[Wittrs] Re: Bogus Claim 5: Searle Contradicts Himself
- From: "SWM" <SWMirsky@xxxxxxx>
- To: wittrsamr@xxxxxxxxxxxxx
- Date: Thu, 22 Apr 2010 01:53:05 -0000
--- In Wittrs@xxxxxxxxxxxxxxx, Justintruth <wittrsamr@...> wrote:
> I am confused. Can someone help me?
>
> What I don't understand is the idea of syntax in Searle's work.
It's probably better to let those who count themselves Searleans respond first
(before I try) because, on my view, Searle is the one confused on the use of
"syntax".
> It
> seems to me that I can write a computer program which when executed
> processes the syntax of a language - as in the CR. But I can also
> write a program which when executed does not process the syntax of a
> language.
I think Searle is using "syntax" in a more generic sense than just as
linguistic syntax. He means formal rules (or the following of same in a
mechanistic or rote way). The syntax of language would thus be an example of a
type of syntax but would not be what he means in all cases.
> I can digitize signals for example (light, sound, chemical)
> and extract and process the information in them provided I have the
> right transducers. I do this in my work. Most computer programs that I
> know of do not process the syntax of a language as input. So why does
> the CR apply in general to computers?
Searle generalizes "syntax" to include ANY mechanistic processing including the
algorithms of programs. But then he goes on to suggest that the actual physical
implementation of such syntax is syntax, too, while claiming that syntax is
abstract, without causal efficacy. Later in his arguments he introduces the
notion that even what he calls "syntax" isn't syntax unless it's viewed as such
by a mind, thus you can't have syntax without first having semantics (because
minds have semantics and are characterizes as minds by the fact that they have
semantics).
> I noticed that in Searle's paper
> he was looking at a computer program that did in fact process the
> syntax of a language and that gave also output answers in that syntax.
> That was not a typical program. I also noticed that on the last page
> I think he seemed to expand the applicability but doesn't his argument
> - assuming it is true completely - only apply to computer programs
> that only process language syntax as input?
>
I would say no. He uses the CR (a system that responds to Chinese queries with
Chinese answers) to stand in for all computational processes even though the
underlying programming steps to achieve even that level of responding might
involve more than accurate the symbol matching he has specked the CR to be
capable of.
> In other words: Is it: "Searle's actual claim (from the CRA) is that
> computers can't be conscious merely by virtue of running a program."
> or "Searle's actual claim (from the CRA) is that computers can't be
> conscious merely by virtue of running a program that process the
> syntax of a language." ?
>
I would say it's the former, not the latter. But that reflects the ambiguous
usage of a term like "syntax".
> With respect to this question I have been struggling to understand
> whether all information is inherently syntactical. Is that what he
> thinks?
He thinks that syntax isn't enough to get understanding, contra Dennett who
suggests that what understanding finally is is the interplay of a complex
system of information processing at more basic levels which, when combined,
have the effect of being what we would recognize as understanding. It's the
difference between understanding as a elementary process level property and
understanding as a system level property.
> For example if light bounces off of the objects in a room we
> can derive information from those signals about the objects. In fact,
> for us the experience itself of those objects is really an
> interpretation of experiences caused by the signals combined with the
> "apriori" or "genetic" ability to render a three dimensional object
> world from that input. I know personally that it is possible to
> reinterpret the world in a way where objects no longer are experienced
> so there is definitely not enough information in the input signals we
> have to conclude that the world consists of objects (in fact its no
> longer scientifically credible either) but still that does not mean
> that those inputs do not contain the information required to conclude
> that they are consistent with a particular objective interpretation.
> Is that information, the light in a room full of things for example,
> "syntactical" in Searle's ideas?
>
A good question!
> The information stored in any signal interpretable as being about the
> presence of objects... is it always syntactical? It must be encoded
> somehow. If I look at the light coming through a pinhole in a room
> that contains a table and a lamp it will be different if I put a ball
> on that table. There is therefore information in the light about the
> room. Does that mean that the light signal contains a "syntax" in the
> way Searle means it? It seems that there is no "dictionary" for such
> signals.
>
Also a good point.
> The second problem I have is with the notion of consciousness as a
> property.
It's just a choice of terms for discussion purposes. I don't like using
"property" like this either but, over the years and in the course of these
discussions I have allowed myself to adopt that particular terminology because
others use it.
> One analogy of consciousness to matter from earlier posts
> was the analogy of the turning of a wheel to the wheel. Another was
> the liquidity of water which was not due to any component or part of
> the water but due to the aggregate system. But it seems to me that
> both of those cases are false analogies because in both of those cases
> the relative motion of the material involved is what is needed to
> predict the effect. Water is liquid because the molecules are not held
> in rigid arrangement by the forces between them whenever the average
> energy of the molecules is great enough for them to escape from a bond
> between one and another.
But it seems like experience in general -
> both consciousness and its correlates - cannot be predicted by motion
> alone.
It's not a perfect analogy because subjective experience is not the same as
observed phenomena and wetness and turning are observed phenomena. Subjectness
is experienced in a different way, a way that is not easily descibable as
"observed" (though we could use THAT term under certain scenario, as when we
speak of the Zen student observing his or her own thoughts during meditation).
>This means that not only syntactical processing but any
> detailed description of any physical device no matter how complex will
> not result in a prediction of consciousness or of its correlates the
> so called "quallia".
I'm not sure I would agree with that. Say we learned what particular events in
the brain were causal for particular brain states we called conscious. In that
case we could predict the instance of consciousness by observing the initiation
of the relevant brain events. If we could ascertain just what those brain
events are doing, then theoretically at least we could use other physical
processes to perform the tasks and thus produce consciousness on other
platforms.
> So the problem of syntax becomes completely
> irrelevant. The problem is that the description of any physical motion
> of any system of material particles does not result in the prediction
> that an experience will occur. Some additional principle must be used.
Why shouldn't identification of relevant brain proceses and an understanding of
them enable predictability?
> Something like: "Assemble a series of molecules in a certain way and
> start them moving and experiences of such and such will be had by that
> assembly."
>
> In summary the distinction between computer programs that process only
> the syntax of a language
By, on Searle's view, performing other, more basic syntactical operations . . .
>seems to be occurring but that distinction
> (between syntactic processing and non-syntactic processing) does not
> exist at the physical level.
Agreed. Once you get to the physical level there is no different in principle
between the different platforms of operations, even if some platforms are more
suited than others for performing certain tasks.
> It might be better to say that the CRA
> really states that any material arrangement that causes experience,
> when provided exclusively with the syntax of a language as input, will
> not be able to become aware of information not contained in that
> syntax. Its kind of an argument against ESP not that matter can cause
> consciousness. Right?
> ==========================================
I think the issue of "information" also needs to be sorted out linguistically.
But insofar as the CR and the CRA are about what a computational platform can
do, the man in the room is just a proxy, his consciousness irrelevant to the
room's capabilities except insofar as it is needed to enable him to play a CPU.
What is really going on is that Searle is claiming that no system made up of
the constituents that make up the CR (syntactical -- in the broad sense of that
term -- processing) could ever be conscious.
I think this finally depends on whether we conceive of consciousness (and all
the features associated with it) as occurring at a process level (one or more
of the contituent elements within the CR) or at a system level (in which case
the CR is insufficiently specked to enable consciousness).
SWM
=========================================
Need Something? Check here: http://ludwig.squarespace.com/wittrslinks/
Other related posts:
