Hume on why we use induction
In the last forum, which no longer exists, I've made numerous topics on the problem of induction, trying to justify it (to the annoyance of many members). Given all the negative responses in my previous discussions about induction I hesitate to bring one up now. So today I am being more modest, I no longer have that ambition to try to justify induction. I agree that induction can't be justified using deduction, and to try to justify it inductively would be arguing in a circle. I'm framing my discussion not as an attack on Hume, but a desire to understand him.
That out of the way, let me begin. By Hume's logic, any account of nature, which includes human nature, that goes beyond observations and our memory is totally without justification. You can't say anything justifiably about human nature other than the combined observations of our actions and our thoughts and feelings, all of which happened in the past. You can't even generalize to all humans.
However, in An Enquiry Concerning Human Understanding, Hume states:
[b][i]Suppose, again, that he has acquired more experience, and has lived so long in the world as to have observed familiar objects or events to be constantly conjoined together; what is the consequence of this experience? He immediately infers the existence of one object from the appearance of the other. Yet he has not, by all his experience, acquired any idea or knowledge of the secret power by which the one object produces the other; nor is it, by any process of reasoning, he is engaged to draw this inference. But still he finds himself determined to draw it: And though he should be convinced that his understanding has no part in the operation, he would nevertheless continue in the same course of thinking. There is some other principle which determines him to form such a conclusion.
36. This principle is Custom or Habit. For wherever the repetition of any particular act or operation produces a propensity to renew the same act or operation, without being impelled by any reasoning or process of the understanding, we always say, that this propensity is the effect of Custom. By employing that word, we pretend not to have given the ultimate reason of such a propensity. We only point out a principle of human nature, which is universally acknowledged, and which is well known by its effects.[/i][/b]
What I find odd is that right after Hume talks about how induction is unjustified he seems to go on and use it. I think if Hume were to be consistent he would say induction is unjustified and just leave it at that, end of story. He doesn't do that, rather he tries to give an account on why humans use induction generally. Hume claims we use induction because of "custom" or a "habit" (his words) where we are so used to things happening in a predictable way that we come to expect nature to continue to be predictable. My question is, how does he know that? Has he observed the use of induction in all people? What warrant gives Hume the right generalize that all humans and animals possess this induction instinct, past, present, and future? By Hume's account there is no justification for such generalities. Why should we believe him then?
From what I see, it's not only that Hume uses induction when he argues that induction is unjustified, Hume goes on to use induction pointing out evidence that our use of induction is on shaky ground (habit, custom). It's also important to note that Hume is not doing a reductio ad absurdum on induction either. He's not saying, "assuming induction is rational when we use it in such and such case we find out that it is not rational by its own account, thus contradicting our assumption that it is rational". Rather it seems Hume is using induction as if it were a perfectly fine method to show that, after all, it's a rather arbitrary method (habit, custom).
I find it hard to believe that Hume's inconsistency with respect to his use of induction would go unnoticed by major philosophers. So I ask: do philosophers agree that Hume is inconsistent in his Enquiry? if not, how do you explain away Hume's supposed inconsistency?
That out of the way, let me begin. By Hume's logic, any account of nature, which includes human nature, that goes beyond observations and our memory is totally without justification. You can't say anything justifiably about human nature other than the combined observations of our actions and our thoughts and feelings, all of which happened in the past. You can't even generalize to all humans.
However, in An Enquiry Concerning Human Understanding, Hume states:
[b][i]Suppose, again, that he has acquired more experience, and has lived so long in the world as to have observed familiar objects or events to be constantly conjoined together; what is the consequence of this experience? He immediately infers the existence of one object from the appearance of the other. Yet he has not, by all his experience, acquired any idea or knowledge of the secret power by which the one object produces the other; nor is it, by any process of reasoning, he is engaged to draw this inference. But still he finds himself determined to draw it: And though he should be convinced that his understanding has no part in the operation, he would nevertheless continue in the same course of thinking. There is some other principle which determines him to form such a conclusion.
36. This principle is Custom or Habit. For wherever the repetition of any particular act or operation produces a propensity to renew the same act or operation, without being impelled by any reasoning or process of the understanding, we always say, that this propensity is the effect of Custom. By employing that word, we pretend not to have given the ultimate reason of such a propensity. We only point out a principle of human nature, which is universally acknowledged, and which is well known by its effects.[/i][/b]
What I find odd is that right after Hume talks about how induction is unjustified he seems to go on and use it. I think if Hume were to be consistent he would say induction is unjustified and just leave it at that, end of story. He doesn't do that, rather he tries to give an account on why humans use induction generally. Hume claims we use induction because of "custom" or a "habit" (his words) where we are so used to things happening in a predictable way that we come to expect nature to continue to be predictable. My question is, how does he know that? Has he observed the use of induction in all people? What warrant gives Hume the right generalize that all humans and animals possess this induction instinct, past, present, and future? By Hume's account there is no justification for such generalities. Why should we believe him then?
From what I see, it's not only that Hume uses induction when he argues that induction is unjustified, Hume goes on to use induction pointing out evidence that our use of induction is on shaky ground (habit, custom). It's also important to note that Hume is not doing a reductio ad absurdum on induction either. He's not saying, "assuming induction is rational when we use it in such and such case we find out that it is not rational by its own account, thus contradicting our assumption that it is rational". Rather it seems Hume is using induction as if it were a perfectly fine method to show that, after all, it's a rather arbitrary method (habit, custom).
I find it hard to believe that Hume's inconsistency with respect to his use of induction would go unnoticed by major philosophers. So I ask: do philosophers agree that Hume is inconsistent in his Enquiry? if not, how do you explain away Hume's supposed inconsistency?
Comments (32)
But even if he were inconsistent on that point and really used induction to conclude that Custom is a universal principle, that wouldn't show that there is no problem of induction, but rather that it's sometimes hard to see the problem, even for Hume!
I think Hume's point was to show the arbitrariness in our propensity to draw generalization. The words "habit" and "custom" are his words. Why does he use those words? What is he trying to convey?
It seems like there's no justification for his conclusion that we all have that propensity. I see no reason to agree with him.
Quoting leo
Granted, but doesn't Hume have his reason why he calls the propensity a "habit" when he could have just stuck with the word "propensity"?
Quoting leo
I think there two propositions here regarding induction:
1. The use of induction cannot be justified. (Logical claim.)
2. Our use of induction is based on something arbitrary such as a habit, or a custom. (Empirical claim.)
I do agree that first proposition is really a problem. However I don't see how the second one can really be demonstrated.
Furthermore, if induction was such a mistake, then how come Bob Newheart has not been inducted into the TV Hall of Fame yet?
P.s. I have not condcted proper research for the background of my post. Thus, I am not sure if Induction ovens exist, and/or if Bob Newheart is in the Tv Hall of Fame or not. But you get the gist of my intent.
And that is the only kind of justification which Hume is seeking to deny to induction. He isn't trying to discredit induction as an enterprise, he is only asking how, in general, it might get done. Rather than inventing some infallible method, he quite sensibly looks at how animals generalise beyond their experience. You can question the flow of sources backing up his biological and psychological claims about the role of habit, but you can't say he is inconsistent. Quite the opposite. He is probably reaching these claims through the exercise of precisely those habits of mind that he ends up claiming are efficacious (though fallible) for such a purpose.
Quoting Purple Pond
Nope! He approves of it.
Quoting Purple Pond
Nope! He loves it.
Quoting Purple Pond
Nope! Not arbitrary, depends on experience.
My reading of Hume gives me the impression that he denies that there's any kind of justification of induction. What kind of justification do you have in mind that he would approve of?
Quoting bongo fury
Hume may believe that induction is efficacious, however that doesn't excuse him from coughing up justification in order to persuade others. Put it this way, how is he supposed to convince me of his inductive conclusion is accurate when he puts serious doubts on the very process of induction?
Quoting bongo fury
He approves of it in the sense that when he stops philosophizing he forgets all his skeptical doubts and goes on with his day. It doesn't mean that he thinks that there is any rational reason why induction will continue to work.
Quoting bongo fury
I specifically mentioned his words, "custom", and "habit". To me it makes it seem like Hume is implying that our inductions are quite arbitrary. You say, "depends on experience" but for Hume there is no warrant for extrapolating beyond what we observe and remember. If you say that he does warrant such things, I'm dying to know what his warrant is.
Habit, custom.
Quoting Purple Pond
Why not? Do you not advocate science, on the grounds of its success?
Quoting Purple Pond
He doesn't put doubts on the process. Like I said, he only wants to recommend that you give up the futile search for a justification in deduction.
Quoting Purple Pond
Which, if I remember rightly, he implies was a relatively louche existence. But if he was a scientist, it would mean getting on with his science. Following inductive habits and customs, of course.
Quoting Purple Pond
Yes, it does. Just not a guarantee.
Quoting Purple Pond
He definitely doesn't want you to think that way.
Quoting Purple Pond
You mean extrapolating deductively? Yes, that's right. No warrant for induction in deduction from experience.
Quoting Purple Pond
Habit, custom.
So, we readily assume causality when "one object produces the other" ...
Quoting Purple Pond
... but only when it habitually does so.
In other words, when you regularly see two kinds of events following each other, you will start assuming that the first event causes the second.
Quoting Purple Pond
In fact, we may not know why the cause causes the consequence, but since it does, we (used to) call it "Custom".
If you systematize this empirical approach, you get experimental testing, which is indeed the core knowledge-justification method of science.
Quoting Purple Pond
(later on) Pavlov's dog:
Classical conditioning (also known as Pavlovian or respondent conditioning) refers to a learning procedure in which a biologically potent stimulus (e.g. food) is paired with a previously neutral stimulus (e.g. a bell). It also refers to the learning process that results from this pairing, through which the neutral stimulus comes to elicit a response (e.g. salivation) that is usually similar to the one elicited by the potent stimulus.
If you repeatedly show to people experiment tests in which a dog will associate ringing a bell, with food (the dog will start salivating), you will condition these people into believing that repeated occurrence of such two events will lead to a learning process in which these two events will be understood/misunderstood to be cause and consequence, i.e. causality.
In the context of Karl Popper's falsificationism, the generalization from a mere sample is permitted. It will not be proof but merely evidence for the causal pairing, until a counterexample is produced somewhere.
We have long ago abandoned verificationism:
Logical positivists within the Vienna Circle recognized quickly that the verifiability criterion was too stringent. Notably, all universal generalizations are empirically unverifiable, such that, under verificationism, vast domains of science and reason, such as scientific hypothesis, would be rendered meaningless.
Pavlov's dog is very reasonable, and the dog's approach is epistemically absolutely sound. Pavlov's dog is a Popperian falsificationist who rejects verificationism, which he undoubtedly dismisses for being an unrealistic requirement, which would prevent him as a dog from learning anything at all.
Quoting Purple Pond
Hume is not inconsistent. The laws of nature can only be learned by using the laws of learning in nature. Pavlov's dog is not a verificationist who first wants to verify all possible cases. Pavlov's dog is a falsificationist, just like science is.
You mean the habit of the mind to associate two or more things that are constantly conjoined, justifies our belief that they will be constantly conjoined in the future? Why? It's not obvious that our habits of association in the past are relevant to what will happen in the future. Am I missing some point?
Quoting bongo fury
Because mine and Hume's personal beliefs about the efficacy of induction are not of any importance in this discussion. We're here to consider arguments not personal beliefs. The point is that Hume needs to provide justification for his use of induction irrespective of his or my personal beliefs about induction.
Quoting bongo fury
But what's left? You say habits and customs, but first I need you to explain exactly how they justify our use of induction.
Quoting bongo fury
You mean lend probable support? You would then have to assign probabilities to events. How do you do that without assuming that the past will be like the future?
He says: For wherever the repetition of any particular act or operation produces a propensity to renew the same act or operation, without being impelled by any reasoning or process of the understanding, we always say, that this propensity is the effect of Custom
I take it that he refers to things we do without thinking about them, out of repetition. Like tying one's shoes or walking, there is no reasoning that goes into them when we do them, we're not thinking about how to tie our shoes or how to walk, we do them out of "habit" or "custom".
Like we infer generalities from particular instances without thinking that we're doing it, if everytime we observe some object A we observe the object B nearby, then out of repetition the next time we observe object A we assume the object B is nearby, out of "habit" or "custom".
The words "habit" or "custom" are not meant to be a justification of induction, he's just relating our habit of generalizing an observation from previous instances of that observation to our habit of renewing an act from previous instances of that act.
In a sense, when we observe object A, we're renewing the observation of object B even if we haven't actually observed it yet. Like we renew the act of walking because it worked the previous times that we did it, even though in principle it's possible that the ground is going to suddenly disappear or that gravity is going to be suddenly 1000 times stronger, but we act as if the way we walked in the past is still going to work, it's just what we do.
And we act as if the correlations we identified in our past observations will keep being the case. Until we find an instance where the correlation stops working, then we come up with a new habit.
Quoting Purple Pond
There is no more justification to his conclusion that we all have that propensity, than there is justification to the conclusion that object B will be there when we observe object A, or that the way we walked in the past will keep working in the future, but we act as if the conclusion is valid until we find a counterexample. Hume couldn't think of anyone who didn't have that propensity, so out of habit he said that we all have that propensity. You don't have to agree with him, but you can make your own observations and reach your conclusions out of your own habits.
Quoting Purple Pond
I wouldn't say the first proposition is only a logical claim, it is also empirical, if we were omniscient I suppose we could justify induction in some cases. It is our empirical observations showing that some generalizations we made are not always valid even though we thought they were, that lead us to say that we can't know whether a generalization is valid even if we think it is.
Regarding the second proposition, I wouldn't say Hume is attempting to prove that induction is based on something arbitrary. Rather he's using induction to relate induction to what we usually call habit or custom.
Is it a problem to use induction to say that we cannot know whether the generalizations reached through induction are universally and eternally valid? I wouldn't say so. The conclusions he reaches can be taken as temporarily valid until we find a counterexample. Until then, it works, just like tying our shoes or walking works.
Taking a step back, it seems to me your desire to justify induction is the desire to find universal and eternal truths. But does it really matter if we can't find them? What if we can't find them because there is no such thing as "the way the world is", but rather the world changes through us? If we participate in shaping the world through our will, then there are no universal and eternal truths to be found out there, rather if they exist they would be found in here, in us. Such a truth wouldn't be arrived at through induction, but through something else, say a profound feeling. Or it could even be that there is no universal and eternal truth, but you could still hang on to your own truths, your own beliefs, your own habits or customs, which could work long enough for you that you don't have to care whether they aren't eternal outside of you, as long as they are eternal to you.
It justifies our provisional (falsifiable) theories, yes.
Quoting Purple Pond
Seems to work.
Quoting Purple Pond
Eh?? You don't choose to adopt such habits of association, then, in your daily life?
Quoting Purple Pond
You're not alone in finding the whole situation to be a scandal.
Quoting Purple Pond
Yes, arguments, not only beliefs. Including arguments about whether deductive argument from observationally verifiable premises is sufficient to support induction. Hume argues it is not.
Quoting Purple Pond
Which is that it seems to work.
Quoting Purple Pond
They seem to work.
Quoting Purple Pond
No, because as you say:
Quoting Purple Pond
So I mean, rather, it lends inductive (fallible, less than certain) support to a hypothesis which may or may not be framed in probabilistic terms. But which, if it is so framed, will - as I think you are rightly pointing out - imply a universal hypothesis about the appropriateness of a particular statistical or probabilistic framework.
Quoting alcontali
Which points up that the question too rarely asked in this kind of discussion is: "which are the good habits and customs... the good inductive hypotheses thereby produced?". Falsifiable ones, yes. Let's assume that. But that leaves an embarrassment of riches to choose from. Unlike the dog, we may want to decide (and formulate) what are the best habits to follow.
Yes, as to the universality of Popper's falsificationism, you can even subject Pavlov's dogs, who are clearly falsificationist animals, to repeated attempts at falsification, in order to find a counterexample of a dog who is not.
In that sense, falsificationism itself can be subjected to the rigours of falsificationism.
Falsficationism clearly "eats its own dog food".
I meant the alleged appropriateness (complete and utter) of some particular probabilistic framework in the inductively (and therefore uncertainly) projected state of affairs wider than the evidence. I didn't mean some test of falsificationism as a doctrine. I was clarifying for Purple that Hume is not appealing to probability to justify induction.
Pardon me if I'm being too serious here. Or yet unclear.
Isn't he saying that, although induction is unjustified and unjustifiable, we use it anyway because we have no choice? When there's nothing better available, we use induction, Occam's Razor, and all manner of other rules of thumb (guesswork). Is there an alternative? :chin:
An interesting tidbit: If you think about it, there's only one way for nature to be uniform, and an infinite ways it can vary.
No such luck...
https://en.wikipedia.org/wiki/New_riddle_of_induction
https://thephilosophyforum.com/discussion/comment/437915
I'm not quite sure how this corresponds to what I said. :chin: What's the point you're making?
Quoting Pattern-chaser
Hume says we use induction because we are compelled to use it by our very nature, and not because there is no other method.
Quoting Pattern-chaser
The problem with that is, according to Hume, there's no reason to think that induction, or any other rules of thumb, would be better, for example, than consulting a psychic, or any other attempt to predict the future.
I can't really disagree with that. Deduction is the ideal-world tool, but cannot always be correctly applied. The obvious example of this is when there is simply insufficient evidence to derive the things we're after. So, if deduction can't be used, we look for alternatives. Because we have no real choice. And induction is one of those alternatives. But deduction it ain't, and it does not carry with it the authority of deduction, any more than Occam's Razor does. But I agree with Hume on the difficulty of rating the efficacy of these lesser alternatives to deduction. I believe it to be next to impossible. So what do we do? We guess, as we always do. :blush:
The empirical interpretation is to view the idea of induction as being distinct from the idea of repetition, whereby 'induction is habit' is viewed as an a posteriori empirically contingent assertion correlating two distinct ideas, say, the internal mental state of expectation and its association to external observations of repetition. As you point out, this interpretation appears to be self-undermining, since according to Hume no empirically contingent proposition can have universal justification, which in this case can lead to semantic skepticism concerning the very meaning of induction.
The grammatical interpretation is that induction is defined directly in terms of observed repetition. In which case, 'induction is habit' is a deflationist assertion, i.e an analytic a priori definition of induction without empirical implications, as opposed to being an empirically contingent assertion. In which case semantic skepticism regarding the meaning of induction might be considered to have been circumvented, assuming one is happy to accept the notion of repetition as being an empirical notion that can ground the rational idea of induction without presuming it's existence. Alternatively, if one rejects the idea that repetition is an empirical notion, one might instead be willing to accept the converse, namely that idea that induction is a directly observable mental state (at least in terms of an inward experience of anticipation) and that it is this 'experience of induction' that serves to ground the idea of repetition within the faculty of reason .
Yet however complicated the phenomenology of induction is, presumably we can at least draw the conclusion that induction doesn't have a purely empirical justification nor a purely rational justification.
Unfortunately neither the empirical nor the grammatical interpretation fits with either our counterfactual intuitions concerning induction; For if a cause of type 'A' is understood to entail an effect of type 'B' when situated within a background context 'C', then an absence of 'A' implies the absence of 'B' whenever 'C' remains the same as in the former case. And whilst there are many cases in which we directly observe frequencies of conjoined absence, we mostly appeal to counterfactual reasoning rather than to observation to infer the consequences of absent causes (e.g. what doesn't happen when a nuclear bomb fails to detonate).
At the very least, we cannot argue for causation without appealing to circularity or to counterfactual situations embedded within vague background contexts, for which we cannot explicitly provide a constructive argument, and in which our observations and logic are inseparably interwoven.
Induction is saying "That swan is white, therefore all swans are white", isn't it? How much more complicated must we get?
This is incorrect. Hume is pointing out the error of human understanding.
Please read this again. And again.
Your opening post is really good. I don't have time tonight to expand on this. But you just demonstrated the kind of thinking Hume is critiquing above.
I'll come back to this again.
Or as we call it these days "conditioning.' Pavlov's dog does not reason, but becomes accustomed habitually to expect food on hearing a bell. It is the foundation of learning, but has no logical basis.
Hume's philosophy is a critique of reason, that can be summed up in two matching limits.
You cannot derive an 'ought' from an 'is'.
You cannot derive a 'will be' from a 'has been'.
For some reason people who are happy to assent to the former often have difficulty with the latter.
... or induction, of course.
Quoting unenlightened
... except by induction.
Quoting unenlightened
... if logic means deduction (as it usually does).
Quoting unenlightened
... by deduction.
Quoting unenlightened
Because they assume logic and reason are deductive?
Indeed, but giving it another name does not give it a justification. Do you think you have done that? Or that someone else has?
it is true
why? tell us, do
it happened before
it'll happen therefore
chicken, "I was fed every morning"
today my family mourning
what is induction?
forever a fraction
never a whole, all
like deduction, standing tall
I merely think I understand Hume to have pointed out that justification (or reason or logic or derivation or inference) is sometimes deductive but just as often inductive (habitual or associative).
https://thephilosophyforum.com/discussion/comment/331201
I don't think that's completely accurate.
It seems you agree with the uncertainty of induction. I'm not an expert but Hume is probably suggesting an explanation for the situation. He proposes habit/custom as one but, from your post, never claims that it's the only explanation for it. There must be an explanation right? What could it be? Habit/custom? What else?
Quoting Hume, 36.
My emphasis.
:ok: :up:
So, here Hume is practically 'chastising' the "habit" of ordinary way of looking at things. Whom is Hume referring to? Who is it that draws an inference without first negotiating a process of reasoning that does not involve his own experience? Can you think of a philosopher who does that?
I can think of a school of thought that could satisfy Hume's dissatisfaction here. Realism.
I can't help but wonder if you have read The Science of Scientific Inference by Wesley Salmon. It is quite good.
Also, I am not a fan of David Hume. I started a thread here asking "What advance in epistemological or metaphysical knowledge did David Hume bring us?" I invite you to come check it out.
https://thephilosophyforum.com/discussion/6682/what-advance-in-epistemological-or-metaphysical-knowledge-did-david-hume-bring-us