[Date Prev][Date Next][Thread Prev][Thread Next]   [Date Index] [Thread Index] [Author Index] 

Re: Observers/surrogacy

Judith,

Yes and no. It certainly clarifies one aspect: that no models are perfect and that each observer will have a unique perspective on reality. But it doesn't clarify the implication that there can be observation (or interaction) with no associated model at all - something I would have thought he would reject. If this sentence is meant only in the sense you state, then it does not relate to that deeper issue. In that case, his meaning was "Observers are not [accurate or equivalent] models of each other." But that seems like too much modification of his phrasing. He seems to be talking about some case where observers do not model each other at all - a distinction between observing and modeling.

-jjk

Judith Rosen wrote:
On observers/observations/surrogacies: He wrote; "Not all observers are
models of each other."

      
What my father meant by this is always going to be ambiguous but I do
believe I know what he meant: He was saying that what is observed is "in the
eye of the beholder" and no two beholders, or observers, are going to see
things the same way. It refers to the idea of surrogacy; or substituting one
observer for another, because if you are using formalisms to represent these
observers, then the obvious difference between two human beings will not be
as apparent. He was musing on the fact that any observation that is recorded
is automatically contaminated by the specific observer who recorded it-- a
inescapable fact that has been the bane of science since the dawn of human
consciousness. The observation is then further contaminated by the person
interpreting it, and so on. He was also musing on the fact that formalisms
leave out crucial information that can completely change the seemingly
obvious result.

Many of the notes near that quote had to do with surrogacy. One discussed
using rats as surrogates for humans in medical experiments and so forth.
Surrogacy is a concept that is so widespread in human life that it becomes
an assumption. People tend to forget that there's a big "IF" at the
beginning of the whole exercise: "IF rats are close enough surrogates for
humans in this capacity, THEN drug A could be a carcinogen..." or whatever
the experiment is about. That can be very misleading with regard to analysis
of the data (where you then have the joy of further complications with
different people interpreting the data!).

With scientists observing some natural phenomenon and trying to understand
it, the "if" would be along the lines of "IF I am seeing this clearly enough
and in its entirety..."

Does that help at all?

Judith






  
I don't know what this means. What is the criteria for observers that
are models of each other? One speculative interpretation, might be that
in order to establish a modeling relationship, or I should say when
establishing a modeling relationship, one necessarily establishes a
"larger system." There implies a hierarchy (although not a unique one)
where a larger system can contain a model of a smaller one. We may then
ask how a smaller system's internal model can model larger systems. This
could be the purpose for his statement - a question in this regard.

It seems clear from other aspects of the view/theory that all these
systems are connected in some way. A big theme of Rosen's was how larger
systems DO causally affect the ontology of smaller ones, violating the
usual taboo on this in traditional science. In other words the existence
of parts may partially be owed to factors in the whole. This was the
case with protein folding examples. The explanation here seems to be
that larger-smaller hierarchies exist naturally (aside from any
scientific model), and that the smaller system models, say those
employed by an animal, are not formally restricted just to the encodings
contained in the animal itself but also extend up and down the
hierarchy. As an example, my model of the world might include real-time
influences from the environment that exists independently of me, i.e.,
the encodings embedded in the environment rather than an internal memory
storage system. In that case, the entailments exist up and down the
hierarchy.

In other words, this way of seeing nature in terms of modeling
relationships, eliminates the physical reductionist hierarchy, but then
spreads any system through an infinite hierarchy of modeling relations.
This implied connectedness is very ecological, and thus appealing in
that regard (everything is connected to everything else), but it means
that sub-systems would at least partially entail their larger systems.
So again, which obsrvers are not models of each other?

All I can think of is the case of an observation model that has no
decoding entailment loop - like a thought that is never tested or acted
upon. One can argue that this is ultimately an impossibility, so I would
question the statement. Could his intention have been to consider the
proof of this impossibility????

Another possibility is that he was focusing on the "each other" part.
Perhaps all observers are models of what they observe, but there is the
case where this is not mutual - that the observed may not be observing
back. That would seem defensible. So, if I observe animals from behind a
blind, they may not be aware of my presence. That would seem like an
obvious case, but a relative one. In an absolute sense, even my blind
observation does feed back to the subject through future changes in my
behavior and alterations in the environment, management schemes, etc. So
ultimately even this case is entailed.