Slides from 9/15/11

No new reading for Tuesday.
Write about McKinsey’s paper
(Chapter 4) for next Thursday.
Narrow Content and Cognitive
Science
Broadly speaking, cognitive science is
the study of the mechanisms that
produce intelligent behavior (this
includes the study of perception,
language-use, theory construction,
and memory, among other things).
The standard procedure is to construct
models of the internal process that
produces intelligent behavior.
What Is a Model
-a specification of basic elements (e.g., neurons or
atomic symbols in a code)
-a specification of basic operations of the system
(say, copying symbols from one location to
another)
-a specification of the architecture of the internal
system (e.g., where things can be stored, which
parts of the system can communicate with which
other parts)
-a specification of the rules or procedures (e.g.,
algorithms) that lead from one state of the internal
system to other states
Where Does Broad Content Fit
In?
Broad content of basic elements (symbols
defined perhaps only by bare relations of
sameness and difference) is fixed by
causal relations or some kind of
correspondence to external properties,
individuals, or kinds
This allows symbol strings to have truthconditions or correctness conditions in
relation to the external world
Narrow Content?
In a given individual, the narrow
content of a symbol might be thought
of as the way in which the symbol
interacts with other symbols.
But, whether or not this is truly content
might be a stipulative matter. It’s not
clear that the overall causal role of the
symbol is itself representational.
Combine Sawyer’s distinction with the
cognitive scientific approach: two people
have the same concept iff both have some
symbol or other that has the same broad
content.
Even when different people share an atomic
concept, though, that concept may well
appear in different sets of stored symbol
strings. (They may have different
conceptions.) The differences among
these sets cause differences in behavior.
There’s no need to invoke narrow content
to explain such differences.
These differing roles matter in cognitive
science; they predict subjects’
differing responses, reaction times,
learning curves, and so on.
They might also be thought of as
narrow contents insofar as they
produce different self-reports about
one’s own states. But one could also
focus on the different broad contents
of the symbol strings, without
invoking any narrow content at all.
Narrow narrow content
Narrow content is whatever is shared by
physical duplicates.
It is a function (in the mathematical sense)
from environments to broad contents.
It’s thought to be necessary to play the
local causal role in the production of
behavior.
Sawyer’s objections
-It’s not content (functions aren’t
representations).
-It’s derivative upon broad content (what is
the function defined over? broad
contents).
-It’s not necessary for an explanatory
psychology (mechanisms might be
necessary to implement psychological
laws, but psychological laws aren’t about
local mechanisms).
Epistemic narrow content
Something like the subject’s first- person
perspective or the evidence that the
subject has direct access to. For all
subjects knew in 1600, water could have
turned out to be H2O or XYZ.
Sawyer: What form does such evidence
take? Is there a sufficiently rich, neutral
language in which it can be stated or
conceived of?
Thoroughly narrow content
On this view, all psychological content
is narrow.
True propositional attitude ascriptions
assert a (sufficiently close) match
between the sense of the expression
used to attribute the attitude and the
content of the attitude.
Sawyer’s objections
If content is to play a role in
psychological explanation, it must
be shared by different subjects. On
Segal’s view, there is too much
variation in individual narrow
content for there to be any viable
scientific (or explanatory)
psychology.