Jonathan Waskansmells and emotions, due to seizures or electrocortical stimulation. Patient M.M., for instance,
reported the following: “‘I had ... a familiar memory, in an office somewhere. I could see the
desks. I was there, and someone was calling to me, a man leaning on a desk with a pencil in his
hand’” (Penfield 1958: 59). This suggests that (in humans) brain events suffice to generate qualia,
that qualia are thus constituted by these events. If so, invoking body and world in the robot case
would seem no more helpful or warranted.
One response to this worry is to opt for disjunctivism (Noë 2006), the view that there
are two independent classes of qualia: the (aberrant) internal kind and the normal (partly)
external kind. Still, one wonders: If there is a purely neural cause of qualia in the former
case (with no relevant external objects), why add to the story—indeed, parsimony dictates
otherwise—in the latter?^3 Surely the same neural events underpin qualia (e.g., experienced
redness) in both cases.
Some maintain, however, that the extra-neural component of qualia is not the outside world
in the usual sense, but the contents of our representations of it (Lycan 2001; Tye and Byrne 2006).
Contents seem a good candidate for this role because they are widely considered to be (at least
partly) external to the brain, part of the natural order, and at times present when their target
objects are absent. On this way of thinking, if I experience a red tomato before me and there
is none, there still is a (ex hypothesi external) mental state with content (e.g., red tomato there).
Extrapolating to robots, if their internal states have appropriate (external) contents, they too can
be expected to have qualia.
However, as a strategy for explaining qualia, content externalism appears as a kind of parlor
trick as soon as we ask: Exactly what part of the external physical world produces, and thereby
explains, the qualitative character of our experience of a red tomato? The experience is real.
Now point to it! Lycan (2001) floats the idea that the relevant content is “a nonactual physical
individual” or a set of possible worlds, but these intangible abstracta are poor candidates for the
physical producers of, and thus nonstarters as explanations for, qualia. Perhaps recognizing this,
Tye (1999) opts for the view that the seeming inability of contents to physically explain qualia is
itself best explained away with a variant on the cognitive illusion strategy.
7 Look Up for QualiaThere is, admittedly, a certain pull to the idea that we are looking in the wrong place for qualia
and that if we continue this course we will remain befuddled, confounded, and unable to rep-
licate qualia in our robotic contrivances. But rather than looking out, perhaps we should be
looking up, and not just past Leibniz’ structural level, but also past Searle’s programming level.
Consider all of the different ways of building a computer that implements MSWord. Starting
at the bottom, there are various materials (e.g., from copper wire to nanotubes) that can be used
to implement particular architectures (e.g., Turing, von Neumann, super-scalar)—that is, each of
the latter is multiply realizable. Likewise, a program written in a particular language (e.g., C++,
Lisp, Visual Basic) can be built from any of these lower-level architectures.
But have we now reached the pinnacle? Perhaps not. For instance, what if a relatively high-
level C++ program realizes a Java virtual machine (a simulation of a type of machine), which
runs a Java program, which implements the Minecraft virtual environment (a 3D sandbox
game), which (through its proprietary array of blocks, redstone circuits, repeaters, and the like)
implements a machine that approximates a universal Turing machine, which implements Visual
Basic and, finally, MSWord? This is possible, at least in principle. There can thus be layers upon
layers upon layers such that we may never be sure that we’ve reached the pinnacle, assuming
there is one. And at each new layer we encounter entities and activities with properties that are