Re: Relational "Space"

[Tim Gwinn <***>]

Judith,

Yes, I was using 'measurement' in a broader sense than the more typical
sense you were using. I tend to think of 'measurements' as general kinds of
mappings (where quantitative measurements are just those particular ones
that happen to map to some number set such as reals or rationals), in which
case it becomes largely synonymous with the term 'encoding'. So, sorry for
causing confusion.

As to 'metaphor', your father explicitly distinguished metaphors from
models, where metaphors are occur in modeling relations minus the encodings.
[LI 64-66] So, a metaphor is a model-like thing which allows us to make
predictions (decodings) and to impute things back to the natural system. But
we do so without being able to encode back from the natural system: we give
up the ability to verify these imputations.

Going back to the original thread, if we take the hypothesis that there is
an unperceivable primary structure and unperceivable conformation which is
causally responsible for phenomena, and we build a formal representation of
that primary structure and conformation, then I think that we have created a
metaphor, not a model. Because they are both unperceivable, there is no way
to do any encoding from the primary structure or conformation to the formal
side of the modeling relation. Thus, we do not have a commuting modeling
relation, and our formal representation is not a model, but a metaphor. How
does that sound?

Regards,
Tim

> -----Original Message-----
> From: ROSEN Forum [mailto:*** Behalf Of Judith
> Rosen
> Sent: Thursday, March 18, 2004 10:41 AM
> To: ***
> Subject: Re: Relational "Space"
>
>
> Tim, we're speaking in tongues here. I realized that our definitions are
> completely different when you said "a measurement doesn't have to be
> quantitative".  My understanding of the word "measurement" connotes
> quantitative analysis of something. Therefore, it looks like my statement
> that "we can prove the existence of something we cannot directly
> perceive by
> observing the behavior of the phenomena it generates" means (in
> your terms)
> that "we can measure it indirectly, using the phenomena it generates".
>
> If that's so, then we are in agreement.
>
> My father said, in response to the accusation that theory is "abstract":
> "There is nothing MORE abstract than a MEASUREMENT." He was putting
> measurement into the realm of experimental science, from whence the
> accusation of abstraction was coming. His use of the word usually referred
> to quantitative analysis of some kind.
>
> However, I was struck by something new you said in your post about the
> modelling relation:
>
> > I very much disagree with this. What about the modeling relation - the
> > "habitat of all epistemology"? Without measurements, there is
> no encoding
> in
> > a modeling relation, and it cannot commute. If we only have the
> decoding,
> > then the result is that we have metaphors, not models.
>
> The modelling relation is a wide ranging area that goes from very
> precisely
> measured (in my sense of the word) aspects in a precise model (say, of a
> simple system or of a subsystem from a larger or complex system) to very
> broadly defined models, such as my father's model for a complex system. He
> didn't quantify anything in that diagram except the existence of
> the system
> and the existence of its abilities, in a particular arrangement of
> relationships. That model has been attacked for stating the existence of
> things that cannot be precisely quantified. These kinds of
> imprecise (as in
> unquantified) models are, indeed, a form of metaphor or analogy. Why would
> the fact that a model is analogous to the system it represents
> make it "not
> a model" in your opinion? Isn't that one way to define what a model IS?
> Further; how do you get to your logic of encoding and decoding in your
> analysis? I don't see how there "is no encoding" in a modelling relation
> without measurements. That's a non-sequitur. Even if I change the word
> measurements, in the above statement, to not mean 'quantitative analysis'
> but something else, even a vague model such as my father's (M,R)-System
> 'commutes'. It has encoding. And it is analogous.
>
> Again, I suspect that you and I are defining things differently from each
> other.
>
> Judith
>