Ants and robots, parlour games and steam drills 9
Both in philosophy and in the cognitive sciences, it is a relatively recent but
now well-established proposition to seriously consider a constitutive role of an
organism’s environment to his cognitive abilities and their accomplishment. So,
as to the first question, I will present a take on the role of the environment in cognition that shares the notion of such a constitutive role but claims distinctiveness
in how it conceives of the information encountered in the environment. More precisely, I will seek to rehabilitate the notion of natural information introduced by
Fred Dretske (1981) against the altogether reasonable criticisms brought forward
by advocates of 4E cognition who, following Ruth Millikan (2001), argue for a
local and probabilistic rather than nomologically governed character of natural
information. In fact, these criticisms will be helpful in carving out the main point
to be defended here. In the service of that defence, I will align Dretskean information with a reading of the notion of information for perception in the ecological
approach to perception (Gibson 1979).
As to the second question, both sides involved in the aforementioned debates
either confine their considerations of the role of the environment in cognition to
the natural environment or they have a restricted view of the role and importance
of artefacts in human environments. However, throughout all of human history
and much of prehistory, artefacts have been invented and used that remain deeply
involved in how human beings perceive and act in their environments. One may
think of carvings, figurines, pictures and inscriptions back then or books, broadcast media, phones and computers now. Basically, one may think of all things
manufactured that do not primarily aim at altering physical conditions in the environment but at providing information in some way. Their function is to augment,
supplement, substitute or, in some cases, even first enable, human cognitive functions. It will be helpful to conceptually distinguish between these types of contributions by artefacts, which I will henceforth refer to as “cognitive artefacts”. If
one considers language an artefact, it will be the paradigm of strongly constitutive
cognitive artefacts.
My main focus in this context, however, will be on a subset of cognitive artefacts whose very existence owes to Turing’s work, namely digital computers.
Among these, I will focus on computer-based artefacts that are directly concerned
with their users’ perception of, and interaction with, their environments. I will
hence seek to decouple my account of those artefacts from the often poorly specified criterion of human-likeness that holds sway over much of what followed
upon classical AI – including parts of 4E cognition. I will try to demonstrate that,
with respect to human cognition, the most notable effect of digital computers and
their manifold descendants is not that they might be able, or become able, to think
or to provide a simile thereof, but that they alter the way we perceive, think and
act. What they do is to modify our informational environments. Whereas all technologies have the effect of materially changing human environments and perhaps
also of affecting our perception of and interactions with these environments, those
technologies which directly operate on information have a capacity of equally
directly modifying the patterns of available informational relations by which we
steer. Despite the involvement of artefacts, and despite being technologically
Both in philosophy and in the cognitive sciences, it is a relatively recent but
now well-established proposition to seriously consider a constitutive role of an
organism’s environment to his cognitive abilities and their accomplishment. So,
as to the first question, I will present a take on the role of the environment in cognition that shares the notion of such a constitutive role but claims distinctiveness
in how it conceives of the information encountered in the environment. More precisely, I will seek to rehabilitate the notion of natural information introduced by
Fred Dretske (1981) against the altogether reasonable criticisms brought forward
by advocates of 4E cognition who, following Ruth Millikan (2001), argue for a
local and probabilistic rather than nomologically governed character of natural
information. In fact, these criticisms will be helpful in carving out the main point
to be defended here. In the service of that defence, I will align Dretskean information with a reading of the notion of information for perception in the ecological
approach to perception (Gibson 1979).
As to the second question, both sides involved in the aforementioned debates
either confine their considerations of the role of the environment in cognition to
the natural environment or they have a restricted view of the role and importance
of artefacts in human environments. However, throughout all of human history
and much of prehistory, artefacts have been invented and used that remain deeply
involved in how human beings perceive and act in their environments. One may
think of carvings, figurines, pictures and inscriptions back then or books, broadcast media, phones and computers now. Basically, one may think of all things
manufactured that do not primarily aim at altering physical conditions in the environment but at providing information in some way. Their function is to augment,
supplement, substitute or, in some cases, even first enable, human cognitive functions. It will be helpful to conceptually distinguish between these types of contributions by artefacts, which I will henceforth refer to as “cognitive artefacts”. If
one considers language an artefact, it will be the paradigm of strongly constitutive
cognitive artefacts.
My main focus in this context, however, will be on a subset of cognitive artefacts whose very existence owes to Turing’s work, namely digital computers.
Among these, I will focus on computer-based artefacts that are directly concerned
with their users’ perception of, and interaction with, their environments. I will
hence seek to decouple my account of those artefacts from the often poorly specified criterion of human-likeness that holds sway over much of what followed
upon classical AI – including parts of 4E cognition. I will try to demonstrate that,
with respect to human cognition, the most notable effect of digital computers and
their manifold descendants is not that they might be able, or become able, to think
or to provide a simile thereof, but that they alter the way we perceive, think and
act. What they do is to modify our informational environments. Whereas all technologies have the effect of materially changing human environments and perhaps
also of affecting our perception of and interactions with these environments, those
technologies which directly operate on information have a capacity of equally
directly modifying the patterns of available informational relations by which we
steer. Despite the involvement of artefacts, and despite being technologically