SUPERVENIENCE
— 850—
facts to vary independently of the physical facts of
our world when considered in toto. Furthermore,
since this holds for any logically possible physical
duplicate of our world, it follows that biological
properties logically (globally) supervene on physi-
cal properties.
This is a remarkable result. One may well
wonder whether under such broad conditions of
global supervenience there can be any property at
all that could fail to supervene on the (micro-)
physical facts of an entire world. If not, physicalist
materialism would carry the day. This brings us to
the second distinction mentioned above, the dis-
tinction between logical and natural superve-
nience. This is the more interesting distinction be-
cause it leads straight into highly controversial
territory. Generally speaking, B-properties natu-
rally supervene on A-properties if any two natu-
rally possible situations with the same A-properties
also have the same B-properties. In other words, in
the case of natural supervenience, the B-facts are
nomologically, though not logically, implied by the
A-facts. That is to say, in possible worlds that are
governed by the same natural laws as the actual
world, the A-facts naturally necessitate the B-facts
(assuming natural supervenience). Clearly, natural
possibility is much stricter than logical possibility.
For example, a (stable) cubic kilometer of ura-
nium-235 is logically, but not naturally, possible.
The critical question then is: Are there any (higher-
level) properties that accompany the physical facts
in all naturally possible worlds without being fixed
by the physical facts in all logically possible
worlds? And the controversial answer given by
some philosophers is: Consciousness is such a
property. On the one hand, they argue, conscious-
ness at least naturally supervenes on the physical
facts because any two physically identical creatures
in the natural world will presumably have qualita-
tively identical phenomenal experiences. Never-
theless, these philosophers go on to argue, con-
sciousness fails to supervene logically on the
physical facts of our world. Here they appeal to
two famous thought experiments. It seems entirely
conceivable that a creature physically identical to a
conscious creature might lack consciousness alto-
gether (like a zombie), or might have experiences
qualitatively very different from ours (they might
have so-called inverted qualia, so that, for exam-
ple, they might have our sensation of phenomenal
red when looking at the sky). Therefore, if these
two intuitions hold, materialism is false. A full ac-
count of the physical facts of our world, including
a specification of the minutest details of the distri-
bution of its microphysical properties over space
and time, would yet leave entirely undetermined
the quality, even the existence, of the phenomenal
properties of our world.
Principle of multiple realizability
As discussed above, supervenience may yield on-
tological dependency relations of varying degrees.
But how do things stand with respect to the other
philosophical benefit reputedly reaped from this
recently developed notion of supervenience, that
of barring reducibility? This becomes apparent
when one focuses on the converse relation implied
in the definition of supervenience. Assuming su-
pervenience, while two situations cannot differ in
their B-properties without a corresponding differ-
ence in their A-properties, the converse does not
hold. That is to say, type-identical B-situations may
be realized by an indefinite variety of type-
different A-situations. In other words, the notion of
supervenience brings in its wake an important
corollary notion, that of multiple realizability. And
again, this feature holds special interest for the phi-
losophy of mind because multiple realizability is
just what we expect in the mental realm. Pain in
humans may be realized by C-fibers firing, while in
dolphins it may be realized by D-fibers firing with-
out ceasing to be just another simple instance of
pain. Indeed, the situation may be vastly more di-
verse and confusing at the physical level than this
example suggests: Your headache may be physi-
cally realized differently from mine, as, for that
matter, may my headache today versus my
headache tomorrow. Similarly, the property of
being a monetary transaction, which is a unitary
concept at the level of economics, may be physi-
cally realized by a wide variety of physical events
lacking any perspicuity or explanatory integrity at
the level of physics. Accordingly, the predicates of
a given special science will only map onto predi-
cates of physics that are at best wildly disjunctive.
Thus, inasmuch as supervenience entails multiple
realizability, higher-level supervening properties
turn out to be irreducible. In psychology, in par-
ticular, it has been argued, there cannot be any
type-type identities between mental properties and
the physical properties realizing them. Nor, conse-
quently, is there any room for strict psychophysical