122 Maarten Franssen
account.” Millikan’s PF version of the etiological theory and Krohs’s SD variant of the
CR theory are currently the only theories that can account for the attributions of functions
to biological items as well as to social items and to artifacts. The way the PF theory attri-
butes functions to artifacts is, however, quite abstract and is in a sense a by-product of the
theory’s task of naturalizing intentionality, for which it was developed.^25 Boorse claims
that his GC theory is a third theory capable of dealing with biological items and artifacts
in one stroke, but as it seems to me, the theory faces great diffi culties in accounting for
the functions of technical artifacts since there are no plausible candidates for the goal-
directed systems of which both artifacts and their users would be components. The SD
theory currently looks like the best candidate the CR perspective has to offer in order to
defend function as a special sort of causal role, which is necessary if the concept is to
escape elimination. To gain wider support, I suggest that the theory’s central notion of
“type fi xation” be given a stronger naturalistic footing.
Acknowledgments
I am grateful to the attendees of the workshop on comparative philosophy of technical
artifacts and biological organisms at the Konrad Lorenz Institute and in particular to Ulrich
Krohs and Tim Lewens for their comments and suggestions.
Notes
- It may be questioned whether the distinction between answering a “how” question versus answering a “why”
question is the best way to characterize the different outlooks of the two approaches. Etiological theories are
perhaps better seen as also answering, by the attribution of function, a “how” question: how did it come about
that this item now exists having this particular feature? The main difference, then, is the adoption of a synchronic
(CR) versus a diachronic (PF) point of view as the one required for the purpose of explanation. This is not the
place to elaborate this point, however. - The way in which I now apply this twofold classifi cation to existing theories may not be as universally
accepted as is the primary distinction. - The case of nonliving goal-directed systems is complex; societies and cultures do not live but are commonly
considered to be able to survive and reproduce. McLaughlin (2001) seems to be of the opinion that the functions
that biologists and social scientists assign are just these contributions to the survival and reproduction of organ-
isms and societies. Wouters (2003), however, sees a difference between goal-contribution functions and causal-
role functions in biological practice. - Throughout this chapter, small letters refer to tokens and capital letters to types in all cases where the distinc-
tion is relevant. - I am ignoring here any reason why people would want to distinguish types from kinds.
- A second difference is that Davies thinks that a sixth clause is necessary in order to have normative statements
follow logically from the defi nition. Further on in this section (7.3) and in section 7.4, I discuss that this is not
the way that etiological theories seek to recover the normativity of functions. As the extra clause does not affect
both his argument and my rejection of it, I have left it out. - Actually the leading PF and SE theorists seem to be confused concerning the importance of the distinction
between functional and (roughly) morphological types or kinds. Millikan (1984: 17) says: “It is the “proper
function” of a thing that puts it in a biological category,” which, I think, is simply false for most, if not all, bio-