From newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rutgers!cs.utexas.edu!usc!rpi!think.com!yale.edu!ira.uka.de!chx400!bernina!neptune!santas Tue Mar 24 09:56:43 EST 1992
Article 4544 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!ists!helios.physics.utoronto.ca!news-server.csri.toronto.edu!rutgers!cs.utexas.edu!usc!rpi!think.com!yale.edu!ira.uka.de!chx400!bernina!neptune!santas
>From: santas@inf.ethz.ch (Philip Santas)
Newsgroups: comp.ai.philosophy
Subject: Re: Definition of understanding
Message-ID: <1992Mar17.221557.18924@neptune.inf.ethz.ch>
Date: 17 Mar 92 22:15:57 GMT
References: <1992Mar12.233945.9244@psych.toronto.edu> <1992Mar13.143003.26132@neptune.inf.ethz.ch> <1992Mar16.222403.26423@psych.toronto.edu>
Sender: news@neptune.inf.ethz.ch (Mr News)
Organization: Dept. Informatik, Swiss Federal Institute of Technology (ETH)
Lines: 163
Nntp-Posting-Host: spica.inf.ethz.ch


In article <1992Mar16.222403.26423@psych.toronto.edu> michael@psych.toronto.edu (Michael Gemar) writes:
>In article <1992Mar13.143003.26132@neptune.inf.ethz.ch> santas@inf.ethz.ch (Philip Santas) writes:
>>
>>michael@psych.toronto.edu (Michael Gemar) writes:
>>> santas@inf.ethz.ch (Philip Santas) writes:

PS:
>>They can refer to another Type, to another Set of objects, to Text, to
>>a space in the memory, to a drive, to sound, and whatever you want.

MG:
>I think we are talking at *completely* different levels.  The computer
>may very well be able to *differentiate* between the different things
>a variable can "refer" to, just as I can tell the difference between
>the strings "brillig" and "slithy".  But that does not mean that
>these distinctions carry any *semantic* meaning.  They may very well
>cause *behavioural* differences in the computer.  But they have no
>meaning as the term has been technically defined by almost all
>philosophers.

Just by curiosity and scientific interest, can you explain to me the
meaning of a semantic net? What are the nodes?

MG:
>>>differentiated, but there is nothing in the computer to say that they
>>>refer to distance and spring constants as opposed to potential and
>>>capacitance, or apples and oranges.  Just naming the type "Spring Constant"
>>>doesn't do squat.
>>>Distance x = displacmentFromEquilibrium
>>>does not tell the computer what "distance" and "displacement from equilibrium"
>>>*is*!  I could have just as easily typed:
>>>Qaatlus x = GwornsBleebArack
>>>and the program would *still* compute *both* Potential *and* electrostatic
>>>energy.

PS:
>>>>Notice that there are many human languages and you can say the same thing
>>>>with differnt words and syntax. What I do not understand is how can a
>>>>you calculate two irrelevant things by evaluating just one function.

MG:
>The whole point of the above example is that they two systems are
>*functionally identical*.  Indeed, there is a whole *discipline* devoted
>to examining the relations among physical systems that are functionally
>identical - it is called "systems theory".

Excellent. There are sentences in the human language that are semantically
identical, so what? People from the domain of linguistics believe that
the european languages are the same powerful. Is this used as argument against
human language?

By telling me that the information which differentiates two types is only their
naming, when there is obviously a namber of procedures, variables and objects
related with these types, which help the building of the semantical model,
you put the whole faculty of computer science into existential problems...

PS:
>> Notice that type information
>>is just one form of information: there are some more like membership,
>>satisfaction of inequalities etc. that help us to deal with well defined
>>and bad defined (although not all levels) of concepts.

MG:
>*WHAT* concepts?  All you have are variables with labels, connected to
>other variables with labels.  They have *no* privileged interpretation.
>(Could *you* tell the difference between a potential energy calculating
>program and a electrostatic energy calculating program based *solely*
>on the *relations* between the variables (remembering that variable
>names are arbitrary).

PS:
>>You mix variables with types and other information related with them.
>>If your variable is of type Dinstance, it cannot be of type Resistance
>>except if you have created a relationship between these types (something
>>obvious for types Distance and Magnitude).

MG:
>How does the computer know that when I type a "12", I mean Distance and
>not Potential Difference?  It's *just* a number...!

Aha. So for you a number is NOT a number but something else.
Are you serious?

In 'low' level languages like C you can do declarations
Distance a=12;
Then a has no chance of being Potential differnce.
In more advanced languages you can do even more complex things.

MG:
>>>C++ may have the ability to distinguish between different variables.  It
>>>does *not* have the ability to provide semantics for those variables.

PS:
>>Their membership to a class (or type, or set), their sharing of functions
>>and other variables etc is a form of semantics.

MG:
>No, this is all syntactic information, as syntax has been traditionally
>defined.

In other words you say that languages like C++ change their syntax
dynamically since the user can create new datatypes. I have never
thought of this...

Once again: with the usage of types you build models. This is the semantics
of your program.

PS:
>>What do you mean by "meaning to the symbols"? This is a crucial problem
>>I suppose.

MG:
>I mean "meaning" in that they have one unambigious interpretation, just like
>my thoughts do.

In other words computer programs have not unambigious interpretations.
This is again an interesting unverified point which of course I do NOT share.

MG:
>This is completely opaque to me.

The classical argument...

PS:
>>It is not a matter of belief, but of logical argumentation, search for
>>rational possibilities, and rejection of contradictions...

MG:
>...all of which, as far as I can see, deny the possibility of a C++
>program having semantics.

Beg your pardon? 

BTW, I have never said that C++ is adequate for AI programs.
I gave some examples for reasons of simplicity.

MG:
>While I appreciate your attempt to get at the issues discussed here, it
>seems to me that you are somewhat unclear as to how certain terms are
>used.  I would suggest that you look at some references that discuss
>these issues in more detail.  Until we are both clear on the basic
>terminology, I don't think it is worthwhile to pursue this on the Net.

It seems that you have a strange attitude against computer programs
and the things you can model with them. If you still believe that
a program cannot have semantics (when CS text books argue for the
opposite) I think that we have a serious problem of missconception.

On the other hand you can protest against all those who write papers
and explain algorithms and physical systems with the usage of
computer programs.

Philip Santas

--------------------------------------------------------------------------------
email: santas@inf.ethz.ch				 Philip Santas
Mail: Dept. Informatik				Department of Computer Science
      ETH-Zentrum			  Swiss Federal Institute of Technology
      CH-8092 Zurich				       Zurich, Switzerland
      Switzerland
Phone: +41-1-2547391
      


