Program? John Searle

topsalmonAI and Robotics

Feb 23, 2014 (3 years and 5 months ago)


Is the Brain’s mind a computer
Program? John Searle

Can a machine think?

Can it have conscious thoughts as we have?

Can it be said to think just by virtue of
implementing a computer program?

Is the program by itself constitutive of

g AI view point:


by designing the right programs, minds
are being created. Believe that the Turing test
(devised by Alan Turing) can determine the
success or failure of attempts to create
thinking machines.

Turing Test:

If computer can perform in s
a way that an expert can’t distinguish it from a
human with a certain cognitive ability, then
the computer also has that ability.

Hence, goal
> design programs that simulate
human cognition in order to pass Turing Test.
The resultant program would be

a mind, not a
model of one.

Thinking is just the manipulation of symbols.

‘The mind is to the brain as the program is to
the computer’.

Weak AI view point:

Computer models are useful for studying the

Chinese Room:


Baskets full of Chinese


Rule book (in English) for manipulation of

People outside hand in bunches of symbols.

Manipulate according to rule book.

Hand out results.

Could satisfy Turing test without
understanding Chinese. Hence
> manipulating
symbols, but attac
hing no meaning. (Like a

So manipulating symbols is not enough to
guarantee understanding, perception etc.

Searle goes on to make a number of

Programs are defined In terms of symbol
manipulation & the symbols are purely

Computer programs are formal (syntactic)

Programs can be run on different hardwares.
Symbols are manipulated without reference
to meanings. Symbols can stand for

Human minds have mental contents

So just having sy
ntax is not enough

Syntax by itself is neither constitutive of nor
sufficient for semantics


Programs are neither constitutive of nor
sufficient for minds

Therefore, Searle says, strong AI is false.

We can simulate the brain on some levels, the
n thinks, but this does not mean that SM
is thinking.

Strong AI says that SM is thinking because
that is all thinking is. Churchlands mis
understanding of this

think that strong AI
says that
one day computers will think
, but
strong AI says that
they DO

New parallel technologies modelled on neural
networks, useful models, all parallel can still
be done serially. Chinese room still refutes it.

Simulation no more real than simulation of
other things e.g. ...

Brains cause minds

Any other syste
m capable of causing minds
would have to have causal powers (at least)
equivalent to those of brains

Any artefact that produced mental
phenomena, any artificial brain, would have
to be able to duplicate the specific causal
powers of brains, and it could
not do that just
by running a formal program.

The way that human brains actually produce
mental phenomena cannot be solely by virtue
of running a computer program.

Thesis of strong AI

any system, whatever it’s
made of, must have thoughts and feelings
provided it implements the right program with
the right inputs and outputs.

Some objections to Searle’s ideas:

1. In the Chinese room you really do
understand Chinese, even though you don’t
know it.

2. You don’t understand Chinese, but there is
an unc
onscious sub
system in you that does.

3. You don’t understand Chinese but the
whole room does.

4. Semantics don’t exist anyway; there is only

5. You are not really running the computer

you only think you are.

Could a machine Think? P
aul &
Patricia Churchland

What is their view of AI?

Classical AI

Reply to Searles Chinese room argument with
the Luminous Room

Electricity and magnetism are forces

The essential property of light is luminance

Forces by themselves are neither const
of nor sufficient for luminance


Electricity and magnetism are neither
constitutive nor sufficient for light.

SM machines are just the wrong architecture
for the job.

How does the brain achieve cognition?

Neuro sciences reveal a lot abo
ut the brain.

Clearly the brain is making systematic use of
... computational advantages. But it need not
be the only physical system capable of doing
so. AI, in a non
biological but massively
parallel machine, remains a compelling and
discernible prosp