Eating Dirt etc

[Daniel Barnes]

Moved from another thread:

If we conjecture that bread is food but dirt is not, how do we test -- i.e., attempt to falsify -- the latter point? It might be said that we tried eating dirt as children, spat it out, and decided it wasn't fit to eat. Okay, but how do we know that no dirt is fit to eat? Why do we stop after a few instances of eating dirt? Why don't we keep sampling dirt in an effort to falsify the hypothesis that no dirt is fit to eat?

We don't do this for a simple reason: From our experience with a few cases of eating dirt, we conclude, via inductive reasoning, that those unpleasant experiences will repeat themselves in any future attempts to eat dirt. We reach a conclusion about the nature of dirt from a limited number of particular cases.

In other words, if inductive reasoning is never valid, there would be no way to falsify a hypothesis or even render it improbable. For just because x was incompatible with our hypothesis in the past would be no reason to assume that instances similar to x would be incompatible with our hypothesis in the future. We would be required to "test" our hypothesis over and over again, indefinitely, with additional instances of x, and then we could only say that this particular x is compatible or incompatible with a hypothesis. We could never (legitimately) generalize about all x's. We could only say that the dirt we have eaten so far is not food. We could never justifiably say that dirt we haven't tried yet is not food. We could never learn from experience. We would need to keep eating dirt over and over again in an effort to falsify our hypothesis that dirt is not food, and, logically speaking, this process would never end.

Now, this is an infinite regress.

Just returning to this issue now I have a moment, as you seem to think this is a good argument against Popper's hypothetico-deductive theory.

It isn't. In fact it is just another typical example of why enumerative-induction doesn't work. As my previous reply obviously wasn't very clear, I will go through it point by point to show exactly where and how it goes wrong. This may also help other interested readers who are not as familiar with Popper as you are, so I will write it accordingly.

You start with a conjecture:

If we conjecture that bread is food but dirt is not, how do we test -- i.e., attempt to falsify -- the latter point?

You then propose a simple test:

It might be said that we tried eating dirt as children, spat it out, and decided it wasn't fit to eat.

So far, so Popperian. But here's where it goes off the rails:

Okay, but how do we know that no dirt is fit to eat?

Ah ha. But according to the problem of induction, we can never know for sure that the universal theory "no dirt is fit to eat" is true. And just as well, because it turns out that it's false anyway - see here, and here. Further, part of the problem is the way your hypothesis has been framed. As it stands - "no dirt is fit to eat" is in any practical sense unfalsifiable - you'd effectively be trying to prove a negative, as you'd indeed have to sample all dirt that ever existed or will exist in order make such a proof. This is irrelevant, as Critical Rationalism never tries to prove anything - only to falsify. So it turns out your example is just standard fallacious proof-by-enumerative-induction coming in via the back door. Now, an important part of Popper's recommendation is to avoid these sorts of problems in the first place by trying to frame hypotheses so they are more easily falsifiable; that is, by making them testable. So you could just as easily make your hypothesis "all dirt is food", for this could just as well be the infant's vague, unspoken hunch when she puts the dirt in her mouth rather than "no dirt is food." This hypothesis now becomes testable, and can be falsified by modus tollens as only a single counterexample - one taste - is necessary from a logical point of view ( though if you want to try more, that's up to you...). More about that below.

You then wonder:

Why do we stop after a few instances of eating dirt? Why don't we keep sampling dirt in an effort to falsify the hypothesis that no dirt is fit to eat?

As I mention above, the answer is in Popper's well-known adoption of the logical form modus tollens, which he built his notion of falsifiability around. Using the modus tollens, we only need one counter example - a single black swan - to falsify the universal claim that "all swans are white". Likewise "all dirt is food". Thus the scenario of tasting dirt forever is not entailed by, nor has anything to do with, the Popperian approach. We can now replace our falsified hypothesis "all dirt is food" with a new one - for example, "some dirt is not food" - and act with the appropriate caution, or with an appropriate sense of exploration and adventure :-) in future. And all this is achieved without any recourse whatsoever to any enumerative-induction approach. So the rest of your post is naturally incorrect.

Hopefully that clears that up.

As for clinging to one's beliefs, do you cling to your belief that Popper was right in this area? Do you believe that Popper's arguments are justified? Have you attempted to falsify them? If so, by what means?

No, I hold them lightly as ever. No, but they have withstood many tougher tests than yours above, so so far so good :-). Yes, many times; by debating them in places like this that are not likely to easily accept them, for just one example.

PS: I have tried to make this explanation as clear as possible. If you still profoundly disagree with it and want to debate it at length, I propose starting another thread so Robert's thread doesn't get hijacked.

[Daniel Barnes]

This is George's reply to the above:

If we conjecture that bread is food but dirt is not, how do we test -- i.e., attempt to falsify -- the latter point? It might be said that we tried eating dirt as children, spat it out, and decided it wasn't fit to eat. Okay, but how do we know that no dirt is fit to eat? Why do we stop after a few instances of eating dirt? Why don't we keep sampling dirt in an effort to falsify the hypothesis that no dirt is fit to eat?

We don't do this for a simple reason: From our experience with a few cases of eating dirt, we conclude, via inductive reasoning, that those unpleasant experiences will repeat themselves in any future attempts to eat dirt. We reach a conclusion about the nature of dirt from a limited number of particular cases.

In other words, if inductive reasoning is never valid, there would be no way to falsify a hypothesis or even render it improbable. For just because x was incompatible with our hypothesis in the past would be no reason to assume that instances similar to x would be incompatible with our hypothesis in the future. We would be required to "test" our hypothesis over and over again, indefinitely, with additional instances of x, and then we could only say that this particular x is compatible or incompatible with a hypothesis. We could never (legitimately) generalize about all x's. We could only say that the dirt we have eaten so far is not food. We could never justifiably say that dirt we haven't tried yet is not food. We could never learn from experience. We would need to keep eating dirt over and over again in an effort to falsify our hypothesis that dirt is not food, and, logically speaking, this process would never end.

Now, this is an infinite regress.

Just returning to this issue now I have a moment, as you seem to think this is a good argument against Popper's hypothetico-deductive theory.

It isn't. In fact it is just another typical example of why enumerative-induction doesn't work. As my previous reply obviously wasn't very clear, I will go through it point by point to show exactly where and how it goes wrong. This may also help other interested readers who are not as familiar with Popper as you are.

You start with a conjecture:

If we conjecture that bread is food but dirt is not, how do we test -- i.e., attempt to falsify -- the latter point?

You then propose a simple test:

It might be said that we tried eating dirt as children, spat it out, and decided it wasn't fit to eat.

So far, so Popperian. But here's where it goes off the rails:

Okay, but how do we know that no dirt is fit to eat?

Ah ha. But according to the problem of induction, we can never know for sure that the universal theory "no dirt is fit to eat" is true.

In my discussion the statement "no dirt is fit to eat" did not function as a universal theory. It was a generalization based on particular experiences of attempting to eat dirt. Generalizations can have exceptions. Generalizations function as presumptions -- in some cases very strong presumptions -- that serve as cognitive shortcuts. They enable us to form reasonable -- not infallible -- expectations without having to repeat past experiences over and over again.

And just as well, because it turns out that it's false anyway! - see here, and here. Further, part of the problem is the way your hypothesis has been framed.

This is an old tactic. After an inductivist states a generalization based on particular instances, the critic points to exceptions, or apparent exceptions, to falsify the generalization. But a generalization is not the same thing as a theory, and inductive generalizations, which can be accepted with varying degrees of probability, often remain accurate despite a few exceptions.

Popper often uses this tactic. For example, he takes Hume's example that "bread nourishes" and claims that this generalization "was tragically refuted when bread baked in the usual way practically wiped out a French village, due to an outbreak of ergotism." (Objective Knowledge, p. 97)

This shows a failure to understand the nature of generalizations, in contrast to "theories" or "hypotheses." To generalize that "bread nourishes" does not mean that we never regard bread and other foods as unfit to eat in some circumstances. We wouldn't normally eat spoiled food, for example, or raw pork. Popper's obsession with viewing every inductive generalization as a universal theory of some kind led him into a number of unnecessary problems.

Another common tactic used by Popper is illustrated in the following passage:

It is a common fact (whatever philosophers may say) that we are commonsensically certain that the sun will rise over London tomorrow. Yet we do not know it for certain. There are millions of possibilities which may prevent it....But even a necessary hope is not objective knowledge, though it may dispose us to belief.

In other words, those rules which are still used by philosophers as standard examples of inductive rules (and of reliability) are all false, even when they are quite good approximations to the truth.

So the rules of induction can yield "quite good approximations to the truth"? I will settle for this, as would most other inductivists, especially since Popper himself sometimes says that approximations of truth are the most we can ever hope to attain.

Popper's discussion here is very confusing, and things get even more confusing when Popper goes on to say:

Genuine induction by repetition does not exist. What looks like induction is hypothetical reasoning, well tested an well corroborated and in agreement with common sense.

Note that Popper refers to "induction by repetition." This is often called induction by mere enumeration, and this is precisely the sort of thing that Harriman, along with virtually every other inductivist, rejects. If Popper believes that induction consists of adding up particular cases and then, with no additional reasoning (whether implicit or explicit), forming a generalization, then he is attacking a type of "induction" that no one has ever defended.

Let's return to Popper's example of the sun rising tomorrow. Does any reasonably intelligent person who makes this inductive generalization really not understand that certain cosmic events could occur that would prevent the sun from rising tomorrow? No, of course not. When people say this, they are not proposing a "theory," much less a scientific theory; rather, they are expressing a reasonable expectation based on a generalization from past experiences.

There are a number of other things I would need to discuss to show how confused Popper's approach is, but I will confine myself to a few brief remarks.

1) According to Popper (OK, p. 34), "all science, and all philosophy, are enlightened common sense." This is wrong. Although philosophy typically relies on "common sense," science, especially physics, does not. This is why the specific methods of verification used in science differ dramatically from the methods used in philosophy. Popper's confusion in this matter leads him in a futile quest to explain how a formal method of verification used in science -- Daniel calls it the hypothetical-deductive model, but this can be misleading when applied to Popper -- applies to philosophy and even everyday life as well. This is a serious error.

2) Popper's remarks about "objective knowledge" sometimes conflict. At times he suggests that objective knowledge is possible, whereas at other times he says that objective knowledge is merely an ideal that can only be approximated to one degree or another. The problem here is that Popper often equates "knowledge" with absolute certainty, or even infallible certainty. Thus if we don't know something for certain, we don't have any knowledge at all. This is one reason he can say that we often form reasonable beliefs based on induction (only he doesn't want to call this "induction," even though everyone else under the sun has) and still not have "knowledge."

(3) To add to the confusion, Popper has some excellent discussions where he criticizes the old notion, which goes back to the ancient Greeks, that knowledge must be "certain" before it can qualify as knowledge. Indeed, he makes a number of comments that could easily be translated into the language of Rand's contextualism. (These are among my favorite discussions by Popper.)

(4) The upshot of all this is that the problem of induction -- which Popper claims to have solved, "though negatively" (OK, p. 94) -- is inextricably tied to his theory of objective knowledge, and the confusion exhibited in the latter seeps into his discussion of the former. (Daniel: Please don't tell me that Popper distinguishes between the questions "How can we justify induction?" and "Is induction at all justifiable?" I know this, but I have to cut some corners to keep this post to a reasonable length. This distinction is not relevant to the points I am making here.)

(5) At various times, when Popper comes very close to defending the traditional theory of induction, merely restating it in a different form, he sometimes calls attention to the similarity, only to seek refuge in his favorite excuse, viz., that he doesn't want to quibble over the meaning of words. At some point he even says, in effect, Well, if you want to call my approach "induction," then call it induction. Okay, I call it not only induction in some instances but also the traditional theory of induction. As noted previously, Popper focuses on induction by mere enumeration, and no inductivist that I can recall ever defended this approach.

Now back to Daniel's post:

As it stands - "no dirt is fit to eat" is in any practical sense unfalsifiable - you'd effectively be trying to prove a negative, as you'd indeed have to sample all dirt that ever existed or will exist in order make such a proof.

Earlier you gave two links to show dirt that apparently is fit to eat, yet now you claim that "no dirt is fit to eat" cannot be falsified, an least not in any practical sense. So what were your counter-examples? -- impractical falsifications?

I will need to continue this later.

Ghs

Edited September 9, 2010 by Daniel Barnes

[Daniel Barnes]

In my discussion the statement "no dirt is fit to eat" did not function as a universal theory. It was a generalization based on particular experiences of attempting to eat dirt. Generalizations can have exceptions. Generalizations function as presumptions -- in some cases very strong presumptions -- that serve as cognitive shortcuts. They enable us to form reasonable -- not infallible -- expectations without having to repeat past experiences over and over again.

Well, that's all very well, but Popper is talking about universal theories, and not vague generalisations. If you're not talking about induction in terms of providing universal theories or laws, but merely vague uncertainties that amount to something like "Some dirt might be edible, some not", rather than an absolute "NO dirt is fit to eat" then there's no real problem in the first place.

Now back to Daniel's post:

As it stands - "no dirt is fit to eat" is in any practical sense unfalsifiable - you'd effectively be trying to prove a negative, as you'd indeed have to sample all dirt that ever existed or will exist in order make such a proof.

Earlier you gave two links to show dirt that apparently is fit to eat, yet now you claim that "no dirt is fit to eat" cannot be falsified, an least not in any practical sense. So what were your counter-examples? -- impractical falsifications?

LOL! I apologise, that was silly of me. My example does in fact falsify "no dirt is fit to eat" if we consider it to be a universal statement ie one that is just as true in the future as it was in the past. In fact it's easy to falsify considered like that. The child could easily have found the very first bite nutritious, falsifying it immediately.

When I wrote that I was actually thinking of your comment here, italics my emphasis:

GHs:

In other words, if inductive reasoning is never valid, there would be no way to falsify a hypothesis or even render it improbable. For just because x was incompatible with our hypothesis in the past would be no reason to assume that instances similar to x would be incompatible with our hypothesis in the future. We would be required to "test" our hypothesis over and over again, indefinitely, with additional instances of x, and then we could only say that this particular x is compatible or incompatible with a hypothesis. We could never (legitimately) generalize about all x's. We could only say that the dirt we have eaten so far is not food. We could never justifiably say that dirt we haven't tried yet is not food. We could never learn from experience. We would need to keep eating dirt over and over again in an effort to falsify our hypothesis that dirt is not food, and, logically speaking, this process would never end.

In the italicised part you describe an ongoing search for falsification that is simply not necessary for a universal statement. I simply took your example too seriously, forgetting myself for a moment that a universal statement applies to the future as much as the past. So if we falsify it now, there's absolutely no need to keep testing it into the future. It's failed! In contrast, your vague generalisation, of course, might be true today but false tomorrow, or whatever, so you can keep eating dirt till the cows come home, or not, or whatever. Such generalizing commits you to very little, and you can always just make it even vaguer as you come across contrary examples. Anyway: Duh! Brain fart. So I stand corrected. But sadly, my blunder doesn't help your argument - there is still no need for any enumerative-inductive-like "infinite regress" with Popper, as he is dealing with universal statements or laws, so you are quite wrong I think.

If all you are discussing is vague generalisations rather than universal laws then this has little to do with Popper, and we are clearly talking at cross purposes. As such, confusions such as my one above are bound to creep in.

Edited September 9, 2010 by Daniel Barnes

[Rich Engle]

[in the italicised part you describe an ongoing search for falsification that is simply not necessary for a universal statement. I simply took your example too seriously, forgetting myself for a moment that a universal statement applies to the future as much as the past. So if we falsify it now, there's absolutely no need to keep testing it into the future. It's failed! In contrast, your vague generalisation, of course, might be true today but false tomorrow, or whatever, so you can keep eating dirt till the cows come home, or not, or whatever. Such generalizing commits you to very little, and you can always just make it even vaguer as you come across contrary examples. Anyway: Duh! Brain fart. So I stand corrected. But sadly, my blunder doesn't help your argument - there is still no need for any enumerative-inductive-like "infinite regress" with Popper, as he is dealing with universal statements or laws, so you are quite wrong I think.

I always knew that you are good for more than one thing, or another, George.

rde

[George H. Smith]

In my discussion the statement "no dirt is fit to eat" did not function as a universal theory. It was a generalization based on particular experiences of attempting to eat dirt. Generalizations can have exceptions. Generalizations function as presumptions -- in some cases very strong presumptions -- that serve as cognitive shortcuts. They enable us to form reasonable -- not infallible -- expectations without having to repeat past experiences over and over again.

Well, that's all very well, but Popper is talking about universal theories, and not vague generalisations. If you're not talking about induction in terms of providing universal theories or laws, but merely vague uncertainties that amount to something like "Some dirt might be edible, some not", rather than an absolute "NO dirt is fit to eat" then there's no real problem in the first place.

If you read Popper's accounts -- and the accounts of Hume, on whom Popper relies in this context -- you will find that their arguments are aimed at all generalizations based on experience, not merely at hypotheses and theories. As Popper puts it in Objective Knowledge(p. 89):

Hume's logical problem of induction is the problem whether we are entitled to infer unobserved cases from observed cases, however many; or "unknown" (unaccepted) statements from "known" (accepted) statements, however many.

How could this be any more clear? The whole point of Hume's argument is to show that we can never rationally conclude that x will occur in the future just because we have observed x to occur in the past. Popper agrees with this criticism, and it has nothing inherently to do with theories and hypotheses. Popper's problem is that he frequently insists on treating empirical generalizations as if they were theories or hypotheses. This is a huge mistake.

If all you are discussing is vague generalisations rather than universal laws then this has little to do with Popper, and we are clearly talking at cross purposes. As such, confusions such as my one above are bound to creep in.

I claim that empirical generalizations based on inductive reasoning are rationally warranted, if only as matters of probability. Popper, in contrast, claims that inductive reasoning is never warranted, that it cannot even yield probability in any degree.

You and I may be talking at cross purposes, but I was not talking at cross purposes when it came to Popper.

If you wish to concede that inductive reasoning is sometimes warranted, insofar as it enables us to form reasonable generalizations, then you have given away the Popperian farm. Such a concession strikes at the very heart of his argument against induction, as well as the argument of Hume.

Ghs

[kiaer.ts]

What would be the purpose of forming the conclusion that eating dirt is never a good idea? Would not, "eating dirt is generally a bad idea," be sufficient and natural?

The induction would be justified by experience and the realization of the causally relevant fact that dirt is weathered stone mixed in various proportions with decayed organic matter, feces, and often a significant portion of protozoans, bacteria and invertebrates.

A child might reasonably doubt a parent's unexplained admonition not to eat dirt "because it's dirty." But surely the child will eventually learn how dirt is formed, and what happens to food and dead bodies left out in the rain.

This "dilemma" is an artificial one.

Edited September 9, 2010 by Ted Keer

[George H. Smith]

Aaargh! I just lost another unfinished message. Now I have to retype this damned thing.

If you wish to concede that inductive reasoning is sometimes warranted, insofar as it enables us to form reasonable generalizations, then you have given away the Popperian farm. Such a concession strikes at the very heart of his argument against induction, as well as the argument of Hume.

Consider this passage by Popper from Conjectures and Refutations, p. 42. (Italics are in Popper's original.)

Hume, I felt,was perfectly right in pointing out that induction cannot be logically justified. He held that there can be no valid logical arguments allowing us to establish "that those instances, of which we have had no experience, resemble those, of which we have had experience." Consequently, "even after the observation of the frequent or constant conjunction of objects, we have no reason to draw any inference concerning any object beyond those of which we have experienced." For "shou'd it be said that we have experience" -- experience teaching us that objects constantly conjoined with certain other objects continue to be so conjoined -- then, Hume says, "I wou'd renew my question, why from this experience we form any conclusion beyond those past instances, of which we have had experience...."

Note what Hume is saying here. He is saying that no empirical generalization based on past experience is rationally warranted. Hume never refers to theories or hypotheses; he didn't have those in mind. Instead, he specifically says that "even after the observation of the frequent or constant conjunction of objects, we have no reason to draw any inference concerning any object beyond those of which we have experienced." (My italics.)

But now observe what Popper does. After quoting Hume, Popper writes:

As a result we can say that theories can never be inferred from observation statements, or rationally justified by them. (My italics.}

Popper is correct, of course. If Hume is right, then no theory can indeed be inferred from, or rationally justified by, observation statements (i.e.,. appeals to previous experience). But Hume is right only if it is true that we can never legitimately generalize in any fashion from previous experiences.

Hume understood the force of his argument against induction, namely, that there is no rational warrant to believe that our future experiences will even resemble our past experiences, so empirical generalizations are never warranted.

Popper agrees with Hume on this, and he then applies Hume's reasoning to theories. Thus if Daniel concedes that inductive reasoning can be legitimately used to form empirical generalizations and that Popper's argument applies only to universal theories, then Daniel has cut the entire Humean ground from under Popper's argument and thereby collapsed that argument.

Ghs

[George H. Smith]

This post puzzled Ted, and for good reason. I mistakenly thought I was responding to Daniel. See Ted's protest below and my correction and apology.

What would be the purpose of forming the conclusion that eating dirt is never a good idea?

There is a context for such generalizations. When we say that eating dirt is never a good idea, we mean "dirt" of the same (or relevantly similar) kind as the dirt we tried eating in the past. If we encounter a new and different type of dirt, then we might need to revise our generalization. This is the way knowledge progresses.

So-called exceptions to empirical generalizations are sometimes not really exceptions at all. The fact that we may accidentally drink poisoned water does not affect the generalization that water is good for us, because in making this generalization we are talking about the same kind of water that we have drunk in the past. We are not talking about poisoned water. Only a pedant would insist that every qualification and possible exception, which are typically implicit and understood, be included in every empirical generalization.

Would not, "eating dirt is generally a bad idea," be sufficient and natural?

Yes, of course you could say this. But even this generalization would be based on inductive reasoning -- you know, the kind of reasoning that Popper claims is never valid.

Nevertheless, suppose you are about clean a very dirty cat litter box. You say, "Gee, I wonder if I should eat some of this cat litter; it might be real tasty and good for me to boot." I then say to you, "No eating cat litter is always a bad idea." Would you then correct me and say, "Eating cat litter is generally a bad idea. This might be an exception."

I know you don't like the word "context," but there is a context to all empirical generalizations, and if you ignore that context, anything goes.

The induction would be justified by experience and the realization of the causally relevant fact that dirt is weathered stone mixed in various proportions with decayed organic matter.

An induction justified by experience? Causally relevant facts?

Welcome, at long last, to the non-Humean, non-Popperian world of the inductivist.

Ghs

[kiaer.ts]

Yes, of course you could say this. But even this generalization would be based on inductive reasoning -- you know, the kind of reasoning that Popper claims is never valid.

Yes, and Popper's and Hume's inductions can, by their own arguments, be dismissed out of hand.

[kiaer.ts]

I know you don't like the word "context," but there is a context to all empirical generalizations, and if you ignore that context, anything goes.

. . .

An induction justified by experience? Casually relevant facts?

Welcome, at long last, to the non-Humean, non-Popperian world of the inductivist.

Huh? I think you have me confused with someone who is not me. I have been an "inductivist" for as far back as I can remember and fully endorse both the scientific method and Rand's epistemology.

[Selene]

Folks:

I thought I remembered some data about eating dirt:

"Is eating dirt so dangerous that we should always be watching our children for this behavior? According to doctors who study infectious diseases and poisonings, we should instead watch for several hazards that contribute to the risk. And there are some reasons to think that eating the occasional pinch of good clean dirt benefits the immune system.

Eating dirt isn't always a medical issue. Pica, pronounced PIE-kah, is a disorder in which people persistently eat things that are not food, including ice, hair, ashes, laundry starch, or soil. Pica is from the Latin word for magpie, a bird said to pick up and eat anything. But a pinch of dirt isn't enough to qualify, according to categories developed in the Atlanta meeting by experts gathered by the U.S. Agency for Toxic Substances and Disease Registry:

• Normal soil consumption: up to about 500 mg a day of soil consumption is considered normal in children up to 3 years old. Two pennies weigh about 500 mg.
• Soil pica: children who persistently eat more than 1 g a day of soil may be suffering from soil pica.

Dr. Jeremy Friedman, the division head of Paediatric Medicine at The Hospital for Sick Children (SickKids) in Toronto, agrees that most children eat soil as a normal exploratory phase. "It's important to distinguish between normal kids from 9 months to 2 years of age who experiment with putting things in their mouths, and older children with pica who will hunt down some soil and consume it," he said."

http://www.aboutkids...oryID=news-poh6 <<<<link to the article

Adam

wiping the mud from his chin Post Script:

From Monday July 13, 1942 TIME Magazine

"Many a homesick or sardonic Northern Negro, writing to Southern friends, says "Ship me a bag of good dirt to eat." Sometimes he means it. Even in the Yazoo-Mississippi Delta, Negroes and whites send requests to their upcountry friends for a bit of red clay, declaring that black Delta soil is "right bad eating." In certain parts of Mississippi, poor whites will walk miles for a spoonful of dirt from a favorite bank of clay, because it "tastes sour, like a lemon." In other sections of the South, some top their meals with a savory tablespoon of dirt, believing that it is "good for them," despite its constipating effects.What makes people eat dirt?

Two scientists at Mississippi State College, Drs. Dorothy Dickins and Robert N. Ford, think they know why: lack of iron in the diet. Last fortnight Science News Letter reported some results of their investigation among 207 Negro school children in Oktibbeha County, Miss. At least a quarter of the children admitted eating dirt. Most of the dirt-eaters had less of the iron-rich foods, such as molasses, mustard greens, liver, in their diet than did the non-dirt-eaters. And as far as the scientists could find out, the craving for dirt (known as geophagia) has nothing to do with hookworm, as many doctors firmly believe, for hookworm is very rare in geophagous Oktibbeha County."

Read more: http://www.time.com/time/magazine/article/0,9171,884551,00.html#ixzz0z3esvdNi

Gimme that Ole Time Religion and some Ole Time Dirt!

Edited September 9, 2010 by Selene

[George H. Smith]

I know you don't like the word "context," but there is a context to all empirical generalizations, and if you ignore that context, anything goes.

. . .

An induction justified by experience? Casually relevant facts?

Welcome, at long last, to the non-Humean, non-Popperian world of the inductivist.

Huh? I think you have me confused with someone who is not me. I have been an "inductivist" for as far back as I can remember and fully endorse both the scientific method and Rand's epistemology.

Shit! I somehow got it into my head that I was responding to Daniel. ( I thought his concession came way too easily.) Hence my comments about "context." I wouldn't have needed to lecture you about such elementary matters. Sorry about that.

I won't revise the post, because it would make your correction curious, but -- for those reading this -- I fucked up! The other points still stand, of course, though I might have presented them somewhat differently if I knew to whom I was responding.

This is not a good way to start off the day.

Ghs

[Daniel Barnes]

This seems to be the nub of your disagreement. Once again, I will break it down to illustrate where I think it goes wrong, and hopefully I won't make too many trivial blunders in pursuit of it this time...;-)

The whole point of Hume's argument is to show that we can never rationally conclude that x will occur in the future just because we have observed x to occur in the past.

Right so far.

Popper agrees with this criticism, and it has nothing inherently to do with theories and hypotheses. Popper's problem is that he frequently insists on treating empirical generalizations as if they were theories or hypotheses. This is a huge mistake.

This is where it might be going wrong. Popper's introduction of hypotheses is to help rescue the situation from Hume's dilemma.

To illustrate I will quote Popper's restatement of Hume - for it is his reformulation of the problem that makes it potentially solvable.

(Karl Popper, "Conjectural Knowledge: My Solution to the Problem of Induction", p7, Objective Knowledge)

5. The Logical Problem of Induction: Restatement and Solution.

In accordance with what has just been said...[his principle of translating subjective psychological terms like "belief" etc into objective terms, such as "statement"]....I have to restate Hume's [logical problem of induction] in an objective or logical mode of speech.

To this end I replace Hume's "instances of which we have experience" by "test statements" - that is, singular statements describing observable events ("observation statements", or "basic statements"); and "instances of which we have no experience" by "explanatory universal theories."

I formulate Hume's logical problem of induction as follows:

L1: Can the claim that an explanatory universal theory is true be justified by "empirical reasons"; that is, by assuming the truth of certain test statements or observation statements (which, it may be said, are "based on experience")?

My answer to the problem is the same as Hume's: No, it cannot; no number of true test statements would justify the claim that an explanatory universal theory is true.

But there is a second logical problem, L2, which is a generalization of L1. It is obtained from L1 merely by replacing the words "is true" by the words "is true or that it is false."

L2: Can the claim that an explanatory universal theory is true or that it is false be justified by "empirical reasons"; that is, can the assumption of the truth of test statements justify either the claim that a universal theory is true or the claim that it is false?

To this problem, my answer is positive: Yes, the assumption of the truth of test statements sometimes allows us to justify the claim that an explanatory universal theory is false.

So there we have it in a nutshell. It is quite clear Popper is addressing the issue of how Hume's problem stands in the way of determining true explanatory universal theories. We also see how he introduces, in L2, the hypothetical turn in order to arrive at his solution of falsifiability. For there is only an issue if we try to claim such a theory based on the truth of individual test statements is true - not if it is possibly true or false.

Thus it seems to me L2 has about the same claim to truth status as your "empirical generalizations", which might also turn out to be false. That is, none. Likewise, there is no problem for Popperians if they make similarly vague "empirical" or "reasonable generalizations" such as "It usually rains in winter" or "the bus stops here at 8 o'clock" or "I will probably get presents this Christmas" . This is simply loose, colloquial speech - no-one expects even an inductivist to produce, say, their probability workings for such comments - and in no way tries to pretend they are somehow "warranted", or are intended as a universal truth claim that droughts can never happen in winter nor the bus is never late. So your point below:

If you wish to concede that inductive reasoning is sometimes warranted, insofar as it enables us to form reasonable generalizations, then you have given away the Popperian farm. Such a concession strikes at the very heart of his argument against induction, as well as the argument of Hume.

...is another clear miss, like your "infinite regress" argument before it.

But perhaps I have misinterpreted your "legitimate" or "reasonable" or "empirical generalisation", and there is in fact some kind of specific principle behind it that is the equivalent of Popper's logical formulation above. If so, could you please outline how it operates re the truth content of these.

Edited September 9, 2010 by Daniel Barnes

[Daniel Barnes]

I claim that empirical generalizations based on inductive reasoning are rationally warranted, if only as matters of probability. Popper, in contrast, claims that inductive reasoning is never warranted, that it cannot even yield probability in any degree.

Incidentally, Popper follows Hume in this. And they are both quite right.

If you mean "probability" in any more meaningful way than the entirely vague and hypothetical sense I mentioned above - that is, if you want to seriously justify one of your "empirical generalisations" claim to truth by an appeal to an actual probability - then perhaps you might want to specify that probability and show your workings.

Edited September 9, 2010 by Daniel Barnes

[George H. Smith]

This seems to be the nub of your disagreement. Once again, I will break it down to illustrate where I think it goes wrong, and hopefully I won't make too many trivial blunders in pursuit of it this time...;-)

The whole point of Hume's argument is to show that we can never rationally conclude that x will occur in the future just because we have observed x to occur in the past.

Right so far.

Popper agrees with this criticism, and it has nothing inherently to do with theories and hypotheses. Popper's problem is that he frequently insists on treating empirical generalizations as if they were theories or hypotheses. This is a huge mistake.

This is where it might be going wrong. Popper's introduction of hypotheses is to help rescue the situation from Hume's dilemma.

To illustrate I will quote Popper's restatement of Hume - for it is his reformulation of the problem that makes it potentially solvable.

(Karl Popper, "Conjectural Knowledge: My Solution to the Problem of Induction", p7, Objective Knowledge)

5. The Logical Problem of Induction: Restatement and Solution.

In accordance with what has just been said...[his principle of translating subjective psychological terms like "belief" etc into objective terms, such as "statement"]....I have to restate Hume's [logical problem of induction] in an objective or logical mode of speech.

To this end I replace Hume's "instances of which we have experience" by "test statements" - that is, singular statements describing observable events ("observation statements", or "basic statements"); and "instances of which we have no experience" by "explanatory universal theories."

I formulate Hume's logical problem of induction as follows:

L1: Can the claim that an explanatory universal theory is true be justified by "empirical reasons"; that is, by assuming the truth of certain test statements or observation statements (which, it may be said, are "based on experience")?

My answer to the problem is the same as Hume's: No, it cannot; no number of true test statements would justify the claim that an explanatory universal theory is true.

1) As I illustrated in a post written earlier today, a post in which Popper quotes Hume, this was not Hume's argument. Hume says absolutely nothing about explanatory universal theories, or theories of any kind. This is Popper's application of Hume's argument to the realm of theory. "Instances of which we have no experience" are not the same as "explanatory universal theories." Popper can replace the former with the latter is he likes, but he severely distorts Hume's argument in the process.

You are going to have to read some Hume if you wish to educate yourself in this matter. Popper usually cites Hume's early Treatise of Human Nature, but you can find a more succinct statement of the same argument in An Enquiry Concerning Human Understanding. Hume regarded the latter work as containing his definitive statements on philosophical issues.

2) There was no need for Popper to "rescue" us from Hume's problem because there is no problem to begin with, as Hume conceived it. Even Popper didn't accept Hume's linking of the problems of induction and causation, but Hume's rejection of causation is absolutely fundamental to his points about induction. Hume's argument about the lack of a causal connection between observed events constitutes the foundation of his argument that inductive inferences are invalid. Hume understood that inductive reasoning presupposes causal reasoning, but Popper did not.

(3) A while back I noted that no inductivist has ever claimed that we can justify universal theories by appealing to inductive reasoning alone. (Many inductivists in the physical sciences have defended the hypothetico-deductive model.) Thus if Popper's supposedly radical argument is that no universal theory can be justified via induction, then has merely stated a truism that no inductivist would deny. The only difference is that Popper introduces a lot of needless technicalities in order to prove this truism.

So there we have it in a nutshell. It is quite clear Popper is addressing the issue of how Hume's problem stands in the way of determining true explanatory universal theories.

I addressed this very point earlier today. Again: Hume's problem is not Popper's problem, and if Popper is only able to solve his problem (the one relating to theories rather than all empirical generalizations), then he has not solved Hume's problem. Popper will only have shown how universal scientific theories can be corroborated without appealing to induction. Fine, this may very well be true -- it certainly is not an original point in any case -- but Hume's problem was much more fundamental, broader in scope, and remains untouched by Popper's supposed solution.

Ghs

[George H. Smith]

I posted this reply early this morning on the "David Harriman's Book" thread, while Daniel was in the midst of starting this thread and transferring other posts to it. I am reposting it here so that all of the major posts will be available on this thread. The latter part of this post contains some important statements about my view of induction.

So far, so Popperian. But here's where it goes off the rails:

Okay, but how do we know that no dirt is fit to eat?

...Further, part of the problem is the way your hypothesis has been framed. As it stands - "no dirt is fit to eat" is in any practical sense unfalsifiable - you'd effectively be trying to prove a negative, as you'd indeed have to sample all dirt that ever existed or will exist in order make such a proof. This is irrelevant, as Critical Rationalism never tries to prove anything - only to falsify.

You have completely shifted the context in which I originally offered the generalization "no dirt is fit to eat."

I was discussing how we learn that some things, which we call "food," are fit to eat, whereas some things (e.g., dirt) are not. I contended that we form inductive generalizations based on our past experiences, and that these generalizations are reliable. We have tried eating certain things, such as apples, and found that they satisfy our hunger and will keep us alive (or at least not hurt us), whereas we have tried, perhaps as children, eating dirt and concluded that no dirt is fit to eat. Each of these can be construed as value judgments, viz., that apples are good to eat, and dirt is bad to eat. (For the sake of convenience and simplicity, I will sometimes use this terminology.)

My main argument is that it is quite absurd to construe this inductive reasoning in the Popperian manner of conjectures and refutations. We do not form a hypothesis that applies are good and then attempt to refute this, nor do we form a hypothesis that dirt is bad and attempt to refute that. This Popperian model of scientific reasoning is not how we usually reason in everyday life, and the latter is no less reasonable and reliable because of this.

I highlighted a particular problem with dirt on the Popperian model, namely, that if inductive reasoning is never justified, then we could never legitimately generalize that dirt is bad. If "dirt is bad" (or, per my original formulation, "no dirt is fit to eat") is construed as a Popperian theory, then it could never be falsified. This means that a rational person could not rely on his previous experiences of eating dirt as a child, since it might be the case that some dirt, including dirt of exactly the same kind as the dirt we had sampled previously, is good. And in this event we would need to keep sampling dirt throughout out lives in order to reassure ourselves that no dirt is fit to eat. Never, on the Popperian model, would we reach the point where we could rationally conclude that we shouldn't eat dirt. All we could say is that the dirt we had eaten in the past was bad, but we could not generalize about dirt we have never tried.

In the passage quoted above, Daniel objects to how my "hypothesis" about dirt is framed, but keep in mind that I do not regard "dirt is bad" (or "no dirt is fit to eat") as a hypothesis at all. This is an empirical generalization based on a limited number of previous experiences, not a hypothesis or theory. My generalization does not propose an explanation of why dirt is not fit to eat, nor it is a universal law that claims to subsume each and every possible case, without exception.

Daniel makes a couple odd claims. First, he claims that "no dirt is fit to eat" is not falsifiable in any practical sense (even though he himself gave some apparent counter-examples). Yet this was exactly my point. Construed as a theory instead of as an empirical generalization, "no dirt is fit to eat," is not (normally) falsifiable, and this is why, on the Popperian model, we would need to keep sampling dirt for the rest of out lives.

Things get curiouser and curiouser as Daniel goes on to say: "Critical Rationalism never tries to prove anything - only to falsify." Exactly. The inability, as a practical matter, to falsify the theory that "no dirt is fit to eat" means that we would never have a good reason not to try eating dirt again, and again, and again....

Permit me to interject some common sense here:

Why do we generalize that dirt is bad to eat on the basis of a few samples? Upon eating dirt as children, did we merely think, in effect: This dirt is bad; that dirt is bad; therefore all dirt is bad? No, we did not merely enumerate instances. Rather, after a few samples, we thought, "This stuff is dirt, and it is not good to eat. So when I encounter other samples of the same kind , I will know that it is also dirt and not fit to eat." In other words, we made a conceptual identification of the substance we call "dirt" and rationally concluded that other substances with the same nature will have the same negative effects.

In sum: We reasoned from particular instances to a generalization about the nature of dirt, and we then applied this generalization to instances as yet untried. This is pretty much what J.S. Mill had in mind when he said that induction consists of reasoning from particulars to particulars by means of an intervening generalization. (I discussed the problem of exceptions in my last post.)

Given these explanations, I don't see the need to respond to the remainder of Daniel's post. I would merely end up repeating myself.

Ghs

[George H. Smith]

This is another post about Popper that I posted early this morning on the "David Harriman's Book" thread. I believe this completes the transfer of relevant posts.

A few general comments about Popper and then I need to get some sleep....

I have been reading Popper for many years, and I am still frequently unclear about the status of his Conjectures and Refutations Model -- or CRM for short.

Specifically, is CRM a prescription for how we should reason? Or is it a description of how we actually do reason when we reason correctly?

At times Popper prescribes CRM not only a valid method of reasoning but as the only valid method of reasoning. This gives rise to a major objection that I have stated before, namely, that even if CRM is the only proper method for experimental science -- and that's a big "if" -- it does not follow that CRM is the only proper method for other fields, or even an appropriate method. (In The Poverty of Historicism, Popper attempts to extend CRM to history and the social sciences, and he does some serious tap dancing in the process.)

At other times, especially in his discussions of induction, Popper presents CRM as a description of how people actually do reason, whether they know it or not. Thus, when Popper claims to have solved the problem of induction, he means that he has shown that inductive reasoning has never played a role in the justification of knowledge -- for one thing, we never "justify" knowledge at all -- so induction is irrelevant. Although induction has served as a method of discovery (we sometimes get good ideas from it), it has never served as a method of justification or even of corroboration. What inductivists thought they were doing in the latter case was not really induction at all but some version of CRM.

This approach can get extremely bizarre at times. It verges on a type of cognitive psychoanalysis, to wit: In arriving at sound conclusions, you thought you were doing x, but I know better. I know that you were actually doing CRM. We see shades of this approach in my current debate with Daniel.

It is Popper's fixation on CRM as overarching grand theory, one applicable to every case and every discipline, that sometimes imparts a cult-like quality to his followers, who will twist and turn to show how every type of valid reasoning is ultimately a variant of CRM.

I may have simplified matters a bit here, but not by much.

Ghs

[George H. Smith]

I claim that empirical generalizations based on inductive reasoning are rationally warranted, if only as matters of probability. Popper, in contrast, claims that inductive reasoning is never warranted, that it cannot even yield probability in any degree.

Incidentally, Popper follows Hume in this. And they are both quite right.

If you mean "probability" in any more meaningful way than the entirely vague and hypothetical sense I mentioned above - that is, if you want to seriously justify one of your "empirical generalisations" claim to truth by an appeal to an actual probability - then perhaps you might want to specify that probability and show your workings.

This is yet another example of the pretentious meanderings of a Popperian who thinks that only the technical use of terms should be taken seriously, and that scientific methods and models are paradigms that should be applied to everyday life.

If I say that you will probably die if you shoot yourself in the head, my meaning is perfectly clear. I don't need to refer to a calculus of probability and come up with an exact number and then "show my workings"; such a thing would be impossible in any case.

Different types of justification and calculation are appropriate in different contexts. There is no one correct method.

Ghs

[Daniel Barnes]

1) As I illustrated in a post written earlier today, a post in which Popper quotes Hume, this was not Hume's argument. Hume says absolutely nothing about explanatory universal theories, or theories of any kind. This is Popper's application of Hume's argument to the realm of theory. "Instances of which we have no experience" are not the same as "explanatory universal theories." Popper can replace the former with the latter is he likes, but he severely distorts Hume's argument in the process.

When you say this "was not Hume's argument", surely you can't be trying to claim it is unrelated - that Hume's problem somehow doesn't apply to universal theories, or even theories of any kind? Because universal theories make predictions about "instances of which we have no experience", or unobserved events, as well as observed ones.

Are you really sure about that? Actually, skipping to the bottom, it seems you're not after all. So perhaps you might rephrase.

2) There was no need for Popper to "rescue" us from Hume's problem because there is no problem to begin with, as Hume conceived it. Even Popper didn't accept Hume's linking of the problems of induction and causation, but Hume's rejection of causation is absolutely fundamental to his points about induction. Hume's argument about the lack of a causal connection between observed events constitutes the foundation of his argument that inductive inferences are invalid. Hume understood that inductive reasoning presupposes causal reasoning, but Popper did not.

But Popper points out that the logical problem of induction is actually behind the problem of causation. He explains that in Hume there is "no sensational basis for the idea of [causal] necessity...the nearest to it which is observable is regular succession. But if the regular succession of two events were 'necessary' then it would also have to take place with certainty, not only among observed instances but also among unobserved ones. This is, essentially, the way in which the logical problem of induction enters Hume's subjectivist discussion of causation, his bucket-theoretical search for the origin or the basis of the idea of necessity." (Popper, Objective Knowledge p88).

I addressed this very point earlier today. Again: Hume's problem is not Popper's problem, and if Popper is only able to solve his problem (the one relating to theories rather than all empirical generalizations), then he has not solved Hume's problem. Popper will only have shown how universal scientific theories can be corroborated without appealing to induction. Fine, this may very well be true -- it certainly is not an original point in any case -- but Hume's problem was much more fundamental, broader in scope, and remains untouched by Popper's supposed solution.

So you claim. However, you talk rather a lot about "empirical generalizations" and "probabilities" and "hypothetico-deductive" models that are defended by inductivists such as yourself. But you have yet to supply any details of the aforegoing: say, how the truth content of these "empirical generalizations" is established, how the "probabilities" of these generalizations are calculated, and how a "hypothetico-deductive" model you would defend improves on the version offered by Popper.

I think it's about time you put these on the table so we can clearly see why you consider them a superior solution.

Also, are you still sticking by your "infinite regress" argument?

Edited September 10, 2010 by Daniel Barnes

[George H. Smith]

1) As I illustrated in a post written earlier today, a post in which Popper quotes Hume, this was not Hume's argument. Hume says absolutely nothing about explanatory universal theories, or theories of any kind. This is Popper's application of Hume's argument to the realm of theory. "Instances of which we have no experience" are not the same as "explanatory universal theories." Popper can replace the former with the latter is he likes, but he severely distorts Hume's argument in the process.

When you say this "was not Hume's argument", surely you can't be trying to claim it is unrelated - that Hume's problem somehow doesn't apply to universal theories, or even theories of any kind? Because universal theories make predictions about "instances of which we have no experience", or unobserved events, as well as observed ones.

I have explained this at least two times now, and I'm not not going to continue this discussion if you don't extend the courtesy of reading my explanations with a modicum of care.

One last time: Yes, of course, Popper's argument is related to Hume's argument. It is a specific application of Hume's argument to the realm of universal scientific theories. But if Popper can avoid Hume's problem only in the case of universal theories, then he has not solved Hume's problem. Hume's problem is far more fundamental than how we can verify or corroborate universal theories. (Nor does Popper "solve" Hume's problem in the realm of science; he merely does an end-run around it by claiming that induction is not needed.)

2) There was no need for Popper to "rescue" us from Hume's problem because there is no problem to begin with, as Hume conceived it. Even Popper didn't accept Hume's linking of the problems of induction and causation, but Hume's rejection of causation is absolutely fundamental to his points about induction. Hume's argument about the lack of a causal connection between observed events constitutes the foundation of his argument that inductive inferences are invalid. Hume understood that inductive reasoning presupposes causal reasoning, but Popper did not.

But Popper points out that the logical problem of induction is actually behind the problem of causation....

I am very well aware of Popper's claim. He is wrong. If you want to know about Hume's argument, then read Hume, not Popper.

I addressed this very point earlier today. Again: Hume's problem is not Popper's problem, and if Popper is only able to solve his problem (the one relating to theories rather than all empirical generalizations), then he has not solved Hume's problem. Popper will only have shown how universal scientific theories can be corroborated without appealing to induction. Fine, this may very well be true -- it certainly is not an original point in any case -- but Hume's problem was much more fundamental, broader in scope, and remains untouched by Popper's supposed solution.

So you claim. However, you talk rather a lot about "empirical generalizations" and "probabilities" and "hypothetico-deductive" models that are defended by inductivists such as yourself. But you have yet to supply any details of the aforegoing: say, how the truth content of these "empirical generalizations" is established, how the "probabilities" of these generalizations are calculated, and how a "hypothetico-deductive" model you would defend improves on the version offered by Popper.

I have shown that empirical inductive generalizations have a rational foundation, and I have given a number of explanatory examples. That's all I need to do. If you have a problem with my arguments, then address my arguments. I am not required to give a full-blown theory of empirical generalizations. I need only address the fundamentals, and I have done this several times.

Moreover, I said nothing about my version of the hypothetico-deductive model being better than Popper's. I said that he does not employ the standard model. In many ways I prefer Popper's version.

Nevertheless, as I have repeatedly pointed out, it is absurd to demand that we employ this model (in whatever form) in the business of everyday life, where reliable empirical generalizations work just fine. If this were not the case, we wouldn't live very long. We would be killed while walking in front of buses and trains and in many other ways. The empirical generalizations we employ every day are (generally) reliable because inductive reasoning is valid. Period.

Similar things do indeed behave in similar ways in similar circumstances. This is the causal foundation of inductive reasoning and the resulting empirical generalizations. No need for Popperian conjectures and refutations or the hypothetico-deductive model. Popper sometimes claimed to be a fan of common sense, but he missed the boat on this one. That's why I previously characterized his arguments against induction as silly. I was serious.

Ghs

[Daniel Barnes]

I claim that empirical generalizations based on inductive reasoning are rationally warranted, if only as matters of probability. Popper, in contrast, claims that inductive reasoning is never warranted, that it cannot even yield probability in any degree.

Incidentally, Popper follows Hume in this. And they are both quite right.

If you mean "probability" in any more meaningful way than the entirely vague and hypothetical sense I mentioned above - that is, if you want to seriously justify one of your "empirical generalisations" claim to truth by an appeal to an actual probability - then perhaps you might want to specify that probability and show your workings.

This is yet another example of the pretentious meanderings of a Popperian who thinks that only the technical use of terms should be taken seriously, and that scientific methods and models are paradigms that should be applied to everyday life.

If I say that you will probably die if you shoot yourself in the head, my meaning is perfectly clear. I don't need to refer to a calculus of probability and come up with an exact number and then "show my workings"; such a thing would be impossible in any case.

Clearly you didn't read this from my earlier post

Thus it seems to me [Popper's] L2 has about the same claim to truth status as your "empirical generalizations", which might also turn out to be false. That is, none. Likewise, there is no problem for Popperians if they make similarly vague "empirical" or "reasonable generalizations" such as "It usually rains in winter" or "the bus stops here at 8 o'clock" or "I will probably get presents this Christmas" . This is simply loose, colloquial speech - no-one expects even an inductivist to produce, say, their probability workings for such comments - and in no way tries to pretend they are somehow "warranted", or are intended as a universal truth claim that droughts can never happen in winter nor the bus is never late.

So your accusation is quite incorrect.

Different types of justification and calculation are appropriate in different contexts. There is no one correct method.

I agree. You can use whatever method you like on whatever problem you like! There's not even a requirement to be rational. What you do is up to you. No Popperian policeman is going to force you to falsify. However, we are talking about logically valid methods. If your "cognitive shortcuts" or "empirical generalizations" or whatever you will call them next don't have any logical validity, then they are as helpless against Hume's problem as any universal theory. You can make pretentious, meandering statements about them being "rationally warranted", or "justified", or whatever - but that is all just hot air, as we both know what the real problem is.

Further, if you are going to come out with language like this already cited above:

I claim that empirical generalizations based on inductive reasoning are rationally warranted, if only as matters of probability. Popper, in contrast, claims that inductive reasoning is never warranted, that it cannot even yield probability in any degree.

...claiming that, "in contrast" to Popper, there is some actual "degree" of predictive probability that your rationally-warranted-empirical-generalizations-based-on-inductive-reasoning will "yield", it is understandable that the reader might be given quite the wrong impression.

Given that you obviously can "yield" nothing like a "degree" of probability yourself either - in practice there is no "contrast" at all, as both yours and Popper's claims would be zero - here's how you perhaps should have written your passage.

GHs revised: "I claim that empirical generalizations based on inductive reasoning are rationally warranted, whilst accepting they are logically invalid, if only as matters of probability, even though this too was demonstrated to be invalid by Hume. Thus, when I use "probability", or "degree" I can really only mean it in the vaguest possible sense of the words, with no claim to any actual degree of probability. Much in common my own position, Popper claims that inductive reasoning cannot yield probability in any degree either, although, like myself, he will happily use phrases like "probably" or "in all probability" in a merely vague, colloquial sense that commits him to nothing."

Edited September 10, 2010 by Daniel Barnes

[George H. Smith]

I claim that empirical generalizations based on inductive reasoning are rationally warranted, if only as matters of probability. Popper, in contrast, claims that inductive reasoning is never warranted, that it cannot even yield probability in any degree.

Incidentally, Popper follows Hume in this. And they are both quite right.

If you mean "probability" in any more meaningful way than the entirely vague and hypothetical sense I mentioned above - that is, if you want to seriously justify one of your "empirical generalisations" claim to truth by an appeal to an actual probability - then perhaps you might want to specify that probability and show your workings.

This is yet another example of the pretentious meanderings of a Popperian who thinks that only the technical use of terms should be taken seriously, and that scientific methods and models are paradigms that should be applied to everyday life.

If I say that you will probably die if you shoot yourself in the head, my meaning is perfectly clear. I don't need to refer to a calculus of probability and come up with an exact number and then "show my workings"; such a thing would be impossible in any case.

Clearly you didn't read this from my earlier post

Thus it seems to me [Popper's] L2 has about the same claim to truth status as your "empirical generalizations", which might also turn out to be false. That is, none. Likewise, there is no problem for Popperians if they make similarly vague "empirical" or "reasonable generalizations" such as "It usually rains in winter" or "the bus stops here at 8 o'clock" or "I will probably get presents this Christmas" . This is simply loose, colloquial speech - no-one expects even an inductivist to produce, say, their probability workings for such comments - and in no way tries to pretend they are somehow "warranted", or are intended as a universal truth claim that droughts can never happen in winter nor the bus is never late.

So your accusation is quite incorrect.

The issue here is whether inductive reasoning is ever rationally warranted. Hume said never, absolutely never, not even in regard to the empirical generalizations that we make in everyday life. So let's clear this up: Do you agree with Hume?

More specifically, do you agree with Hume that we have no rational basis whatsoever for assuming that future events will even resemble past events?

Ignore the rest of this post, if you like, but please address this issue. And don't hide behind Popper's skirts. Explain your own position.

Different types of justification and calculation are appropriate in different contexts. There is no one correct method.

I agree. You can use whatever method you like on whatever problem you like! There's not even a requirement to be rational. What you do is up to you. No Popperian policeman is going to force you to falsify. However, we are talking about logically valid methods. If your "cognitive shortcuts" or "empirical generalizations" or whatever you will call them next don't have any logical validity, then they are as helpless against Hume's problem as any universal theory. You can make pretentious, meandering statements about them being "rationally warranted", or "justified", or whatever - but that is all just hot air, as we both know what the real problem is.

I said that different types of justification are appropriate in different contexts. How could you possibly construe this to mean that we may choose whatever methods we like?

Inductive reasoning is "logically valid," and I have explained why on several occasions. But you have ignored those parts of my posts.

Given that you obviously can "yield" nothing like a "degree" of probability yourself either - in practice there is no "contrast" at all, as both yours and Popper's claims would be zero....

If you shoot yourself in the head, would you say that the probability that you will die is zero because we cannot specify precisely what that probability is in quantitative terms? As for degrees, very probable is a degree. The attempt to quantify such matters would be witch-doctory -- the sort of thing that impresses people with a mystical view of numbers and mathematical calculations that they do not understand..

Ghs

[anthony]

Gents,

To interject, and inject a dose of "everyday life" into this enjoyable debate.

From the viewpoint of a practised inductivist, I'd like to suggest that the method of induction is a constantly self-checking mechanism, which narrows down those pesky probabilities, into possibilities, and then close-to-definites.

All the more effective, if one is rigorously using deduction to chuck out pet hypotheses that fail.

In short,I appreciate the terms "empirical generalizations", and "cognitive shortcuts."(Ghs)

As a last thought, it has often seemed to me that trained empiricists and deductivists -eg,logicians and scientists - appear to deny, distrust,and resent induction as a tool, because it's too'common'or 'easy'; not merely because it may be unreliable.

This has to be self-defeating IMO.

Thanks

Tony

[George H. Smith]

Gents,

To interject, and inject a dose of "everyday life" into this enjoyable debate.

From the viewpoint of a practised inductivist, I'd like to suggest that the method of induction is a constantly self-checking mechanism, which narrows down those pesky probabilities, into possibilities, and then close-to-definites.

All the more effective, if one is rigorously using deduction to chuck out pet hypotheses that fail.

In short,I appreciate the terms "empirical generalizations", and "cognitive shortcuts."(Ghs)

As a last thought, it has often seemed to me that trained empiricists and deductivists -eg,logicians and scientists - appear to deny, distrust,and resent induction as a tool, because it's too'common'or 'easy'; not merely because it may be unreliable.

This has to be self-defeating IMO.

Thanks

Tony

I have been discussing induction in the context of everyday life, or what is sometimes called "common sense." If we employ different cognitive standards in the practical reasoning of everyday life than we use in the physical sciences, this is because we are more concerned with concretes -- How should I act in this or that circumstance? -- than we are with universal laws. This does not mean, however, that practical reasoning is somehow less valid than theoretical or scientific reasoning.

There is a very interesting discussion of "common sense" in Bernard Lonergan's book, Insight: A Study of Human Understanding (rev. ed., Harper and Row, 1978). After devoting many chapters to the nature of scientific and mathematical knowledge (with chapter titles like "The Problem Peculiar to Physics" and "The Description of Space and Time"), Lonergan moves on to other types of "intelligence" that produce their own "insights."

According to Lonergan (pp. 176-77), common sense differs from science "in the meaning it attaches to analogies and generalizations." He continues:

Common sense [in contrast to science] never aspires to universally valid knowledge and it never attempts exhaustive communication. Its concern is the concrete and particular. Its function is to master each situation as it arises.

Roy Childs turned me on to Lonergan's lengthy (750 pages) and difficult book many years ago. Lonergan is a Catholic philosopher who doesn't toe any orthodox line. Insight has been praised as one of the most original books in its field. Here is what Stephen Toulmin, a highly esteemed philosopher of science, said about it:

Insight is a masterly work, whose importance reaches far beyond the boundaries of theology and Catholic philosophy. It has much to say of interest and significance to cognitive psychologists and students of epistemology. Lonergan's careful scrutiny of the procedures by which we put our creative intelligences to work is precise, lucid, and fascinating.

Ghs

[George H. Smith]

I have been rereading J.S. Mill's famous and influential discussion of induction in A System of Logic: Ratiocinative and Inductive . His succinct description of inductive reasoning is as good as any I have seen:

Induction, then, is that operation of the mind by which we infer that which we know to be true in a particular case or cases, will be true in all cases which resemble the former in certain assignable respects. In other words, induction is the process by which we conclude that what is true of certain individuals of a class is true of the whole class, or that which is true at certain times will be true in similar circumstances at all times. (Book III, Chapter II.)

Mill goes on to point out that induction presupposes causation:

We must first observe that there is a principle implied in the very statement of what induction is; an assumption with regard to the course of nature and the order of the universe; namely, that there are such things in nature as parallel cases; that what happens once will, under a sufficient degree of similarity of circumstances, recur. This, I say, is an assumption involved in every case of induction.

...This universal fact, which is our warrant for all inferences from experience, has been described by different philosophers in different forms of language; that the course of nature is uniform; that the universe is governed by general laws; and the like. (Book III, Chapter III.)

This last passage indirectly illustrates why Hume's critique of causation was the linchpin of his rejection of inductive reasoning.

Mill also makes a point that I had overlooked in my posts on this topic, namely, that inductive inferences are not limited to future events:

We believe that fire will burn tomorrow because it burned today and yesterday; but we believe, on precisely the same grounds, that it burned before we were born, and that it burns this very day in Cochin-China. It is not from the past to the future, as past and future, that we infer, but from the known to the unknown; from facts observed to facts unobserved; from what we have perceived, or been directly conscious of, to what has not come within our experience. In this last predicament is the whole region of the future; but also the vastly greater portion of the present and of the past.

Ghs