74 Informational environments
up the things he encounters into practical kinds that are suitable to the fulfilment
of his purposes (see NI-4b). Residual ambiguities might assume some importance
if and when conditions in the environment become unstable or change in ways
that the receiver cannot easily anticipate and accommodate into the practically
determined domains of the signals that he relies on. In such situations, the conditions of stability (NI-4c) and unequivocality (NI-4d) of natural information have
been violated and call for some kind of adaptation or, less biologistically speaking, readjustment.
Resurrection at last
If I have to summarise the discussion in the preceding chapters in two brief points
that pick up on the key authors on which I built my account, it will be these: first,
despite acknowledging the irreducibility of situations of underspecification for a
perceiving organism, both Dretske and, albeit less unequivocally, Gibson, establish natural information as a specification relation. Second, despite his reference
to knowledge, the Dretskean notion of information from which my inquiry commenced partly parallels Gibson’s deflationary spirit with respect to knowledge
and mental representation – which is ultimately the more coherent view.
It is these two parallels that I see as deserving to be defended or, as the title
of Chapter 2 had it, “resurrected”. At the very beginning of Knowledge and the
Flow of Information, Dretske says: “In the beginning there was information. The
word came later” (1981, vii). He thus anticipates that the historically and systematically prior thing for organisms is to “selectively exploit” information in
the environment to adaptive ends, that is the guidance of activities. Dretske correctly identifies Gibsonian information as “higher-order invariants in a temporal
series of signals”, but then continues to refer to sensory representation (Dretske
1981, 145) – a notion that Gibson would not have subscribed to. In the endnote
to that passage, he also mentions that a distinction has been made between Gibsonian information for perception and information under the mathematical paradigm, including by Gibson himself (Dretske 1981, 255, n10). He then goes on
to quote a passage from The Senses Considered as Perceptual Systems in which
Gibson says that “a property of the stimulus is univocally related to a property
of the object by virtue of physical laws” (Gibson 1966, 187). It seems, however,
that Dretske quotes this passage in order to actually relativise the very distinction
between Gibsonian and mathematical information – which should not be surprising because Dretske himself apparently believes his account of information to be
in line with both Gibson’s and the mathematical paradigm.
In this attempt to align himself with two concepts of information that are
demonstrably disparate, one can detect, first, the tension between Dretske’s deflationary view of information and his epistemological aims and, second, the subtle
but significant difference between Dretske’s and Gibson’s accounts: on Dretske’s
view, informational relations are invariant relations between world affairs that
could, in principle, be detected by a receiver, whatever his constitution and abilities might be. These relations can be analysed in isolation from each other, and
up the things he encounters into practical kinds that are suitable to the fulfilment
of his purposes (see NI-4b). Residual ambiguities might assume some importance
if and when conditions in the environment become unstable or change in ways
that the receiver cannot easily anticipate and accommodate into the practically
determined domains of the signals that he relies on. In such situations, the conditions of stability (NI-4c) and unequivocality (NI-4d) of natural information have
been violated and call for some kind of adaptation or, less biologistically speaking, readjustment.
Resurrection at last
If I have to summarise the discussion in the preceding chapters in two brief points
that pick up on the key authors on which I built my account, it will be these: first,
despite acknowledging the irreducibility of situations of underspecification for a
perceiving organism, both Dretske and, albeit less unequivocally, Gibson, establish natural information as a specification relation. Second, despite his reference
to knowledge, the Dretskean notion of information from which my inquiry commenced partly parallels Gibson’s deflationary spirit with respect to knowledge
and mental representation – which is ultimately the more coherent view.
It is these two parallels that I see as deserving to be defended or, as the title
of Chapter 2 had it, “resurrected”. At the very beginning of Knowledge and the
Flow of Information, Dretske says: “In the beginning there was information. The
word came later” (1981, vii). He thus anticipates that the historically and systematically prior thing for organisms is to “selectively exploit” information in
the environment to adaptive ends, that is the guidance of activities. Dretske correctly identifies Gibsonian information as “higher-order invariants in a temporal
series of signals”, but then continues to refer to sensory representation (Dretske
1981, 145) – a notion that Gibson would not have subscribed to. In the endnote
to that passage, he also mentions that a distinction has been made between Gibsonian information for perception and information under the mathematical paradigm, including by Gibson himself (Dretske 1981, 255, n10). He then goes on
to quote a passage from The Senses Considered as Perceptual Systems in which
Gibson says that “a property of the stimulus is univocally related to a property
of the object by virtue of physical laws” (Gibson 1966, 187). It seems, however,
that Dretske quotes this passage in order to actually relativise the very distinction
between Gibsonian and mathematical information – which should not be surprising because Dretske himself apparently believes his account of information to be
in line with both Gibson’s and the mathematical paradigm.
In this attempt to align himself with two concepts of information that are
demonstrably disparate, one can detect, first, the tension between Dretske’s deflationary view of information and his epistemological aims and, second, the subtle
but significant difference between Dretske’s and Gibson’s accounts: on Dretske’s
view, informational relations are invariant relations between world affairs that
could, in principle, be detected by a receiver, whatever his constitution and abilities might be. These relations can be analysed in isolation from each other, and