For a lengthy critical treatment of Popper (which I recall being good but which I haven’t read in years),try this.
For another trashing of Popper, Lakatos, Kuhn and Feyerabend, click this thang.
I heard about this from my control theory professor. A philosopher named Karl Popper (He’s a famous one, as far as I have heard) holds the position that all theories are only ‘true’ until falsified. In fact he hold that theories are not really true, but that they are hypotheses that still need to be taken with a grain of salt.
He also holds that only theories that could be falsified are scientific. Stuff like ‘Intelligent Design’, for example, isn’t.
My problem with this is: This theory cannot be falsified. It has to be taken on faith.
It seems like a twist of Hume’s argument to me: “You can never know the truth, and this is true.”
But he says that if you falsify a theory, then you have to discard it, which is reasonable. But still he adds Hume’s element of uncertainty into it.
Any thoughts?
Let’s take a specific theory: Newtonian Mechanics.
We know Newtonian Mechanics is wrong. We’ve falsified it. But we don’t toss it because it is still useful. Newtonian Mechanics is good enough to get us to the moon and back, and it is good enough for very wide range of applications.
Of course you can’t use Newtonian Mechanics for something like the GPS, because it *is* wrong and you have to bring in General Relativity.
Newtonian Mechanics is useful even if falsified.
The actual case is that we know that every one of our physical theories is falsified by some piece of empirical data. We don’t take this to mean that we don’t have physical theories at all, which your characterization of Popper’s theory would require we do.
I suppose we could save Popper’s theory by requiring that every physical theory have a range of validity and some sort of specification of error so that as long as the physical data falls within the range of error we’ll say it is okay. But Popper is really concerned with some notion of absolute truth and discarding any theory when the first falsehood is found.
What is really needed is some sort of verificationist account where accumulated evidence supports a theory and gives us confidence in its application (which maps a little better to what humans actually do).
But Popper is really concerned with some notion of absolute truth and discarding any theory when the first falsehood is found.
One of the consequences of Popperian nihilism is that you also cannot objectively falsify a theory, leading to the impossibility of knowledge. The idea behind a falsifier is that a certain observation would be made when a theory predicts that there is an X at some place, and then because there is no X you know the theory is false. This depends on the observation, which in science is usually indirect because it involves an instrument. Instruments are founded on scientific theories and they work because the underlying theory is true (or they don’t, because the theory is false). But Popperian logic precludes saying that the underlying theory of the instrument is true – all you can say is that it hasn’t yet been falsified. Every supposed falsifying observation is therefore due to the falsity of the theory being tested, or the falsity of the theory of the instrument. Consequently, observations are insecure in Popperworld.
What is really needed is some sort of verificationist account where accumulated evidence supports a theory and gives us confidence in its application (which maps a little better to what humans actually do).
We have that.
One of the consequences of Popperian nihilism is that you also cannot objectively falsify a theory, leading to the impossibility of knowledge. The idea behind a falsifier is that a certain observation would be made when a theory predicts that there is an X at some place, and then because there is no X you know the theory is false. This depends on the observation, which in science is usually indirect because it involves an instrument. Instruments are founded on scientific theories and they work because the underlying theory is true (or they don’t, because the theory is false). But Popperian logic precludes saying that the underlying theory of the instrument is true – all you can say is that it hasn’t yet been falsified. Every supposed falsifying observation is therefore due to the falsity of the theory being tested, or the falsity of the theory of the instrument. Consequently, observations are insecure in Popperworld.We have that.
For measuring instruments:
A strict Popperian would say that we have to treat *all* of the theories that apply to the measuring instrument (i.e. both the theory being tested and the theory of the instrument itself) as a single theory being tested. That is we aren’t allowed the luxury of separating the two. So if we don’t see what we expect then we know that the theory:
(Theory of operation of the instrument) & (Theory under investigation)
is false although we cannot be sure whether
Theory of operation of the instrument
or
Theory under investigation
or both are false. In fact we shouldn’t even presume they can be nicely separated as I have just done above.
So basically this isn’t really an objection to a strict Popperian, it might be an odd way of thinking about things to most people, but it isn’t problematic in the way you indicated.
The only observations that exist in Popper-world are basically everyday sense observations. So I might set up an apparatus that turns on a red light when the temperature goes above a certain value. I am never observing the temperature, I am always observing the red light. Thus if I set up an experiment that according to theory should cause the red light to turn on in ten minutes (that is in ten minutes the temperature should exceed the bound), my observation is only that the light went on in ten minutes or didn’t go on in ten minutes. If it didn’t go on then I know that something was wrong with the sum total of all the theories involved.
Of course this might lead to the need to include a psychological theory of the observer into the whole mix and whether they are going to see the red light, or think it was green or whatever.
As for the theory we have:
What verificationist theory do we have?
So basically this isn’t really an objection to a strict Popperian, it might be an odd way of thinking about things to most people, but it isn’t problematic in the way you indicated.
I think it is still problematic, when you have an epistemology that can’t come up with anything stronger than “something is wrong”. If all you can say for sure is “I don’t know if this is false” and “something is false”, then you don’t have knowledge.
What verificationist theory do we have?
Chapter 5 of OPAR gives the basics principles, especially the discussion of “certainty”.
I think it is still problematic, when you have an epistemology that can’t come up with anything stronger than “something is wrong”. If all you can say for sure is “I don’t know if this is false” and “something is false”, then you don’t have knowledge.Chapter 5 of OPAR gives the basics principles, especially the discussion of “certainty”.
You definitely have a much weakened notion of knowledge. That in itself isn’t an objection though. The strict Popperian is going to say “Sorry, you’re idea of a stronger knowledge is just wishful thinking and unjustifiable. This weakened knowledge is the best we can do.” We obviously can’t live in a dream world demanding things we can’t have.
So the onus is on the person who thinks we can have stronger knowledge to give a good account of that knowledge.
As for a theory of verificationism, I had in mind something with a little more technical rigor whereby I can compare in a strict way two competing theories both of which have some support in empirical data, as well as some data contradicting the theory, and decide which is the better verified. Or maybe given a piece of empirical data I can decide which theory it better supports. I don’t recall OPAR doing anything of this kind of elaborated rigor.
One of the consequences of Popperian nihilism is that you also cannot objectively falsify a theory, leading to the impossibility of knowledge.
Very good point. I was never able to put this in words. Thanks fot the link, too. 
If that’s your only problem with it, you’re missing the more important point: Popper’s theory isn’t true. There’s no reason to accept. There are absolutely no grounds for concluding that we can only know what’s false and not what’s true.
For a lengthy critical treatment of Popper (which I recall being good but which I haven’t read in years),try this.
Thanks for the link. I haven’t read the text by now, but I will read it tomorrow. 
I just wanted to say with this that Popper contradicts himself. This means that he is wrong. But thanks for pointing it out more clearly.
I remember another member bringing up “falsifiability” in another thread, here, nice to know what philosopher is responsible for this silliness.
You’re welcome! This guy was heavily influenced by Popper.
You definitely have a much weakened notion of knowledge. That in itself isn’t an objection though.
It is an objection, but not one that a Popperian would accept. The root issue seems to be a nearly axiomatic principle that knowledge is an invalid concept, from their POV.
So the onus is on the person who thinks we can have stronger knowledge to give a good account of that knowledge.
It looks that way, except that I don’t know of anything that can budge them from their primary principle, that knowledge of truth is impossible and that at most you can have knowledge that an untruth exists. Since their theory of knowledge is unfalsifiable, of course, I don’t know what would constitute a refutation of Popperian epistemology.
As for a theory of verificationism, I had in mind something with a little more technical rigor whereby I can compare in a strict way two competing theories both of which have some support in empirical data, as well as some data contradicting the theory, and decide which is the better verified. Or maybe given a piece of empirical data I can decide which theory it better supports. I don’t recall OPAR doing anything of this kind of elaborated rigor.
What kind of rigor do you find lacking in OPAR (specifically ch. 5)? It’s true that Peikoff does not give an elaborate case study, because he’s presenting a philosophy rather than an account of a scientific domain. If you mean, how do you decide between extensionally distinct theories where the test case is presently beyond our reach and therefore both theories fit the known facts equally, then (1) try to do the crucial experiment and (2) don’t decide.
It is an objection, but not one that a Popperian would accept. The root issue seems to be a nearly axiomatic principle that knowledge is an invalid concept, from their POV.It looks that way, except that I don’t know of anything that can budge them from their primary principle, that knowledge of truth is impossible and that at most you can have knowledge that an untruth exists. Since their theory of knowledge is unfalsifiable, of course, I don’t know what would constitute a refutation of Popperian epistemology.What kind of rigor do you find lacking in OPAR (specifically ch. 5)? It’s true that Peikoff does not give an elaborate case study, because he’s presenting a philosophy rather than an account of a scientific domain. If you mean, how do you decide between extensionally distinct theories where the test case is presently beyond our reach and therefore both theories fit the known facts equally, then (1) try to do the crucial experiment and (2) don’t decide.
The idea behind a verificationist account is to be able to resolve something along the lines of the following:
-
We have two theories A and B which do not have a do-able crucial experiment (say it is far beyond our current technical level
-
We have a body of data which supports both theories, but actually deviates from both of them in various ways
-
We want to decide nonetheless which theory is better supported by the body of data
The verificationist would want some sort of calculus (which would probably be some sub-discipline of statistics) which would resolve this sort of issue, as well as provide some sort of measure of confidence in a single theory given a body of data.
I guess the key idea here is *confidence* in a theory given the data. There is no such thing as confidence in a popperian account. The theory either guilty or innocent until proven guilty (as it were).
As we established above, the popperian has to say we currently have *NO* scientific theory worth speaking of, as they have all been shown to be wrong by the evidence. Of course we believe all those theories are good for something, the point of verificationism is to give some idea of how good they are.
The alternative you offer to just say “we don’t know which one is right” and throw our hands up in the air seems like giving up as well.
An aside:
It just occurred to me that I’m using the term “verificationism” which is a term one usually sees used with some form of logical positivism (where they say that the meaning of a statement is the means of verifying it, which obviously is not at all what I’m talking about). I don’t want there to be any confusion that I consider this usage of “verificationism” as distinct from that usage.
The idea behind a verificationist account is to be able to resolve something along the lines of the following:
Okay, before we go there, I just want to point out that that is not an appropriate definition, in the context of naming a theory that dimetrically opposes Popper. Popper takes the purpose of scientific inquiry to be to “show to be false”: a verificationist takes the purpose of scientific inquiry to be to “show to be true”. That means, for example, that experiments which confirm a theory are probative, not pointless (a la Popper). Couple this with the concept of “certainty”, and you have a means to prove that a theory is true.
- We want to decide nonetheless which theory is better supported by the body of data
The point is that neither theory is “better supported” by the data, so no metric can or should say “A is better because it scores more points”. Bear in mind that a scientist is not compelled to only consider two pre-constructed package deals, so if you know that A fails to explain behavior under extreme pressures and B fails to explain behavior under extreme cold, then you need to devise C which factors out both sets of mistakes (which you can presumably do since you know the nature of the errors).
As we established above, the popperian has to say we currently have *NO* scientific theory worth speaking of, as they have all been shown to be wrong by the evidence.
Apart from the basic nihilism of Popperian epistemology, this is the other vile problem that I see with their view of science. It is sour-grape science, throwing out the baby, the bathwater, the tub, and burning down the house just in case there is a spill. It is basically the temper-tantrum elevated to the status of a theory of science. The verificationist says “Look, A messes up with pressures and B messes up with temperature, but they otherwise get the facts right. Now what really makes these theories different is that A has this mass term inversely related to Q and B has that energy term logarithmically related to Z, so if you switch these things around and restate W, they you get a new theory C that handles both cases”. The Popperian, however, say “See, I still know nothing!”.
Okay, I am not an expert on Objectivism, but I cannot understand the Objectivist position on the concept of falsifiability. The general consensus among people on this board seems to be that science should not be falsifiable. I disagree, and here’s why:
For a scientific theory to be “falsifiable” means merely that you must be able to state the conditions under which it would be proven false. People on this board take it to mean that a theory must actually be proven false. This is not the way that Karl Popper or Thomas Kuhn use the term.
So, a scientific theory is valid if and only if you are able to state the conditions under which it would be disproven. Thus, the theory of gravity is valid because we can state the conditions under which it would be proven wrong. If I dropped my pencil and it just hovered in the air, gravity would be proven wrong. I have just demonstrated how the theory of gravity is falsifiable, but I have not proven it wrong because such a phenomenon has not been observed.
Freud’s theory of psychoanalysis is not falsifiable. Any evidence you come up with to prove psychoanalysis wrong will just elicit some whacked-out response about how you have repressed sexuality that stems from your toilet training when you were 3.
A better example is that of Creation. You can present a Creationist with evidence for Evolution, perhaps by carbon-dating a dinosaur bone to show that it is millions of years old. What will the Creationist say? “God made it that way.” So, after all the evidence that exists of the world being billions of years old, a Creationist can evade reality simply by saying that God intentionally made it to look older than it really is. There is no way to disprove such a theory. Since God exists outside of the universe, then there is no evidence which cannot be shrugged off with a simple “God made it that way on purpose.” This theory is not falsifiable, because any Creationist will be incapable of stating conditions under which he would be proven wrong. Thus, it is not a valid theory.
So, assuming that my definition of falsification is correct (which I’m about 99% sure it is), tell me where I’m wrong. If my definition is incorrect, please explain.
It’s not your definition they’re quibbling with, so how do your comments apply to this discussion?
So, assuming that my definition of falsification is correct (which I’m about 99% sure it is), tell me where I’m wrong.
Being able to state the conditions under which something would be disproven is a far cry from actually being able to disprove it. That’s my point, and it’s a distinction that I think other people in this discussion have failed to make.
Okay, I am not an expert on Objectivism, but I cannot understand the Objectivist position on the concept of falsifiability. The general consensus among people on this board seems to be that science should not be falsifiable. I disagree, and here’s why:
For a scientific theory to be “falsifiable” means merely that you must be able to state the conditions under which it would be proven false. People on this board take it to mean that a theory must actually be proven false. This is not the way that Karl Popper or Thomas Kuhn use the term.
No, you misunderstand. Popper set up falsifiability as the criterion or definition of science. This (and certain of its consequences, as I mentioned in my earlier post) isn’t good enough.
It isn’t sufficient, but it’s still necessary.
For a scientific theory to be “falsifiable” means merely that you must be able to state the conditions under which it would be proven false.
The word “falsifiable” means “capable of being shown to be false”. In order to show that a statement is false, it has to be false. That, of course of “show” but not “claim” – you can claim that something is false without it actually being false. But in order to show that something is false, it has to actually be false.
People on this board take it to mean that a theory must actually be proven false.
No, that’s clearly not true. That’s the difference between the potential and the actual – it may be possible to demonstrate something, even though it has not already been done.
This is not the way that Karl Popper or Thomas Kuhn use the term.
Can you point me to a couple places where Kuhn uses the term “falsifiable”? I just don’t recall that being his bag – he’s more into the “science is all social ediface” routine. As for Popper, it’s true that he uses the word “falsifiable” in a special way, but I won’t use the word incorrectly; however, you will notice, I hope, that in post #24 and other places I’ve identified the valid concept of question-begging as being the evil that is supposedly guarded against by “falsificationism”. The guy who started all of this was apparently following some mystic rather than Popper, so the focus on Popper’s errors hasn’t been that strong – it would probably be best to focus this specifically on the epistemological crimes of Popper, if you really want to pursue this.
So, a scientific theory is valid if and only if you are able to state the conditions under which it would be disproven.
Or proven: but wait, I hate to pick, but you don’t have the Popperian view stated correctly. Falsifiability is Popper’s so-called “criterion of demarcation”. It has nothing to due with validity. A theory which is false is not valid.
Nobody here is arguing for Freudian psychology or divine creation. The issue is scientific epistemology – is it possible it have scientific knowledge. Popper does not correctly identify the principle that distinguishes contentless verbiage like Freudian psychology or Creationism from real scientific theories like Newtonian gravity or evolution. The real crime of Popperianism is not the term “falsification”, but the fact that it is thoroughly nihilistic. I don’t want a science with falsifiable statements, I want a science with testable, _ true _ statements. If you can show that all of my statements are false, then I would not have any knowledge, which would suck.
You have science with testable statements. In science, however, things can only be disproven. You cannot empirically prove that gravity works the way we think it does by any other means than disproving all of the alternatives. Given that our explanation seems to be the best, we accept it as true, but it can’t be proven.
You have science with testable statements. In science, however, things can only be disproven. You cannot empirically prove that gravity works the way we think it does by any other means than disproving all of the alternatives. Given that our explanation seems to be the best, we accept it as true, but it can’t be proven.
This is the part that I’m saying is wrong. Things can be proven in science. Do you understand why Popperian epistemology doesn’t even allow you to disprove a scientific statement?
DavidOdden’s statements are consistent with what I’ve read/seen on this topic.
I’ve seen Popper’s nihilistic epistemology applied in my field (finance), leading to destructive results. In effect, it tells academics to abandon induction as a means to knowledge, because the only real knowledge you can have is of what has been proven false. Implied by this falsafiability concept is that all “testable”, (including highly rationalistic) theories that have not yet been proven false, are essentially equal. This rationalistic, anti-inductive bias has encouraged finance academics to spend most of their time working on existing databases and statistical methods, each hoping to disprove (falsify) each others’ rationalistic hypotheses. Overall, very little time is spent going out into the finance world, talking with investors, companies, and other market participants trying to formulate concepts and principles based upon actual observations of the world.
Ultimately, Popper’s approach is anti-knowledge, anti-certainty, anti-Aristotle. While I can see how some could be attracted to its surface appeal to “hard-headed” scientific method, it takes the exact opposite approach required of a healthy epistemological method. And there’s nothing good in “falsifiability” that would not already be required by the process of integration (namely, recognizing that contradictions do not exist).
Things can be proven in science.
How? How can you prove that gravity works the way we think it does? I think we can be fairly certain, but I won’t rule out the possibility that we’re wrong. If the history of science has taught us anything, it’s that our most basic scientific beliefs often turn out to be wrong. Newtonian physics may very well be correct, but I don’t see how we can claim to know for absolute certain that it won’t be replaced with something more consistent, in the future.
Do you understand why Popperian epistemology doesn’t even allow you to disprove a scientific statement?
…not really. I was afraid that this conversation would quickly rise above my head, and it appears that I was right.