gence, and by extension against any psychological or neurological theory which
describes the mind as a computer. Computers are created by humans who have
minds, and the inputs and outputs of computers are always interpreted by
human consciousness. To suggest that a computer might think is only a façon
de parler. It is we who project a homunculus into the computer, a human mind
which decides how the pattern of electronic branchings is to be interpreted as
meaningful ideas. Things and ideas are irreducible realms; there simply is no
way to go from an external description of the one to the inner meanings of
the other.
The Leibniz-Searle argument might seem to hold against sociological re-
ductions of ideas as well; the observable economic and political behaviors of
classes and states are not of the same kind as the ideas they purport to explain.
Yet Leibniz points to a clue: the connection between ideas and the spatial world
of human bodies is endlessly mysterious if these are truly distinct substances;
ideas and bodies are correlated because they are aspects of a single kind of
being. Naturally one will not find ideas amid the machinery of the brain or
the computer if one is looking for an idea-thing amid material-things. Ideas
are not thing-like at all, except insofar as we represent them in symbols written
on materials such as paper, but are first of all communication, which is to say
interaction among bodily humans. To enter into the physical brain (or inside
the computer) is precisely the wrong way to perceive ideas; for ideas are in the
process of communication between one thinker and another, and we perceive
the ideas of another brain only by having them communicated to us. It is the
same with oneself: one perceives one’s own ideas only insofar as one is in a
communicative mode. There is no thinking except as aftermath or preparation
of communication. Thinkers do not antedate communication, and the commu-
nicative process creates the thinkers as nodes of the process.
The difficulty of reducing ideas to political economy is not found in this
direction. Economic and political activity is not merely physical but mindful
as well, above all because it is social. The force of the anti-reductionist point
is that certain kinds of ideas which we are interested in explaining cannot be
explained by referring to social action where that kind of communication does
not take place. There are regions of sociological reduction where the explana-
tion is crude and unsuccessful. Economic and political macro-structures do not
explain much about abstract ideas, because such ideas exist only where there
is a network of intellectuals focused on their own arguments and accumulat-
ing their own conceptual baggage train. It is the inner structure of these
intellectual networks which shapes ideas, by their patterns of vertical chains
across the generations and their horizontal alliances and oppositions. Reduc-
tion is an error not because we are making a primitive category mistake about
ideas and things, but because we look for a pattern of communicative action2 •^ Introduction