Happy New Year, Tony
Bill
laid on me over Hume and induction earlier this year. You can drub me
again without breaking you New Year's resolution. I enjoyed being
trounced so much, that I want to relive the experience. In brief...
You and I and Hume agree that there is no logical foundation for
induction.
You conclude from this that induction can give you no valid knowledge
about the world.
Hume and I conclude that our (inductively derived) knowledge of the
world, while quite obviously valid, is not based on logical deduction.
[The above is my recapitulation of the drubbing you drubbed me
earlier]
Based on induction, I suspect that if you take the time to try to find
support for your position from Hume's writings, that you will cite
places in which he clearly shows that there is no logical foundation
for induction, and then declare victory. But we all agree that there's
no logical foundation for induction. It will be more difficult for you
to find the claim from Hume that induction does not lead to valid
knowledge of the world - remember, it is you, not he, who thinks that
for knowledge to be valid it should be supported by deductive logic.
[The above is a prediction of your response, assuming you make a
substantive response and don't simply declare victory.]
I think that you dislike induction because it can lead you to wrong
conclusions about the world. Indeed it can. You are uncomfortable with
the possibility of error. That's where you differ from the scientists.
Base your knowledge of the world on a deductive system that is free
from the possibility of error, and you will know nothing whatsoever
about the real world. Base your knowledge of the world on induction,
susceptible as it is to error, and you will know a lot of things, any
one of which might be incorrect. But the longer you work at it the
smaller you can make the errors.
[This is an only semi-serious psychological conjecture about why you
hold the position you do - my inductive prediction is that you very
likely will respond to this bit, and ignore the rest of the post.]
P.S. It's OK with me if you do break your New Year's resolution -
you've drubbed and thrashed and demolished me so much already that I'm
stunned and senseless; a few more choice epithets thrown my way can't
do me any more damage.
T Pagano
<broger...@gmail.com> wrote:
>Merry Christmas, Tony.
.. . .and back at ya.
> I do have warm memories of the drubbing you
>laid on me over Hume and induction earlier this year. You can drub me
>again without breaking you New Year's resolution. I enjoyed being
>trounced so much, that I want to relive the experience.
While I did rub it in, it was one of the more interesting exchanges
I've had in a while. You actually took the time to go into the
original works of Hume. You forced me to do the same.
>In brief...
>
>You and I and Hume agree that there is no logical foundation for
>induction.
Not exactly. You denied that Hume layed the foundation, I produced
the necessary quotes to prove otherwise, and you graciously conceded.
Not only is there no foundation for induction it is logically invalid.
>You conclude from this that induction can give you no valid knowledge
>about the world.
I don't conclude this because I don't know what "valid knowledg" is.
Instead, I argue that every inductive argument is logically invalid.
That is, the conclusion goes beyond what is contained in the premises.
This does not mean that the conclusion is necessarily false only that
we have no way of knowing its truth value from an invalid inductive
argument.
In a valid deductive argument the truth of the premises is transmitted
to its conclusion.
>Hume and I conclude that our (inductively derived) knowledge of the
>world, while quite obviously valid, is not based on logical deduction.
Wrong again. Hume concluded that inductive arguments were invalid and
as such there was no way to evaluate the truth value of any such
conclusion. As you recall Hume also concluded that one could not even
assign some probability of truth to any such conclusion.
>[The above is my recapitulation of the drubbing you drubbed me
>earlier]
>
>Based on induction, I suspect that if you take the time to try to find
>support for your position from Hume's writings, that you will cite
>places in which he clearly shows that there is no logical foundation
>for induction, and then declare victory.
You have a very short memory. You claimed that Hume did not show that
induction was logically invalid. I found the appropriate quotes that
demonstrated otherwise. It was over this dispute that I claimed
victory. I don't claim victory over the facts themselves.
Induction is a logically invalid form of reasoning. Many great
philosophers and logicians have tried---and failed---to remediate
this. Yet even though this reasoning is logically invalid modern
atheists employ it as if it were valid. This is irrational.
> But we all agree that there's
>no logical foundation for induction. It will be more difficult for you
>to find the claim from Hume that induction does not lead to valid
>knowledge of the world - remember, it is you, not he, who thinks that
>for knowledge to be valid it should be supported by deductive logic.
We don't search for "valid" knowledge, we search for "true" knowledge.
"Truth" and "validity" are not synonyms.
Hume's argument addressed the method of reasoning alone. Hume implied
that there is no way to know from the form of an invalid inductive
argument whether or not the conclusion (new knowledge) was true or
even probably true.
We can all make conjectural statements about nature. But we also want
to know which conjectures, conclusions, knowledge are true and discard
those that are false. Hume simply argued that inductive reasoning
offers not the slightest help in making such a determination.
>
>[The above is a prediction of your response, assuming you make a
>substantive response and don't simply declare victory.]
>
>I think that you dislike induction because it can lead you to wrong
>conclusions about the world. Indeed it can. You are uncomfortable with
>the possibility of error.
I dislike it because it leads nowhere. Invalid inductive methodology
misleads people into believing that level of corroborative evidence
makes an inductive conclusion probably true. Hume showed that this
was mistaken and NO ONE has shown Hume wrong.
>That's where you differ from the scientists.
>Base your knowledge of the world on a deductive system that is free
>from the possibility of error, and you will know nothing whatsoever
>about the real world. Base your knowledge of the world on induction,
>susceptible as it is to error, and you will know a lot of things, any
>one of which might be incorrect. But the longer you work at it the
>smaller you can make the errors.
This suffers from the mistake of presuming that genuinely new theories
about the world are necessarily derived inductively. Nothing could be
further from the truth.
>[This is an only semi-serious psychological conjecture about why you
>hold the position you do - my inductive prediction is that you very
>likely will respond to this bit, and ignore the rest of the post.]
You acknowledge that inductive arguments are logically invalid. Yet
you also are of the opinion that this form of reasoning can help you
sort between true and false conclusions. These statements are
inconsistent with one another. To hold them both---as you apparently
do-----makes you irrational.
>P.S. It's OK with me if you do break your New Year's resolution -
>you've drubbed and thrashed and demolished me so much already that I'm
>stunned and senseless; a few more choice epithets thrown my way can't
>do me any more damage.
I don't have to. . .the truth is devastating enough without throwing a
single stone.
Happy New Year,
T Pagano
Bill
> On Mon, 27 Dec 2010 16:40:15 -0800 (PST), Bill
>
> >Merry Christmas, Tony.
>
> .. . .and back at ya.
>
> > I do have warm memories of the drubbing you
> >laid on me over Hume and induction earlier this year. You can drub me
> >again without breaking you New Year's resolution. I enjoyed being
> >trounced so much, that I want to relive the experience.
>
> While I did rub it in, it was one of the more interesting exchanges
> I've had in a while. You actually took the time to go into the
> original works of Hume. You forced me to do the same.
>
> >In brief...
>
> >You and I and Hume agree that there is no logical foundation for
> >induction.
.
>
> Not exactly. You denied that Hume layed the foundation, I produced
> the necessary quotes to prove otherwise, and you graciously conceded.
I denied that Hume laid what foundation? I honestly don't get your
drift here. What I said back in June was "We agree that induction is
not supported by logic. That is indeed what Hume said. No argument
from me there."
>
> Not only is there no foundation for induction it is logically invalid.
>
> >You conclude from this that induction can give you no valid knowledge
> >about the world.
>
> I don't conclude this because I don't know what "valid knowledg" is.
>
> Instead, I argue that every inductive argument is logically invalid.
> That is, the conclusion goes beyond what is contained in the premises.
> This does not mean that the conclusion is necessarily false only that
> we have no way of knowing its truth value from an invalid inductive
> argument.
.
>
> In a valid deductive argument the truth of the premises is transmitted
> to its conclusion.
That's certainly true. However, unless the premises are true by
definition, in which case you are restricted to statements about
mathematics, deduction is useless. How do you propose to get to a
"true premise" about the physical world?
>
> >Hume and I conclude that our (inductively derived) knowledge of the
> >world, while quite obviously valid, is not based on logical deduction.
>
> Wrong again. Hume concluded that inductive arguments were invalid and
> as such there was no way to evaluate the truth value of any such
> conclusion. As you recall Hume also concluded that one could not even
> assign some probability of truth to any such conclusion.
>
> >[The above is my recapitulation of the drubbing you drubbed me
> >earlier]
>
> >Based on induction, I suspect that if you take the time to try to find
> >support for your position from Hume's writings, that you will cite
> >places in which he clearly shows that there is no logical foundation
> >for induction, and then declare victory.
.
>
> You have a very short memory. You claimed that Hume did not show that
> induction was logically invalid. I found the appropriate quotes that
> demonstrated otherwise. It was over this dispute that I claimed
> victory. I don't claim victory over the facts themselves.
You should get a job as a revisionist historian - you'd be great. What
I said back then was "We agree that induction is not supported by
logic. That is indeed what Hume said. No argument from me there." I
agree and have always agreed that induction is logically invalid.
<snip - I've long since stipulated that induction is not logically
justified>
>
> We don't search for "valid" knowledge, we search for "true" knowledge.
> "Truth" and "validity" are not synonyms.
OK, Tony, how do you find "true knowledge?"
>
<snip - but I'll come back to it if there's not enough to talk about
elsewhere.>
.
>
> You acknowledge that inductive arguments are logically invalid. Yet
> you also are of the opinion that this form of reasoning can help you
> sort between true and false conclusions. These statements are
> inconsistent with one another. To hold them both---as you apparently
> do-----makes you irrational.
Now we're getting somewhere. I'm irrational. Exactly Hume's point.
Bingo. There's a reason he called it "A Treatise Concerning Human
Nature." His large scale argument is against those philosophers, like
Plato and Descartes, who saw human nature as fundamentally rational.
Hume argues that the most basic way we learn what we need to know
about the world in order to survive in it, a method which we share
with young children and animals, is completely unsupported by logic.
Our life in the world and all of the knowledge we have about it is
derived from "habit" (as I trust you know he never called it induction
himself). Everything that we know about the world is unsupported by
logic. He did not conclude that we don't know anything or that we
should not trust induction, even to the point of making life and death
decisions based on it; what he concluded is that human nature is not
fundamentally rational. Elsewhere in the Treatise he makes similar
arguments about ethics. Our moral judgments are based not on reason
but on sentiment. He does not conclude that morality is meaningless or
should be disregarded, only that, like knowledge, its basis is non-
rational.
So we're stuck. Induction is clearly not founded on logic. But try
living without it. You can say in words that what induction teaches
you gives you not a hint about the truth of the world, but Hume and I
will challenge you to live 5 minutes without making constant use of
induction. You can minimize the value of "inductive knowledge," but to
downplay the importance of the form of knowledge on which you
constantly depend for your daily survival seems like posturing.
Deduction is not a useful alternative; it can tell you nothing about
the real world, and even in mathematics, where it works best, it is
incomplete.
Friar Broccoli
> Plato and Descartes, who saw human nature as fundamentally rational.
> Hume argues that the most basic way we learn what we need to know
> about the world in order to survive in it, a method which we share
> with young children and animals, is completely unsupported by logic.
> Our life in the world and all of the knowledge we have about it is
> derived from "habit" (as I trust you know he never called it induction
> himself). Everything that we know about the world is unsupported by
> logic.
I have not read Hume or followed the many similar arguments here, but
it seems to me that Hume is wrong and induction is supported by
logic. How else can we get a good approximation of what is true about
the world without observing the world and making predictions based on
what we have repeatedly seen in the past? (I am aware that this last
is your point.)
Since logic cannot be built on anything except induction/habit, how
can Hume argue it is: "unsupported by logic"?
Ernest Major
<a63232fe-b60c-428f...@l8g2000yqh.googlegroups.com>,
Friar Broccoli <eli...@gmail.com> writes
>On Dec 27, 10:40�pm, Bill <brogers31...@gmail.com> wrote:
>
>> Plato and Descartes, who saw human nature as fundamentally rational.
>> Hume argues that the most basic way we learn what we need to know
>> about the world in order to survive in it, a method which we share
>> with young children and animals, is completely unsupported by logic.
>> Our life in the world and all of the knowledge we have about it is
>> derived from "habit" (as I trust you know he never called it induction
>> himself). Everything that we know about the world is unsupported by
>> logic.
>
>I have not read Hume or followed the many similar arguments here, but
>it seems to me that Hume is wrong and induction is supported by
>logic. How else can we get a good approximation of what is true about
>the world without observing the world and making predictions based on
>what we have repeatedly seen in the past? (I am aware that this last
>is your point.)
That is induction being supported by observation, not logic. Being
supported by (deductive) logic would mean that there were axioms from
which you could deduce that past experience is an infallible guide to
future events.
>
>Since logic cannot be built on anything except induction/habit, how
>can Hume argue it is: "unsupported by logic"?
>
--
alias Ernest Major
John S. Wilkins
You are equivocating upon "logic". Of course we use inductive logic:
Hume never denied it. He was making the point that we cannot *justify*
induction through the use of deductive logic, because it requires
assumptions that are not deductive.
This misinterpretation of Hume is based, I think, on a misreport of
Popper's point about falsification, which relies on a deductive claim:
modus tollens (show that a generalisation is false if it is) is all you
can do to an inductive generalisation. Some people think this means that
induction canot work. Certainly Popper was uninterested in it (he has no
logic of discovery, contrary to the English title of his magnum opus),
but Hume was not. He said that reason is not how we form ideas based on
experience, and he was right. But he also said that when we are done
doing philosophy we should continue to live.
--
John S. Wilkins, Associate, Philosophy, University of Sydney
http://evolvingthoughts.net
But al be that he was a philosophre,
Yet hadde he but litel gold in cofre
Bill
> On Dec 27, 10:40 pm, Bill <brogers31...@gmail.com> wrote:
>
> > Plato and Descartes, who saw human nature as fundamentally rational.
> > Hume argues that the most basic way we learn what we need to know
> > about the world in order to survive in it, a method which we share
> > with young children and animals, is completely unsupported by logic.
> > Our life in the world and all of the knowledge we have about it is
> > derived from "habit" (as I trust you know he never called it induction
> > himself). Everything that we know about the world is unsupported by
> > logic.
>
> I have not read Hume
Then indeed you are missing out on one of philosophy's great
pleasures. I'd not suggest "A Treatise on Human Nature," though - he
made the same arguments later in "An Enquiry Concerning Human
Understanding" and "An Enquiry Concerning the Principles of Morals".
The later books are more clearly written. His "Natural History of
Religion" is also fun to read.
>or followed the many similar arguments here, but
> it seems to me that Hume is wrong and induction is supported by
> logic. How else can we get a good approximation of what is true about
> the world without observing the world and making predictions based on
> what we have repeatedly seen in the past? (I am aware that this last
> is your point.)
>
> Since logic cannot be built on anything except induction/habit, how
> can Hume argue it is: "unsupported by logic"?
Hume is correct, I think; there is no valid, deductive logical
syllogism that gets you from "the sun rose every day for the past 50
years" to "the sun will rise tomorrow." His point is precisely that
what we know about the world we do not know by means of logical
syllogisms.
Friar Broccoli
.
> Hume is correct, I think; there is no valid, deductive logical
> syllogism that gets you from "the sun rose every day for the past 50
> years" to "the sun will rise tomorrow." His point is precisely that
> what we know about the world we do not know by means of logical
> syllogisms.
I'd like to be convinced of this. It seems to me that the
(inductively derived) rules about mass and gravity allow us to state
as a premise that the earth is orbiting the sun, and thus that, in the
absence of a nearby black hole, an act of God, or a reversion to the
reality of LastThurday, we can deduce that the sun will rise tomorrow
morning. (Note that those three exceptions might also change the
relationship of the opposite to the adjacent in a triangle)
Friar Broccoli
> Friar Broccoli <elia...@gmail.com> wrote:
> > On Dec 27, 10:40 pm, Bill <brogers31...@gmail.com> wrote:
>
> > > Plato and Descartes, who saw human nature as fundamentally rational.
> > > Hume argues that the most basic way we learn what we need to know
> > > about the world in order to survive in it, a method which we share
> > > with young children and animals, is completely unsupported by logic.
> > > Our life in the world and all of the knowledge we have about it is
> > > derived from "habit" (as I trust you know he never called it induction
> > > himself). Everything that we know about the world is unsupported by
> > > logic.
>
> > I have not read Hume or followed the many similar arguments here, but
> > it seems to me that Hume is wrong and induction is supported by
> > logic. �How else can we get a good approximation of what is true about
> > the world without observing the world and making predictions based on
> > what we have repeatedly seen in the past? �(I am aware that this last
> > is your point.)
>
> > Since logic cannot be built on anything except induction/habit, how
> > can Hume argue it is: �"unsupported by logic"?
.
> You are equivocating upon "logic". Of course we use inductive logic:
> Hume never denied it. He was making the point that we cannot *justify*
> induction through the use of deductive logic, because it requires
> assumptions that are not deductive.
This appears empty. All deduction depends on inductively derived
premises. So this is equivalent to saying that no logic can be
justified.
> This misinterpretation of Hume is based, I think, on a misreport of
> Popper's point about falsification, which relies on a deductive claim:
> modus tollens (show that a generalisation is false if it is) is all you
> can do to an inductive generalisation.
Only if you are using an exclusive true/false template. You can also
show that it is more and more likely to be true. Also all scientific
generalisations are at least based on induction.
> Some people think this means that
> induction canot work. Certainly Popper was uninterested in it (he has no
> logic of discovery, contrary to the English title of his magnum opus),
Be interested in an explanation of what this means. You appear to be
saying that Popper was uninterested in induction (ie. how hypothesis
are formed).
> but Hume was not. He said that reason is not how we form ideas based on
> experience, and he was right.
Again, I'm not sure what this could mean. "Experience" can only lead
to inductive conclusions.
> But he also said that when we are done
> doing philosophy we should continue to live.
I am beginning to hope that an understanding of falsification is
within (or maybe just beyond) my grasp :-)
Bill
Well, short of sending you to read Hume, let me try. A successful
syllogism is of the form
Major premise: All men are mortal
Minor premise: Socrates is a man
Conclusion: Socrates is mortal.
The only problem with that syllogism is confirming the truth of the
major premise. If it were math, you could define the premises as true.
But in the real world, how do you show that all men are mortal?
Billions of them have died, certainly, but Hume's point is that there
is no deductive syllogism ....
Major premise: Billions of men have died
Minor premise: Friar Broccoli is a man
Conclusion: Friar broccoli is mortal
Of course, it is entirely natural and reasonable to assume that all
men are in fact mortal, and that the sun will come up tomorrow, but
that assumption cannot be shown to be true by a deductive logical
argument.
We cannot deduce that the sun will rise tomorrow. We unavoidably and
irresistibly induce that it will do so. And we'd be fools not to.
John S. Wilkins
Really? All logic is derived from induction? That is a rather large
claim. The *use* of logic may be derived from induction (it was valid
last time I used it, so I'll use modus ponens again), but the validity
itself is not, sort of by definition.
As to whether the logical forms can be justified (noncircularly?
Circular reasoning is considered unreasonable, but not from induction
about past circularities); it is arguable that logic, en se, is *not*
justified, but does the justifying (setting up Hume's problem).
>
>
> > This misinterpretation of Hume is based, I think, on a misreport of
> > Popper's point about falsification, which relies on a deductive claim:
> > modus tollens (show that a generalisation is false if it is) is all you
> > can do to an inductive generalisation.
>
> Only if you are using an exclusive true/false template. You can also
> show that it is more and more likely to be true. Also all scientific
> generalisations are at least based on induction.
A petitio. Science is not thereby a justification of induction; for
that, you would need independent evidence that science is truth
tracking, and that does not happen to exist. And what's this about the
law of the excluded middle? Are you trying to establish supervaluative,
or worse, paraconsistent logics? Or merely multivalued logics?
>
> > Some people think this means that
> > induction canot work. Certainly Popper was uninterested in it (he has no
> > logic of discovery, contrary to the English title of his magnum opus),
>
> Be interested in an explanation of what this means. You appear to be
> saying that Popper was uninterested in induction (ie. how hypothesis
> are formed).
That is indeed what I am saying. Popper basically said: conjecture how
you like; dream it, invent it, guess, and so forth. The real test is the
refutability of the hypothesis (vide his book title Conjectures and
Refutations, which was his title, unlike Logic of Scientific Discovery,
which was called The Logic of Research in German). He literally had no
logic of discovery.
>
> > but Hume was not. He said that reason is not how we form ideas based on
> > experience, and he was right.
>
> Again, I'm not sure what this could mean. "Experience" can only lead
> to inductive conclusions.
Err, yes. That is what I said. For Hume, reason = logic (classical,
Aristotelian, syllogistic logic). But we do not reason to form our ideas
that are experiential. Hume's empiricism is not deductive.
>
> > But he also said that when we are done
> > doing philosophy we should continue to live.
>
> I am beginning to hope that an understanding of falsification is
> within (or maybe just beyond) my grasp :-)
Faslification is simple; so simple anyone can misunderstand it. The more
falsifiable (i.e., refutable on the basis of observations) a hypothesis
is, the more empirical content it has. If it is highly falsifiable, *and
has not been falsified*, it is a most scientific hypothesis.
Unfortunately, Popper's general scheme would cripple science and makes
no sense of some of the most crucial moments in scientific progress
(like, for example Newton's theories of light or gravity).
jillery
I really am enjoying this dialog. You are doing well to explain these
subtle distinctions well enough that I understand what you're saying.
But if you don't mind, you just said something that has me just a bit
confused. IIUC the actual question at hand is if induction is
logical. If so, then your statement above begs the question. Yes, we
can't deduce the Sun will rise tomorrow (actually it doesn't but
that's another topic). We infer it. But why is that inference not
logical?
And while I'm on it, does any of this relate to abduction?
Burkhard
Burkhard
> subtle distinctions well enough that I understand what you're saying.
> But if you don't mind, you just said something that has me just a bit
> logical. 锟絀f so, then your statement above begs the question. 锟結es, we
> can't deduce the Sun will rise tomorrow (actually it doesn't but
> logical?
>
> And while I'm on it, does any of this relate to abduction?
Of course, and also to the panspermia thread: I was abducted by Aliens
on Monday and they anally probed me. I was abducted by aliens on
Tuesday and they anally probed me....Inductive inference: all aliens
are buggers.
Seriously though: abduction is a form of inference to the best
explanation. From knowing that all M are P1....Pn. If S is also
P1...Pn, I can infer that probably, S is M So from knowing that the
sprinkler always makes my garden fence wet, the observation that my
garden fence is wet, I abduce that it was the sprinkler that cause the
observation.
Of course, this is a defeasible inference. If I also learn that it was
raining that day, I might revise any conclusion, so the truth of the
conclusion is not guaranteed.
It plays an important role in "diagnostic disciplines, from Holmes to
House - you might find this paper mildly amusing for a computational
approach to abductive reasoning in crime investigations, the relevant
part is section 3ff :o) (there is also a version that applies it to
Agatha Christie's "and then tehre was none", but it is ot onthe web
yet) ;
http://www2.warwick.ac.uk/fac/soc/law/elj/jilt/2007_1/schafer_keppens/
Abduction can be treated rigorously in a probabilistic Bayesian
setting- and there is a follow up of taht paper which does just that.
Which would bring me to the point that was missed so far: even though
induction is not valid in terms of deductive logic, we can justify
often a inductive inference probabilistically. Patrick Maher, in
"Betting in theories" has extended the notion to the rational decision
of acceptance of theories, so we need not be quite as sceptical as
Hume I'd say about that rationality of our inductive theory choices.
jillery
> > subtle distinctions well enough that I understand what you're saying.
> > But if you don't mind, you just said something that has me just a bit
> > logical. If so, then your statement above begs the question. Yes, we
> > can't deduce the Sun will rise tomorrow (actually it doesn't but
> > logical?
>
> > And while I'm on it, does any of this relate to abduction?
>
> Of course, and also to the panspermia thread: I was abducted by Aliens
> on Monday and they anally probed me. I was abducted by aliens on
> Tuesday and they anally probed me....Inductive inference: all aliens
> are buggers.
Yeah, I reacted the same way when I first heard the term :)
> Seriously though: abduction is a form of inference to the best
> explanation. From knowing that all M are �P1....Pn. �If �S is also
> P1...Pn, I can infer that probably, S is M So from knowing that the
> sprinkler always makes my garden fence wet, the observation that my
> garden fence is wet, I abduce that it was the sprinkler that cause the
> observation.
> Of course, this is a defeasible inference. If I also learn that it was
> raining that day, I might revise any conclusion, so the truth of the
> conclusion is not guaranteed.
I'm lost already. Why is this an abduction and not an induction? I
thought the difference was abduction was identifying the most likely
from a set of possible inductions. Either way, would abduction be
considered "not logical" for the same reasons induction is considered
"not logical"?
> It plays an important role in "diagnostic disciplines, from Holmes to
> House - you might find this paper mildly amusing for a computational
> approach to abductive reasoning in crime investigations, the relevant
> part is section 3ff �:o) �(there is also a version that applies it to
> Agatha Christie's "and then tehre was none", but it is ot onthe web
> yet) ;http://www2.warwick.ac.uk/fac/soc/law/elj/jilt/2007_1/schafer_keppens/
>
> Abduction can be treated rigorously in a probabilistic Bayesian
> setting- and there is a follow up of taht paper which does just that.
> Which would bring me to the point that was missed so far: �even though
> induction is not valid in terms of deductive logic, we can � justify
> often a inductive inference probabilistically. Patrick Maher, in
> "Betting in theories" has extended the notion to the rational decision
> of acceptance of theories, so we need not be quite as sceptical as
> Hume I'd say about that rationality of our inductive theory choices.
Thank you for this. But I still don't know why induction might be
considered "not logical". Must logic be based strictly on deduction?
Also, does the noun "induction" and "inference" mean similar things?
Mark Isaak
> On Dec 28, 9:23 am, T Pagano <not.va...@address.net> wrote:
>> On Mon, 27 Dec 2010 16:40:15 -0800 (PST), Bill
>>
>> <brogers31...@gmail.com> wrote:
>> >Merry Christmas, Tony.
>>
>> .. . .and back at ya.
>>
>> > I do have warm memories of the drubbing you
>> >laid on me over Hume and induction earlier this year. You can drub me
>> >again without breaking you New Year's resolution. I enjoyed being
>> >trounced so much, that I want to relive the experience.
>>
>> While I did rub it in, it was one of the more interesting exchanges
>> I've had in a while. You actually took the time to go into the
>> original works of Hume. You forced me to do the same.
>>
>> >In brief...
>>
>> >You and I and Hume agree that there is no logical foundation for
>> >induction.
>>
>> Not exactly. You denied that Hume layed the foundation, I produced
>> the necessary quotes to prove otherwise, and you graciously conceded.
>
> I denied that Hume laid what foundation? I honestly don't get your drift
> here. What I said back in June was "We agree that induction is not
> supported by logic. That is indeed what Hume said. No argument from me
> there."
>
>
>> Not only is there no foundation for induction it is logically invalid.
>>
>> >You conclude from this that induction can give you no valid knowledge
>> >about the world.
>>
>> I don't conclude this because I don't know what "valid knowledg" is.
>>
>> Instead, I argue that every inductive argument is logically invalid.
>> That is, the conclusion goes beyond what is contained in the premises.
>> This does not mean that the conclusion is necessarily false only that
>> we have no way of knowing its truth value from an invalid inductive
>> argument.
>
>
>> In a valid deductive argument the truth of the premises is transmitted
>> to its conclusion.
>
> That's certainly true. However, unless the premises are true by
> definition, in which case you are restricted to statements about
> mathematics, deduction is useless. How do you propose to get to a "true
> premise" about the physical world?
> [...]
To perhaps throw a curve into this discussion, I will remind people that
not all logic is Aristotelian. What happens when we bring Bayes to the
party? It is not (deductively) logical for me to conclude, from the fact
that I have seen it do so for the last fifty years, that the sun will
certainly rise tomorrow, but may I deduce that the chance of it not
rising are less than 1/18,000?
--
Mark Isaak eciton (at) earthlink (dot) net
"It is certain, from experience, that the smallest grain of natural
honesty and benevolence has more effect on men's conduct, than the most
pompous views suggested by theological theories and systems." - D. Hume
Burkhard
He is already there and came with me, get your own partner :o)
and also in response to a post by Jillery that I can't find: It comes
down to semantics. IF you have a certain conception of logic, these are
not logically valid inferences (and abduction in particular often looks
like a fallacy - affirming the consequent) But if you have a more
liberal notion of logic, and include in particular, then you coudl say
that these inferences are justified relative to such a theory
Richard Harter
wrote:
Stated that baldly, you may make that deduction, but you would do so
incorrectly. The applicability of probability arguments depends upon
assumptions supplied by induction. If the great deceiver is stage
managing events all bests are off.
M. S. Goodrich
(snip)
Then according to Cox and Jaynes, you end up with a generalization or
extension of Aristotelian logic which yields a calculus of
probability
(vice frequency) for propositions in the closed interval [0,1].
Happy New Year All,
-Mike Goodrich
Go Hokies, beat the Cardinal!
Richard Clayton
Hot dog, welcome back to t.o, Mike!
(And I didn't know you were a Virginia Tech fan.)
--
[The address listed is a spam trap. To reply, take off every zig.]
Richard Clayton
"I keep six honest serving men (they taught me all I knew); their names
are What and Why and When and How and Where and Who." — Rudyard Kipling
M. S. Goodrich
wrote:
Thank you for your warm greeting, Richard.
> (And I didn't know you were a Virginia Tech fan.)
>
Yes indeed, and a big one, having been an undergraduate there.
Friar Broccoli
.
Because abduction reasons:
from an outcome to (potentially many) possible antecedents.
while induction references:
many specific observations of:
an antecedent associated with (a presumably caused) outcome
Note the direction change antecedent -> outcome.
Note also that abduction is a form of guessing from a pool of
apparently possible causes, while induction is an association based on
many observed correlations between one antecedent and one outcome.
> I thought the difference was abduction was identifying the most likely
> from a set of possible inductions. Either way, would abduction be
> considered "not logical" for the same reasons induction is considered
> "not logical"?
Abduction is a form of weak deduction from pre-existing knowledge.
Induction is an inference (often to causality) from observed
correlations.
Thus induction is essentially statistical, while abduction is
essentially logical. Induction normally fails because the association
is found not to be causal, while abduction normally fails because a
different possible cause (anticipated or not) was found to be the true
cause. So induction generally considers a single correlation, while
abduction implicitly assumes many possible correlations, even if only
one (say rainfall) is known.
> > It plays an important role in "diagnostic disciplines, from Holmes to
> > House - you might find this paper mildly amusing for a computational
> > approach to abductive reasoning in crime investigations, the relevant
> > part is section 3ff :o) (there is also a version that applies it to
> > Agatha Christie's "and then tehre was none", but it is ot onthe web
> > yet) ;http://www2.warwick.ac.uk/fac/soc/law/elj/jilt/2007_1/schafer_keppens/
>
> > Abduction can be treated rigorously in a probabilistic Bayesian
> > setting- and there is a follow up of taht paper which does just that.
> > Which would bring me to the point that was missed so far: even though
> > induction is not valid in terms of deductive logic, we can justify
> > often a inductive inference probabilistically. Patrick Maher, in
> > "Betting in theories" has extended the notion to the rational decision
> > of acceptance of theories, so we need not be quite as sceptical as
> > Hume I'd say about that rationality of our inductive theory choices.
.
> Thank you for this. But I still don't know why induction might be
> considered "not logical". Must logic be based strictly on deduction?
> Also, does the noun "induction" and "inference" mean similar things?
Inference refers to a class of processes. Deduction, induction,
abduction etc. are members of the class.
Friar Broccoli
I got your original statement backward (almost normal for me). Your:
"what we know about the world we do not know by means of logical
syllogisms."
became in my mind:
what we KNOW about the world we know by means of logical syllogisms.
So we are all just back to the old we don't know anything with
certainty, just that some things are very very likely to be true and
others (like Noah's ark) are so unlikely as to be absurd.
Friar Broccoli
.
.
Here you appear to be responding to a powerful argument I did not
make, but which I now feel foolish for having missed: namely that
deduction itself is based on and justifiable only by induction. Thus
it is silly to place deduction above induction in the pantheon of
logical forms. (The argument I was intending to make was that the
premises used in every [useful] instance of deductive arguments are
inductively derived)
As you recognize in the following paragraph modus ponens is just a
form of deduction, and thus its use here is circular.
Thus your argument seems to rest on a definition "sort of", but
definitions need not refer to anything that's really out there like
unicorns, so I don't think that's going anywhere.
> As to whether the logical forms can be justified (noncircularly?
> Circular reasoning is considered unreasonable, but not from induction
> about past circularities); it is arguable that logic, en se, is *not*
> justified, but does the justifying (setting up Hume's problem).
I've heard that Hume concluded that the problem was completely
insoluble to the point where the scientific process on which modern
society rests would be shown to be unworkable because different
observes simply would not agree that the "same" results they were
seeing were in fact the "same". (If we don't all agree the sun is
rising, induction fails.)
If you agree with this, did he talk about establishing measurement
procedures prior to performing tests or looking at the data? Or
alternatively what is the modern answer?: keep creationists out of
labs.
>>> This misinterpretation of Hume is based, I think, on a misreport of
>>> Popper's point about falsification, which relies on a deductive claim:
>>> modus tollens (show that a generalisation is false if it is) is all you
>>> can do to an inductive generalisation.
>
>> Only if you are using an exclusive true/false template. You can also
>> show that it is more and more likely to be true. Also all scientific
>> generalisations are at least based on induction.
.
> A petitio. Science is not thereby a justification of induction; for
> that, you would need independent evidence that science is truth
> tracking, and that does not happen to exist.
I don't think I am assuming my conclusion, except in the sense that
all deductive arguments assume their conclusions.
We know by induction that science eliminates a vast array of false
propositions, and thereby leads us closer and closer to the truth
track. The usability of the truth that remains is the "independent
evidence that science is truth tracking".
> And what's this about the law of the excluded middle? Are you
> trying to establish supervaluative, or worse, paraconsistent
> logics? Or merely multivalued logics?
I'm not even going to try and unpack that. I know from past
discussions that all logical systems collapse from internal
contradictions. If I stray too far from the practical, induction
tells me to expect a gang rape.
>>> Some people think this means that
>>> induction canot work. Certainly Popper was uninterested in it (he has no
>>> logic of discovery, contrary to the English title of his magnum opus),
>
>> Be interested in an explanation of what this means. You appear to be
>> saying that Popper was uninterested in induction (ie. how hypothesis
>> are formed).
>
> That is indeed what I am saying. Popper basically said: conjecture how
> you like; dream it, invent it, guess, and so forth. The real test is the
> refutability of the hypothesis (vide his book title Conjectures and
> Refutations, which was his title, unlike Logic of Scientific Discovery,
> which was called The Logic of Research in German). He literally had no
> logic of discovery.
>
>>> but Hume was not. He said that reason is not how we form ideas based on
>>> experience, and he was right.
>
>> Again, I'm not sure what this could mean. "Experience" can only lead
>> to inductive conclusions.
>
> Err, yes. That is what I said. For Hume, reason = logic (classical,
> Aristotelian, syllogistic logic). But we do not reason to form our ideas
> that are experiential. Hume's empiricism is not deductive.
>
>>> But he also said that when we are done
>>> doing philosophy we should continue to live.
>
>> I am beginning to hope that an understanding of falsification is
>> within (or maybe just beyond) my grasp :-)
.
> Faslification is simple; so simple anyone can misunderstand it. The more
> falsifiable (i.e., refutable on the basis of observations) a hypothesis
> is, the more empirical content it has. If it is highly falsifiable, *and
> has not been falsified*, it is a most scientific hypothesis.
>
> Unfortunately, Popper's general scheme would cripple science and makes
> no sense of some of the most crucial moments in scientific progress
> (like, for example Newton's theories of light or gravity).
I assume that Newton's gravity cannot be falsified because it is so
obvious that there can be no evidence against it?
And that his theory of light was just a model on which others could
build - ie it was somewhat true and somewhat false - and thus not
subject to a true/false model of falsification?
Bill
> On Dec 29 2010, 2:28 am, j...@wilkins.id.au (John S. Wilkins) wrote:
>
> > As to whether the logical forms can be justified (noncircularly?
> > Circular reasoning is considered unreasonable, but not from induction
> > about past circularities); it is arguable that logic, en se, is *not*
> > justified, but does the justifying (setting up Hume's problem).
>
> I've heard that Hume concluded that the problem was completely
> insoluble to the point where the scientific process on which modern
> society rests would be shown to be unworkable because different
> observes simply would not agree that the "same" results they were
> seeing were in fact the "same". (If we don't all agree the sun is
> rising, induction fails.)
I don't believe Hume ever said anything like that. I mean, he never
said anything like "the problem was completely insoluble......were in
fact the "same." Hume never talked about "the problem of induction."
He did not think that it was wrong to expect that "similar causes
should have similar effects." He was not criticizing induction, he was
criticizing the view that man is primarily a rational animal guided by
reason (by which he meant something like deductive logic). He was not
in any way an epistemiological nihilist.
>
> If you agree with this, did he talk about establishing measurement
> procedures prior to performing tests or looking at the data? Or
> alternatively what is the modern answer?: keep creationists out of
> labs.
>
<snip>
T Pagano
<broger...@gmail.com> wrote:
>On Dec 28, 9:23�am, T Pagano <not.va...@address.net> wrote:
>> On Mon, 27 Dec 2010 16:40:15 -0800 (PST), Bill
>>
>> <brogers31...@gmail.com> wrote:
snip
>> Not exactly. �You denied that Hume layed the foundation, I produced
>> the necessary quotes to prove otherwise, and you graciously conceded.
>
>I denied that Hume laid what foundation? I honestly don't get your
>drift here.
You denied that Hume addressed induction's inadequacy at all. You
even attempted to produce Hume quotes which put induction in a good
light. You failed. You challenged me to produce the quotes from Hume
addressing the problem of induction. I suspect you doubted that I
could produce them. Nonetheless I did produce and you were rebuffed.
It was only then that you claimed to have seen the light.
> What I said back in June was "We agree that induction is
>not supported by logic. That is indeed what Hume said. No argument
>from me there."
You agreed to nothing until I proved that Hume (with decisive quotes)
uncovered induction's inadequacy.
snip
>>
>> In a valid deductive argument the truth of the premises is transmitted
>> to its conclusion.
>
>That's certainly true. However, unless the premises are true by
>definition, in which case you are restricted to statements about
>mathematics, deduction is useless. How do you propose to get to a
>"true premise" about the physical world?
Nonsense if scientists could not provisionally presume the truth of
their observation statements or their reports of experimental results
then no scientific activity could be conducted. So without any
difficulty (or fear) an investigator can take those observation
statements and reports of experimental results and form deductive
arguments (albeit provisional ones).
And so long as the deductive argument is valid the truth of the
premises is transmitted to the conclusion. In other words we can be
sure of the truth value of the conclusion merely from the form of the
argument. In an inductive arguement it is irrelevent whether the
premises are true or not; we have no way of knowing from the form of
the argument whether the conclusion is true or even probably true.
>
>You should get a job as a revisionist historian - you'd be great. What
>I said back then was "We agree that induction is not supported by
>logic. That is indeed what Hume said. No argument from me there." I
>agree and have always agreed that induction is logically invalid.
>
><snip - I've long since stipulated that induction is not logically
>justified>
Again this is poppy cock. You never saw the light until I finally
waved the necessary Hume quotes in front of your face.
>>
>> We don't search for "valid" knowledge, we search for "true" knowledge.
>> "Truth" and "validity" are not synonyms. �
>
>OK, Tony, how do you find "true knowledge?"
[Note: I should point out that the context of this discussion is with
regard to Universal theories. ]
When I wrote that "we" search for true knowledge I was not referring
to modern secular/atheists. Generally Christians search for truth not
secular/atheists.
Modern secular/Atheists search for verified (or justified) knowledge
not true knowledge. That is, they do exactly what Hume discovered
canNOT be done. That is, they incorrectly believe that observations
that can be made can justify/verify/make probable observations which
cannot be made.
snip
>>
>> You acknowledge that inductive arguments are logically invalid. �Yet
>> you also are of the opinion that this form of reasoning can help you
>> sort between true and false conclusions. �These statements are
>> inconsistent with one another. �To hold them both---as you apparently
>> do-----makes you irrational.
>
>Now we're getting somewhere. I'm irrational. Exactly Hume's point.
>Bingo. There's a reason he called it "A Treatise Concerning Human
>Nature." His large scale argument is against those philosophers, like
>Plato and Descartes, who saw human nature as fundamentally rational.
>Hume argues that the most basic way we learn what we need to know
>about the world in order to survive in it, a method which we share
>with young children and animals, is completely unsupported by logic.
>Our life in the world and all of the knowledge we have about it is
>derived from "habit" (as I trust you know he never called it induction
>himself). Everything that we know about the world is unsupported by
>logic. He did not conclude that we don't know anything or that we
>should not trust induction, even to the point of making life and death
>decisions based on it; what he concluded is that human nature is not
>fundamentally rational. Elsewhere in the Treatise he makes similar
>arguments about ethics. Our moral judgments are based not on reason
>but on sentiment. He does not conclude that morality is meaningless or
>should be disregarded, only that, like knowledge, its basis is non-
>rational.
I think this is a great analysis and I agree with it, in some
respects. Hume was essentially driven into irrationalism over what he
perceived as an irreconcilable conflict. On the one hand Hume
recognized that we learn via simple generalization, while on the
other hand he discovered that we cannot generalize from what we can
see to that which we cannot see.
The conflict drove Hume into irrationalism. What he failed to realize
is that simple generalization forms the basis for action not a
mechanism for generating knowledge.
>So we're stuck. Induction is clearly not founded on logic. But try
>living without it. You can say in words that what induction teaches
>you gives you not a hint about the truth of the world, but Hume and I
>will challenge you to live 5 minutes without making constant use of
>induction. You can minimize the value of "inductive knowledge," but to
>downplay the importance of the form of knowledge on which you
>constantly depend for your daily survival seems like posturing.
You make the same mistake which drove Hume into irrationalism.
Certainly if I observe a 10 lb steel ball drop from a third story
window every 10 seconds for several hours I would be wise not to stand
under the window. This generalization forms a basis of action not a
mechanism for generating or testing new ideas or knowledge or
theories.
>Deduction is not a useful alternative; it can tell you nothing about
>the real world, and even in mathematics, where it works best, it is
>incomplete.
It is the atheist who claims that genuinely new knowledge is generated
from some mechanistic, inductive method not I. As you point out
deduction is a method of illuminating what we already know. How man
comes to create/invent/generate genuinely new knowledge is mostly
unknown. And while our background knowledge may play a part in these
leaps of insight they are never mere deductions from it.
Furthermore, the beauty of deduction is that it can provisionally tell
us when a theory is false. In order for a theory to be scientific it
must prohibit at least one event from occurring. If the event is
observed the theory is (provisionally) falsified. This is pure
deduction.
Regards,
T Pagano
T Pagano
>On Mon, 27 Dec 2010 19:40:39 -0800, Bill wrote:
>
>> On Dec 28, 9:23 am, T Pagano <not.va...@address.net> wrote:
>>> On Mon, 27 Dec 2010 16:40:15 -0800 (PST), Bill
>>>
>>> <brogers31...@gmail.com> wrote:
>>> >Merry Christmas, Tony.
snip
>>> In a valid deductive argument the truth of the premises is transmitted
>>> to its conclusion.
>>
>> That's certainly true. However, unless the premises are true by
>> definition, in which case you are restricted to statements about
>> mathematics, deduction is useless. How do you propose to get to a "true
>> premise" about the physical world?
>> [...]
>
>To perhaps throw a curve into this discussion, I will remind people that
>not all logic is Aristotelian. What happens when we bring Bayes to the
>party?
While Bayesians are often deductivists, Bayesianism has virtually
nothing to do with separating theories into true and false ones. It
is generally limited to assessing probabilies given some evidence.
Since there is no proof that the more probable theory is necessarily
closer to the truth one wonders why anyone would apply Bayesianism to
assessing the verisimilitude of some set of competing theories.
It's not clear that Bayesianism is an improvement on or even an aid to
the inadequacy of inductiive reasoning.
> It is not (deductively) logical for me to conclude, from the fact
>that I have seen it do so for the last fifty years, that the sun will
>certainly rise tomorrow, but may I deduce that the chance of it not
>rising are less than 1/18,000?
This demonstrates that simple generalization forms the basis for
action not as mechanism for reasoning or as a means of generating or
justifying or verifying theories of any content.
That the sun has risen for the last 50 years would form a rational
basis for action for me to pack my sun glasses for my trip to work the
next morning. Nonetheless secularists regularly employ this same sort
of reasoning to justify/verify that which has never been seen or
cannot be seen. Hume showed that this is useless. The Problem of
Induction is one of the great unsolved philosophical problems.
I had always hoped that Wilkins would educate us. . . .
Regards,
T Pagano
John Stockwell
> On Mon, 27 Dec 2010 19:40:39 -0800 (PST), Bill
>
> >On Dec 28, 9:23 am, T Pagano <not.va...@address.net> wrote:
> >> On Mon, 27 Dec 2010 16:40:15 -0800 (PST), Bill
> Pagano said in part:
> It is the atheist who claims that genuinely new knowledge is generated
> from some mechanistic, inductive method not I. � �As you point out
> deduction is a method of illuminating what we already know. �How man
> comes to create/invent/generate genuinely new knowledge is mostly
> unknown. �And while our background knowledge may play a part in these
> leaps of insight they are never mere deductions from it.
Actually, Tony, inductive methods are one tool. As any of us know in
science, we invent theories, they are not handed to us. So, claiming
that "atheists say thus and such" is basically a lie on your part.
One level of theorizing consists of identifying (or inventing) laws
to reduce large numbers of observations into simple structures.
This need not be mathematical. The classification schemes of taxonomy
and of stratigraphy, with the geologic time scale being a major
examples.
>
> Furthermore, the beauty of deduction is that it can provisionally tell
> us when a theory is false. �In order for a theory to be scientific it
> must prohibit at least one event from occurring. �If the event is
> observed the theory is (provisionally) falsified. �This is pure
> deduction.
Actually, no. It would be nice if all theories were so clear cut. You
are forgetting experimental and observational error. All observations
are statistical objects. The experimenter has to determine what the
acceptable agreement between theoretical prediction and experimental
result is.
An example was Kepler's reduction of the motion of Mars into an
ellipse. He had a schema based on Platonic solids, but he found the
error between the fit with this law to be unacceptable, and eventually
decided on an ellipse as the appropriate "law" to adopt to reduce the
data.
>
> Regards,
> T Pagano
-John
John Stockwell
The mathematical theory of statistics and probability was invented to
deal with experimental and observational error, which are not handled
in classical logic. In classical logic, there is no such thing as a
partial
truth, or nearly true item. Yet, in science we see such phenomena all
of the time.
Bayes theorem can be used to address the notion of improved confidence
through a formalization of a type of induction. Induction isn't logic,
in the sense of a consistent system. But it is a more realistic model
of how we actually figure out things, or more appropriately, how
we communicated our reasoning in a way that is testable by others.
>
> > It is not (deductively) logical for me to conclude, from the fact
> >that I have seen it do so for the last fifty years, that the sun will
> >certainly rise tomorrow, but may I deduce that the chance of it not
> >rising are less than 1/18,000?
>
> This demonstrates that simple generalization forms the basis for
> action not as mechanism for reasoning or as a means of generating or
> justifying or verifying theories of any content.
>
> That the sun has risen for the last 50 years would form a rational
> basis for action for me to pack my sun glasses for my trip to work the
> next morning. �Nonetheless secularists regularly employ this same sort
> of reasoning to justify/verify that which has never been seen or
> cannot be seen. �Hume showed that this is useless. �The Problem of
> Induction is one of the great unsolved philosophical problems. �
>
> I had always hoped that Wilkins would educate us. . . .
Science largely deals with things that we cannot observe. In lieu
of observation we have the predictions of theories. The fact that we
can calculate the positions of the planets with great precision, and
have
an understanding of the processes by which planetary motions operate
that give us confidence regarding not only the sun appearing in the
morning
sky tomorrow, but that we can calculate the time of that rising to any
degree of precision that we need.
All of these things have become part of our uniform experience of the
universe, and have never been violated. Which is why claims that such
laws have been violated are to be taken with a grain of salt.
>
> Regards,
> T Pagano
-John
Bill
> On Mon, 27 Dec 2010 19:40:39 -0800 (PST), Bill
>
I take a pragmatic approach, like William James - the truth is what
works. So if an idea is a good basis for action it is, to that extent,
true. That's all the truth that's on offer. It is positively weird
that you think an idea might be such a good basis for action that one
can stake one's life on it, and yet consider it impossible to make any
meaningful claim for the truth of that idea.
> >Deduction is not a useful alternative; it can tell you nothing about
> >the real world, and even in mathematics, where it works best, it is
> >incomplete.
>
> It is the atheist who claims that genuinely new knowledge is generated
> from some mechanistic, inductive method not I. � �As you point out
> deduction is a method of illuminating what we already know. �How man
> comes to create/invent/generate genuinely new knowledge is mostly
> unknown. �
>And while our background knowledge may play a part in these
> leaps of insight they are never mere deductions from it.
I want some of that background knowledge of yours Tony, you know that
kind that just appears and is not subject to the problem of induction
or to any skeptical review. Last time I asked you for some you told me
to go to any library. That was hardly a serious answer. Where does
background knowledge come from and how do you know it's true?
John S. Wilkins
John Stockwell <john.1...@gmail.com> wrote:
> On Jan 4, 6:58 pm, T Pagano <not.va...@address.net> wrote:
...
> > That the sun has risen for the last 50 years would form a rational
> > basis for action for me to pack my sun glasses for my trip to work the
> > next morning. Nonetheless secularists regularly employ this same sort
> > of reasoning to justify/verify that which has never been seen or
> > cannot be seen. Hume showed that this is useless. The Problem of
> > Induction is one of the great unsolved philosophical problems.
> >
> > I had always hoped that Wilkins would educate us. . . .
I don't need to. Induction has been well discussed, and solutions
offered, in the literature. [Philosophical problems aren't solved,
anyway, they are resolved or abandoned, as Dewey noted.]
The problem of induction is not justification but the straight rule
problem. That is, as Hume said that we must presume uniformity of
causation for it to work, the question is how, from a position of
uncertainty, we can do this? I think that we bootstrap induction (and if
we didn't, we'd know precisely nothing), by refining the classes of
things that we can take to be uniform, and eliminating those that
experience tells us do not behave uniformly.
I suggest anyone interested in the actual issue reads this paper:
Godfrey-Smith, P. 2003. Goodman's Problem and Scientific Methodology.
The Journal of Philosophy 100:573-590.
If Pagano responds, will someone quote him so I can see if he says
anything interesting?
>
> Science largely deals with things that we cannot observe. In lieu of
> observation we have the predictions of theories. The fact that we can
> calculate the positions of the planets with great precision, and have an
> understanding of the processes by which planetary motions operate that
> give us confidence regarding not only the sun appearing in the morning sky
> tomorrow, but that we can calculate the time of that rising to any degree
> of precision that we need.
>
> All of these things have become part of our uniform experience of the
> universe, and have never been violated. Which is why claims that such laws
> have been violated are to be taken with a grain of salt.
--
Burkhard
Because it works. see Patrick Maher, Betting on Theories, Cambridge
1998 - it shows under what conditions it is rational to chose on
theory over another
Friar Broccoli
> Piggybacking
>
>
>
> John Stockwell <john.19071...@gmail.com> wrote:
> > On Jan 4, 6:58 pm, T Pagano <not.va...@address.net> wrote:
> ...
> > > That the sun has risen for the last 50 years would form a rational
> > > basis for action for me to pack my sun glasses for my trip to work the
> > > next morning. �Nonetheless secularists regularly employ this same sort
> > > of reasoning to justify/verify that which has never been seen or
> > > cannot be seen. �Hume showed that this is useless. �The Problem of
> > > Induction is one of the great unsolved philosophical problems. �
>
> > > I had always hoped that Wilkins would educate us. . . .
>
> I don't need to. Induction has been well discussed, and solutions
> offered, in the literature. [Philosophical problems aren't solved,
> anyway, they are resolved or abandoned, as Dewey noted.]
>
> The problem of induction is not justification but the straight rule
> problem. That is, as Hume said that we must presume uniformity of
> causation for it to work, the question is how, from a position of
> uncertainty, we can do this? I think that we bootstrap induction (and if
> we didn't, we'd know precisely nothing), by refining the classes of
> things that we can take to be uniform, and eliminating those that
> experience tells us do not behave uniformly.
>
> I suggest anyone interested in the actual issue reads this paper:
>
> Godfrey-Smith, P. 2003. Goodman's Problem and Scientific Methodology.
> The Journal of Philosophy 100:573-590.
I think this is a link to the paper (the page numbering is correct)
http://www.people.fas.harvard.edu/~pgs/PGS-Grue-04.pdf
Should this paper (which looks terribly abstract to me) allow me to
understand the general rules needed to go from the observation that
the sun rises every morning to the conclusion (after invoking
additional [sampled?] observations about gravity) that the earth
orbits (a location inside) the sun and not the other way around - or
does this deal with some other aspect of the problem of induction?
> If Pagano responds, will someone quote him so I can see if he says
> anything interesting?
I will.
>
>
>
> > Science largely deals with things that we cannot observe. In lieu of
> > observation we have the predictions of theories. The fact that we can
> > calculate the positions of the planets with great precision, and have an
> > understanding of the processes by which planetary motions operate that
> > give us confidence regarding not only the sun appearing in the morning sky
> > tomorrow, but that we can calculate the time of that rising to any degree
> > of precision that we need.
>
> > All of these things have become part of our uniform experience of the
> > universe, and have never been violated. Which is why claims that such laws
> > have been violated are to be taken with a grain of salt.
>
> --
> John S. Wilkins, Associate, Philosophy, University of Sydneyhttp://evolvingthoughts.net
Steven L.
"Burkhard" <b.sc...@ed.ac.uk> wrote in message
news:3590e4df-468f-446c...@32g2000yqz.googlegroups.com:
To me, it's not "Bayesianism" as much as it is the mathematics of
statistical sampling.
If ten different public opinion polls from ten different pollsters all
show candidate Smith ahead of candidate Jones by a wide margin,
candidate Jones should not expect to win the election if it were held
any time soon.
When you do scientific experiments, you're in effect taking statistical
samples of the universe for study. And the more samples you take, the
closer the mean of those sample aspects will get to the population
aspect. Ditto for proportions (like Smith 60% of respondents vs. Jones
40%).
It won't work as well for devising general theories of abiogenesis.
Because all we have are biased samples--the history of life on just one
planet, all of which descended from a common ancestor.
Truly universal laws of biology will have to wait until we learn more
about life in the Universe beyond the Earth.
Just as we learned more about our own Sun's history from studying other
stars.
-- Steven L.
Friar Broccoli
> On Jan 4, 10:44 pm, j...@wilkins.id.au (John S. Wilkins) wrote:
.[snip]
> > The problem of induction is not justification but the straight rule
> > problem. That is, as Hume said that we must presume uniformity of
> > causation for it to work, the question is how, from a position of
> > uncertainty, we can do this? I think that we bootstrap induction (and if
> > we didn't, we'd know precisely nothing), by refining the classes of
> > things that we can take to be uniform, and eliminating those that
> > experience tells us do not behave uniformly.
>
> > I suggest anyone interested in the actual issue reads this paper:
>
> > Godfrey-Smith, P. 2003. Goodman's Problem and Scientific Methodology.
> > The Journal of Philosophy 100:573-590.
>
> I think this is a link to the paper (the page numbering is correct)http://www.people.fas.harvard.edu/~pgs/PGS-Grue-04.pdf
>
> Should this paper (which looks terribly abstract to me) allow me to
> understand the general rules needed to go from the observation that
> the sun rises every morning to the conclusion (after invoking
> additional [sampled?] observations about gravity) that the earth
> orbits (a location inside) the sun and not the other way around - or
> does this deal with some other aspect of the problem of induction?
To answer my own question: yes - pg 587.
jillery
Thank you for letting me crib your notes. That's the kind of answers
I was looking for.
Walter Bushell
<c6b789e5-73cc-4ea7...@w17g2000yqh.googlegroups.com>,
John Stockwell <john.1...@gmail.com> wrote:
>
> The mathematical theory of statistics and probability was invented to
> deal with experimental and observational error, which are not handled
> in classical logic. In classical logic, there is no such thing as a
> partial truth, or nearly true item. Yet, in science we see such phenomen
> of the time.
IIUC, it was invented by gamblers to deal with their specific needs.
Later it was extended, but it needed an immediate payoff to finance
itself. Like the personal computer revolution was financed by gammers
and the internet by the p0rn customers.
--
The Chinese pretend their goods are good and we pretend our money
is good, or is it the reverse?
Robert Grumbine
> In article
><c6b789e5-73cc-4ea7...@w17g2000yqh.googlegroups.com>,
> John Stockwell <john.1...@gmail.com> wrote:
>
>>
>> The mathematical theory of statistics and probability was invented to
>> deal with experimental and observational error, which are not handled
>> in classical logic. In classical logic, there is no such thing as a
>> partial truth, or nearly true item. Yet, in science we see such phenomen
>> of the time.
>
> IIUC, it was invented by gamblers to deal with their specific needs.
> Later it was extended, but it needed an immediate payoff to finance
> itself. Like the personal computer revolution was financed by gammers
> and the internet by the p0rn customers.
Invented by a friend (Pascal) of a gambler. Said gambler looking,
of course, to get an edge on the house.
The observational error reduction aspect comes a century-plus later,
with Gauss, though one or more Bernoullis were working on it earlier.
--
Robert Grumbine http://moregrumbinescience.blogspot.com/ Science blog
Sagredo (Galileo Galilei) "You present these recondite matters with too much
evidence and ease; this great facility makes them less appreciated than they
would be had they been presented in a more abstruse manner." Two New Sciences
Friar Broccoli
> Friar Broccoli <elia...@gmail.com> wrote:
[snipping]
> > > Some people think this means that
> > > induction canot work. Certainly Popper was uninterested in it (he has no
> > > logic of discovery, contrary to the English title of his magnum opus),
>
> > Be interested in an explanation of what this means. You appear to be
> > saying that Popper was uninterested in induction (ie. how hypothesis
> > are formed).
.
> That is indeed what I am saying. Popper basically said: conjecture how
> you like; dream it, invent it, guess, and so forth. The real test is the
> refutability of the hypothesis (vide his book title Conjectures and
> Refutations, which was his title, unlike Logic of Scientific Discovery,
> which was called The Logic of Research in German). He literally had no
> logic of discovery.
In this month's Scientific American there is an article on Robotic
Scientists (two of which the author has built, one named Adam).
On page 76 he says:
"We therefore designed Adam to devise efficient experiments that test
hypotheses cheaply and quickly. To achieve this goal, Adam assumes
that every hypotheses has a probability of being true. This
assumption is controversial, and some philosophers such as Karl Popper
have denied that hypotheses can have associated probabilities."
To me, this appears to imply that Popper thought there is an absolute
truth out there AND people can know what it is, which would mean that
scientific "knowledge" is absolute and not provisional.
I'm sure my inference here cannot be correct, but it is not
immediately obvious to me how Popper could have believed that
hypotheses don't have probabilities if he understood that the
hypotheses finally selected to form part of a theory was not an
absolute truth.
Would you be kind enough to suggest how you think Popper put these
ideas together?
Bill
It sounds like the robot, Adam, is a Bayesian - that is, he considers
the prior probability that a hypothesis is true, then does an
experiment, and calculates the posterior probability that the
hypothesis is true, given the experimental result. In practice, this
sort of comes down to "extraordinary claims require extraordinary
evidence." To reject the possibility of assigning prior probabilities
to hypotheses is not to deny that theories are provisional, it is, if
anything, a step in the other direction, claiming that there is no way
whatsoever to assign probabilities to hypotheses. At most one could
say that, after the experiment the theory is more likely to be true
than it was before the experiment, but that one could not assign any
absolute probability that the theory was true at all. In Popper's
view, repeated failed attempts to falsify a theory make it more likely
to be true than it was before the failed falsifications, but there is
no way to assign a value to the chance that the theory is, in fact,
true.
John S. Wilkins
Oh, gods, that's thirty years ago! I can't be expected to remember all
that :-)
OK, as far as I do recall it, and without going to my library and
reading for twenty minutes, Popper held that a statement is either true
or false. So a false statement cannot be *more* true than another false
statement. But it *can* be more "truthlike", or in his terminology, have
more verisimiltude.
Popper's view on probabilities is discussed at SEP well:
http://plato.stanford.edu/entries/popper/#ProKnoVer
Michael Siemon
jo...@wilkins.id.au (John S. Wilkins) wrote:
...
> OK, as far as I do recall it, and without going to my library and
> reading for twenty minutes, Popper held that a statement is either true
> or false. So a false statement cannot be *more* true than another false
> statement. But it *can* be more "truthlike", or in his terminology, have
> more verisimiltude.
>
> Popper's view on probabilities is discussed at SEP well:
>
> http://plato.stanford.edu/entries/popper/#ProKnoVer
I think the current idiom is Stephen Colbert's "truthy". :-)
John S. Wilkins
Michael Siemon
jo...@wilkins.id.au (John S. Wilkins) wrote:
> Michael Siemon <mlsi...@sonic.net> wrote:
>
> > In article <1juv8f5.1j1idyhapgsznN%jo...@wilkins.id.au>,
> > jo...@wilkins.id.au (John S. Wilkins) wrote:
> > ...
> >
> > > OK, as far as I do recall it, and without going to my library and
> > > reading for twenty minutes, Popper held that a statement is either true
> > > or false. So a false statement cannot be *more* true than another false
> > > statement. But it *can* be more "truthlike", or in his terminology, have
> > > more verisimiltude.
> > >
> > > Popper's view on probabilities is discussed at SEP well:
> > >
> > > http://plato.stanford.edu/entries/popper/#ProKnoVer
> >
> > I think the current idiom is Stephen Colbert's "truthy". :-)
>
> I'd love to think that Colbert was satirising Popper...
I would too, and it is (remotely!) possible -- but I doubt it.
John S. Wilkins
> In article <1juvshe.106cq96rv31maN%jo...@wilkins.id.au>,
> jo...@wilkins.id.au (John S. Wilkins) wrote:
>
> > Michael Siemon <mlsi...@sonic.net> wrote:
> >
> > > In article <1juv8f5.1j1idyhapgsznN%jo...@wilkins.id.au>,
> > > jo...@wilkins.id.au (John S. Wilkins) wrote:
> > > ...
> > >
> > > > OK, as far as I do recall it, and without going to my library and
> > > > reading for twenty minutes, Popper held that a statement is either true
> > > > or false. So a false statement cannot be *more* true than another false
> > > > statement. But it *can* be more "truthlike", or in his terminology, have
> > > > more verisimiltude.
> > > >
> > > > Popper's view on probabilities is discussed at SEP well:
> > > >
> > > > http://plato.stanford.edu/entries/popper/#ProKnoVer
> > >
> > > I think the current idiom is Stephen Colbert's "truthy". :-)
> >
> > I'd love to think that Colbert was satirising Popper...
>
> I would too, and it is (remotely!) possible -- but I doubt it.
He studied philosophy at Hampton-Sydney College...
Mitchell Coffey
> Michael Siemon <mlsie...@sonic.net> wrote:
> > In article <1juvshe.106cq96rv31maN%j...@wilkins.id.au>,
> > j...@wilkins.id.au (John S. Wilkins) wrote:
>
> > > Michael Siemon <mlsie...@sonic.net> wrote:
>
> > > > In article <1juv8f5.1j1idyhapgsznN%j...@wilkins.id.au>,
> > > > j...@wilkins.id.au (John S. Wilkins) wrote:
> > > > ...
>
> > > > > OK, as far as I do recall it, and without going to my library and
> > > > > reading for twenty minutes, Popper held that a statement is either true
> > > > > or false. So a false statement cannot be *more* true than another false
> > > > > statement. But it *can* be more "truthlike", or in his terminology, have
> > > > > more verisimiltude.
>
> > > > > Popper's view on probabilities is discussed at SEP well:
>
> > > > >http://plato.stanford.edu/entries/popper/#ProKnoVer
>
> > > > I think the current idiom is Stephen Colbert's "truthy". :-)
>
> > > I'd love to think that Colbert was satirising Popper...
>
> > I would too, and it is (remotely!) possible -- but I doubt it.
>
> He studied philosophy at Hampton-Sydney College...
Colbert seems to be satirizing the American Right's contempt for
conclusions based on facts and logical analysis, rather than gut-
feeling and authority.
Mitchell