72 Informational environments
and differences relevant to his purposes without the benefit of an additional property that would help him to type {s
i
,. . ., s
n
} together in an unequivocal way and
thus warrant R’s acts of identification. Any misidentification of an individual will
be a misidentification only in relation to the practical purposes of R and only if
these practical purposes are adversely affected by that misidentification. The practical purposes of the competitors of California Condors will not be affected when
an Andean Condor accidentally appears in Californian skies, given their phenotypical and behavioural similarities, whereas the practical purposes of a California
Condor looking for a mate and the practical purposes of the California Condor
Recovery Programme are likely to be affected.
If we restrict our view to practical purposes, the latter observations will equally
hold for either species concept. Millikan (2004, Chapter 18) applies the term
“practical kinds” when referring to those situations in which a difference between
species or other types of things or even between one and the same vs. different
individuals is practically irrelevant to R since the effects on R’s purposes and
behaviours of treating them as identical are negligible. Thus, it will also be irrelevant in practical terms whether the species involved is a historical individual or
an aggregate of individuals, as long as R carves them up in proper alignment with
his practical purposes. The California and the Andean Condor form one “practical kind” for the competing scavenger, but two different practical kinds for other
Condors, and two natural kinds for the field biologist and the conservationist.
Even the distinction between repeated encounters with one and the same individual and with other members of the same species may not be important under
all circumstances (see Millikan 2004, 219f ).
However, the theoretical implications of the choice of species concepts may be
notable, and some higher-level practical considerations may flow from it. Choosing species, understood as historical individuals, as the paradigm of domainbound natural information might serve to mix up the local and the probabilistic
character of natural information C . Assuming species to be individuals implicitly
suggests that the character of natural information C amounts to being probabilistic
qua locally bound (see p. 65 in this chapter, see also NI-2b and 2c). This is no
minor point, and it takes some explaining.
The extension of a domain of a species as a historical individual would be
marked by its incipience at t 0 , its population dynamics, its patterns of migration
and its extinction at t n . Prima facie, the sources of all possible equivocation in
the definition of a domain of a species would hence lie with, first, determining its
time of incipience, second, the lines of descent that flow from it, third, its population distribution and fourth, its time of extinction (which might be the easiest to
identify if it has happened already). These are all things that R could be mistaken
about, but not the species itself, as it were. The rule that determines when a species is a species is supposed to be unequivocal and observer independent, and
hence the membership conditions of individuals pertaining to that species are supposed to be unequivocal and observer independent, too.
On this view, it should be a straightforward matter to account for the possibility
that one may encounter, on the one hand, individuals that belong to the species in
and differences relevant to his purposes without the benefit of an additional property that would help him to type {s
i
,. . ., s
n
} together in an unequivocal way and
thus warrant R’s acts of identification. Any misidentification of an individual will
be a misidentification only in relation to the practical purposes of R and only if
these practical purposes are adversely affected by that misidentification. The practical purposes of the competitors of California Condors will not be affected when
an Andean Condor accidentally appears in Californian skies, given their phenotypical and behavioural similarities, whereas the practical purposes of a California
Condor looking for a mate and the practical purposes of the California Condor
Recovery Programme are likely to be affected.
If we restrict our view to practical purposes, the latter observations will equally
hold for either species concept. Millikan (2004, Chapter 18) applies the term
“practical kinds” when referring to those situations in which a difference between
species or other types of things or even between one and the same vs. different
individuals is practically irrelevant to R since the effects on R’s purposes and
behaviours of treating them as identical are negligible. Thus, it will also be irrelevant in practical terms whether the species involved is a historical individual or
an aggregate of individuals, as long as R carves them up in proper alignment with
his practical purposes. The California and the Andean Condor form one “practical kind” for the competing scavenger, but two different practical kinds for other
Condors, and two natural kinds for the field biologist and the conservationist.
Even the distinction between repeated encounters with one and the same individual and with other members of the same species may not be important under
all circumstances (see Millikan 2004, 219f ).
However, the theoretical implications of the choice of species concepts may be
notable, and some higher-level practical considerations may flow from it. Choosing species, understood as historical individuals, as the paradigm of domainbound natural information might serve to mix up the local and the probabilistic
character of natural information C . Assuming species to be individuals implicitly
suggests that the character of natural information C amounts to being probabilistic
qua locally bound (see p. 65 in this chapter, see also NI-2b and 2c). This is no
minor point, and it takes some explaining.
The extension of a domain of a species as a historical individual would be
marked by its incipience at t 0 , its population dynamics, its patterns of migration
and its extinction at t n . Prima facie, the sources of all possible equivocation in
the definition of a domain of a species would hence lie with, first, determining its
time of incipience, second, the lines of descent that flow from it, third, its population distribution and fourth, its time of extinction (which might be the easiest to
identify if it has happened already). These are all things that R could be mistaken
about, but not the species itself, as it were. The rule that determines when a species is a species is supposed to be unequivocal and observer independent, and
hence the membership conditions of individuals pertaining to that species are supposed to be unequivocal and observer independent, too.
On this view, it should be a straightforward matter to account for the possibility
that one may encounter, on the one hand, individuals that belong to the species in