178 J.J. Haldane
of electrons and protons’ (chapter 3, p. 153). The reductionism remains, as
does my objection that it is unwarranted by empirical observation and theory.
On the contrary we should suppose that the irreducibility of explanatorily
rich and powerful biological theory is evidence for the reality of biological
entities and powers – including teleological ones. This is in no way incompat-
ible with the claim that such entities are composed of electrons and protons;
butcompositionis not identity. We need to distinguish in living things, as in
artworks, between the medium of realization and that which is realized in it,
and to acknowledge that both are real.
My next step was to argue that natural ‘mechanico-evolutionary processes’
are not sufficient to explain the emergence of living things, reproductive
species and thinking animals. Smart addresses the transition from the non-
reproductive to the reproductive by suggesting that there is no problem for
naturalism and that a hypothesis of this sort is to be preferred. Why is
naturalism unembarrassed? Because, Smart supposes, there is no difficulty
in principle to the natural emergence of replication: ‘Why could not a self-
replicating molecule come about through the coming together of a number
of non-replicating molecules?’ (chapter 3, p. 152). Well, first of all self-
replication is not sufficient for evolution. The latter requires reproduction
involving the coming to be of entities sufficiently like their predecessors to be
continuants of their basic nature but sufficiently different to allow for selec-
tion, in point of varying degrees of adaptation, to take place. And what
reproduction requires is a highly adaptive mechanism of the sort which it has
been the goal of evolutionary theory to explain. Given teleology, evolutionary
theory has a role; the question is whether it is intelligible to suppose that it
could be the whole story.
Second, however, is the fact that there is no satisfactory naturalistic
account even of primitive replication. There is nothing unintelligible about
the idea which Smart mentions, that life on earth may have begun in conse-
quence of organic molecules having arrived here from interstellar space. The
problem with this suggestion, which Smart does not himself endorse, is that
it is regressive. In answer to the question of ultimate origin Smart offers what
is in effect a ‘why not?’ reply. Why could replicating molecules not just arise
from non-replicating ones – be it that this occurrence may have a very low
probability? My objection, however, was not to this being more or less likely,
but to the very idea that there could be a natural explanation of the emergence
of replicators from non-replicators. Indeed, this is just the sort of case that
illustrates the notion of radical emergence. Some feature F is radically emer-
gent if it is novel in a subject S (i.e. if it is not just a linear combination of
instances of the same property type, as the size of a quilt is a linear function
of the sizes of its constituent squares), and no naturalistic theory of the
components of S can predict or explain F.^6 Thus my claim is that the power