540 Tim Lewens
offers an additional promise not merely of characterising organic development as
directed towards a target, but of explaining why organic development takes the
pattern it does because it is directed at some target. So we might say that nutrients
are recruited into a seedling via the roots because the seedling has the attainment
of the adult form as a goal.
Today such goal-based theories are highly unfashionable ([Boorse, 2002] is an
exception). Obviously the characterisation of organic development, or the main-
tenance of organic integrity, as directed towards a goal is almost irresistible. The
question for goal-based theories of function is whether such characterisations can
be justified as anything more than anthropomorphic projection, and specifically
whether one can give a satisfactory account of what makes it the case that a pro-
cess is directed towards some end in a way that does not fall back on something like
the selected effects account. I do not propose to run through all the objections to
goal-directed theories of function here, but I will give a selective review. Nissen’s
[1997] thorough assessment of the literature on teleology is a good source for those
wishing to know more, and I have drawn on it in the following paragraphs.
Crucial to viewing development as goal-directed is the possibility of goal failure.
If, for example, we want to characterise seedlings, or embryos, as directed towards
the attainment of adult form, then we must be able to give an account that can fix
the adult form (or the capacity for reproduction) as an organism’s developmental
target even when, because of disease, or some other internal or external influence,
it does not reach that target. This condition immediately rules out a number of
accounts of goal-directedness. We might say that organisms are directed towards
whatever final states they will, as a matter of fact, assume. Goal-directedness is
here accounted for merely in terms of actually realised ends. Here, however, we
face a problem of how to account for goal failure. If a particular hen embryo fails
to develop into a hen, shouldn’t we say that it failed in its goal? But we cannot
say this if we identify the goal only with how that embryo turned out as a matter
of fact.
This cannot be fixed by saying that the embryo is directed towards whatever
final state it is most likely to arrive at. This rules out the coherence of congenital
defects, which are facts about an embryo that make it unlikely to realise the state
it ought to realise, and it also forces us to say that some environmental factors that
make attaining the adult state very unlikely do not perturb the organism from the
adult state, but instead change the organism’s goal. Perhaps all of these problems
can be fixed if we say instead that an embryo is directed towards whatever outcome
most other embryos of the same type actually arrive at. Yet the problem here is
that in many cases we want to say that organisms are directed at ends which most
of them fail to realise. Most acorns fail to turn into oaks, yet that should not rule
out our identifying the production of an oak as an acorn’s goal.
One might also worry that goal-directed accounts of function will suffer from
problems of promiscuity. Physical and chemical systems tend to assume equilib-
rium states, and here, too, one need not identify the state they aim at with the
state which, as a matter of fact, they will assume. A ball bearing that is rolling