Resurrecting Dretskean information 23
The “restrictiveness critique”, as brought forward by authors such as Ruth
Millikan (2001; 2004) and Skyrms (2010) maintains, as Dretske later acknowledged, that setting admission criteria to the realm of informational relations as strict
as Dretske originally did might leave us with an account of information that is quite
removed from the ways in which organisms successfully perceive and act under less
tightly bounded but more natural conditions. Reliance on relations between source
and signal that are not nomologically governed but more probabilistic and local
in kind is a common practice in nature. Some, and arguably most, conditions in
nature will hold only within locally or temporally circumscribed domains, possibly
with no method available to the organism for determining the boundaries of these
domains, let alone the degree of reliability of some putatively informative signal.
If, on the one hand, one requires a strict, nomologically rooted covariance between
the presence of r and s being F for there to be information at all, and if, on the other
hand, it shall be acknowledged that natural organisms live under less-than-ideal,
variable and often uncooperative conditions, the probabilities considered on either
side, although related, have different reference points. Information as a probabilistic
affair, in the manner introduced in IN-1 and discussed by the mathematical theorists,
concerns only the question of how many binary decisions or “bits” are required to
reduce a given number of antecedent possibilities to one determinate state at the
source, and how much redundancy will have to be added at the source in order to
compensate for noise-induced loss of that information during transmission, so as to
avoid or minimise equivocation on the receiving side.
In contrast, the probabilities introduced by Dretske in his definition of informational content concern the question of whether a signal transmits information on a
certain world affair, namely that s is indeed F. The question here is whether some
affair r is related to what happens at s with sufficient reliability and regularity:
The conditional probabilities used to compute noise, equivocation, and
amount of transmitted information (and therefore the conditional probabilities defining the informational content of the signal) are all determined by the
lawful relations that exist between source and signal.
(Dretske 1981, 77)
That conditions at the source lawfully affect the signalling relation is the one
core presupposition of the mathematical theory of communication – which is
not further explained or elucidated in that theory. This presupposition is one key
reason for Dretske to actually mobilise that theory, as he most clearly states in
(Dretske 1983, 56): “For what this theory [= communication theory] tells us is
that the amount of information at r about s is a function of the degree of lawful
(nomic) dependence between conditions at these two points.” Still, that dependence between conditions at s and r tells us something only about the measure of
probabilities in a signalling relation.
The difference in focus here can be illustrated as follows: there may well be an
r-token that signals “conditions at s unchanged”. This would be an informationally rather bland affair on the mathematical theory, as the probability at t 0 of s
The “restrictiveness critique”, as brought forward by authors such as Ruth
Millikan (2001; 2004) and Skyrms (2010) maintains, as Dretske later acknowledged, that setting admission criteria to the realm of informational relations as strict
as Dretske originally did might leave us with an account of information that is quite
removed from the ways in which organisms successfully perceive and act under less
tightly bounded but more natural conditions. Reliance on relations between source
and signal that are not nomologically governed but more probabilistic and local
in kind is a common practice in nature. Some, and arguably most, conditions in
nature will hold only within locally or temporally circumscribed domains, possibly
with no method available to the organism for determining the boundaries of these
domains, let alone the degree of reliability of some putatively informative signal.
If, on the one hand, one requires a strict, nomologically rooted covariance between
the presence of r and s being F for there to be information at all, and if, on the other
hand, it shall be acknowledged that natural organisms live under less-than-ideal,
variable and often uncooperative conditions, the probabilities considered on either
side, although related, have different reference points. Information as a probabilistic
affair, in the manner introduced in IN-1 and discussed by the mathematical theorists,
concerns only the question of how many binary decisions or “bits” are required to
reduce a given number of antecedent possibilities to one determinate state at the
source, and how much redundancy will have to be added at the source in order to
compensate for noise-induced loss of that information during transmission, so as to
avoid or minimise equivocation on the receiving side.
In contrast, the probabilities introduced by Dretske in his definition of informational content concern the question of whether a signal transmits information on a
certain world affair, namely that s is indeed F. The question here is whether some
affair r is related to what happens at s with sufficient reliability and regularity:
The conditional probabilities used to compute noise, equivocation, and
amount of transmitted information (and therefore the conditional probabilities defining the informational content of the signal) are all determined by the
lawful relations that exist between source and signal.
(Dretske 1981, 77)
That conditions at the source lawfully affect the signalling relation is the one
core presupposition of the mathematical theory of communication – which is
not further explained or elucidated in that theory. This presupposition is one key
reason for Dretske to actually mobilise that theory, as he most clearly states in
(Dretske 1983, 56): “For what this theory [= communication theory] tells us is
that the amount of information at r about s is a function of the degree of lawful
(nomic) dependence between conditions at these two points.” Still, that dependence between conditions at s and r tells us something only about the measure of
probabilities in a signalling relation.
The difference in focus here can be illustrated as follows: there may well be an
r-token that signals “conditions at s unchanged”. This would be an informationally rather bland affair on the mathematical theory, as the probability at t 0 of s