Slashdot Mirror
Have Some Physicists Abandoned the Empirical Method?
HughPickens.com writes: Adam Frank and Marcelo Gleiser write in the NY Times that two leading researchers, George Ellis and Joseph Silk, recently published a controversial piece called "Scientific Method: Defend the Integrity of Physics," that criticized a newfound willingness among some scientists to explicitly set aside the need for experimental confirmation of today's most ambitious cosmic theories — so long as those theories are "sufficiently elegant and explanatory." Whether or not you agree with them, Ellis and Silk have identified a mounting concern in fundamental physics: Today, our most ambitious science can seem at odds with the empirical methodology that has historically given physics its credibility.
Quoting: "Chief among the 'elegance will suffice' advocates are some string theorists. Because string theory is supposedly the 'only game in town' capable of unifying the four fundamental forces, they believe that it must contain a grain of truth even though it relies on extra dimensions that we can never observe. Some cosmologists, too, are seeking to abandon experimental verification of grand hypotheses that invoke imperceptible domains such as the kaleidoscopic multiverse (comprising myriad universes), the 'many worlds' version of quantum reality (in which observations spawn parallel branches of reality) and pre-Big Bang concepts. These unprovable hypotheses are quite different from those that relate directly to the real world and that are testable through observations — such as the standard model of particle physics and the existence of dark matter and dark energy. As we see it, theoretical physics risks becoming a no-man's-land between mathematics, physics and philosophy that does not truly meet the requirements of any."
Richard Dawid argues that physics, or at least parts of it, are about to enter an era of post-empirical science. "How are we to determine whether a theory is true if it cannot be validated experimentally," ask Frank and Gleiser. "Are superstrings and the multiverse, painstakingly theorized by hundreds of brilliant scientists, anything more than modern-day epicycles?"
Quoting: "Chief among the 'elegance will suffice' advocates are some string theorists. Because string theory is supposedly the 'only game in town' capable of unifying the four fundamental forces, they believe that it must contain a grain of truth even though it relies on extra dimensions that we can never observe. Some cosmologists, too, are seeking to abandon experimental verification of grand hypotheses that invoke imperceptible domains such as the kaleidoscopic multiverse (comprising myriad universes), the 'many worlds' version of quantum reality (in which observations spawn parallel branches of reality) and pre-Big Bang concepts. These unprovable hypotheses are quite different from those that relate directly to the real world and that are testable through observations — such as the standard model of particle physics and the existence of dark matter and dark energy. As we see it, theoretical physics risks becoming a no-man's-land between mathematics, physics and philosophy that does not truly meet the requirements of any."
Richard Dawid argues that physics, or at least parts of it, are about to enter an era of post-empirical science. "How are we to determine whether a theory is true if it cannot be validated experimentally," ask Frank and Gleiser. "Are superstrings and the multiverse, painstakingly theorized by hundreds of brilliant scientists, anything more than modern-day epicycles?"
Social Scientists regularly jettison reproducible, empirical methods of gathering data in favor of cherry-picking to push ideological outcomes in their studies. Biologists are starting to do the same. It's all about producing the right "narrative" now. "Problematic" facts will be discarded, and any semblance of objective truth will be damned.
That's probably the biggest fundamental danger with something like String Theory that currently has no method to test; if it effectively becomes dogma over centuries or even decades in the way that the Earth-centric Universe had then it's very, very difficult to undo that even when a new hypothesis with actual compelling evidence is crafted. The problem isn't even necessarily among scientists either, though they can have their doubts, but in a public that doesn't understand the scientific process and is unwilling to accept a scientific challenge to their deeply-held world views and religious perspectives.
I suppose that's why I want scientists to continue working on other hypotheses for the explanation of the fundamental structure of the Universe, so that scientists and the public don't make too many assumptions about what's right versus what's still up for debate.
Do not look into laser with remaining eye.
In many ways, we're not so different than the Greeks trying to measure the speed of sound - without good clocks. The Standard Model may be the Taj Mahal of empericism. It has plenty of predictive power, based on lots and lots of observations, but is almost completely lacking in explanitory power. Weight of the electron? Value of the strong nuclear force? Meh, ask me a hard one. Modern physicists today are like people who work in a restaraunt but can never leave. "Damn", they say to each other, "it gets busy like that every day at 12:05. Watch, it'll do it again tomorrow." - such is the state of physics tday - all observation, no explanation, no true understanding why things are that way. Chemistry doesn't have that problem, they can pretty much explain ALL of chemistry from first principles. Of course the Plank distance and energy plus uncertanty put an absolute limit on observation, so, in some ways, we can never get out of the restaraunt.
Cosmologists are in the same boat, just at different scales. Untestable assumptions like "space is uniformly flat everywhere" and "the gravitational constant and the speed of light is the same everywhere and everywhen" lead us to conundrums like Dark Matter(TM) and Dark Energy(TM) just to put a label on the things we are ignorant about because the sums (and assumptions) don't add up.
TL;DR If you have an explanatory story where the math works, I, for one, don't care, and don't think it matters if your story amounts to "it's turtles all the way down". At least then we can start trying to dream up experiments and observations to prove or disprove it. But physics current OBSESSION with empericism has us painted into a corner where we don't have (and may never be able to have) the equipment to come up with defintive answers directed from observation rather than directed from theory.
Mod parent up - he is right on about science becoming a religion. I'm working in astronomical instrumentation (France), and I'm shocked to hear stuff like "we believe that the atmosphere follows von Karman model". Who the f%$# told you that?! You have no evidence of this, yet developing methods based on this shaky assumption. And then they are surprised that it does not work...
No one bothers with preproducibility anymore. I caught my colleagues repeatedly at cheating and outright cooking up results. The articles are written in a way that it is impossible to replicate the methods, let alone simulation/experimental results. I know couple of articles in mathematics that contain non-working algorithms. Those were "peer"-reviewed articles in good journals...
Another desese is called simulations. Good Lord, if only you knew how primitive and wrong those simulations are! I'm sitting right now and for 2 freaking months I'm trying to make a simulator (written by my boss) to work. When it does, those simulations are nowhere near the results he pulished in the paper. Half the questions about the code he answers like "we found THIS to work, but we don't know why". F%$#ing great! Am I in the laboratory of astrophysics, or maybe in astrology and homeopathy lab?
And don't get me even started about those so-called "soft sciences" like sociology...
I have a different version of the question.
Has the general public set aside empiricism as a standard against which to judge funding appropriations in the name of fundamental scientific progress?
I say no.
The public has not set aside empiricism as part of the social contract through which public money is directed at research institutions. Once the public understands how tenuous empiricism has become among research physicists, the tiny trickle we already provide will only get smaller.
So here's my message to all the modernist physicists out there ready to bury Karl Popper (there were one or two in this year's Edge question): speculate all you want about the non-falsifiable multiverse, but use the Templeton Foundation to fund your chalk supply, and whisper sweet nothings on bent knees so that they also fund your chalk boards, bean bag chairs, and baloney sandwich cafeteria.
It's not like public research funds have nowhere else to go. Proteomics, as difficult as it is, has not yet broken free of its empirical yoke (the complexity of this field begins with the water molecule, and ramps up from there).
We should start with the auto-immune diseases which ought to be simpler systems—if, in fact, they are indeed auto-immune diseases after all.
There really ought to be an entire chapter in Kahneman's next book devoted to the human psychology quirk through which an otherwise sensible person willingly exchanges ten of twenty free physical parameters for 10^500 fiendishly complex initial conditions and calls it a good deal.
There's no problem at all with being a mathematician or a philosopher of science. I'm a physicist, and I don't think any of my colleagues would argue that these fields should go away or that physicists shouldn't work in them. Emmy Noether is a great example of how people outside physics can help develop new physics.
But... relativity wasn't accepted until it was tested. Neither should any other theory coming out of advanced mathematics. Simply being around for a long time is not enough to move a set of math from clever speculation into physics. We've been down this path before. Allowing foundational theories to be integrated into the rest of physics without verification might end up fine, or it might waste the careers of a generation of physicists. Today, that also might mean many billions of dollars of funding and significant public trust.
You say this like there's some cabal deciding on 'allowing foundational theories into the rest of physics without verification'.
If you look at the Dec. 2014 Nature article that sparked the NYTimes article, you'd see that the concern there isn't even about the conduct of science itself -- it's about the worry that apparent dissent among scientists will fuel anti-scientism. So we'd better work out these 'what's experimentally verifiable' questions far away from the inquiring public.
There's no real worry that somehow the world's best and brightest physicists have forgotten about falsifiability.
in academics.
The problem is that there is a huge oversupply of Ph.D. and Ph.D. candidate labor for the number of positions available pure academic research and institutions.
People that offer incremental improvements or work—just lab stuff, just more data, just duplication research, or slight variations to tease out empirical nuances—are a dime a dozen and struggle to differentiate themselves. Real science is often workman-like and laborious.
On the other hand, if young Ph.D. candidates and people weaving their way through the identity-building process that is a Ph.D. focus on conceptual innovation and performances—ideas, narratives, radical departures—then they are seen as doing something "new" and "innovative" (which somehow has become what science is about in popular discourse, which creeps into academic discourse), and something that sells better in the presses and to the public when the monographs come out, enabling "crossover" works and coverage that is much more lucrative than straight empirical work that gets buried in the journals or small print runs. They also more attendees at the conferences, and by virtue of interviewing and appearing more, get more coverage for the institution that hires them, often doing more to drive prestige and enrollments.
I think market forces play into this in a significant way.
STOP . AMERICA . NOW
We may be running into a fundamental limitation of the scientific method. By Goedel's impossibility theorem, for a given set of logical rules and axioms, there will always exist logical constructs which are unprovable - i.e. cannot be determined to be true or false. "This sentence is false." Is it true or false? For a physical analogue, if everything we know about the speed of light is correct, there is no way to ever observe what is inside a black hole. Yet obviously there is something inside. The fact that we can't observe it within the limitations imposed by physics doesn't mean it doesn't exist. It falls into a class of physical phenomena which can't be tested by the scientific method.
I've been trying to explain this to people this for years. Logical truth isn't binary. Every hypothesis (theory, speculation, faith, whatever you want to call it) doesn't resolve into being true or false. They resolve into true, false, or cannot be determined. While by the standards of the scientific method something in the "cannot be determined" category cannot be shown to be true, it is equally erroneous to decide it is false. So contradictory to what many rationalists think, there are actually three categories of belief - scientific truth (belief in things provable by science), superstition (belief in things disproven by science), and what for lack of a better word I'll call supposition (belief in things that cannot be proven nor disproven by science). String theory probably falls into the third category, which is why it's wrong to lump string theorists with the second category simply because it can't be proven by the scientific method.
From a mathematical standpoint, even if string theory is wrong, if it can come up with accurate predictions, it can still be useful and worth pursuing. Early astronomers believed the planets were painted onto spheres, and attempted to define their motion across the sky as a circle within a circle, offset slightly from the observer While we now know that their method was wrong, the predictions these circles came up with were pretty accurate. Basically it's like using a nth order polynomial to fit a set of data points. The phenomenon that created those data points probably isn't a nth order polynomial, but if the polynomial can accurately predict in-between data points, then it doesn't really matter that it's wrong. You just need to be careful of extrapolating outside the data points, or expecting more accuracy from your prediction than the accuracy of your data points.