Problem of Induction

Ender · Post by **Ender** » 2008-05-21 07:09pm

Hey, a little help. A poll about whetehr you believe in religion or science has grown into an argument with a fool. It started with claiming that science requires faith. I pointed out that since faith is belief without evidence and science is drawing conclusions based off evidence, the two were dynamic opposites. Another fool pipes up challenging that I must be using Dawkin's definition of faith. He got hit with the dictionary definition and then claimed my statement that since science was based off evidence I was only pushing ti back a step and then went on a rant about how Hume and the Problem of Induction mean the science still requires faith.

And as Hume himself has said, "To say [the inference that the future will be like the past] is experimental [i.e., based on experience], is begging the question. For all inferences from experience suppose, as their foundation, that the future will resemble the past, and that similar powers will be conjoined with similar sensible qualities. If there be any suspicion that the course of nature may change, and that the past may be no rule for the future, all experience becomes useless, and can give rise to no inference or conclusion. It is impossible, therefore, that any arguments from experience can prove this resemblance of the past to the future; since all these arguments are founded on the supposition of that resemblance."

Thus, even the scientist has to rely on an unwarranted principle in order to continue doing science.

I responded

This is one of the greatest displays of ignorance of the scientific method I have ever seen. The scientific method accounts for the problem of induction in its revision stage - if the results are different from hypothesis then you modify the initial hypothesis. It does not require past principles to be the same at all. If they were to change then you will have discovered new variables in your experiment and need to examine them as well. No assumptions are required to conduct an experiment via the scientific method, in fact in many cases the scientists are hoping for different results.

He came back with

Gosh, what a sweetheart. Anyways, are you saying that the scientific method doesn't assume the uniformity of nature? And that the future will resemble the past? If so, then it still has to answer to the problem of induction. Furthermore, an experiment isn't in a constant state of flux, like Schroedinger's cat. Somewhere along the lines, a final conclusion has to be drawn. But, according to Hume, whether it's the law of gravity or the second law of thermodynamics, or whatever else, there's no logically justified reason to believe that any of these principles will still be applicable a year from now, or that they may apply to unobserved objects. So you missed the point by a country mile, but nice try regardless.

I came back by pointing out that the scientific method is a process, not a premise, and pointing to the fact that conclusions are not fixed, citing Newton vs Relativity and the fact that the physical constants are still tested again and again.

Any good responses to this? I vaguely recall that a few simple experiments will prove that things will be consistent, verifying casuality, etc. But I was hoping for somethign a bit more concrete.

Rye · Post by **Rye** » 2008-05-21 07:57pm

There's a couple of responses to this. The short hand is that all knowledge is gained from the past, and while it may be logically possible for the future to suddenly become wildly different, like the sky turns to blancmange, the universe becomes Richard Nixon, etc, it's not logically proposable as a possible future, because it has no epistemological basis, since all knowledge is gathered from the past.

Secondly, a consistent universe offers predictions that a chaos-verse doesn't. With a consistent universe, you can predict where the sun will rise tomorrow. You can use a computer, a system which only works because the system is consistent. You can predict how many times a coin will land heads up etc. Given that it actually works and has useful predictions, athere's no pragmatic or epistemological reason to doubt it.

The long hand way of explaining it is reading up on Bayesian induction.

Terralthra · Post by **Terralthra** » 2008-05-21 08:02pm

Um, he's actually right, on one level. The scientific method is predicated on the universe being the same next time as it is this time. This is an element of faith, in Hume's definition, as there is no logically deductive manner in which to arrive at the conclusion that the universe will react in the same manner next time as it does now. It's pure induction, which can never be 100% certain.

The trick being, it's a short step from abandoning that premise to pure skeptical solipsism. I suppose it's possible to live as if causation and consistent reality are non-existent, but I highly doubt the person arguing with you actually does. Science requires the article of faith that the universe will continue to act in the manner it has so far with respect to physical laws, causation, etc. It's not exactly a huge leap of faith, given that so far, as far as we can tell, it has. It's a very pragmatic assumption to make.

Post by **Surlethe** » 2008-05-21 09:53pm

Discard the notions of surety and induction. Instead, focus on the fact that the scientific method is an algorithm for generating confidence in a hypothesis.

An alternative tack is to admit that, yes, the scientific method requires some faith in the consistency of reality. But religion requires more faith: namely, every religious person has faith in the consistency of reality (by dint of operating as a person); but they also believe in something more as well.

Rye · Post by **Rye** » 2008-05-22 07:12am

It's not by faith that you accept your senses and intuitively grasp logic. They evolved that way because they produced a reliable means of survival. If you flip a coin, you don't need faith to predict that half the time it will land heads up, you'd definitely need faith to say that it would never land, since all knowledge up to that point has been reliable and useful, making such a claim irrational to propose in the first place.

Since a consistent behaving universe offers predictions with observable outcomes, you can test that hypothesis all you want. It is beyond reasonable doubt, it has become irrational (and faith-based) to seriously doubt its truth rather than merely being open to new information.

Post by **Alyrium Denryle** » 2008-05-22 01:01pm

Your best response is to the acknowledge the problem of induction, but mitigate it.

1. Science does not seek to prove a hypothesis true. Technically speaking it is trying to disprove it. It is much easier using inductive logic to prove that a proposition is not true.

2. Science also only makes probabilistic statements. Scientists do not actually claim that the universe is going to be the same tomorrow as it was yesterday, only that it probably will be based upon every observation we have ever done. It requires a certain type of the faith, yes. But it is a different type of faith and the idiot you are debating with is using a fallacy of equivocation. When a word like faith has two different dictionary meanings, one does not get to use one definition to argue that a person also uses the other.

1. Confident belief in the truth, value, or trustworthiness of a person, idea, or thing.
2. Belief that does not rest on logical proof or material evidence. See Synonyms at belief, trust.
3. Loyalty to a person or thing; allegiance: keeping faith with one's supporters.
4. often Faith Christianity The theological virtue defined as secure belief in God and a trusting acceptance of God's will.
5. The body of dogma of a religion: the Muslim faith.
6. A set of principles or beliefs.

Definition 1 is the the one scientists have. A certain methodological trust in the idea that the universe is consistent that is necessary to function.

He is trying to equivocate that definition with definition 2 or 5, which it is not. He is being an intellectually dishonest little shitbag.

Terralthra · Post by **Terralthra** » 2008-05-22 01:48pm

Zuul wrote:It's not by faith that you accept your senses and intuitively grasp logic. They evolved that way because they produced a reliable means of survival. If you flip a coin, you don't need faith to predict that half the time it will land heads up, you'd definitely need faith to say that it would never land, since all knowledge up to that point has been reliable and useful, making such a claim irrational to propose in the first place.

Since a consistent behaving universe offers predictions with observable outcomes, you can test that hypothesis all you want. It is beyond reasonable doubt, it has become irrational (and faith-based) to seriously doubt its truth rather than merely being open to new information.

You just said, essentially, that evidence from the past will be predictive of the future because evidence from the past shows that to be true. That's begging the question.

Rye · Post by **Rye** » 2008-05-22 02:11pm

Terralthra wrote: You just said, essentially, that evidence from the past will be predictive of the future because evidence from the past shows that to be true. That's begging the question.

No, I said that it's irrational to propose a prediction that'd be impossible to predict from the current evidence. All evidence from the past was the future relative to some other point in the past. Why wouldn't that trend continue beyond this present? There's no knowledge to suggest the "temporal existential doubt" is reasonable, tenable or believable, it is just metaphysical sophistry.

Terralthra · Post by **Terralthra** » 2008-05-22 02:39pm

Zuul wrote:
Terralthra wrote: You just said, essentially, that evidence from the past will be predictive of the future because evidence from the past shows that to be true. That's begging the question.
No, I said that it's irrational to propose a prediction that'd be impossible to predict from the current evidence. All evidence from the past was the future relative to some other point in the past. Why wouldn't that trend continue beyond this present? There's no knowledge to suggest the "temporal existential doubt" is reasonable, tenable or believable, it is just metaphysical sophistry.

You just did it again. You can't use "current knowledge" (which is shorthand for "the evidence from the past") to show that evidence from the past will predict the future. That's either circular logic or begging the question, depending on how you look at it. The knowledge we have from "evidence from the past" is a purely inductive body, and can NOT make 100% predictions.

Of course there is no evidence that the universe is not consistent. The point is that there can not be evidence that the universe is not consistent, by definition. Moreover, taking a consistent universe as a given and then saying that it's reasonable to assume the universe is consistent is pure logical garbage, and calling the lack of that massive assumption 'sophistry' is in itself, sophistic.

Kanastrous · Post by **Kanastrous** » 2008-05-22 04:16pm

I'm mostly unequipped for this kind of question, but I leads me to ask whether the fact that the observable consistent behavior of an electron (for example) is tied to the nature of the electron, and therefore to suggest that tomorrow electrons might suddenly begin behaving differently, would mean that tomorrow they might suddenly alter their nature. And with evidence over a long span of time that electrons have not altered their fundamental nature, isn't the burden on the other party, to demonstrate as reasonable the claim that they might, plus a mechanism as to how they would do it?

Wyrm · Post by **Wyrm** » 2008-05-22 05:58pm

Terralthra, there is a world of difference between taking the consistency of the universe as a reasonable conclusion based on its past consistency, and proving that the universe is consistent based on this past consistency. You are correct that using the past consistency of the universe to prove the future consistency of the universe is a circular argument (which is the same thing as begging the question, by the way).

But this is not what Zuul claimed. He said that it was irrational to assume that the universe will suddenly go bonkers tomorrow, because the past evidence doesn't support such a proposition. This is a decision theoretic argument. It doesn't make sense to wake up tomorrow and try to figure out the way the world works all over again from scratch, because that would be irrational to behave like all the rules have been thrown out the window unless you find a good reason to believe such (for instance, your coffee taking the consistency of rubber), given that it's never done so before. Especially since assuming the contrary will cost you dearly.

Saying that there is no absolute, mathematical proof that the universe will be consistent in the future is fine and even, strictly speaking, valid. But to then to put the rational choice of treating the universe as if it will continue to do so on equal level to wholly unsupported faith in the Almighty is insane.

Rye · Post by **Rye** » 2008-05-22 07:07pm

Terralthra wrote: You just did it again. You can't use "current knowledge" (which is shorthand for "the evidence from the past") to show that evidence from the past will predict the future. That's either circular logic or begging the question, depending on how you look at it. The knowledge we have from "evidence from the past" is a purely inductive body, and can NOT make 100% predictions.

That's because most human knowledge is imprecise. Nobody with sense would require metaphysical certainty that the sun will rise in the east tomorrow in order to predict that it will. Likewise, metaphysical uncertainty is not sufficient reason to claim that it won't or reasonably doubt that it will.

Of course there is no evidence that the universe is not consistent. The point is that there can not be evidence that the universe is not consistent, by definition.

I don't see how that follows.

Moreover, taking a consistent universe as a given and then saying that it's reasonable to assume the universe is consistent is pure logical garbage,

Taking a consistent universe as an observation and then saying it's reasonable to conclude the universe is consistent is not logical garbage, it is the basis of all knowledge. To contend otherwise is to steal the concept; you use language and concepts from knowledge that you fundamentally deny the significance of once you accept an inconsistent universe as potentially as likely as a consistent one.

and calling the lack of that massive assumption 'sophistry' is in itself, sophistic.

It is sophistry, since it's impossible to conclude from reality, even if it's true, it is a notion couched in perfectly defined ignorance to forever protect it. It is "metawankery".

Ender · Post by **Ender** » 2008-05-22 08:56pm

Ok, I went with Surthele's recommendations of shifting it to insisting that science is a process rather then a body of facts and that the scientific method is a way of generating confidence for a theory. My opponent has taking the interesting role of declaring these statements to be false. And I don't mean just disagreeing, he flat out said that the scientific method does not generate confidence in a theory. He has also launched in to declaring that we do not replicate large scale phenomena on a small scale, so I'm hammering him on things like WTF are thermonuclear warheads then.

Post by **Alyrium Denryle** » 2008-05-23 12:43pm

Ender wrote:Ok, I went with Surthele's recommendations of shifting it to insisting that science is a process rather then a body of facts and that the scientific method is a way of generating confidence for a theory. My opponent has taking the interesting role of declaring these statements to be false. And I don't mean just disagreeing, he flat out said that the scientific method does not generate confidence in a theory. He has also launched in to declaring that we do not replicate large scale phenomena on a small scale, so I'm hammering him on things like WTF are thermonuclear warheads then.

To which the guy that does science for a living responds that he is flat out wrong and having a Philosophy of Science Problem.

Science does not generate facts. It collects facts, collates them into a concete observation about the universe, then goes about designing an experiments or observation regimes to test an explanation for those facts against said facts. It is very simple The more repeats of these tests using different methods, testing different aspects of our proposed explanation, the more confident we grow in said explanation because the goal of the rest is to potentially prove our explanation wrong or in need of revision. The more times it fails to be rejected, the more confident in it we are. The probability of it being true asymptotes toward 100%

Post by **Darth Wong** » 2008-05-23 12:58pm

Ender wrote:Ok, I went with Surthele's recommendations of shifting it to insisting that science is a process rather then a body of facts and that the scientific method is a way of generating confidence for a theory. My opponent has taking the interesting role of declaring these statements to be false. And I don't mean just disagreeing, he flat out said that the scientific method does not generate confidence in a theory. He has also launched in to declaring that we do not replicate large scale phenomena on a small scale, so I'm hammering him on things like WTF are thermonuclear warheads then.

How does he explain the fact that we still call them "theories" instead of "truths"?

Religions do not promote "theories". They claim to possess 'truths".

Science, on the other hand, promotes theories and attempts to justify those theories by showing that they are very good at predicting empirical outcomes, in the same manner that one justifies a new machine by turning it on and showing that it works. It's not about "truth"; it's about effectiveness and accuracy.

Anti-science people always try to attack science by pretending that it claims to have some hold on absolute "truths", the way religion does. It only uses the word "truth" in the most colloquial sense of the word (the same one we use when we give true/false answers to historical trivia questions). The reality is that science is a descriptive modeling system, and a highly effective one, as demonstrated by its application in engineering. It only becomes "truth" because most people agree that empirical evidence is a valid way to seek the truth.

And that's the rub: push an anti-science person long enough, and he will eventually admit that he doesn't think empirical evidence is superior to personal revelation.

Moron wrote:Gosh, what a sweetheart. Anyways, are you saying that the scientific method doesn't assume the uniformity of nature?

The uniformity of nature can be thought of as another theory: one which could be disproven by conducting highly controlled experiments and getting wildly inconsistent results despite every input factor being held constant. The fact that this doesn't happen means that we have fairly high confidence in the uniformity of nature, hence we use that as an assumption when doing other work. It does not require "faith" as religions do.

Ender · Post by **Ender** » 2008-05-23 01:19pm

Ok, guy came back late last night ranting that I was still unable to prove what faith and science were despite providing the definitions and then launched in on comments I and another poster made about logic - he started arguing that I needed to prove logic was a human concept, that I needed to prove it shares qualities with math, arguing that it should be subjective and a few other bits of miscellaneous BS. It was late and I was tired, so I just pointed out every single logical fallacy he made in challenging my statements on logic, said something to the effect of "and you will respond to this with more ad nauseum replies because you can't help yourself. And that is why you are my bitch" so now his ego means he has to settle for losing the argument because he will not admit he is my bitch. Like Mike said it eventually came down to him rejecting empirical evidence in favor of personal reality (in this case the definitions of words), so most people are just bitching about the fact that the argument is taking place.

However in doing a bit more poking around I found out more about Karl Popper and his refutation of Hume's argument. Given that no one cited him here I presume there is something wrong with his argument but nothing is leaping out at me at first blush. Is Popper's argument flawed or incomplete?

Illuminatus Primus · Post by **Illuminatus Primus** » 2008-05-23 01:34pm

Even if science requires some basic assumptions given as true without evidence, the resultant body of understanding conforms very well to human experience of the universe and allows very well controlled, very precise manipulations of it (engineering). Its a MUCH more reasonable assumption than there is an invisible man in the sky no one can see touch or feel or in my closet or room or whatever which improves our genuine understanding of the universe not one iota.

Terralthra · Post by **Terralthra** » 2008-05-23 01:38pm

Ender wrote:However in doing a bit more poking around I found out more about Karl Popper and his refutation of Hume's argument. Given that no one cited him here I presume there is something wrong with his argument but nothing is leaping out at me at first blush. Is Popper's argument flawed or incomplete?

Popper's response to the problem of induction is accurate, but doesn't really answer Hume's point. He addresses the pragmatic question "Why do you believe and act as if xyz?" not the the epistemological question "How can you know xyz?" To my knowledge, no philosopher has been able to satisfactorily answer the problem of induction on anything other than a pragmatic level, Zuul's broken-record posturing aside.

Post by **Darth Wong** » 2008-05-23 01:40pm

I once made a post pointing out some problems with Hume's argument, without relying on Popper:

http://bbs.stardestroyer.net/viewtopic.php?t=54383

That's not to say that there's anything wrong with Popper's argument, it's just that it gives up ground which doesn't need to be given up. Hume's argument assumes that inductive logic is always invalid: an assumption which has no more inherent justification than the assumption that it is always correct.

Wyrm · Post by **Wyrm** » 2008-05-23 01:56pm

Ender wrote:Ok, I went with Surthele's recommendations of shifting it to insisting that science is a process rather then a body of facts and that the scientific method is a way of generating confidence for a theory. My opponent has taking the interesting role of declaring these statements to be false. And I don't mean just disagreeing, he flat out said that the scientific method does not generate confidence in a theory.

Which is demonstratably false. "Confidence" obeys all of the axioms of probability, and the methods of science essentially act as repeated trips through Bayes' Theorem, if you equate "prior/posterior confidence in hypothesis" = "prior/posterior probability of hypothesis", and "hypothesis strongly predicts/weakly predicts or does not predict observation" = "hypothesis assigns much/little or no probability to observation".

The methods of science are more organic than explicit Bayesian updates. In both, a hypothesis that is predictive of an observed outcome (in the language of probability, assigns high probability to the outcome) is awarded with more confidence; a hypothesis that is antipredictive of that outcome (assigns low to 0 probability) is penalized with less confidence. A Bayesian calculation of confidence simply puts an exact value on how much your confidence is adjusted (up or down).

Theories are simply hypotheses that have gained so much confidence, that it is irrational (decision theory, again) to not to take them as a given without some serious evidence to the contrary.

Post by **Surlethe** » 2008-05-23 01:56pm

Darth Wong wrote:That's not to say that there's anything wrong with Popper's argument, it's just that it gives up ground which doesn't need to be given up. Hume's argument assumes that inductive logic is always invalid: an assumption which has no more inherent justification than the assumption that it is always correct.

I think this doesn't do justice to the original argument, which, if I read it correctly, posits that inductive logic is simply not deductively valid. It seems more economical to argue against the binary truth values or to argue that science is not inherently a deductive system (as opposed to mathematics or theology) than to argue against that premise.

Post by **Surlethe** » 2008-05-23 02:01pm

Darth Wong wrote:Anti-science people always try to attack science by pretending that it claims to have some hold on absolute "truths", the way religion does. It only uses the word "truth" in the most colloquial sense of the word (the same one we use when we give true/false answers to historical trivia questions). The reality is that science is a descriptive modeling system, and a highly effective one, as demonstrated by its application in engineering.

One can think of the scientific method as an algorithm that takes the body of empirical evidence and spits out models that account for the evidence as it accumulates. In another sense, think of the space of all the possible worlds and start with one as a model of reality; then as empirical evidence accumulates, the scientific method will generate a (Cauchy) sequence of worlds within that space.

It only becomes "truth" because most people agree that empirical evidence is a valid way to seek the truth.

On this point, it might be best to appeal to experience; the basic level of 'truth' that we take so for granted we don't even question it is the reality of the world around us. We know it's true because ... of the empirical evidence that our senses provide us. Even the solipsistic types act as though it's true.

Illuminatus Primus · Post by **Illuminatus Primus** » 2008-05-23 02:12pm

Yeah, but that's the problem with hide-bound religious types, they want to insist their faith is on the same level of intuition. We've all experienced this excuse (even though its clearly untrue, namely that people believe all sorts of distinctly different gods and superstitions, and they always do so based on what their parents believe - these are cultural values, not the deep level of truly universal human intuition).

Adrian Laguna · Post by **Adrian Laguna** » 2008-05-23 02:15pm

Hume's argument in a nutshell: You cannot have absolute proof of anything.

Short reply to it (which Hume himself agreed with): We don't need absolute proof of anything. It is acceptable to make some assumptions about the world, such as time flow, symmetry, and predictability, then build our understanding of the universe upon them. For the simple reason that this method comes naturally to human perception, and that it works.

I think the very last part if of great importance. Typically in an argument you don't want to just defend your position, but also mount an effective attack against the other's position. In this particular case, the best attack is to point out that science has given us modern industrial society, then ask what does [insert brand of sophist bullshit being peddled] has given us? Where are its great accomplishments? Even the Cathedrals build for the Glory of God are held-up by scientific principles, not the prayers of the faithful.

Rye · Post by **Rye** » 2008-05-23 02:18pm

Surlethe wrote: On this point, it might be best to appeal to experience; the basic level of 'truth' that we take so for granted we don't even question it is the reality of the world around us. We know it's true because ... of the empirical evidence that our senses provide us. Even the solipsistic types act as though it's true.

I'd also point out there's no other way of learning of the notion of truth without it. If you're asking what the truth of the matter is, you had to get that concept from your sensory experience of life.