish conscious rational control from spontaneity. Experiments on this second kind of volition include Wegner’s work as well as studies by Jeannerod, Goschke, Frensch, Knoblich and others ([40], [14], [13], [16], [18]). The focus in all these experiments is on the question of how far our conscious intentions to act and conscious evaluation of our actions are influential in causally bringing about behaviour. Finally we will bring these two different senses of volition into contact with the venerable debate between incompatibilists and compatibilists. We will attempt to map the incompatibilist intuition of the importance of spontaneity on fast implicit decisions of the bodily self, and the compatibilist notion of higher order volitions on the action monitoring function of the narrative self. Finally the paper will try to use its findings to construct an argument for an empirically informed compatibilism.

Autonomy, The Bodily Self and Pathologies of the Self

Emphasizing the role of narration for autonomy risks loosing touch with our deepest values as embodied beings. It is not plausible that we always know explicitly, when we form an intention, whether this intention is really in line with what we deeply believe to be valuable. We might for example convince ourselves that killing people is acceptable, if we do it for the right reasons, but might still refrain from pulling the trigger, if put to the test, because of some inchoate feeling that killing people is always wrong. This problem is used by Michael Bratman [2] to attack theories of autonomy that rely on explicit reasons. According to Bratman, such theories cannot account for these phenomena, because they have to assume that responsibility can only be ascribed if the subject acts on the best explicitly available reason. This paper will try to address Bratman’s complaint. One promising route to answer the challenge might be to introduce the notion of “narrative heuristics”. If the person understands herself as someone who tends to act more in line with what she believes to be the best when she follows her gut feelings, then she has a good heuristic reason to trust her gut feelings.
A second challenge to our conception of autonomy comes from pathologies of the self. It might be the case that in several pathologies (e.g., schizophrenia), it is not narrative monitoring that is disturbed, but a far more fundamental process (see Frith & Blackmore’s feed forward model theory). Nevertheless individuals suffering from these pathologies have a very strong feeling of loss of autonomy; they feel that they are only passive spectators who have nothing to do with the movements of their bodies, even though they are doing what they wanted to do and know cognitively that it is them who are bringing about the movement. As well as presenting a challenge to our claim that it is the narrative self who is autonomous, these findings also seem to provide some support FOR the theory that the narrative self developed out of the bodily self.

The Social Ecology, Autonomy and Reason

The connection between the narrative self and interactive consciousness described by IP-01 entails that there is a social component to autonomy as we describe it. The traditional conception of autonomy as the individual’s capacity to act spontaneously, independently of the social ecology of reason-giving practices has provided social institutions with a rationale for delegating responsibility for intentional action and its consequences exclusively to the individual – for example, in responding to speech [20]. The possibility of recognizing that responsibility for certain effects is often shared among various agents can thus be overlooked, even when this view would be more accurate causally and more effective in serving widely shared aims. For example, recent research on stereotypes using Greenwald's implicit association tests [16] shows that, under specific circumstances, even very unprejudiced persons automatically and unconsciously behave differently towards members of certain minority groups. These unconscious control layers are not present in the classical picture of the responsible human individual. Our conception of responsibility and indeed of autonomy on the other hand can do justice to this distinctive importance of social ecology in human cognition and action. This importance hinges on two factors: On the one hand the narrative self as a feature of the interactive consciousness arises in a social context. In this respect our better understanding of the psychological mechanisms of narrative reasons might entail support for a communitarian understanding of autonomy-constituting reasons. On the other hand, as in the case of stereotypes, our actions are not only influenced on the autonomy-constituting reason-giving level, but as well on the autonomy-undermining subpersonal level.

Empirically informed conceptual analysis which draws in equal measure on past and present scientific and philosophical work.

The University of Edinburgh has always taken pride in its interdisciplinary approach which makes us an ideal department to participate in a project of this kind. The department of Philosophy has strong ties with Psychology and Informatics which we plan to exploit in developing the research outlined above. We will create core research community which will include other members of our department with expertise in Philosophy of Mind and Moral Psychology. 2006/7 will also see the inauguration of the world’s first interdisciplinary Masters programme on Embodied Cognition. The material taught as part of this programme will complement the work of the PhD and Post-Doc we plan to recruit for our project.

Year 1: We will produce articles (single or multi-authored according to emerging interests) on:

• Varieties of Interactionism
• The Dual Visual Systems Challenge
• Two Kinds of Volition

Year 2: Spring 2008 will see Edinburgh host a CONTACT conference on consciousness and the will (to be co-organised by Dr Till Vierkant, Prof Andy Clark, and Dr Julian Kiverstein). Editing by Vierkant and Kiverstein and Clark of collection of papers on Consciousness and the Will plus articles on:

• Locating Consciousness
• The Bodily Self and Action
• Interactive Consciousness and Autonomy

Year 3: Final editing and writing of introduction for volume on Consciousness and Will, send it to press. Articles on:

• Emotion and the Bodily Self
• The Extended Mind and Consciousness
• The Bodily Self, the Narrative Self and Volition

Throughout: Prof Clark, the Post-Doc and PhD will work collaboratively on above topics according to their research interests. Both the post-doc and (to a lesser extent) the PhD will be encouraged to publish papers, Post-Doc will be expected to publish at least one paper per year in addition to work on edited collection. Both will be required to participate in other CRP events. Visits and exchanges with other CRPs will be organised.

Deliverables/Milestones

Primary deliverables will be journal articles and publications in high-profile edited volumes on the above topics. Collaborative papers are also anticipated with AP-01, IP-01, IP-02 and IP-04. A first milestone will be the first of the CONTACT conferences to be hosted at Edinburgh in spring 2008 (see also section on CONTACT meetings). This will form the basis for an edited collection of papers which we expect to be in press by the end of the project. Edinburgh will also host the PPNB graduate/post-doc conference in 2008 (see the section on documentation for further details).