- seCtIon one: tHe PRoBLem
essential to understanding consciousness: ‘Consciousness is in the first place not
a matter of “I think that” but of “I can” ’ (1945/2002, p. 159). O’Regan and Noë argue
that classical theories of vision do not explain how the existence of an internal
representation can give rise to visual consciousness (another version of the hard
problem). In their theory, the hard problem is avoided because ‘The outside world
serves as its own, external, representation’. Instead of being about building rep-
resentational models, ‘seeing is a way of acting. It is a particular way of exploring
the environment’ (2001, p. 939).
More specifically, an organism has the experience of seeing when it masters the
governing laws of sensorimotor contingencies – that is, when it develops skills
for extracting information from the world; for interacting with the visual input
and exploiting the ways in which that input changes with eye movements, body
movements, blinks, and other actions. On this view what you see is those aspects
of the scene that you are currently ‘visually manipulating’. If you do not interact
with the world, you see nothing. When you stop
manipulating some aspect of the world, it drops
back into nothingness.
As with Rensink’s virtual representation, what
remains between saccades is not a picture of the
world, but the information needed for further explo-
ration. A study by Karn and Hayhoe (2000) confirmed
that spatial information required to control eye
movements is retained across saccades. Could this
be sufficient to give an illusion of continuity and
stability?
This theory is radically counter-intuitive, not least
because seeing does not feel like manipulating tem-
porary aspects of the world that then disappear, but
O’Regan likens it to the light inside your fridge. Every
time you open the door the light is on. Then you
close it and wonder whether it’s still on. So you open
it and look again. It’s still on. So it is with the world:
it is always there when you look, so it’s easy to think
you have a constant detailed representation of it.
Sensorimotor theory is dramatically different from
most existing theories of perception, but is closely
related to theories of embodied or enactive cogni-
tion (Chapter 8). It is similar to the idea of perception
as a kind of ‘reaching out’ (Humphrey, 1992, 2006),
to theories stressing the interdependence of percep-
tion and action (Hurley, 1998), to J. J. Gibson’s (1979)
ecological approach to perception, and further back
to Merleau-Ponty’s idea that ‘consciousness is noth-
ing other than the dialectic of milieu and action’
(1942/1965, pp. 168–169). Seeing does not mean
building representations of the world that can then
be acted upon; rather, seeing, attending, and experi-
encing are all kinds of action.
‘seeing is a way of acting
[. . .] of exploring the
environment’
(O’Regan and Noë, 2001,
p. 939)
‘sensations are
representations of
something we do’
(Humphrey, 2016, p. 117)
seeInG oR BLInD? A tHoUGHt
exPeRIment to test
sensoRImotoR tHeoRY
According to o’Regan and noë’s (2001, p.
1020) sensorimotor theory, ‘perception is
constituted by mastery of sensorimotor con-
tingencies’. seeing means manipulating the
contingencies between action and input, such
as moving one’s eyes and getting changed
visual input. A thought experiment suggests
a bizarre consequence of this theory.
C
on
C
e
P
t
3.2
FIGURE 3.10A • Alva at start