Scientists mostly don't know what a "theory" is, but should they?
In a thought-provoking rant, the Effect Measure science blog challenges the idea that science students should (or can!) know what a "fact," a "theory" and a "model" are. These are not simple, settled concepts, but rather areas of hotly contested debate.
Most scientists also can't properly say what distinguishes science from pseudoscience, say what a fact is, give a satisfying rendition of what a theory is, etc. The problem is that Mr. Williams can't do it either, because there is no agreement on these knotty matters. He seems to think these are simple and settled questions. But finding adequate criteria that separate science from pseudoscience, the so-called Demarcation Problem, remains an unsettled question in the philosophy of science. It is so difficult, and possibly so fruitless, that many philosophers have ceased to be concerned with it. Similarly, what is a "fact"?...A rant about science educatorsA wag once commented (and I have quoted here often) that to expect a scientist to understand the philosophy of science is like expecting a fish to understand hydrodynamics. I guess the same thing goes for science educators, although it is less excusable. Mr. Williams seems to be under the impression that these extremely difficult foundational issues are settled and should be common knowledge for all scientists.
the latest
latest episodes
But.. but... How can this be? Richard Dawkins knows with absolute certainty. If there is no Law the sky will open and all will be lost.
I'm doing a an undergrad science degree at the moment and we've had plenty of opportunities to study history and philosophy of science. We were lucky to have a few very clever lecturers who spoke about demarcation, underdetermination, logical positivism and all sorts of similar topics.
I always found it very interesting, but a lot of my cohort don't appear to. I think it's often very difficult to discuss without appearing particularly abstract. Discussing the role of inductive "logic" in empirical research and trying to make heads or tails of whether this is authentically logical is probably a little airy for some people.
With regards to the demarcation problem in particular though - it's certainly a thorny one. I personally like the idea that science is at least partially socially constructed (not as much so as people like Feyerabend would suggest, but more in the manner of say Kuhn or Thagard).
Lastly, i'm incredibly surprised to see this here - it really does seem like a topic that's a little bit ignored.
Thanks BoingBoing for reminding me why I read this site :)
(and pardon the wall of text comment)
shades of "lies-to-children" - where to begin?
I suppose the problem lies in starting out in the confidence you will "get" somewhere - as opposed to understanding the the structure you are building must inevitably become fuel for the next enterprise.
It would be enough if they could just separate observation from conjecture.
One of the best course I took in college was a Philosophy of Science course.
But, I don't agree that distinguishing science from pseudo science is difficult.
Science involves creating a *falsifiable* hypothesis, testing it, and publishing the results in peer reviewed journal.
Most pseudo science fails at the first hurdle.
This is fascinating. I know quite a few people in the sciences, and I've always found that the smartest ones, the ones I respect the most, are always debating this sort of thing among themselves. By contrast, the ones who are most sure of their own ideas, who see their work as flawless and beyond criticism, tend to be the ones who only think in proscribed ways and are often critical of my arts background.
It would be delightful to see Pierre Bourdieu's field of cultural production taught in a science class. Maybe even bring Derridian post-structural analysis to bear on these sorts of problems. Sure, it might not lead anywhere, and it might just be a waste of time, but what if it weren't?
It would be nice to see the arts and the sciences talking to one another more often. The best and brightest of each side tend to have a great deal of interest in what the others are doing, I find, but there's still a large segment of the population who tends to pick one camp and stick with it. It would be wonderful if we could all move beyond that.
Pseudo-science seems aimed at the gullible. Usually for reasons of profit.
Science - unfortunately not always pure - is usually aimed at truth for its own sake.
(At least observable truths. ;)
i's true that when this science thing came along, quite a few of the brethren just shifted over a seat from the priesthood. Just a different god to worship, same uncritical mindset.
What troubles me more is the inability of the media (and hence the general public) to understand the difference between a hypothesis and a theory.
But yes, I've been on the front lines and observed those studying to become primary and secondary school science teachers. All I can say is, be afraid, be very afraid. And this in a state that has a reputation for good public education. I shudder to think what goes in states less education and/or science friendly.
And a bit of philosophy is useful, BUT, have you ever sat in on a faculty committee that includes a member of a philosophy department? Oh gawd, just shoot me. It's a guaran-damn-tee that the committee will never accomplish anything.
This ties directly into a lovely essay from The Journal of Cell Science that I read this morning, titled "The Importance of Stupidity in Scientific Research" read it here....
http://jcs.biologists.org/cgi/content/full/121/11/1771
Donopolis
If you can check it, it's science. If you can't, it's not.
If a lot of people have checked it and it seems like it works 99% of the time, it's a theory. If only few people have checked it, it's a hypothesis.
This is basic, grade school level science. How anyone misses it and can still make it to high school is beyond me.
The whole point of science rests on the concept of testable hypotheses. If a hypotheses can be tested empirically, either experimentally or through observatioal data, then it's science. The more extraordinary the claim made, the more rigorous the proof needed. Pseudoscience seeks to justify itself either through inadequate research or claims that the scientific process is flawed.
I apolosise in advance for this nit-picking, but "Scientists mostly don't know what a "theory" is, nor should they?" is not a question, so should not have a question mark. If it was "Scientists mostly don't know what a "theory" is - should they?" that would be fine. But just because something is asked with an interrogative tone, does not make it a question. I usually use a 'floating' q mark after a couple of stops - "nor should they..?" - to achieve a questioning tone.
Surely this has got to be the most annoying comment for quite some time..?
@9
Oh man, never let the philosophers talk about philosophy--that's what the sociologists and English professors are for.
I second Mark Levitt @5 - real science involves producing a falsifiable hypothesis, testing it repeatedly, and getting peer review and replication of the results of experiments.
Pseudoscience hijacks the seeming authority of science.
This contrasts with /merely bad/ science, wherein the overall work structure of science is followed but one - or a few - mistakes are made without the researcher understanding that a mistake was made.
A poor scientist will take into account criticism and improve his/her work on the next go-round.
A pseudoscientist has done no scientific work, or very little work that might appear scientific, and will not address, incorporate, or adapt to criticism. They will also often claim there is a social agenda to suppress their findings.
"To expect a scientist to understand the philosophy of science is like expecting a fish to understand a subject that has little to do with actually living like a fish." Ba doom schnick!
Try the veal -- I'll be here all week.
Gravity is not a simple, settled concept, but rather an area of hotly contested debate...
... but that doesn't mean it's unreasonable to ask a high school physics student to learn Newton's law.
Science is all about approximation, kids. If a simple definition of "theory" is enough to teach your average future voter the difference between the Theory of Evolution and the Theory of Intelligent Design, that's the one they need to be learning first. They can go to college and argue about the details later.
#7: Plenty of great science has been done in the name of profit. Bell Labs, geophysicists working at the oil companies...
@Jewels Vern, too right!
I mean a few other people here have mentioned being able to "check" something - but how well can you check anything, when we're unable to seperate out and test any individual theory on its own. Sometimes I suppose the line is cut and dried but often, it appears not.
I think a good example is actually Quantum Mechanics, where when you think about stuff like the Forman thesis, you can see that the rationale for picking one interpretation of the results over another is not always clear or even necessarily totally logical.
Let alone the "truth" of the matter :P
Well, I drive a car but cannot say how it works internally, should I?
I major in a basic science (chemistry) and yet had no idea there was any such debate on such issues (well science/pseudoscience I was aware of).
I was taught with absolute certainty what makes a theory. It was explained very simply and matter-of-factly as being a hypothesis which has held to be observably correct after rigorous empirical testing.
That I was unaware this was anything other than canonical fact certainly surprises me.
There are no facts - a "fact" is the best theory.
It's not even a "fact" that right now you are sitting at your computer reading this. How do you know that you're not in the Matrix, and that everything you're experiencing is real and not a simulation? You know it because that the world is real is the best explanation, the most parsimonious. Scientists find the best explanation by the mechanisms mentioned above - "producing a falsifiable hypothesis, testing it repeatedly, and getting peer review and replication of the results of experiments." Such mechanisms don't produce "facts," they produce things considered facts because they're good theories.
When creationists say that evolution is just a "theory," they're right. But it's a much better theory than creationism. Just because everything is a theory doesn't mean we have to teach the crazy ones.
Science teachers should be able to explain this.
My problem is that resigning these concepts off as debated ideas makes it very easy for people who don't appreciate the debate to make EXTRAORDINARILY stupid statements and missteps.
Taking the philosophyof economics (my major) was very enlightening for me, because I've learned to see that there are certainly grey areas in these definitions.... possibly even worth studying. However, in my experience, most of the debate around defining a model or fact would be unhelpful even if resolved.
I think instead, we should have loose definitions and be aware that there are multiple cases at the boundaries of these definitions.
As a parallel, let's take the concept of a chair. You might have just pictured something with foor legs, but is that required for a chair? Could an exceptionally large crate be a chair? A large mushroom? A person? It could go either way, but clarifying the definition of chair to include or exclude mushrooms does nothing to help people sit.
I think it's interesting stuff, but mostly pointless beyond a certain level of abstraction.
Sounds like the latest volley in the Science Wars.
I remember as a 19 year old taking my second philosophy class that it was a whole different type of thought and I wasn't getting it. The first PHL class was Logic 101 and that was fine for me; everything clicked as it related to everyday life and experience as a mechanical engineer. But my brain at that time was not ready to question the fact or "fact." I do wish there had been a more immersive program for the geeks like me back then to show the two sides of science; philosophical and I guess more empirical.
Evolution at best is a "Model" - It can not be a theory because it resides on past, unobserved ...
same as the model of creation.
Both are a matter of "Faith"
surprised?
In the field where I work, molecular biology, there is a sorry lack of facts and a whole lot of theories and hypothesis. And I have the feeling that most of my co-workers share the same understanding of this. And I think most students that finish their studies and go into science soon will get an understanding of the differences.
I also think that the "publics" conception of scientists as prone to cling to their theories, is wrong. It is every scientists wet dream to falsify a well-documented theory.
Maybe this misunderstanding between fact and theory is an artifact of the english language (as you might understand from my writing, I am not from an english-speaking country). In my language, Norwegian, the phrase "scientific fact" carries A LOT of weight, and is not something people tend to use in a discussion, not even a scientific one.
When I was a grad student (in physics) I had some discouraging interactions with philosophers-- I concluded that expecting philosophers to understand science is the result of a sort of misunderstanding.
I remember, in particular, a bunch of physicists went to a lecture by a Wittgensteinian on 'causality'. The thesis of the lecture was that sometimes an effect may precede a cause because there's a form of causal language that doesn't specify the temporal sequence of cause and effect.
Not to make fun of all that, (hah!) but-- one should bear in mind that scientists and philosophers will generally have different concerns.
...oh, and #24....stop being silly. The artifacts of evolution can be observed all over the DNA of all living organisms. It is a theory, based upon observed and measured data. Creationism is just a statement.
Dogma is based on the absurd assumption that human perception defines reality.
Science says "We don't know, but we can try to find out."
@5, 11, & 15:
The idea that science is distinguished by testability and/or falsifiability doesn't work. Astrophysics is an example: no experiments there, but it's definitely science. Of course it builds on other sciences, but it's mostly about theory and observation. If you expand the notion of testability to include astrophysics, then you'll include lots of other things that clearly aren't science. This is exactly what makes the demarcation problem so hard, and it's absolutely not trivial.
Corey
@24: That's fatuous. Evolutionary theory is robust, well-supported, and rife with successful predictions. Creation "theory" is not.
Evolutionary theory is about as much a matter of "faith" as computers, gravity, and believing the sun will rise tomorrow morning.
Thank you for playing. NEXT!
#27 - not at all. If we had a Large Sun Collider we'd be able to test/falsify Astrophysical theories no problem. Building the test apparatus is the problem, but conceiving of a viable test apparatus is not.
This sets it nicely apart from the astrologists and crystal healers of the world.
@29:
But you run into problems if you're allowed to appeal to conceivable, but not actual, testing procedures. What's to stop a crystal healer from saying that this field is also falsifiable, if only we had the Large Aura Collider? It can't just be that we "know" how to build a Large Sun Collider, but not a Large Aura Collider.
Science involves creating a *falsifiable* hypothesis
Ah yes, the lure of scientism is strong. God is after all, unconscious.
Both are a matter of "Faith"
No. Evolution is indeed a theory and ID is religious faith. Your solipsism is full of fail.
scientists and philosophers will generally have different concerns.
Exactly. Science is pragmatic and asks how questions. "How does electromagnetism work?" Philosophy asks why and what questions. "What do you mean by freedom?"
DONOPOLIS: It's a pity that the writer of that essay doesn't know the difference between stupidity and ignorance. Good scientists come to terms with their ignorance, and use it to find good research questions. But they are certainly not stupid.
The real problem alluded to in the original article is the state of science teaching in our schools. Here's a fabulous article on what's really wrong with it, which does a nice job of sorting out the mess:
http://coehp.uark.edu/pase/TheMythsOfScience.pdf
Seems like the quantum folks can agree on the repeatable experimental data but not on the implications or meaning of same. And what about the role of the observer? The materialist mindset is dying. The committee can't seem to agree once you get to the bleeding edge. Its the best system we've got but there's a bunch of phenomena that the current models fail to explain convincingly.
You know who else wasn't satisfied with current theories?
Anyone whoever made a breakthrough.
I love this topic, but have no formal education, so let me throw out my understanding gathered from my OJT. Please be kind to my mistakes; I am taking this chance to further my own understanding. I don’t get much opportunity to discuss these ideas since retiring.
Facts are observable events. Facts are repeatable observations. Accepted facts, are the same for all observers. Facts change in as much as our understanding of our observations or our observational skills improve. So, yes it is a fact that I am sitting here, but does this mean I am anything more than a thought? Does this define reality? No. It only means the vast majority of us agree with the observation.
Facts are not truth; truth is absolute and rare. The only universal truth, imho, is we don’t know the truth. Facts and truth are not self-evident. With continued and refined observation, we may accept facts as truth, but this often proves dangerous. Facts are fluid; truth is not. A model is our current understanding of a system based on theory, supported by observations, and which makes predictions.
A hypothesis is a working theory based on facts. Generally, everyone agrees on the facts, but does not agree on our interpretation of the facts or on the accuracy of our observations. A hypothesis is a framework built of facts. This framework attempts to order facts, explain the relationships of facts, and further our understanding of the universe. As we further our understanding of current observations, observe new facts, or our observational skills improve, our understanding changes and the hypothesis is refuted and revised, sometimes discarded.
Theories are fact based, refutable, and verifiable understandings of vetted facts. Theories produce predictions that confirm the theory. The Theory of Special Relativity predicts time dilation; time dilation is confirmed by observation and must be accounted for in the operation of the Global Positioning System.
Theories may change to accommodate facts, but new facts will not require us discard the theory, as we might discard a hypothesis. My understanding is that if new facts cause us to discard a theory, then our theory was a hypothesis.
Cory, I'm a little surprised you posted this. While it's not a bad idea to reexamine old ideas periodically, in this case the writer merely seems to deny centuries of previous work on the philosophy of science, rather than adding anything to it.
The public may not know what science is and how it works, but scientists understand the core of their methodology quite well.
Understanding methodology is not the same as understanding these concepts.
yay! Dictionaries all around!
The concept of theory is itself a theory - it's a scientific model which stands up to experimental evidence (thus far). The model doesn't have to be infallible to be useable. However it is necessary to understand the limitations of the model one is using, in order to transcend it. Now I think I'll go make some fake snot.
What, so now we're expecting our scientists to be philosophers? They don't need to know any of this theoretical bullshit to correctly deduce facts and create accurate models.
I can't help thinking that the writer of the article protests too much. He doesn't present any evidence that Williams has a simplistic notion of the definitions of terms such as theory, fact, etc. And sure, there are problems defining anything, but what Williams seems to be pointing out is that many students aren't even in the ballpark in attempting to define these terms. Since scientists are going to be dealing with things like the "theory" of evolution you'd hope they'd have more sophistication than to equate the term "theory" with the popular understanding of the word as "unproven ideas". Idiot creationists make that mistake -- science students shouldn't.
This is something that continually frustrates me. Damn near nobody knows what the word "theory" means in the context of science. Most people think it means some kind of guess; it's actually just the opposite. The vernacular use of "theory" is almost exactly the inverse of its scientific meaning.
Here's the basic definitions behind the primary use (most people leave out the words in parentheses):
(Empirical) data is collected measurements about the world. (It may be inaccurate, but it can't be made up)
A (falsifiable) hypothesis is a statement about the world that is either true or false, and for which there exists a test to determine whether it is true. It is a syntactical definition: it makes no comment about whether the statement is true or false, just that it must be one or the other, and that it is possible to find out which. A statement is not a hypothesis if it could be a "partially" or a "maybe", although sometimes scientists write a vague description of a hypothesis rather than being precise, and expect the reader to work out the details.
A model is a collection of hypotheses with a common theme, optionally including some consequences of these hypotheses which can be derived through (formal) logic. (Mathematics as a subject is primarily concerned with determining whether or not a consequence can be derived from a set of hypotheses using logic; see there for a precise definition)
A useful model is a model which makes predictions about the world. If the hypotheses of the model are all true, then the predictions will be accurate.
A theory is a model supported by data that indicates its hypotheses are all true.
Hence, a useful theory is a theory which makes accurate predictions about the world.
Science is concerned with the study and creation of useful theories.
As a related term, a scientific law is a (well-known) hypothesis supported by data. A theory will normally contain at least one law, since it's very hard to reason about the world without using any.
Scientists may disagree on the amount of support required to consider something a theory, and about what is included in 'the world'. There are other ways to define these terms, but they're all more or less equivalent; I like this one because it's comparatively easy to determine whether or not something qualifies as a theory. The so-called "demarcation problem" is only unsettled in the sense that people can't agree on which precise definition to use. In practice, it doesn't really matter which one you use.
I think these ideas are less important as aids to understanding real science than as vaccinations against "Snow Crash"-esque viral bogosities such as creationism and pseudo-scientific claims. We're dangerously susceptible if even scientists don't understand why it's wrong to say that evolution is merely a theory and therefore creationism should also be taught.
I'm working on a degree in Philosophy of Science right now, and also TAing for a class on the representation of science in the popular media, so maybe I can be of some help.
First, about the Demarcation Problem. Judging from the responses above, most of the commentators seem to divide science from pseudoscience along lines of falsifiability. While Karl Popper, the main progenitor of the importance of falsifiability, had a good idea going, it was not the end of debate. There has been progress regarding the Demarcation Problem in philosophy of science in the intervening 50 years, and a lot of it has been helpfully informed by psychological and sociological research. In this direction, T.S. Kuhn's The Structure of Scientific Revolutions is only the tip of the iceberg, but I would encourage everyone who has so far commented to give the book a try. For something more recent and technical, I recommend Phil Kitcher's Advancement of Science. The enterprise of science is difficult to define, but is defined by much more than the pursuit of falsifiable tests.
Next, on to whether it is important for scientists and science educators to know about issues in the history of philosophy of science. I think that knowledge of those issues are important whenever a scientist conducts research that are relevant to societal interests. When a scientist's work is sufficiently esoteric as to resist attempts at effective communication (hello, Quantum Physics!), knowledge of philosophical issues diminishes in importance relative to the knowledge required to effectively continue work. That is an issue of practicality. When the work is meant to be communicated, though (and this is ESPECIALLY important for science educators), it is important to realize that terms like theory, fact, and value are regularly contested and still in the process of development.
On that note, I'll be here all week. Cheers.
I highly recommend the book "Genesis and Development of a Scientific Fact" by Ludwik Fleck. I read it back in college and it has proved to be provocative and illuminating even now as a scientist at a university.
Coreymaley:
I wish I did not have to say this so bluntly, however: You are wrong.
There are many experiments set up and performed by astrophysicists. Those experiments may be done in a particle scintillation chamber deep in the Alps, they may be done with gravitometric satellites, thay may be done with gravitometric stations on the surface of the Earth.
Some of the best astrophysical experiments are done by allowing planets, far-off in space, to perform them for us - this is how we have experimentally verified many of Einstein's conjectures with regards to gravity and mass.
Pioneer 10 & 11, Voyager 1 & 2, Ulysses, Cassini, and New Horizons - not to mention Galileo - are all astrophysical experimental platforms, not to mention the massive radiotelescope arrays in the southwestern United States and in Arecibo.
I would suggest that you read the Kitzmiller v. Dover decision - you seem to have confused "astrophysics" with "astrology" as to what is in need of a scope-expansion to be accepted as a science.
#24 eggman:
Evolution at best is a "Model" - It can not be a theory because it resides on past, unobserved ...
Evolution is tested all the time. All botany and biology students eventually partake in such tests.
When religious science students of talent leave their colleges for university graduate schools they often experience crises of faith, especially if their schools were conservative or fundamentalist. It can be a wrenching experience when the notion that evolution is a universal process and is EVERYWHERE overtakes them. I have seen such students in crisis and it ain't a pretty sight.
When I studied anthropology (admittedly many years ago) I was surprised at how little the scientific method mattered to most cultural anthropologists; their method was more akin to library science or bookkeeping. Evolution was a dirty word to them, and as a consequence of their reactionary influence cultural anthropology was considered by many to be a joke. Things are better now, but I shall never forget once being called a ''gossip columnist'' by a geologist.
At WPI I took a required sociology course that went over what each of these things are.
yeah Buddy, there should be a term for the moment when the realization sweeps over the formerly blindly religious that there is a greater world.
having read the article linked - is it not the educators job to teach his students these terms, not despair that they do not know them. bit weird that. Also the comments by the blogger are quite ignorant of the depth of knowledge of those in the scientific community, in a fairly shocking way.
For instance: Theories can be shown to be false - but NOT proved true. Once all alternative theories are falsified whatever is left standing is the most likely fact, while still subject to continual testing for falsehood... (Sadly this is why the anti-evolution gibberish has more staying power than it deserves.)
I have a PhD in Marine Science from a University in the Gulf Coast of the US, and for my comprehensive exams the propositions of Thomas Kuhn, Paul Feyerabend and Karl Popper etc regarding the nature of scientific inquiry, figured heavily. I would assume that most universities expect their graduates to have a working knowledge of the philosophy of science as well. What else would they teach us but how to develop a testable hypothesis based on observation of nature and how to experimentally or otherwise test it's validity, in a statistically powerful manner.
Would guess from the lack of responses from other PhDs that they feel this entire discussion is a bit puerile or just another attempt to erode science in american culture...
sudo science: science playing God?
Yay, it's time for the special logic! The special logic is for things that are not everyday or true. The special logic is for convincing ourselves of things we must believe!
Generally the kind of inference that allows us to, say, tell if our shoelaces are tied, suffices for scientists. Though they do need to be periodically reminded of the finer points of statistics.
Anyway, carry on!
Falsificationism alone isn't enough of a demarcation criterion.
Consider crystal healing. Crystal healing is certainly falsifiable under certain conditions: you do a double-blind test. Provided the practitioners have a concrete theory of their practice (that's the bit Popper didn't like about Freud, etc. - they don't), you can falsify it - you'll certainly find it false, but it is falsifiable.
So, we must introduce another criterion: science must be both falsifiable, and not falsified.
This seems reasonable enough, until you consider that any theory may be falsified at any point in the future. Then the question of demarcation degrades into a question of confirmation, not just falsification - how sure can I be of this particular theory? If you want to retain scientific truths, you quickly end up trying to mount a defence of induction.
Thus, you can't just blindly tell people that science is based solely on the "simple, settled concept" of falsification.
A similar argument applies against empiricism as a criterion: there are plenty of junk pseudosciences that involve vast amounts of empirical observation. Astrology is one: it uses empirical data and observations about the position of the planets and moon. Yet we certainly don't want to admit it as a science.
So, our demarcation must involve more than just "science happens by observation", or "science is based on empiricism".
So, I hereby call out @11, @12, @15, @32, @38 and @44. You seem to think you have a viable demarcation between science and psuedoscience, but your models are inaccurate (they don't reflect how science actually works), and insufficient (they don't provide a good demarcation). If you think I'm wrong, if you think your demarcation still works, by all means, get it published. You'll either be famous, or your argument will be ripped to shreds. I wonder which?
I see this one far too often.
The problem is that the two of you have not agreed on what the word 'evolution' means, and you're using it to reference different things. It's a fairly vague term. You normally then proceed to argue past each other, accomplishing nothing.
'Evolution' encompasses many theories, models, and outright guesses. Some, like the work of Darwin, are extremely solid and backed by large amounts of data. Some, like the development of the dinosaurs and their relationship to modern creatures, are horrifyingly thin on evidence and involve far too much conjecture by the standards of modern science.
People who want to believe in one thing or another point to examples from either end of the spectrum and claim that makes their view correct. I don't think they even want to find out the truth. It's very tiresome.
Be specific. Say what you mean, not a vague description of what you mean. Leave imprecise words like "evolution" for the realm of science fiction. Don't let people get away with nonsense like "I believe in evolution". Once you get people down to "I believe that fruit flies will not adapt to a different food source" or "I believe that people are getting smarter with each generation" then it becomes painfully obvious who the crackpots are.
sooner or later (sooner), scientists need philosophers to define "rigor". Then they can get on with "for a given value of true". Priests just clutter things up.
First of all, there are no "facts" in science; only measurements. Measurements have systemic and random errors which prevent them from qualifying as facts.
Second, there are no "theories" in science; only hypotheses that have not been disproved. Newton's "theory of gravity" is an approximation to relativity, which in turn is an approximation to the real world.
A Model is a mathematical description of a hypothesis that predicts what measurements should be obtained if the hypothesis is correct. The scientific method consists of developing hypotheses, modeling them, and then testing the models with measurements and discarding the invalid hypotheses.
Mathematics is at the core of science, because it is a universal language that can be used to describe measurements, models, and hypotheses. Mathematics itself is not a "fact" any more than a scientific hypothesis, it is merely a tool that begins with specific axioms and can be used with great success to make sense of the world. On the other hand, mathematics is as real as any scientific observation. Both physical observations and mathematics rely on the human mind to interpret them. In this sense, a mind may be just as sure of performing physical measurements as it is of performing mental arithmetic and other mathematical operations. The point is that the scientific method, relying on mathematics, is so far the best meta-model that humans have discovered for surviving and learning about the world. Many people confuse science and mathematics with truth, when instead they should treat the best science and mathematics available as only the best known model, and continue to search for something better.
My guess is that this latter concept of skepticism is what is lacking in people who don't understand the nuances of facts, theories, and models. They search for truth and expect to find it, instead of a continually better approximation through science.
notjustatheory.com
Do you need it simpler ?
@48:
I would say that the examples you mentioned are not instances of experiments, but tests of models. A particular model can lead to testable predictions, but that's quite different than performing experiments on the actual phenomenon of interest. That's all perfectly good science; the issue is that a naive conception of falsifiability does not always work in examples of real science, because we cannot directly do the experiments.
Another example would be neuroscience in humans: we just can't slice up people's brains to do the kinds of experiments that would lead to a vindication of falsifiability claims. However, we can model human brains using the brains of other animals, and make predictions about what would happen if certain lesions were to occur. You can say that, when a person has some accident resulting in a brain lesion, we've let "nature" do the experiment for us, but that's a rather gratuitous notion of experiment: I don't see a principled way to distinguish that from observation. There are a lot of experiments that one would need to perform that "nature" is just not going to perform for us.
Regarding Definitions of Terms:
Asuffield @#44: Beautifully said. As short as possible, but no shorter. Is it original, or are you paraphrasing? If the latter, I'd love to know your source.
Synchronym @#58: Cool. notjustatheory.com/ nicely captures that part of Asuffield's "summary of science" which is desperately lacking in the public awareness. Let's all make printouts and hang them by the water cooler!
Regarding Demarcation:
I don't know if it can be easily generalized from a detection statistic for cranks on Usenet science groups to a tool for demarcation between science and pseudoscience, but I always get a chuckle out of rereading the Crackpot Index. It's funny because it's true.
Regarding Scientists and the PoS:
In my opinion, undergrad science students should concern themselves with learning science, not philosophy (excepting of course that they are taking a PHIL minor or double major). It is not until you've completed (or nearly completed) your undergraduate studies that you become sufficiently well-versed in the practice of science for the type of meta-analysis offered by Philosophy of Science to be really meaningful or relevant (and therefore interesting). Airpillo @#20, take heed. Even then it's not guaranteed: I knew people in university who basically faked their way through a science degree by being really good at memorizing the solutions to all the sample questions in their textbooks, and never bothered to try developing a scientific mindset at all.
A corollary to this opinion is that I don't usually have much patience for philosophers of science who weren't first scientists. Wispsmoke @#46, I'm sorry, but as heartfelt as your post was, I actually thought it provided observational evidence in support of the hypothesis presented by Pipenta @#9. What the heck does the "relevance to societal interests" of one's field have to do with whether or not one, as a scientist, ought to have a firm understanding of PoS?
The goal of a scientist is to produce good scientific results. The goal of a philosopher of science is to think about the process of science, including whether there are any hidden weaknesses inherent in the way that science is currently practiced, and if so, to try to identify strategies for mitigating these weaknesses. The goal of a science educator is to teach the basics of science---as it is currently practiced---to the masses. That means covering the stuff listed in Asuffield's post (@#44), it doesn't mean the kind of (again, I apologize for my bias) abstract wishy-washy gedankeneksperiments that philosophers sit around entertaining during their faculty colloquia.
As fascinating---and important---as the philosophical ramifications (on all of mathematics) are from something like Gödel's Incompleteness Theorem, it really has no bearing in the context of a lay-educator trying to explain that the Pythagorean Theorem generalizes from 2 to 3 (or even N!) dimensions. Ditto for PoS and science educators in general. It's exactly the kind of uncontrolled spewing off of philosophical skepticism which is endemic among philosophers rather than scientists (however well-intentioned for the sake of intellectual rigor) that has enabled the creationists to put the shiny new suit of Intelligent Design on their arguments and pass it off to the ignorant masses as being equivalent in stature and a viable alternative to Darwinism. (Please note, I said "ignorant", not "stupid".)
Hmm... In the time it took me to regurgitate this particular diatribe, it appears that Takuan @#56 beat me to the punch, and with much more succinctness. Damn.
Are there multiple theories that can account for the same facts? (hint - yes) If all theories are are under-determined how would one choose between them?
W.V.O. Quine's "Two Dogmas of Empiricism" leaps to mind. But if Quine is in then Popper is out (and good riddance if you ask me).
I looked into your soul while you were preoccupied.
I had a friend at university who took a survey round academics in biology and biochemistry.
It have a number of binary questions relating to what a gene actually is.
Unfortunately I wouldn't trust my memory on what any of them actually were. Something like "Is inherited" was probably one of the least esoteric, and depending how you look at that there's as much information theory as biology in there.
He showed me the results and there wasn't any particular sign of agreement.
Doesn't mean they weren't good at what they did.
#55 - you forgot to slap our cheeks with one of your pedant leather gloves.
it is customary to do honour to proper names when dueling, numerals are naff.
Asuffield @#56:
I know this thread is getting longer and longer, but your post brings to mind a classic joke:
Socrates goes out for a stroll one day with one of his colleagues. Suddenly they begin to hear heated voices in the distance, and as they continue their walk the voices get louder and louder (as they get closer and closer in distance), and it becomes apparent that they belong to two people who are engaged in a passionate debate. When eventually the voices are at their loudest, and there is still no visible evidence of their owners, Socrates and his colleague stop and look around in bewilderment. It soon becomes clear that the debaters are standing on the roofs of their respective houses, across the road from one another. Sensing an opportunity for intellectual sustenance, the two philosophers decide to observe for a while. After listening for five minutes or so, Socrates remarks to his colleague: "You know, they're never going to arrive at a consensus." To which his colleague responds: "Why not?". Socrates' reply: "They're arguing from different premises."
Ba doom schnick! Try the veal, etc.
(Thanks to Evagelia R. for telling me this joke so long ago that I didn't even get it at the time.)
I really had a hard time reading the comments on this.. seems like a few too many posters wiki'd "the demarcation problem" and ran with that as a first shot approach to the problem that the original author in The Scientist was trying to address. It is completely meaningless in this discussion...
#44 has done your dirty work here, these are the terms and meanings of them, as widely accepted in Science. There is no need for semantics here, outside of some very small circles there is no debate as to what we mean when we talk about facts/laws, hypotheses, theories, etc.
Getting back to the discussion, the blog author has simply missed the point of Mr. Williams and gone off on some tangents that are inexplicable.
"I can agree with the first sentence and not the second. There is certainly a difference between "a fully fledged scientific theory that is backed by evidence and accepted by the scientific community and a speculative guess" but it isn't demonstrated by teaching children the philosophy of science. Most scientists are quite competent in doing science even though they aren't versed in the philosophical arguments about what it is they are doing. If they bothered with those questions most would likely become paralyzed."
Williams wasn't positing that these educators should know the finer philosophical debates that entertain the minds of a few logic die-hards, he was simply stating that many of these 'scientists' don't actually understand the basic terms that make up the degrees of strength in defining something, in the realm of science. Without repeating #44 verbatim, lets just say that these are the fundamental tenants of science, most of which are introduced conceptually in middle school. Evolution is accepted as a theory because it has evidence to support its hypotheses, such as the fossil record (thats really the big one). Evolution is less strong as theory than relativity, which has far more and compelling evidence to support it. Still, both are elevated to the realm of theory because of the evidence that supports them.
Pseudoscience is science in which things (theories) that do not have a large body of evidence to support their hypotheses are pushed through without standard rigor or even a testable hypothesis. And by pushed through I mean reported in major media outlets as a real and present danger. That was kind of the point of the issues that Williams brings up. Hope I helped clear up what was being debated here, and would love to hear an argument as to why science educators (or scientists for that matter) should not have to know the fact/law -> hypothesis -> theory progression and be able to explain it clearly. (Btw, Williams said most of them did know, but an unnerving % did not define it clearly)
In reply to message #37
"Facts are not truth; truth is absolute and rare. The only universal truth, imho, is we don’t know the truth. Facts and truth are not self-evident."
This is not exactly true. There is a category of facts that are true by definition, and then there are other facts of somewhat less certainty that are based on observations, measurements, etc.
Philosophers of science have terminology for these different kinds of facts, and I'll leave it up to you to look those up if you want to know.
For an example that is true by definition: a triangle is a closed geometrical figure that has three sides and three vertices.
DAW
I can see lots of posts from scientists on here, but not many from philosophers. I will take a step forward, and identity my self as a Philosophy graduate (MA), who started off his degree in the Physics faculty before transferring. I try not to admit that too much...
From my recollections of the Philosophy of Science class, the difference between science and pseudo-science is that real science tries to repeatedly test it's basic hypotheses and the theories dependent on them. No theory or hypothesis should be untestable and sacrosanct.
The way pseudoscience works is by assuming a certain set of premises, i.e. people are influenced by the alighment of stars and planets from when they were born (astrology), or that the human mind is split into three parts (psychoanalysis) and then move from those to add detail through experimentation and observation.
If there is a problem with this view, then I haven't managed to come up with it. There are untestable hypotheses included in the scientific process - that things can be proved by experimentation, that facts exist on a true/false binary scale, with no room in-between, and suchlike, but these are assumptions about reality, not about the way we are trying to explain it.
I do always try and keep in mind the problem of the Logical Positivists when discussing this. The main thesis they expounded was that every idea or theory must be proved, and lots of time was spent on the arguments to show that this was possible. The only problem? They were never able to prove that theory...
Since I have nothing insightful to add to this discussion itself, just wanted to say this has been one of the most fascinating comment threads I've read in a while.
It makes me happy to see so many smart folks having a spirited and intelligent discussion, and I've learned a few new things (and it's not even 10 am!).
Thanks to all the participants — especially asuffield (#44) — for your insights.
As a multivariate empirical scientist my thoughts trend away from the strictly logical toward the statistical. Hypothetical Truth is statistical based upon the data pool, confidence criteria, correlation standards and assumptions. I accept that pure logic has huge weaknesses associated with multiple meanings, paradox, and the shallow nature of words and metaphor.
I prefer logic confirmed by multiple perspectives, not just by shallow mental logic but confirming intuition, confirming visual logic, confirming gut feel, confirming analogy models, confirming intuition, confirming statistical experiments.
Thus, for me, a pseudo scientist is one who:
1. Insufficiently provides a basis for statistical experimentation and independent confirmation of his hypothesis.
2. Does not use Occam's razor to choose the simplest hypothesis or does not explain why.
3. When faced with nonconfirming statistics stubornly remains attached to his false hypothesis.
I think the demarcation problem is not most salient when presented in something like evolution but perhaps something like quantum mechanics or super-string theory...
Partially because the fields of interest/study are (relatively) new, but also partially because the things being described often are not directly observable.
Which means theories about how to interpret results from machines or instruments and theories about how even to get valid data. Then you run into a bunch of other problems.
As Palindromic (#68) mentioned, a few people tend to wiki things, but in a lot of cases - it's not the demarcation problem alone that is a cause for concern. It's the culmination of a lack of understanding about the structure of science, about generalisations, problems with induction, issues with underdetermination. Sometimes it's hard to show a salient "problem" with science utilising just one of these in isolation. These are interesting and current questions for scientists though. Understanding why we operate under the premises we do is, I think, an important part of the continued pursuit of knowledge.
Science consists of:
1) The process of gathering evidence
2) The process (blackest art) of making conjectural models about how pieces of evidence relate;
3) The process (black art) of using a conjecture to represent the data in a hypothesis, which can convey all accumulated data via the conjecture (with the null hypothesis being the trivial example)
4) The competitive proving (from Latin probare, to test) of all presented hypotheses for probability of predicting future results, done practically by experimental method or theoretically by measurement for MDLI (see "Minimum Description Length Induction, Bayesianism and Kolmogorov Complexity", by Paul M. B. Vitanyi and Ming Li);
5) The body of hypotheses presently proven best, known as "theory". Note that which hypothesis is considered "theory" may potentially change, based on the arrival of additional evidence or a new hypothesis. (The sports-obsessed might consider it akin to a title belt, subject to new ownership based on another match or a new contender.)
Using the theoretical form for #4, science thus reduces to a branch of philosophy which assumes the validity of formal logic for rules of inference, that the Zermelo-Fraenkel set axioms (absent choice) may be self-consistently asserted, and that Reality is related to Evidence via (arbitrary) finite set of finite rules. Since there is no non-circular means to "prove" these, they must essentially be asserted as valid via Faith; this, however, is where the requirements of Faith end for Science. All other "rules" of science (such as "cause and effect") are not themselves propositions of Faith, but inferences based on present evidence via the primary tenets, and subject to change (as QM has for C&E) if the balance of evidence changes. Popperian Falsification may be refined (or reduced) to the requirement that a tentative hypothesis be other than the null hypothesis.
(Of course, the math for the proofs of this is beyond the usual high-school level. Changing that might or might not be desirable.)
Psuedoscience usually violates one of the requirements, or mandates the use of additional philosophical prior assumptions beyond those listed.
As a student currently finishing his Ph.D. in Physics, I can say: these issues need to be addressed. Formally. In a curriculum-based format.
science keeps you alive.
#74
So intelligent thou seem, perhaps the equal of the mightiest of philosophers who have argued endlessly and worthlessly these many centuries over the triviality of paradox and nuance. Did not I suggest the need to confirm by alternate means the shallowness of thy formal logic? Wouldst thou continue a hypotheses given intuitive failure? Better to find out why then to continue the fools errand. Wouldst thou continue the use of Bayesianism and Kolmogorov Complexity given the limited significance these actually may provide? Would thou neglect the null hypothesis by proclaiming it trivial? Does black art mean to you a science only thou are qualified for? To what hast thou ego contributed by this defense of thy logical mind? Do you therefore mean only the exclusively logical mind need be those that do pursue the science? To these questions I address the field and not so you exclusively, given your excessive logical proclivity.
JOB007, He may be wrong, you may be right, but I think I would have let the big guy kick your ass.
FOETUSNAIL, It's about pseudo science in the zone between easy science, easy philosophy, easy theology where nothing is easy because the definitions break down depending on the perspective. Some try to project back 14 billion years to the big bank or bounce or multiverse or wormhole nonsense or whatever, far beyond the data base (maybe 2000 years crudely recorded knowledge) we actually have. Some try to base known knowledge on literal interpretation of ancient text by politicians on 40 day hot desert fasts. Some try to create knowledge by math or logical syllogism without connecting reality to the definitions.
OK, I'm skeptical and could be more diplomatic. Why should anyone trust anyone? The historical data supports the hypothesis that all persons or propositions are untrustworthy untill proven otherwise.
Show the data, assumptions, projections, correlations, outliers, and conclusions, otherwise it's pseudo science or worse.
Palindromic @68, is reading off single correct answers to questions of this caliber sufficient to understand them? Is the sole purpose of conversation to produce a correct answer and then move on?
I don't see any benefit in being quite so dismissive of a benevolent, complex, and civil discussion of how we understand the process of science. There will inevitably be some discussion of near-understandings, and the relative value of different approaches to understanding. I don't see that as invalid. When travelers have diverse starting points but a common destination, they're bound to compare sets of driving instructions. I hold that the overall tendency of the conversation is convergence on a better understanding.
DN'T KNW F Y PPL R SRS R VRYN S JST PLNLY NSN R STPD! MYB F Y, YR HGHNSSS F "SCNTFCLLY BRGHTNSS" ND PSD-NTLLCTL XTRVGNZ, WLD KNW SMTHNG LS THN WHT S BVS ND LMNTL, Y ND YR MRNC SCNTSTS WLD KNW THT FR THS SMPL CNCPTS F FCT, THRY, ND MDL, CRRSPND VRY SMPL LNGSTC - D Y KNW WHT T S R D LS ND T XPLN WHT T S - THSRS DFNTNS. G S THM. BT HY, 'M JST BLDY FRGNR TRY T PT P FR VR 4 YRS WTH DTC PSD-DCTRS ND PSD-NTLLCTLS N STPD MRCN NVRSTY. HV FN BNG PSD.
And even if Job007 were completely right on every point, which I'm not seeing, it wouldn't warrant such prickliness.
Aeneias, we don't do that.
Consider my favorite definition of science:
"Science is the total cumulative results of an open-ended search for natural explanations of observable phenomena."
adapted from from draft education standards from the State of Kansas.
* Science is the "total cumulative results" which means that science is not just the latest findings. It includes all the findings of all the studies. Sometimes the findings from different studies are inconsistent and you have to harmonize them or recognize that uncertainty remains.
* Science is an "open ended search" which means that science is never complete. Our understanding increases, but we will never know it all. Science is a never-ending process of discovering more and more about the world. Answering one question often raises another question.
* Science requires "natural explanations" which means that our explanations must come from the natural world, not the world of magic, religion, myth, superstition, meta-physics, astrology, gods, extra-terrestrials, faith, etc.
* Science involves "observable phenomena" which means that our our hypothesis must be testable. Our experiments must delve into phenomena which we can measure, which mean they must involve changes we can see, hear, feel, etc, or at least things we have the technology to measure such changes. If we can't measure the things that change as a result of our experiments then we can't test our hypotheses.
might I recommend these three books?
http://en.wikipedia.org/wiki/The_Science_of_Discworld
Domster:
I never said that the definition of science I put forward was based on "Falsificationism alone ..." I never said it was based on "... a defence of induction" or "... empiricism alone".
I said, and I quote:
"... real science involves producing a falsifiable hypothesis, testing it repeatedly, and getting peer review and replication of the results of experiments."
There are THREE parts to the statement, and you have addressed only TWO, and each of those SEPARATELY.
You state that "So, our demarcation must involve more than just "science happens by observation", or "science is based on empiricism"."
-- and then go on to "call out" - at the very least - one statement of the philosophy of science that clearly transcends mere observation, mere empiricism.
You have committed a fallacy by producing a strawman of the statement of philosophy I put forward, by dividing the statement :
"... real science involves producing a falsifiable hypothesis, testing it repeatedly, and getting peer review and replication of the results of experiments."
into two /separate/ items and declaiming them as if they did not, in the case of the operation of science, depend one upon the other for it to properly be considered the scientific method.
-- Or, perhaps, you merely have an axe to grind and failed to read and consider as a whole, the /entire/ statement :
"... real science involves producing a falsifiable hypothesis, testing it repeatedly, and getting peer review and replication of the results of experiments."
No semi-colons. No full-stops. I do try to be as clear as possible when I write, yet Damned I shall be the day I do proclaim that the sky is blue unto the entire world and someone doth not cry "HELLEBLAU!? ICH KANN NICHT SEHEN" in the spirit of contrarianism.
Also, you seem to be of the opinion that the open-ended-ness of the scientific method - the ability to falsify a hypothesis, to overthrow a theory in favour of a better one - is some sort of /failing/ in the philosophy behind it.
You are wrong.
Science is /meant/ to be changeable by new evidence. Science is /meant/ to be changeable by new data, by new hypotheses, by new models that better fit the observed and available facts.
---doopdedoojustbeinghelpfulProseMaTron3000----
So intelligent thou seem'st,
pr'aps the equal of the mightiest
of philosophers
who, these many centuries
have argued endlessly and without worth
on the triviality of paradox and nuance.
Did I not suggest the need,
by alternate means,
to confirm the shallowness
thy formal logic doth betray?
Wouldst thou an hypotheses given, continue unto intuitive failure?
Better to find out why
than to play the fool upon errand.
Wouldst thou continue
Bayesianism and Kolmogorov Complexity
given the limited significance these actually may ensign?
Wouldst thou neglect
the null hypothesis
by proclaiming it trivial?
Dost thou find a science only thou are qualified for a black art?
To what hast thy ego contributed by this defense of thy reasoning mind?
Do you conclude thereby that only the exclusively logical mind need be those that do pursue the science?
These questions I do address unto the field and not so you exclusively - given your excessive love of logic.
---doopdedoo justbeing helpfulProseMaTron3000---
JOB007: Did not I suggest the need to confirm by alternate means the shallowness of thy formal logic?
Yes, with your reference to "multiple meanings, paradox, and the shallow nature of words and metaphor". This is why the preferred testing is by Minimum Description Length Induction, which requires mathematical expression.
If you have a paradox in ZF, publish already.
JOB007: Wouldst thou continue a hypotheses given intuitive failure?
Possibly as part of semi-exhaustive search; encountering the Banach-Tarski sphere dissection leaves one with much less respect for intuition. If someone has taken the trouble to examine a conjecture, express data via hypothesis, and asserts it has a minimum induction length lower than an alternative, I'd certainly look at the work. Proof verification is only P-hard, IIR.
JOB007: Wouldst thou continue the use of Bayesianism and Kolmogorov Complexity given the limited significance these actually may provide?
Errr... actually, you should check that paper; B&KC yield Minimum Description Length Induction as a mathematically rigorous form of "Occam's Razor"... which you mentioned as a requirement for science. This allows Occam's razor to be an inference of the very possibility of inference, rather than asserted as a primary proposition. Also, using the formal expression of Minimum Description Length Induction makes it clear why "God diddit" does not qualify as the "simplest" explanation for anything.
JOB007: Would thou neglect the null hypothesis by proclaiming it trivial?
No. That it is trivial means it should always be the very first competitor to any tentative hypothesis; anything that can't beat the null hypotheses is a waste of time to try taking on any others in the field.
JOB007: Does black art mean to you a science only thou are qualified for?
No. It means that the step can be anywhere in RE-hard. ("Blackest art" indicates it is RE-Complete hard.) These steps may be performed in any way someone sees fit, as long as they accomplish the task. Waiting for the results to be delivered inscribed on golden tablets by seraphim choir courier team is acceptable (though in the event, most professional associations may have ethical guidelines applicable to co-authorship listing for publication).
JOB007: To what hast thou ego contributed by this defense of thy logical mind?
"Thy ego"... and I didn't claim to be logical. See above regarding Seraphim choir, my favorite method for getting a hypothesis.
What I've contributed is the expression of this in English, rather than mathematics. The mathematics is far more precise.
JOB007: Do you therefore mean only the exclusively logical mind need be those that do pursue the science?
Would you consider someone waiting for Seraphim choir delivery "exclusively logical"? The non-exclusively logical mind need merely be willing to have the results tested by the (hypothetical) logical mind.
For me, I tend towards thinking that art largely involves a subjective exploration of the universe and that science largely involves the objective exploration of the universe.
Perhaps truly objective exploration is impossible - in which case, perhaps the quest for objectivity is key to the sciences? Is this where philosophy steps in?
From this perspective, a psuedo-science would fit somewhere between these poles - either because it can't, or won't, measure up to the traditional tests associated with scientific enquiry and objectivity.
I'm not a trained scientist - my head stuffed on your mantle piece isn't likely to be worth that much; but please shoot me anyway ... I need to learn.