Susan Hurley and Alva Noë
Distinguish comparative
and absolute explanatory gaps: why does
neural activity in a particular area of cortex have this qualitative
expression rather than that, and why does it have any qualitative
expression at all? Our general
strategy: focus on comparative gaps,
both intermodal and intramodal; set absolute gap aside. Address comparative gaps not by inward
scrutiny but rather by expanding our gaze to include relations between brain,
body and environment.
We introduce a distinction between cortical dominance and cortical
deference, and apply it to various examples of neural plasticity in which
input is rerouted intermodally or intramodally to nonstandard cortical
targets. In some cases but not others,
cortical activity 'defers' to the nonstandard sources of input. We ask why,
consider some possible explanations, and propose a dynamic sensorimotor
hypothesis. We believe that the distinction is important and worthy of further
study, both philosophical and empirical, whether or not our hypothesis turns
out to be correct. In particular, the
question of how the distinction should be explained is linked to intermodal and
intramodal comparative gaps.
1.
The distinction
introduced: cortical dominance vs.
cortical deference
When
inputs are neurally rerouted, does the area of cortex activated by a
nonstandard source of input dominate the nonstandard input source and retain
its normal qualitative expression or does it defer to the nonstandard source of
input? Dominance may be assumed as the
norm. But: 1) deference occurs also; 2) neural supervenience does not entail
dominance.
2.
The
distinction schematized.
A
& B: peripheral sources of input
1
& 2: cortical areas normally
activated by inputs from A and B, respectively
Rerouting: A instead of B projects to 2
When
2 is activated by A, is its qualitative expression the B feeling (dominance,
illusory) or the A feeling (deference, veridical)?
A/B
comparison can be intermodal or intramodal
Qualitative A-feeling ??
B-feeling: dominance
Expression: A-feeling: deference
Of
Cortical 1 2
Area:
Input A
B
Source:
3. The distinction at work.
Intramodal
dominance: referral of touch to face to
phantom arm
Intermodal
dominance: color-graphemic synaesthesia
(under one interpretation)
Intermodal
deference: visual to auditory rerouting
in ferrets; early blind Braille readers
(Intramodal deference:
see section 7 below)
4. How can the distinction be explained?
Intermodal deference/intramodal dominance? Not satisfactory: there seems to be intermodal dominance, and even if intermodal
rerouting is only necessary for deference, we’d want to know why.
Early deference/late dominance? Is there early dominance in congenital
phantoms? Again, even if early
rerouting only necessary for deference, we want to know why.
Both
these suggestions explanatorily shallow.
Extend
deference to rerouting between distal and peripheral sources of input, not
between peripheral source and cortical target.
TVSS: Distal source of visual input rerouted to peripheral source of tactile input, and on to somatosensory cortex, with arguably visual qualitative expression. This only occurs if subject controls movement of camera, not if someone else does or if he is moved passively. Switch from tactile pattern of dynamic sensorimotor (DSM) contingencies to visual pattern explains intermodal deference in TVSS; know-how governed by characteristically visual patterns of DSM contingency makes TVSS-perception qualitatively visual.
Not
remapping from input sources but changes in DSM contingencies drives changes in
qualitative expression. Distinct
patterns of DSM contingencies characterize different modalities and different
qualitative characters within modalities.
Intermodal
changes in qualitative expression driven higher-order changes in relations
between DSM mappings. Difference
modalities governed by different characteristic patterns of DSM contingency.
The link between such DSM patterns and
qualitative character is intelligible, not brute. Relevance to intermodal comparative explanatory gaps. Does the point extend to intramodal
comparative gaps? Yes.
Sources
of visual and proprioceptive inputs:
left arm and right arm
Cortical
target areas: left and right visual
cortex and left and right proprioceptive cortex
Goggles
produce external remapping between sources of input and visual cortex by
left-right reversing goggles: conflict
between co-stimulated areas of visual and proprioceptive cortex. Old DSM contingencies no longer hold. Taylor:
veridical visual adaptation with practice. Deferential intramodal change in qualitative expression of visual
cortex. DSM hypothesis: deference driven by loss and regaining of
DSM know-how.
Harris: nonveridical proprioceptive adaptation to
reversing goggles, not veridical visual adaptation. Also, familiarization makes vision seem normal. Still a kind of intramodal deference, driven
by change in DSM contingencies. But
which kind of deference does DSM view predict?
Regaining
know-how when DSM contingencies change requires intermodal harmony and ability
to negotiate environment. Veridical visual adaptation à la Taylor does
both. If adaptation instead involves
illusory proprioception, à la Harris, then a canceling secondary adaptation or
illusion is also required to restore full know-how: your wedding ring on your right hand REALLY looks leftward to
you, but comes to seem to look rightward.
This secondary adaptation would restore your know-how, but at cost of
denying self-evidence of the qualities of your experience. DSM view predicts the two illusions cancel
out qualitatively as well as practically.
Qualitative
expression cannot be explained just in terms of area of cortex active, nor
additionally in terms of character of peripheral input sources. Rerouting, internal or external, changes
pattern of DSM contingencies in which given areas of cortex participate. Deference reflects agents’ know-how in
relation to patterns of DSM contingency characteristic of specific modalities
or qualities but using nonstandard neural paths.
DSM
approach regards deference as the norm.
It predicts dominance where agent is passive or where rerouted input
generates dangling cortical activity, not tied into patterns of DSM
contingency.
DSM
approach predicts deference for ferret and Braille cases, dominance for phantom
case. Further prediction: if phantom patient stroked own face, while
watching in mirror, would get deference instead. Explanation of dominance in synaesthesia: clue from abnormal covert priming?
DSM
hypothesis promising but not proven, worth further attention.
Main aims: draw dominance/deference distinction, relate
it to comparative explanatory gaps, raise question of how to explain
distinction. Have also proposed an
expanded gaze strategy and a DSM hypothesis.
An
empirical account can scratch explanatory gap itches. DSM approach compatible with neural supervenience, but more explanatory. Characteristic patterns of DSM contingency
are more qualitatively scrutable than NCCs (taken out of context of DSM
patterns in which they function).
Neural supervenience may be true, but inward scrutiny still be the wrong
way to address explanatory gaps.
Very preliminary notes
for second short talk on work in progress:
How the DSM hypothesis is
compatible with neural supervenience but more explanatory.
11. Neural supervenience of qualitative
character as orthodoxy.
General
form: no mental difference without a
nonmental difference. To be fully
specified supervenience claims require some parameter tweaking: are duplicates covered in same, all, or
nearby worlds? What is relevant
boundary? Which types of mental
properties covered: content vs.
qualitative character?
Here: neural supervenience of qualitative
character across near worlds. Nonmental
neural duplicate in same qualitative states as twins in near worlds.
Neural
supervenience not disputed, but it no more explains our distinction than it
addresses comparative explanatory gaps.
Neural supervenience does not predict cortical dominance or rule out
deference reflecting extraneural sources of input.
Distinguish
two reasons for claiming that supervenience is compatible with cortical
deference.
If
we extend the boundary within which qualitative character supervenes on
intrinsic properties, we can preserve supervenience. Not our reason: neural
supervenience is compatible with deference to extraneural sources of input.
Our
reason: Rerouting can induce changes in
intrinsic neural properties of cortical activity even at the level of single
cells. This can happen in cases of
deference just as much as in cases of dominance. Thus deference reflecting even extraneural sources does not
challenge neural supervenience.
14. Why
the distinction needs explanation whether or not qualitative expression
supervenes on neural properties.
Neural
supervenience, even neural correlates of consciousness, don’t explain why we
get deference in some cases, dominance in others, any more than they address
the comparative gap question: why do
these neural properties have this qualitative expression rather than that?
Objection: Just use the table of neural correlates of
consciousness (once we have it) to look up which neural properties have which
qualitative expressions. There may be no further explanation.
Neural supervenience though true may mislead us by
directing our attention inwardly to intrinsic neural properties, which are
qualitatively inscrutable, thus leading straight to the comparative explanatory
gaps. But we may be looking in the
wrong place. How hard it is to bridge
explanatory gap may depend on how the nonphenomenal side of the gap is
conceived.
Some invoke protopsychons or fundamental
psychophysical laws at this point.
Our
strategy: expand our gaze to include
DSM contingencies, complex and active relations of embodied organisms to their
environments, as well as neural activity.
Our DSM hypothesis has the potential, if correct, to scratch explanatory
itches in a way that neural supervenience, though correct, does not.