60 Françoise Longy
scientifi c background is needed to come up with a plausible mechanism. Often a layman’s
theoretical background is enough to arrive at a sensible hypothesis. It is often possible for
the layman to grasp something relatively simple that may be a part (a submechanism, let
us say) or a rough sketch of the much more complex mechanism responsible for the high-
level causal relationship. Not much is needed to sketch in broad outline a mechanism that
may explain how a property like the human preference for CFR corn may have produced
steadily a high ratio of CFR corn in cultivated land.
It must be stressed that only such a top-down explanation, an explanation relying on a
mechanism or structure visible only when contemplating matters from a certain level of
abstraction, can meet the task of explaining a kind of stability that does not result from
physicochemical laws. Without such a top-down perspective such stability will usually be
inexplicable. To continue with our example, without a mechanism explaining how the
better taste or digestibility of CFR corn could have acted steadily in favor of CFR corn,
the probable instability of genetic and climatic conditions in the period concerned would
make the long-lasting presence of CFR corn a mysterious and highly improbable fact. A
causal explanation remaining at a lower level will not in general be able to account for
the stable causal dependency the function points to, no matter how detailed it is. If, for
example, you knew whether the genetic makeup plus the actual growth conditions of the
type of wheat cultivated most at present induced the presence of the CFR feature, you
would be in a situation to explain why 95 percent, let us say, of the studied wheat would
produce CFR corn. However, this would not explain why it was also roughly the same
before (95 percent of cultivated wheat producing CFR corn) when the genetic pool of
cultivated wheat and the growth conditions were somewhat different. It would also give
you no reason to expect the same or another ratio for the cultivated wheat not yet studied,
or for cultivated wheat three hundred years from now, if a very small change in growth
conditions or in the genetic pool were to modify the carbohydrate–fi ber ratio of the
corn.
I conjecture that this point is an aspect that could, once properly elaborated, explain and
justify the importance of functional explanations as a particular type of causal explanation.
Functions provide the basis for an important sort of top-down explanations. They supply
the right causal frame for explaining a type of phenomena that cannot be explained satis-
factorily bottom-up. Such are the phenomena that rely on the existence of complex mecha-
nisms that produce stable causal connections while different levels of reality (physical,
biological, psychological, etc.) are involved. However, this is not the place to develop this
point further. The conclusion I want to draw here concerns only the interpretation of the
diffi culties encountered in seeking to demarcate biological functions from artifactual and
cultural ones. Such diffi culties do not result from a lack of clarity at the conceptual level.
They derive from the fact that some functions point to stable causal connections that
depend on very complex multilevel mechanisms. The attempt to separate the biological
elements from the artifactual or the cultural ones would inevitably result in making every