544 Tim Lewens
us about the organic world. What allows us to identify this, rather than any
other equally unlikely arrangement of organic matter, as a goal, including the
arrangements of matter that constitute dead organisms, and which we tend to
think of as uninteresting and unworthy of special explanation? This account posits
internal developmental plans, but as yet has offered no account of what makes a
plan directed towards one kind of outcome rather than another. Indeed, we might
worry that it is only the fact that we find organisms puzzling that leads us to see
adult states as in need of privileged explanation at all.
7 KANTIAN PROJECTIVISMThere is a line of argument that can be found in (or perhaps projected onto)
Kant, which makes a virtue of these concerns. Kant suggests that we must view
organisms in teleological terms — that is, as the product of something analogous
to a rational plan — if we are to do biology at all. But Kant also dismisses the
thought that, for humans at least, this reveals anything about the true explanation
of organic form. Viewing organisms as directed towards particular goals is merely
heuristic — it leads us towards the proper explanation of organic form, without
containing any genuine explanation itself.^2
Here is my best effort to use Kant’s work to fashion a defensible heuristic view
of goal-directedness (for a detailed philosophical treatment of Kant’s work on
biology I recommend [McLaughlin, 1990]). Let us begin with the problems that
face a statistical account of goal failure; many young organisms do not, in fact,
develop to be adults. For many species, the great majority of sperm and eggs will
never combine to produce organisms at all, and many embryos will not survive
beyond early developmental stages. The embryologist could, of course, attempt
to give mechanical explanations for every one of these outcomes in terms of the
interactions of the various causal contributors to development, and the ways in
which some patterns of interaction yield short-lived entities, others yield longer-
lived entities, and still others yield nothing at all. Yet such a catalogue would be
unmanageable; the embryologist instead will choose to focus on some small set of
outcomes as the ‘normal’ ones, cataloguing mechanical processes of development
according to how they contribute to that normal development. Not only would the
catalogue be unmanageable, the classification of some organisms’ developmental
trajectories as short-lived and others as long-lived itself presupposes some kind
of standard for what development should look like. Development is development
(^2) A note for Kantians: in this section I am taking Kant’s contrast between ‘constitutive’ and
‘regulative’ principles as roughly equivalent to a contrast between principles that reflect the law-
like workings of nature, and principles that reflect the methodological rules that investigators
should follow to uncover nature’s workings. Joan Steigerwald has warned me that this way of
glossing the division between the constitutive and the regulative, as well as the consequent divi-
sion it leads to between the heuristic role of teleological principles and the genuinely explanatory
role of mechanical principles, may fail to do justice to the complexities of Kant’s thought. I hope
that readers will indulge an exploration of a loosely Kantian approach to biological teleology,
even if that approach is far from Kant’s.