352 Alex Rosenberg
An even vaster disjunction of nucleic acid sequences than the actual sequence will
work just as well, or indeed just as poorly to constitute the functional hemoglobin
gene (and probably will do so in the future, given environmental contingencies and
mutational randomness). Just think of the alternative introns that could separate
exon regions of the sequence (and may do so in the future, given mutation and
variation). And then there are all the promoter and repressor genes, and their
alternative sequences, not to mention the genes for producing the relevant riboso-
mal protein-synthesizing organelles, all equally necessary for the production of the
hemoglobin protein, and so claiming as much right to be parts of the functional
hemoglobin gene as the primary sequence of the coding region of structural gene
itself. Just as the actual disjunction is too complex to state, and yet not biologi-
cally exhaustive of the ways to code for a working hemoglobin protein, so also all
these other contributory sequences don’t exhaust the actual biological alternatives,
and so make the macromolecular definition of the functional hemoglobin gene a
will-o-the wisp.
In other words, being a hemoglobin molecule “supervenes”, in the philosopher’s
term, on being a particular sequence of amino acids, even though there is no
complete specification possible or scientifically fruitful of all the alternative par-
ticular sequences of amino acids that could constitute (i.e. realize) the function
of the hemoglobin molecule in oxygen transport. Roughly, a biological property,
P, supervenes on some (presumably complex) physical and/or chemical property
or other, Q, if and only if anything else that has physical/chemical property Q,
must also have biological property P, (see [Rosenberg, 1978]). There is among
philosophers a fairly sustained debate about the force of the “must” in this for-
mulation. Does the supervenience of the biological on the physical/chemical have
to obtain in virtue of natural laws, or even some stronger sort of metaphysical
necessity. As many philosophers argue, biological properties are “local” — make
implicit but ineliminable reference to a particular place and time (the Earth since
3.5 billion years ago). Thus, it may be that biological properties are only locally
supervenient, a much weaker thesis than one which makes it a matter of general
law every where and always in the universe, (see “Concepts of Supervenience” in
[Kim, 1993]).
When a biological property is supervenient on more than one complex physi-
cal/chemical property then it is also a multiply realized property. The superve-
nience of the biological on the physical is a way of expressing the thesis of physi-
calism. The blindness of natural selection to differences in structure is what turns
the supervenience of the biological on the physical into the multiple realization of
the biological by the physical. This structural diversity explains why no simple
identification of molecular genes with the genes of population genetics of the sort
post-positivist reduction requires is possible. More generally, the reason there are
no laws in biology is thus the same reason there are no bridge-principles of the
sort post-positivist reduction requires (This result will be even less surprising in
light of the post-positivist realization that most bridge principles in science will
be laws, not definitions.)