The domains of natural information 63
informed about the presence of the object that he will need to recognise in order
to persist. By virtue of its regularity, the seeming mismatch will constitute an
informational relation in its own right. Again by virtue of the regularity of these
relations, the inclusion of organism-involving relations is perfectly compatible
with the notion of information being an objective affair. It may be true that the
informational relation in question would not exist in absence of the receiver, but
only in the same trivial sense in which any informational relation depends on the
existence of certain world affairs.
The regularities involved in this complex mapping relation will have been
established in a history of interaction, on individual or population levels, with
those states of affairs. The seeming departure from the values of the physical variables involved in such situations is not at all capricious but depends on processes
in which that deviation acquired an adaptive function that consists precisely in
enabling the organism to get a grasp of the conditions at the source that are relevant to him and his kind. The probabilistic element in natural information, on
this level, is not one of random distributions in which, for example, an object
appears n degrees brighter than measured at t 0 , exactly as bright as measured at t 1 ,
n degrees darker at t 2 and so forth, but in the frequency distribution of perceptual
situations under which the function of perceiving objects of that kind in a certain
way evolved – notwithstanding the possibility of false positives or false negatives
in perception, and allowing for some degree of variability in perceived qualities
over a series of perceptual events.
The preceding discussion should have served to show that it is the situation,
the constitution, the abilities and the purposes of the receiving organism that are
indispensable to the definition of the reference class for the information involved.
I will try to articulate the upshot of that discussion, and add more detail to clause
NI-3 in particular, by stating a set of conditions for the successful use of natural
information by its receivers:
(NI-4) The detectability condition: different signals {r 1 ,. . ., r n } must be discernible to some determinate degree for some receiver R, where
(a) the conditions at s to be detected by R are determined with respect to
the constitution and abilities of R – the condition of the intrinsic definition of the reference class for the mapping relations of {r 1 ,. . ., r n };
(b) the mapping relations between {r 1 ,. . ., r n } and the conditions at
the source make a difference to R in terms of the fulfilment of R’s
specific purposes – the condition of relevance;
(c) the conditions for information uptake are such that signal tokens
{r 1 ,. . ., r n } occurring in accordance with the respective conditions
at s remain detectable for R over time – the condition of stability;
(d) the frequency for R of recording r 1 where s is not F, or where some
other r n occurs instead of r 1 or where no r at all occurs remains
below a threshold that is critical to the fulfilment of R’s purposes –
the condition of unequivocality.
informed about the presence of the object that he will need to recognise in order
to persist. By virtue of its regularity, the seeming mismatch will constitute an
informational relation in its own right. Again by virtue of the regularity of these
relations, the inclusion of organism-involving relations is perfectly compatible
with the notion of information being an objective affair. It may be true that the
informational relation in question would not exist in absence of the receiver, but
only in the same trivial sense in which any informational relation depends on the
existence of certain world affairs.
The regularities involved in this complex mapping relation will have been
established in a history of interaction, on individual or population levels, with
those states of affairs. The seeming departure from the values of the physical variables involved in such situations is not at all capricious but depends on processes
in which that deviation acquired an adaptive function that consists precisely in
enabling the organism to get a grasp of the conditions at the source that are relevant to him and his kind. The probabilistic element in natural information, on
this level, is not one of random distributions in which, for example, an object
appears n degrees brighter than measured at t 0 , exactly as bright as measured at t 1 ,
n degrees darker at t 2 and so forth, but in the frequency distribution of perceptual
situations under which the function of perceiving objects of that kind in a certain
way evolved – notwithstanding the possibility of false positives or false negatives
in perception, and allowing for some degree of variability in perceived qualities
over a series of perceptual events.
The preceding discussion should have served to show that it is the situation,
the constitution, the abilities and the purposes of the receiving organism that are
indispensable to the definition of the reference class for the information involved.
I will try to articulate the upshot of that discussion, and add more detail to clause
NI-3 in particular, by stating a set of conditions for the successful use of natural
information by its receivers:
(NI-4) The detectability condition: different signals {r 1 ,. . ., r n } must be discernible to some determinate degree for some receiver R, where
(a) the conditions at s to be detected by R are determined with respect to
the constitution and abilities of R – the condition of the intrinsic definition of the reference class for the mapping relations of {r 1 ,. . ., r n };
(b) the mapping relations between {r 1 ,. . ., r n } and the conditions at
the source make a difference to R in terms of the fulfilment of R’s
specific purposes – the condition of relevance;
(c) the conditions for information uptake are such that signal tokens
{r 1 ,. . ., r n } occurring in accordance with the respective conditions
at s remain detectable for R over time – the condition of stability;
(d) the frequency for R of recording r 1 where s is not F, or where some
other r n occurs instead of r 1 or where no r at all occurs remains
below a threshold that is critical to the fulfilment of R’s purposes –
the condition of unequivocality.