|
| Group: |
sci.philosophy.meta |
[ Post new message ] |
|
|
 |
| Author: |
Haines Brown |
| Subject: |
Re: Universality as warrant for relative truth value |
| Body: |
Neil W Rickert <rickert+nn@cs.niu.edu> writes:
> Haines Brown <brownh@teufel.hartford-hwp.com> writes: >Neil W Rickert
><rickert+nn@cs.niu.edu> writes:
>
> Lots of snippage. I only quote enough to provide context for my
> responses.
Thank you.
>>Your paragraph puzzles me. Are in you fact saying that theories lack
>>truth value, and that this lack of truth value is not only
>>conventional wisdom, but also common sense?
>
> I am saying that the conventional wisdom takes theories to be true. I
> don't think we can generalize about common sense, because it does not
> seem to be common. That is, what one person derives from common sense
> can be different from what another derives. Or, to say it
> differently, common sense is common in that it is not rare - just
> about everybody exhibits common sense. But it is not common in the
> sense of being shared (something held in common by all).
Obviously I offered a hypothesis that was speculative and would be hard
to prove. Merely intuition. I suspect one consideration is how broadly
we use the word "theory". I was using it very loosely. My wife and I are
following a car on the highway late at night that is weaving. I say,
"That driver must be drunk". That's a hypothesis. Suddenly there's an
accident and reading about it in the paper the next day, the police
found that the driver was in fact intoxicated. Aha, I say, my suspicion
turned out to be true. I suppose my hypothesis was common sense, for
that what I saw most often corresponds to the kind of driving that we
associate with driving while under the influence.
> Suppose I develop a theory of automobile driving. Part of that theory
> will be the rule that we drive on the right side of the road (except
> for one-way roads).
But we do differ here about a theory. I take theory to be necessarily
explanatory. Your example, however, is descriptive, citing one of the
conditions (the rule) under which driving is done. If you had said that
in the US we generally drive on the right and suggest that this is
because of the law, then we have an explanatory theory.
> If the British were to develop a theory it would be similar, except
> that it would say that we drive on the left side of the road.
>
> If theories have truth value, which of those theories is true - the
> British one, or ours?
Actually, the explanatory theory would be the same, only the specific
conditions differ.
> My view is that theories are pragmatic constructs, and that saying the
> theory is true adds nothing (though it might be good marketing
> strategy). What matters is that practice be consistent with the
> theory.
If I said 1+1-3, that would be a false statement; if I said 1+1-2, that
would be true. Yes, these are not theories, but are an example where the
two statements have a property of truth value. There's nothing pragmatic
about this example, for we are not talking about oranges, but about
math. If we remove truth value from these two examples, we would end up
with mathematical anarchy.
A pragmatic test of truth seems to have several problems. First, it is
deductive, and deductive logic applies to precious few situations (the
lab, not the real world). In the real world, outcomes are not
unequivocally predictable from knowledge of an initial state. How can
one use a pragmatic test for an emergent system? Another problem is that
the test of an outcome is based on its observation and therefore
excludes such unobservables such as causal powers. For example, the
structure of a social system might be such that the real power of an
interest to shape the course of history is systemically blocked, so that
the observed outcome does not correspond to the observed initial
state. That is, the pragmatic test is empiricist and subject to all the
criticisms launched at empiricism. Another problem, is that it is well
known that for any given set of phenomena, there is a range of possible
explanations, and so which of them is validated by the observed outcome?
Of course, you might infer from that outcome which of the possible
theories was more likely to be correct, but then we are back to relative
truth value, and what is being tested is not the truth of the correct
theory, but a choice among competing theories. There is also an
objection that the pragmatic test can point to undesirable outcomes. In
Germany of 1943, one would have to say that Hitler was on the right
track. One could not infer from the situation at that time that his
entire project was fundamentally flawed because at that time he was
enjoying success. Another way of putting this is that it fixates on the
actual past and downplays the role of innovation; might makes right; it
is profoundly conservative. There are many other objections to a
pragmatic test for truth.
>>My dictionary definition for "theory" uses the example of the general
>>principles behind musical practice. So are you suggesting that it is
>>obvious on the face of it that the general principles of music lack
>>truth value in that music has no general principles?
>
> I certainly think that music has no general principles. But you will
> probably misunderstand that. By way of comparison, I also have been
> known to say that natural languages have no grammar, and that there
> are no laws of nature.
I take a course in "musical theory". What am I studying, then, if not
the general principles of music? What are scales, pitches, rhythm,
melodies, chords, chord progressions, phrases? It is obvious to any
musician that without some music theory, nobody will progress from
playing music to understanding and even writing it. What can you
possibly mean that music has no general principles? There is enormous
number of such principles and they are of considerable complexity.
By the way, not that in this example, theory is not really explanatory,
but comes closer to your rules of the road. It is often said that
scientific theory is necessarily explanatory, not descriptive, but I'm
not entirely confident this is true.
>>Doesn't common sense suggest that there are general principles in
>>music to which the term musical theory refers?
>
> That's an entirely different question, unrelated to whether music has
> general principles, though people often take them to be related.
I don't see that it's a different question unless by the word "theory"
you mean that its content must be coherent or organized (related). That
may not be true. My theory that the driver was drunk was ad hoc,
immediately and narrowly causal. In evolutionary sciences (meteorology,
cosmology, evolutionary biology, etc.), explanations are ad hoc and do
not rely primarily on covering law explanation. An explanation takes
into account boundary or background conditions and a multiplicity of
causal influences. We encounter not only the N-Body Problem, but the
fact that all things in principle are open processes, and so any theory
must be equivocal, fuzzy. Added to this, take, for example, the theory
of a thermodynamic engine: it involves the interdependence of opposite
processes, and so the theory must take unobservables into account if it
is to gain any coherence, but such coherence is non-empirical.
> Humans organize their world. In organizing our world of music, we
> introduce principles. I am saying that those are human principles for
> how we should organize music (or our study of music), and that those
> principles are not a natural property of music itself, independent of
> our organizing.
Music is, of course, unnatural to begin with. Do the principles of music
reduce to human whim? Of course not. It is the exercise of freedom
within constraints imposed by sound and the ear.
> Getting back to scientific theories, let's use the example of special
> relativity that you mentioned. SR contains or implies the statement
> E=mc^2. I am not at all questioning whether that statement is true,
> or has a truth value. It is the theory as a whole that I say is
> neither true nor false, hence has no truth value.
>
> The theory introduces terminology, and determines the meaning of the
> terms introduced. It also sets standards for evaluating the truth of
> statements made that use the terminology. Under the standards of
> truth set by SR (as a theory), E=mc^2 turns out to be a true
> statement.
>
> As I see it, SR (as a theory) is not making statements about the
> world. Rather, it is making statements about its own terminology,
> about how to use that terminology - roughly speaking, it is making
> statements about empirical practices to be followed when using the
> theory.
Sorry, I don't follow, You accept the truth of a formula that expresses
the relation of some variables and a constant. Then STR has a formula,
the Lorentz Transformation, where γ = 1/(1-(v²/c²)). Just another
formula with two variables and one constant. What's the difference. You
accept the first as a theory, but not the second. Why?
STR is conventionally taken to be statements about the world. I'll bet
99% of all scientists would agree. So why take the doubts of a
philosopher seriously? Is it that philosophers have something to say
about science? Well, of course they do, but these philosophers _of
science_ don't have the problem you mention.
I would take philosophers of science a lot more seriously than
philosophers outside the sciences (Rorty?). I'd rather follow Tarski
(his 1933 paper "one of the two most important developments in
mathematical logic in the first half of [the 20th] century". He provided
a linguistic analysis for what came to be known as the correspondence
theory of truth (the relation to Tarski's position is a bit contested)
to show that correspondence did not entail some questionable
metaphysics. The correspondence interpretation of Tarski's theory
underwrites the naturalistic conceptions of both realists and
empiricists.
Now a linguistic critique of the correspondence theory of truth is
certainly possible, Why should we consider it? a) contradictions in the
correspondence theory, b) advantages in the alternative theory. We would
be wise to employ these tests for our own private preferences, but we
might also have private reasons to prefer a theory. This is fine, but
when we engage others and present our alternative view, assuming that is
is not a conventional view within the discipline of which we speak
(science in this case), we must always justify our position by 1)
providing arguments in its favor that refer to considerations (a) and
(b), or by 2) citing a well recognized authority in the field. Rorty's
expertise lies in literature, humanities and philosophy, not in science
or the philosophy of science. We ought to read him for our private
amusement or edification, but we surely can't present his theory in the
context of science without sufficient justification.
I have the pleasure of holding some off-the-wall views, but when I
trundle them out in a context other than where such views are
conventional, I must always do so in the language and with the values of
the milieu I address. If for some reason I can't manage that, such as my
ignorance, my injecting my off-the-wall views will only invite
indifference or ridicule. It will have no rhetorical effect whatsoever,
which after all, is what is required of dialog.
> I think my view is not too different from that expressed by
> C.I. Lewis, "A Pragmatic Conception of the A Priori" (a 1923 article
> in J. of Philosophy, and reprinted in the book "A Priori Knowledge
> (ed. Paul K. Moser)).
Interesting, I don't know this article, nor have I ever read anything by
Lewis. When he had influence after World War II, I was studying science,
but had no interest in the philosophy of science, and so missed him. His
reputation apparently declined precipitously after the War, and for your
own development, you might look in to why. I suspect Carl Hempel took
him seriously, and might explore Hempel a bit.
>>The final thing I find strange about your paragraph is the suggestion
>>that I should not swallow conventional wisdom. Why not?
>
> You should question it. You might still follow it, even though you
> question it. But questioning is good.
Undoubtedly. My ability to launch a viable critique of anything is very
limited. There's a (narrow) field in which I can speak with some
assurance; there's a few fields were at least I hopefully would not
appear the fool, but in most fields I'm profoundly handicapped. When I'm
looking into these fields, I'm obliged to accept the current consensus
as probably true, for doing so enables me to learn a bit. For me to
raise a question can't be to challenge others in a field in which I lack
expertise, but only to learn more. So I follow conventional wisdom
except where I see problems with it and think I have something
substantial to offer as an alternative.
> What I disagree with is the view often seen in theological and
> philosophical literature, that seems to treat truth as something that
> is completely human independent, that we just have to somehow
> discover.
Gosh, I don't know anyone who would say that. It is certainly not the
prevailing notion of truth in the sciences. Positivism perhaps, but
positivism is dead. Actually, I'm not sure if positivism suggested that
truth can be or is entirely objective.
>>> I believe that correspondence theory could not possibly be a
>>> definition of "truth".
>
>>As I showed, it is in fact presumed by the dictionary definition.
>
> Dictionary definitions (so called) don't actually define. I expect a
> definition to give some explicit principle that can be applied.
> Otherwise it is not a definition. Perhaps my idea of "definition" is
> influenced by my being a mathematician.
News to me. I thought dictionaries defined words. True, the definitions
tend to be empiricist or functionalist, but that's another issue. I take
a dictionary to define words as they are commonly used. Why should a
definition engage some principle?
I pick this up from that great font of wisdom, Wikipedia:
There are two types of definitions: A descriptive definition provides
to a term a meaning which is in general use. A stipulative definition
of a term carries a meaning which a speaker wants it to convey for the
purpose of his or her discourse. Thus, the term may be new, or a
stipulative definition may prescribe a new meaning to a term which is
already in use. A descriptive definition can be shown to be "right"
or "wrong" by comparison to usage, but a stipulative definition
cannot.
What is the mathematician's use of the word "definition"?
> Berkeley's theory of reality was idealism. According to Berkeley,
> reality exists only as ideas in the mind, and there is no physical
> existence. I am not understanding why Berkeley's ontology would be
> significantly different from the ontology of a realist.
I presume because the realist says that theoretical objects,
unobservables (and also observables) are ontologically real, and
Berkeley denies it.
>>Sigh ;-). I don't understand what you mean by saying that truth does
>>not matter when it comes to scientific theories. It obviously means a
>>lot to those who propose them, for they think of themselves in a quest
>>for truth.
>
> Many scientists consider themselves pragmatists, rather than truth
> seekers.
True, many pragmatists seem agnostic about the truth of their
explanations. But, assuming them not to be insane, are they not
nevertheless looking for truth, if only, "Eureka, it works!" That is, I
doubt any sane scientist feels the aim of his scientific career is to
write fiction.
> For sure, scientists want to make truthful statements about the world.
> But what does that have to do with the truth of the scientific
> theories themselves?
OK. I concede the point. A pragmatist is concerned with truth, although
agnostic about the truth value of his theories. Their truth is only that
they allow prediction, they work.
>>Who (besides yourself) says the gas laws are false?
>
> I think most physicists would agree with me on that.
Not in a commonsensical manner, I suspect. Again, resorting to
Wikipedia:
The gas laws are a set of laws that describe the relationship between
thermodynamic temperature (T), absolute pressure (P) and volume (V) of
gases. They are a loose collection of rules developed between the late
Renaissance and early 19th century.
Do you mean the laws are idealizations (represent a boundary case)? I
can think of no other reason why one might cast doubt on them. But are
ideal boundary cases "false"? In integral calculus, the whole thing is
ideal, of course, but it can entail boundary definitions, and there's no
suggestion that I know of that a solution for them is in error.
> But why call it "transmission of knowledge", when it is information,
> not knowledge, that is being transmitted?
Well, yes, it is certainly a transmission of information. But when I
inform you that the moon consists of green cheese, I am offering a
statement that constrains the properties of the moon. Or we can think of
information as the entropy state of a closed system. If I convey to you
the information that the moon consists of cheese or that a given system
has a certain information entropy, I'm conveying a fact having truth
value that can be empirically tested. Knowledge is a practical or
theoretical understanding of a subject, what is know of it, facts
concerning it. These also have truth value. While the two terms are not
quite the same, they are so broad and diverse in meaning that I don't
know that we can sharply distinguish them without imposing our favorite
definitions. If I >>make a statement that everyone, including myself,
know to be false, but >>actually turns out by some fluke to correspond
with reality, wouldn't >>that be a true statement which is a property
intrinsic to statements and >>independent of our beliefs about those
statements.
>
> That's a tricky question. Making a statement is an intentional
> action. If your intention was to make a false statement, but what you
> said was actually true, perhaps that would make your statement doubly
> false in the sense that it didn't even match your intentions in making
> that statement.
That's true, but not "doubly". My intentions are incidental because most
people usually listen to what I say rather than read my mind (it may be
different in a court of law where my credibility as a witness is
challenged). You start out with a presumption that truth lies in
language, not the relation of language and world, but you don't offer
any reason why I should accept that presumption myself. Because you do
not, my own bias stands.
>>Not at all clear here. So let me recall the context. I had originally
>>asked for justifications for the coherence theory suggestion that the
>>potential truth value of a theory is a function of its universality.
>
> Yes, sorry. I don't really have much to say about coherence theories,
> other than that I am skeptical of them.
Why skeptical? You strike me as naturally sympathetic to such
theories. People use coherence theories all the time, but I have trouble
finding any developed defense of them. Some time ago I asked about the
meaning of "robust" in this group, and got no useful replies. I'll have
to look into this theory (about which I'm ignorance) more deeply as time
allows, for it is clearly important and much used.
> The broader sense of "proof" is that of a persuasive argument. But
> whether a person is persuaded will vary from person to person. You
> can say what you want about justification, but unless it persuades me,
> I won't count it.
Of course.
>>> But it cannot do either unless you have first developed a notion of
>>> correspondence that is independent of your concept of truth.
>
>>This seems circular. If I define truth as the relation of statements
>>and the world where the statement corresponds to that reality in some
>>sense, then where's the problem?
>
> It is so vague as to not be definitive, and therefore cannot be a
> definition.
Ah, so that's your problem with this. But because there are different
ways to assess coherence surely does mean they all must be false
;-). Some ways are quite precise and specific. For example, a great deal
has been done with model theory. It gives me a headache, but at least it
aims at precision.
>>You seem to say that to define truth presumes the existence of a
>>transcendental Truth, where in fact truth is only a label we invent to
>>point to an attribute of the statement of the relationship between
>>specific statements and their referents.
>
> I have no problem with truth as a relation between statements and
> their referents. But I cannot find that in the correspondence theory.
?? That's the definition of correspondence theory. Let me give a simple
minded example (of congruence correspondence) borrowed from Wikipedia:
The cat is on the mat. Congruence means that our words refer to the real
entities of cat and mat, and their relation is what I state it to
be. I'm not advocating this (I'm a process realist and don't believe our
the referents of our statements can be reduced to objects of
observation), but the issue is whether the posited correspondence is at
all vague. The statement posits the existence of a cat and a mat, and we
can in fact observe them; it posits their relation, and that too we can
clearly perceive. So there is a correspondence.
> There is a substantial literature on truth, that attempts to define it
> as correspondence, and which presupposes that we have already settled
> questions about reference. And then there is another substantial
> literature on reference, that attempts to define reference in terms of
> an already presumed understanding of truth. I am simply pointing out
> that the emperor has no clothes. The whole thing is completely
> circular, and actually tells us nothing about either truth or
> reference.
Understood. When we have a correspondence theory, our statement posits
the existence of things or their structure, and the issue is, these are
being posited as truths, when in fact truth is what we are trying to
ascertain.
I can think of objections to this proposal:
a) It seems to rest on the proposition that the word "cat" in my example
statement is equivalent to the statement "There is a cat". However, it
is not. The word cat is an index, comparable to my pointing to it. To
say "is on" refers to a kind of relationship, not a statement of
fact. It is only their assembly into the statement that we have
something that we can compare with the world observed. True, both the
indices and the relation term involve low level observational or
ontological theories, so let me turn next to that.
b) To refer to the cat and mat and to posit their relation (or even if
we feel my point (a) is entirely incorrect about these terms not being
truth statements, there remains that our entire existence depends on the
relative truthfulness of much of the knowledge we acquired from society
or through personal experience. If the totality of our knowledge was in
fact randomly either true or false, we would effectively be insane. I'll
assume this is intuitively obvious. In other words, we have an
existential need for a presumably largely true fund of knowledge. If
these knowledges are embedded in my statement about the cat and mat, we
would have to admit that the truth of it has a conventionalist
foundation, but we can't do without it.
c) When we find a correspondence between our statement and the world, we
are not calling for any absolute or unequivocal truth, for an
approximate and rough an ready truth will do. If you object to my
example that the cat is not really on the mat, for there are atoms that
separate them, my reply would be, "You quibble, Sir". We are not seeking
absolute certainty, as in math, but a rough sense that the
correspondence is more truthful than not.
While I see much utility in a correspondence (and coherence) theory of
truth, let me remind you that I privilege another theory of truth, which
is centered on an action theory (not pragmatism). It is kind of like
Lakatos' position, but in many respects not.
> Now take a scientific theory. But don't take it as just a
> self-contained set of propositions. Instead, look at it in the
> context of science. That context includes the laboratory training
> that students of science take, and where they learn the practices of
> the particular science. I think you will see that a scientific
> theory, looked at in this manner, is actually concerned with defining
> terms and establishing reference for those terms. If you look at
> special relativity, the purpose of the thought experiments that
> Einstein gave was to help establish reference, as it was to be used
> under SR. Looked at that way, you will see that the role of a
> scientific theory is to setup the reference, such that it will then
> become possible to have a truth relation between statements and their
> referents. The theory is prior to being able to make the statements
> for which we want there to be an attribute of truth.
Don't follow this. My (dim) recollection of STR was that there were
contradictions in the received theory, and the Michelson-Morley
experiment forced them to the surface, and Einstein employed abductive
reasoning to generate a hypothesis which could resolve the
contradiction. But you need to explicate your general point above before
I can quite grasp it. It is true that the lab is an artificial construct
designed to yield the very artificial (one-sided) truth we call general
laws. But the positivist lab is hardly a model of normal scientific
practice, and more than is deductive reasoning the principle logic
(despite what kids are taught in school, for it had ideological utility
having little to do with science).
> It requires a transcendental notion of truth, only when it depends on
> a transcendental notion of correspondence. And the literature on
> correspondence theory mostly seems to be dependent on a transcendental
> notion of correspondence.
This comes as a surprise to me on both counts. Does this, for example
apply to my example of the cat being on the mat? The notion that
anything is transcendental strikes me as really bizarre (I hold to
materialistic monism, like nearly everyone else).
> Sure, the ruler does not define "true", but it does define reference
> (or this particular case of reference). My objection to the
> correspondence theory is that you cannot define truth in terms of
> reference, if reference is itself undefined. And once reference is
> defined, that implicitly defines truth without a separate theory or
> definition of truth being required.
And why can't we define truth in relation to this artificial standard of
the ruler. There is a correspondence between the height of the desk and
the number of units on the ruler. The ruler is not a truth any more than
the desk. By statement that the desk is 30" high remains relative to the
man-made ruler, but it is nevertheless true. If I told you that my desk
is 30" high, it would be bizarre to replay that this is not so because a
ruler is not transcendental. The reference here, the ruler, is defined
as a social convention, and the object of the ruler does not deconstruct
simply because its markings are a social convention. If I say, this is a
ruler", that of course is a truth statement, and so my measurement of my
desk entails auxiliary beliefs, but these beliefs I hold as true because
they are a condition for my taking the measure, and the utility of a
ruler for this is transparent and uncontested.
> Okay. But here you are establishing correspondence with a social
> convention (the use of a measuring stick). What does that have to do
> with correspondence with reality?
Aha! I've been using the word "world" here without troubling to define
it. The world is everything outside consciousness: the moon, breakfast,
the ruler, my brain. A social convention is real, is it not? Try
violating them and you get into trouble.
> The conventional wisdom sees science and knowledge as having to do
> with determining whether propositions are true. I see science and
> knowledge as having to do with finding ways of forming propositions,
> which could not even have existed without the kind of activity that
> humans engage in.
And how would you justify such a position? I don't mean to be cruel, but
an alternative perspective than the conventional one used by scientists
is of absolutely no interest or use (except mildly as a kind of
intellectual aphrodisiac) to anyone but yourself. The value of your
alternative perspective lies _entirely_ in its justification, such as
showing unequivocally, and in terms of the language understood by
scientists, that your perspective is obviously better than the
convention.
I apologize for the haste with which I am forced to reply to you.
--
Haines Brown, KB1GRM
|
|
|
|
| Topic: |
Universality as warrant for relative truth value |
|
| Message: |
Author |
Date |
|
|
*Message 1* |
Haines Brown |
Tue, 15 Jul 2008, 5:07 pm |
|