Adversarial collaborations were originally proposed as a solution to solve the arguments between the believers in the extrasensory perception on one side and the sane people on the other side. If you write a paper along with a guy who was abducted by the extraterrestrial aliens and talks to them through telepathy, that will surely lead to your mutual cultural enrichment and the joint work will clarify everything! ;-)
Expect no apologies here: I am mocking the concept of such joint papers because a joint paper assumes – and implies – that the co-authors believe to be peers in some intellectual sense. And a competent scientist simply can't consider a nutcase believing in supernatural phenomena as his intellectual peer. Also, as the Wikipedia article explains, adversarial papers are most possible when they're not needed (when the two sides actually agree about some basic axioms and methodology) and least possible when they're needed (when the differences in the methodology, axioms, and priorities are deep).
Now, Tetragraviton proposes a similar procedure to bring order to the passionate discussions about the interpretation of quantum mechanics; battles between MOND and dark matter; and the existence of de Sitter space vacua within string theory.
Can it work?
As all readers know, my main message is the message of love and peace. And I am really thrilled by the prospect of bringing more love and peace to the world. At an emotional level, I love the adversarial collaboration as a method to end the hatred and wars.
But as a part of the scientific method, I think that adversarial collaborations are just rubbish. Opposing sides simply cannot write a coherent paper they may sign off because their views about fundamental – and most other – things contradict each other. So if there's anything important in the paper, the plan
For the paper to get published, both sides have to sign off on it: they both have to agree that everything in the paper is true.just cannot be realized. It contradicts mathematical logic.
This prevents either side from cheating, or from coming back later with made-up objections: if a point in the paper is wrong, one side or the other is bound to catch it.Many fundamental propositions that the opposing side is making are fundamentally and often demonstrably wrong. So if they say these important things about Nature or the theories, they will simply be wrong from my viewpoint. And vice versa, I guess. At most, we could write true sentences saying "A said that something in Nature obeys XY, B said that something in Nature doesn't obey XY".
That "he said she said" genre wouldn't result in a very interesting paper. You can get the same by gluing their papers together and adding "he said" or "she said" in front of every sentence.
A paragraph says:
One advantage of adversarial collaborations is in preventing accusations of bias. The debate between dark matter and MOND-like proposals is filled with these kinds of accusations: claims that one group or another is ignoring important data, being dishonest about the parameters they need to fit, or applying standards of proof they would never require of their own pet theory. Adversarial collaboration prevents these kinds of accusations: whatever comes out of an adversarial collaboration, both sides would make sure the other side didn’t bias it.What Tetragraviton is missing is that "dishonesty about the parameters" (and basically all other analogous complaints) is a matter of a subjective judgement – or it is at least presented as a matter of subjective judgement by a side that is either being fair or not. But whether it is too dishonest to overlook some data, to claim that some experimenters are wrong, and so on – this is a part of the disagreements. So the sides simply can't find common ground if both of them have at least some internal consistency.
One more magic feature of his adversarial collaborations is claimed to be the following:
Another advantage of adversarial collaborations is that they make it much harder for one side to move the goalposts, or to accuse the other side of moving the goalposts.But this is a deep misunderstanding of what's going on. "Moving the goalposts" is a rhetorical method to negatively describe something done by the antagonist that, from our perspective, seems to weaken his position – or "we" want to make it look bad. But "moving the goalposts" isn't a mortal sin!
Instead, "moving the goalposts" is just an unflattering name for an inseparable part of the scientific method. As long as both sides are pursuing the scientific method, both of them are sometimes moving the goalposts, although they will prefer to describe the exact the same thing using more positive words. The positive way to describe it is to adjust the theories, knowledge, predictions, and expectations to the modified evidence when new evidence arrives.
A scientist has to do such things, otherwise he is blind to the data and evidence – i.e. he is not a scientist.
If and when you believe that the proposition AB is almost certainly true, you may list five pieces of evidence in favor of AB, for example E1,E2,E3,E4,E5. But E5 may be proven wrong. You still have E1,E2,E3,E4 and because you still consider them important, you may keep your belief in AB and you adjust the argumentation, recipes to make progress, methods to test things, and other things.
Meanwhile, your opponent has used F1,F2,F3,F4,F5,F6 and F5,F6 were proven wrong. He also has to move the goalposts. When things are changing, everything that the scientists believe, write in their theories, plan to do, and predict may be modified in general. Some of these modifications may be interpreted as "moving the goalposts". But "moving the goalposts" isn't something that may be "completely banned in science". Theories in science – and all statements that scientists make about science and Nature – have to adapt to the new evidence etc. These changes may sometimes be viewed or interpreted negatively – and the negative interpretation may be fair or unfair. Whether it's fair or unfair will depend on the side and their other assumptions about all the questions.
If the sides are honest at least at some internal level, they "move the goalposts" because they think they're forced to do it by something that they consider more important – or more guaranteed to be true – than the "ban on the moving the goalposts". In other words, your antagonist may think that the "end justifies the means". We have often used the phrase "the end justifies the means" to describe the executions of the opposition organized by the communists. "The end justifies the means" is also a negatively sounding phrase.
But just like "moving the goalposts", "the end justifies the means" has something legitimate in it, too. Even people who are much more decent than the Stalinists sometimes do a "smaller sin" in order to avoid a "bigger one". Some people wanted to save thousands of people from additional terrorist attacks planned by Osama bin Laden, so they sent killers to deal with Osama. Is it right to hire guns to kill somebody? It's probably not a good thing in isolation. But there was a grander justification.
In a similar way, the U.S. government nuked two Japanese cities. Is it something you expect from a saint? Probably not. It may be an even bigger sin to quickly vaporize tens of thousands of people than to drive an SUV. But the nukes have arguably saved millions of lives because they persuaded Japan to surrender – which could otherwise be very unlikely for that stubborn kamikaze nation.
What Tetragraviton seems to misunderstand is that all these "sins" that he presents as "absolute" are just relative sins, or even neutral events that are sometimes negatively painted, and how much they're negative depends on other assumptions – and the two sides have very different assumptions. So these accusations and negative descriptions of the antagonists' steps are really unavoidable. They follow from the different axiomatic framework. And the axiomatic frameworks may sometimes get modified – they may be replaced by vaguely similar but unequivalent ones. This is not the end of the world. The laws of physics believed by the same physics community have been changing for centuries – sometimes abruptly, sometimes incrementally.
You know, as I see it, the anti-quantum zealots misunderstand – either due to their stupidity or their stubbornness – almost all the conceptually important experiments, laws, principles, arguments, generalizations. Instead, the thesis "the world must be realist" i.e. basically classical is the most crucial and stable axiom of their whole world view. So of course they will be moving goal posts. They will find every partial failure of their work and theories irrelevant because "the world must be realist" is more important than anything from their viewpoint – just like Allah is more important than anything else according to the Muslims.
People who understand that the world has to be described by a non-realist theory – the correct one is called quantum mechanics – simply don't share this kind of religion with the anti-quantum zealots. According to them, physics is supposed to produce valid predictions about Nature. If a theory can't do it, it is a serious problem and such a problem – or many such problems – are surely in principle capable of falsifying an assumption, including the dogma "the world must be realist". And that's what happened already in the mid 1920s. Every physicist who was still up to his job understood that realism was dead before the 1920s ended. Anti-quantum zealots have their Allah and they're not willing to admit any evidence that shows that this Allah was shown to be a piece of šit more than 90 years ago.
This is the simple fundamental reason why they have been saying and writing idiotic things and tons of worthless books for more than 90 years. It makes no sense to co-write papers with them, you would basically become one of them. The primary reason for all the disagreements is very simple. They're bigots – and everything else follows from that through consistency. If you're an intellectually semi-honest bigot who believes that a ludicrous dogma is the most important truth in the world, it's your moral duty to spend your life by writing complete garbage, just like what all the anti-quantum zealots do.
Now, I think that the two other disputes that Tetragraviton mentioned are far more open than the foundations of quantum mechanics. Dark matter seems solid but there's some potential for a dark-matter-free, (generalized) MOND theory, to describe the data by more predictive equations. This more predictive theory seems to be too predictive and it seems to be falsified for that reason, although it looked pretty good at some level. But the falsification isn't waterproof. And there are really intriguing properties of the MOND-like explanations.
Like in the case of the foundations of quantum mechanics, the dispute results from the different fundamental assumptions. MOND enthusiasts simply think that the hypothesis about the extra invisible matter with a distribution that may be chosen arbitrarily to fit the data is too bad – it's cheating. Dark matter folks think it's no cheating – it's pretty much guaranteed that there is some matter that we can't see. If the addition of arbitrary dark matter is assumed to be a "mortal sin", of course you may forgive less mortal sins or failures to the MOND theory and its defenders.
In the case of the foundations of quantum mechanics, I think it's right to say that every scientist who is worth the name understands that the dogma "the Universe has to be classical" is something that simply doesn't belong to physics or science and the people who place it above everything are bigots. They're really led to ignore virtually all evidence and arguments – and even in the absence of any theory (let alone a natural and universally valid one), they're willing to say that their picture is better than the non-realist quantum mechanics. Concerning the predictions, they just take the predictions by another theory, the non-realist quantum mechanics, and with a poker face, they tell you (the ludicrous assertion) that their completely different theory may just obviously produce the same predictions. So they may steal all predictions from quantum mechanics, do no calculations based on their theory and logical framework, and claim that this procedure strengthens the case and kills quantum mechanics they have robbed. By the way, Muslims operate similarly, that's why the analogy is so good.
To write a paper along with an anti-quantum zealot would be at most a meaningless stunt. It would be like a plan to write a paper along with a chimp – except that the average chimp says many fewer stupid things than the average anti-quantum zealot.
But in the case of dark matter, it's already less clear. A MOND-like theory without dark matter could really be vastly more predictive. If it could work, the addition of arbitrary dark matter could be compared to creationism – lots of fine-tuning of God's plans is needed to explain something that a more predictive theory predicts basically automatically or unequivocally. Equivalently, if you lived in a world where dark matter is suddenly proven not to be needed, you could surely present the ancestors's belief in dark matter as a belief in ghosts – something that floats around us, cannot be seen, but still affects our lives. Of course that there is some reason why MOND remains attractive.
You can't really objectively quantify the relative importance of various arguments – such as "it would be so much better to have a more predictive theory without dark matter" vs "MOND seems to predict wrong motion of this or that cluster or dwarf galaxy or anything else". People are differently calibrated. So some people think that arguments of some kind are more important than others, and vice versa. Of course, at the end, at most one side is correct. But the debate between "dark matter and MOND" isn't "totally" settled yet which is why it's unsurprising that you may find some people – perhaps even rather intelligent people – who keep on defending MOND. They effectively assume MOND, adjust their interpretations of everything else accordingly, and their picture isn't ludicrously indefensible yet. That's why some of these arguments may continue.
Just to be sure, I find dark matter more plausible and I also think that the defenders of MOND are less intelligent in average – surely less skillful in particle physics and related things.
Now, take the de Sitter vacua in string theory. It would be better if we knew it but we just don't know it. I know the papers on both sides rather well. They really differ in the character of the work. The "de Sitter vacua are everywhere", e.g. all the papers based on KKLT, seem to assume that one has to write papers full of formulae and constructions that qualitatively seem to be using proper string theory everywhere and sometimes they use one approximation or another, some qualitative guess extracted from low-energy effective field theory, and so on.
I never found KKLT – even the anti de Sitter part of it – as a construction of something that looked canonically important, that had to be the way it was, and so on. It looked like a rather random construction which didn't overlap with the nice features of the Standard Model etc. too much. But I had enough time to think about it and I think that the many anti de Sitter flux vacua probably exist as solutions of the universal equations of string/M-theory – they exist "somewhere in the landscape".
Now, the vacua may be lifted to de Sitter, we heard, by the addition of anti-D3-branes etc. Another level of the seemingly arbitrary complications that don't obviously agree with Nature's empirical fingerprints. It's plausible that the vacua may be lifted but it's not certain. I've seen arguments on both sides and I remain rather undecided – close to 50-50 odds. You know, those 50-50 odds may easily change to 20-80 or 80-20 when I appreciate some argument, and then I may found that appreciation as a temporary intoxication. It makes no sense to distinguish 50-50 and 80-20 too much. Both of them really mean that we can't be sure. I know that there are people on both sides whose subjective certainty is much closer to 100-0 or 0-100. But I think that they overestimate the robustness of their arguments.
I think we can't be sure about the validity of the proposed de Sitter vacua in string theory. The really solid constructions seem to ban de Sitter. On the other hand, there could be very understandable reasons why de Sitter vacua want to look illegitimate even if they're fine – in particular, really doable calculations where approximations work and stability is guaranteed in the full theory are only possible with spacetime supersymmetry that is incompatible with de Sitter. To summarize, I think it's important for physicists to be allowed to believe both frameworks – and work on them. String theory without de Sitter vacua may still be compatible with the observations if there's some quintessence etc.
These pictures are qualitatively different. They differ in their conclusions about what's unique, what's predictable about physics, what's hard to calculate, what the bulk of the future research will resemble, how much energy you should spend with details in the situation where some important uncertainties persist, they just differ in almost everything. The difference isn't as brutal as the difference between string theory and loop quantum gravity, for example. But the difference is rather deep, anyway. If Christianity and Islam represent string theory and loop quantum gravity, "the landscape of de Sitter string vacua" and "quintessence in a simplified stringy landscape" may be compared to the Orthodox Christianity and Catholicism (let me assume that Persian guys may go Catholic, Cumrun LOL). It's possible that a Protestant branch will soon split off Cumrun's quintessence picture, too. ;-)
Science just deals with uncertainty and every sufficiently exciting cutting-edge research question allows some deep split of the community into subgroups that profoundly differ in some of their fundamental assumptions. When the progress continues, one of these groups will shrink or it will be increasingly obviously composed of dumber or less honest people. At some moment, you may effectively imagine that the losing side either gets peacefully converted or jumps the shark or implodes or hits the singularity. Well, the real world is a šittier place than that but the differences aren't interesting from a scientific viewpoint, only from a sociological one. As long as neither sides defends absolute dogmas, the truth will prevail at the end, when enough evidence tilts the balance to one side.
But I think we're not really there. The claims that Vafa or Silverstein have to make about papers on their side and competing side are not yet "obviously wrong" or "truly indefensible". You don't need the same degree of bigotry and blindness as the anti-quantum zealots need. It's a completely different story.
So I think that the disagreements are unavoidable, some of the behavior that looks really dishonest looks unavoidable to the side that is committing it because that side thinks it's obliged to respect some true axioms – which our side doesn't consider true, important, established, or neither. But all the researchers should keep their freedom to pursue the competing programs.
There must still be quality filters but the quality filters should only use benchmarks that are established. So the graduate students – whether they will choose Cumrun's or KKLT-style picture of cosmology (I am presenting them as equal but I do expect that Cumrun's program will continue to be much smaller than the "Stanford school") – will have to master the standard quantum field theory, basics of string theory that every sensible expert agrees with, and write a thesis that doesn't look stupid or obviously wrong, and so on. Because the controversial things may look wrong to one-half of experts, it may be a good idea to write a thesis about a less controversial topic! If you can only do controversial things, many smart people may have very good reasons not to consider you an expert, at least up to the moment when you defend your groundbreaking theory and become an unlikely hero.
But once a physicist passes the tests that are independent of the disputed questions, he or she should be free to choose the program and the axioms. Cumrun's and KKLT-like axioms are equally acceptable. The arguments will probably remain unavoidable up to the moment when one side really finds some evidence that settles the discussion. I do think that both Silverstein and Vafa and others would admit that they were wrong if they got a really solid proof of the opposite viewpoint.
From the sidelines, one thing that frustrates me watching string theorists debate whether the theory can describe de Sitter space is that they rarely articulate what it would take to decisively show that a particular model gives rise to de Sitter.What Tetragraviton writes here is just a wishful thinking. To decisively show that a particular model yields a de Sitter vacuum, you need the complete, not approximate, definition of string/M-theory that may be demonstrated to include all the known vacua, approximations, and formalisms within their accuracy – or which may authoritatively show that some of them were inconsistent for some reason.
When you have the complete definition of string/M-theory, then you partly solve it and rigorously prove or disprove the assertion that a particular de Sitter solution or any de Sitter solution obeys these ultimate stringy laws of Nature.
It was rather easy for an articulate man like me to articulate what it would take to decisively decide about the fate of a/the de Sitter vacuum in string theory. A problem is that to actually do these things is much harder than to articulate this program. So it's rather likely that we won't have the totally precise and universal definition of string/M-theory in coming years. All the arguments and calculations will therefore have to rely on some approximations or assumptions that aren't waterproof, and no proof that a/the de Sitter solution exists or doesn't exist may be completely decisive!
The uncertainty simply is a part of the scientific process. If things were certain, the questions would be settled and there would be no reason to study them. So I think it's fundamentally wrong for Tetragraviton to be so existentially scared of the disagreements and arguments – and to propose crazy magic cures for something that he undeservedly considers a serious problem that doesn't belong to science.
You know, Tetragraviton, you apparently want to liberate science from uncertainty, battles, and adjustments of theories following the new evidence, among other things. But uncertainty, battles, and adjustments of theories in the wake of new evidence are some of the key defining attributes of the scientific method! You should better get used to them. In particular, science is a battle – between sufficiently sharply formulated competing ideas and theories (and often their champions, too). People who have a psychological problem with this confrontational character of science may try a hippie club or a lipstick instead.