Your thoughts on Hume's case against induction?

[aynfan]

"When on innumerable occasions we observe certain experiences succeeding others, we naturally feel under similar circumstances in the future like events or causes will be followed by like effects. ... only custom or habit may validly be said to serve as the foundation for this causal idea."17 There is no guarantee, no matter how accustomed we may have become to certain sequential events of the past that the sequence will necessarily repeat itself. He concluded that the "whole of our science assumes the regularity of nature - assumes the future will be like the past ..."

[msb]

Is there a question here?

[DPW]

"When on innumerable occasions we observe certain experiences succeeding others, we naturally feel under similar circumstances in the future like events or causes will be followed by like effects. ... only custom or habit may validly be said to serve as the foundation for this causal idea."17 There is no guarantee, no matter how accustomed we may have become to certain sequential events of the past that the sequence will necessarily repeat itself. He concluded that the "whole of our science assumes the regularity of nature - assumes the future will be like the past ..."

Yeah, and he was wrong. EDIT: Uh, not about that last part (although it isn't an assumption) - about the first part.

[aynfan]

Is he correct in doubting the predictablity of future events?

[Spearmint]

I'd say that Hume's argument here is correct, it just doesn't have the implications that many beiieve it does. All hes really saying is that its impossible to obtain knowledge of some kind of 'transcendental' casuality rather than the sequencing between events that we obtain via our senses. Since Objectivism rejects outright the possibility of non-contextual knowledge such as this, it isnt really a problem.

The other part of Hume's argument is (to simplify) that you can't validate induction by means of deduction. This is also correct - you can't validate deduction by means of induction either. The fallacy comes when the claim is made that this somehow makes induction invalid, when in fact both induction and deduction are primary but independent sources of gaining knowledge. Technically, you couldn't follow a deductive argument without relying on induction anyway ("how do I know these symbols mean what I think they do? I remember reading their definitions in a book, but how can I prove my memory is valid?" "Did I actually prove this argument five minutes ago? I'm almost certain that I remember doing so, but I can't be sure!")

[danielshrugged]

I was going to answer your question, but I disappeared into thin air first. I hate when that happens.

[Spearmint]

Is he correct in doubting the predictablity of future events?

It depends what you mean by 'doubt'. It's never possible to achieve absolute 100% certainty, but it would be extremely irrational to doubt when faced with a 99.9999% chance. It's often claimed that the inability to achieve 100% certainty means that human's can never obtain 'true knowledge', but all it actually represents is the fact that humans aren't omniscient. If you have no rational grounds on which to doubt something then it is correct to say that you have certainty - to claim otherwise is to commit the fallacy that Rand accused Kant of commiting, namely to assert that because all knowledge must be contextual to a given consciousness, it pales when compared to "real" knowledge which is somehow obtained without the use of means of perception (in other words, to forget that consciousness has identity).

[DavidOdden]

It depends what you mean by 'doubt'. It's never possible to achieve absolute 100% certainty, but it would be extremely irrational to doubt when faced with a 99.9999% chance.

Well, actually it is possible to possible to achieve absolute 100% certainty. You're imposing a requirement of infallibility on certainty, which is why it seems to be impossible to be certain. Certainty simply means that the evidence is so compelling that there are no reasonable grounds for doubting. Furthermore, if there is some reason to doubt, then it is not reasonable to deny that reason.

If you have no rational grounds on which to doubt something then it is correct to say that you have certainty - to claim otherwise is to commit the fallacy that Rand accused Kant of commiting, namely to assert that because all knowledge must be contextual to a given consciousness, it pales when compared to "real" knowledge which is somehow obtained without the use of means of perception (in other words, to forget that consciousness has identity).

Which means that 100% certainty is possible. I think the view that certainty has to do with "chances" is where the problem lies. Certainty doesn't come by accident or random chance: it's a statement about a proposition and the sum of your knowledge. So there is no element of chance -- but there is the fact that we are not omniscient to contend with.

{Footnote: if by using the word "absolute in the first quote you mean "certainty regardless of context", then it's true that there is no such thing as 100% absolute certainty, or even absolute .01% certainty, because "a knowledge state that assumes no knowledge" is a contradiction.}

[Spearmint]

Well, actually it is possible to possible to achieve absolute 100% certainty. You're imposing a requirement of infallibility on certainty, which is why it seems to be impossible to be certain. Certainty simply means that the evidence is so compelling that there are no reasonable grounds for doubting. Furthermore, if there is some reason to doubt, then it is not reasonable to deny that reason.

But 100% absolute certainty _does_ mean infalliabilty, which is why its unobtainable in most contexts (I exclude trivial situations where 'contextless certainty' is possible, such as when evaluating "there are no square circles" and so on). If one of your beliefs turns out to have been wrong, then you couldnt have been 100% certain about it unless youre redefining the meaning of the word.

I think our disagreement is that you're thinking of certainty as being some kind of binary quantity, where a person either "is certain" or "isnt certain". However, I dont think this is a particularly good way of describing certainty. For instance, I'm certain that I have a pint of milk in my fridge. However I'm even _more_ certain that my computer wont blow up with in the next 5 seconds, and my certainty that gravity wont reverse ,causing things to fall upwards, even exceeds my certainty of the previous 2 events. As such, it makes more sense to think of certainty as being some kind of scale, such as

Absolute 100% --------------------------------------------------------- 0% Lack of 

 

[F*ckCommunism]

if you have 0% certainty, then aren't you certain you are uncertain?

[Spearmint]

if you have 0% certainty, then aren't you certain you are uncertain?

I'd say 0% certainty would just be complete lack of knowledge with absolutely no rational justfication for believing any particular hypothesis. For instance if a stranger stopped me in street and asked me to guess his name, then I'd say I had something close to 0% certainty in any particular guess that I came up with (I say 'close to 0%' rather than 0%, because although I'd have no rational justifcation for believing his name was "Stuart", I would have more certainty of it being "Stuart" than of it being something like "Zycholpikkantis")

[Betsy]

Well, actually it is possible to possible to achieve absolute 100% certainty. You're imposing a requirement of infallibility on certainty, which is why it seems to be impossible to be certain.

I think absolute certainty is possible in induction for the same reason that "A is A" is absolutely certain.

Betsy's Law for Inductive Certainty - "If you know, or can discover, enough about an entity to reduce your inductive conclusion about it to a statement of identity, you can be certain."

[oldsalt]

Spearmint: One can be certain -- within context. Knowledge is contextual, don't forget. There can be no such thing as possibility (or probability) without certainty. How would you even assess such a thing. It's like saying that there is gray, but no black or no white.

I know with a certainty that Existence Exists. I know with a certainty that A is A. I know with a certainty that if a metal projectile of a certain size(s) hits my flesh at a certain volocity, that it will damage my body. I know with a certainty that if I suddenly blow 90% of my left ventricle, I will die. I know with a certainty that if I leap off of the top of the Empire State Building, I will be flattened out of all recognition upon hitting the ground.

I now know, with 100% certainty, that if a Boeing 767, loaded with fuel, hits a building designed, engineered, and constructed in a particular way, that that building will fall to the ground. I've seen it happen twice and I do not ever have to see it again to know with a certainty that it will happen again under the same circumstances. I don't need any tradition to know this, nor do I need a vote of the people to know this. I know it from direct experience of reality.

[Fred Weiss]

To assume a range of certainty (or probability) in regard to something presupposes some certainty or you end up in an infinite regress. For example, if you were to say that you are 90% certain of something, how do you know it is 90% and not, say, 80%? If you are unsure about it being 90%, then you have added an additional layer of probability onto your probabilty. What do you say then? That you are 90% sure that it is 90% probable. And so on.

If we can't be certain of anything, then we can't know *anything*. It totally demolishes knowledge down to its roots. There is no remnant of "probability" which can salvage it.

If you just hold in mind some things you know to be certain, and toward which there is absolutely no doubt, it will help you to grasp this idea.

My own favorites (but you could pick any of 1,000's of other examples) are:

A cow cannot jump over the moon.

The earth orbits the sun.

The WTC was destroyed on 9/11.

It is not enough to say, "Well, I can imagine a cow jumping over the moon" or "I can imagine the earth doing pirouettes around Jupiter, instead of orbiting the sun", so therefore it's possible. You can imagine anything you wish, but reality isn't a Walt Disney cartoon. You can imagine these things *only by dropping the context of all of human knowledge* and in these particular instances, all of our knowledge of physics.

Fred Weiss

[DavidOdden]

I think our disagreement is that you're thinking of certainty as being some kind of binary quantity, where a person either "is certain" or "isnt certain".

I'm actually thinking of conclusions as being certain, not people. It's a small but important difference. But you're absolutely correct that I hold certainty to be an is / isn't concept. In this respect, I agree with Peikoff in OPAR Ch. 5.

However, I dont think this is a particularly good way of describing certainty. For instance, I'm certain that I have a pint of milk in my fridge. However I'm even _more_ certain that my computer wont blow up with in the next 5 seconds, and my certainty that gravity wont reverse ,causing things to fall upwards, even exceeds my certainty of the previous 2 events.

What you say here is consistent with the view of certainty as an emotive state, one hopefully but not necessarily related to observation. In other words, you consider "certainty" to be exactly the same thing as "the scale of evidence". If that is your view, then it's very straight-forward that you can never have 100% certainty because there is no upper bound on how much evidence there could be (not just that you are aware of, but that does exist and that could exist in any mind). There are huge problems with trying to define certainty as based on a Platonic "universal context", so I'll assume that you don't mean such a thing. But then I don't know what you do mean by certainty.

If the context where this scale is judged is one's own knowledge context, then you can have 100% certainty: all of your knowledge points to the conclusion, as I siad and as Peikoff says (this is not argumentum ad verecundiam, just mentioning the relevant literature). If you don't mean the context of an individual's knowledge, then I can't figure out what context you have in mind. And if you don't relate certainty to knowledge, then I don't have any idea what you mean by certainty.

where some things are more certain that others. The ends of the scale, 0% and 100% certainty, are normally unreachable

I don't understand why 0% certainty is unreachable in your epistemology.

(since humans are omnscient)

I didn't know that. Does that refute the claim?

[DPW]

But 100% absolute certainty _does_ mean infalliabilty, which is why its unobtainable in most contexts (I exclude trivial situations where 'contextless certainty' is possible, such as when evaluating "there are no square circles" and so on).

But that makes no sense at all. If we were infallible, we wouldn't need the concept "certain". What you're saying, therefore, is the fact we need the concept "certain" proves certainty doesn't exist. Doesn't make much sense, if you ask me.

I would say we need the concept certainty because we are fallible - we need it to distinguish between proved ideas and unproved ideas when those ideas are of a type that are validated through an accumulation of evidence over time. To say an idea is certain is to say that one should no longer doubt it. One can't doubt it because doubt, in a rational man's mind, must be based on evidence (or a lack of evidence).

To doubt that which is certain is the same as accepting the arbitrary, and includes all the destructive epistemological consequences that follow from that error.

If one of your beliefs turns out to have been wrong, then you couldnt have been 100% certain about it unless youre redefining the meaning of the word.

Our meanings are the same: a certain idea is one free from doubt. What you're saying is that we must doubt everything (to some degree) because we are fallible, but that's like saying we should all get divorced since some marriages fail.

I think our disagreement is that you're thinking of certainty as being some kind of binary quantity, where a person either "is certain" or "isnt certain". However, I dont think this is a particularly good way of describing certainty.

See, that's the difference between your approach and mine. I don't look for what I think is a good way of describing something - I try to figure out what something is. And look where your approach leads you: to the idea that the best way of describing something is to deny its existence!

[aynfan]

Hume is saying, in effect, that just because the sun comes up everyday, we can not say with certainy that it will rise tomorrow; that anomolies can and do arise in the physical universal over which we have no control and no foreknowledge.

[DPW]

Hume is saying, in effect, that just because the sun comes up everyday, we can not say with certainy that it will rise tomorrow; that anomolies can and do arise in the physical universal over which we have no control and no foreknowledge.

Sure, and that's exactly what is covered by context. When I say, "I am certain the sun will rise tomorrow, in my present context of knowledge," what I am saying is, given all the conditions so far discovered, the sun must "rise" tomorrow. If it doesn't, I know there is a new condition(s) at work, which I will attempt to discover. Notice that if the sun doesn't rise, that does not invalidate my previous conclusion. Under those conditions, the sun would have risen.

[aynfan]

Bravo Don.

[Spearmint]

Sure, and that's exactly what is covered by context.  When I say, "I am certain the sun will rise tomorrow, in my present context of knowledge," what I am saying is, given all the conditions so far discovered, the sun must "rise" tomorrow.  If it doesn't, I know there is a new condition(s) at work, which I will attempt to discover.  Notice that if the sun doesn't rise, that does not invalidate my previous conclusion.  Under those conditions, the sun would have risen.

I don't think youre using 'certainty' in the way that most people use the term. "Certainty" would relate to your degree of believe that you HAVE discovered all the relevant conditions, and that your context of knowledge isnt to expand to the point of invalidating what you currently know (within the new context)

For instance, I am absolutely certain that 2 + 2 = 4. There are no 'new conditions' that are going to ever invalidate this knowledge, nor is there anything that I could possibly discover about the universe that is going to stop 2 + 2 equalling 4. If you are admitting possible new conditions under which your knowledge may not hold then you cannot claim absolute certainty, unless you are redefining the word.

Using what I am assuming your definitions are, I would translate Hume as saying "No matter what we discover about the empirical universe, it is always possible that our context of knowledge will be expanded in ways we cannot now predict", which I think is in agreement with what you have said.

[Spearmint]

DPW

But that makes no sense at all.  If we were infallible, we wouldn't need the concept "certain".  What you're saying, therefore, is the fact we need the concept "certain" proves certainty doesn't exist.  Doesn't make much sense, if you ask me.

I agree with you here. Certainty is essential precisely because we are infalliable - its a way of describing degrees of belief. If we were infallible then we would just say that we were either 0% or 100% about everything in the wolrd (depending on whether it was false or true respectively) - in a way certainty would become synonymous with knowledge. As this is not the case, certainty is useful for describing how confident we are about something - in other words, how much faith we have in our predictions.

Our meanings are the same: a certain idea is one free from doubt.  What you're saying is that we must doubt everything (to some degree) because we are fallible, but that's like saying we should all get divorced since some marriages fail.

No. I explicitly said that it would be irrational to doubt something of which you are 99% certain. The fact that something isnt 100% certain does not mean that it is open to doubt. I do not have doubts that I will make it home safely whenever I leave my house, although obviously I cannot claim 100% certainty on this fact.

[Spearmint]

Fred

To assume a range of certainty (or probability) in regard to something  presupposes some certainty or you end up in an infinite regress. For example, if you were to say that you are 90% certain of something, how do you know it is 90% and not, say, 80%? If you are unsure about it being 90%, then you have added an additional layer of probability onto your probabilty. What do you say then? That you are 90% sure that it is 90% probable. And so on.

Well I wasn't meaning that people should/could assign exact numerical values to how certain they were about something, I was just trying to get across that your certainty about something would be measured relative to your certainty about everything else (as in 'more' certain or 'less' certain) rather than as a binary relation where you either had certainty or you didnt. Numbers help to make this point, but I'm not saying people should actually think in terms of them.

Let's take 'love' as an analogy. Love isnt a binary thing either - its not a case of either you love someone or you dont, because its possible to love two people yet love one more than the other. I would describe love in the same way as I described certainty, namely to think of it as being a line with one end corresponding to 'absolute true love!" and the other end being "no love whatsoever", which could be represented by 100% love and 0% love respectively. In practice, you wouldnt ever think something like "Well I loved my ex-girlfiriend 88.3543% yet I only love you 82.2%", you'd just think "I love you slightly less than my ex girlfriend". I was using numbers to illustrate this principle, not as a way I'd expect people to actually function.

(edit: and like certainty, you wouldnt need '100% love' in order to claim that you loved someone. If you only loved a person 93% or so it would still be perfectly rational to claim that there was love. To remove the numbers, I might love my wife, get divorced, and then meet someone that I end up loving even more. Does this mean that I didnt 'actually' love my wife because it was possible for me to feel a greater degree of love towards someone else? No, it just meant that I didnt actually have '100% love' in the first case, so to speak).

[DPW]

I don't think youre using 'certainty' in the way that most people use the term.

Yeah, and it leads to Hume and every other skeptical error.

"Certainty" would relate to your degree of believe that you HAVE discovered all the relevant conditions, and that your context of knowledge isnt to expand to the point of invalidating what you currently know (within the new context)

You keep saying this and I'm asking you, why? Why must that, which we both agree is generally impossible, be the definition of certainty? The only reason you've so far offered is, "That's how people use the term." Well, I don't know about you, but I'm certain that's a stupid reason.

For instance, I am absolutely certain that 2 + 2 = 4. There are no 'new conditions' that are going to ever invalidate this knowledge, nor is there anything that I could possibly discover about the universe that is going to stop 2 + 2 equalling 4. If you are admitting possible new conditions under which your knowledge may not hold then you cannot claim absolute certainty, unless you are redefining the word.

I thought I was clear on this point. New knowledge does not invalidate previous knowledge. "The sun will rise under a given set of conditions" is not contradicted by "The sun will not rise under a different set of conditions."

This is not to say context makes us infallible. Nothing makes us infallible. But fallability results, not from the expansion of our context of knowledge, but deviation from the proper method of gaining knowledge.

In neither case, however, can certainty demand omniscience and infallbility and remain a meaningful concept. That's why I can't imagine why you are fighting so hard to define certainty in such a way as to make it useless, meaningless, non-referential.

I've justified my use, the Objectivist use, of certainty by explaining its cogntive role. I could also point out its moral role (i.e., its role in human action). You can't wipe that aside by saying, "I don't like using certainty that way." There is no way around it, my friend: context and fallability are not the enemies of certainty, but its precondition.

Using what I am assuming your definitions are, I would translate Hume as saying "No matter what we discover about the empirical universe, it is always possible that our context of knowledge will be expanded in ways we cannot now predict", which I think is in agreement with what you have said.

That I agree with completely. The question is, what conclusions do we draw from that fact? You seem to say, "We can know nothing much for certain." I say, "Certainty must take into account context." The problem with your point of view is that it's wrong. Of that, you can be certain.

[DavidOdden]

For instance, I am absolutely certain that 2 + 2 = 4. There are no 'new conditions' that are going to ever invalidate this knowledge

BTW, in base 3 arithmetic, 2+2=11. In base 4 math, 2+2=10. So I guess you mean, in the context of decimal arithmetic. I'm not certain whether you knew that.

You haven't explained what you think certainty is. Is it an emotion? While you're explaining that, can you explain why it is irational to doubt a conclusion that is 99% certain, but rational to doubt a conclusion that is 98% certain? Or if doubting at the 98% level is still irrational, is it rational to start doubting when you're 97% certain. How do you relate the concept of certainty to rationality?

[nimble]

david: the bases dont actually matter. They will on paper. But not in a real situation. You will always be able to say "I have 2 apples, if i get 2 more I will have 4."