by Jerry O’Driscoll
Some recent controversies move me to take up the topic within the limitations of a blog post. Many years ago (1956), Fritz Machlup ably addressed the issue in an essay titled “The Inferiority Complex of the Social Sciences.” He rejected limiting the term science to particular subject matters or methods. He concluded that “there is no epistemologically defensible borderline short of the widest meaning of scientific method, defined in the Encyclopedia Brittanica as ‘any mode of investigations by which impartial and systematic knowledge is acquired.’”
I endorse Machlup’s broad definition of science as any systematic study of a subject. As he observed in a footnote, the German Wissenschaft is more inclusive: “the historians of literature, the philologists, the philosophers, the mathematicians, the sociologists, they are all scientists (Wissenschaftler).” In French, science is knowledge and one can speak of la science infuse, intuitive knowledge. La science de l’art is simply the systematic study of art.
A common barb to launch against an opponent is that he espouses religion, not science. By what criteria? Is it his subject matter or method? As Machlup reminds us, neither is a legitimate criterion for separating science from non-science. The charge is an exercise in confusion. There certainly can be a science of religion, i.e., a systematic study of religion. For contemporary examples, look at the works of the sociologist Rodney Stark. Moreover, the barb assumes religion itself cannot be reason based. Stark’s The Victory of Reason is an antidote for that unsubstantiated belief.
By all means let us disagree. But let us shed our inferiority complexes and not wrap ourselves in the mantle of science as defense. That is scientism, not science.
ThinkMarkets 
Isn’t the “impartial” part of Brittanica’s definition the problematic part, particulary for the social sciences? My own educational background is in sociology/anthropology and looking back, there was nothing impartial about them at all. The theoretical foundations of both seemed to me, during my time at college, to have definate leftist philosophical underpinnings and nothing I’ve come across recently has disuaded me of this view.
Couldn’t one also be systematic but completely wrong in their interpretations of a given phenomenon? How else would one account for the persistance of bad ideas?
Great post, Jerry. Machlup nailed it on this as on so many other issues. Methodology of economics premised on a bogus model of what physicists do isn’t likely to improve our understanding, let alone pratice. Methodology premised on a correct model of what physicists do may be no better if the methodologist forgets that matter doesn’t talk.
The term “science” is political.
Each side seeks to appropriate the mantle for itself while denying it to opponents. With no change in respective views, all that occurs is a shift of categorisation. However, this has powerful political implications with regard to respect, authority, and funding.
Perhaps the thing to encourage is a broad definition of “science,” because like inflation, it might eventually render the term worthless as a political tool.
While I agree the social sciences should not feel second rate compared to physics, I think this comment does not capture what I think is crucial about science as distinguished from other kinds of knowing.
First, science is not done by a single scientist because it is a process of testing and discovery where a crucial condition is that a claim can be examined by others. Challenges and testing are vital. Leonardo da Vinci was a scienTIST, because he systematically studied things, but he did not do scienCE as we think of it today because he did not make his results available for examination.
Perhaps an economic equivalent will be two people engaging in a barter trade of found goods after encountering one another in a wilderness, and who then pass, on never to see one another again, while immediately consuming what they obtained (fish and berries). They did engage in an economic transaction but hardly participated in “the market” The network of relationships was absent.
Science is also a process of discovery. Euclid was a scientist in this sense and his geometry delineated a scientific program, but as I understand it, there is no active science of Euclidian geometry because nothing in it seems open to new discoveries. Science is a moving target in terms of content, an expanding ball where the outside boundaries are uncertain and so open to scientific study but the interior is as ‘hard’ and certain as we can imagine, and so no longer attracts the interest of scientists. Of course, as Einstein showed, sometimes the ball is much fuzzier than anyone imagined.
Because of this social dimension, science is not only systematic study, it is systematic study open to rational challenge. It is a discovery process. Popper went too far I think, because I don’t think refutation is quite as impersonally objective a process as he claims. Michael Polanyi has a crucial insight about that. But he had a point.
I would agree that if a proposition cannot be tested by some relatively impersonal observation, then it is not science. If the tests it passes are uninteresting or have other more accepted explanations that fit as well, the claim may be scientific in a formal sense, but will likely not attract the attention of scientists and so not become part of the process of scientific discovery.
One of the great challenges in social science is agreeing on standards for what constitutes a reasonable rebuttal of a theory or claimed observation. But logically I think the problem is identical to the same issue of physics.
I am trying to keep the social sciences firmly in what English speakers call ‘science”and not within the to me much larger category of Wissenschaft. To me, the words are not quite equivalent.
I appreciate all the comments and particularly Gus diZerega’s thoughtful post. He raises an interesting point about the social nature of science. His formulation is a bit too Popperian for my taste. The same issue arose at TheAustrianEconomists, where the topic is also being discussed today. Pete Boettke put up a quote by Mises on the difference between theory and history, which I think is on point. Gus diZerega is correct, fo course, that as conventionally used “science” is narrower than Wissenschaft. I endorsed the broader as opposed to narrower usage.
Gus,
Let’s say you’re right. Aren’t you then saying science is a spontaneous order? Now, if that’s right, isn’t it hard to know how it really works? All our methodology and philosophy of science is pretty dubious, because it does not adequately consider the complexity of the spontaneous order we call “science.” It would then seem to be a rather bold conjecture to say, “if a proposition cannot be tested by some relatively impersonal observation, then it is not science.” Perhaps the curious task of science studies is to demonstrate to men how little they really know about what they imagine to be “science”?
I am saying science is a spontaneous order. What we somewhat misleadingly call the scientific method serves a function rather similar to money. Prices give information. So do the practices of scientists. Just like solid money, practices that get lots of credibility are taken more seriously than those that do not.
Because science’s underlying value is universal agreement about reliable knowledge concerning the material world, its standards grew out of the pursuit. A means for making a case is stronger the more universally accepted it is. Measurement is universally accepted in principle, though particular measurements can be challenged. The same holds for prediction and experiment. No philosophers need be involved.
As I understand the history of science, they ‘spontaneously’ emerged as standards because people concerned with studying physical phenomena found them persuasive.
I think, like markets, we are looking at signals promoting persuasion in complex phenomena. Details are simplified, hopefully without too much loss in assisting participants and their goals. Prices contribute overwhelmingly to that in markets, but sometimes over simplify – which is why, for example, some people like to “buy locally” even if prices are the same or even if they pay more for the local source. But usually price information is sufficient, and simplify tasks of cooperation. Mises showed how money emerged ‘spontaneously.’
The standards for agreement are fuzzier in science because the values reflected in its pursuit are different and more complex. They require agreement not on do I prefer X to Y and you prefer Y to X, regardless of why, but rather on explanations in physical terms for why something is observed – or what we might observe because of this agreement. But these standards emerged from the pursuit of as wide a range of agreement among competent observers (in the eyes of the scientific community) as possible. Measurement, prediction, and experiment won the day. Reason plays an important role as well.
Of course these standards are not neutral. They make assumptions about the nature of reality – and to the degree reality does not accommodate itself to these standards, others will hopefully emerge or the science that emerges will be seriously deficient. Human phenomena do not as easily accommodate themselves because meaning to the actor is such a important issue. Physical science has no room for inner meaning. This complicates things immensely.
But IMPERSONAL human phenomena that are complex adaptive systems, such as markets, do share a great deal with biological complex adaptive systems, such as evolution. And so economics can be a science. Ironically, it was the early social sciences that developed this insight – that was later adapted by Darwin. Ecologists trod a similar path.
All these scientific standards are particularly useful for subjecting claims to tests that all concerned will regard as fair, while making minimal claims as to what all parties must regard as acceptable criteria.
Nonlinear systems such as evolution and markets necessarily give up predictive power but have explanatory power great enough that those involved in science generally accept them as legitimate scientific endeavors. But because prediction is more difficult to assess people are less willing to accept that a given instance truly is evidence a claim is wrong, and so in the social sciences many theories continually vie for acceptance for a far longer time compared to most natural science.
So just like money in economics, the standards for scientific cooperation can arise spontaneously, given what people are trying to do. The quest for agreement lies at the base of both.
Any way, that’s how I understand it until I’m persuaded by another argument.
Gus
Okay we seem to have similar ideas here, Gus, but I don’t see how you’ve addressed my question. Here’s what I was trying to ask: Isn’t it a bold conjecture to say, “if a proposition cannot be tested by some relatively impersonal observation, then it is not science”? Note that your statement refers to each *proposition* that might be considered for acceptance within science.
You say, “Measurement, prediction, and experiment won the day.” Yes, pretty much, in physics. Evolutionary biology had come to have a good bit of measurement, prediction, and experiment. But what were their roles in Darwin’s Origins? I recall lots of observations in that book, but I don’t recall any predictions or experiments. I’m don’t recall numerical measurement in D’s 1959 book, but I do seem to recall lots of morphological comparisons, which might be considered a kind of measurement.
The critical phrase in Gus’ comment is the requirement for “explanations in physical terms for why something is observed.” That is not explantion in the social sciences. I also endorse Roger’s point.
OK, so let’s say we embrace the broad Wissenschaft definition for science. After all, words just mean what we agree they mean. So then, what should we call more narrowly defined endeavors? Is there a special name for that vein of inquiry that concerns itself exclusively with non-falsifiable claims? Or is there a word that describes the difference between work that has a long history of generating knowledge that can be turned into useful technology, as opposed to work that doesn’t?
Gus write, “Nonlinear systems such as evolution and markets necessarily give up predictive power…”
I’m not familiar with the evidence for that statement. Can you point me in the right direction?
A possible alternative explanation is that nonlinear systems need to be understood in enormous detail before their behavior may be accurately predicted.
Robert,
Check out the computability literature on that one. Here’s a way to start:
http://organizationsandmarkets.com/2009/03/02/computable-entrepreneurship/
I recall Hayek observing that, after writing his essays on the errors of scientism, Popper convinced him that scientists don’t really do what they say they do. Popper’s students went further to debunk the positivist model. So I’m not sure how useful the “narrow science” concept is. In my opinion, if narrow science is the paradigm, then economics is not a science. (Or, as Jim Buchanan phrased it, it’s an “almost science.”) For these and other reasons, the broad definition is more useful.
Science began with physical phenomena. The methods it developed were kept to physical description. When many scientists later tried to apply these methods to social phenomena they often resulted in reductionism of a fairly crude sort. I am not arguing as some might think that purely physical descriptions make up science as such. They only make up physical science.
But the original methods developed in that context because science sought reliable knowledge and those methods seemed to produce it best in the kinds of issues they explored. THIS is my point – not the methods but the problems. The problem was how to winnow out explanations that were inferior to others. The sciences of complex adaptive systems – to use today’s terminology – developed different ways of attacking issues of measurement and prediction compared to physics.
The social sciences were in many ways rooted in the Scottish Enlightenment and centered on spontaneous order phenomena. Adam Smith first used that method to examine how language developed – I don’t have the source in front of me – sorry. But it’s true. Later Smith wrote W of N.
That kind of method was later applied to evolution. Measurement exists in both the social sciences and in evolution, but initially it was qualitative. In evolution at first it was crude – visible variations in bones and such – but that’s what first made it possible to link species in chains. This kind of measurement is pretty sophisticated these days – but it’s a continuum of development from Darwin and Wallace.
As far as prediction is concerned, evolutionary theory was at first an example of what Hayek called “pattern prediction.” Because experiment was not possible at the time (Darwin did not know the mechanisms for change, for example) only predictions about what had already happened bt was currently inknown were possible. Evolutionary theory hit those predictions out of the ball park. Italso proved to link neatly with other fields where measurement was more developed, such as geology, and those other fields did have predictve power – such as where gold or oil might most reliably be found.
Now that has been supplemented with prediction of what would be found in the past if you looked at the right aged rocks laid down in the right circumstances. (The discovery of tiktaalik is an example of such prediction in action.)It’s a developmental sequence.
In the social sciences what Hayek would call “pattern prediction” seems always to have been a feature. The details cannot be predicted but fairly strong predictions can be made of many large scale phenomena and why. Rent control, anyone? It’s not 100% because data have to be interpreted by actors and we cannot know what opportunities they (or we) will see in the future especially because so much perception is contextual.
So what we have that physical scientists do not have is some understanding of the inner minds of people – but the predictive power that gives is limited. I am sure this is old hat to most readers.
Of course seeing patterns also required interpretation – but I’ve already argued that physical science requires interpretation as well. The point is to minimize the scope of untestable interpretation, not eliminate it. Polanyi is good here.
My point in all this is to say the goal of science wherever it exists is to find reliable knowledge and the only way to do that is find a way to winnow out unreliable knowledge. It’s harder with us social scientists, but not impossible. Ironically it was these methods developed in the nascent social sciences that gave biology a lift. We do not need to think of ourselves as science’s poor relations when the biological sciences learned one of their most fundamental concept from early social scientists.
For a practical example, as my work on democratic polities as spontaneous orders advanced I was able to shed light on why the “democratic peace” existed – liberal democracies do not fight one another in wars and they are the only kind of modern polity – perhaps any polity – where that is true. This is a pattern prediction that grew out of my study. It also predicts under what circumstances that peace might break down and what circumstances would strengthen these tendencies. So far as I know there are no exceptions to this peace – which is pretty good for the social sciences where exceptions abound. I would suggest these predictions are significant, measureable, and able to be rebutted.
Roger Koppl wrote,
This is a great point. I just recently noted on another thread that one problem with the scientistic-positivistic view is that on its own terms it’s not science, since this theory is not itself testable. Mises writes in Ultimate Foundation of Economic Science, p. 5:
See also on this topic Hans-Hermann Hoppe’s Economic Science and the Austrian Method, In Defense of Extreme Rationalism, p. 199, and The Economics and Ethics of Private Property, pp. 271-72.
There used to be a thing called “social studies” which aptly describe much of what goes on in sociology, political science, anthropology, big chunks of psychology, and a portion of economics. If science is nothing more than systematic explorations, then much of the humanities and fine arts qualify as well…and then the term “science” means what?
Gus,
Huh? Why are you lecturing on how evolutionary biology has measurement, prediction, and experiment? I said that already. I explicitly said, “Evolutionary biology had come [meant “has come”] to have a good bit of measurement, prediction, and experiment.”
You’re dodging my question: “But what were their roles in Darwin’s Origins?” And please don’t repeat about “qualitative” measurements, which I already said was in Origins.
I think there was ZERO quantitative measurement in Origins, ZERO prediction, and ZERO experiment. I also think it was thoroughly “scientific” work. What then of your statement that every *proposition* must be testable to be scientific? It just isn’t so. In fact your latest post seems to back off that claim by making your point something far less restrictive. “My point in all this is to say the goal of science wherever it exists is to find reliable knowledge and the only way to do that is find a way to winnow out unreliable knowledge.” That’s not real restrictive.
I’m hammering this testability thing for a reason. I like evidence and testability is an entirely good thing. Three cheers for testability. But demarcation criteria are hard to identify. And if you think science is spontaneous order, as I do, then we should be very hesitant indeed to identify one simple (simplistic) thing as the demarcation criterion. You seemed to say that *every proposition* had to be testable. Now we just aim at “reliability.” Thus are rather different points.
Jerry makes a fundamental point when he challenges demarcation criteria, though using different lingo. I think he is right and I think it supports Jerry’s position to point out that science in the more narrow sense is a spontaneous order.
I endorse Machlup’s broad definition of science as any systematic study of a subject.
So that’s it? Systematicity is the only criterion? Then isn’t astrology a science too, because of its “systematic methods”?
If so, what’s the point of even having a term for science? You should just lump everything into “making claims”.
Silas, I note you don’t propose an alternative definition. Methinks me detects scientism.
Darwin had plenty of prediction as to what we would NOT find in the rocks. Plenty. If someone had found a skeleton of a horse or a human nestled in among dinosaurs, or a modern fish in pre-Cambrian rocks, that would have been a pretty big cannon shot through Darwin’s argument. Evolution can potentially be rebutted through observation.
Negative prediction is still prediction, and for pattern explanations may be all we have.
I argued we had a shift from qualitative to quantitative measurement in evolutionary biology not, as you seem to suggest, a sudden shift from something lacking any prediction and measurement “ZERO” to prediction and measurement. That’s our disagreement if I understand you. Perhaps you can show me where this sudden shift in evolutionary biology took place?
Qualitative distinctions can turn into quantitative distinctions quite easily. The following isn’t from Darwin, but I think makes this point. As I recall, in his “Seeing Like a State,” James C. Scott describes a doctor who could spot syphilis at very early stages, when patients seemed asymptomatic to others. For a long time no one, not even the doctor, knew just how. It turned out he noticed, without quite being explicitly aware of it, very slight movements in their eyes. Those movements then became a diagnostic tool. This is as clear a shift from qualitative measurement to quantitative measurement as anyone could ask for.
So I do not think I am avoiding anything here.
I may have been unclear in my terminology, so let me be quite explicit. If something is utterly untestable or unable to even in principle be disconfirmed, it may be true but it is not science. Period.
Science is discovery and discovery requires the possibility of error and the means to avoid at least some of them, which means being able to recognize them.
Reason and logic can and do play an important role – but they are more persuasive when they are demonstrated by observable evidence. And persuasiveness is what it’s all about. Einstein thought he had come up with a persuasive thought experiment against quantum physics because, as he showed, it required nonlocality to exist, which he said was impossible. Many physicists of the time thought he had a strong point – until someone tested for nonlocality and found it to exist, even though it seems impossible. At the time of his thought experiment it apparently could not be tested. When it could be, it was – and that settled that issue. Of course people will disagree when results are unclear – hence the tendency to privilege quantitative methods when they can be devised.
Historically scientists were interested in contestable theories and explanations, and sought means to test them. Not as philosophers, but pragmatically. How do we choose between alternative explanations? We do not know what is really true. We have explanations that seem mutually exclusive. Reasonable people will seek to devise and perfect tests to winnow the field. What is so controversial about that?
Perhaps you can name me a scientific proposition that cannot be subjected to an attempted rebuttal by looking at some possible observable event. I’d be curious as to what it might be.
Silas, well said.
Stephan, why not define science according to the kinds of claims it deals with, and the kinds of claims it does not? For example, some say that science concerns itself exclusively with falsifiable claims. This distinguishes it from astrology.
More importantly, dealing only with falsifiable claims makes progress possible. If non-falsifiable claims are admitted into your ‘systematic method’ then inevitably you will find that you are confronted with mutually exclusive, non-falsifiable explanations for a phenomenon. How do you distinguish between them? You can’t, because neither claim can be disproved! You end up arguing endlessly…Doesn’t that sound like the social sciences, or at least like economics?
Limiting your line of inquiry to falsifiable claims means that you can distinguish between mutually exclusive claims, by demonstrating a contradiction in one claim or the other. All you have to do is invent an experiment that shows the contradiction, and that claim is ripe for the scrap heap (or at least for the rework heap).
So what is science? Science is winnowing down claims (by disproving the ones that can be disproved) until the ones that are left the claims that are most difficult to disprove preciseley because they are closest to the truth!
Roger, Thank you for the pointer, but do you have an unprotected link to the paper?
Robert J.: “Stephan, why not define science according to the kinds of claims it deals with, and the kinds of claims it does not? For example, some say that science concerns itself exclusively with falsifiable claims. This distinguishes it from astrology.”
If you assert “science concerns itself exclusively with falsifiable claims” and expect us to rely on this, it’s presumably true, and scientific. But it’s not falsifiable. So you have a contradiction.
Stephan,
Actually the statement “science concerns itself exclusively with falsifiable claims” IS falsifiable, if you regard it as a truth claim. All you need to do is find an instance of science concerning itself with a non-falsifiable claim (which is theoretically possible), and you’ve falsified this claim.
But treating this statement as a truth claim is a complete mistake. It is a statement of definition, not a factual claim.
Gus,
I don’t recall any predictions about the fossil record in Origins. Maybe I just forgot, but I am moderately confident they are simply not there. And yet it was a very persuasive work of science.
Now you want to turn the tables on me by challenging me to cough up a scientific proposition that is not testable. No fair. I did not deny that science tends to move toward “measurement, prediction, and experiment” and thus also toward testing. I just wanted to deny (and I continue to deny) that “if a proposition cannot be tested by some relatively impersonal observation, then it is not science.” I’m with Percy Bridgeman who said something like “apply your mind to the problem, no holds barred.” As our conversation reveals, lots of stuff starts out looking non-testable and then ends up testable. I would have thought that fact alone would have been sufficient to show your strong remark was in fact too strong. But it seems not.
It’s a bit of a dicey matter to cough up an example for you because we haven’t talked about direct vs. indirect testing. The Lakatosian hard core is not testable by definition, and yet it is subject to a kind of indirect test if the research program is not “progressive.” If such indirect testing is possible, it’s hard to know if any proposition is not testable. However that may be, I think indirect testing comes well shy of the proposition-by-proposition view I keep quoting form your earlier comment. So it’s not real clear what you might and might not consider a legitimate example. The ambiguity is increased because you speak of “attempted rebuttal” rather than “falsification.” Anyway, let me see if any of these pass muster with you:
1. “Man acts.”
2. The non-satiety axiom in neoclassical choice theory
3. Any of the axioms of Euclidean geometry
4. Uniformitarianism in geology
5. Einstein’s critique of “simultaneity” in his first paper on special relativity.
6. Arrow’s impossibility proof
7. Use of the envelope theorem to show that the minima of short-run average cost curves do not generally touch the long-run average cost curve.
Any of those work for you, Gus?
Robert,
I’m not sure which paper you mean. If you mean my paper on “computable entrepreneurship,” it shows up at Entrepreneur.com: http://www.entrepreneur.com/tradejournals/article/185611029_1.html
Hm, something’s up. One of my posts is not showing up. Robert, if you mean my paper on computable entrepreneurship, there is an ungated version here:
http://www.entrepreneur.com/tradejournals/article/185611029_1.html
Hope this one shows up properly!
Roger,
1. The term ‘acts’ needs definition before the statement can be evaluated for falsifiability.
3.The Euclidean axioms are not factual claims, they are definitions that, if accepted, can be used as premises for theorems.
6. Arrow’s impossibility theorem is not a factual claim. A theorem is a system of logical relationships. Whether the theorem is correct depends only upon consistency with definitions.
Robert J: “treating this statement as a truth claim is a complete mistake. It is a statement of definition, not a factual claim.”
If it’s not a truth claim, why should anyone abide by it or give it any heed?
Some here may find these posts of interest:
The Other Fields of Praxeology: War, Games, Voting… and Ethics? and Extreme Praxeology.
Stephan,
Really? You didn’t notice that the title of the post is “What is Science”? That’s a question of definition. Statements of definition are relevant.
The striking thing is that nobody ever explains what, for them, would be an acceptable criterion of demarcation for science. Some people want a description of the social institution of science, other search for a description of how scientists think about what they do, and yet others seek to describe what scientists actually do whether they know it or not.
The criterion of falsifiability straddles the line between this last kind of description and a proposal for conduct. Popper had an evolutionary theory of knowledge which describes how knowledge emerges objectively. In one sense, his criterion of falsifiability was a description of how knowledge develops, irrespective of the beliefs of scientists or social institutions. But in another sense it was a *proposal* for how scientists should conduct themselves, since it would more closely accord with the evolutionat way in which knowledge actually develops.
Science, and knowledge more broadly, is a spontaneous order, but it is also an evolutionary process. A feedback process selects among competing hypotheses and conjectures, and with our methods we should endevour to open ourselves more fully to that feedback, or in other words, “let our theories die in our stead.”
Oh, I forgot, probably the most popular standard which any criterion of demarcation should satisfy is this:
People who don’t like economics tend to conventiently fall upon a demarcation of science in which economics is unscientific, while people who do like economics tends to veer toward meaninglessly broad demarcations which include economics. The same pattern is then repeated for sociology, religion, psychology, etc.
Lee, you boil it down nicely, but I have to take issue with your statement that “People who don’t like economics tend to conventiently fall upon a demarcation of science in which economics is unscientific…”
I love economics!
Robert, I used the word “tend”! That way my claim was less falsifiable, see?
Lee, I know I know! You’ve got the angles covered.
I guess, really, I just hope you’re wrong that people who like economics are more likely to reject demarcation on falsifiability.
What’s the fascination with calling something ‘science’ anyway? As has been pointed out by most everyone here, there are a lot of endeavors out there that are useful and valuable, even if they don’t fit the Popperian definition of science.
Robert,
Do you really want to say that only “factual claims” count as “science”? Einstein’s thought experiments are great examples because they are brilliant bits of physics. Everyone acknowledges physics as “science.” But they are pure flights with no “empirical content.” Yes, yes, the theories of special and general relativity most definitely had empirical content. But not his critique of simultaneity. If your demarcation criterion says that this classical bit of physics is not science, then perhaps the problem is with your demarcation criterion.
Roger – you and I may end up with a little book out of all this.
You write: “I don’t recall any predictions about the fossil record in Origins. Maybe I just forgot, but I am moderately confident they are simply not there. And yet it was a very persuasive work of science.”
I think you are really dodging the point Roger. Not only did Darwin refer to fossils, his argument had definite empirical implications that could be tested – and were.
You also write: “… I’m with Percy Bridgeman who said something like “apply your mind to the problem, no holds barred.”
Good advice so long as mind is defined in the broadest sense – but that’s a whole new discussion. But that isn’t science.
Science involves persuading scientists, and that involves more than just applying your mind. It also involves giving others good reasons to take your explanation seriously instead of someone else’s. It is not enough to say you don’t like the reasoning or the conclusions – you need to show why. This is a very complex issue but it necessarily involves showing a shortcoming in the competitive explanation. Sometimes competing explanations will vie back and forth because there is never a knock out blow – and then, ultimately, there may be. I think the evidence is conclusive that among seriously contested theories the knock out blows are empirical.
You also write “As our conversation reveals, lots of stuff starts out looking non-testable and then ends up testable. I would have thought that fact alone would have been sufficient to show your strong remark was in fact too strong. But it seems not.”
Actually I think it proves my point, and that makes me think we may be arguing past one another. Something may not yet be testable, but we can envision testable circumstances – and then seek to make those circumstances real so we can test it. It’s what made it possible to come to a scientific conclusion about Einstein’s argument. Until then we had two brilliant theories and the brilliant men advocating them arguing without any means of directly putting their theories to a test.
If you have no problem with saying theories might compete in science but never win out so long as they cannot be tested, I’ll go with that. But until they can, science will more or less stand still in that area. It grows and knowledge expands when some competitors can be retired from the field.
You write regarding my challenge “The ambiguity is increased because you speak of ‘attempted rebuttal’ rather than ‘falsification.’
But Roger, “Attempted rebuttal” MEANS subjecting to a challenge it might fail. That means a test that might falsify it in the eyes of a competent observer.
You give me the following examples:
1. “Man acts.”
To me that’s a description, and not a very interesting one without additional observations (that can be falsified) or assumptions (that can lead to falsifiable conclusions). It’s a starting point – what can you do with it that’s interesting? It’s pretty boring as it stands. Any being with even a rudimentary knowledge of cause and effect acts. Dogs act.
Mise and Rothbard, to name two Austrians, made a lot of stated and unstated additional assumptions to generate their systems. They separated means from ends analytically and said nothing important in describing human action was lost by doing so – I disagree – and they left unexamined how “man learns” even though it is crucial to their analysis.
2. The non-satiety axiom in neoclassical choice theory
Not being an economist I must ask you to translate.
3. Any of the axioms of Euclidean geometry
Those are assumptions he used to do his proofs. Doing the proofs back when they had not been done was a scientific endeavor. They increased our fund of reliable knowledge. They could be disproven. Different assumptions yield different geometry. Applications reveal which geometries are more applicable to the world in which we find ourselves. We now know Euclidean geometry is a special case.
4. Uniformitarianism in geology
That is an assumption. It may be true, it may not. In geological science we use it to help develop theories WHICH MAY BE FALSE and are demonstrably so if key assumptions turn out to be mistaken. As it turns out many uniformitarian assumptions are proven false – as with what we now know about how continental plates move about – something uniformitarians long denied was possible.
5. Einstein’s critique of “simultaneity” in his first paper on special relativity.
Translate please.
6. Arrow’s impossibility proof
Translate please.
7. Use of the envelope theorem to show that the minima of short-run average cost curves do not generally touch the long-run average cost curve.
Translate please.
Based on what I understand with enough confidene to comment on, you are giving either definitions or assumptions that are then used to develop theoretical explanations that themselves are open to testing. That is what differentiates them from astrology, which I suspect NO scientist regards as scientific. Why is that so, Roger?. I know astrologers who think long and hard about their subject. Do you think astrology is a science? If not, why?
You seem wedded to the notion that science is concerned with Truth. I think that is an error. As people scientists are often so concerned – it’s what motivates them. It’s what motivates me. But science as a process never finds Truth, it only eliminates error.
Science is a journey that may never end, and we can contribute to further steps along that journey, or not. As knowledge grows what seems certain today becomes contestable tomorrow and perhaps is rebutted the day after. Perhaps you conflate the intentions of scientists with the achievements of science.
The growth of reliable knowledge DEMANDS there be a way to eliminate less reliable knowledge. Otherwise we just have undifferentiated growth of theories and explanations and you choose what seems nicest to you. It;s harder to give knock out blows in the social sciences – that’s why we still have Marxists.
Robert Johnson wrote:
“or example, some say that science concerns itself exclusively with falsifiable claims.”
But this is rubbish, and has been well known as rubbish for several decades. Philosophy of science has made some progress since Popper, you know!
Robert,
My comments seems to have gone lost in cyberspace. Here is what I tried to post earlier today:
Do you really want to say that only “factual claims” count as “science”? Einstein’s thought experiments are great examples because they are brilliant bits of physics. Everyone acknowledges physics as “science.” But they are pure flights with no “empirical content.” Yes, yes, the theories of special and general relativity most definitely had empirical content. But not his critique of simultaneity. If your demarcation criterion says that this classical bit of physics is not science, then perhaps the problem is with your demarcation criterion.
Roger,
OK, so you prefer a different definition for the term ‘science’. I’m alright with that. So how about my earlier question:
“…let’s say we embrace the broad Wissenschaft definition for science. After all, words just mean what we agree they mean. So then, what should we call more narrowly defined endeavors? Is there a special name for that vein of inquiry that concerns itself exclusively with non-falsifiable claims? Or is there a word that describes the difference between work that has a long history of generating knowledge that can be turned into useful technology, as opposed to work that doesn’t?”
Gene,
“Rubbish” is not an argument. Nor is a general and unsupported statement that Popper has been superseded.
In any case, the argument is about the definition of a term. What is the advantage of giving the term ‘science’ a more inclusive definition over a less inclusive one?
Does this all have to do with gaining a title of respectability? If the world agreed, say, that much of economics is not strictly science, would economics as a field of study be diminished?
Roger,
Thank you for the ungated link to your paper. Looks like some interesting reading that is likely to teach me a few things!
Robert,
I think there is a difference between “science” and “wissenschaft,” however we label it. Sure. I’m just pretty chary of demarcation criteria, especially simple ones. If science is a spontaneous order, the we should social scientists about what distinguishes “science” from other activities. There may be no one thing or no particular cluster of the things. The traffic is crazier in Rome than Copenhagen, but I would not attribute that real difference to the “hot blooded” Italian character or Danish “rigidity.” Things are just lots more textured than that. I think it’s the same with demarcation criteria. It’s not that there is nothing to distinguish, but simple (simplistic) criteria might all be wrong. I think they are all wrong.
Gus,
I, um, . . . Wow, I . . . I mean Einstein is . . . Not even Arrow’s impossibility result? Okay, well, um, here’s the Wiki link:
http://en.wikipedia.org/wiki/Arrow's_impossibility_theorem
Anyway, you seem to say now that some propositions can be science if they are not testable, as long as they are somehow linked to propositions that are testable. But that’s not what you said before. You’re moving the goal posts.
Roger-
It is very very dirty pool to give a long list of technical issues in different fields and expect someone to answer them all. Even when I have heard of them I want to know something about them in detail before I respond. That you have not answered my responses, and prefer belittlement, means we are reaching diminishing returns.
Given your handling of Darwin I have little confidence that the ones I wont take the time to look up are any better than the ones I know about.
And I am not moving the goal posts. I will answer that laimin detail if another reader thinks I am and can give an example – otherwise, I’ve made my case and people can make of it what they will.
gus
Lee, if you’re still following all of this, have you got a blog or webpage?
Everyone else, I’ve enjoyed the discussion! I have learned a few things in the past 24 hours or so.
Parting shots:
Historians, mathematicians, musicians, etc. don’t tend to refer to themselves as scientists, nor describe their respective fields as the Science of ______. They might throw in an ‘-ology’ here or there, but usually that’s all. Why is it different for economists, students of politics, sociologists, etc? Why is it important to lay claim to the title ‘science’?
On the other hand, if economists, for example, wish to be called scientists, why do they tend to neglect the use of the idea of falsifiability to further their study? Is it because this approach is difficult or fruitless when applied to economics? I’m asking honestly, since I’m not an economist.
Finally, do economists, sociologists, etc. admit that while physics has changed the world by generating incredibly powerful, useful, and irrefutable knowledge, economics, etc. has not produced comparable achievement? Do economists recognize that relatively little practical technology has been developed using economic insights?
Now, I don’t blame economics for that. And I only half blame economists. The other half of the blame lies with the fact that economists are not nearly as free to experiment (and harvest the resulting data) that physicists are. Face it, economists are data poor when compared against the problems they wish to solve.
However, I DO blame economists for endlessly arguing over non-falsifiable claims! And I DO blame economists for publishing ‘studies’ that defy repeatability! I DO blame economists for claiming that it’s alright to be sloppy, biased, and deliberately arcane!
One last thing. I never directly answered Roger’s question about whether I really want to limit science to only factual claims. My answer is yes. It’s debatable because it’s just a question of definition, but the way I see it Science is exclusively about processes that take place in the real world. There is no Science of Middle Earth precisely because Middle Earth is exterior to the real world. So when Einstein conducted thought experiments that abstracted from reality, he was, in my opinion, doing something other than science. Maybe it was philosophy, maybe it was mathematics clothed in imagery. HOWEVER, when Einstein did science, he demonstrably understood the nature and significance of the falsifiable claim.
Robert Johnson wrote: “Is there a special name for that vein of inquiry that concerns itself exclusively with non-falsifiable claims?”
Yes, Robert, that vein of inquiry is called “nothing,” because there is no such vein. It has been, what, fifty years since Kuhn demonstrated that Newton’s three laws of motion are not all falsifiable together — you can falsify any one of them, but only by assuming the other two. It has been perhaps sixty years since Polanyi demonstrated that crystallography is an a priori science. It has been fifty or so years since Feyerabend illustrated how all scientific theories are “born falsified.” It was in the seventies and eighties that Grünbaum and Salmon demonstrated that Popper could not even defend science as rational per his own criteria.
There is no shame in not being aware of these results — not everyone can or even should be expected to keep current with the findings in the history and philosophy of science. But, if one has not done so, one certainly can and should be called to task for pontificating to others about the philosophy of science!
Gene,
Newton’s axioms of motion are statements of definition…you know it’s not even worth going there because you haven’t offered an argument, you’ve just hidden behind appeals to authority. Why don’t YOU make the argument?
Yes, Gus, I confess I was vexed by your “Translate please” requests. But perhaps you should recall that it was you who challenged me to come up with even one scientific *proposition* that was not falsifiable.
And, Gus, the thing is that I was playing completely fair. The Einstein thing is not obscure. I don’t particularly expect you to be able to reproduce the argument. But did expect you would know that I referred to one of his thought experiments, that those thought experiments challenged basic Newtonian notions such as simultaneity, and that they were all arm-chair stuff and yet high points in the history of physics. That is totally fair and I was genuinely surprised when you seemed to blank out on it.
You challenged me by saying,
—–
Perhaps you can name me a scientific proposition that cannot be subjected to an attempted rebuttal by looking at some possible observable event. I’d be curious as to what it might be.
——-
I said right off, “No fair” in part because I knew that your request would drag us into lengthy exchanges over any proposition I might offer, and that’s just where your reply pointed. Yes, that got my goat.
Robert,
It seems to me that if you go that way, excluding Einstein’s thought experiments from science, then you are stipulating a definition of science that is inconsistent with common usage and excludes from “science” some of the best of what we have always thought of as science. There’s no logical error in doing so as far as I can see, but it would seem to imply that you’re imposing a requirement independent of actual scientific practice. If science is so wonderful, as I believe, then we should know what made it so in fact before imposing restrictions on what will count as “science.” IMHO anyway.
Roger,
I agree on all points! And I think it’s actually a really good point that we don’t want to marginalize productive methods of inquiry. Honestly I have some more thinking to do about the question, ‘What is Science’. I’m not sure that I’m entirely satisfied with the definition I’ve argued for here.
And in fairness to Gene, I should get more familiar with the work that’s already been done on this question by people more competent than myself. But, that is why I’ve engaged on the topic with all of you. You’ve given me reasons to think harder and more deeply, and suggested places to start reading.
Thank you!
By “proposition” – if I used that word – I meant theory/finding/argument. Something scientists regarded as a contribution to human knowledge. I thought it was obvious in the context of the argument what I meant. I now think I made a tactical error by not continually referring back to science as a self-organizing, or complex adaptive, system, as I initially did. When we look at it from this perspective we see that means of falsification are as necessary in science as means for a product to be abandoned are necessary in market economics. Both are essential features of the process.
Scientists, like entrepreneurs, use a wide variety of methods to arrive at theories, and any theory of rationality and mind needs to be able to account for them. Many important insights arrived as dreams – like the periodic table.
But considered as a social process that expands reliable knowledge, science is something both bigger and narrower. Bigger because it seeks consensus about the nature of the world, narrower because in seeking that consensus it needs to focus on arguments and methods other reasonable people find persuasive. This exerts pressure to the most clear and universal standards: measurement, prediction, experiment being the three major ones because pretty much any rational person would be impressed by them.
Since agreement is a central part of the goal and since many theories compete, this is necessary. Otherwise it would be like the market where no business ever went out of business, where there was no effective competition or like a theocracy where the dominant ‘product’ was protected against competition.
So from my perspective the bottom line is what other scientists find persuasive, not what philosophers say. Consequently “demarcation” as a philosophical issue does not interest me – but science considered as a complex adaptive system that grew from people seeking agreement about the nature of reality interests me a great deal.
All spontaneous orders function on both positive and negative feedback- some things are amplified, some are dampened. Like the demise of Edsel cars and tape recorders in the market, the death of theories is a feature of the system. That means they must be able to be abandoned for good reasons, or else science is unable to expand reliable knowledge.
What are good reasons to abandon a theory? The reasons are complex when we look at case studies because data often do not speak for themselves, and many theories initially gain in interest because they are beautiful or interesting or suggest new avenues of research. But in competing with existing theories I think the record shows if they cannot be subjected to tests and challenges that reasonable people in the field regard as telling, the theory never triumphs over its competition.
We get instead parallel worlds of study. This is a big problem in the social sciences, as is pretty apparent in economics, and from what I gather, not unknown in the natural sciences. When parallel worlds exist in a science, I suspect that is when more purely political and organizational factors dominate impersonal standards of evaluation. Different departments hire their own supporters, schools of thought emerge based in particular schools, and the like. This is common in the social sciences not because they are not science, but because definitive refutations of a theory are so hard to get. It is why I was so excited about the democratic peace phenomena – because no other theory dominant in political science could either predict it or explain it very well.
When refutations cannot reach near universal agreement the field fractures or stagnates. The physical sciences have less of that – not none but a lot less – because they can more easily rebut bad or flawed theories.
This is not a Popperian model. It is a Polanyian one that acknowledges there is no purely objective knowledge, at least nothing that is particularly interesting scientifically, that as scientists we seek to minimize the personal component, and that the model of spontaneous order captures this better than any other. It also acknowledges that individual scientists and research organizations and schools of thought do not necessarily harmonize their interests with science as a process of disovery – and that explains why standards of testing are so important.
I think this approach recognizes the difficulty of what constitutes falsification in many social science fields while keeping it as a central feature of science as such. I think it also acknowledges at least some social science as science – without watering the word down in the eyes of physicists.
Roger – Regarding Einstein, I already covered his thought experiments in an argument you ignored. Adding another example does not change my point. His thought experiments did not solve problems in the eyes of other scientists – they led to conjectures/theories/insights that could be TESTED. When they passed the tests then people bought in, especially in revealing unexpected results, then his theories caught on. If they could never be tested or suggest new observations, given his eminence they would probably be remembered – but they would not have revolutionized physics.
Robert Johnson – I did not answer your question about complex adaptive systems because someone else gave a reference and I was focused on Roger’s criticisms. My apologies for seeming to have ignored you. Two initial places to look are Albert-Laszlo Barabasi “Linked,” (Plume) and Yochai Benkler, “The Wealth of Networks,” (Yale).
g.
The conversation over “testing” and “falsifiability” really ended with Kuhn and Hayek. Popper HIMSELF directly conceded the failure of his program in a passage in his memoir were Popper concedes that Hayek is right and he is wrong about the scientific status of complex sciences like Darwinian biology, where the Popperian “falsifiability” criterion utterly fails. Popper had long insisted that Darwinian biology wasn’t science — until Hayek persuaded Popper that the thing to be rejected wasn’t biology, but Popper’s own demarcation principle.
If you look at Kuhn & Hayek & Quine & Wittgenstein, what you’ll see is that all of our ideas and pictures of the world are deeply intertwined, and “interpretations” become simply perceptions, and “theories” become part of our background understanding, i.e. the “empirical” becomes “a prior”, this happens at a brain level (Hayek, Edelman), and it is build into the logical structure of language (Wittgenstein, Quine), and it is part of how science becomes part of our understanding (Kuhn).
What Greg said.
I think Greg was referring in part to Hayek’s argument in The Sensory Order that science tends toward an apriori system. Anyway, I’m sure he’s say that’s something could have mentioned to further boost his point.
Greg,
The story you tell about Popper and Hayek doesn’t make any sense, because Popper regularly stressed that just because something wasn’t falsifiable, didn’t mean it couldn’t be true, valuable, interesting, or useful. So when Popper said that something is unscientific, that does not imply that he is rejecting it. On the contrary, he was among the original evolutionary epistemologists and proposed that all forms of knowledge are an extension of biological evolution. By *categorising* evolutionary theory as unscientific, Popper was merely acknowledging that it couldn’t be contradicted by observation. In short, we can never observe a “less fit” organism surviving in the place of a “more fit” organism, since whatever survives is by definition “more fit.” Particular stories of evolution, however, can be falsified, such as the present story of human evolution, and I don’t believe Popper ever intended to suggest that this wasn’t so.
Greg – Popper said a lot of things weren’t falsifiable that were in fact falsifiable. Evolutionary theory IS falsifiable – but does not give predictive power about the future. It cannot be tested in that way.
Falsifiability does not equal successful predictions about the future. Those are two different issues. Another way of putting it – evolutionary theory has an impressive series of successful predictions about future finds in the fossil record.
If the fossil record does not support Darwin, evolution will have very serious problems. His theory predicts that it will – and it has. Does anyone in their right mind believe differently today?
When we get to processes that do not give us much in the way of relativity impersonal measurements it gets much harder.
But the context in which they act shapes the kinds of actions they can take. Here is where I think spontaneous order in society comes in. It shapes the context impersonally compared to what ultimately motivates individuals. I have to pay attention to price data if I want to at for long in the market, even if I am uninterested in wealth.
China began adopting market approaches in place of bureaucratic decision making and it took off. That’s a more controversial argument than the fossil record, but, it follows from those who argued IF China does A then B will happen. And it did. Sounds a lot like a prediction to me.
Polanyi argued that the success or failure of a theory depended on the judgment of qualified practitioners who understood the field. They combined the evidence before them with the experience of having worked within it. I think Kuhn and many who came after him fit neatly into a Polanyian viewpoint that targets the community of evaluation as crucial. They can be wrong, but it’s all we’ve got.
If I remember, Hayek himself said the test for theories of complex phenomena was that they consistently outperformed the alternatives. They could not predict future events, but they predicted future patterns of events. That is a much softer prediction than what we see in physics, but it gives us a standard which, if not met, can lead to serious rethinking, modification, or abandonment of the theory.
In other words, Greg, Popper wouldn’t have been forced into a dilemma between rejecting falsification and rejecting biological evolution, because the criterion of falsification was never intended as a rule by which to reject theories — but merely categorise them.
Perhaps Popper did face this conundrum at some point (because people’s views change over time), but it is not evident from my reading that he ever faced it, because even from his early work on falsification is was never a criterion by which to reject something entirely. In his later life, I recall him referring to it as a “rule of thumb” — falsifiability was a desirable but not necessary quality.
Lee, you’re a voice of reason. Got a blog?
Gus, thank you for the pointers! I’ve added them to my list.
Gus,
There is some semantic confusion here.
I think, when Popper said that evolutionary theory was unfalsifiable, he was referring to the abstract theory itself. In short, we cannot observe a less fit something surviving in the place of a more fit something, since whatever survives is more fit by definition.
Popper described evolutionary theory as part of a metaphysical research program. Darwinian biology is an application of evolutionary theory which is falsifiable, since, with the right auxiliary assumptions, it makes definite predictions about the world which we can test it by. However, if ever a particular story of evolution is falsified, there is always another evolutionary story that can take its place, and in this sense evolutionary theory itself — stripped of specifics — is unfalsifiable.
Robert,
Sorry, I don’t have a blog. I should state one though and stop intruding in the comments of others.
Lee, yeah you should! But don’t stop commenting either.
The following link is nonsense, but I had to post it any way:
Thank you Lee. I stand corrected regarding that distinction. I haven’t read Popper in years and years.
From what you say, as soon as evolutionary theory is applied to the physical world, it becomes falsifiable – yes? Only when it exists so abstracted from the physical world that no empirical question can arise to confront it does it become unscientific.
If I understand you, I agree.
Scattered thoughts:
-When asking “what is science?”, I think we’re really asking, “what methods of acquiring knowledge should we regard as respectable?” It’s not enough just to be “systematic” or to “give it your all”. There have to be constraints that tell you when you’re just groping in the dark, or playing echo-chamber, or re-labeling ignorance as knowledge.
-The falsifiability criterion was on the right track, and its shortcomings are corrected by Bayesian inference. See Yudkowsky’s “A Technical Explanation of Technical Explanation”. This resolves the incessant complaints about falsibility and Quine’s holism argument: By building a belief network (graph of conditional dependence between beliefs), you connect theories to observations. Applying Bayesian updating, you can see whether new observations require you to believe you made an observation error, your parameters are wrong, or the theory needs to be scapped. It also shows how untestable propositions can nonetheless be falsifiable by virtue of their logical dependence on other observables. (e.g. why you can believe something still exists even when it leaves your light cone)
-I agree strongly with this comment from Gus_diZerega. For example he’s right that the seemingly self-evident truths actually carry subtle empirical assumptions that are necessary any time you try to actually use them. And he made the point in enough of the examples that moving to the topics he wasn’t familiar with would have been pointless.
-Crystallography is most certainly *not* an a priori science like Austrian economics claims to be. It makes inferences based on the results of experiments. Yes, as any information theorist will tell you, “you have to make assumptions to make inferences”, and crystallography is and was predicated on assumptions. But this is different from saying that you can “a priori” deduce those assumptions such that you know they’ll never change. If the assumptions crystallography went in with were no good, they would eventually have to modify them, with good reason (see point about bayesian reasoning above).
(Perhaps Polanyi’s argument wasn’t very good, which is why it isn’t more widely known and cited as an example of how such an obviously empirical science isn’t.)
-It’s a bit disturbing how the Austrians here, *when defending economics*, basically revert to “what really matters is if you try your best”, “who knows what’s really science anyway?”, “people just call something not-science because they don’t like it”. Sheesh.
Silas Barta’s is the 63rd comment on my post. It’s hard to keep track at this point, but I don’t recall anyone saying anything like what Barta attributes to “Austrians” in his final paragraph. In my quote of Machlup, he was citing the Encyclopedia Brittanica on the meaning of science. And the requirement was for “impartial and systematic” knowledge. As a number of comments have noted, the “impartial” part is crucial. That requirement helps explain why science is usually social, requires evidence of some kind, and so on.
@Jerry_O’Driscoll: For my charge of “what really matters is you try your best, see Roger_Koppl’s comment (11/5/09-9:27am)on what he considers science: “I’m with Percy Bridgeman who said something like “apply your mind to the problem, no holds barred.” ”
As a number of comments have noted, the “impartial” part is crucial.
Well, you didn’t seem to think it was crucial when you said: “I endorse Machlup’s broad definition of science as any systematic study of a subject.”
But regardless, that’s a question-begging criterion. The problem is to identify *which methods* reliably filter out partiality. After all, astrologers could just as easily say that they’re impartially deriving a priori truths about the relationship between stars and human action.
But then, once you start coming up with criteria for what makes something impartial, you start coming up with things that sound remarkably close to the “falsifiability” standard, because, without any attachment to some objective test, anyone is free to claim impartiality, and sufficiency of their evidence.
My original post was clear and the first comment seized on the “impartial” aspect. So, please. You link “impartial” with “falsifiability.” Talk about question-begging. I go back to Micahel Polanyi’s Personal Knowledge on the difficulty in even the hard sciences — even in mathematics — of getting agreement among scientists on what constitutes evidence or proof. In practice, what one does in economics is counter complex theories with equally complex theories (econometrics).
My original post was clear and the first comment seized on the “impartial” aspect.
No, it wasn’t. You clipped off the “impartial” part, but regardless, I explained why that assumes away the problem because you need to define a process that will *prevent* partiality.
You link “impartial” with “falsifiability.” Talk about question-begging.
??? Do you even know what question-begging *means*? It’s when you assume your conclusion. It’s what I didn’t do. What I said was, if you want to ensure impartiality, you will have to rely on some objective standard, which in practice means something like falsifiability because it’s harder to sneak your partisan views into objective observations than in arguments.
I go back to Micahel Polanyi’s Personal Knowledge on the difficulty in even the hard sciences — even in mathematics — of getting agreement among scientists on what constitutes evidence or proof.
Well, now you do. But in any case, that’s evidence of the poor epistemological training of scientist and Polanyi’s having written before the Bayesian revival in the sciences.
Are you capable of disagreeing without being disagreeable?
Jerry, ad hominem?
No, Jerry, unfortunately Silas is not. This fact is well known to the posters at many, many other web sites.
Silas, there is no Bayesian “revival” in science. There is one amongst people who wish to reduce science to a mechanical procedure.
Robert, are you aware that appeals to authority are perfectly valid when the authorities in question are, in fact, true authorities? (That doesn’t mean they are conclusive, of course.) I am pointing you to literature showing you why this “falsifiability” criterion is a dead letter. I am certainly not going to make up for the gaps in your education by rehearsing the entire literature for you in the comment thread of a blog!
Lee, so Greg mentions what actually happened between Popper and Hayek, and you respond “it doesn’t make any sense”?! So this is history falsifiable by the criterion, “Lee can’t quite believe this,” hey? Or, on the other hand, perhaps Popper understood the issue a little better than you do?
Gus wrote: “His thought experiments did not solve problems in the eyes of other scientists – they led to conjectures/theories/insights that could be TESTED. When they passed the tests then people bought in, especially in revealing unexpected results, then his theories caught on.”
This is false history of science. The theories of relativity were widely accepted before they had been tested at all, and, as Polanyi points out, there were many, many contrary results in the 1920s, all of which were broadly ignored. (As was proper — the theories were too good to let a few hundred ‘falsifications’ change anyone’s mind!)
Gene,
The problems with appeals to authority are, 1) as you point out, not everyone may be familiar with the work of the authority, 2) the ‘authority’ may actually not be one (see Silas’ comments on crystallography), and 3) it’s a substitute for actually making an argument.
It’s easy, and pointless, to simply say ‘other people have shown you’re wrong’. But if you present an argument then we can discuss it’s merits and flaws.
For example, you bring up Kuhn’s analysis of Newtons Axioms of Motion. I counter that this doesn’t defeat my argument at all because Newton’s Axiom’s are not statements of fact, they are statements of definition.
See, that’s how discussion works. If you have a position, just explain it! Then we can talk about it.
Gene,
Based on your comments about the acceptance of Relativity before it had been well substantiated, I think you and I are likely arguing about different topics (another reason why relying only on appeals to authority can be problematic – by not clearly stating your own position you make it difficult to be sure that we are actually talking to the same points).
Frankly, I’m not terribly interested in how an hypothesis gains popular appeal. I’m not even arguing about what’s true and what isn’t. To me, what is science and what isn’t depends on the process.
So, again, what is YOUR argument? The title of the post is ‘What is Science’. Do you have a proposed answer?
Yes, Robert, I agree, Silas is certainly not an authority on crystallography! (Whereas Polanyi, a renowned physical chemist, was.)
As far as your request that I review the last 50 years of the philosophy of science for you, well… no thanks. As your evasion of my Newtonian example shows, that would be pointless anyway. I am only making these posts to indicate to others who may be reading that the current state of the philosophy of science is nothing like what all of this talk of falsification would indicate it to be, and certainly not to try to convince you of anything!
Gene,
That’s kind of my point. Instead of telling anyone what YOUR position is, you just keep referring to authority.
You might notice that no one else in the discussion has been taking that approach. Most everyone else has been arguing for a position, with occasional references to specific works as support. We’ve been enjoying the give and take, and the chance to engage with opposing arguments.
I understand that kind of discussion isn’t for everyone.
I’ll have to revisit my reading on the history of the theory, Gene. As I remember the conclusive event was the Michelson-Morley experiment. Until then, as I said I think, his thought experiments had proved very interesting/suggestive – but had not won widespread acceptance. But I read all this stuff, incl Polanyi, many years ago My memory is sometimes unreliable.
But it may be nuance. Part of my argument – in citing Polanyi – was that judgments as to what counts and what doesn’t is the final decision of the community involved. That no measurement is completely objective because as he said, they all depend on personal interpetation of the meaning of the measurements. I’ll check Polanyi again.
Gus
Gene,
A lot of modern philosophy of science doesn’t interest me; they aren’t trying to solve the problems that drew me to the subject in the first place. The only people who did were Popper and his students. The majority of criticism of the criterion of falsification is irrelevent to me (like Quine’s wholism or Feyerabend’s unmethdology), because they accuse the falsifiability criterion of failing to solve “problems” that I never expected any criterion of science to solve. In other words, I rarely can fault their logic, but disagree it is a problem we should be trying to solve. The underlying difference and root of the confusion, in my opinion, is nonjustificationism — different problems and goals.
In any case, Popper affirmed on many occassions that unfalsifiable hypotheses were not necessarily false, useless, or uninteresting; he also developed a theory of metaphysical research programs. The idea that Popper would be forced into a dilemma between rejecting his criterion of falsifiability or evolutionary theory doesn’t make sense, because it is evident from Popper’s other writings that he was comfortable with unfalsifiable hypotheses. Of course, perhaps in the development of Popper’s ideas there was a time when he faced that dilemma, but it would seem to have been a temporary misstep rather than representative of general position. I would certainly be curious to see Greg’s source, because, given everything else Popper wrote, it just doesn’t make any sense!
Gus, the Michelson-Morley predated the Special Theory by some 18 years. Although the “Whig history of science” has tried to portray Einstein’s theory as a response to the experiment, he himself said the experiment played almost no role in his thought at the time. And, in the 20s, the “aether wind” was measured repeatedly, and the results were simply (and rightly!) ignored.
Here:
‘When asked by his assistant what his reaction would have been if general relativity had not been confirmed by Eddington and Dyson in 1919, Einstein famously made the quip: “Then I would feel sorry for the dear Lord. The theory is correct anyway.”‘
As this article mentions, the first precision tests of relativity occurred in the 50s and 60s, by which time the theory was very well accepted already.
Thanks for the link, Gene. I am currently without access to a university library and with only a fraction of the books I once had, so I appreciate it. Plus I’m really rusty on the history.
That said, the links you gave don’t seem to me to say anything different than what I suggested in the quote you criticized.
Einstein’s General Theory initially solved a problem – the precession of the perihelion of Mercury, and at a theoretical level unified Newtonian theory with his Special Theory of relativity. So – one empirical problem solved and two dissimilar theories harmonized, if the link’s account is accurate. That’s grounds for the scientific community to take it seriously especially given Einstein’s reputation in the field. He presented it in 1915, around 10 years after he had published his first path-breaking work.
Wikipedia says “The perihelion precession of Mercury was the first evidence that general relativity is correct. Sir Arthur Stanley Eddington’s 1919 expedition in which he confirmed Einstein’s prediction for the deflection of light by the Sun helped to cement the status of general relativity as a likely true theory.” (http://en.wikipedia.org/wiki/History_of_general_relativity)
Wikipedia might be wrong, but I doubt it would be on this kind of thing.
Notice the word “likely.” He was taken seriously. Some were convinced, some weren’t. Most were interested. Additional tests cemented the theory as a very important one.
Einstein gave three tests that his theory would provide unexpected results in answering –
1. The perihelion precession of Mercury’s orbit
2. The deflection of light by the sun
3. Gravitational redshift of light
It passed all the tests well enough in the judgment of the physics community to be increasingly accepted – and as their ability to measure and test has gotten better, it continues to do so.
If it had not passed these tests in the opinion of physicists, it would not be accepted as reliable knowledge for trying to study the world.
Gene, this seems in harmony with what I wrote. He arrived at his theories through thought experiments. He had enormous confidence in them – hence the quip you cite. But he could only convince the wider physics community by proposing ways the theory made a difference in what they studied. Ad that is what I think is crucial about science as a spontaneous order – it can not be analyzed simply by looking at individuals. The context in which they work must be treated at least as seriously.
And I have always emphasized that my focus is science as a community, a system of relationships. Not as individual scientists.
Maybe that is the underlying issue.
Gus
I think it was Buckminster Fuller who said “You uncover what is when you get rid of what isn’t.”