If we accept a new philosophy that promotes selecting theories based on something other than facts, why stop at physics? I envision a future in which climate scientists choose models according to criteria some philosopher dreamed up. The thought makes me sweat.
That kind of reasoning normally makes me want to physically beat people with a copy of the complete works of Plato, but I shall refrain on this occasion Because Context. Anyway the answer to "where will it stop ?" is always, as John Oliver put it, somewhere.
The philosophers are certainly right that we use criteria other than observational adequacy to formulate theories. That science operates by generating and subsequently testing hypotheses is only part of the story. Testing all possible hypotheses is simply infeasible; hence most of the scientific enterprise today—from academic degrees to peer review to guidelines for scientific conduct—is dedicated to identifying good hypotheses to begin with... It doesn’t relieve us from experimental test, but it’s an operational necessity to even get to experimental test.
In the foundations of physics, therefore, we have always chosen theories on grounds other than experimental test. We have to, because often our aim is not to explain existing data but to develop theories that we hope will later be tested—if we can convince someone to do it. But how are we supposed to decide what theory to work on before it’s been tested? And how are experimentalists to decide which theory is worth testing? Of course we use non-empirical assessment. It’s just that, in contrast to Richard, I don’t think the criteria we use are very philosophical. Rather, they’re mostly social and aesthetic. And I doubt they are self-correcting.
Not sure why it's "Richard" here and not "Dawid" ? Are they on a first-name basis ? Meh. I agree with the sentiment though : we generally choose theories because we like them, not for any higher philosophical reasoning. However (reversing the order of the original text for my own
He claims that certain criteria that are not based on observations are also philosophically sound, and he concludes that the scientific method must be amended so that hypotheses can be evaluated on purely theoretical grounds. Richard’s examples for this non-empirical evaluation—arguments commonly made by string theorists in favour of their theory—are (1) the absence of alternative explanations, (2) the use of mathematics that has worked before, and (3) the discovery of unexpected connections.
... those all seem like pretty good criteria to me. A theory that cannot be empirically tested but which is intended to be so eventually is a sort of pro-science. Mathematics itself is like this, and no-one blames mathematicians for not doing empirical tests, and woe betide anyone who say's they're irrational.
String theory is currently the most popular idea for a unified theory of the [fundamental physics] interactions. It posits that the universe and all its content is made of small vibrating strings that may be closed back on themselves or have loose ends, may stretch or curl up, may split or merge. And that explains everything: matter, space-time, and, yes, you too. At least that’s the idea. String theory has to date no experimental evidence speaking for it... Arguments from beauty have failed us in the past, and I worry I am witnessing another failure right now.
I don't like string theory, I think it's over-hyped. But I think it falls in more of the grey area between mathematics and science at this point, rather than being something which is (as implied above) obviously wrong.
“So what?” you may say. “Hasn’t it always worked out in the end?” It has. But leaving aside that we could be further along had scientists not been distracted by beauty, physics has changed—and keeps on changing. In the past, we muddled through because data forced theoretical physicists to revise ill-conceived aesthetic ideals. But increasingly we first need theories to decide which experiments are most likely to reveal new phenomena, experiments that then take decades and billions of dollars to carry out. Data don’t come to us anymore—we have to know where to get them, and we can’t afford to search everywhere. Hence, the more difficult new experiments become, the more care theorists must take to not sleepwalk into a dead end while caught up in a beautiful dream. New demands require new methods. But which methods? I hope the philosophers have a plan.
I think there's some Historian's Fallacy at work here. I know a lot of people who are convinced that satellite galaxies orbit in planes around their hosts. They do all kinds of fantastically elaborate theories to demonstrate this, and I'm absolutely persuaded that at least two of them are much more intelligent than me. Nonetheless, they are wrong - the planes just don't exist. What I'm not convinced of is that if they hadn't followed this wrong idea they'd have made some other, more useful contribution to the field instead - it is at least equally possible they'd have become stuck on some other wrong idea instead. Equally, perhaps my own conviction that planes don't exist is holding me back from making a useful contribution to the field. Or, to take it to extremes, I could suggest that all theoretical physicists should be forced to work on cancer research or something. The problem is there's little to suggest they'd be any good at it and megatonnes (literally) of evidence that pure research eventually leads to practical consequences.
There's only so far you can go. Once you've made your arguments and tried to persuade other researchers of their follies, you can't really do more than that. You have to accept other people's views and let them get on with things : ultimately you just disagree.
https://www.scientificamerican.com/article/a-theory-with-no-strings-attached-can-beautiful-physics-be-wrong-excerpt/
I expect my copy of Lost In Math to arrive tomorrow. In the meantime, I feel compelled to consider "Theoretical and Applied Beauty" as a topic of further study. A series of arxiv.org papers may be forthcoming. The first might be "Theoretical Prose and Poetry: What Beauty Does the Data Expose? Can the Experiments Be Reproduced?"
ReplyDelete