Matt Strassler has been trying hard to make the boundary look more fuzzy or non-existent. But the boundary is still there.
There are findings in physics
- that could be planned, whose relevance is immediately obvious, but that require hard work that someone unsurprisingly does and someone should pay for via grants
- that are unexpected, qualitatively change the landscape and thinking, redirect the character of the future research, but still may be supported by solid evidence.
Let me begin with a comment by "dude" that I subscribe to.
Incidentally, dude is a physicist who works in a physics town called Son-of-the-Queen – Thousand-Kilos in the U.S. state named Fashionable T-shirt (I hope that you know enough U.S. geography) and I am not quite sure about his or her identity. When something is obvious or obviously wrong, he or she doesn't hesitate to point it out. So I learned I was doped when I implicitly wrote that squarks were fermions :-).
The right explanation was that I woke up early to see a paper I had wanted to see for days and my brain wasn't still fully running when I was writing that sentence, but the general message of dude's criticism was of course right. My mistake was very silly.
On Matt Strassler's blog, dude wrote:
Such luxurious prizes are given to people who have done something that few others could do. Many of the recipients of the prize definitely answer to this criterion. Take any of them and imagine he did not exist, well, the world would look quite differentAs you could have noticed at the top, I completely agree with that. That doesn't mean that I don't admire the people who are the world's best professionals in higher-loop QCD computations. Thousands of insights, skills, facts, and formulae that I don't know are hardwired in their brains. Nevertheless, I am confident that I know enough to have a good idea about the "big picture" of their work and it seems clear that all the obstacles that ever emerge in their work may be overcome in rather straigthforward ways. They are constrained by lots of things that we know and they develop procedures to sail in between these constraints, using the method of trial and error (and other methods).
This is not true for hardcore higher-loop QCD computations. Somebody gotta do it, and if not them then somebody else would have stepped up.
Matt Strassler disagrees with dude and thinks that dude's blasphemy is exactly the reason why there have to be prizes for the hard workers, to make them feel "equal" to the revolutionaries. Well, they're not quite equal. The world of modern physics would be totally different if we didn't know there exists M-theory, supersymmetry, that the Universe probably inflated some time ago, it could have large dimensions, quantum gravity obeyed the holographic principle, and so on. However, if someone didn't find a way to define jets that is accurate enough for two-loop calculations, someone else would find a different definition of jets that would be comparably accurate later.
I can't believe that Strassler doesn't see the difference between the importance and scarcity of these two types of contributions to physics. It seems more plausible to think that he sees the difference as clearly as everyone else does but he just decided to promote an egalitarian perspective in physics whether or not it is defensible.
While glorifying the BlackHat team (nothing whatsoever against them, of course, it's just that the things "for them" aren't infinite), Strassler writes:
I have to follow-up on this one, it’s such a stupid remark. Why do you think the BlackHat people are some of the leaders in the field of calculations? Because they were smart enough to invent new methods! It’s not as though just anyone can do that — you need both the smarts of good theorists and the years of experience with calculations to recognize something clever that will actually work.It's probably fair. You need some of the best mathematicians-practitioners and years of work. But it's still an "engineering" job and you're bound to find such people and the solutions if you offer them good enough conditions (I don't mean just a salary but I do also mean a salary). Many similarly good mathematicians-practitioners are found in the experimental particle physics teams or hedge funds or other places. There is a large enough set of smart people who are replaceable, however.
But my point is the following one. To ignite conceptual breakthroughs in physics, you need much more than that. You need people who are able to think outside the box, at the right places outside the box, and maybe sometimes closer to the center of the box than where anyone else has ever been (i.e. deeply inside the box), people who have intuition that can't be planned, and you're never sure whether such people exist on the job market at all. Maybe such people are born once in 10 years or 20 years or 100 years, depending on the importance of the breakthrough we are talking about. They're just a higher category. Perhaps they also needed some good luck but if you see someone who is able to have good luck so many times, you should start to doubt that the "good luck" is the right explanation.
To argue that there's no difference between the Milner Prize winners and the higher-loop QCD professionals, Strassler offers this tricky argument:
And it’s not an accident that the BlackHat founders are the same people who have made major contributions, while this has all been going on, to both our understanding of the maximally supersymmetric gauge theory and the maximally supersymmetric gravity theory, including ones that the Milner prize winners Maldacena and Arkani-Hamed have often referred to in their own work.So these higher-loop physicists are at the very top because they've been cited by Arkani-Hamed and Maldacena, LOL. Is Strassler serious or was it just a joke? I have lots of citations from Witten, Maldacena, and two papers co-written with Arkani-Hamed, among other things, but I don't claim and I have never claimed to be or to have been their full-fledged peer. People are citing other people's work and the citation doesn't mean that the "citer" and the "citee" are exactly at the same level.
Theorists who write more important papers or less important papers also refer to experimenters but that doesn't mean that every experiment that has been done has been as remarkable as a theoretical breakthrough and the experimenters had to be unusually exceptional physicists. Experiments simply may be done. Many of them could be done by pretty much any trained physicist. To a large extent, complicated numerical calculations are analogous to experiments. They may be done. It requires a significant amount of expertise and lots of hard work but they're "predictable". In particular, if you allow the results to be approximate, there's always a way to improve the accuracy or the number of loops (which are related). It's an evolutionary process, something that differs from a revolutionary event.
Bernhard stood on Strassler's side with this comment:
dude,I think it's at least debatable whether Milner is a scientist. It depends how you define it. At any rate, many people surely care about what he thinks, at least because of those millions. Among the people who have been doing different things for more than a decade, Milner seems to be unusually aware about what's shaking, and that's another reason why I do care.
People working with these computer programs are actually very much acknowledged for their work (just not by Milner, but he is not a scientist so who cares?) and the world would indeed look quite different if they didn’t exist. Can you imagine the world without Pythia? Without HERWIG? Remember, writing these programs is a bit of an art as well as result of years of theoretical research. They carry the imprint of their fathers.
Can I imagine a world without Pythia? Yes, it would be a world that differs from ours in details but not in the grand features. There wouldn't be a Pythia but there would be a Kuthia (Czech readers may recall that the main unit of time is a sepynda). And maybe there would be two programs, Kut and Hia, that would approximately play the same role as Pythia when combined. The details of the program would surely be different. More efficient than Pythia in some respects, less efficient in others.
So I agree that Pythia or HERWIG carry the fingerprint of their creators. Most programs do. But just the fact that something carries someone's fingerprints doesn't make it critically important or revolutionary. Competent physicists have different fingerprints and the main point I want to make is that one's fingerprints aren't necessarily "much better" than someone else's fingerprints. So it's just irrational to worship someone's fingerprints just because they're there. It's a typical feature of an engineering work that there's a lot of ambiguity about the way how things may be done, different people choose different solutions, and their products therefore carry fingerprints.
And indeed, the reason why conceptual discoveries such as cosmic inflation, string theory, or large extra dimensions are more important is that the main ideas do not carry any fingerprints of their discoverers! That's really why they're important. The discoverers have found something that is objectively important, not just something that the best solution according to one particular person with one random fingerprint.
Spanish engineers could have been building landfills in Spain and they carried their fingerprints. On the other hand, Columbus has found a new continent and the new continent doesn't really carry his fingerprints. It's not even named after him. The richest part of the continent doesn't even consider itself to be a Spanish or British colony anymore. ;-) But that's exactly why the New World was important. Its importance went well beyond one person's idiosyncrasies. The era of exploration still demanded some bold people who were doing things that were needed for such discoveries.
Bernhard says that the creation of such programs is also an art. I agree with that. One needs to be an artist of a sort to find such solutions etc. But artists are creative workers who produce new things that are not unique but it's exactly the uniqueness and universal importance and inevitability that we cherish in physics. So being an artist in the sense of making lots of choices that could also be done differently is exactly one of the signs that a physicist is mainly a hard worker and not a revolutionary. So Bernhard's logic is really upside down. It's hard to say whether it's his reasoning or his primary values that are upside down.
In a subsequent comment, Strassler defends the hard workers in QCD against the evil barbarian string theorists in this way:
While I completely agree with this comment, I think the problem with it is that it doesn’t communicate across the knowledge barrier. Most string theorists think Pythia is just a silly computer program that simulates what any good theorist could do with a pen and paper. That’s because they don’t understand how much non-trivial physics insight goes into understanding the formation of jets (URL), [...]Lots of knowledge about detailed physics goes into Pythia; most of this knowledge about elementary interactions, parton distributions, showers, and fragmentation were found previously by others and the Pythia creators combined them. It's an impressive piece of hard work that could have only been created when the expertise and trials and errors of lots of smart and hard-working folks was combined. But that's exactly why it's not fundamental, earth-shaking science. It's the union of previously known "qualitative knowledge" found by someone else and lots of new partial insights, technicalities that are guaranteed to be found if one works hard enough.
[...] or into the very observation that quantum mechanical processes in QCD can be simulated using a classical computer program, [...]That's a totally unrealistic accusation. String theorists (and others) know that their brains work pretty much like classical gadgets – conventional computers – but they may still find out what happens in a system by pure thinking or "simulations" so it would be silly to think that computers with much higher frequencies and larger memories aren't useful. Of course that they're useful. From a mathematical viewpoint, the probability amplitudes are "ordinary" integrals (path integrals) that may be calculated. Parts of these expressions can be conveniently calculated in one way, parts of them are more conveniently translated to different concepts and opposite limits and these parts may be merged and interpolated in various ways. There's a lot of flexibility and lots of things have to be decided in a sensible way to do it right but that's exactly why it's not fundamental science. It's a form of engineering.
[...] or into the observation that there are things about the proton that we can learn in one class of experiments and then apply in a completely different class. [...]Much of research in string theory or any other sufficiently abstract discipline of science is about learning things about one object and applying the knowledge elsewhere. It's crazy to suggest that string theorists are unfamiliar with this intellectual procedure; they're really the best ones in the world when it comes to the application of this concept because string theory is the longest intellectual bridge between objects and situations that are seemingly as separated as possible.
[...] All of these are non-obvious facts about QCD that someone really smart had to prove. [...]Right. There's been lots of hard work and lots of things have been proved. Others haven't been proved. But exactly because there are lots of such similarly important insights, they must be treated statistically and none of them may be said to qualitatively transform all of physics.
[...] String theorists tend to assume that once you have the equations, the rest is just details; but almost none of them have actually read the very challenging and brilliant QCD literature from the 1970s and 1980s that made all of this stuff possible. [...]This sentence is wrong at many levels.
First, it is nonsensical to say that string theory research is only about the search for the defining equations of the theory (although yes, I surely find this part of the research to be among the most fundamental ones). The easiest way to prove this proposition of mine is to notice a simple fact, namely that no one knows "the" primordial universal equations of all of string theory. We don't even know whether such universal equations exist.
String theory as we know it today is a "manifold" composed of many overlapping "patches", each of which is understood as a systematic expansion (often to all orders of perturbation theory) or a non-perturbatively exact set of equations that however prevent us from changing the asymptotic structure of the spacetime (i.e. they can't be extrapolated into another patch). But each of the patches uses the same kind of "phenomenology" that the people who study detailed QCD phenomena are familiar with. Again, string theorists are not unfamiliar with this mode of thinking; they are doing it all the time. Almost none of the important insights in string theory of the last decades were about "finding totally new elementary equations". Instead, the important insights were about relationships between the equations i.e. about effects that these equations surprisingly implied.
So Strassler has completely mischaracterized what the research of string theory actually looks like. But there is another, equally serious problem with his musings: he still tries to deny that there is a difference between shocking, earth-shaking, conceptual, qualitative, surprising discoveries on one side and collections of predictable contributions boiling down to lots of hard work and expertise on the other side.
Even though string theorists are ultimately doing "the same kind of research" as Strassler describes – they usually develop concept, tools, and methods needed to understand the physical behavior of known equations – they still distinguish revolutionary events from incremental evolutionary progress. They distinguish it not only outside string theory; they distinguish it within string theory, too. People in every meaningful scientific discipline make this distinction, too. When it comes to these basic features such as the existence of a "hierarchy of importance", string theory is just another scientific discipline.
So once one (Argentine guy) discovers the AdS/CFT correspondence and perhaps once some other guys outline the bulk interpretation of the boundary correlation functions, it's obvious to almost everyone that one may try to calculate these objects in many example theories and to various degrees of precision. So someone inevitably does lots of these things – that's what the 8,000 papers referring to Maldacena's AdS/CFT paper are about – and none of these papers' authors ever tries to claim that his or her paper is exactly as important as Maldacena's paper (unless there is some exception, and a clear exception for which this statement would be indisputably legitimate is still being waited for). It would simply be ludicrous.
[...] Almost none of them are keeping up with the problem of multi-scale calculations, which is again an issue of fundamental theoretical importance in quantum field theory. [...]OK, I totally disagree. It's exactly one of the subdisciplines that isn't fundamental, almost by definition, because it is just a conglomerate of several fundamental categories of insights that are just merged and interpolated in a satisfactory way. Multi-scale calculations aren't fundamental physics for the same reason why the transport of genetically modified horses by aircraft isn't fundamental: there are many "combined problems" of this kind and one is solving some particular technicalities that arise from their combinations. There probably doesn't exist any surprising "holy grail" here, just the gradual evolutionary improvements of the "transfer of GM horses by air" technology. But (relatively) more fundamental issues are being solved by those who construct aircrafts of genetically modified horses themselves. ;-)
Hard calculations of this kind are important and must be done but they are just not fundamental. Fundamental insights have to be insights that may be formulated and repeatedly used and that wouldn't be true if they were modified. A methodology to do calculations is just a methodology to do a calculation. It's not unique, it's not fundamental, it's just an engineering solution to a "homework exercise" that had to exist if the underlying equations were truly consistent. Of course, even among these ideas and tricks, one finds more important (or universal) ones and less important (or universal) ones.
[...] And until very recently, almost none of them understood that the formation of jets is a problem in conformal field theory; this was well-known in some QCD circles (well, certainly I knew it, and I was on the borderline between the subfields) but I think Maldacena was the first string theorist to make it known to the string theory community. [...]This insight tries to pretend to be completely true, novel, fundamental, and surprising, but it's really neither. Everyone has known from the 1970s that the formation of jets is a process that depends on the strong coupling in a gauge theory, one for which the perturbative i.e. weakly coupled QCD isn't enough. And everyone has known since the 1970s that conformal field theories are the "starting" point, rudimentary enough quantum field theories. And nontrivial (interacting) conformal field theories are the ultimate model for strongly coupled quantum field theories (e.g. strongly coupled QCD).
However, the real-world QCD just isn't exactly conformally invariant so predictions of a conformal field theory won't agree with the observations exactly. Some features will depend on the conformal physics, other observed phenomena depend on the theory's being non-conformal (well, the bottom quark, for example, always carries the same mass). This much has been known for nearly four decades, too.
So Strassler's claim isn't really accurate as it stands. It's analogous to the commercial mentioned by Feynman, "Wesson oil isn't absorbed by the food" which is partly right and partly wrong, depending on the temperature but not depending on the brand of the oil. So the commercial is intrinsically misleading. Analogously, to make Strassler's claim right, one has to separate the features that depend on the conformal dynamics and features that depend on the deviations of QCD from conformality. There is a fuzzy border somewhere in between the assumptions – this much has been known from the 1970s as well – and the location of this boundary is known much more accurately than it was known decades ago. But that's a result of hard detailed work, not qualitative breakthroughs.
So when it comes to the validity of Strassler's claim (and similarly most claims in the large body of scientific research as we know it), the devil is in the details. But that's different from fundamental breakthroughs because in the case of the fundamental breakthroughs, the devil is in their primary big-picture identity, not in the detail. The explosive power of a revolutionary insight in physics doesn't depend on the details. Quite on the contrary, it has its own "life support" and it opens many detailed questions that may be studied by others.
Matt Strassler doesn't understand how to distinguish vital, precious discoveries in physics from more or less minor hard work.
And that's the memo.
0 comments:
Post a Comment