RT list: review of Carruthers

Greetings. I paste below my review of Carruthers. A interesting book indeed.

Peter Carruthers The architecture of the mind, Oxford: Oxford University
Press, 2006, xviii + 462 pages. ISBN: 978-019920708-4.
Reviewed by Alessandro Capone (University of Palermo).

The volume ‘The architecture of the mind’ by Peter Carruthers denotes
great erudition and intelligence, and is certainly the basis for future
research on the modularity of the mind.
        According to Carruthers, the mind consists of a number of modules,
functional components which are autonomous as far as their processing
mechanisms are concerned, and sufficiently isolable and separate to allow
the whole mind to execute its work, even if one of its components (or
modules) has been damaged (say, the case of a person affected by a
stroke, who will lose some of the functions of her mind (e.g. writing
abilities), but not all of them). Modularity is also considered in
relation to evolution, and the modular mind is considered capable of
allowing modifications of one of its modules, or of adding a module as a
consequence of adaptive changes in response to environmental challenges.
        In the picture of massive modularity presented by Carruthers, each module
is in constant dialogue with the other modules, in the sense that it can
query a number of modules for information. Each module is switched on by
a limited type of input (so there are heavy restrictions to the type of
input a module can process, but surely a module can receive its input
from different modules, as in the case of the practical reasoning module,
which takes as input both desires and beliefs). Carruthers rules out
categorically that each module has access to the internal operations of
the other modules, as this would involve a simulation process imposing
an excessive burden on processing. However, he admits the possibility
that each module can search the archives of other modules.
The same module can be located in more than one area of the brain, but is
strictly correlated with a certain function. If two functions are
structurally different, in that they involve different types of
operations, then they will correlate with distinct modules. A module may
consist of a number of sub-modules. For example, the mind-reading module
consists of a module that deals with automatic inferences, as the ones
studied by relevance theorists, and a reasoning module that deals with
non-automatic, conscious and relatively slow inferences.
 Modules may share a number of parts, and it is predicted that modules
situated in areas of the brain that are sufficiently close will share a
greater number of parts than modules situated at a distance, since the
relatively slow propagation of the information between modules may offset
the advantage of having parts that are shared. Carruthers believes that
the same packages may underpin distinct modules. (This may be a
drawback. After all, the modular architecture has the advantage of
insulating one region of the brain from another and of preventing damage
in one area from affecting the normal functioning of the remaining
modules. If modules share parts, then how should we explain the fact that
 damage in one module will not affect the functioning of another? By
saying that only modules situated in areas of the brain that are in
close contact share parts, Carruthers limits the problem, as damage in
one area will have consequences ONLY on neighboring areas).

Carruthers claims that central intelligence is the result of the
interaction among modules, as integrated by further predispositions.
Carruthers believes that central intelligence consists of two systems. In
system 1, operations are highly specialized, and this is the reason why
its output is fast. System 2 is responsible for steps in reasoning that
are conscious. And that is what the global architecture does for us when
supplemented with mental rehearsal and language.
        Carruthers believes that there is a module that integrates information
coming from distinct modules. He does not believe that this is the
mind-reading module, but thinks that this is the language module. In
fact, the language module is already capable of interfacing with modules
using distinct algorithms. This is a reasonable hypothesis. Notice what
happens in an utterance of:

(1) Mary is drawing a triangle.

Here the speaker and the hearer are integrating information from the
perceptual module (the reference of ‘Mary’) and of the geometry module (a
triangle activates geometrical information) .

According to Carruthers, it is the syntax of language, as well as its
recursive and hierarchical organization, that allows the integration of
different types of information coming from distinct modules. For example,
the fact that we can insert multiple modifiers under an NP node allows us
to integrate information from different modules. NPs typically allow us
to integrate information coming from perception (see also Jaszczolt 1999
on default semantics) with other type of information.

Another very important part of Carruthers’ research is constituted by his
ideas on practical reasoning and mind-reading, namely that human beings,
but also certain apes, are able to rehearse an action in their minds,
broadcasting the result to central/conceptual modules, which give further
beliefs and motivational states as output. A child who holds a banana in
his hand may associate the banana with the concept of a telephone and then
rehearse in imagination the action of phoning his grandmother, which
provokes pleasure and thus determines the continuation of the action of
rehearsing.

The book is very rich in ideas. The bias in favor of the modular mind
probably leads the author to concentrate on behaviors that are automatic
and fast. The heuristics used in behavior are pretty simple, but there is
the risk that a certain amount of determinism will pervade our view of
human action. In fact, the author discovers a number of automatisms, such
as the choice of partners in human relationships, the use of fast and
frugal heuristics, etc..

The constant comparison of the human mind with those of apes and
invertebrates has obvious limits. After all, human beings are free to do
what they want to do, and they can be what they want to be. Human beings
can use this infinite freedom to indulge in capricious, unmotivated
actions, of which animals are not capable. I can decide whether to act or
not to act simply by tossing a coin. Human actions are capable of being
totally arbitrary and unmotivated. They are also capable of reflective
processes like writing a treatise like Kant’s ‘Critique of Pure reason’.
        Anyway, Carruthers agrees that the claim that humans have a degree of
control over themselves that other animals don’t is correct, and he
argues that it is explicable within the modular framework. That is what a
normative module can get us. Beliefs about norms (either for action or
for reasoning) can give rise to a novel intrinsic motivation, which often
issues in novel behavior.

Another limit of the book is that while it considers communication as an
important way of propagating information and technological inventions, it
does not fully explore theories of communication that are alternative to
Sperber & Wilson’s. For example, in the theory of Dascal (2003) ample
space is devoted to rationality and its influence on language. (It is
true, however, that the book discusses ideas by authors like Chomsky,
Pinker, Dennett, and Vygotsky).

Summing up, I found this book very instructive and thought-provoking. I am
sure it will have fruitful ramifications in due time and it will serve to
throw further light on pragmatic theories of communication. Especially
promising is the idea that a module – presumably the language module – is
dedicated to integrating information coming from distinct modules. Seen in
this light, an utterance can be seen as a merger representation à la
Jaszczolt (2005), in which semantic, pragmatic and encyclopedic
information merge through slots provided by the syntax present in the
sentence or by syntactic constituents that are added at the level of the
merger representation in order to make a thought plausible and maximally
rational. Since perceptual, mind-reading, geometrical, etc. modules may be
involved in communication of information, syntax is what allows the
integration of the various types of information, and it is possible that
syntactic constituents are added in the merger representation through the
pragmatics of language in order to give room to constituents of thought
such as modes of presentation of referents (see Jaszczolt 1999), thus
integrating the theory of mind module with the perceptual module.

References

Chomsky, N. 1957. Syntactic structures. Berlin: Mouton.
Dascal, M. 2003. Interpretation and understanding. Amsterdam: John Benjamins.
Dennett, D. 1991. Consciousness explained. London: Penguin.
Huang, Y. Pragmatics. Oxford: OUP.
Jaszczolt, K.1999. Discourse, beliefs and intentions. Oxford: Elsevier.
Jaszczolt, K. 2005. Default semantics. Foundations of a compositional
theory of acts of communication. Oxford: OUP.
Pinker, S. 1997. How the mind works. London: Penguin.
Sperber, D. and Wilson, D. 1986. Relevance. Oxford: Blackwell.
Vygotsky, L. 1961. Thought and language. Cambridge, Ma: MIT Press.

Reviewer’s address:

Private:
Dr Alessandro Capone
Via San Francesco P. 105
98051 Barcellona PG ME
Italy

Department of Philosophy
University of Palermo
Viale delle Scienze
Palermo - Italy
Received on Fri Jul 17 23:15:36 2009