What Graduate School in Theoretical Physics is Really Like

I’m about to head off for a short New Year’s vacation in West Texas, but wanted to recommend a wonderful article that just appeared at Nautilus. It’s a memoir by Bob Henderson (who I met when he wrote about me, see here), appearing under the title What Does Any of This Have To Do with Physics? (although the title of the web-page, What Graduate School in Theoretical Physics is Really Like, is more descriptive).

Henderson was a graduate student at Rochester in theoretical physics, working with S.G. Rajeev. He later went to work on Wall Street, and more recently in journalism. His Nautilus piece is the best explanation I’ve ever seen of what it’s like to start working in this field as a graduate student, should certainly be required reading for anyone thinking of going into the subject. It’s also somewhat of a profile of Rajeev, who has worked on a wide variety of topics in theoretical physics.

One of the main themes of the piece is Henderson’s thinking about how and why he left theoretical physics, why he “quit”. Something to keep in mind is that this kind of decision is what most people who get Ph.Ds in the subject end up facing. There are 5-10 times more people getting Ph.Ds in this field than there are permanent positions doing research in it, so the career path starts out with a game of musical chairs that you are highly likely to lose. Different people make the choice to quit the game and do something else at different points and in different ways.

Henderson does an excellent job also of explaining what the real problem is with doing this kind of research: that of figuring out what the right thing to calculate is. For everyone, but especially for those at the beginning of a career, the subject is a huge collections of topics one doesn’t understand. One has to somehow choose a direction to pursue, and it most likely won’t go anywhere:

Writers talk of the terror of facing a blank page, but it’s no different for theorists like Rajeev trying to choose which path to take. There are an infinite number to choose from, and most go nowhere or back from where you came. The clock is always ticking and you spend so much time in the dark that it can make you not only question your path, but your own self worth. It can make you feel stupid.

Sticking with this and making a career of it involve some combination of good luck (being in the right place at the right time), ability, self-confidence, not having a family to support, and a host of other factors. As Rajeev explains to him:

Without naming names, he ticked through a catalog of his contemporaries who’d succeeded in theoretical physics even without having the towering mathematical intellect that I was sure it took and that Rajeev surely has. They’d made it, Rajeev explained, by focusing on problems that played to their strengths, or by taking advantage of computers, or by collaborating with peers who had complementary skills. Some socially gifted but not so mathematically talented types had gone quite far this way, earned a lot of renown.

Anyway, the whole piece is well-worth reading. Another recently published Nautilus piece that I learned about from a link on this one is The Universes of a Woman in Science. It’s by Kate Marvel, who shares with Henderson (and hundreds if not thousands of others…) the experience of getting a theoretical physics Ph. D. (string cosmology in her case), and then leaving the subject for another field (in her case, climate science, which she blogs about here).

This entry was posted in Uncategorized. Bookmark the permalink.

47 Responses to What Graduate School in Theoretical Physics is Really Like

  1. Tim May says:

    Two poignant stories.

    Things were bad when I was in physics in the 70s, but seem much worse now. I cannot imagine that Ph.D. students think they will become career professors.

    Myself, I once cooked steaks for Feynman in Isla Vista in 1973. We sat around in our typically dingy student apartment and Feynman opined, with great humor, about his adventures. More seriously, he said the “easy days” of discoveries and major new theories were just about over.

    We asked him where he thought the new excitement, or revolutionary spirit, might be.

    He said “computers.”

    A year later I opted to delay going to grad school for a while and joined a then-small company, Intel Corporation. I learned it was heavily based on some good physics and I was constantly challenged. Physicists were not too common (EEs and ChemEs were more common), so I did some good stuff. In 1977 I discovered that a single bit error problem which was a problem in dynamic RAMS was caused mostly by alpha particles from low levels of uranium and thorium in packages. And to a much lesser extent, then, by cosmic rays.

    Just three years after sitting down on cushions in our apartment with Richard Feynman. What an inspiration he was.

    Oh, and I also got to work at Intel with Gordon Moore, Andy Grove, Craig Barrett, and Bob Noyce. A stellar group. I was fortunate to work for Craig Barrett who worked for Andy Grove who worked for Gordon Moore. My results and summaries went right up to the top every day…like working on the Manhattan Project. Heady days.

    And now, merging black holes. Personally, the failure of SUSY is not that big a deal. Something else will come along.

    Happy New Year!

  2. Pingback: Ninth Level Ireland » Blog Archive » What Graduate School in Theoretical Physics is Really Like

  3. Chris Oakley says:

    Interesting article, although it should be pointed out that you do not need relativity to get nonsensical, infinite answers in quantum field theory. Second-quantised non-relativistic QM does this perfectly well. Also I, too would get angry or suicidal if required to work 15-hour days on quantum gravity in one dimension.

  4. Low Math, Meekly Interacting says:

    Surely one of the more foolish questions you’re likely to get, but what the Hell: Why is all this labor needed in theory? My background is purely experimental, and I’ve mixed with experimentalist grad students in the areas of biology, chemistry and physical chemistry. In these disciplines, it was quite obvious what grad students were for: slave labor in the laboratory. The more the better. Cheaper even than technicians and so terrified half the time they’d do any dirty job they were asked to just to avoid being forgotten entirely. And there’s no end of dirty jobs to do. But, aside from grading and other pedagogical grunt work, why have so many unneeded theory grad students? What else do they do for the institution?

  5. AcademicLurker says:


    I’ve heard it said that really experienced physicists often guess the answer to their problem/question using heuristic arguments, and then do the math to make sure their guess was correct. Maybe that’s the role of theory students, at least before they start an independent project. “I think the answer is X. Go check it out.”

  6. cgh says:

    In 1999 I started work on my PhD. It was in physics. I had some grant money and was simultaneously publishing. The work on my thesis wasn’t going as fast as it could of, probably because of this, but I felt productive. I was meeting regularly with a mathematician who was not my advisor but was working independently with me on related ideas. He is / was well known in the NY area for RH and some analysis work. One day, I suppose in retrospect, I was complaining. The gist of my complaint was about what I was going to do after I finished. He responded that if I did this to be employed I didn’t get it and was in for a rude awakening. To him, it wasn’t about employment, it was about math, and solving problems. He was an established mathematician with tenure, then, now emeritus. He didn’t have to take endless strings of adjunct positions in the boondocks. When he was working on his PhD in the 60’s if you thought hard and did good work things came to you. One didn’t need to think, at least as much, about food, of family, or being near family. He spent his entire career between 3 NY universities, 2 in Manhattan within 20 miles of where he grew up and raised his family, without giving it much thought. So in 2001 I went to Wall Street. I shared an office with a Caltech postdoc that came from the control and dyn systems group. I exchanged the hype of strings for, then, fast PDE solvers and, today, the hype of AI, deep and machine learning.

  7. Jeff M says:

    I started my PhD in 1986, in math. Back then, anyone in math with a doctorate from a good school got an academic job. I finished in 92, the year the job market collapsed. But even that year, most people ended up with academic jobs, just not at the level they expected. I have tenure, I get to do some research, it’s not the research 1 place I expected (my thesis was published in Mathematische Annalen, in a normal year I probably would have gotten a much better offer) but I’m happy. It’s much worse now, even in math. We’re hiring this year, we put our ad in really late, 2 weeks ago, and we’re already close to 200 applications. Very qualified people, including people with doctorates from places like Berkeley and Oxford, with publications in good journals. cgh is right about the old days, all my professors basically got their jobs when their advisors called someone to tell them to hire them. I have a funny feeling I might know the person he’s talking about, since I got my doctorate in the city…

  8. X says:

    Interesting article by Henderson, with things that all of us who have gone through the PhD experience can relate to. But what I found even more interesting was the insight it gave into the working life of his advisor Rajeev. He seems to have been extraordinarily generous with the time and effort he put into his students to bring them along on his research path. And his ‘rewards’ for that seem to have been pretty mixed. Imagine what it must have been like for him when, after all that time and effort, Henderson started hating him (which Rajeev and any advisor in that situation would definitely pick up on), rejected his research path and went off to try find his own.

    I’m not blaming Henderson for that, since it was a difficult situation for him and understandable that he wanted to find/generate sufficient output to justify the input and hardship from his perspective. Relations between PhD students and their advisors are often difficult and complex, since both are investing so much and the outputs they seek are not always aligned. But this illustrates a reality of the working life of faculty which can be pretty draining.

    So I think those considering doing a PhD in this field should not only look at Henderson’s experience as a student but also the close-up description of Rajeev’s working life as a prof, assuming finding a faculty job is the person’s ultimate plan/hope when embarking on this path.
    In a comment above, cgh recounts how a maths prof told him that “it wasn’t about employment, it was about math, and solving problems.” That’s true for the motivation of people who go down this deprived and often dead-end path, but it’s a far from accurate description of what the job really is for those “lucky” enough to land a faculty position.

    The reality is that it is first and foremost an educator job. Teaching classes, guiding graduate students, supervising undergraduate student projects, doing outreach activities in high schools, etc…(plus various administrative things). Research is something you basically do in your spare time, e.g. waking up at 3am to have a bit of spare time for it as Rajeev did in the story. (And this is at research universities such as Rochester; at liberal arts colleges getting research done is even more challenging.)

    Young people aspiring to a faculty job in this field because they “want to do research”, and think that “doing a bit of teaching is bearable and a fair trade-off for the freedom to do research” have misunderstood the nature of the job. It is an education industry job rather than research job. They should ask themselves if they really want to spend their working life in the education industry rather than, e.g., doing graphics software design for Apple, working on automated car driving for Tesla, or the myriad of other interesting and intellectually challenging things that smart and hardworking young people can aspire to do. If doing research is really so important you can always go work on Wall Street, retire rich at 50, and spend the rest of your life doing research without constraints or other commitments.
    (I’m speaking from experience here as someone who reached the “holy grail” of a faculty position at a research university and discovered that the job was not quite what I imagined it would be…)

    It was very refreshing how Rajeev acknowledged that “Some socially gifted but not so mathematically talented types had gone quite far this way, earned a lot of renown.” For quite a few people (myself included), one of the motivations for going down this path was the expectation that our lack of interest and talent for “sucking up” wouldn’t matter so much compared to careers in the business world or elsewhere. This is a myth though – it is no different than other walks of life in this regard.

    Re. LMMI’s question above about the need for PhD student “labor” in fields like theoretical particle physics: No it is generally not needed for the research, in fact it is often a time and effort-consuming burden on the prof to drag the students along his/her research path, as Henderson’s story illustrates. One explanation for why it happens anyway (perhaps the main one at many institutions) is that the number of PhDs produced is a metric on which individual faculty and departments as a whole are evaluated.
    I remember one faculty meeting where the dean excitedly told us about a new opportunity for PhD funding, and urged us to “go and grab this opportunity” – presumably by finding students to lure into the PhD path. There was no discussion of whether there was a need for more PhDs or what their future career prospects would be…

  9. A reader says:

    It is a beautiful and soulful article/essay.
    The key message for me is “Control your emotions.” It takes time. Sometimes it is too late.
    Between “Pursuing the Holy Grail” and “Shutting up and calculate,” is it not possible to find some middle ground?
    Like … “Be part of an epic human endeavour (we are primates), enjoy trying to understand/solve a problem, be ready to move on when it is over?”

    That said, from afar, and even though he obviously can write very well, I really wish that Bob Henderson gets a chance to go back to working on that he evidently loves: Physics.

  10. Jeff M says:


    I think what you describe, professors as educators, research a sideline, is sometimes true, and sometimes not. It is certainly true for anyone who teaches somewhere like I do, though even here it’s possible to do it differently. One of my colleagues, who just retired, used to only ever teach all sections of the easiest course he could get, and paid as little attention to teaching as he could. You don’t need to do much prep to teach what is really high school level math. And from what I’ve been told, there are plenty of people at research 1 schools like Rochester who are nothing like Rajeev, and who don’t even really care much about their advisees, much less the undergrads. As to retiring early and doing research on your own, it’s a thought I had when I didn’t think I would get an academic job. Decided I would sign a contract with myself that when I had x dollars in the bank I had to retire. Sometimes I look back and think it might have worked out OK, now that I’m stuck teaching summers to pay my kids school tuition. I could be relaxing in my Tribeca loft, sipping single malt, and thinking about math. Actually, back when I was finishing grad school, my wife and I lived in a giant loft in Williamsburg, but boy was the city different then.

  11. vmarko says:

    I have to say that I am not as fascinated by the Henderson article as many here are. To me it sounded like a story of a person having a naive, romanticized point of view on research in theoretical physics, only to be faced with a down-to-earth reality that has very little in common with his romanticized view. Einstein looked at a building through a window, and figured out general relativity — yeah, right. Everyone who thinks research can be done like that is certainly going to be very disappointed. And Rajeev was very very right at the end — Henderson’s issue was not lack of ability, lack of open faculty positions, lack of quality or lack of effort. His issue was precisely that he decided to quit. Research was not what he naively imagined it to be, and he caved when he got under some pressure. It was lack of dedication.

    In order to survive in a highly competitive research area such as hep-th, a person needs (aside from some luck, math abilities, good mentor, etc.) one crucial character trait not mentioned in the article — stubbornness. One simply needs to be more hardheaded, and not get too distracted by the reality being different from what one imagined.

    Regarding mentoring and collaborating with a supervisor, Rajeev seems to be a good mentor, judging from what Henderson wrote in the article. In order to teach a student enough to do research (in hep-th at least), a supervisor needs to dedicate to each student at least one afternoon of uninterrupted discussion in front of a blackboard every week or two. My supervisors did it with me, I try to do it with my students. Those mentors that don’t do that are plain and simply irresponsible, IMO. And it usually shows, in the number of years their students take to reach PhD’s, and their research performance afterwards.

    As for teaching vs. research, my anecdotal evidence is quite the opposite of what X is describing — I am in a research-only institute, doing quantum gravity, and I have literally no teaching duties (and never had any during my career, except when I substitute for a colleague). Moreover, recently I actually asked to get involved in teaching, at the graduate level. The reason is that I need more fresh PhD students as collaborators, to share the workload of my research. In a research-only institution I don’t have any contact with PhD students, so it’s hard to recruit them.

    So anecdotal evidence can vary greatly, it all depends where you are. I’ve heard some people dreading research-only institutions like IAS, Princeton. No contact with students is a terrible handicap when doing research, they claim, as there are no fresh ideas and excitement that young motivated people bring along. And I tend to agree — one needs to find some sweet spot between teaching and research, since each of the two extremes will have some drawbacks.

    Best, 🙂

  12. Low Math, Meekly Interacting says:

    Re. X: Ah, I see. They’re adornments. I know enough about academia that should have been obvious. Thanks.

  13. Why says:

    It is really confusing when you think about it. On one hand, the need for specialising in one of the so many new disciplines in theoretical physics seems to increasingly recquire new PhDs and post docs and on the other hand the increase of the population seems to require more tenured professors to be hired to teach.
    So why all this crisis in theoretical physics job market?

  14. X says:

    Jeff M,

    “there are plenty of people at research 1 schools like Rochester who are nothing like Rajeev, and who don’t even really care much about their advisees, much less the undergrads”

    That’s no doubt true, but it’s difficult to get away with as a tenure-track faculty, and probably getting harder for tenured faculty too. With rising tuition costs students are being regarded more and more as paying customers, and making them satisfied seems to be becoming a higher priority for universities.

    At the university I was at (ranked in the top 100 in the world in several rankings), teaching performance was important and given equal weight with research performance in our annual evaluations. There were people who were denied tenure despite having done well in research because their teaching sucked according to the student evaluations. And I’m pretty sure that if someone got tenure and then proceeded to be a terrible teacher thereafter, the powers-that-be would still find a way to push him/her out… (In fact there was a tenured prof who the powers-that-be didn’t like for some reason, and they made his position in the dept so uncomfortable that he “voluntarily” left…)
    In any case, I felt morally obliged to at least try to be a good and conscientious teacher/lecturer.

    As for the actual time and effort that the teaching side requires, it depends enormously on the details of the teaching assignments as you mentioned. There were some people who taught the same couple of courses every year, requiring zero preparation time and allowing them to get high student evaluation scores after fine-tuning the courses for so long. Those tended to be people who were favored by the powers-that-be for whatever reason… There were also assignments that required preparing course materials from scratch (including powerpoint lecture slides, handout notes, tutorial question & solution sets, online quizzes etc) for a new course or because the previously used course materials were disastrously bad. Less favored faculty such as myself tended to get those assignments…

    So unless the person is very favored and fortunate with his/her teaching assignments (and has great PhD students who don’t require much time or cause much stress), or doesn’t care about being a good teacher and is able to get away with not caring about it, I think the reality of the job even at most research universities is being first and foremost an educator, with research as a sideline.

    On the topic of gaining financial freedom as a path to being free to do research in future, Wall Street is not the only route. Marrying a successful businesswoman with her own company is another way, one that I was fortunate to follow (more by accident than plan…) These days I’m free to do whatever I want once the housework is done. 🙂

  15. colorado says:

    For a person coming from condensed matter theory, it is amazing that professors in HEP theory still do actual calculations and not rely on their students and postdocs to enlarge their publication list. In my field, a PhD advisor would not give an open problem to a student or postdoc without hoping to get a coauthorship. In fact, if you discuss more than an hour about what you are doing with a senior person, he or she would expect his/her name on the paper.

  16. future says:

    what is sad and ominous for the future of “physics” is that several people want to co-opt “theoretical physics” to mean high-energy theory / particle theory or at the very least, something quantum.

    many of the best theoretical physicists i know (having spent my time exclusively at the most “famous” institutions around the country) don’t work on quantum problems at all – and certainly not high energy physics. They work on condensed matter problems – quantum and classical, soft matter and increasingly, biology.

    The real test for physics departments is whether they quickly redefine physics to be these exciting areas with actual deep ideas – or restrict “theoretical physics” to mean rehashing of 40 year old high-energy ideas.

    The exciting new theoretical physics in biology, classical non-linear physics and condensed matter will thrive for sure.. but it may end up in biology and engineering departments in 50 years from now and “physics departments” will become the Nokia to their Apple.

  17. X says:


    Faculty in HEP theory definitely rely on postdocs to generate publications (and PhD students too, to the extent that they are able). In fact it is a main way they keep up a ‘respectable’ rate of research output, considering that their own time for doing research is often very limited. Research projects are farmed out to postdocs and students who do most of the actual research work, while the prof’s role is mainly ‘research manager’. I guess this is the standard model throughout academic sciences. At any rate, the people who were productive and successful in research where I was all followed this model to some degree. (Personally I consider it a horrible model and refused to play that game…)

  18. Why says:

    To those thinking that it is a good idea to work in business until retirement and then start doing research: can you give me one single example from the history of physics where somebody made a significant contribution to physics after retiring?

  19. Another Anon says:

    That’s a great piece of writing by Bob Henderson. I think it answers his question of where it all went wrong: his talents lay in being a writer, not being a physicist.

  20. Thomas Larsson says:


    Although not exactly physics, an example would be Leopold Kronecker, who retired from the family business to pursue a career in mathematics. Of course, most of us cannot afford to retire at the age of 32.

  21. Jeff M says:

    The whole “research manager” thing I find amazing. I appreciate it’s very common in the sciences, but in math from my experience it’s very different. A postdoc working with a senior researcher would only put the senior persons name on a paper if they wrote it together. It wouldn’t happen if the senior person suggested a topic, and then made some suggestions while the postdoc was working on it. The senior person would get thanked, sure, but not writing credit. When I started work on my thesis, my advisor suggested something, which I went off and worked on. After a few months, I realized I could show that what my advisor thought was true wasn’t true, but I had an idea what should be, and that it was interesting. But I had no clue how to prove it. Neither did my advisor. So he asked around, and a friend suggested an idea he had once had for a related problem, which he never ended up working out, since he figured out an easier way to do what he wanted. So I worked it out, all the details, and proved what I wanted. I thanked the (very famous) guy who suggested the basic technique, but he certainly didn’t get writing credit, and in fact the technical lemma is known as “M***’s lemma” in the field. I did do all the hard work on it….

  22. Oldster says:

    Why: Does Copernicus count? I think I recall from Arthur Koestler’s book, “The Sleepwalkers”, that he was actually on his deathbed before he saw the first printed copy of his book. He’d been somewhat known as an advocate for the Heliocentric Theory prior to that, at least locally. And maybe that doesn’t count as “retirement”, even though he was dying …

  23. Why says:

    Thomas Larsson,
    Thank you for your answer. But that was an exception that confirms the “rule”: one generally does not produce anything significant after age 60. I am not well aware about teaching in the U.S. but why somebody like Henderson does not teach at high school after getting a PhD and have plenty of time for independent research instead of waiting until age 50 or 60 to retire before starting research? Besides, I guess one does not need a PhD to teach in a high school, so why all this pressure on physicists who really love to teach and do physics?

  24. Jeff M says:


    Teaching high school is hard. Doesn’t leave any time for research. And leaves no time for things like going to talks and conferences and such, except over the summer. The high school teachers I know are some of the hardest working people around.

  25. vmarko says:


    Teaching at a university and teaching in high school are two completely different things. In high school, the course material is very easy, but you typically have several classes per day with 20-30 pupils in each class. You need to give classes basically every day and every week, which means no time to attend meetings, workshops, conferences etc (as Jeff said). In addition, you need to provide exams to all those pupils, and grade them, which takes a huge amount of time. In other words, its intellectually easy, but the workload is enormous.

    In contrast, at a university, the course material requires some preparation, but you typically have one or two courses per semester, which amounts to giving lectures twice a week, to 20-30 students in total. You have the ability to skip or reschedule the lectures to make room for travel, and you have a teaching assistant who does all the grunt work of working the students through the problem books and grading their exams. So it’s intellectually less trivial, but the workload is nowhere near the amount of high school.

    Best, 🙂

  26. A reader says:

    Why, your question is interesting. I don’t know whether it is OT or not, but … IMO, it is not very useful to repeat/promote that myth(?), that somebody can do research working all alone. Is there any modern example (say after 1946) of a theoretical physicist or a mathematician who has been able to do proper research without being part of/interacting with a community of peers? See for example 3. How is mathematical understanding communicated? in that essay of William Thurston, On proof and progress in mathematics, which Peter Woit mentioned a long time ago: http://www.ams.org/journals/bull/1994-30-02/S0273-0979-1994-00502-6/S0273-0979-1994-00502-6.pdf (arxiv.org/pdf/math/9404236v1.pdf)

  27. Peter Woit says:

    A Reader,
    Perelman came up with his proof of the Poincare Conjecture while not having an academic job (I don’t know how he was supporting himself, certainly not by working at another job).

  28. Dragster says:

    And, out of curiosity, what percentage of tenured professors do you feel have made a ‘significant contribution to physics’?
    Seems retired amateurs aren’t necessarily lagging so far behind if you want to use that metric. Perhaps they just feel less compunction to publish mediocre papers.

  29. Why says:

    Thank you all for your comments and answers.

  30. Petite Kabylie says:

    Through Why’s question and others replys and comments I understand now better why unifying the laws of Nature is not the holy grail for a theoretical physicist as is a tenured professorship!

    Thank you!

  31. A reader says:

    Peter, thank you!

  32. srp says:

    Great article and informative thread here. Thanks.

  33. Pavel Krapivsky says:

    Perelman quit his job at Steklov Institute at the end of 2005, long after he submitted his proof to arXiv. I watched a Russian documentary about Perelman where some of his colleagues were guessing that his mother pension plus support from his sister were crucial. The wiki article about Perelman refers to articles in Russian press claiming that from 2014 Perelman is working, perhaps part time, in Sweden. His sister apparently works in Stockholm, so hopefully Perelman is surrounding by family and happy.

  34. Yi-Zen Chu says:

    Jeff M’s comments above (January 2, 2017 at 9:54 am) reinforces why I hold mathematicians with higher intellectual regard than I do theoretical physicists.
    Jeff M: How do mathematicians maintain this high level of intellectual and scientific integrity — i.e., where only the ones who did the actual work get on the paper? Is this unspoken culture, or are there concrete incentives in place that support and encourage such behavior? As the job market/funding withers and competition increases, is there any discussion regarding how the math community can continue to maintain such high standards?
    I have been wondering for a while now, why major journals in physics (Physical Review, JHEP, JCAP, etc.) do not, at the minimum, require authors to enumerate their individual contributions in some detail. This would level the playing field at least a tiny bit for people who do not wish to play the game of riding on other people’s hard work, or be exploited by others. (Much more, of course, can be said regarding the credit structure of theoretical physics.)
    Lest Peter boots me for being off-topic, let me mention this is highly relevant to the experience of being a graduate student in theoretical physics. It is indeed true that the (rarely challenged) cultural norm of theoretical physics is that, if the adviser came up with the idea, he/she would most definitely show up on the paper — regardless of how much work he/she had put into the actual implementation. It’s almost as if the details aren’t important; when we know. in science, this is untrue. Furthermore, the student often ends up being mentored by postdocs and other grad students, especially if the adviser is very senior. When the project succeeds, the adviser gets to put it on his/her CV, which will be viewed by funding agencies and colleagues as yet another contribution to his/her field. With the money obtained, the adviser is able to pay his/her summer salary, his/her standing with school administration advances further — and quite importantly, he/she gets to travel around to interact with distant colleagues; build his/her network; write yet more papers simply by “chatting” with others; etc. What the graduate student may not realize at first, is that many “ideas,” “insights,” or “open questions” in theoretical physics are bounced around among the people in the know/loop; and so while the student may wonder in awe how the heck his/her adviser came up with these research directions, the basic framework/questions are often quite well known to researchers who have been in the field for a while. On the other hand, when the project progresses very slowly because of genuine technical/scientific issues or even fails completely, the burden of the ensuing consequences — say, the lack of postdoc opportunities because of the low rate of publication — oftentimes falls squarely on the lower ranking members, particularly the graduate student. How this is an ethical reward structure, especially given the dire job situation of our times and given we are supposed to be “scientists,” is something that has boggled me for many years. (I also believe these issues are intimately tied to the reliability of the theoretical physics research literature.)

  35. chethan krishnan says:

    I am sure there are plenty of senior guys who piggy-back on the blood, sweat and tears of grad students and postdocs, but that doesn’t mean that there is no value to the leadership, intuition and experience that a senior person brings to the table.

    The skill of asking the key questions and having the key ideas that solve problems requires maturity. Together with a bit of (delusional?) confidence, a sense of direction, and an ability to not panic while fumbling in the dark. This last bit is what happened to Henderson as he beautifully narrates in his piece.

    In Jeff M’s example, think the senior person deserved some credit. He had a key input in solving the problem. He didn’t want the credit, but that doesn’t change that fact. It is also a tribute to Jeff M’s integrity that he is open about it.

    For the purpose of full disclosure, I have a permanent job. But I eat my greens and I calculate, dammit! 🙂

  36. piscator says:

    Some comments on this.

    1. In the big picture of science in terms of authorship, theoretical physics is pretty close to maths. Go look at any lab subject to see papers where the senior author *really* has no intellectual contribution to the subject (conventionally the last name).

    2. A lot of the comments here miss the fact that a paper by a graduate student on an OK-but-not-revolutionary topic gets *more* impact with a senior person’s name on it than if by themselves – this helps the student not hinders them, and senior folks’ reputation in theoretical physics is not normally based on being senior author on OK papers done primarily by graduate students.

    3. Problems given to graduate students are often of the ‘I know how to do it if not the precise result, these are the broad methods that will work, this is the broad structure of the result.’ The ‘idea’ is a large part of the work, and normally the advisor will also be writing parts of the paper as well (relating the work to the wider context of the subject etc).

    4. Probably the single biggest failing of graduate students is to have an overly-narrow focus on the precise technical details of their own calculations, at the expense of understanding *why* something is important (and therefore focus on calculating *things that matter*).

    5. Scientific contributions are on a logarithmic scale. Most advisor/student papers are as much training projects as they are advances in science. Reputations are based on big results – and important, open and easily doable problems don’t get assigned to graduate students, they get done as quickly as possible by anyone who encounters them.

  37. Bernd says:


    at least in some areas, #1 is not true anymore. In some groups, it is normal for pure theory papers to have about a dozen of authors. This is crazy of course. However, people involved in this game receive much higher citation statistics so they have an easier life on the job market, perpetuating this practice.

    While #2 might be true for the impact of the particular paper, it is not helpful to the individual student. If you have a single author paper, it will look much better on your CV than if all your papers are together with your PhD advisor.

  38. piscator says:


    My more detailed statement would be that the optimal positioning for a student on the job market is to have one or two papers that are by themselves or with other students, and the rest coauthored with senior people. So I broadly agree with you here, ideally a student in TP will neither have all papers by themselves or all papers with their advisor.

  39. Richard says:

    An interesting, enlightening — and ultimately * depressing * thread: as I made my way through it (N.B. I read the Henderson piece in its entirety before doing so.), I could not shake the increasingly intense feeling that the garment — to the weaving of which such a thread best lends itself — is a shroud….

  40. Jeff M says:

    Some general comments. First, there aren’t really any incentives in math in relation to giving author credit, it’s just the way it is. I think the fact that there is much less money in math is part of it. It’s also maybe that it’s hard to imagine a math paper with 15 authors if you’ve ever written a math paper. It’s pretty hard to imagine a paper with 4 authors, although I know a few (they tend to be from conferences, where some people got together over lunch and wrote something). As far as the senior person who helped me on my thesis, he of course did get credit, just not author credit. I thanked him, something like “I would like to thank so and so for suggesting this technique, from an unpublished manuscript.” In the published version and the in house thesis both. Very common in math papers.

  41. colorado says:

    I do agree that collecting data or calculating numbers is not doing science. Figuring out which quantity to measure or calculate, coming up with the way to do such measurement/calculation, and analyzing the data/numbers is the real part of the science in action. In my experience, many students and postdocs do most of these things, while the “PI” just tells his/her minions to work on a particular area. If the PI is paying the students and postdocs, then it’s alright that they are expected to work together with the PI. My gripe is that the PI’s are never working, and they expect their work to be done by the students and postdocs.

  42. Radioactive says:

    I don’t know if there was really a problem with Henderson or his advisor here¸ more that the inspirational material he read didn’t describe the reality of doing theoretical physics. Kind of like watching the Olympics and assuming that you’d be standing on the podium without realising most of your life would be training, and that the vast majority of people who try, fail.

    Who doesn’t want to solve the mysteries of the universe? That’s simply not enough.

  43. Justin says:

    How difficult is it for a particle physics graduate to switch into mathematics where I think the employment prospects are much better? At my undergraduate institution (only a top 70 school) the phd math graduates had decent employment rates at very small unknown four year colleges. Maybe this isn’t what one would consider glamorous, but if the physics grads can’t find any physics jobs, isn’t the switch to math an option they could consider?

  44. Jeff M says:


    I can tell you my school we wouldn’t hire someone with a physics PhD, we have way to many math PhD applicants to consider it. I’m a state U (second level) I tend to think liberal arts schools are the same. Even in math people have trouble getting an academic job, though if you are willing to teach anywhere and have good teaching refs you can often get something. We’re hiring this year, we got our ad in late (state budget issues), it only got posted about christmas, and I just checked, we already have 218 applicants. Plenty of them with degrees from top 20 schools.

  45. bks says:

    The problem is neither theoretical physics nor capricious advisers, but rather the mind-numbing, interminably gray weather of Rochester. I relapse to seasonal affective disorder just thinking of my 15 years there. It is no surprise that in the third paragraph Henderson has his epiphany under the first blue sky in the wake of one of Rochester, New York’s typically brutal winters.

  46. Yi-Zen Chu says:

    Chethan Krishnan (January 4, 2017 at 5:12 am): Senior scientists most certainly possess the potential to contribute intellectually to the papers their names appear on — from the ideas on and guidance towards the important questions; to providing invaluable insights and problem solving approaches along the course of research; to suggesting the appropriate checks that ought to be performed after an arduous calculation; etc. I personally view these as a vital part of the mentor-ship a senior scientist should be providing to his/her students and postdocs. My main point is, given the contrast between the reward structure faced by established scientists and those by the students/postdocs — as well as the dearth of jobs/over population of students and postdocs — how frequently does this mentor-ship scenario actually play out? (Expecting senior physicists to perform their own independent calculations is largely out of the question these days; this is why those who still write papers on their own certainly command my respect.)
    On the other hand, I should have clarified that my observations are somewhat US-centric, having spent graduate school and postdoc years here. In particular, I have little sense of how the Indian academic/theoretical physics system works. Finally, please do not view my comments as a reckless attack against all senior theorists. They are, instead, based on a genuine desire to uphold high standards of scientific honesty. (It is because I take these issue serious, that is why I do not post these comments anonymously — as you probably can recognize, as a lowly postdoc, my comments do carry some level of personal risk if they are misconstrued.) Mature scientists should be able to look to other scientific communities — in my case at hand, mathematicians — to adopt whatever good practices they have, so as to help safeguard the integrity of one’s own scientific community.
    piscator (January 4, 2017 at 5:25 am): Your #2 — that getting senior people on papers benefits the junior ones — is, strictly speaking, a political statement and not a scientific one. Now, I am just as guilty as the next person for clicking on an arXiv paper because it has a famous name on it. But isn’t this precisely why it is an issue of integrity that we should be sensitive to? In addition to the intellectual discipline we should exercise, to judge a paper by its merits and not by who wrote it — should we not question who actually did the work and implemented the ideas involved? (This is why I suggested all journals should, at the minimum, require authors to enumerate in some detail their individual contributions.) Why are we unfairly rewarding people who have infinite job security while watering down the credit of those who likely have to quit some time down the road? I would also point out, the more everyone buys into this mode of operation, the more the incentives are stacked in favor of established people and against the students/postdocs.
    Do you recognize your #4 describing the failure of the adviser involved too? In other words, when a student “fails” to sift out the important physics from the complicated technicalities, is it not the adviser’s job to steer him/her back on course? I believe the cultural norms and ethical standards of theoretical physics oftentimes place unfair and unrealistic expectations on graduate students. While academia pays lip-service to the notion of “diversity” — and from what I can tell many physicists do claim to subscribe to liberal and/or progressive values — the reality on the ground is really quite different. How can we claim to truly value a diverse range of viewpoints if we do not seriously invest the effort to nurture and support budding scientists, i.e., the graduate students/postdocs?
    Jeff M (January 4, 2017 at 3:44 pm): Incentives need not be very explicit “if you do X you will receive Y,” although the money/funding you mentioned likely serves as a significant driver of behavior among more senior theoretical physicists. Given we are social animals, cultural norms exert powerful pressures on us all — sometimes for the good. If, for instance, mathematicians “grow up” watching their advisers being extremely scrupulous about giving proper credit and finding multiple ways to check their work for errors — these practices will likely be passed on to the next generation. On the other hand, if questions such as “Can we take limits to check if this messy answer makes physical sense?” come up rather infrequently; while students/postdocs witness it is the people who publish O(10-15 papers) per year that ultimately get hired — then these are rather concrete disincentives against upholding basic standards of carefulness/thoroughness. In theoretical physics, “bend-over-backwards” attitudes are essentially extinct; while fashion-chasing and publication-rate-amplification is high on most people’s agenda. As I’ve alluded to before, given the current climate, I worry about how reliable the literature is. (The sloppy nature of peer review does not help either — but here, I am wary about straying too far off-topic…)

  47. anon says:

    It seems utterly depressing that the best minds of the last two generations (at least) have been guided by circumstances and a changing culture into squandering their talents on toys (e.g. social media) and predatory finance.

Comments are closed.