62 Informational environments
If he has no such means of detection, he must remain confined to living within the
boundaries of the region in which these conditions hold. I will discuss this point,
introduced by Millikan as the “domains” of natural information, in more detail in
the next section.
While NI-2 considers the question of specificity vs. probability of relations
within a reference class as being dependent on the constitution and abilities of a
receiver, and hence refers to NI-3, the receiver condition stated in this clause also
caters for the intrinsic nature of natural information to a perceiving organism:
informational relations may not always and under all circumstances involve a
straightforward, organism-independent matching between a received signal and
the values of the physical variables at the source. As the discussion of perceptual
illusions in the previous chapter should have shown, a signal may be found to
never match the value of a variable at the source and still be useful to a perceiving
organism. Getting the values of some physical variables at the source wrong may,
under certain circumstances, actually be in the service of getting something more
important about that source right, namely what can or should be done with respect
to conditions at that source. If an – external – observer of the informational relation
limits his perspective to conditions at the source and types of environment-bound
signals related to that source, the relevant pattern will not be properly detectable
for him to begin with. Only to an external observer who does thus not account for
the intrinsic, organism-involving mappings that are part of the informational relations, will there be an informational bias involved in perception. To the perceiving
organism, the reference class of the signals may well comprise of tokens mapping
distal conditions in indirect but highly reliable ways.
The correspondence to the physical variables does, however, not typically
become explicit to the perceiving organism in the perceptual process. Otherwise,
the perceptual tokens would have to carry a mark of how they map onto the physical variable under a defined range of conditions – something like an indicator of
the complex affair “the object looks n degrees brighter now that the total amount
of visible light decreased by m lm”. Although there is no such indicator, the informational mappings will remain intact to the organism. If, in this kind of case, a
bias in perceived qualities were to result in two different values for some physical
variable being perceived equally, such as equal perceived brightness for variantly
luminous objects, an ecological argument might be applicable (or even necessary)
to account for such situations: the equivocation either does not matter, in that it
does not, or not significantly, interfere with the attainment of the organism’s goals,
or the other type of object or lighting condition does, as a matter of empirical fact,
rarely or never occur in R’s environment, where that matter of fact should be due
to one or a number of natural regularities in their own right.
Hence, there is no “too bright” in the perception of some object if the appropriate perception of the conditions at the source – the affair to which the perceiving
organism is supposed to respond and hence the affordance that it presents to him –
depends on perceiving the object in question as being brighter than the measured
luminance values would suggest. Although the organism might thus appear to
be misinformed about the luminance values at the source, he will be correctly
If he has no such means of detection, he must remain confined to living within the
boundaries of the region in which these conditions hold. I will discuss this point,
introduced by Millikan as the “domains” of natural information, in more detail in
the next section.
While NI-2 considers the question of specificity vs. probability of relations
within a reference class as being dependent on the constitution and abilities of a
receiver, and hence refers to NI-3, the receiver condition stated in this clause also
caters for the intrinsic nature of natural information to a perceiving organism:
informational relations may not always and under all circumstances involve a
straightforward, organism-independent matching between a received signal and
the values of the physical variables at the source. As the discussion of perceptual
illusions in the previous chapter should have shown, a signal may be found to
never match the value of a variable at the source and still be useful to a perceiving
organism. Getting the values of some physical variables at the source wrong may,
under certain circumstances, actually be in the service of getting something more
important about that source right, namely what can or should be done with respect
to conditions at that source. If an – external – observer of the informational relation
limits his perspective to conditions at the source and types of environment-bound
signals related to that source, the relevant pattern will not be properly detectable
for him to begin with. Only to an external observer who does thus not account for
the intrinsic, organism-involving mappings that are part of the informational relations, will there be an informational bias involved in perception. To the perceiving
organism, the reference class of the signals may well comprise of tokens mapping
distal conditions in indirect but highly reliable ways.
The correspondence to the physical variables does, however, not typically
become explicit to the perceiving organism in the perceptual process. Otherwise,
the perceptual tokens would have to carry a mark of how they map onto the physical variable under a defined range of conditions – something like an indicator of
the complex affair “the object looks n degrees brighter now that the total amount
of visible light decreased by m lm”. Although there is no such indicator, the informational mappings will remain intact to the organism. If, in this kind of case, a
bias in perceived qualities were to result in two different values for some physical
variable being perceived equally, such as equal perceived brightness for variantly
luminous objects, an ecological argument might be applicable (or even necessary)
to account for such situations: the equivocation either does not matter, in that it
does not, or not significantly, interfere with the attainment of the organism’s goals,
or the other type of object or lighting condition does, as a matter of empirical fact,
rarely or never occur in R’s environment, where that matter of fact should be due
to one or a number of natural regularities in their own right.
Hence, there is no “too bright” in the perception of some object if the appropriate perception of the conditions at the source – the affair to which the perceiving
organism is supposed to respond and hence the affordance that it presents to him –
depends on perceiving the object in question as being brighter than the measured
luminance values would suggest. Although the organism might thus appear to
be misinformed about the luminance values at the source, he will be correctly