Monday, July 25, 2016

Experimental physicists shouldn't be trained as amateur theorists

Theoretical work is up to theorists who must do it with some standards

Tommaso Dorigo of CMS is an experimental particle physicist and in his two recent blog posts, he shares two problems that students at their institutions are tortured with. The daily problema numero uno wants them to calculate a probability from an inelastic cross section which is boring but more or less comprehensible and well-defined.

Dorigo just returned from a trip to Malta.

The problema numero due is less clear. I won't change anything about the spirit if I simplify the problem like this:
A collider produces two high-energy electrons, above \(50\GeV\). Think in all possible ways and tell us all possible explanations related to accelerators, detectors, as well as physical theories what's going on.
Cool. I have highly mixed feelings about such a vague, overarching task. On one hand, I do think that a very good experimenter such as Enrico Fermi is capable of producing answers of this broad type – and very good answers. And the problem isn't "too dramatically" different from the rather practical, "know everything" problems I was solving in the PhD qualifying exams at Rutgers – and I am too modest to hide that I got great results in the oral modern exam, good (A) results in the oral classical exam, and the best historical score in the written exam. ;-)

On the other hand, I don't think that there are too many Enrico Fermis in Italy these days – and even outside Italy – and the idea that a big part of the Italian HEP students are Enrico Fermis look even more implausible to me. The problem described by Dorigo is simply more vague and speculative than the problems that look appropriate.

Unsurprisingly, this "problem numero duo" was sufficiently vague and unfocused that it attracted no comments. What should one talk about if he's told to discuss "everything he knows or thinks about particle physics, with a special focus on high-energy electrons"? Well, there have been no comments except for mine. Let me repost my conversation with Dorigo.

LM: Sorry, Tommaso, but the kind of "problems"
...Consider all thinkable physics- and detector-related sources of the detected signature and discuss their chance to be the true source of the observation. Make all the credible assumptions you need about the system you deem appropriate.
are very poor from a pedagogic viewpoint. It's like "be a renaissance man who knows everything, just like the great Tommaso Dorigo". It's not possible to consider "all thinkable sources" in any kind of rigor – some of them are understood well, others are speculative or due to personal prejudices. So it would be impossible to grade such a thing. And at the end, the implicit assumption that Tommaso Dorigo himself is a renaissance man is sort of silly, especially because you basically never include "new physics" among the thinkable sources of observations and even if you did, you just have no idea which new physics is more likely and why.

TD: Hi Lumo, sorry but you should think a bit more about the task that the selection committee has in front of them. They have to select experimental physicists, not theorists. An experimental physicist should be able to produce an approximate answer, by considering the things worth mentioning, and omit the inessential ones.

Prejudices are good – they betray the background of the person that is being examined. This question is not too different in style to the 42 questions that were asked at the last big INFN exam, 11 years ago.

LM: Dear Tommaso, your order to "consider all thinkable things" is even more wrong and *especially* wrong if the student is supposed to be an experimenter because an experimenter simply shouldn't be solving the task of "thinking in all possible ways". He is not qualified (and usually talented) for such things – and that's also why it's not the kind of an activity that he will be doing in his job.

An experimenter is doing a complex but largely mechanical job "within a box" – at most, as a clever engineer, he is inventing ways how to test things assuming that one is within a box. He isn't supposed to understand why the box is what it is (derivation or justification of current or new theories) and his work doesn't depend on having opinions whether the box is a good one or a bad one and whether it should be replaced by another one and how likely it is.

High-energy electrons at a collider are *obviously* a chance that new physics has been discovered. That's why the colliders are being built. They *want* to and *should* find all the cracks in the existing theories of physics that are accessible within a given budget and a kind of a machine. What the possible theories explaining new phenomena with high-energy electrons are isn't something that an experimenter should answer, it's a difficult eclectic question for a theorist. And theorists wouldn't agree "what kind of a theory" should explain new high-energy electron events. They have their priorities and ways to read the available evidence. But they're not hired or rated or graded for their opinions. Theorists are hired and rated for the ability to mentally create, derive, calculate, and construct arguments. You want to change that – you want an experimenter to be an amateur theorist who is graded for his opinions, not for actual theoretical work that he isn't producing.

A task like that could only be about the knowledge of the "box" – about all the reasons why the Standard Model could produce such a thing, including all the phenomena that appear in the dirty real world of the experiment. But [because the work on many distinct parts of an experiment is usually left to the specialists,] it's questionable whether it's helpful to formulate tasks that ask the student to consider *all* these dirty things at the same moment. It's almost like a question "tell us everything you know about experimental particle physics". There are lots of things and what are the priorities is unavoidably largely subjective and dependent on the context. It's bad to grade such things because the examiner is still imposing his personal opinions onto the students.

You know, what I am generally afraid of in this pedagogic style of yours is that you want to upgrade yourself to a kind of a duce – and grade students for the degree to which they agree with you. But that's exactly what science is *not* – and you wouldn't get an A if you were graded by real experts in these parts of thinking yourself. Science works independently of duces, including duces like you. It is building on the objective evidence and by telling the students to show their thinking about almost everything and the balance between all these ideas, you are clearly shifting science away from the cold and unquestionable realm of objective evidence to the realm of personal opinions – and you (and every experimental physicist) is simply too small a man for such an ambitious position.

LM: added for TRF: Note that I was respectful towards Dorigo's nation – for example, I have used the original term "duce" instead of the derivative term "Führer". ;-)

But back to the issue. All of our disagreement is basically about "one idea" in some way. But we express it in various ways. For example, I haven't commented on Dorigo's comment that INFN has been using similar problems in many other exams (well, that's even worse if there's something fundamentally wrong with the problems) and especially his quote:
Prejudices are good – they betray the background of the person that is being examined.
Prejudices – that indeed betray the background – aren't "good" in science. Prejudices are unavoidable and the approach of people (theorists and experimenters) unavoidably depends to their background and prejudices to one extent or another. And some backgrounds may perhaps be more helpful in achieving a certain advance in science which is why the diversity of backgrounds could sometimes help – although I am generally skeptical about similar claims. But a key insight of Dorigo's countrymate Galileo Galilei that Tommaso Dorigo clearly misunderstands is that the whole scientific method is a clever and careful way to make the conclusions as independent of the prejudices as possible. And that's why rational people think of science (and Galileo) so highly! The processes included in the scientific method simply "dilute" the effect of the prejudices to homeopathically low levels at the end – the final conclusions are almost all about the cold hard evidence that is independent of the prejudices.

So prejudices may always exist to some extent (and I often like the fun of the interactions of the backgrounds, as my repeated references to Dorigo's Italian nationality in this very blog post also show) but science, if it is done well, is treating them as annoying stinky pests. Feminists and reverse racists may celebrate the influence of backgrounds and their diversity for their own sake (and so do all other racists, nationalists, staunch religious believers, and all other herd animals) but a person who thinks as a scientist never does. Science maximally minimizes the concentration and influence of these arbitrary cultural and psychological pests – and that's why science is more successful than all the previous philosophies, religions, nationalisms and group thinks combined. Dorigo worships the opposite situation and he worships the dependence on the background, so he is simply not approaching these fundamental issues as a scientist.

Backgrounds and prejudices may be fun but if a talk or an argumentation depends on them too much, this fun is not good science. I am sure that everyone whom I have ever met and who was considered a good scientist – at all institutions in the U.S. and my homeland – agrees with that basic assertion. Too bad Dorigo doesn't.

As I have explained, the question "what kind of new physics produces high-energy electrons" is a typical question that should be answered by a high-energy physics phenomenologist, not an experimenter. Similarly, the detector part of Dorigo's question should be answered by experimenters – but it's OK when they become specialized (someone understands calorimeters, others are great at the pile-up effect). And the phenomenologists will disagree about the most likely type of new physics that may lead to some observation! The answer hasn't been scientifically established and people have different guesses which concepts are natural or likely to change physics in the future etc. Their being good or the right men to be hired isn't determined by their agreement with each other or with some universal authorities. Phenomenologists and theorists are people who are good at calculating, organizing derivations, arguments, inventing new schemes and ideas etc. It doesn't matter whether they agree with some previous authorities.

Obviously, the Italian student facing Dorigo's question can't write a sensible scientific paper that would consider all possible schemes of new physics that produce high-energy electrons at the level of standards and rigor that is expected in particle physics – he would have to cover almost all ideas in high-energy physics phenomenology. This student is basically order to offer just the final answers – his opinions and his prejudices, without the corresponding argumentation or calculation. And that's what the student must be unavoidably graded for.

That's just bad, bad, bad. In practice, the student will be graded for the similarity of his prejudices and Dorigo's prejudices. And that's simply not how the grading works in the scientific method or any meritocracy. If a question hasn't been settled by convincing scientific evidence, the student shouldn't be graded for his opinions on that question. If he is, then this process can't systematically select good students, especially because Tommaso Dorigo – the ultimate benchmark in this pedagogic scheme – would get a grade rather distant from an A from some true experts if he were asked about any questions that depend on theory, especially theory and phenomenology beyond the Standard Model. For the Italian institutions, to produce clones (and often even worse, imperfect clones) of such a Dorigo is just counterproductive.

Dorigo's "pedagogic style" is ultimately rather similar to the postmodern education systems that don't teach hard facts, methods, algorithms, and well-defined and justifiable principles but that indoctrinate children and students and make them parrot some ideological and political dogmas (usually kitschy politically correct lies). Instead of learning how to integrate functions, children spend more time by writing "creative" (identical to each other) essays how the European Union is wonderful and how they dream about visiting Jean-Claude Juncker's rectum on a sunny day (outside) in the future.

That's not a good way to train experts in anything. And Dorigo's method is sadly very similar. Needless to say, my main specific worry is that people like Dorigo are ultimately allowed by the system to impose their absolutely idiotic opinions about the theory – something they know almost nothing about – on the students. It seems plausible that an Italian student who is much brighter than Dorigo – to make things extreme, imagine someone like me – will be removed by the Italian system because that has been hijacked by duces such as Dorigo who have completely misunderstood the lessons by giants such as Galileo. In this way, the corrupt Italian system may be terminating lots of Galileos, Majoranas, and Venezianos, among others, while it is producing lots of small marginal appendices to Dorigo who is ultimately just a little Šmoit himself.

By the way, after the blog post above was published, a new comment was posted by another user nicknamed "Please Respect Anonymity" on Dorigo's blog (the post about the "problema numero due"):
Based on credible assumptions about the system, a possible theory is that some not-very-good physicist expert of leptons, photons and missing energy must win the "concorso". So the commissars choose a undefined question with a subjective answer about leptons, photons and missing energy.

PS: the question would be very good in a different system.
Exactly, amen to that. Perhaps at a school with 500 clones of Enrico Fermi (among commissars and students) who were made up-to-date, the question would be great. But in the realistic system we know, with the names we know, it's pretty likely that the question brings at most noise to the grades etc. Or something worse.

Dorigo's reply is that the question is great and the right answer is that there had to be a Z-boson or two W-bosons with either one or two photons, or a spurious signal in the calorimeter. I think that only a small fraction of the true experts would answer the question exactly in this way. Moreover, the answer completely omits any new physics possibility, as all of Dorigo's thinking. It's extremely problematic for experimenters not to have an idea what a discovery of new physics would actually look like – or to encourage them to believe that it can't happen – because that's the main achievement that the lucky ones among them will make.

Experimenters simply must be able to look for signals of new proposed theories, at least some of them – whether they "like" the new theories (as a group or individually) or not. Whether they like them or not should be an irrelevant curiosity because they are simply not experts in these matters so this expert work shouldn't be left to them. Experimenters' opinions about particular new theories' "a priori value" should be as irrelevant as the opinion of the U.S. president's toilet janitor about the Israeli-Palestinian conflict. She can have an opinion but unless the system is broken, the opinion won't affect the U.S. policies much. If she believes that there can't be an excrement at a wrong place that needs to be cleaned after the visit by a Palestinian official, and that's why she doesn't clean it, well, she must be fired and replaced by another one (be sure that that may happen). The same is true for a CMS experimenter who is hired to look for new physics but is unable to look for SUSY because of some psychological problems.

No comments:

Post a Comment