Functions 543of artefacts, we specify what a plan is for by reference to the intentions of the
architect; for organisms we will say that the developmental plan, itself nothing
more than some arrangement of developmental materials that will help to build
the organism, specifies the form of the adult organism in virtue of some aspect of
its evolutionary history. For most theorists this will turn on what the elements of
the developmental plan have been selected for. The adequacy of this kind of view
will depend in the main on how we assess the selected effects account of functions,
hence this move in effect gives up on any distinctive account of goal-directedness
as the basis for function claims, relying instead on accounts of function of the form
we examined in the first half of this essay.
An alternative is to think of a developmental plan as an organic analogue of an
intentional state, whose properties are specified as those which are adequate to ac-
count for the repeated formation of the adult organism itself, and whose existence
we should believe in simply on the grounds that it offers a good explanation of
regular adult development. Paul Griffiths reminds me that Ernst Mayr may have
held something like this view. Mayr defends the notion of informationally rich
‘programs’ that account for the goal-directedness of both individual development
and individual behaviours. He acknowledges that natural selection is capable of
‘producing and perfecting’ these programs [Mayr, 1961, 1504], but he is resistant
to making their content and existence logically dependent on their evolutionary
history. Mayr’s later analysis of a program as ‘coded or prearranged information
that controls a process (or behavior) leading it to a given end’ [Mayr, 1974, 102]
does not eliminate the problematic notion of goal-directedness, for the outcome
that is encoded in some information-bearing structure is not merely the outcome
the structure actually promotes. Perhaps, then, we should take the developmental
plan or program as primitive, while offering an inference to the best explanation
as the argument for its existence. It certainly is a puzzle how, in spite of a chang-
ing environment, an adult cow and bull will nonetheless be able to produce a calf
that will grow up to resemble the adults. This remains a puzzle in spite of hints
that discoveries in genetics have solved the problem of inheritance: no one thinks
that genes are sufficient to explain trans-generational resemblance, for what effects
genes have are contingent on environmental background, so that it appears that all
the elements of development need to be placed appropriately for reliable develop-
ment to occur. It seems we have no reason to think that chance alone would place
the elements of development in the right pattern generation after generation, and
nor are we aware of any general laws of nature that might guarantee that these
materials will be structured so as to facilitate development. So it is legitimate
to suppose that some sort of guiding plan must exist, if we are to explain the
phenomena of development at all.
The main problem with this kind of account is its methodological limitations; we
explain the regularity of development merely by positing some kind of plan able to
bring about such development, and thereby forestall further investigation into the
material causes of regular development. The account also offers no enlightenment
regarding the question of goal identification. Adult development is what puzzles