Identities
The View from Nowhere
“True” is a sign that something is to be done,
for inferring is a doing.
(Sellars 1991b, 206)
Philosophy, said Wilfrid Sellars, is the attempt “to understand how things in the broadest possible sense of
the term hang together in the broadest possible sense of
the term.” (Sellars 1991a, 1) Despite its apparent vagueness, this is as good a way of encapsulating the concerns
of philosophy as anyone has ever given, since we can
specify what the “broadest possible sense” of the terms
“things” and “hang together” is here. For Sellars, “things
in the broadest possible sense” covers everything from
theorems to fermions. By the same token, the philosophical sense of “hanging together” should furnish an insight
into the link between things as disparate as logical norms
and elementary particles. The philosophical vision ought
not only to encompass but also to explain the intrication of conceptual ideality and physical reality. Is this
to reiterate an antiquated dualism? No. A dualism is a
distinction that fails to explain the connection between
the terms it distinguishes. Philosophy discriminates, it
distinguishes and separates, but always with a view to
ultimate integration. In this regard, philosophy discriminates precisely in order to avoid dualism. The animus
towards dualism should not excuse insensitivity towards
distinction. To distinguish between the normative and
the factual is not to promulgate dualism once it is understood that this distinction furnishes the precondition
for understanding the intrication of the conceptual and
the physical; an intrication that is constitutive of what we
call “reality.” Philosophy is synoptic in that it strives to
reconcile a basic disjunction in our conception of reality. This disjunction is a consequence of the fundamental
conceptual discrepancy bequeathed to us by philosophical modernity. If Sellars’ work (unlike that of many of his
analytic contemporaries) retain its contemporaneity for
us today, fifty years after the bulk of it was written, this
is because, over and above its sometimes forbidding difficulty, it represents one of the most sustained attempts to
think through the implications of a fundamental diremption which extends into our very conception of what we
are. This is the diremption between our self-understanding as rational subjects and our scientific understanding
of ourselves as physical objects. Throughout his work,
Sellars sought to arbitrate the conflict between these two
increasingly divergent images of man-in-the-the-world:
7
Ray
Brassier
Journal for Politics, Gender and Culture Vol. 8 / No. 2 / Summer 2011
Ray Brassier
8
the manifest image of man as a self-conscious rational
agent and the scientific image of man as a “complex
physical system.” Yet Sellars was careful not to portray
this divergence as a conflict between naïve pre-theoretical common-sense and sophisticated theoretical reason.
Rather, he insisted it be understood in terms of the tension between the disciplined and critical refinement of
common-sense through which a perennial tradition of
philosophical reflection has taught us to conceive of ourselves as rational beings bound by conceptual norms; and
the methodical extrapolation from ordinary perception
through which modern science has taught us to explain
manifest phenomena by postulating increasingly complex systems of imperceptible entities (e.g., molecules,
electro-magnetic radiation, gravitational fields, etc.). In
this regard, the fundamental contrast at issue is one between man’s manifest self-image as a rule-bound rational
agent participating in but not governed by the realm of
physical law, and man conceived through the optic of
natural science as a “complex physical system” whose
capacity for agency can ultimately be accounted for in
terms of concatenations of spatio-temporal causation.
Yet there is a persistent ambiguity in Sellars’ account
of the relation between manifest and scientific images.
On one hand, he seems to insist that the philosophical
task is to recognize the parity of the two images. The
acknowledgement of parity follows from the realization that the images are not in fact competing over the
same territory. Philosophy can adjudicate between the
competing claims of the manifest and scientific images
by distinguishing the normative privileges of the former
from the ontological rights of the latter. Thus, apparently
undermining his commitment to parity, Sellars upholds
The View from Nowhere
the priority of the scientific image by famously insisting that “in the dimension of describing and explaining
the world, science is the measure of all things, of what
is, that it is, and of what is not, that it is not.” (Sellars
1991, 173) This apparent inconsistency can be defused
once we recognize that the commitment to parity and the
commitment to priority operate at two distinct levels: that
of conceptual interpretation (giving and asking for reasons) and that of ontological description and explanation.
Parity at the level of conceptual interpretation is compatible with priority at the level of ontological description
and explanation. The claim for parity follows from the
recognition that the manifest image furnishes us with the
fundamental framework in terms of which we understand
ourselves as “concept mongers,”1 creatures continually
engaged in giving and asking for reasons. But we are
able to do things with concepts precisely insofar as concepts are able to do things to us. It is this capacity to be
gripped by concepts that makes us answerable to conceptual norms. And it is this susceptibility to norms that
makes us subjects. The manifest image is indispensable
insofar as it provides the structure within which we exercise our capacity for rational thought. Hence the parity
between images: both are governed by the norm of truth,
understood as maximally warranted assertion, despite the
conceptual incommensurability between manifest and
scientific truth claims. Yet the manifest image remains
indispensable as the originary medium for the normative.
To the extent that this normative framework does not survive, Sellars warned, “man himself would not survive.”
(Sellars 1991a, 18) But it is man qua rational agent, not
anthropological object, which Sellars wishes to safeguard here. The manifest image remains indispensable
because it provides us with the necessary conceptual resources we require in order to make sense of ourselves as
Identities
cognitive activity, and with it science’s investigation of
reality, would become pointless. Is this to say that the
manifest image subordinates the ends of enquiry to human interests? Does the manifest image predetermine our
understanding of what a person is? I think the answer to
both questions is no.
Sellars aligns himself with a rationalist lineage that postulates an intimate link between rationality and subjective
agency. It is encapsulated in this Sellarsian dictum:
“‘True’ is a sign that something is to be done, for inferring is a doing.” The capacity to draw inferences requires
the ability to be bound by a rule. This binding is spontaneously undertaken by a subject, not passively submitted
to by an object. The agent is a subject precisely insofar as
she is able to submit to a rule. Our capacity to do things
with concepts presupposes that concepts can do things to
us. Our grasp of a concept requires that we be gripped by
the concept. But if rationality is indissociable from subjectivity, and subjectivity is synonymous with selfhood,
does this mean that the capacity for rationality requires
the existence of selves? Does the institution of rationality
necessitate the canonization of selfhood? Not if we learn
to distinguish the normative realm of subjective rationality from the phenomenological domain of conscious
experience. To acknowledge a constitutive link between
subjectivity and rationality is not to preclude the possibility of rationally investigating the biological roots of
subjectivity. Indeed, maintaining the integrity of rationality arguably obliges us to examine its material basis.
Philosophers seeking to uphold the privileges of rationality cannot but acknowledge the cognitive authority of
the empirical science that is perhaps its most impressive
offspring. Among its most promising manifestations is
9
persons, that is to say, concept-governed creatures continually engaged in giving and asking for reasons. It is
not privileged because of what it describes and explains,
but because it renders us susceptible to the force of reasons. It is the medium for the normative commitments
that underwrite our ability to change our minds about
things, to revise our beliefs in the face of new evidence
and correct our understanding when confronted with a
superior argument. In this regard, science itself grows
out of the manifest image precisely insofar as it constitutes a self-correcting enterprise. Indeed, for Sellars, a
proto-scientific theory lies at the heart of the normative
structure of the manifest image. We had to learn to postulate thoughts as unobservable inner episodes in order
to explain publicly observable speech. Only in doing so
did we acquire the ability to understand ourselves as rational agents operating in the concept-governed space of
reasons. Once ushered into this normative dimension, we
developed ever more sophisticated resources for describing and explaining what we observe in terms of what we
do not observe. Thus Sellars is a resolutely modern philosopher in his insistence that normativity is not found
but made. The rational compunction enshrined in the
manifest image is the source of our ability to continually revise our beliefs, and this revisability has proven
crucial in facilitating the ongoing expansion of the scientific image. Once this is acknowledged, it seems we are
bound to conclude that science cannot lead us to abandon
our manifest self-conception as rationally responsible
agents, since to do so would be to abandon the source
of the imperative to revise. It is our manifest self-understanding as persons that furnishes us, qua community
of rational agents, with the ultimate horizon of rational
purposiveness with regard to which we are motivated
to try to understand the world. Shorn of this horizon, all
Journal for Politics, Gender and Culture Vol. 8 / No. 2 / Summer 2011
Ray Brassier
cognitive neurobiology, which, as its name implies, investigates the neurobiological mechanisms responsible
for generating subjective experience. Does this threaten
the integrity of conceptual rationality? It does not, so
long as we distinguish the phenomenon of selfhood from
the function of the subject. We must learn to dissociate
subjectivity from selfhood and realize that if, as Sellars
put it, inferring is an act - the distillation of the subjectivity of reason - then reason itself enjoins the destitution of
selfhood.
10
*
It is instructive to contrast Sellars’ account of conceptual
parity and explanatory priority between the manifest and
scientific images with Jürgen Habermas’ recent attempt to
adjudicate the relation between the factual and normative
in a controversy over the implications of cognitive neurobiology. In a 2008 paper entitled “The Language-Game
of Responsible Agency and the Problem of Free-Will,”
Habermas invokes the Sellarsian schema in order to
refute what he sees as the attempt by contemporary neuroscientists to undermine the norm of rational agency
which plays such a fundamental role not only in ethical
and political theorizing, but also in legal and psychiatric discourse. (Habermas 2008, 13-50) Habermas’ text
is largely concerned with responding to a manifesto in
which eleven distinguished German neuroscientists claim
that our ordinary concept of “free-will” is on the verge of
being overthrown by recent advances in cognitive neurobiology. As Habermas himself notes, “neurologists
expect the results of their research to lead to a profound
revision in our self-understanding.” (ibid., 14) According
The View from Nowhere
to these neuroscientists themselves: “We stand at the
threshold of seeing our image of ourselves considerably shaken in the foreseeable future” (Elger et al 2004,
37). The Sellarsian resonances of both formulations are
striking. But Habermas accuses the neuroscientists who
would deploy the methods of natural scientific investigation to explain some of the fundamental features of our
manifest self-conception - specifically, our understanding of ourselves as agents - of illegitimately extending
the resources of objectification beyond their proper remit. For Habermas, the attempt to study first-person
subjective experience from the third-person, objectifying
viewpoint, involves the theorist in a performative contradiction, since objectification presupposes participation in
an intersubjectively instituted system of linguistic practices whose normative valence conditions the scientist’s
cognitive activity. Attempts to interrogate the normative
status of agency within the manifest image unwittingly
undermine the very concept in whose name every rational investigation is ultimately undertaken, since it is the
collectively instantiated norm of agency that provides
the rationale for producing “truer,” more accurate descriptions of reality in the first place. Thus, according to
Habermas, attempts to explain agency naturalistically fail
because “the social constitution of the human mind which
unfolds within interpersonal relationships can be made
accessible only from the perspective of participants and
cannot be captured from the perspective of an observer
who objectivates everything into an event in the world.”
(Habermas 2008, 34) Habermas characterizes this intersubjective domain of rational validity as the dimension of
“objective mind,” which cannot be understood in terms
of the phenomenological profiles of the community of
conscious selves comprised in it. Accordingly, it is the
intrinsically intersubjective status of the normative realm
Identities
It is not the subjectivity of our conscious life that distinguishes humans from other creatures but the intentional
stance and the interlocking of the intersubjective relations
between persons with an objectivating attitude to something
in the world. The linguistic socialization of consciousness
and the intentional relation to the world are mutually constitutive in the circular sense that each presupposes the
other conceptually. (ibid., 35)
The objectivity of social mind is grounded in the relation of reciprocal presupposition between an inherently
linguistic (and hence constitutively social) consciousness
and the cognitive relation to the world. For Habermas,
the interdependence between language and intentionality
implies not only that neither can be studied independently of the other, but more strongly, that neither can be
intelligibly distinguished from the other. Here Habermas
certainly echoes Sellars, whose attack on “the myth of the
given” challenges the idealist attempt to ground “originary” intentionality in transcendental consciousness.
Consciousness construed as originary condition of givenness becomes an unexplained explainer. This brand of
transcendental idealism is inimical to naturalism, since if
consciousness is the originary condition of objectivation,
of which science is one instance, it follows that science
cannot investigate consciousness. Upending this idealist order of explanation, Sellars roots the intentionality
of the mental in socially instantiated linguistic practice.
While the normative order retains a quasi-transcendental
status, its linguistic embodiment allows us to understand how it is embedded in the empirical order. Thus,
while Sellars maintains the irreducibly normative status
of intentionality, the fact that it is always linguistically
embodied allows us to investigate when or how this normative dimension might have arisen in the course of our
evolutionary and social history.
Habermas, for his part, rightly emphasizes the necessity of distinguishing the normative from the natural, or
reasons from causes, and accurately diagnoses the contradictions and confusions attendant upon any pre-emptive
collapse of the former into the latter. But because his account is so largely reactive, unlike Sellars, he is unable to
propose any positive account of the intrication between
concepts and causes. Conflating naturalism with empiricism, Habermas upholds Sellars’s distinction at the cost
of eliding its scientific realist corollary, viz., that mind,
and hence the normative order, possesses a neurobiological as well as socio-historical conditions of emergence.
As a result, Habermas pre-emptively disqualifies by
conceptual fiat every scientific attempt to describe and
explain the transition from pre-linguistic to linguistic
consciousness, from the sub-personal to the personal,
and from neurobiology to culture. For Habermas, the explanatory resources required in order to provide such an
account threaten to cost too much: they would incur a
self-objectification which would irrevocably estrange us
from ourselves. As he puts it: “The limits of naturalistic
self-objectification are trespassed when persons describe
themselves in such a way that they cannot recognize
themselves as persons anymore” (ibid., 25). Such an
11
that precludes any attempt to account for its operation
or genesis in terms of entities or processes simpler than
the system itself. Neither the phenomenological nor neurobiological profiling of participants can be cited as a
constituting condition for this socially “objective mind”
since it is the source of the capacity for intentional objectivation presupposed by both:
Journal for Politics, Gender and Culture Vol. 8 / No. 2 / Summer 2011
Ray Brassier
objectification of the human, Habermas maintains, would
bring about a “fictionalization” of selfhood which would
conjure “the image of a consciousness that hangs like
a marionette from an inscrutable criss-cross of strings”
(ibid., 24). Yet such depersonalization remains impossible, Habermas contends, because it could only come
about through the attainment of a hypothetical “view
from nowhere” which science cannot realize:
12
The resistance to a naturalistic self-description stemming
from our self-understanding as persons is explained by the
fact that there is no getting round a dualism of epistemic
perspectives that must interlock in order to make it possible
for the mind, situated as it is within the world, to get an
orienting overview of its own situation. Even the gaze of a
purportedly absolute observer cannot sever the ties to one
standpoint in particular, namely that of a counterfactually
extended argumentation community. (ibid., 35)
This dualism of epistemic perspectives invoked by
Habermas is the dualism of observer and participant.
And in fact, Habermas recodes the Sellarsian distinction
between manifest and scientific images in terms of a dualism of theory and practice wherein the former indexes
the objectifying stance of scientific naturalism while the
latter expresses subjective participation in intersubjective
discourse (the “argumentation community”). Yet even as
Habermas insists on the complementarity of scientific theory and discursive practice, he inscribes the former within
a horizon of conceptual possibility entirely delimited by
the latter. Thus, he insists, “the conceptual constitution
of domains of enquiry, the construction of designs and
measurements, and the experimental production of data
are all rooted in pre-scientific practice” (ibid., 38). Yet as
Habermas knows, there is a crucial difference between
The View from Nowhere
methodological priority and nomological dependence,
and the fact that pre-scientific practice enjoys chronological precedence over scientific theorizing in no way entails
that the latter is logically dependent upon or reducible to
the former. In his determination to ward off the naturalistic
dissolution of the normative, Habermas resorts to an instrumentalization of science - of the sort Sellars repeatedly
warned against - which inadvertently suggests that nothing
we learn about ourselves from the perspective of scientific
theory could force us to revise the content of our subjective or “participatory” self-understanding. Habermas’
epistemological dualism of objectifying theory and discursive practice is in many ways an exacerbation of the more
familiar dualism of first and third-person perspectives in
Anglo-American philosophy of mind. Ultimately, the dualism of epistemic perspectives seems to point toward the
conceptual impossibility of arriving at a synoptic vision
that would finally bridge the gap between the conceptual and the natural, or the subjective and the objective.
What Anglo-American philosophy characterizes as the
“explanatory gap” between mind and brain, or first and
third person perspectives, Habermas rashly inflates into
a “pragmatic contradiction” between the neuroscientist’s
practico-discursive reliance on intersubjectively instituted
semantic norms and her conceptual disavowal of those
conditions in her theoretical propositions.
*
Is it possible to describe and explain the correlation
between first-person experience and neurobiological
processes without lapsing into the sort of conceptual incoherence denounced by Habermas? In Being No One: The
Identities
representation. However, for Metzinger, the representing
or “vehicule” does not have its boundaries at the skin of
the organism but can extend out into the environment from
which it extracts a represented “content.” Consequently, in
Metzinger’s account, the representing may be defined as
“internal” to the representational system even when it is
constituted by spatially external events. Moreover, where
much philosophy of mind tends to hypostatize the vehicule/content distinction, with the result that vehicule and
content are construed as distinct entities which can then all
too easily be interpreted as instances of mental or physical events respectively, Metzinger insists that representing
and represented be conceived as conjoined aspects of a
single informational process whose deep-structure needs
to be mapped according to five distinct levels of analysis:
phenomenological, representational sensu stricto, information-computational, functional, and neurobiological.
Although each level of representational structure remains conceptually distinct, its autonomy is constrained
by the minimal requirement that any “slice” of the representational process remains correlated with events at the
neurobiological level. Thus, rather than trying to directly
identify the mental with the physical, Metzinger maintains the relative irreducibility of these distinct levels of
description, carefully distinguishing the structural properties and features specific to each, while insisting that every
representational state invariably supervenes upon the neurobiological level - the guiding hypothesis being that there
must always be minimally sufficient neural correlates for
every representational state, even in those cases where we
are not yet in a position to identify them.
On the basis of this characterization of conscious states
as a variety of representational states, Metzinger is able
13
Self-Model Theory of Subjectivity (Metzinger 2004; originally published in 2003, four years prior to Habermas’
article), Thomas Metzinger describes and explains in
principle how normatively regulated social interaction
between conscious selves supervenes upon un-conscious,
sub-symbolic neurobiological processes. Moreover,
Metzinger does so by explaining how the phenomenon of selfhood, and hence the first-person subjective
perspective, can be understood as arising out of subpersonal representational mechanisms. First however, it
is necessary to stave off a potential misunderstanding.
Although unequivocally naturalistic in its methodology
and uncompromisingly “materialist” in tenor, Metzinger
does not adopt the kind of straightforwardly “reductionist” strategy espoused by traditional mind-brain identity
theories, whether in their strong versions, where identity
is construed as obtaining between mental and physical
types, or in their weaker formulations, where the identity in question is merely between mental and physical
tokens.2 Rather than postulating direct token or type
identities between psychological and neurological states,
Metzinger proceeds by elaborating a naturalized theory
of representation wherein the latter is construed as a
dynamic process involving three distinct types of state internal representations, which are always unconscious;
mental representations, which are only sometimes conscious; and phenomenal representations, which are
always conscious. Furthermore, every representational
state comprises a relation between a representing - i.e.,
the concrete internal state of the system - and a represented - the particular feature of the world or of the system
itself about which the representational state carries information. In many ways, Metzinger’s distinction between
representing and represented corresponds to the familiar
distinction between the “vehicule” and the “content” of
Journal for Politics, Gender and Culture Vol. 8 / No. 2 / Summer 2011
Ray Brassier
to propose a novel account of the nature of conscious experience as a special case of phenomenal representation
in which an individual information processing system
generates a reality-model. At its simplest level then,
consciousness can be defined as obtaining whenever a
representational system generates a phenomenal world
model: “Conscious experience then consists in the activation of a coherent and transparent world model within
a window of presence.” (ibid., 213) Metzinger goes on to
specify three minimal constraints for the experience of
phenomenal consciousness:
1. Presentationality, or the generation of a window of
temporal presence through which the system represents the world.
2. Globality, or the availability of information for
guided attention, cognitive reference, and control of
action.
14
3. Transparency, defined as “inversely proportional to
the introspective degree of attentional availability of
earlier processing stages.” (2004, 165)
Transparency, the third constraint, is arguably the most
significant for Metzinger’s entire account. Here again, it
is important to distinguish it from more familiar philosophical definitions of “transparency” in terms of the
inaccessibility of vehicule as opposed to content properties (or of the properties of the representing as opposed
to those of the represented). Metzinger refuses this orthodox construal of transparency because, once again,
it encourages the temptation to reify the distinction
between content and vehicule in terms of traditional distinctions between the mental and the physical. Thus, the
The View from Nowhere
mental would be defined as transparent in contradistinction to the opacity of the physical. But on Metzinger’s
account, it is simply not the case that the representational
vehicule is a physical entity while its represented content is mental: both vehicule and content, representing
and represented, are indissociable aspects of an informational continuum wherein each can switch role and serve
as content or vehicule for another, higher order representation. Consequently, transparency is fundamentally
a phenomenological rather than epistemological notion:
phenomenal content is not epistemic content: “The transparency of phenomenal representations is cognitively
impenetrable; phenomenal knowledge is not identical
to conceptual or propositional knowledge.” (ibid., 174)
Accordingly, the fact that something is phenomenologically transparent does not entail that it is cognitively
accessible to the system itself; as we shall see, the reverse
is far more often liable to be the case. In fact, phenomenal
transparency implies the unavailability of the representational character of the contents of conscious experience:
Truly transparent phenomenal representations force a conscious system to functionally become a naïve realist with
regard to their contents: whatever is transparently represented is experienced as real and as undoubtedly existing
by this system. (ibid., 167)
Thus, in a move strikingly redolent of Kant, Metzinger
characterizes what U.T. Place originally identified as
“the phenomenological fallacy” - “the mistaken idea
that descriptions of the appearances of things are descriptions of actual state of affairs in a mysterious inner
environment” (Place 1970, 42) - in terms of the abstraction of the represented from the process of representation.
Transparency understood as the occlusion of the process
Identities
Once consciousness is minimally defined as the activation of an integrated world-model within a window of
presence, then self-consciousness can be defined as the
activation of a phenomenal self-model (PSM) nested
within this world-model: “A self-model is a model of the
very representational system that is currently activating
it within itself.” (ibid., 302) Metzinger identifies three regards in which the system may benefit from the ability to
consciously represent its own states to itself:
1. The possession of phenomenal states clearly increases the flexibility of the system’s behavioural profile by
amplifying its sensitivity to context and its capacity for
discrimination.
2. The PSM “not only allows a system to make choices
about itself but adds an internal context to the overall
conscious model of reality under which the system operates.” (ibid., 308)
3. Lastly, the PSM exerts an important causal influence, not only by differentiating but also by integrating
the system’s behavioural profile. Thus, “as one’s
bodily movements for the first time became globally
available as one’s own movements, the foundations
for agency and autonomy are laid. A specific subset of
events perceived in the world can now for the first time
be treated as systematically correlated self-generated
events.” (ibid., 309)
Through the PSM, a system becomes able to treat itself as a second-order intentional system - one capable
of entertaining beliefs about its own beliefs3 - and is
thereby transformed from something merely exhibiting
behaviour into an entity capable of exerting the sort of
self-regulation characteristic of what we call “agency.”
Accordingly, given any system for which the constraints
of presentationality, globality, and transparency obtain,
the acquirement of a PSM will necessarily entail the
emergence of a phenomenal self. Yet the latter is not an
autonomous or independent entity but merely the represented of a phenomenal representation. Moreover, it
is precisely the system’s lack of access to the process
through which it generates its own self-model that engenders the condition of “autoepistemic closure” whereby
the represented of the system’s self-representation occludes the representing that gave rise to it:
Phenomenal selfhood results from autoepistemic closure
in a self-representing system; it is a lack of information
… The phenomenal property of selfhood is constituted by
transparent, non-epistemic self-representation - and it is on
this level of representationalist analysis that the refutation
of the corresponding phenomenological fallacy becomes
truly radical, because it has a straightforward ontological
interpretation: no such things as selves exist in the world …
What exists are information processing systems engaged in
the transparent process of phenomenal self-modelling. All
that can be explained by the phenomenological notion of a
“self” can also be explained using the representationalist
notion of a transparent self-model. (ibid., 337)
15
of representation to the benefit of its phenomenal contents
encourages the system to remain a “naïve realist” about
what it experiences. It generates the subjective impression
of phenomenological immediacy. As a result, phenomenal
transparency, which is among the defining features of the
subjective experience of conscious immediacy, is in fact “a
special form of darkness.” (Metzinger 2004, 169)
Journal for Politics, Gender and Culture Vol. 8 / No. 2 / Summer 2011
Ray Brassier
16
Ultimately, the PSM is simply the shadow cast by the
occlusion of global, attentionally available information
about the workings of the system. But why should this
transparency have come about? Metzinger’s answer is
that autoepistemic closure is imposed by the need to minimize the amount of computational resources required in
order to make system-related information consciously
available. Transparent self-modelling provides systemic
information without generating a potentially debilitating
regress of recursive self-modelling, for if the system had
to include every representing involved in generating its
self-represented within the latter, then it would also have
to incorporate within it the representing required in order
to generate this new, second-order self-represented, and
so on ad infinitum. Phenomenal transparency is a cheap
way of minimizing the neurocomputationally exorbitant
cost of representational opacity.
Metzinger concludes by summarizing his principal
claim in terms of three heuristic metaphors: the neurophenomenological cave; the phenomenal map; and
total simulational immersion. The first is a reworking of
Plato’s allegory of the cave. Recall that according to the
latter, the human mind’s relationship to reality is akin to
that of a prisoner held captive in a cave - the prisoner
has never seen anything but the shadows cast onto the
wall facing her by puppet-simulacra of objects which are
paraded in front of the fire that is burning behind her. In
Metzinger’s version of this Platonic allegory, the cave is
the physical organism or information processing system
as a whole; the fire its neurocomputational dynamics; the
puppet-simulacra of objects its mental representings; and
the shadows cast on the cave wall its phenomenal representeds. But according to Metzinger, there is no prisoner
The View from Nowhere
in the cave; indeed there is no-one there at all. The conscious self is not an entity but a shadow; not an individual
object, but rather the ongoing process of shading through
which a multidimensional neurocomputational representation is projected as a much lower dimensional
phenomenal model onto the surface provide by the system’s world-model. Thus the PSM is not the shadow of a
captive individual, nor the avatar of a supposedly authentic or even “transcendental” subject beneath or behind
the conscious individual, but rather a shadow cast by the
cave as a whole: “It is the physical organism as a whole,
including all of its brain, its cognitive activity, and its
social relationships, that is projecting inward from all directions at the same time … The cave shadow is there,
the cave itself is empty.” (ibid., 550)
In Metzinger’s second metaphor, phenomenal experience constitutes a dynamic, multidimensional map of
the world. And like the maps in subway stations, the
phenomenal world model features a little red arrow in
it that allows the user to locate herself within the map.
The PSM is analogous to this little red arrow saying “You
are here:” “Mental self-models are the little red arrows
that help a phenomenal geographer to navigate her own
complex mental map of reality by once again depicting
a subset of her own properties for herself.” (ibid., 552)
But whereas the red arrow in the subway map is opaque
to the map user, and hence explicitly apprehended by her
as a representation, the PSM is transparent: its status as a
representation is occluded for the system because of the
introspective unavailability of all those earlier processing
stages through which it has been produced. Yet this is not to
say that we are mistakenly identifying ourselves with our
own PSM - there can be no question of misidentification
Identities
As long as the pilot is needed to navigate the world, the
puppet-shadow dances on the wall of the neurophenomenological caveman’s phenomenal state-space. As soon as
the system does not need a globally available self-model,
it simply turns it off. Together with the model, the conscious experience of selfhood disappears. Sleep is the little
brother of death. (ibid., 558)
Ultimately then, Metzinger explains the phenomenological experience of selfhood as a specific type of
representational content: the self is the represented of a
phenomenally transparent self-model. But it is not necessary to postulate the existence of entities called “selves”
over and above the dynamic web of relations between the
complex physical system known as the human organism,
its internal representational economy, and its physical environment. All the salient cognitive and phenomenal data
can be accounted for in terms of the PSM. Is this then to
say that the notion of “the self” as an autonomous reality can be dispensed with and relegated to the dustbin of
intellectual history? Before we address this question and
some of the objections voiced against Metzinger’s thesis,
let us consider some further implications of the latter.
According to Metzinger, even if it is the case that we cannot
help experiencing ourselves as “selves” and find it impossible to phenomenologically imagine selfless experience,
the latter remains an epistemic possibility. Clearly, organisms can satisfy the minimal constraints for phenomenal
consciousness (presentationality, globality, transparency)
without being in possession of a PSM. Undoubtedly,
many forms of animal life provide instances of selfless
consciousness in this sense. But they remain incapable
of generating sophisticated conceptual representations of
themselves and their world. Thus, for Metzinger, the philosophically interesting question is whether it is possible
to envisage systems capable of generating sophisticated
conceptual representations of themselves and their world
without the benefit of a PSM. Metzinger suggests that such
systems are indeed envisageable, but would have to be
characterized as systems whose representational models
have been rendered fully opaque. Recall that phenomenal transparency is a function of epistemic darkness: for
any representation, its degree of transparency is inversely
proportional to the degree of available epistemic information about the representational processes that preceded its
instantiation. But it is possible to imagine systems endowed with the same cognitive capacities as humans,
but for whom the transparency constraint, specifically
as pertaining to the PSM, would not obtain. Thus, “earlier processing stages would be attentionally available
17
here since the PSM is all we are. There is no transcendental or noumenal self who could mistakenly identify itself
with the phenomenal self since, as Metzinger insists,
the cave is empty. But its multidimensional neural selfimage generates a condition of “full immersion.” Thus,
in the third and last of Metzinger’s heuristic metaphors,
the PSM operates like a total simulation: “A total flight
simulator is a self-modelling aeroplane that has always
flown without a pilot and has generated a complex internal image of itself within its own internal flight simulator.”
(ibid., 557) The PSM is this internal image which functions as an invisible interface for the interaction between
system and world. And just as the total flight simulator
generates its own virtual pilot, the human brain activates
its PSM when it requires a representational instrument to
integrate, monitor, predict, and remember the activities
of the system as a whole:
Journal for Politics, Gender and Culture Vol. 8 / No. 2 / Summer 2011
Ray Brassier
18
for all partitions of its conscious self-representation; it
would continuously recognize it as a representational
construct, as an internally generated internal structure.”
(ibid., 565) Such a system would possess a system-model
without instantiating selfhood. It would retain the functional advantages of possessing a coherent self-model
(integration, monitoring, prediction, memory) but without experiencing itself as a self. It would be burdened
with an additional computational load, which it would
have to find some way of discharging without getting
trapped into infinite loops of self-representation, but if it
could find some means of solving this problem without
resorting to the transparency solution, then this would
indeed constitute an example of a cognitive system operating with a non-phenomenologically centred model of
reality. Such a system would be nemocentric: it would
satisfy a sufficiently rich set of constraints for conscious
experience without exemplifying phenomenal selfhood.
It would quite likely remain functionally egocentric, in
order to satisfy the requirements of biological adaptation, but it would remain phenomenologically selfless.
Moreover, such a system’s reality-model would be richer
in informational content than our own, because at every
stage of processing, more information about earlier processing stages would be globally available for the system
as a whole. Thus such a system would instantiate what
Metzinger calls a “first-object” perspective because it
would experience its own phenomenal self-model not
only as a represented but also and simultaneously as a
representing. It would be aware of the representational
vehicule as well as of the represented content.
There is an interesting comparison to be made between this hypothetical nemocentric perspective and
The View from Nowhere
the transcendental perspective of pure phenomenological consciousness as effected by what Husserl called
the “transcendental reduction.” The goal of the latter is
to “bracket off” or suspend the assumption of the autonomous reality of objects in order to isolate the ideal
objectifying acts through which intentional consciousness generates its objective correlates. Obviously, in
Husserl’s idealist schema, this reduction is carried out
by and for a transcendental subject, the better to separate the world-less realm of intentional consciousness as
originary source and locus for the possibility of scientific objectification. By way of contrast, the hypothesis
of the nemocentric perspective suggested by Metzinger
is one in which the representational process’s reincorporation into the represented object serves to foreground
the sub-personal dimension of neurocomputational processing that underlies objectifying representation, and
hence the objective processes through which objectivity
is partly produced. Over and above its status as a phenomenological anomaly, the hypothesis of nemocentric
consciousness provides a possible model for the new
type of experience that could be engendered were scientists to succeed in objectifying their own neurobiological
processes of objectification. The nemocentric subject of
a hypothetically completed neuroscience in which all the
possible neural correlates of representational states have
been identified would provide an empirically situated and
biologically embodied locus for the exhaustively objective “view from nowhere,” which Habermas and others
have denounced as a conceptual impossibility. Yet here,
as Metzinger’s work suggests, empirical possibility outstrips a priori stipulations of conceivability. In railing
against the possibility of the mind’s complete theoretical
self-objectification, Habermas inadvertently reiterates
the conflation of personhood as conceptual norm with
Identities
[N]euroscientific enlightenment about the illusion of free
will crosses the conceptual border into self-objectification
… For this shift in the naturalization of the mind dissolves
the perspective from which alone an increase in knowledge
could be experienced as emancipation from constraints.
(Habermas 2008, 24)
But what Habermas fails to see is how the genitive in the
proposition “self-knowledge is not knowledge of the self”
is at once subjective and objective: if the subject is not a
self, then the subject who knows herself to be selfless is
neither the proprietor of this knowledge (since it is not hers)
nor its object (since there is no-one to know). Ultimately,
Habermas’ inability to articulate the distinction between
theoretical objectification and discursive practice ends up
promulgating a dualism of theory and practice, objective
and subjective, which results from the refusal to acknowledge their interpenetration. For as Sellars so clearly saw, it is
precisely the norm-governed domain of subjective practice
that demands the conceptual integration of the subjective
and the objective, reasons and causes, in the obligation to
attain a maximally integrated understanding of the world
and our position within it as creatures who are at once conceptually motivated and cause-governed. Unlike Sellars,
Habermas pushes the irreducibility of the normative to the
point where it generates a schism within the conceptual
order in the form of a dualism of the normative and the natural. Lacking any understanding of the interplay between
subjective practice and objective explanation, Habermas’
account of rationality becomes internally contradictory: it
seeks to defend rationality by excluding a key part of it,
viz., the naturalistic explanation of empirical subjectivity,
which can only increase, not compromise, our understanding of the conceptual, both in its distinction and emergence
from the empirical. Disregarding the imperative to understand the latter, Habermas posits a distinction that he reifies
into a substantive dualism of reasons and causes.
*
Critics have objected that the notion of “self” which
Metzinger claims to have eliminated is a straw man: Hume,
Kant and Nietzsche had already demolished this (supposedly) Cartesian conception of the self as an autonomous
metaphysical substance. Others have responded to his
work by insisting that phenomenology in the Husserlian
tradition abjures precisely this metaphysical reification
of the self: phenomenology construes the subjectivity of conscious experience in terms of a pre-reflective
dimension of ipseity according to which phenomenal
experience is necessarily “owned.” One of Metzinger’s
phenomenological critics, Dan Zahavi, insists that it is
in terms of the unobjectifiable “mineness” of conscious
experience - which Heidegger called Jemeinigkeit - that
selfhood ought to be understood once liberated from its
metaphysical reification as res cogitans:
Whether a certain experience is experienced as mine or not
does not depend on something apart from the experience,
19
selfhood as phenomenological reality - the very confusion he initially sought to denounce. Here we have an
example of what could be called “the philosopher’s fallacy:” a failure of imagination paraded as an insight into
necessity.4 Habermas refuses to envisage the possibility of a convergence between self-objectification and
self-knowledge because he continues to assume that selfknowledge must be knowledge of the self:
Journal for Politics, Gender and Culture Vol. 8 / No. 2 / Summer 2011
Ray Brassier
but on the givenness of the experience. If the experience is
given to me in a first-personal mode of presentation, it is
experienced as my experience, otherwise not. To be conscious of oneself, is consequently not to capture a pure self
that exists in separation from the stream of consciousness,
rather it just entails being conscious of an experience in
its first-personal mode of givenness. In short, the self referred to is not something standing beyond or opposed to
the stream of experiences, rather it is a feature or function
of their givenness. It is the invariant dimension of first-personal givenness in the multitude of changing experiences.
(Zahavi 2005, 9)
It is this focus on the allegedly transcendental dimension
of “givenness” (which is “ontological,” as opposed to the
merely “ontic” given) that distinguishes phenomenology from psychology, and phenomenological experience
stricto sensu from any merely empirical cataloguing of
introspectively accessible psychic states or processes.
Indeed, Zahavi cites Husserl approvingly to the effect
that the phenomenological domain is “neither psychic
nor physical:”
20
Rather, phenomenology is interested in the very dimension of givenness or appearance and seeks to explore its
essential structures and conditions of possibility. Such an
investigation is beyond any divide between psychical interiority and physical exteriority, since it is an investigation
of the dimension in which any object - be it external or
internal - manifests itself. (ibid., 14)”
Thus Zahavi insists that for phenomenology, the self is
not something given - it is precisely never something
given at the level of content of experience - but rather
the form of givenness or of experience as such. This
form is precisely what Heidegger called eigentlichkeit
The View from Nowhere
or “mineness:” the owning of experience. Consequently,
Zahavi contests Metzinger’s use of the PSM theory of
subjectivity to explain the fracturing of selfhood and the
anomalous phenomenologies involved in pathologies
such as anosognosia, schizophrenia, and Cotard’s syndrome. He objects that even in cases of thought insertion,
where the subject experiences thoughts that she ascribes
to another, she continues to own the experience, since her
very estrangement from the thought reveals how, even in
disavowing that the thought is hers, she continues to own
the experience in which this estrangement is registered
and this disavowal occurs. Thus, Zahavi insists, selfhood
remains an ineluctable phenomenological feature of the
form of the given, rather than of its content. The schizophrenic continues to experience alien thought episodes
as occurring to her, rather than to someone else: “Rather
than involving a lack of a sense of ownership, passivity phenomena like thought insertions involve a lack of
a sense of authorship (or self-agency) and a misattribution of agency to someone or something else.” (ibid., 6)
Zahavi demotes subjective agency to the level of empirical content, the better to elevate selfhood into a formal
condition of experience. Accordingly, he concludes,
even schizophrenic depersonalization presupposes this
irreducible proprietary relation to experience, which phenomenology identifies as this dimension of “ownness.”
But who owns experience? What remains of the self once
it has been de-substantialized and transposed to the level
of form? If phenomenological selfhood pertains to the
form rather than the content of experience, then what
formal property (or set of properties) can we invoke to
identify an experience as our own, or discriminate one
self from another? What characteristics distinguish my
Identities
“given” in the same way as a pub conversation or a religious experience? The fact that saccades and lesions
can be turned into intentional correlates of consciousness
does not make them “phenomena” in the same sense in
which conversations and sensations are said to be. Just
as unconscious phenomena can be viewed as intentional
correlates, conscious phenomena can be turned into objects and investigated from the third person perspective.
The former is no more a vindication of phenomenology
than the latter is of naturalism. Playing on the inherent
ambiguity of the word “phenomena,” Zahavi elides the
distinction between intentional and conscious phenomena and reduces the former to the level of the latter. But
he adduces no argument for the claim that phenomenological “givenness” remains irreducible to psychological
and/or cognitive experience; he simply stipulates it.
Ultimately, the claim that givenness itself must be accepted as an undeniable datum is merely the most radical
version of the myth attacked by Sellars.6 On the one hand,
subjectivity understood as “mineness” is precisely an aspect of experience that Metzinger is at pains to describe
and explain via his PSM theory. Having relinquished the
metaphysical postulate of a noumenal self subsisting behind or beyond appearances, the phenomenologist cannot
then maintain that the reality proper to the experiencing
self is more than just an experience. To understand the
subject as a structurally necessary condition of experience in the Kantian sense is precisely not to construe it as
a self exercising a proprietary grip over its experiences,
since the Kantian subject is an impersonal function, not a
titled individual proprietor endowed with deeds of ownership. The relation between subjective condition and
conditioned object does not map onto the relation between
21
experience from yours at the level of phenomenological
form? The problem is that everything that distinguishes
my self from yours subsists at the level of experienced
content, not the form of experiencing. Phenomenology
inflates selfhood into a structurally necessary property of
experience, the invariant form for the givenness of the
given, when precisely what distinguishes my self from
yours is something given, rather than its givenness. To
insist that it is given to me, rather than to you, is simply to beg the question as to the identity of the dative,
by reiterating a distinction experienced at the level of
given content and projecting it back onto the form of
its givenness. So what is the explanatory worth of the
phenomenological postulate according to which selfhood is a formally necessary property of experience? In
descriptive terms, all that distinguishes the phenomenological postulate of “mineness” as originary form from
the self-model theory of subjectivity is the fact that the
former stipulates as a necessary condition of experience
a phenomenon that the latter derives as a conditioned
experience. Instead of providing some property or set
of properties, whether conceptual, qualitative, or experiential, that would mark the difference between the
phenomenological structures governing the possibility
of appearance and those of its phenomenal counterparts,
which can be accounted for in terms of the sub-personal
mechanisms mapped by Metzinger, Zahavi invokes a dimension of givenness which, although defined using all
those features of phenomenal consciousness accounted
for by the PSM, is nevertheless “neither physical nor
psychical.”5 Moreover, the claim that this givenness
provides the dimension wherein any object “whether
internal or external” must manifest itself remains unpersuasive: in what sense does a saccadic eye movement or a
lesion of the occipital lobe appear as phenomenologically
Journal for Politics, Gender and Culture Vol. 8 / No. 2 / Summer 2011
Ray Brassier
22
proprietary self and owned experience. Questions as to
the reality of experience are undoubtedly metaphysical.
Zahavi denounces Metzinger’s denial of the existence
of selves as a dubious piece of scientistic metaphysics.
But Zahavi cannot then proclaim the indubitable reality
of selfhood simply because it is given as an experienced
content. For as both Metzinger and Sellars point out,
phenomenal transparency is not epistemic transparency.
To insist on the epistemic authority of conscious experiences is to reiterate the dogmatic pre-Kantian postulate
according to which experiences are cognitively self-authenticating. It is one thing to insist, as Descartes did,
that where phenomenal seeming is concerned, doubt is
inappropriate, since there can be no appearance-reality
distinction of the sort subject to epistemological adjudication. But where doubt is inappropriate, so is certainty.
The corollary of the admission that we cannot doubt how
things seem is the recognition that we cannot be certain
of it either, since certainty is doubt’s epistemic obverse.
It is as inadmissible to proclaim the indubitable epistemic
authority of phenomenal experience as to denounce it as
illusory.
Thus, just as Metzinger exposes phenomenal transparency as a kind of epistemic blindness, Sellars (like Kant
before him) insists that self-knowledge is mediated by
knowledge of objects. The phenomenon that Metzinger
describes and explains subtends the epistemic assumption that Sellars diagnoses and analyses in his critique
of the given. Zahavi reiterates this assumption when he
insists that “At its most primitive, self-consciousness is
simply a question of having first-personal access to one’s
own consciousness; it is a question of the first-personal
givenness or manifestation of experiential life.” (ibid.,
The View from Nowhere
7) Self-knowledge certainly comprises a dimension of
non-inferential immediacy that endows us with a privileged epistemic access to our own internal states, but
only within certain limits, since the immediacy of selfknowledge is itself the result of conceptual mediation
and cannot be evoked to ratify the appeal to an allegedly intuitive, pre-conceptual self-acquaintance. The
prejudice that immediacy is not the result of a mediating
self-relation seduces us into absolutizing phenomenal experience. Phenomenology’s absolutization of givenness
as such is the most extreme variant of the myth dismantled by Sellars.
Consequently, Zahavi is no more entitled to infer the reality of selfhood from its experience than Metzinger is to
deny it. Here it is important to bear in mind the distinction between different levels of analysis: concepts are not
phenomena. The concept of the subject, understood as a
rational agent responsible for its utterances and actions,
is a constraint acquired via enculturation. The moral to be
drawn from Metzinger’s work here is that subjectivity is
not a natural phenomenon in the way in which selfhood
is. But Metzinger need not even deny the reality of the
self (we might say that self-models are “real” in some
suitably qualified sense - though justifying this would require working out a full blown metaphysics), only the
phenomenological postulate of its absolute explanatory
priority. He draws a metaphysical conclusion where a
methodological one would be more apt: the self-model
theory of subjectivity describes and explains the phenomenon of selfhood in a way that allows it to be reintegrated
into the domain investigated by the natural sciences. It
forces us to revise our concept of what a self is. But this
does not warrant the elimination of the category of agent,
Identities
Notes:
1. The phrase is Robert Brandom’s.
2. For canonical statements of the position, see the first four papers
by Herbert Feigl, U.T. Place, J.J.C. Smart and David Armstrong
in Borst 1970, 33-79. See also Armstrong 1968, Feigl 1967, and
Smart 1963. Donald Davidson’s “Mental Events” is the classic
statement of the case for token identity (Davidson 2011).
3. “Let us define a second-order intentional system as one to which
we ascribe not only simple beliefs, desires, and other intentions,
but beliefs, desires, and other intentions about beliefs, desires,
and other intentions” (Dennett 1978, 273).
4. Daniel Dennett was of course the first to identify this fallacy.
5. The claim that for phenomenology consciousness is neither psychical nor physical is of course made by Husserl in the second
volume of his Logical Investigations. Zahavi (2005) cites it approvingly on p. 13.
6. “Many things have been said to be ‘given’: sense contents, material
objects, universals, propositions, real connections, first principles,
even givenness itself” (Sellars 1991, 127; my emphasis).
References:
Armstrong, D. M. 1968. A materialist theory of mind. London:
Routledge.
Borst, C.V., ed. 1970. The mind/brain identity theory. London:
Palgrave MacMillan.
Davidson, Donald. 2001. Mental events. In Essays on action and
events, 2nd edition, 207-228. Oxford: Clarendon Press.
Dennett, Daniel. 1978. Brainstorms: philosophical essays on mind
and psychology. Harmondsworth: Penguin.
Elger, C. E., A. D. Friederici, C. Koch, H. Luhmann, C. von der Malsburg,
R. Menzel, H. Monyer, F. Rösler, G. Roth, H. Scheich, and W.
Singer. 2004. Das Manifest: Elf führende Neurowissenschaftler
über Gegenwart und Zukunft der Hirnforschung [The Manifesto:
Eleven prominent neuroscientists on the present state and future
of brain Research]. Gehirn und Geist, 6, October 13, http://www.
gehirn-und-geist.de/artikel/839085.
Feigl, Herbert. 1967. The “mental” and the “physical.” Minneapolis:
University of Minnesota Press.
Habermas, Jürgen. 2008. The language game of responsible agency
and the problem of free will: How can epistemic dualism be reconciled with ontological monism? Philosophical Explorations
10, 1: 13-50.
Metzinger, Thomas. 2004. Being no one: The self-model theory of
subjectivity. Cambridge: MIT Press.
Place, U.T. 1970. Is consciousness a brain process? In The mind/
brain identity theory, ed. C.V. Borst. London: MacMillan.
Sellars, Wilfrid. 1991. Empiricism and the philosophy of mind.
In Science, perception, and reality, 127-196. Atascadero:
Ridgeview Publishing Co.
___. 1991a. Philosophy and the scientific image of man. In
Science, perception, and reality, 1-40. Atascadero: Ridgeview
Publishing Co.
___. 1991b. Truth and “correspondence.” In Science, Perception,
and Reality, 197-224. Atascadero: Ridgeview Publishing Co.
Smart, J.J.C. 1963. Philosophy and scientific realism. London:
Routledge and Kegan Paul.
Zahavi, Dan. 2005. Being someone. In Psyche: An Interdisciplinary
Journal of Research on Consciousness, 11.5: 1-20. http://theassc.org/files/assc/2611.pdf (accessed October 1, 2011).
23
since an agent is not a self. An agent is a physical entity gripped by concepts: a bridge between two reasons,
a function implemented by causal processes but distinct
from them. And the proper metaphysical framework for
explaining the neurobiological bases of subjective experience is that of a scientific realism rooted in an account
of conceptual normativity that supervenes on, but cannot
be identified with, socially instantiated and historically
mediated linguistic practices.
Journal for Politics, Gender and Culture Vol. 8 / No. 2 / Summer 2011