The extension of the extended mind 113
general). The “ancestors” and “generations”, rather than being organisms themselves, may also be component mechanisms, and they may also be reproducible
forms that can be iteratively used by some individual or a collective thereof, such
as artefacts or linguistic items. The individual tokens of some such reproducible
item have proper functions derived from the functions of the mechanisms that
produce them. It is the direct proper function of those mechanisms to generate
the kind of effects in question if these effects, qua individually bearing certain
relations to the environment with sufficient reliability, form part of the necessary
conditions for the reproduction of the mechanisms themselves and of the systems
that rely on their presence. In turn, the tokens of some reproducible item have
the derived proper function of being adapted to some concrete world affair if and
when one relevant condition for their reproduction lies in their co-occurrence with
that sort of world affair and, in a relevant subset of cases, their role in directing an
organism’s behaviours towards it.
Nothing in this account rules out the possibility of mechanisms that are only
partly based within an organism. In fact, Millikan (1993a, 179) herself considers
the insight that “the organismic process has no skin” fundamental to the science
of psychology.
5
Because all sorts of traits, cognitive and other, are characterised in
terms of their historically acquired functions, and because that history incorporates
whatever turns out to realise the function in question, the locus of the components
involved in realising it is systematically irrelevant, under a two-part proviso: the
organismic traits involved are to be coupled with persistent or reproducible features
of their environment with a degree of reliability sufficient for meeting the reproduction criterion, where that coupling occurs in accordance with a uniform explanation that accounts for the “normal mechanisms” of coupling. Hence, although the
content externalism defended in the aetiological account does not entail a vehicle
externalism, it is methodically impartial to the locus of the vehicles involved.
As to the concrete benefits of an aetiological perspective for extended functionalism, accounting for the historical nature of functions will, first, provide an
argument against the “coupling/constitution” objection brought forward by Fred
Adams and Kenneth Aizawa (2001): in contrast to their objection, instances of
accidental coupling between an organism and some entities in his environment do
not count as coupling proper. Nor will legitimate instances of coupling endow the
individual entities thus coupled with the characteristics of the entire coupled system. These entities play constitutive, historically established roles in the establishment or accomplishment of the functions of the entire system, where these roles
may widely diverge. I will discuss this point in more detail in the next section.
Second, and for related reasons, accounting for the historical nature of functions will provide an argument against the “fleeting vs. persistence” criticism
that has been levelled against extended cognition, first and foremost, by Rupert
(2010). The suspicion here is that environmental extensions, unlike internally
based cognitive capacities, are not persistently coupled with the organism. If some
such extension is accidentally detached from the cognising organism, it seems, the
cognitive system comes apart, only to be re-instantiated if and when that extension is successfully recovered. Clark’s reply to this charge was to emphasise the
general). The “ancestors” and “generations”, rather than being organisms themselves, may also be component mechanisms, and they may also be reproducible
forms that can be iteratively used by some individual or a collective thereof, such
as artefacts or linguistic items. The individual tokens of some such reproducible
item have proper functions derived from the functions of the mechanisms that
produce them. It is the direct proper function of those mechanisms to generate
the kind of effects in question if these effects, qua individually bearing certain
relations to the environment with sufficient reliability, form part of the necessary
conditions for the reproduction of the mechanisms themselves and of the systems
that rely on their presence. In turn, the tokens of some reproducible item have
the derived proper function of being adapted to some concrete world affair if and
when one relevant condition for their reproduction lies in their co-occurrence with
that sort of world affair and, in a relevant subset of cases, their role in directing an
organism’s behaviours towards it.
Nothing in this account rules out the possibility of mechanisms that are only
partly based within an organism. In fact, Millikan (1993a, 179) herself considers
the insight that “the organismic process has no skin” fundamental to the science
of psychology.
5
Because all sorts of traits, cognitive and other, are characterised in
terms of their historically acquired functions, and because that history incorporates
whatever turns out to realise the function in question, the locus of the components
involved in realising it is systematically irrelevant, under a two-part proviso: the
organismic traits involved are to be coupled with persistent or reproducible features
of their environment with a degree of reliability sufficient for meeting the reproduction criterion, where that coupling occurs in accordance with a uniform explanation that accounts for the “normal mechanisms” of coupling. Hence, although the
content externalism defended in the aetiological account does not entail a vehicle
externalism, it is methodically impartial to the locus of the vehicles involved.
As to the concrete benefits of an aetiological perspective for extended functionalism, accounting for the historical nature of functions will, first, provide an
argument against the “coupling/constitution” objection brought forward by Fred
Adams and Kenneth Aizawa (2001): in contrast to their objection, instances of
accidental coupling between an organism and some entities in his environment do
not count as coupling proper. Nor will legitimate instances of coupling endow the
individual entities thus coupled with the characteristics of the entire coupled system. These entities play constitutive, historically established roles in the establishment or accomplishment of the functions of the entire system, where these roles
may widely diverge. I will discuss this point in more detail in the next section.
Second, and for related reasons, accounting for the historical nature of functions will provide an argument against the “fleeting vs. persistence” criticism
that has been levelled against extended cognition, first and foremost, by Rupert
(2010). The suspicion here is that environmental extensions, unlike internally
based cognitive capacities, are not persistently coupled with the organism. If some
such extension is accidentally detached from the cognising organism, it seems, the
cognitive system comes apart, only to be re-instantiated if and when that extension is successfully recovered. Clark’s reply to this charge was to emphasise the