Philosophy of Physics - I
orhan yenen
Following Thomas Brody (The Philosophy Behind Physics,
Springer Verlag, Berlin, 1993), we start with few assumptions (or
axioms or whatever). (A philosopher may call this "a naive
realist's point of view/ontology"). If you do not accept them fine,
do not read the rest. However, without these assumptions we
cannot talk about sciences, let alone the philosophy of sciences.
And the biggest argument for a realist ontology is that science
makes correct predictions more often than one could expect on the
basis of mere chance (James R. Brown, Smoke and Mirrors/How
science reflects reality, Routledge, New York, 1994).
There is a reality outside us that we call "the nature". We want to
understand, describe, and predict the behavior of this "nature" as
much as we can. To do so, we use an "active epistemic cycle, that
is to say we interact with a part of this "nature" and see what
happens.
The epistemic cycle begins with a partial interpretation, a guess,
an initial model. Then we interact with a part the system, and note
the changes on the system due to our actions. We improve our
initial guess, model or interpretation until an adequate fit is
obtained with the changes on the system due to our actions. We
will iterate this "active epistemic cycle" until we get the adequate
fit to the required practical and possible precision. Of course, there
is no guarantee that iterating the epistemic cycle will lead to a
satisfactory model of the part of nature under study. The process
may not converge. In this case we start with a different initial
model and go through the epistemic cycles again, and again and
again.
This process of restarting with an initial model sometimes
necessitates that another epistemic cycle is applied at a higher
level. For this purpose, the higher level cycle needs to have its
own initial model. Once the machinery for higher level epistemic
cycles is in place, the process may be used recursively. (More
about higher-level epistemic cycles later).
After few epistemic cycles, a working partial model can be
achieved. Note that this model is partial and some aspects of the
system under study may be grossly distorted. If an analogy must
be made, our model is like a map. It is made for a purpose. For
instance the map of the underground rapid transit system of a city
may distorts distances, or angles, yet it would be a good model
of the topology of the stations. Different maps of this city will all
refer to different aspects, such as the water-map, electricity
distribution, or surface roads. They may all distort certain
properties of the region they describe. Individually, each map
emphasizes another feature of the same region, and we combine
them to achieve a better understanding.
For converging epistemic cycles, the cycles are repeated until the
model describes the segment of nature under study within the
precision we need for the purpose we have in mind.
Low level models are built with highly specific aims. A higher-
level model is built so with its help many different lower level
models can easily built for a wide variety of different aims. As the
higher level models become more and more general, they can be
used for wider ranges of purposes. To illustrate what we mean by
high/low level epistemic cycles, let's consider epistemic cycles to
describe the motion of a small mass at the end of a string
(pendulum). After few cycles, we may have a working model
which would adequately describe the motion of the pendulum, for
example for small oscillations ( a sine or cos function for the
elongation as a function of time, under some initial conditions).
Similarly we can have few epistemic cycles to adequately describe
the motion of a small mass at the end of a spring. These would be
the lower level epistemic cycles. At a higher level, we may combine
what we already know to make a model to describe all kinds of
oscillations. This in turn may be combined by the description
of the motion of the planets to form an even higher level,
such as classical mechanics, whichin turn may be combined
with the epistemic cycles for describing the radiation (called
classical electricity and magnetism) to form
the classical physics. As we go to higher and higher level
epistemic cycles, at some point, the social institution called "the
physics community" names a model at a given level, "a physical
theory".
The epistemic cycle described here is the only method of acquiring
original new knowledge. Other methods suffer from the defect of
not being verifiable. The application of epistemic cycle
distinguishes a developed theory from pure speculation. The
danger is our refusal to apply another epistemic cycle to an
established theory which then may become an obstacle for further
development and results in a dogma.
In summary, the effect of an action on a segment of nature under
study is predicted from the previous version of the epistemic model
and the correlation between what was predicted and what is
actually found provides further knowledge. The application of
these cycles is an ongoing process and the improvements on the
model are dynamic.
(To be continued)
Orhan Yenen
David Gudeman
]Following Thomas Brody (The Philosophy Behind Physics,
]Springer Verlag, Berlin, 1993), we start with few assumptions (or
]axioms or whatever). (A philosopher may call this "a naive
]realist's point of view/ontology"). If you do not accept them fine,
]do not read the rest. However, without these assumptions we
]cannot talk about sciences, let alone the philosophy of sciences.
This last sentence is demonstrably false. In fact it is trivial to
come up with counter-examples.
]And the biggest argument for a realist ontology is that science
]makes correct predictions more often than one could expect on the
]basis of mere chance...
If that is the biggest argument for realism, then realism is in big
trouble. The argument does not support realism at all. An
anti-realist ontology does not require that things happen by "mere
chance", and a realist ontology does not prohibit the possibility that
things happend by "mere chance". Therefore the question of realism is
completely independent of the question of randomness, and an argument
against randomness can give no support at all to the belief in
realism.
]There is a reality outside us that we call "the nature". We want to
]understand, describe, and predict the behavior of this "nature" as
]much as we can. To do so, we use an "active epistemic cycle, that
]is to say we interact with a part of this "nature" and see what
]happens.
Just what parts of the reality are "outside" us? Sounds? Colors?
Shapes? Textures? Sizes? Distances? Temperatures? Changes? Are
all of these outside us, or just some of them? On what basis do you
decide that some things are "outside"?
]The epistemic cycle begins with a partial interpretation, a guess,
]an initial model. Then we interact with a part the system, and note
]the changes on the system due to our actions. We improve our
]initial guess, model or interpretation until an adequate fit is
]obtained with the changes on the system due to our actions.
What constitutes an "improvement" to the initial guess? How do you
measure the "fit" of a model to a completely unknown quantity?
]After few epistemic cycles, a working partial model can be
]achieved. Note that this model is partial and some aspects of the
]system under study may be grossly distorted. If an analogy must
]be made, our model is like a map. It is made for a purpose. For
]instance the map of the underground rapid transit system of a city
]may distorts distances, or angles, yet it would be a good model
]of the topology of the stations. Different maps of this city will all
]refer to different aspects, such as the water-map, electricity
]distribution, or surface roads. They may all distort certain
]properties of the region they describe. Individually, each map
]emphasizes another feature of the same region, and we combine
]them to achieve a better understanding.
If the models are all distortions of the reality, in what sense to
they give an "understanding" of the reality?
And before they can even be distortions, there has to be some form of
similarity to appeal to. How could a mental image or a mathematical
model be "like" a real physical process? What features could they
have in common in the first place?
]The epistemic cycle described here is the only method of acquiring
]original new knowledge. Other methods suffer from the defect of
]not being verifiable.
How is your method verifiable? How do you verify that your model has
any relationship at all to the underlying reality?
--
David Gudeman
gud...@cs.arizona.edu
orhan yenen
>]Following Thomas Brody (The Philosophy Behind Physics,
>]Springer Verlag, Berlin, 1993), we start with few assumptions (or
>]axioms or whatever). (A philosopher may call this "a naive
>]realist's point of view/ontology"). If you do not accept them fine,
>]do not read the rest. However, without these assumptions we
>]cannot talk about sciences, let alone the philosophy of sciences.
>This last sentence is demonstrably false. In fact it is trivial to
>come up with counter-examples.
May be. What are those trivial counter-examples?
>]And the biggest argument for a realist ontology is that science
>]makes correct predictions more often than one could expect on the
>]basis of mere chance...
>If that is the biggest argument for realism, then realism is in big
>trouble. The argument does not support realism at all. An
>anti-realist ontology does not require that things happen by "mere
>chance", and a realist ontology does not prohibit the possibility that
>things happend by "mere chance". Therefore the question of realism is
>completely independent of the question of randomness, and an argument
>against randomness can give no support at all to the belief in
>realism.
No no, that is not what I meant. What I said is that
predictions on a random basis would not have the same success as
the predictions based on sciences. Am I missing what you said?
>]There is a reality outside us that we call "the nature". We want to
>]understand, describe, and predict the behavior of this "nature" as
>]much as we can. To do so, we use an "active epistemic cycle, that
>]is to say we interact with a part of this "nature" and see what
>]happens.
>Just what parts of the reality are "outside" us? Sounds? Colors?
>Shapes? Textures? Sizes? Distances? Temperatures? Changes? Are
>all of these outside us, or just some of them? On what basis do you
>decide that some things are "outside"?
Well, let me throw the ball back to you. What are things which are not
outside? Shapes, Textures, Distances, Sounds, Colors?...etc?
>]The epistemic cycle begins with a partial interpretation, a guess,
>]an initial model. Then we interact with a part the system, and note
>]the changes on the system due to our actions. We improve our
>]initial guess, model or interpretation until an adequate fit is
>]obtained with the changes on the system due to our actions.
>What constitutes an "improvement" to the initial guess? How do you
>measure the "fit" of a model to a completely unknown quantity?
I am not sure what you mean here. Can you expand a little?
>]After few epistemic cycles, a working partial model can be
>]achieved. Note that this model is partial and some aspects of the
>]system under study may be grossly distorted. If an analogy must
>]be made, our model is like a map. It is made for a purpose. For
>]instance the map of the underground rapid transit system of a city
>]may distorts distances, or angles, yet it would be a good model
>]of the topology of the stations. Different maps of this city will all
>]refer to different aspects, such as the water-map, electricity
>]distribution, or surface roads. They may all distort certain
>]properties of the region they describe. Individually, each map
>]emphasizes another feature of the same region, and we combine
>]them to achieve a better understanding.
>If the models are all distortions of the reality, in what sense to
>they give an "understanding" of the reality?
Look, the text gives an example. The subway map of London distorts
angles, distances etc.. Yet they give an understanding of the
topological relations among stations. For instance Piccadilly Square
is between Leicester Square and Green park etc...They give an
understanding for the purpose in hand, in this example to navigate in
London using the Tube (Metro).
>And before they can even be distortions, there has to be some form of
>similarity to appeal to. How could a mental image or a mathematical
>model be "like" a real physical process? What features could they
>have in common in the first place?
>]The epistemic cycle described here is the only method of acquiring
>]original new knowledge. Other methods suffer from the defect of
>]not being verifiable.
>How is your method verifiable? How do you verify that your model has
>any relationship at all to the underlying reality?
If we go back to the example of the map, anybody who has the map can
go and see for themselves that Piccadilly Circle is indeed between
Leicester and Green Park, and can navigate London using the Tube.
>--
> David Gudeman
>gud...@cs.arizona.edu
Orhan Yenen
David Gudeman
]gud...@cs.arizona.edu (David Gudeman) writes:
]>]Following Thomas Brody (The Philosophy Behind Physics,
]>]Springer Verlag, Berlin, 1993), we start with few assumptions (or
]>]axioms or whatever). (A philosopher may call this "a naive
]>]realist's point of view/ontology"). If you do not accept them fine,
]>]do not read the rest. However, without these assumptions we
]>]cannot talk about sciences, let alone the philosophy of sciences.
]>This last sentence is demonstrably false. In fact it is trivial to
]>come up with counter-examples.
]May be. What are those trivial counter-examples?
Well, I find it impossible to think of an example of science that I
_can't_ do without assuming physical realism. So my counter-examples
are (1) all of physics (2) all of chemistry (3) all of the life
sciences, (4) all of the social sciences. You could narrow the field
by explaining what it is about science that you think requires
physical realism.
]>]And the biggest argument for a realist ontology is that science
]>]makes correct predictions more often than one could expect on the
]>]basis of mere chance...
]>If that is the biggest argument for realism, then realism is in big
]>trouble. The argument does not support realism at all. An
]>anti-realist ontology does not require that things happen by "mere
]>chance", and a realist ontology does not prohibit the possibility that
]>things happend by "mere chance". Therefore the question of realism is
]>completely independent of the question of randomness, and an argument
]>against randomness can give no support at all to the belief in
]>realism.
]No no, that is not what I meant. What I said is that
]predictions on a random basis would not have the same success as
]the predictions based on sciences. Am I missing what you said?
I grant that. What I said is that this observation does not support
a belief in physical realism.
]>Just what parts of the reality are "outside" us? Sounds? Colors?
]>Shapes? Textures? Sizes? Distances? Temperatures? Changes? Are
]>all of these outside us, or just some of them? On what basis do you
]>decide that some things are "outside"?
]Well, let me throw the ball back to you. What are things which are not
]outside? Shapes, Textures, Distances, Sounds, Colors?...etc?
I'm not denying the existence of things "outside" us, I'm exploring
the significance of your interpretation of this. Personally, I don't
think there is any basis for distinguishing between the above things
on the basis of outside/inside. They all _appear_ to be outside
(which I say puts them outside), but modern scientific theory has it
that color is actually a response of the visual organ to certain
mixtures of radiation. That theory puts color somewhere other than
where it is perceived. If color occurs in the eye (or more
accurately, the brain) is it still outside? Is color different from
shape in this sense? (BTW, I'm not questioning the theory of color,
I'm questioning the metaphysical interpretation given to the theory).
]>]The epistemic cycle begins with a partial interpretation, a guess,
]>]an initial model. Then we interact with a part the system, and note
]>]the changes on the system due to our actions. We improve our
]>]initial guess, model or interpretation until an adequate fit is
]>]obtained with the changes on the system due to our actions.
]>What constitutes an "improvement" to the initial guess? How do you
]>measure the "fit" of a model to a completely unknown quantity?
]I am not sure what you mean here. Can you expand a little?
I mean that you are assuming a model of the universe where there is
an external, objective reality E, and you are trying to model E with a
theory T. You are postulating some sort of similarity relation "=1="
such that for some e in E and t in T, e =1= t (or something similar).
My question is, given that you have no knowledge of E other than
theories in the first place, how can you ever gain any confidence at
all that e =1= t for _any_ e and t? That is, you have no way of
observing e itself, you can only use your senses which give you
sensations s of e, and you are making up t based on s. Clearly there
is a relationship =2= that holds between the theory T and the set of
sensations S (that is, exists s t . s =2= t), but without an a priori
knowledge of the relation =3= that holds between S and E, you have no
way of getting from s =2= t to e =1= t.
]>If the models are all distortions of the reality, in what sense to
]>they give an "understanding" of the reality?
]Look, the text gives an example. The subway map of London distorts
]angles, distances etc.. Yet they give an understanding of the
]topological relations among stations. For instance Piccadilly Square
]is between Leicester Square and Green park etc...They give an
]understanding for the purpose in hand, in this example to navigate in
]London using the Tube (Metro).
But we already know what the relation is between the map and the city,
so the example is more of an analogy, and a flawed one. Yes, you can
verify that a map has the same topology as the city, but you can
observe both the map and the city in the same way and make similar
observations that apply to both of them. You can clearly explain what
sorts of relationships hold between them. There is no guessing.
How do you do anything similar in science?
--
David Gudeman
gud...@cs.arizona.edu
orhan yenen
>In article <30cift$2...@crcnis1.unl.edu> ye...@unlinfo.unl.edu (orhan yenen) writes:
>]gud...@cs.arizona.edu (David Gudeman) writes:
>]>]Following Thomas Brody (The Philosophy Behind Physics,
>]>]Springer Verlag, Berlin, 1993), we start with few assumptions (or
>]>]axioms or whatever). (A philosopher may call this "a naive
>]>]realist's point of view/ontology"). If you do not accept them fine,
>]>]do not read the rest. However, without these assumptions we
>]>]cannot talk about sciences, let alone the philosophy of sciences.
>]>This last sentence is demonstrably false. In fact it is trivial to
>]>come up with counter-examples.
>]May be. What are those trivial counter-examples?
>Well, I find it impossible to think of an example of science that I
>_can't_ do without assuming physical realism. So my counter-examples
>are (1) all of physics (2) all of chemistry (3) all of the life
>sciences, (4) all of the social sciences. You could narrow the field
>by explaining what it is about science that you think requires
>physical realism.
Sorry to ask you so many questions, but I do not want to jump to wrong
conclusions without understanding you well. Are you implying that you
can still do sciences while believing solipsism? Or is it something
else? I am obviously as not well versed in metaphysics as you.
If it is solipsism, I know I am, therefore if you are a making of my
own mind, then who cares, everthing is really irrelevant, including
our discussion.
Besides, I said that is an assumption where physicists start.
>]>]And the biggest argument for a realist ontology is that science
>]>]makes correct predictions more often than one could expect on the
>]>]basis of mere chance...
>]>If that is the biggest argument for realism, then realism is in big
>]>trouble. The argument does not support realism at all. An
>]>anti-realist ontology does not require that things happen by "mere
>]>chance", and a realist ontology does not prohibit the possibility that
>]>things happend by "mere chance". Therefore the question of realism is
>]>completely independent of the question of randomness, and an argument
>]>against randomness can give no support at all to the belief in
>]>realism.
>]No no, that is not what I meant. What I said is that
>]predictions on a random basis would not have the same success as
>]the predictions based on sciences. Am I missing what you said?
>I grant that. What I said is that this observation does not support
>a belief in physical realism.
How is that?
>]>Just what parts of the reality are "outside" us? Sounds? Colors?
>]>Shapes? Textures? Sizes? Distances? Temperatures? Changes? Are
>]>all of these outside us, or just some of them? On what basis do you
>]>decide that some things are "outside"?
>]Well, let me throw the ball back to you. What are things which are not
>]outside? Shapes, Textures, Distances, Sounds, Colors?...etc?
>I'm not denying the existence of things "outside" us, I'm exploring
>the significance of your interpretation of this. Personally, I don't
>think there is any basis for distinguishing between the above things
>on the basis of outside/inside. They all _appear_ to be outside
>(which I say puts them outside), but modern scientific theory has it
>that color is actually a response of the visual organ to certain
>mixtures of radiation. That theory puts color somewhere other than
>where it is perceived. If color occurs in the eye (or more
>accurately, the brain) is it still outside? Is color different from
>shape in this sense? (BTW, I'm not questioning the theory of color,
>I'm questioning the metaphysical interpretation given to the theory).
I agree that the differentiation between sense impressions and
represantations is impossible, as Einstein also said. (A. Einstein,
Physics and Reality, in Ideas and Opinions, Crown Publishers, 1954,
p.290) My own understanding is along these lines. I start by assuming
that my sense impressions are due to exterior objects. In other words
I make a mapping (in the mathematical set theory sense)between the senses,
and the objects. The justification for such a mapping is that, that way
we "orient ourselves in the labyrinth of sense impressions" (quotation
belongs to A.Einstein, loc.cit.).
>]>]The epistemic cycle begins with a partial interpretation, a guess,
>]>]an initial model. Then we interact with a part of the system, and note
>]>]the changes on the system due to our actions. We improve our
>]>]initial guess, model or interpretation until an adequate fit is
>]>]obtained with the changes on the system due to our actions.
>]>What constitutes an "improvement" to the initial guess? How do you
>]>measure the "fit" of a model to a completely unknown quantity?
>]I am not sure what you mean here. Can you expand a little?
>I mean that you are assuming a model of the universe where there is
>an external, objective reality E, and you are trying to model E with a
>theory T. You are postulating some sort of similarity relation "=1="
>such that for some e in E and t in T, e =1= t (or something similar).
>My question is, given that you have no knowledge of E other than
>theories in the first place, how can you ever gain any confidence at
>all that e =1= t for _any_ e and t? That is, you have no way of
>observing e itself, you can only use your senses which give you
>sensations s of e, and you are making up t based on s. Clearly there
>is a relationship =2= that holds between the theory T and the set of
>sensations S (that is, exists s t . s =2= t), but without an a priori
>knowledge of the relation =3= that holds between S and E, you have no
>way of getting from s =2= t to e =1= t.
I guess you are close. First I do not try to model E with a theory T,
because it may not be unique. I deal with e in E, and for any e in E
my sensations defines what e is. So e and s are actually equivalent.
If I have s =2= t, that means I know e =2= t.
>]>If the models are all distortions of the reality, in what sense to
>]>they give an "understanding" of the reality?
>]Look, the text gives an example. The subway map of London distorts
>]angles, distances etc.. Yet they give an understanding of the
>]topological relations among stations. For instance Piccadilly Square
>]is between Leicester Square and Green park etc...They give an
>]understanding for the purpose in hand, in this example to navigate in
>]London using the Tube (Metro).
>But we already know what the relation is between the map and the city,
>so the example is more of an analogy, and a flawed one. Yes, you can
>verify that a map has the same topology as the city, but you can
>observe both the map and the city in the same way and make similar
Not necessarily, for exemple a blind person would observe the city and
the map differently, the two observations would not be identical. In
the city case he/she may use also her hearing.
>observations that apply to both of them. You can clearly explain what
>sorts of relationships hold between them. There is no guessing.
>How do you do anything similar in science?
Once your model is made, there is no more guessing for the user in
sciences either. The guessing is during the making of the map. You, for
instance, may guess that Piccadilly is between..., then go and see for
yourself. If your initial guess is wrong you improve on it. In the
example of London's map, the improvements are discontinuous, and
binary. But as I said, this is an analogy (my words), and a flawed
one (your words).
David Dhuyvetter
>No no, that is not what I meant. What I said is that
>predictions on a random basis would not have the same success as
>the predictions based on sciences. Am I missing what you said?
How do we verify your claim? Controlled experiment? That would seem to be
cheating. If we have to use the methods of science to verify the basis of
science then we should be able to use the methods of religion to verify the
basis of religion. Any such argument, however, would, I suspect, go over like
a lead balloon.
It would be much simpler if we just left ontology out of the discussion. You
don't have to be a realist to understand/appreciate/use science. It is
exactly when science is used to prop up someone's ontology that the claims
that science is a religion become true. I much prefer science as a dance. It
may or may not correspond to some independent "reality," but it is beautiful
to watch regardless.
>Well, let me throw the ball back to you. What are things which are not
>outside? Shapes, Textures, Distances, Sounds, Colors?...etc?
Wasn't "the outside" presented as an axiom? Who cares if it "really" exists?
Lets look at the neat things that can be done based on the assumption that
there is an "outside." If someone comes along with other neat things that can
be done based on different assumptions, I want to look at those too, but I
won't claim that those things are science.
>>]The epistemic cycle described here is the only method of acquiring
>>]original new knowledge. Other methods suffer from the defect of
>>]not being verifiable.
Yow! OK. For this discussion "knowledge" is shorthand for "knowledge within
the system of science?" Then this is reasonable. Shorthand like this,
however, leads people to think that science == ontology == religion.
>If we go back to the example of the map, anybody who has the map can
>go and see for themselves that Piccadilly Circle is indeed between
>Leicester and Green Park, and can navigate London using the Tube.
Ummmm... Here's where your analogy runs into trouble. Lots of people get
lost while following a map. In fact some people who can't handle maps can
navigate just fine with other methods. (Men don't use maps or ask directions
and they sometimes get where they are going. ;) )
David
-----
David Dhuyvetter
david.t.d...@sandiegoca.ncr.com
David Gudeman
]gud...@cs.arizona.edu (David Gudeman) writes:
]>In article <30cift$2...@crcnis1.unl.edu> ye...@unlinfo.unl.edu (orhan yenen) writes:
]>]gud...@cs.arizona.edu (David Gudeman) writes:
]>]>]Following Thomas Brody (The Philosophy Behind Physics,
]>]>]Springer Verlag, Berlin, 1993), we start with few assumptions (or
]>]>]axioms or whatever). (A philosopher may call this "a naive
]>]>]realist's point of view/ontology"). If you do not accept them fine,
]>]>]do not read the rest. However, without these assumptions we
]>]>]cannot talk about sciences, let alone the philosophy of sciences.
]>]>This last sentence is demonstrably false. In fact it is trivial to
]>]>come up with counter-examples.
]>]May be. What are those trivial counter-examples?
]>Well, I find it impossible to think of an example of science that I
]>_can't_ do without assuming physical realism. So my counter-examples
]>are (1) all of physics (2) all of chemistry (3) all of the life
]>sciences, (4) all of the social sciences. You could narrow the field
]>by explaining what it is about science that you think requires
]>physical realism.
]Sorry to ask you so many questions, but I do not want to jump to wrong
]conclusions without understanding you well. Are you implying that you
]can still do sciences while believing solipsism? Or is it something
]else? I am obviously as not well versed in metaphysics as you.
]If it is solipsism, I know I am, therefore if you are a making of my
]own mind, then who cares, everthing is really irrelevant, including
]our discussion.
I was not implying that a solipsist can do science, but it is true.
And I don't know why it would be any less relevant to the solipsist
than the science of a realist is to a realist. A solipsist is just as
interested in not having a bridge collapse as a realist is. The only
difference is that the solipsist denies any distinction between his
experience of the collapsing bridge and the event of the collapsing
bridge. But just because he thinks that his experience is identical
to reality, that does not in any way lessen the painful consequences
that are experienced.
What I was getting at is that science is done by analyzing and
organizing experiences. It makes no difference what you consider the
experiences to be experiences _of_. You might be firmly convinced
that your entire life is nothing but a dream, and that "real" reality
and the "real" laws of nature are nothing like what you experience in
your life. Yet that has no influence on the benefits that come from
using scientific methodology to learn about the dream universe. It
does not matter whether you think you are doing experiments on dream
substances or "real" subtances, what matters is that you can come up
with laws that describe the behaviors of the substances that _you_
experience, regardless of their "true" nature.
]Besides, I said that is an assumption where physicists start.
It is an assumption where many physicists start.
]>]>]And the biggest argument for a realist ontology is that science
]>]>]makes correct predictions more often than one could expect on the
]>]>]basis of mere chance...
]>]>If that is the biggest argument for realism, then realism is in big
]>]>trouble. The argument does not support realism at all. An
]>]>anti-realist ontology does not require that things happen by "mere
]>]>chance", and a realist ontology does not prohibit the possibility that
]>]>things happend by "mere chance". Therefore the question of realism is
]>]>completely independent of the question of randomness, and an argument
]>]>against randomness can give no support at all to the belief in
]>]>realism.
]>]No no, that is not what I meant. What I said is that
]>]predictions on a random basis would not have the same success as
]>]the predictions based on sciences. Am I missing what you said?
]>I grant that. What I said is that this observation does not support
]>a belief in physical realism.
]How is that?
I explained above. Realism is neither necessary nor sufficent for
predictability of physical events. Since the two are logically
independent, the truth of one cannot give support of the truth of the
other.
]I agree that the differentiation between sense impressions and
]represantations is impossible, as Einstein also said. (A. Einstein,
]Physics and Reality, in Ideas and Opinions, Crown Publishers, 1954,
]p.290) My own understanding is along these lines. I start by assuming
]that my sense impressions are due to exterior objects. In other words
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This seems to contradict something you say below...
]>I mean that you are assuming a model of the universe where there is
]>an external, objective reality E, and you are trying to model E with a
]>theory T. You are postulating some sort of similarity relation "=1="
]>such that for some e in E and t in T, e =1= t (or something similar).
]>My question is, given that you have no knowledge of E other than
]>theories in the first place, how can you ever gain any confidence at
]>all that e =1= t for _any_ e and t? That is, you have no way of
]>observing e itself, you can only use your senses which give you
]>sensations s of e, and you are making up t based on s. Clearly there
]>is a relationship =2= that holds between the theory T and the set of
]>sensations S (that is, exists s t . s =2= t), but without an a priori
]>knowledge of the relation =3= that holds between S and E, you have no
]>way of getting from s =2= t to e =1= t.
]I guess you are close. First I do not try to model E with a theory T,
]because it may not be unique. I deal with e in E, and for any e in E
]my sensations defines what e is. So e and s are actually equivalent.
]If I have s =2= t, that means I know e =2= t.
I'm having trouble making any sense of this. Of course T is not
unique. You may have many different theories by which you model the
reality. But the e's I'm talking about are things like planets and
pendulums. Things that scientists study. You can't deal with these
things "directly" because you can't put a planet or a pendulum into an
equation. You can't put it down on paper, it exists separately from
your representation of it. You can only deal with representations of
these objects, and for that matter, you can only deal with
representative idealizations of these objects because you cannot know
all there is to know about the object. Therefore you often deal with
point-like planets or pendulums with frictionless pivots.
And when you make any general (as opposed to a particular) statement
about planets or pendulums, you are giving a theory T. The objects of
the theory are not real planets or pendulums, they are
generalizations. That is, your theory T is not about a _particular_
planet or pendulum, it is about planets or pendulums in general.
By claiming that s is the same as e, you are contradicting your
earlier statement that "sense impressions are _due to_ exterior
objects" (emphasis mine). Clearly there is a big difference between
that and the statement that sense impressions _are_ exterior objects.
Consider the consequences of saying that a sensation is identical to
reality. Consider looking at a penny, lying flat on the table. As
you change your viewing position, your sensation of the penny changes.
The sensation changes both size and shape, and possible color
(depending on lighting conditions). Is the penny changing in size,
shape, and optical properties? If you answer "yes", then either (1)
you are denying the external reality of the penny, or (2) you are
postulating a world that is very different from the one that physic
usually assumes. If you answer "no", then you have to admit that the
sensation is not identical to the penny.
The question then becomes: what is the relationship between the
sensation of the penny and the externally real penny? It is this
relationship that you have no information about because _all_ of the
information you have is your sensations. You have no direct
information about the penny itself.
--
David Gudeman
gud...@cs.arizona.edu