The Inherent Normativity of Functions in Biology and Technology 123
logical categories. Neander has since backed down from a similar view in her early work, but it is unclear how
far she has moved on. See Neander (2002) for a recent account.
- A pump that is “broken” by being completely smashed to pieces is usually not considered a pump anymore,
nor is any blob at the place where ordinary people have their eyes an eye. Cf. Davies (2001: 177). - This at least is what its proponents claim. Millikan’s PF theory, however, makes heavy use of the word Normal,
meant as a technical term, not just in the notion “Normal explanation,” which she defi nes, but also “Normal
member,” “Normal condition,” “Normal property,” and “Normal production” (Millikan 1984: 24–34). It is tempt-
ing to single out this notion of Normal as grounding the normativity of proper functions, but how would one
substantiate this, since on the one hand, there is no way in which it allows the derivation of “should” or “ought
to” statements and on the other hand, Millikan presents it as just one of the many concepts by which she builds
her naturalistic, i.e., nonnormative, theory of function, and not as the linchpin. Some call this notion of Normal
a normative notion (e.g., Rowlands 1997), but then what does this mean, and is it compatible with Millikan’s
claim to give a naturalistic account of proper function? Although this point deserves to be investigated in more
detail, I do not do so here. - Some things are not an item’s proper function but are necessary accompaniments of its proper function,
such that normative statements are in order: it is not a heart’s proper function to make a bumping noise,
but nevertheless, taking this sort of talk for granted momentarily, a heart ought to make a bumping noise all
right. - Indeed a theory that would extend that far could be accused of committing the naturalistic fallacy. Since the
status of the naturalistic fallacy is very controversial, I do not elaborate this point here. Note, however, that if it
is considered legitimate to say that x ought to do B because it is the proper function of x to do B, this “because”
must be read constitutively, similar to the way it has to be read in “This pawn cannot go to that fi eld because it
is against the rules of chess for it to go there.” But if the relation between proper function and “ought to” is that
intimate, then any naturalization of proper function must run into problems. - In section 7.7 I give an account that has “ought to” statements as central.
- How little thought seems to be given to what we precisely express by these statements can be seen, to take
just one example, by the following quotation from McLaughlin (2001: 4): “Does the attribution of function pre-
suppose a valuation of the end towards which it is a means—at least in the sense that the function bearer is sup-
posed to perform its function?” It is precisely the question what a statement like “x is supposed to perform F”
has to do with a normative notion like value, but McLaughlin seems little interested in posing it. - The two forms are not completely independent. One hardly can be morally justifi ed in expecting something
to happen if on epistemic grounds one believes it to be extremely improbable or impossible. Promises of impos-
sible things should be rejected. This may lead one to question the equation of being morally justifi ed in expecting
something to be the case with having a right to it that this something is the case. See also note 22. - A more accurate discussion would introduce the distinction between manifest and latent function. Supporting
social cohesion is the latent function of the Hopi rain dance, the one that social scientists are interested in.
Additionally the rain dance may be considered as an institution designed and implemented for the purpose of
promoting rainfall, which is its manifest function. We do not as easily say that this institution functions well or
poorly as we do in the case of a country’s legal system, partially because it may be doubted whether the rain
dance was ever designed in the way legal systems are. - Although it is rarely done, I think that one can sometimes be justifi ed in applying the notion of “malfunc-
tion” also to cases where someone uses an object incidentally for some purpose or other. This object may be an
artifact, but one that was designed to be used for a different purpose than it is used for on this occasion. In such
cases, the user of the object can be seen as the designer of his or her own private artifact. - Here it is assumed that wherever the dark variety is frequent, it is because of its superior performance in
being inconspicuous to its predators. It has been challenged whether Kettlewell’s famous experiments from the
1950s indeed showed this to be the case (see Rudge 1999). - Perhaps we fi nd this less obvious in the case of an ideal liver than in the case of an ideal mating dance,
which again testifi es to the special position of bodily organs in biology as soon as normativity is an issue. Would
an ideal mating dance be one that would end in copulation in virtually all cases it was performed? That would
probably not be good for the animal itself, and perhaps also not good for the species. - Raz (1975), however, uses “ought to” in a weaker sense. He says that “p ought to do Z” even when p just
has one reason to do Z.