The domains of natural information 71
they do is to track a correlate of that gradient that also holds with law-like regularity, namely the direction of the Earth’s magnetic field. The lines of the Earth’s
magnetic field are aligned, as for any other magnetic dipole, as field lines, which
can be described by any smaller magnetic particle being placed in the field. These
lines emerge from the magnetic south pole, bend outwards towards the equator
and return to Earth at the north pole. As a result, the lines will point upward relative to the surface of the Earth in the south and downward in the north. The magnetic particles enclosed in the bacteria’s cell membrane align with the field line.
In this sense, magnetic orientation is a lawful proximal correlate of the vertical
gradient of oxygen concentration. This is where the bacteria will invariably move.
However, this direction is inverted from one hemisphere to the other, so “northward” will be “upward” in the south and “downward” in the north. Each specimen
of the respective subspecies of bacteria must be placed on the right hemisphere in
relation to its northward- vs. southward-oriented constitution in order to maintain
the mapping between field direction and the gradient of oxygen concentration on
which it depends. With no supplementary sensory facilities to warn it, the bacterium cannot identify the domain of the downward signal, however lawful the
correlation.
The first and the second of the ways of carving out domains of natural information are the ones Millikan has in mind, whereas the third seems to be implicitly
subsumed under the former two or, more probably, under their preconditions.
4
The first two ways even fall into one if one conceives of species as historical individuals (as in Millikan 2000) rather than extensionally defined populations. The
choice of species concept makes a difference to the interpretation of the nature of
the informational relations involved. Quite obviously, if a species is to be treated
as an individual, and if its being an individual is not considered a theoretical
fiction in the service of explaining biological phenomena, but a real entity held
together by a well-defined set of genealogical relations of descent and proximity,
the observer’s task, when attempting to track the domain of the species, lies in
tracking the domains of the individuals belonging to that species. The domain of
the species specifies the domains of its individuals. Misidentifying some r as a
signal of a q-bird qua q-bird rather than merely of a such-shaped (p- or q-) bird
leaving such-shaped traces, amounts to a failure at correctly recording natural
information related to that individual as a member of the species, although it may
still be correct of that individual as an individual (that bird in my garden last
night, or that bird gracefully circling overhead in the Andean skies).
If, in contrast, species are extensionally defined, the definition of their domains
is observer-dependent in a very specific way: Signals {r
i
,. . ., r
n
} emanating from
different individuals {s
i
,. . ., s
n
}, where the mapping of r
i
onto s
i
, r
j
onto s
j
and so
forth is supposed to be unequivocal on the source side, have to be subsumed by R
under a common type of signals r that correlates with a type of individuals s. That
mapping is not, or not fully, determined by the natural information carried by the
individuals {s
i
,. . ., s
n
}. The signals might be marks of type-identity in properties
for the individuals, but there will also be signals of properties in which these individuals differ. Hence, R is endowed with the task of identifying the similarities
they do is to track a correlate of that gradient that also holds with law-like regularity, namely the direction of the Earth’s magnetic field. The lines of the Earth’s
magnetic field are aligned, as for any other magnetic dipole, as field lines, which
can be described by any smaller magnetic particle being placed in the field. These
lines emerge from the magnetic south pole, bend outwards towards the equator
and return to Earth at the north pole. As a result, the lines will point upward relative to the surface of the Earth in the south and downward in the north. The magnetic particles enclosed in the bacteria’s cell membrane align with the field line.
In this sense, magnetic orientation is a lawful proximal correlate of the vertical
gradient of oxygen concentration. This is where the bacteria will invariably move.
However, this direction is inverted from one hemisphere to the other, so “northward” will be “upward” in the south and “downward” in the north. Each specimen
of the respective subspecies of bacteria must be placed on the right hemisphere in
relation to its northward- vs. southward-oriented constitution in order to maintain
the mapping between field direction and the gradient of oxygen concentration on
which it depends. With no supplementary sensory facilities to warn it, the bacterium cannot identify the domain of the downward signal, however lawful the
correlation.
The first and the second of the ways of carving out domains of natural information are the ones Millikan has in mind, whereas the third seems to be implicitly
subsumed under the former two or, more probably, under their preconditions.
4
The first two ways even fall into one if one conceives of species as historical individuals (as in Millikan 2000) rather than extensionally defined populations. The
choice of species concept makes a difference to the interpretation of the nature of
the informational relations involved. Quite obviously, if a species is to be treated
as an individual, and if its being an individual is not considered a theoretical
fiction in the service of explaining biological phenomena, but a real entity held
together by a well-defined set of genealogical relations of descent and proximity,
the observer’s task, when attempting to track the domain of the species, lies in
tracking the domains of the individuals belonging to that species. The domain of
the species specifies the domains of its individuals. Misidentifying some r as a
signal of a q-bird qua q-bird rather than merely of a such-shaped (p- or q-) bird
leaving such-shaped traces, amounts to a failure at correctly recording natural
information related to that individual as a member of the species, although it may
still be correct of that individual as an individual (that bird in my garden last
night, or that bird gracefully circling overhead in the Andean skies).
If, in contrast, species are extensionally defined, the definition of their domains
is observer-dependent in a very specific way: Signals {r
i
,. . ., r
n
} emanating from
different individuals {s
i
,. . ., s
n
}, where the mapping of r
i
onto s
i
, r
j
onto s
j
and so
forth is supposed to be unequivocal on the source side, have to be subsumed by R
under a common type of signals r that correlates with a type of individuals s. That
mapping is not, or not fully, determined by the natural information carried by the
individuals {s
i
,. . ., s
n
}. The signals might be marks of type-identity in properties
for the individuals, but there will also be signals of properties in which these individuals differ. Hence, R is endowed with the task of identifying the similarities