The domains of natural information 65
be sufficient. It is in this light that NI-4d and the condition of the “determinate
degree” of discernibility in NI-4 should be understood. There is no awareness
required from the receiver of the presence and kind of the informational relations
that he relies on. It will be the privilege of the followers of the complexity- oriented
strategies to develop some degree of awareness, perhaps even a conscious one, of
the informational relations used, but many organisms can do without it under natural conditions. These issues of information-related strategies will be discussed in
more detail in the next chapter.
Informational domains
The final piece to an account of natural information that characterises it as both
specifying and including a local and probabilistic element will be to spell out
the relation between the local and the probabilistic element in more detail. When
moving from information L into her softer but supposedly often more relevant
information C , Millikan appears to target two related ways for that information to
become ‘soft’, which I outlined in the previous section as NI-2b and NI-2c: first,
the regularity I of the signalling relation might turn out to be statistical rather than
nomological in kind; second, there might be regions in space and time – christened “domains” by Millikan – to which that regularity is restricted. The notion
of informational domains in particular will become important in an inquiry into
informational environments in Chapter 6.
(NI-5) The domain-boundedness condition: informational relations may be
bound to certain domains D of some set of signals {r 1 ,. . ., r n } or types of
signals {r 1 ,. . ., r n }, where
(a) D is a spatio-temporally circumscribed region in which type r signals are present and detectable that correlate with some individual s
or type s with a given degree of probability p = n, where 0 < n ≤ 1,
as distinguished from other regions where p < n or p = 0, providing
for local maxima and minima;
(b) the extension of D is demarcated by empirical regularities I e , as
they apply to individuals, species and other spatio-temporally
restricted entities at s;
(c) type r signals may pertain to different domains {D 1 ,. . ., D m } at
once and in non-contradictory fashion, depending on the regularities described in NI-5b;
(d) there is a possibility for a receiver R to keep track of one or various types of signals within D, under conditions described in NI-4,
where R additionally may but need not be able to identify the extension of D.
More often than not, as Millikan’s argument in (2004) goes, the only regularities
at hand for a consumer of natural information are of a statistical kind. For these
be sufficient. It is in this light that NI-4d and the condition of the “determinate
degree” of discernibility in NI-4 should be understood. There is no awareness
required from the receiver of the presence and kind of the informational relations
that he relies on. It will be the privilege of the followers of the complexity- oriented
strategies to develop some degree of awareness, perhaps even a conscious one, of
the informational relations used, but many organisms can do without it under natural conditions. These issues of information-related strategies will be discussed in
more detail in the next chapter.
Informational domains
The final piece to an account of natural information that characterises it as both
specifying and including a local and probabilistic element will be to spell out
the relation between the local and the probabilistic element in more detail. When
moving from information L into her softer but supposedly often more relevant
information C , Millikan appears to target two related ways for that information to
become ‘soft’, which I outlined in the previous section as NI-2b and NI-2c: first,
the regularity I of the signalling relation might turn out to be statistical rather than
nomological in kind; second, there might be regions in space and time – christened “domains” by Millikan – to which that regularity is restricted. The notion
of informational domains in particular will become important in an inquiry into
informational environments in Chapter 6.
(NI-5) The domain-boundedness condition: informational relations may be
bound to certain domains D of some set of signals {r 1 ,. . ., r n } or types of
signals {r 1 ,. . ., r n }, where
(a) D is a spatio-temporally circumscribed region in which type r signals are present and detectable that correlate with some individual s
or type s with a given degree of probability p = n, where 0 < n ≤ 1,
as distinguished from other regions where p < n or p = 0, providing
for local maxima and minima;
(b) the extension of D is demarcated by empirical regularities I e , as
they apply to individuals, species and other spatio-temporally
restricted entities at s;
(c) type r signals may pertain to different domains {D 1 ,. . ., D m } at
once and in non-contradictory fashion, depending on the regularities described in NI-5b;
(d) there is a possibility for a receiver R to keep track of one or various types of signals within D, under conditions described in NI-4,
where R additionally may but need not be able to identify the extension of D.
More often than not, as Millikan’s argument in (2004) goes, the only regularities
at hand for a consumer of natural information are of a statistical kind. For these