Pages

Showing posts sorted by relevance for query quantization procedure. Sort by date Show all posts
Showing posts sorted by relevance for query quantization procedure. Sort by date Show all posts

Saturday, December 20, 2014

Has Loop Quantum Gravity been proved wrong?

Logo of site by name loop insight.
The insight to take away is that you have to
carefully look for those infinities
[Fast track to wisdom: Probably not. But then.]

The Unruh effect is the predicted, but so-far not observed, particle production seen by an accelerated observer in flat space. It is a result obtained using quantum field theory and does not include gravity, and the particles are thermally distributed with a temperature that is proportional to the acceleration. The origin of the particle production is that the notion of particles, like the passage of time, is observer-dependent, and so what is Bob’s vacuum might be Alice’s thermal bath.

The Unruh effect can be related to the Hawking effect, that is the particle production in the gravitational field of a black hole, by use of the equivalence principle. Neither of the two effects has anything to do with quantum gravity. In these calculations, space-time is treated as a fixed background field that has no quantum properties.

Loop Quantum Gravity (LQG) is an approach to quantum gravity that relies on a new identification of space-time degrees of freedom, which can then be quantized without running into the same problems as one does when quantizing perturbations of the metric. Or at least that’s the idea. The quantization prescription depends on two parameters, one is a length scale normally assumed to be of the order of the Planck length, and the other one is a parameter that everybody wishes wasn’t there and which will not be relevant in the following. The point is that LQG is basically a modification of the quantization procedure that depends on the Planck length.

In a recent paper now Hossain and Sadar from India claim that using the loop quantization method does not reproduce the Unruh effect

If this was correct, this would be really bad news for LQG. So of course I had to read the paper, and I am here to report back to you.

The Unruh effect has not been measured yet, but experiments have been done for some while to measure the non-gravitational analog of the Hawking effect. Since the Hawking effect is a consequence of certain transformations in quantum field theory that also apply to other systems, it can be studied in the laboratory. There is some ongoing controversy whether or not it has been measured already, but in my opinion it’s really just a matter of time until they’ve pinned down the experimental uncertainties and will confirm this. It would be theoretically difficult to claim that the Unruh effect does not exist when the Hawking effect does. So, if it’s true what they claim in the paper, then Loop Quantum Gravity, or its quantization method respectively, would be pretty much ruled out, or at least in deep trouble.

What they do in the paper is that they apply the two quantization methods to quantum fields in a fixed background. As is usual in this calculation, the background remains classical. Then they calculate the particle flux that an accelerated observer would see. For this they have to define some operators as limiting cases because they don’t exist the same way for the loop quantization method. They find in the end that while the normal quantization leads the expected thermal spectrum, the result for the loop quantization method is just zero.

I kinda want to believe it, because then at least something would be happening in quantum gravity! But I see a big problem with this computation. To understand it, you first have to know that the result with the normal quantization method isn’t actually a nice thermal distribution, it is infinity. This infinity can be identified by a suitable mathematical procedure, in which case one finds that it is the zero of a delta function in momentum space. Once identified, it can be factored out, and the prefactor of the delta function is the thermal spectrum that you’ve been looking for. One can trace back the physical origin of this infinity to find it is, roughly speaking, that you’ve looked at the flux for an infinite volume.

These types of infinites appear in quantum field theory all over the place and they can be dealt with by a procedure called regularization that is the introduction of a parameter, the “regulator”, whose purpose is to capture the divergences so that they can be cleanly discarded of. The important thing about regularization is that you have to identify the divergences first before you can get rid of them. If you try to divide out an infinite factor from a result that wasn’t divergent, all you get is zero.

What the authors do in the paper is that they use a standard regularization method for the Unruh effect that is commonly used for the normal quantization, and apply this regularization also to the other quantization. Now the loop quantization in some sense already has a regulator, that’s the finite length scale that, when the quantization is applied to space-time, results in a smallest unit of area and volume. If this length scale is first set to zero, and then the regulator is removed, one gets the normal Unruh effect. If one first removes the regulator, the result is apparently zero. (Or so they claim in the paper. I didn’t really check all their approximations of special functions and so on.)

My suspicion therefore is that the result would have been finite to begin with and that the additional regularization is an overkill. The result is zero, basically, because they’ve divided out an infinity too much.

The paper however is very confusingly written and at least I don’t see at first sight what’s wrong with their calculation. I’ve now consulted three people who work on related things and neither of them saw an obvious mistake. I myself don’t care enough about Loop Quantum Gravity to spend more time on this than I already have. The reason I am telling you about this is because there has been absolutely no reaction to this paper. You’d think if colleagues go about and allegedly prove wrong the theory you’re working on, they’d be shouted down in no time! But everybody loop quantum just seems to have ignored this.

So if you’re working on loop quantum gravity, I would appreciate a pointer to a calculation of the Unruh effect that either confirms this result or proves it wrong. And the rest of you I suggest spread word that loop quantum gravity has been proved wrong, because then I’m sure we will get a clarification of this very very quickly ;)

Monday, October 17, 2011

Super Extra Loop Quantum Gravity

In the summer, I noted a recent paper that scored remarkably low on the bullshit index:
    Towards Loop Quantum Supergravity (LQSG)
    Norbert Bodendorfer, Thomas Thiemann, Andreas Thurn

    Bullshit Index : 0.08
    Your text shows no or marginal indications of 'bullshit'-English.

But what is the paper actually about? It is an attempt to make contact between Loop Quantum Gravity (LQG) and Superstring Theory (ST). Both are approaches to a quantization of gravity, one of the big open problems in theoretical physics. LQG directly attacks the problem by a careful selection of variables and quantization procedure. String theory does not only aim at quantizing gravity, but at the same time at unifying also the other 3 interactions of the standard model by taking as fundamental the strings that give it its name. If quantizing gravity and unifying the standard model interactions are actually related problems, then string theorists are wise to attack them together. Yet, we don't know if they are related. In any case, it has turned out that gravity is necessarily contained in ST.

Both theories still struggle to reproduce general relativity and/or the standard model, and to make contact to phenomenology, though for very different reasons. This begs the question how the theories compare to each other, whether they give the same results for selected problems. Unfortunately, so far this has not been possible to answer because LQG has been developed for a 3+1 dimensional space-time, while ST famously or infamously, depending on your perspective, necessitates 6 additional dimensions that then have to be compactified. ST is also, as the name says, supersymmetric. It should be noted that these both features, supersymmetry and extra dimensions, are not optional but mandatory for ST to make sense.

I've always wondered why one hasn't extended LQG to higher dimensions since the idea of extra dimensions is appealing and somebody in the field who should have known better once told me it would be straight forward to do. It is however not so because one of the variables (a certain SU(2) Yang-Mills connection) used in the quantization procedure relies on a property (the equivalence of the fundamental and adjoint representations of SU(2)) that is fulfilled only in 3 spatial dimensions. So it took many years and two brilliant young students, Norbert Bodendorfer and Andreas Thurn, to come up with a variable that could be used in an arbitrary number of dimensions and to work through the maths which, as you can imagine, didn't get easier. It required to work around the difficulty that SO(1,D) is not compact and digging out a technique for gauge unfixing, a procedure that I had never heard of before.

Compared to the difficulty of adding dimensions, going supersymmetric can be done by simply generalizing the appropriate matter content which is contained in the supergravity actions, and constructing a supersymmetry constraint operator.

Taken together, this in principle allows one to compare the super extra loop quantized gravity to string theory, to which supergravity is a low energy approximation, though concrete calculations have yet to follow. One of the tasks on the to-do list the entropy of extremal supersymmetric black holes to see if LQG reproduces the ST results. (Or if not, which might be even more interesting.) Since LQG is a manifestly non perturbative approach, this relation to string theory might also help filling in some blanks in the AdS/CFT correspondence in areas where neither side of the duality is weakly coupled.

Tuesday, September 01, 2015

Loops and Strings and Stuff

If you tie your laces, loops and strings might seem like parts of the same equation, but when it comes to quantum gravity they don’t have much in common. String Theory and Loop Quantum Gravity, both attempts to consistently combine Einstein’s General Relativity with quantum theory, rest on entirely different premises.

String theory posits that everything, including the quanta of the gravitational field, is made up of vibrating strings which are characterized by nothing but their tension. Loop Quantum Gravity is a way to quantize gravity while staying as closely as possible to the quantization methods which have been successful with the other interactions.

The mathematical realization of the two theories is completely different too. The former builds on dynamically interacting strings which give rise to higher dimensional membranes, and leads to a remarkably complex theoretical construct that might or might not actually describe the quantum properties of space and time in our universe. The latter divides up space-time into spacelike slices, and then further chops up the slices into discrete chunks to which quantum properties are assigned. This might or might not describe the quantum properties of space and time in our universe...

String theory and Loop Quantum Gravity also differ in their ambition. While String Theory is meant to be a theory for gravity and all the other interactions – a “Theory of Everything” – Loop Quantum Gravity merely aims at finding a quantum version of gravity, leaving aside the quantum properties of matter.

Needless to say, each side claims their approach is the better one. String theorists argue that taking into account all we know about the other interactions provides additional guidance. Researchers in Loop Quantum Gravity emphasize their modest and minimalist approach that carries on the formerly used quantization methods in the most conservative way possible.

They’ve been arguing for 3 decades now, but maybe there’s an end in sight.

In a little noted paper out last year, Jorge Pullin and Rudolfo Gambini argue that taking into account the interaction of matter on a loop-quantized space-time forces one to use a type of interaction that is very similar to that also found in effective models of string interactions.

The reason is Lorentz-invariance, the symmetry of special relativity. The problem with the quantization in Loop Quantum Gravity comes from the difficulty of making anything discrete Lorentz-invariant and thus compatible with special relativity and, ultimately, general relativity. The splitting of space-time into slices is not a priori a problem as long as you don’t introduce any particular length scale on the resulting slices. Once you do that, you’re stuck with a particular slicing, thus ruining Lorentz-invariance. And if you fix the size of a loop, or the length of a link in a network, that’s exactly what happens.

There has been twenty years of debate whether or not the fate of Lorentz-invariance in Loop Quantum Gravity is really problematic, because it isn’t so clear just exactly how it would make itself noticeable in observations as long as you are dealing with the gravitational sector only. But once you start putting matter on the now quantized space, you have something to calculate.

Pullin and Gambini – both from the field of LQG it must be mentioned! – argue that the Lorentz-invariance violation inevitably creeps into the matter sector if one uses local quantum field theory on the loop quantized space. But that violation of Lorentz-invariance in the matter sector would be in conflict with experiment, so that can’t be correct. Instead they suggest that this problem can be circumvented by using an interaction that is non-local in a particular way, which serves to suppress unwanted contributions that spoil Lorentz-invariance. This non-locality is similar to the non-locality that one finds in low-energy string scattering, where the non-locality is a consequence of the extension of the strings. They write:

“It should be noted that this is the first instance in which loop quantum gravity imposes restrictions on the matter content of the theory. Up to now loop quantum gravity, in contrast to supergravity or string theory, did not appear to impose any restrictions on matter. Here we are seeing that in order to be consistent with Lorentz invariance at small energies, limitations on the types of interactions that can be considered arise.”


In a nutshell it means that they’re acknowledging they have a problem and that the only way to solve it is to inch closer to string theory.

But let me extrapolate their paper, if you allow. It doesn’t stop at the matter sector of course, because if one doesn’t assume a fixed background like they do in the paper one should also have gravitons and these need to have an interaction too. This interaction will suffer from the same problem, unless you cure it by the same means. Consequently, you will in the end have to modify the quantization procedure for gravity itself. And while I’m at it anyway, I think a good way to remedy the problem would be to not force the loops to have a fixed length, but to make them dynamical and give them a tension...

I’ll stop here because I know just enough of both string theory and loop quantum gravity to realize that technically this doesn’t make a lot of sense (among many other things because you don’t quantize loops, they are the quantization), and I have no idea how to make this formally correct. All I want to say is that after thirty years maybe something is finally starting to happen.

Should this come as a surprise?

It shouldn’t if you’ve read my review on Minimal Length Scale Scenarios for Quantum Gravity. As I argued in this review, there aren't many ways you can consistently introduce a minimal length scale into quantum field theory as a low-energy effective approximation. And pretty much the only way you can consistently do it is using particular types of non-local Lagrangians (infinite series, no truncation!) that introduce exponential suppression factors. If you have a theory in which a minimal length appears in any other way, for example by means of deformations of the PoincarĂ© algebra (once argued to arise in Loop Quantum Gravity, now ailing on life-support), you get yourself into deep shit (been there, done that, still got the smell in my nose).

Does that mean that the string is the thing? No, because this doesn’t actually tell you anything specific about the UV completion, except that it must have a well-behaved type of non-local interaction that Loop Quantum Gravity doesn’t seem to bring, or at least it isn’t presently understood how it would. Either way, I find this an interesting development.

The great benefit of writing a blog is that I’m not required to contact “researchers not involved in the study” and ask for an “outside opinion.” It’s also entirely superfluous because I can just tell you myself that the String Theorist said “well, it’s about time” and the Loop Quantum Gravity person said “that’s very controversial and actually there is also this paper and that approach which says something different.” Good thing you have me to be plainly unapologetically annoying ;) My pleasure.

Wednesday, February 13, 2019

When gravity breaks down

[img:clipartmax]
Einstein’s theory of general relativity is more than a hundred years old, but still it gives physicists headaches. Not only are Einstein’s equations hideously difficult to solve, they also clash with physicists other most-cherish achievement, quantum theory.

Problem is, particles have quantum properties. They can, for example, be in two places at once. These particles also have masses, and masses cause gravity. But since gravity does not have quantum properties, no one really knows what’s the gravitational pull of a particle in a quantum superposition. To solve this problem, physicists need a theory of quantum gravity. Or, since Einstein taught us that gravity is really curvature of space-time, physicists need a theory for the quantum properties of space and time.

It’s a hard problem, even for big-brained people like theoretical physicists. They have known since the 1930s that quantum gravity is necessary to bring order into the laws of nature, but 80 years on a solution isn’t anywhere in sight. The major obstacle on the way to progress is the lack of experimental guidance. The effects of quantum gravity are extremely weak and have never been measured, so physicists have only math to rely on. And it’s easy to get lost in math.

The reason it is difficult to obtain observational evidence for quantum gravity is that all presently possible experiments fall into two categories. Either we measure quantum effects – using small and light objects – or we measure gravitational effects – using large and heavy objects. In both cases, quantum gravitational effects are tiny. To see the effects of quantum gravity, you would really need a heavy object that has pronounced quantum properties, and that’s hard to come by.

Physicists do know a few naturally occurring situations where quantum gravity should be relevant. But it is not on short distances, though I often hear that. Non-quantized gravity really fails in situations where energy-densities become large and space-time curvature becomes strong. And let me be clear that what astrophysicists consider “strong” curvature is still “weak” curvature for those working on quantum gravity. In particular, the curvature at a black hole horizon is not remotely strong enough to give rise to noticeable quantum gravitational effects.

Curvature strong enough to cause general relativity to break down, we believe, exists only in the center of black holes and close by the big bang. In both cases the strongly compressed matter has a high density and a pronounced quantum behavior which should give rise to quantum gravitational effects. Unfortunately, we cannot look inside a black hole, and reconstructing what happened at the Big Bang from today’s observation can, with present measurement techniques, not reveal the quantum gravitational behavior.

The regime where quantum gravity becomes relevant should also be reached in particle collisions at extremely high center-of-mass energy. If you had a collider large enough – estimates say that with current technology it would be about the size of the Milky Way – you could focus enough energy into a small region of space to create strong enough curvature. But we are not going to build such a collider any time soon.

Besides strong space-time curvature, there is another case where quantum effects of gravity should become measureable that is often neglected: By creating quantum superpositions of heavy objects. This causes the approximation in which matter has quantum properties but gravity doesn’t (the “semi-classical limit”) to break down, revealing truly quantum effects of gravity. A few experimental groups are currently trying to reach the regime where they might become sensitive to such effects. They still have some orders of magnitude to go, so not quite there yet.

Why don’t physicists study this case closer? As always, it’s hard to say why scientists do one thing and not another. I can only guess it’s because from a theoretical perspective this case is not all that interesting.

I know I said that physicists don’t have a theory of quantum gravity, but that is only partly correct. Gravity can, and has been, quantized using the normal methods of quantization already in the 1960s by Feynman and DeWitt. However, the theory one obtains this way (“perturbative quantum gravity”), breaks down in exactly the strong curvature regime that physicists want to use it (“perturbatively non-renormalizable”). Therefore, this approach is today considered merely a low-energy approximation (“effective theory”) to the yet-to-be-found full theory of quantum gravity (“UV-completion”).

Past the 1960s, almost all research efforts in quantum gravity focused on developing that full theory. The best known approaches are string theory, loop quantum gravity, asymptotic safety, and causal dynamical triangulation. The above mentioned case with heavy objects in quantum superpositions, however, does not induce strong curvature and hence falls into the realm of the boring and supposedly well-understood theory from the 1960s. Ironically, for this reason there are almost no theoretical predictions for such an experiment from either of the major approaches to the full theory of quantum gravity.

Most people in the field presently think that perturbative quantum gravity must be the correct low-energy limit of any theory of quantum gravity. A minority, however, holds that this isn’t so, and members of this club usually quote one or both of the following reasons.

The first objection is philosophical. It does not conceptually make much sense to derive a supposedly more fundamental theory (quantum gravity) from a less fundamental one (non-quantum gravity) because by definition the derived theory is the less fundamental one. Indeed, the quantization procedure for Yang-Mills theories is a logical nightmare. You start with a non-quantum theory, make it more complicated to obtain another theory, though that is not strictly speaking a derivation, and if you then take the classical limit you get a theory that doesn’t have any good interpretation whatsoever. So why did you start from it to begin with it?

Well, the obvious answer is: We do it because it works, and we do it this way because of historical accident not because it makes a lot of sense. Nothing wrong with that for a pragmatist like me, but also not a compelling reason to insist that the same method should apply to gravity.

The second often-named argument against the perturbative quantization is that you do not get atomic physics by quantizing water either. So if you think that gravity is not a fundamental interaction but comes from the collective behavior of a large number of microscopic constituents (think “atoms of space-time”), then quantizing general relativity is simply the wrong thing to do.

Those who take this point of view that gravity is really a bulk-theory for some unknown microscopic constituents follow an approach called “emergent gravity”. It is supported by the (independent) observations of Jacobson, Padmanabhan, and Verlinde, that the laws of gravity can be rewritten so that they appear like thermodynamical laws. My opinion about this flip-flops between “most amazing insight ever” and “curious aside of little relevance,” sometimes several times a day.

Be that as it may, if you think that emergent gravity is the right approach to quantum gravity, then the question where gravity-as-we-know-and-like-it breaks down becomes complicated. It should still break down at high curvature, but there may be further situations where you could see departures from general relativity.

Erik Verlinde, for example, interprets dark matter and dark energy as relics of quantum gravity. If you believe this, we do already have evidence for quantum gravity! Others have suggested that if space-time is made of microscopic constituents, then it may have bulk-properties like viscosity, or result in effects normally associated with crystals like birefringence, or the dispersion of light.

In summary, the expectation that quantum effects of gravity should become relevant for strong space-time curvature is based on an uncontroversial extrapolation and pretty much everyone in the field agrees on it.* In certain approaches to quantum gravity, deviations from general relativity could also become relevant at long distances, low acceleration, or low energies. An often neglected possibility is to probe the effects of quantum gravity with quantum superpositions of heavy objects.

I hope to see experimental evidence for quantum gravity in my lifetime.


* Except me, sometimes.

Friday, June 12, 2015

Where are we on the road to quantum gravity?

Damned if I know! But I got to ask some questions to Lee Smolin which he kindly replied to, and you can read his answers over at Starts with a Bang. If you’re a string theorist you don’t have to read it of course because we already know you’ll hate it.

But I would be acting out of character if not having an answer to the question posed in the title did prevent me from going on and distributing opinions, so here we go. On my postdoctoral path through institutions I’ve passed by string theory and loop quantum gravity, and after some closer inspection stayed at a distance from both because I wanted to do physics and not math. I wanted to describe something in the real world and not spend my days proving convergence theorems or doing stability analyses of imaginary things. I wanted to do something meaningful with my life, and I was – still am – deeply disturbed by how detached quantum gravity is from experiment. So detached in fact one has to wonder if it’s science at all.

That’s why I’ve worked for years on quantum gravity phenomenology. The recent developments in string theory to apply the AdS/CFT duality to the description of strongly coupled systems are another way to make this contact to reality, but then we were talking about quantum gravity.

For me the most interesting theoretical developments in quantum gravity are the ones Lee hasn’t mentioned. There are various emergent gravity scenarios and though I don’t find any of them too convincing, there might be something to the idea that gravity is a statistical effect. And then there is Achim Kempf’s spectral geometry that for all I can see would just fit together very nicely with causal sets. But yeah, there are like two people in the world working on this and they’re flying below the pop sci radar. So you’d probably never have heard of them if it wasn’t for my awesome blog, so listen: Have an eye on Achim Kempf and Raffael Sorkin, they’re both brilliant and their work is totally underappreciated.

Personally, I am not so secretly convinced that the actual reason we haven’t yet figured out which theory of quantum gravity describes our universe is that we haven’t understood quantization. The so-called “problem of time”, the past hypothesis, the measurement problem, the cosmological constant – all this signals to me the problem isn’t gravity, the problem is the quantization prescription itself. And what a strange procedure this is, to take a classical theory and then quantize and second quantize it to obtain something more fundamental. How do we know this procedure isn’t scale dependent? How do we know it works the same at the Planck scale as in our labs? We don’t. Unfortunately, this topic rests at the intersection of quantum gravity and quantum foundations and is dismissed by both sides, unless you count my own small contribution. It’s a research area with only one paper!

Having said that, I found Lee’s answers interesting because I understand better now the optimism behind the quote from his 2001 book, that predicted we’d know the theory of quantum gravity by 2015.

I originally studied mathematics, and it just so happened that the first journal club I ever attended, in '97 or '98, was held by a professor for mathematical physics on the topic of Ashtekar’s variables. I knew some General Relativity and was just taking a class on quantum field theory, and this fit in nicely. It was somewhat over my head but basically the same math and not too difficult to follow. And it all seemed to make much sense! I switched from math to physics and in fact for several years to come I lived under the impression that gravity had been quantized and it wouldn’t take long until somebody calculated exactly what is inside a black hole and how the big bang works. That, however, never happened. And here we are in 2015, still looking to answer the same questions.

I’ll restrain from making a prediction because predicting when we’ll know the theory for quantum gravity is more difficult than finding it in the first place ;o)

Thursday, July 16, 2009

What is Fundamental?

As previously mentioned, I was recently at the FQXi conference on the Azores. FQXi, the "Foundational Questions Institute," has the mission "to catalyze, support, and disseminate research on questions at the foundations of physics and cosmology." I work at an institute whose research is "devoted to foundational issues in theoretical physics." Fundamental, foundational, basic – what do we mean with that? What should we expect from a fundamental theory? What are the foundational questions? This was one of the questions we discussed at the FQXi conference, and while several of the participants contributed, I don’t want to blame any of them for the following summary.

So what is Fundamental?

A theory is fundamental if it cannot be derived from another, more complete, theory. More complete means the theory is applicable to a larger range. Note that a fundamental theory can be derivable from another theory if both are equivalent to each other (though one could plausibly argue then one should consider both the same theory).

Throughout history, the search and discovery of more fundamental theories in the natural sciences has lead to a tremendous amount of progress. That however is not a guarantee it will continue to be the path to progress. The issue is in the expression “cannot be derived” which could mean three different things:

Cannot be derived, version I: not possible in principle.

It might not be possible because it is not possible. Believers in reductionism think this is not the case for the laws of Nature we presently know: they should all follow from one most fundamental "Theory of Everything." While it is true that reductionism proved to be very useful and we thus have good reasons trying to continue it, there is no knowing the laws of Nature always allow a reduction. We would then be left with layers of theories that describe Nature on various scales that cannot ever be derived from each other, and thus have to be considered equally fundamental. While we presently don’t have evidence for this, it is a self-consistent point of view.

In the previous post on Emergence and Reductionism, I explained this is known as “strong emergence:” Emergent features on a higher level require a theory that cannot be derived from the underlying one. We previously discussed the paper “More really is different,” in which Gu et al offer an example for a system that does have emergent features, but it can be proved these are not derivable from the underlying theory. Granted, the system they consider isn’t particularly natural (see discussion on earlier post), but it gives you an impression of what this case means.

Cannot be derived, version II: not possible in practice

It might not be possible to derive emergent features from a more fundamental theory because of practical constraints. For example, it might take more computing power than we will ever have available, or more time than the lifetime of the universe to do it. It might take infinitely precise knowledge of initial conditions; it would make it necessary to measure parameters more precisely than we can plausibly expect ever; it would take a detector the size of the galaxy; etc etc.

Cannot be derived, version III: not yet possible

We might simply not have a derivation because we are too dumb the current knowledge isn’t sufficient, but we might find a derivation with more research.

Okay, now what is fundamental?

The problem is that at any one time we might not know which of these 3 cases we are dealing with. The exception is if we had an actual proof for the impossibility of a derivation. (But then a proof is only as good as its assumption.) We are thus left with our assessment of the situation, which might change with better understanding of the theories we have. In some cases there is a pretty clear consensus on whether a law is fundamental, in other cases it might not be so clear.

Examples

  • Take for example the Tully-Fisher relation. It relates the luminosity of a spiral galaxy with the 4th power of its rotational velocity. It is a useful heuristic relation, extracted from data, and has predictive power. There is no derivation of that relation; yet I doubt any physicist would argue it is a fundamental law. Instead, with increasing understanding of astrophysical processes, we will finally be able to derive it.
  • Stefan came up with an interesting historical example, the Titius-Bode law according to which the distance of planets to the sun grows exponentially with their order in the sequence. The law works pretty well up to Uranus and fails with Neptune, but the far out planets were not known when the law was suggested. People once thought the planets' orbits are fixed by fundamental principles, but with better understanding about the gravitational interaction, the "law" was downgraded to a "rule," or possibly just a coincidence. Though with further knowledge about the dynamics relevant for the formation of solar systems the approximate validity of the relation might be an "emergent" feature one can expect to approximately be valid.
  • Then there is of course the often discussed question whether it is in principle possible to derive all of biology, psychology, sociology and economics from physics and thus physics is the most fundamental of all sciences. Many physicists believe this to be the case. For that reason, one of my profs used to refer to physics as “the queen of sciences” (physics is a female noun in German). But we are far away from practically achieving such a derivation, and we thus do not actually know which of the three cases of “cannot be derived” we are dealing with. Already at the level of proteins things get murky, and we should be considering the option that indeed biology might be as fundamental as physics in the sense that it cannot be derived - cannot be derived in principle, not ever.

One of the reasons why the first case might apply even though reductionism has worked so well over a large range of scales is that in some areas of science the separation of scales might no longer work, and/or there might be no scale that can be used for separation. In physics typically the scale is energy, and we are used to neglect things that happen at energies much higher (wavelengths much smaller) than what we are probing. We know this is a safe procedure backed up by the framework of effective field theories. In contrast, a system like our societies does not simply have higher level organizations constituted out of smaller elements, such that these smaller elements define the "emergent" properties. Instead, these organizations also act back on the elements that they are built of and change their behaviour.

Coming back to physics, there are of course the questions that are hotly discussed at the front of research today, those asking what is fundamental in our present theories. Can the masses of particles in the Standard Model be derived from a more fundamental theory? Are space and time themselves emergent from an underlying theory (generally expected to marry quantum mechanics with general relativity). Is quantum mechanics fundamental, or can the quantization procedure and the measurement prescription be derived from a more complete theory?

I don’t know. But I really, really want to know.

Aside: Some weeks ago Clifford also wrote about the question what is fundamental, anyway? Since he sent me the link to make sure I don’t miss it, I can’t get away without mentioning it. Clifford is mostly concerned with people who use the label “more fundamental” to mean their work is more relevant. While that might happen, people using superlatives to claim their own work (life, opinion) is “more this” or “more that” than others’ is hardly remarkable, and certainly not specific to theoretical physics. The other point Clifford makes is that “Nature recycles good ideas,” meaning that the framework of fundamental theories can often also be found to be useful in non-fundamental areas - and the other way 'round. It is an interesting point, but it addresses more the question where one can find inspiration, not what is actually fundamental.

Bottomline

A theory is fundamental if it cannot be derived from a more complete theory, yet there are different reasons for why we may not be able to derive it: It might not be possible in principle, it might not be possible in practice, or we might not yet have the sufficient knowledge to do it. In general, we do not know which case we are dealing with. Misjudgement of the situation can waste a lot of time and hinder progress. If we wrongly believe a property is not fundamental, we risk searching forever for a more fundamental explanation that doesn't exist. On the other hand, if we believe something is fundamental even though it isn't, our understanding of Nature will remain limited. What is sure though is that understanding always starts with a question.

Monday, October 13, 2014

Does Loop Quantum Cosmology make the black hole information loss problem worse rather than better?

Image Source: Flickr.

Martin Bojowald is one of the originators of Loop Quantum Cosmology (LQC), a model for the universe that makes use of the quantization techniques of Loop Quantum Gravity (LQG). This description of cosmology takes into account effects of quantum gravity and has become very popular during the last decade, because it allows making contact to observation.

The best known finding in LQC is that the Big Bang singularity, which one has in classical general relativity, is replaced by a bounce that takes place when the curvature becomes strong (reaches the Planckian regime). This in return has consequences for example for the spectrum of primordial gravitational waves (that we still hope will at some point emerge out of the foreground dust).

Now rumors reached me from various sources that Martin lost faith that Loop Quantum Cosmology is a viable description of our universe, and indeed he recently put a paper out on the arxiv detailing the problem that he sees.
Information loss, made worse by quantum gravity
Martin Bojowald
arXiv:1409.3157
Loop Quantum Cosmology, to be clear, was never claimed to be strictly speaking derived from Loop Quantum Gravity, though I have frequently noticed that the similarity of the names leads to confusion in the popular science literature. LQC deals with a symmetry-reduced version of LQG, but this symmetry reduction is done before the quantization. In practice this means that in LQC one first simplifies the universe by assuming it is homogeneous and isotropic, and then quantizes the remaining degrees of freedom. Whether or not this treatment leads to the same result that one would get by taking the fully quantized theory and looking for a solution that reproduces the right symmetries is controversial, and to my knowledge this question has never been satisfactorily settled.

Be that as it may, from my perspective and from that of most people working on the topic, LQC is a phenomenological model that is potentially testable and thus interesting in its own right, regardless of its connection to LQG.

It has become apparent however during the last years that if one takes into account perturbations around the homogeneous and isotropic background in LQC then one finds something peculiar: the space-time around the bounce loses its time-coordinate, it becomes Euclidean and is thus just space without time. We discussed this earlier here.

Now the time-coordinate in the space-time that we normally deal with plays a very important role, which is that it allows us to set an initial condition at one moment in time, and then use the equations of motion to predict what will happen at later times. This so called “forward evolution” is a very typical procedure for differential equations in physics, so typical that we often do not think about it very much. Thus I have to emphasize the relevant point is that to determine what happens at some point in space-time one does not have to set an initial condition on a space-time boundary around that point, which would necessitate knowing what happens at some moments into the future, but it is sufficient to know what happened at some moment in the past.

This important property that allows us to set initial conditions in the past to predict the future is not something you get for free in any space-time background. Space-times that obey this property are called “globally hyperbolic”. (Anti-de Sitter space is the probably best known example of a space-time that is not globally hyperbolic, thus the relevance of the boundary in this case.)

In his new paper Martin now points out that if space-time has regions that are Euclidean then the initial value problem becomes problematic. It is then in fact no longer possible to predict the future from a past initial condition. For the case of the Big Bang singularity being replaced by a Euclidean regime, this does not matter so much because we would just set initial conditions after this regime has passed and move on from there. But not so with black holes.

The singularity inside black holes is in LQC then also replaced by a Euclidean regime. This regime only forms in the late stages of collapse and will eventually vanish after the black hole has evaporated. But there being an intermediate Euclidean region has the consequence that whatever is the outcome of the evaporation process depends on the boundary conditions surrounding the Euclidean region. With the intermediate Euclidean region, one can no longer predict from the initial conditions of the matter that formed the black hole what is the outcome of black hole evaporation.

In his paper Martin writes that this makes the black hole information loss considerably worse. The normal black hole information loss problem is that the process of black hole evaporation seems to be irreversible and thus in particular not unitary. The final state of the evaporation is always thermal radiation, regardless of what formed the black hole. Now with the Euclidean region the final state of the black hole evaporation depends on some boundary condition that is not even in principle predictable. We have thus gone from not unitary to not deterministic!

Martin likens this case to that of a naked singularity, a singular region that (in contrast to the normal black hole singularity which is hidden by the horizon) is in full causal contact with space-time. A singularity is where everything ends, but it is also where anything can start. The initial value problem in a space-time with a naked singularity is similarly ill-defined as that in a space-time region with a Euclidean core, Martin argues.

I find this property of black holes in LQC not as worrisome as Martin. The comparison to a naked singularity is not a good one because the defining property of a singularity is that one cannot continue through it. One can however continue through the Euclidean region, it’s just that one needs additional constraints to know how. In fact I can see that what Martin thinks is a bug might be a feature for somebody else, for after all we know that time-evolution in quantum mechanics seems to be non-deterministic indeed.

But even leaving aside this admittedly far-fetched relation, the situation that additional information is necessary on some boundary to the future is not unlike that of the mysterious “stretched horizon” in black hole complementary. Said stretched horizon somehow stores and later releases the information of what fell through it. If the LQC black hole is supposed to solve the black hole information problem, then the same must be happening on the boundary of the Euclidean region. And, yes, that is a teleological constraint. I do not see what theory could possibly lead to it, but I don’t see that it is not possible either.

In summary, I find this development more interesting than troublesome. In contrast to non-unitarity, having a Euclidean core is uncomfortable and certainly unintuitive, but not necessarily inconsistent. I am very curious to see what the community will make out of this -- and I am sure we will hear more about this in the soon future.

Tuesday, May 14, 2019

Quantum Mechanics is wrong. There, I’ve said it.

[Image: needpix.com]


So, you have developed a new theory of quantum mechanics. That is, erm, nice. No, please don’t show it to me. I’m almost certainly too stupid to understand it. You see, I have only a PhD in physics. All that math has certainly screwed up my neural wiring. Yes, I am sorry. But I have a message for you from the depth of abstract math: We know that quantum mechanics is wrong.

Seriously, it’s wrong. It’s as wrong as Newtonian gravity is wrong, as hydrodynamics is wrong, and as spherical cows are wrong. Quantum mechanics is an approximation. It works well in some cases. It does not work well in other cases.

You see, in quantum mechanics we give quantum properties to particles. But we know that, strictly speaking, the interactions between these particles must also have quantum properties. If we give these interactions quantum properties, we call that “second quantization.” It is not used in quantum mechanics. Second quantization results in a larger mathematical framework called “quantum field theory”. The Standard Model of particle physics is a quantum field theory. Sometimes we use the word “quantum theory” to refer to both, quantum mechanics and quantum field theory together.

Moving from quantum mechanics to quantum field theory is more than just a change of name. Quantum field theories inherit many properties from quantum mechanics: Entanglement, uncertainty, the measurement postulate. But they bring new insights – and also new difficulties.

The best known insight brought by quantum field theory is that particles can be created and destroyed, and that each particle has an anti-particle (though some particles are their own anti-particles). Another remarkable consequence of quantum field theory is that the strength of interactions between particles depends on the energy by which one probes the interaction. The strong nuclear force, it turns out, becomes weaker at high energies, an odd behavior that is known as “asymptotic freedom.”

The probably best known difficulty of quantum field theories is that many calculations result in infinity. Infinity, however, is not a very useful prediction. Such results therefore have to be dealt with by a procedure called “renormalization,” whose purpose is to suitably subtract infinity from infinity to get a finite remainder. No, there is nothing wrong with that. It works just fine, thank you.

Quantum field theories lead to other complications. For example, we know how to calculate what happens if two electrons bump into each other and create a bunch of new particles. This is called a “scattering event”. But we don’t know how to calculate what happens if three quarks stick together and form a proton. Well, we do know how to put such calculations on super-computers in an approximation called “lattice QCD”. But really we don’t have good mathematical tools to handle the case. At least not yet.

But let us come back to quantum mechanics. You can use this theory to make predictions for any experiment where the creation and destruction of particles does not play a role. This is the case for all your typical quantum optics experiments, Bell-type tests, quantum cryptography, quantum computing, and so on. It is not merely a matter of doing experiments at low energy, but it also depends on how sensitive you are to the corrections coming from quantum field theory.

So, yes, quantum mechanics is technically wrong. It’s only an approximation to the more complete framework of quantum field theory. But as the statistician George Box summed it up “All models are wrong, but some are useful.” And whatever your misgivings are about quantum mechanics, there is no denying that it is useful. 

Saturday, July 28, 2007

Consistency

I am guilty of frequently using physics speech in daily life, an annoying habit I also noticed among many of my colleagues [1]. You'll find me stating "My brain feels very Boltzmannian today", or "The customer density in this store is too high for my metastable mental balance". I have a friend who calls Chinese take out "the canonical choice" and another friend who, when asked whether he had made a decision, famously explained "I don't yet want my wave-function to collapse". My ex-boyfriend once called it "the physicist's Tourette-syndrome" [2].

One of my favourite physics-speech words is self-consistent. Self-consistency is tightly related to nothing. You know, that "nothing" that causes your wife to conclude her whole life is a disaster, we're all going to die in a nuclear accident, her glasses vanished (again!), and btw that's all your fault (obviously). But if you ask her what's the matter. Well, nothing.
    "There's nothing I hate more than nothing
    Nothing keeps me up at night
    I toss and turn over nothing
    Nothing could cause a great big fight
    Hey -- what's the matter?
    Don't tell me nothing."

~Edie Brickell, Nothing


1. Self-consistent

Science is our attempt to understand the world we live in. We observe and try to find reliable rules upon which to build our expectations. We search for explanations that are useful to make predictions, a framework to understand our environment and shape our future according to our needs. If our observations disagree with our rules, or observations seemingly disagree with each other (I swear I left my glasses in the kitchen), we are irritated and try to find a mistake. Something being in contradiction with itself [3] is what I mean with not self-consistent (What's the matter? - Nothing!).

On a mathematical basis this is very straight forward. E.g. If you assume my mood is given by a real valued continuous function f on the compact interval [now, then] with f(now)f(then) smaller than 0, this isn't self-consistent with the expectation it can do so without having a zero [4]. For more details on my mood, see sidebar.

Self-consistency is a very powerful concept in theoretical physics: if one talks about a probability, that probability better should not be larger than one. If one starts with the axioms of quantum mechanics, it's not self-consistent to talk about a particle's definite position and momentum. The speed of light being observer independent is not compatible with Galileo invariance and the standard addition law for velocities. Instead, self-consistency requires the addition law to be modified. This lead Einstein to develop Special Relativity.

A particularly nice example comes from multi-particle quantum mechanics, where an iterative approach can be used to find a 'self-consistent' solution for the electron distribution e.g. in a crystal or for an atom with many electrons (see self-consistent field method or Hartree-Fock method). A state of several charged particles will not be just a tensor product of the single particles, since the particles interact and influence each other. One starts with the tensor product as a 'guess' and applies the 'rules' of the theory. That is, by solving the Schrödinger equation with the mean- field potential which effectively describes the interaction, a new set of single particle wave functions can be computed. This result will however in general not agree with the initial guess: it is not self-consistent. In this case, one repeats the procedure with using the result as an improved guess. Given that the differential equations behave nicely, this iterative procedure leads one to find a fixed point with the properties that the initial distribution agrees with the resulting one: it is self-consistent.

A similar requirement holds for quantum corrections. A theory that is subject to quantum corrections but whose initial formulation does not take into account the existence of such extra terms is strictly speaking not self-consistent (see also the interesting discussion to our recent post on Phenomenological Quantum Gravity).

There are some subtleties one needs to consider, most importantly that our knowledge is limited in various regards. Self-consistency might only hold under certain assumptions or in certain limiting regimes, like small velocities (relative to the speed of light), large distances (relative to the Planck length) or at energies below a certain threshold. Likewise, not being self-consistent might be the result of having applied a theory outside these limits (typically, using an expansion outside a radius of convergence). In some cases (gravitational backreaction), violations of self-consistency can be negligible.

However, one might argue if it is possible at all to arrive at such a disagreement then at least one of the assumptions was unnecessary to begin with, and could have been replaced by requiring self-consistency. Unfortunately, this is often more easily said than done -- physics is not mathematics. We rarely start with writing down a set of axioms which one could check for self-consistency. Instead, in many cases one starts with little more than a patchwork of hints, and an idea how to connect them. Self-consistency in this case is somewhat more subtle to check. My friends and I often kill each others ideas by working out nonsensical consequences. Here, at least as important as self-consistency is that a theory in physics also has to be consistent with observation.

2. Consistent with Observation

The classical Maxwell-Lorentz theory is self-consistent. However, it is in disagreement with the stability of the atom. According to the classical theory, an electron circling around the nucleus should radiate off energy. The solution to this problem was the development of quantum mechanics. The inconsistency in this case was one with observation. Without quantizing the orbits of the electron, atoms would not be stable, and we would not exist.

This requirement is specific to sciences that describe the real world out there. Such a theory can be 'wrong' (not consistent with observation) even though it is mathematically sound. Sometimes however, these two issues get confused. E.g. in a recent Discover issue, Seth Lloyd wrote:


    "The vast majority of scientific ideas are (a) wrong and (b) useless. The briefest acquaintance with the real world shows that there are some forms of knowledge that will never be made scientific [...] I would bet that 99.8 percent of ideas put forth by scientists are wrong and will never be included in the body of scientific fact. Over the years, I have refereed many papers claiming to invalidate the laws of quantum mechanics. I’ve even written one or two of them myself. All of these papers are wrong. That is actually how it should be: What makes scientific ideas scientific is not that they are right but that they are capable of being proved wrong."

~Seth Lloyd, You know too much


The current issue now had a letter in reply to this article:

    "I was taken aback by Seth Lloyd's assertion that "99.8 percent of ideas put forth by scientists are [probably] wrong" and even more so by his statement that "of the 0.2 percent of ideas that turn out to be correct ... [t]he great majority of them are relatively useless." His thesis omits a basic trait of what we call science -- that it is a continuous fabric, weaving all provable knowledge together [...] we do science for a science sake, because a fundamental principle of science is that we never know when a discovery will be useful"

~Eric Fisher, Springfield, IL.


Well, the majority of my scientific ideas are definitely (a) wrong and (b) useless, but these usually don't end up in a peer review process. However, the reply letter apparently referred to the word 'correct' as 'provable knowledge', and to science as the 'weave' of all that knowledge. It might indeed be that the mathematical framework of a theory that is not consistent with observation turns out to be useful later but that doesn't change the fact that this idea is 'wrong' in the meaning that it does not describe nature. Peer review today seems to be mostly concerned with checking self-consistency, whereas being non-consistent with observation is ironically increasingly tolerated as a 'known problem'. Like, the CC being 120 orders of magnitude too large is a known problem. Oohm, actually the result is just infinity. But, hey, you've turned your integration contour the wrong way, the result is not infinity, but infinity + 2 Pi.

The requirement of consistency with observation was for me the main reason to chose theoretical physics over maths. The world of mathematics, so I found, is too large for me and I got lost in following runaway thoughts, or generalizing concepts just because it was possible. It is the connection to the real world, provided by our observations, that can guide physicists through these possibilities and lead the way. (And, speaking of observations and getting lost, I'd really like to know where my glasses are.)

3. Self-contained

Unlike maths, theoretical physics aims to describes the real world out there. This advantageous guiding principle can also be a weakness when it comes to the quantities we deal with. Mathematics deals with well defined quantities whose properties are examined. In physics one wants to describe nature, and the exact definitions of the quantities are in many cases subject of discussion as well. Consider how our understanding of space and time has changed over the last centuries!

In physics it has often happened that concepts of a theory's constituents only developed with the theory itself (e.g. the notion of a tensor or the Fock-space). As such it happens in physics that one can deal with quantities even though the framework does not itself define them. One might say in such a case the theory is incomplete, or not self-contained.

Due to this complication, I've known more than one mathematician who frowned upon approaches in theoretical physics as too vague, whereas physicists often find mathematical rigour too constraining, and instead prefer to rely on their intuition. Joe Polchinski expressed this as follows:

    "[A] chain of reasoning is only as strong as its weakest step. Rigor generally makes the strongest steps stronger still - to prove something it is necessary to understand the physics very well first - and so it is often not the critical point where the most effort should be applied. [A]nother problem with rigor [is]: it is hard to get it right. If one makes one error the whole thing breaks, whereas a good physical argument is more robust."

~Joe Polchinski, Guest Post at CV


When it comes to formulating an idea, physicists often set different priorities than mathematicians. In some cases it might just not be necessary to define a quantity because one can sit down and measure it (e.g. the PDFs). Or, one can just leave a question open (will be studied in a forthcoming publication) and get a useful theory nevertheless. All of our present theories leave questions open. Despite this being possible, it is unsatisfactory, and the attempt to make a theory self-contained has lead to many insights throughout the history of science.

Newton's dynamics deals with forces, yet there is nothing in this framework that explains the origin of a force. It contains masses, yet does not explain the origin of masses. Maxwell's theory provides an origin of a force (electromagnetic). It has a source term (J), yet it does not explain the dynamics of the source term. This system has to be closed, e.g. with minimal coupling to another field whose dynamics is known. The classical Maxwell-Lorentz theory does this, it is self-contained and self-consistent. However, as mentioned above, this theory is not consistent with observation. Today we know the sources for the electromagnetic field are fermions, they obey the Dirac equation and Fermi statistic. However, if you look at an atom close enough you'll notice that quantum electrodynamics alone also isn't able to describe it satisfactory...

Besides the existence of space and time per se, the number of space-time dimensions is one of these open questions that I find very interesting. It has most often been an additional assumption. An exception is string theory where self-consistency requires space-time to have a certain number of dimensions. However - if it also contains an explanation why we observe only three of them, nobody has yet found it. So again, we are left with open questions.

4. Simple and Natural [5]

The last guiding principle that I want to mention is simplicity, or the question whether one can reduce a messy system of axioms and principles to something more simple. Is there a way to derive the parameters of the standard model from a single unified approach? Is there a way to derive the axioms of quantization? Is there a way to derive that our spacetime has dimension three, or Lorentzian signature?

In my opinion, simplicity is often overrated compared to the first three points I listed. We tend to perceive simplicity as elegance or beauty, concepts we strive to achieve, but these guidelines can turn out to be false friends. If you can find your glasses, look around and you'll notice that the world has many facettes that are neither elegant nor simple (like my husband impatiently waiting for me to finish). Even if you'd expect the underlying laws of nature to be simple, you'll still have to make the case that a certain observable reflects the elementary theory rather than being a potentially very involved consequence of a complex dynamical system, or an emergent feature. A typical example are the average distances of planets from the sun, a Sacred Mystery of the Cosmos that today nobody would try to derive from a theory of first principles (restrictions apply).

Also, we tend to find things simpler the more familiar we are with them, up to the level of completely forgetting about them (did you say something?). E.g. we are so used to starting with a Lagrangian that we tend to forget that its usefulness rests on the validity of the action principle. It is also quite interesting to note that researchers who are familiar with a field often find it 'simple' and 'natural'... I therefore support Tommaso's suggestions to renormalize simplicity to the generalized grandmother.

In this regard I also want to highlight the argument that one can allegedly derive all the parameters in the standard model 'simply' from today's existence of intelligent life. Notwithstanding the additional complication of 'intelligent', could somebody please simply explain 'existence' and 'life'?

Bottomline

Much like classical electrodynamics, Einstein's field equations too have a source term whose dynamics one needs to know. The system can be closed with an equation of state for each component. This theory is self-consistent [6], and it is consistent with all available observations. It reaches its limits if one asks for the microscopic description of the constituents. The transition from the macro- to the microscopic regime can be made for the sources of the gravitational field, but not also for the coupled gravitational field (oh, and then there's the CC, but this is a known problem).

Two theories that yield the same predictions for all observables I'd call equivalent (if you don't like that, accept it as my definition of equivalence.) But our observations are limited, and unlike the case of classical electrodynamics not being consistent with the stability of the atom, there is presently no observational evidence in disagreement with classical gravity.

For me this then raises the question:
    Is there more than one theory that is self-consistent, self-contained and consistent with all present observations?

In a recent comment, Moshe remarked:"To paraphrase Ted Jacobson, you don't quantize the metric for the same reason you don't go about quantizing ocean waves." That sounds certainly reasonable, but if I look at water close enough I will find the spectral lines of the hydrogen atom and evidence for its constituents. And their quantization. To me, this just doesn't satisfactory solve the question what the microscopic structure of the 'medium', here space-time, is.

And what have we learned from all this...?

Let me go back to the start: If you ask a question and the answer is 'Nothing', you most likely asked the wrong question, or misunderstood the answer.

Ah... Stefan found my glasses (don't ask).

See also: Self-Consistency at The Reference Frame


[1]This habit is especially dominant -- and not entirely voluntarily -- among the not native English speakers, whose vocabulary naturally is most developed in the job related area.
[2] Unintentional cursing and uttering of obscenities, called Coprolalia, is actually only a specific feature of the Tourette syndrom.

[3] However, some years ago I was taught the word 'self-consistency' in psychology has a different meaning, it refers to a person accumulating knowledge from his/her own behaviour. A person whose thoughts and actions are in agreement and not in contradiction is called 'clear'. (At least in German. I couldn't find any reference to this online, and I'm not a psychologist, so better don't trust me on that.).
[4] See:
Bolzano's theorem.
[5] "Woman on Window", by F.L. Campello.
For more, see here.
[6] Note that this theory is self-consistent at arbitrary scales as long as you don't ask for the microscopic origin of the sources.



TAGS: , ,

Wednesday, March 21, 2007

Science and Democracy III

[I was reminded that I’ve promised repeatedly to continue the previous posts Science and Democracy I and II. To my own surprise I found an almost finished draft about the danger of using marketplace tactics in scientific research, and I added some recent comments out of the blogoshpere to underline my arguments.]

I vividly recall the first thing my supervisor told me when I was an undergrad: "You have to learn how to sell yourself." Since then I have repeatedly been given well meant career advises how to survive on the scientific marketplace (most of which I ignored, but I’m still around, so I guess I’m not doing too badly).

Many of my friends and colleagues in physics regard these marketplace tactics as an annoying but necessary part of the job. To begin with, this concerns me because I feel that there is a gap between how science is, and how it should be – and an unnecessary gap in addition. But more importantly, the application of economical considerations to scientific research is inappropriate, and the reason why I did not take these career advises is that I don’t want to support strategies that will hinder scientific progress on the long run. So, if you were hoping for some career advises, you're on the wrong blog.

Though comparisons between science and economics often have a grain of truth in them, they are doomed to fail when extended naively. Whether or not you believe in the infallibility of the 'Invisible Hand' [1], scientific theories are not sold like candy bars. If one uses an economical model to analyze the dynamics of research programs, one has to be aware of the limitations of this analogy.

I: The Marketplace of Ideas

The 'marketplace of ideas' is often claimed to act as a self-regulating mechanism that ensures progress in science. It is based upon the believe that all scientific theories when made accessible to the public compete freely among each other, until it eventually turns out which idea is the best description of nature.

Being an optimist, I have no doubts that this works if one looks at the history of mankind over centuries when nature is the ultimate judge on our scientific endeavors. However, there is no reason to believe that this automatically works on shorter time periods as well. On a time scale where we do not have nature to judge (typically a couple of years, maybe up to decades - that is how long grants and employments last), the scientific community is its own judge. The obvious difference to the economical marketplace is that we do not offer our ideas ‘for sale’ to a neutral target group, and depending on whether it is bought or not our product is a success or a disaster.

Ia: The Measurement Problem

No, we are selling our theories inside our own community. And our demand for products can easily be biased if the competitive pressure is high. The situation is significantly worsened by an increasing specialization into many sub-fields and a lack of communication between these fields. Needless to say, the genuine enthusiasm that researchers have for their own field does not improve neutral judgement. If you want to use the analogy to the economical marketplace here, you’d expect it to work even if products are only sold to company managers [2].

This difficulty to find criteria to judge on research programs might not have been a major issue in the previous decades, but it becomes increasingly important if

    a) the community grows to a complex system whose dynamics is little understood (E.g. the increasing influence of 'fashionable topics' is a typical sign for a non-linear feedback effect, the emergence of sub-fields with their own group dynamics is a sign for self-organization)
    b) changes in the sociological, cultural and technological context require adjustment of criteria
    c) financial and peer pressure endangers neutral judgement
    d) timescales that are set through (inappropriate) external constraints.

In other words, it’s a 21st century issue. With increasing complexity, we are left with a decreasing number of people who have an overview on the whole ‘marketplace’. Little people are funded independently of their projects, so their opinions are biased in favor of their own research. Under such circumstances, the ‘marketplace of ideas’ will eventually result in a small number of approaches caught in feedback loops of increasing separation.

As Thomas Dent remarked at 4:27 PM, June 29, 2006:

    It should be obvious that there is no theoretical physics analog of capitalism or the free market. In capitalism there is profit which can be measured objectively, and the one who can make profit wins, the one who cannot must get out of the game. […] Since there is no objective measure of success in theoretical physics, there can never be a free market. Simple.

I would add: there is no such obvious measure on the time scale relevant for funding today. But obvious or not, measures have to be applied, and are applied, and the least we can do is to chose them wisely. Being scientists, it should not be so hard for us to find out how science works best, and to analyze whether the current conditions are optimal.

Ib: Primary goals and Secondary criteria

The primary goal: to support the most promising approaches and researchers, is of little help when you are faced with a 3 inch pile of application documents. Instead, one commonly uses derived secondary criteria that have shown to be useful. There is nothing to object to this procedure, except that the validity of secondary criteria has to be readjusted every now and then. In a time like ours, when the sociological and technological environment changes rapidly, neglect of questioning and re-adjusting applied secondary criteria can result in misleading feedback effects and sub-optimal selection processes.

The best known example might be the citation index and number of publications. These criteria are of course correlated with the originality and quality of the research, but whenever possible, one should ask for primary goals to be met. (I am not telling you this because it is something new, or because I think people in hiring committees are stupid, but to make the matter less abstract.) Other secondary criteria that have grown important over the last decades are e.g. previous employment at well-known institutions, or classifiable work on mainstream topics.

There is an obvious danger in just rewarding those who meet secondary requirements. If these criteria do not exactly match the primary goals, one promotes tactics that are sub-optimal for scientific progress but optimal for career building (see section 'Survival of the Fittest'). If you combine that with the non-linear feedback effect in complex systems, things can easily go seriously wrong.

An issue related to the necessary re-adjustment of secondary criteria to primary goals is to guarantee fairness on the marketplace. The 'Invisible Hand' always needs to be balanced by politics to ensure the marketplace is really 'free' -- this is one of the earliest lessons we have learned from industrialization. If we want the ‘marketplace of ideas’ to optimize progress in science, we have to ensure that every idea gets a chance, irrespective of its origin [3] - the matter of origin plays no role for the question whether an idea is worth supporting.

Ic: Risky Research

Another important point is that supporting risky start-ups is one of the most relevant factors for progress. Unfortunately, this factor is severely neglected by present funding strategies. Sounds familiar? Okay, okay, it’s not my idea:


"Do you want a revolution in science? Do what businesspeople do when they want a technological revolution: Just change the rules a bit […] Create some opportunities for high-risk/high-payoff people […] The technological companies and investment banks use this strategy. Why not try it in academia? The payoff could be discovering how the universe works."

~ Lee Smolin, The Trouble with Physics (p. 331)



Risk averseness is a rather unsurprising consequence of insecurity caused by a lack of communication in a community falling apart into sub-fields. It is also supported by chronically short resources (if we hire anybody, then someone who works on what I find interesting), short-term funding (it takes time to work out new ideas, for more info see e.g. Temporary Display and comments to this post), and by falling for the derived secondary requirement 'If she's interested in what I do, she must be intelligent.'

In the absence of a final judgement by nature on our approaches, it is very short-sighted to discard alternative options. However convinced I am of my own research project, I always have to acknowledge the possibility that it turns out to be a dead end. As Albert Einstein said so nicely "Mathematics are well and good but nature keeps dragging us around by the nose." In the London debate, Nancy Cartwright underlined the need to keep doors open by referring to J.S. Mill's essay 'On Liberty'. She argues that in the absence of nature's judgement the smart thing to do is to not prematurely discard alternative options

    "We need to allow as much liberty as possible (for people in designing their lives) because we don't know what is the best way (to live). And that's in part because we don't know what are all the good alternatives to chose from."


II. Product Placement

Presenting our research results to colleagues is an essential part of our job, in written form as well as in seminars, talks, and discussions. Clearly presented arguments, and well structured seminars are definitely beneficial to progress. However, as with many things in life, it is a matter of balance. Advertisement should not become more important than content. The most entertaining presentation can not make an idea better than it is, and scientific arguments have to remain as honest as possible – even if this means drawing attention to the flaws of the product.

For example, Thomas Larsson remarked

    At 5:53 AM, March 17, 2007,
    There are some good things about the string/LQG conflict, though. Without it, I would not know about the limits of the string black hole prediction, nor would I know that LQG quantization does not work for the harmonic oscillator.

and illustrates with that a problem that arises when researchers feel the pressure to advertise their own work. Today, many praise results in a rather unbalanced way – not because they don’t know better, but because they have to compete with a large number of people. If you put a paper out and don’t have a prominent co-author, a catchy title and exaggerated claims is the way to get others to read it. This tactic is okay for fish sales on mediterranean markets, but it is very dangerous to the standard of scientific research. It leads to rather uncritical status reports in which problems are either not mentioned, or downplayed (and if this shortcoming is pointed out, the author will claim that the problem is obvious and widely known.)

Whether published articles are balanced crucially depends on the referee process [4]. One could say a lot about peer review, but to say the least, it doesn’t always work as it should, and many reports are not as objective as they ought to be. An example that I have repeatedly witnessed myself: when it comes to numerical simulations, it is common practice to point out where the model fits the data very good, and just not to mention the problematic observables. Most often, numerical simulations are hard to check, even if the code is available, and the not-so-good results just don’t get published.

Though this is not strictly speaking wrong, it is just not good practice as it is exactly understanding the failure of an approach that could lead to improvement. However, those scientists who elaborate on difficulties and drawbacks risk being understood as negative, or maybe just not exciting enough, and cause problems for themselves (and probably get the well meant career advise to better sell themselves): Here we have another gap between what would be beneficial for scientific progress (primary goal: understand model), and what is beneficial for the scientific career (secondary goal: hide bugs or declare them as feature).



Now that I think about it, why not include a blurb paragraph to papers with warnings. Like ‘Possible side effects might contain decaying vacua, ghost fields and tachyons.’ Or ‘Do not use this model in Lorentzian signature, and not after the electroweak phase transition. If you consider using it in more than three dimensions, or together with matter fields, please consult a doctor.’

:-) I know, I’m being silly. I apologize, it is far easier to retreat to sarcasm than to come up with constructive criticism.

Were was I?

Uhm, this is another example where marketplace tactics fail in scientific research. We don’t want to sell our theories to as many people as possible and optimize the citation index, but we want to optimize the quality and usefulness of publications.

Another excellent example that shows how advertisement can promote scientific nonsense when secondary criteria (here: holding patents) are in conflict with primary goals (here: quality of research), can be found at the post Micro Black Day.

III. Survival of the Fittest

The survival of the fittest is another catchy phrase (often used by those who profit from the current system) to claim that a natural selection process ensures progress in scientific research. The irony is that those who argue such actually explain why the system fails.

Survival of the fittest doesn’t mean survival of the strongest, the best looking or the most intelligent. It means literally, survivors are those who ‘fit best’. Survivors are those who adapt the behaviour that minimizes existential conflict with the environment.

Now, ask yourself, what is this environment in the context of scientific research? Well, it is our own community with the selection criteria that we apply. If these criteria are not optimal for scientific progress, we don’t only have the possibility but the duty to change it!

The optimization implied in the ‘survival of the fittest’ crucially depends on the environment and available resources. Whether you like the subtitle of Lee's book or not, he makes the important point that we have to ask how science works best – how it works now and here, how it works in this century, in this sociological and cultural environment - and whether the presently applied selection criteria are indeed optimal for progress. Whether the fitness that we reward is actually the fitness that we need. Whether our secondary criteria agree with the primary goals.

We have to blame ourselves if we accept the current conditions even though we know they are not optimal.

    Amara Graps: At 2:51 AM, March 10, 2007
    One reason why the current system has been going on for so long is that scientists are a mild-mannered bunch and are passionate about their work. They are prone to self-abuse to pursue those passions too, being willing to absorb the most degrading conditions.

Repeatedly, I have met colleagues who agree that the situation sucks, but they shrug shoulders and say, that’s just the marketplace. Where does it come from, this believe that passivity is a guarantee for progress?

    amused: Mar 17th, 2007 at 1:17 am
    Of course, that’s hardly a new point in these discussions, and the standard response is to shrug ones shoulders and say “oh well, that’s just market forces”. Which is true, but it’s also relevant to ask whether it is in the best interests of physics. Hopefully it’s not too controversial to suggest that the interests of physics in the long term are best served by ensuring as much as possible that jobs go to the “best” people, regardless of their preferred research topics.

We are scientists. We should be able to analyze the present situation, and to draw consequences. Science is not coming to an end if we fail to meet the challenges that the increasing complexity of our field has brought. But we run in danger to reproduce the failures of the economic marketplace: bubbles of nothing, that are a waste of time, money and energy.

If left without attention, the naĂ¯ve believe that the marketplace will make things right ‘somehow’ can seriously hinder progress. Nature might have supported an approach that failed too early – because it wasn’t advertised well enough, or because the capital investment was simply insufficient to allow it to compete.

Bottomline

There are important differences between the economical and the scientific marketplace. The most obvious ones being the absence of a neutral measure (like profit), and the pitfalls of advertisement.

Currently the ‘marketplace of ideas’ works anything but optimal. Times have changed rapidly, and our community has grown significantly. These changes need to reflect in our organizational structure as well, or we run in danger of getting stuck in a dead end.

And it is easy enough to improve the situation:


  1. Question and doubt. Ask yourself whether the realized strategies are optimal for scientific progress, and if you don't think so, don't shrug shoulders. Don't accept criteria you have been taught are right without taking into account that times have changed.

  2. Analyze. Peer pressure, intense competition, short resources, project-dependent funding and short-term employment favours mainstream, conservative and low-risk work. Be aware of that. Remind yourself and colleagues that 'Good physics has to be open, critical, and responsive'. Research has shown that simply reminding people to think rationally influences their decisions.

  3. Trust yourself. Don't work on topics that you don't genuinely believe are relevant because you are afraid of your reputation. If this work is unavoidable, criticise - even if you are defeated, you make a contribution to science. (Hey - I told you, you're not getting career advises on this blog.)


Hmmm...

It seems, this piece got quite lengthy...

One could write books about it...




Footnote [1]: The 'Invisible Hand' was indroduced by Adam Smith in his book 'The Wealth of Nations' (1776) to describe the self-regulation of the marketplace. From Wikipedia: Many economists claim that the theory of the Invisible Hand states that if each consumer is allowed to choose freely what to buy and each producer is allowed to choose freely what to sell and how to produce it, the market will settle on a product distribution and prices that are beneficial to the entire community. Adam Smith already pointed out that the Invisible Hand's regulation mechanism alone does not guarantee the well-being of the society and needs to be balanced by govenmental guidance "[...] uniformity of [the employee's] stationary life naturally corrupts the courage of his mind [..] His dexterity at his own particular trade seems, in this manner, to be acquired at the expence of his intellectual, social, and martial virtues. But in every improved and civilized society this is the state into which [...] the great body of the people must necessarily fall, unless government takes some pains to prevent it." Nevertheless, this metaphor is often abused to praise the merits of capitalism without given sufficient credits to its limitations. [Back]

Footnote [2]: In addition, there is also the question how our research is presented to the public - who after all pays us to explore the frontiers of our knowledge. This is an important point on its own but should not be mixed up with the question how the community selects promising researchers and research programs. Most people are crucially aware that it requires an appropriate education to judge on the value of very recent developments, and will rely on expert’s opinions for a good reason. The public is neither dumb nor ignorant. I welcome it very much that in the last decades - maybe starting with Hawking’s Brief History of Time - theoretical physics has become more accessible to the public. The resulting discussions of our research among non-experts are regarded by some scientists with concern and skepticism. I am sure it is only a matter of time until our community gets used to this attention and learns how to deal with this kind of feedback. I myself am perfectly sure this communication is inspiring for both sides - and one of the reason why I maintain this blog. [Back]

Footnote [3]: To give a concrete example, research papers should not be judged upon by the author. Researchers should not be selected because of the institutions they have connections to, or the country of origin. Conference invitations should not be made to famous people for the only reason that their name does attract interest – A scientific conference is not a rock concert. [Back]

Footnote [4]: At least one should be careful enought to use 'could' instead of 'does' and 'might' instead of 'will'. You can learn about the importance of weasel words here, in case you followed this discussion about this paper. [Back]



TAGS: , ,