Phenomenology and Reflexivity

Phenomenology and Reflexivity
Getting clear(er) about “selfrepresentational” theories of
Self-representational theories (e.g. Kriegel
2009, Kriegel and Williford 2006, Smith 1986,
• One of the oldest phenomenological ideas about
consciousness is that it has a reflexive or selfpresentational aspect. (precedents: Aristotle,
Brentano, Kant, Husserl, Sartre, Heidegger)
• This aspect might help in explaining the “for-me”ness (subjective character) of consciousness, or its
phenomenal content (qualitative character), or both
• Self-representational theories plausibly solve many
objections to HOT and HOM theories.
• Self-representational theories (first approximation)
hold that: A (representational) state is conscious iff it
represents itself, or some aspect of itself.
Types of self-representational theories
• Structural (proper part): A conscious state has,
as a separable proper part of its structure, an
element that represents all or part of that
• Indexical-involving: A conscious state has a
part that makes indexical (or quasi-indexical)
reference to that very state.
• Structural (improper part): A conscious state
has an essential or inseparable part or aspect
that represents all or part of that state
Things to get clearer on
• What sense of “representation”? (many
linguistic expressions, e.g., are selfrepresenting, but this doesn’t make them
• What is self-representation? How is this
reflexive relation achieved?
• Infinite regress worries
• Token-type problems
Self-representational theories and the
justification of phenomenal beliefs
• It’s reasonable to think that any good reflexive theory of
consciousness must explain the justification of
phenomenal beliefs: for instance, my belief that I am
now having a state with a certain phenomenal character
• Also it seems plausible that the self-representation of
the state should play a role in this justification.
• But on any of these views, all that is available to play a
role in the justification of such a belief is a phenomenalstate token: it is not clear how we get from the selfrepresentation of a token state to general beliefs (about
types and what falls under them).
Self-representation and the
phenomenal concept strategy
• On the phenomenal concept strategy (PCS),
occurrent phenomenal states “figure in” or are
parts of (occurrent) judgments about them,
either as components of concepts used or
directly as parts of the judgments themselves.
• At least some PCS theories are reasonably
construed as self-representational theories: in
figuring in judgments about it a phenomenal
state represents (or presents) itself.
Using self-representation to explain
phenomenal judgments
• “The first-order phenomenal state, being self-presenting in the way
that only phenomenal states can be, figures directly in the belief, as a
mode of presentation of itself…. That is, [the] first-order mental
state-token is deployed, qua predicative constituent of the belief, to
refer to a phenomenal type that the token presents itself as
belonging to” –Horgan, Tienson, and Graham 2006
• If this is correct, then we can explain the justification of phenomenal
judgments in terms of the occurrence (or being-presented) of
phenomenal states: the mere occurrence of the phenomenal state is
sufficient to present it under a “mode of presentation” which can
then be used to make a judgment about it
• However, if “mode of presentation” is anything like a Fregean sense
(i.e. something ideal, with content-conferring properties), this won’t
quite do, since we have no account of HOW something like a (token)
phenomenal state can work as a “mode of presentation” of itself.
Improving the theory: some
phenomenological suggestions
• There are some terminological and conceptual
improvements we can make to the basic selfrepresentational theory by switching to the
terminology of Husserl’s phenomenology
• There are substantive reasons for these
changes, but they can also be taken just as
terminological suggestions, if preferred
First suggestion:
Presentation, not representation
• What we are interested in with reflexive theories
of consciousness is how states can present
themselves to subjects (along with, perhaps, also
presenting external objects). This might be
embedded within a general theory of cognitive
“representation,” but if we do not presuppose
physicalism there is no obvious reason it must be.
• Additionally, talking about presentation rather
then representation avoids the need to posit representational features of the relevant states and
avoids begging questions, e.g. against the
possibility of perceptual direct realism.
Second suggestion: Acts, not states
• In Husserl’s special sense of the term, an “act” is
any experience with intentional (presentational)
content for a subject.
• However, because presentational acts present
something as something, they are inherently (in
one sense) “concept-involving”.
• It might be thought that this begs questions by
assuming a prior “theory of intentionality”.
However, it is reasonable to suppose that an
account of the reflexive structure of
consciousness doesn’t have to solve this problem.
The modified theory
• With these modifications, we have the
following: An act is conscious if and only if it is
an act that (in addition to whatever else it
presents also) presents itself.
• For Husserl, it is definitional that all acts are
conscious. So it might be better to build
reflexiveness into the definition of an act:
A (self-conscious) act is a presentation (of
something) that also presents itself.
• The point of a theory of this kind isn’t to give an
account of sufficient conditions for a state
(conceived, say, in functional or
neurophysiological terms) to be conscious.
• It also isn’t to give an account of how something
that is conceived already as representational
becomes a conscious representation.
• The point is just, given the phenomenologically
plausible datum that conscious acts present
themselves to us in some way, to model this selfpresentation.
Act Self-presentation
• How can we model phenomenologically this selfpresentational structure of the act?
• It is helpful to consider the structure of the
“transcendental epoche”, the means by which we
gain insight into the phenomenological structure
of acts, according to Husserl
• Although the is epoche is used in deliberate,
explicit reflection, we can also gain insight into
the “pre-reflective” but nevertheless reflexive
structure of the acts themselves.
The transcendental epoche
• In the transcendental epoche, insight into the structure
of the act itself is achieved by “bracketing” – or putting
out of commission -- the “positing” of the ordinary
intentional object.
• According to Husserl, this converts the structure of the
act from a “pre-objective” to an “object” status.
• We achieve explicit access to both noetic (really
immanent) and noematic (ideal) aspects of the act.
• The transcendental epoche yields access to the whole
domain of experience, which survives even the
“bracketing” of the whole (spatiotemporal) world
Transcendental epoche and reflexivity
• The explicit reflection that occurs in the performance of the
transcendental epoche is itself only possible because of an
“implicit” reflexivity (or self-presenting aspect) already
there in ordinary conscious acts.
• It is plausible that, if my current act is one of presenting
some content, it is possible (given that I also have a general
concept of experience) for me to know that this act is one
that presents this content.
• Thus, even if the explicit consideration of the (first-order)
act in the epoche is (in some sense) a new act, it is possible
only because this first-order act already (in some way)
presents itself (even if only implicitly and non-objectively)
• Cf. Husserl: “In the very essence of an experience lies
determined not only that, but also whereof it is a
consciousness, and in what determinate or indeterminate
sense it is this.”
• “In the case of an experience directed to
something immanent …(so-called ‘internal’
perception), perception and perceived form
essentially an unmediated unity, that of a single
concrete cogitatio. Here the perceiving includes
its Object in itself in such a manner that it only
can be separated abstractively, only as an
essentially non-selfsufficient moment, from its
Object.” (Husserl 1913, pp. 68-69).
• “…there is a certain, extraordinarily important
two-sidedness in the essence of the sphere of
mental processes, of which we can also say that
in mental processes there is to be distinguished a
subjectively oriented side and an objectively
oriented side…” (Husserl 1913 p. 161)
• Conclusion: Husserl’s theory is a selfpresentational act theory that holds that acts
have an inseparable part -- or essential aspect-that presents the act itself
• This part or aspect presents the act’s noematic
(ideal and content-determining) as well as its
noetic (real) features.
• It is probably best to construe the selfpresentational part itself as an ideal (noematic)
rather than a real part.
Noema and Sense
• In 1969, Dagfinn Føllesdal drew an analogy
between Husserl’s noemata (or noematic senses)
and Fregean senses:
– Noema (or noematic sense) is a generalization of the
notion of meaning (Sinn)
– The intentional act presents its object by means of its
noema (thus the notion of noema also generalizes the
notion of a mode of presentation)
– Noema determines object but not conversely
– One and the same noema may be shared by different
token acts: thus the noema is not a real component of
the token act. It’s, rather, an ideal (or abstract) part.
Another analogy: Epoche and
quotation (Thomasson 2006)
• The kind of explicit reflective access that we get
in performing the epoche might be compared to
quotation or “semantic ascent”
• We “quote” the content in order to get access to
the content-conferring properties of the act, the
properties in virtue of which it has and presents
its content
• Even if most conscious states aren’t explicitly
“quoted” in this way, they must, structurally, be
• Taking both analogies seriously: A (conscious) act
which presents an (ordinary) content also
presents, inexplicitly, the ideal (noematic)
features of itself in virtue of which it presents
that content
• On this kind of view, a conscious act which
presents a content is also implicitly selfpresented as the kind of act that has that kind of
content. Thus, the act, which involves a mode of
presentation of the object, also presents the
mode of presentation itself
• This allows us to make an explicit phenomenal
judgment: “I am having an experience of that
type” (i.e. an experience that presents its
(ordinary) content in that way).
Qualitative Aspect and Transparency
• This accounts for the subjective (for-me-ness) aspect
of consciousness but it might also help to account for
the qualitative aspect as well.
• In particular, if we accept transparency, it is
reasonable to suppose that the self-presenting state,
in presenting itself, presents itself as the kind of state
that presents an object having certain properties
• Thus, the judgment that I am having an experience
of phenomenal type R might simply result from a
reflective modification of the act in which I present
an object with a related or identical property
Self-(re)presentational theories and
the hard problem
• Do self-(re)presentational theories help us
understand the hard problem, in the sense that
they help to explain why there should be at least
an explanatory gap between functional/structural
features and consciousness?
• Ans: Possibly, but only if the self-(re)presentational
aspect of a state (in virtue of which it is conscious)
doesn’t make a functional/causal/structural
• This will be the case iff a state’s presenting itself
isn’t a further functional/structural relationship
(not even one borne to itself)
• On structural (proper part) theories, the part of the
token state that presents the whole state is a real
and separable part, so it seems that if we remove it
we make a structural/functional difference.
• On indexical-involving theories, the indexical
presentation of a state also seems to make a
structural/functional difference
• However, a structural (improper part) theory might
work better. On such a theory, since the selfpresentational part of the state is not a real part, it
plausibly doesn’t make a structural/functional
How can a state present itself without
this self-presentation making a
structural/functional difference?
• We’ll get the right kind of theory if the
features of the state in virtue of which it
presents itself aren’t further
structural/functional parts, but rather formal
• But how do we model this self-presentation
in virtue of formal features?
One possibility: set self-membership
• On this approach, we model a state, A, with the
(first-order) content B as : A= {A, B}, using (nonwell-founded) set membership to model
• We might then hold that extensionally we just
have A and B and the functional relations
between them, but there is an additional formal
relation of presentation
• But it’s hard to see how this formal relation could
be realized
• And it’s not clear that we avoid the infiniteregress problem
A Gödelian analogy?
(cf. Hofstadter 2006)
• Consider the Gödel sentence G:
G  ~Prov(“G”)
• This sentence is essentially “self-quoting” where the
device of quotation is Gödel numbering
• The availability of the quotational device allows it to
present – in addition to the straightforward numbertheoretical claim it makes – itself. We might call this
“implicit” self-quotation
• Roughly speaking, the quotational device yields access
to the (syntactic) properties in virtue of which the
sentence has the (first-order) content it does
• For the Gödel sentence, there is a sense in
which even a complete description of its
intra-systematic (i.e. syntactic) properties
leaves open the question of its truth
• Analogously, if conscious states are implicitly
self-quotational (in something like the Gödel
sense) then even a complete description of
their real, intra-systematic (i.e.
functional/structural) properties might leave
open the question of whether they’re
Further questions
• In the Gödel case, (implicit) self-quotation is
achieved by the device of using Gödel numbers to
encode syntactic properties. What’s the
analogous device in the consciousness case?
• In the Gödel case, we only achieve an insight into
its truth by “jumping out” of the system and
considering systematic properties of the whole.
How do we achieve (the analogue to) this
“jumping out” in the consciousness case?
• Taking up insights of phenomenology, it’s
reasonable to construe conscious states as acts
that (implicitly) present their own ideal contentdetermining properties.
• This kind of self-presentational theory solves
problems that arise for (other) selfrepresentational theories
• If we can make sense of such self-presentation in
virtue of ideal content-determining properties,
we might be able to account for the existence of
an explanatory gap

similar documents