Trapped in the Scientific Doldrums

By Julianne Dalcanton | March 19, 2009 10:00 am

Scientific progress does not always go “Boink“. Instead, it frequently goes “muttermuttermutter@!$%&*mutter”.

The thing about science is that, well, when you’re trying to figure out something new, no one can tell you the answer. There’s no hint in the back of the book, or 24-hour help line. So, if you get stuck, you may well find yourself in the scientific equivalent of a lonely mountain road with a broken axle. And no cell service. In essence, you’ve broken down in a place no one wants to go to.

But, unless you’re ready to call it quits, you somehow have to find your way out. You don’t know how long it’s going to take, or how you’re going to make it, but somehow, you’re going to find your way back to the main road. So, you keep hurling yourself at the problem, trying every MacGyver trick in your arsenal, hoping that something, anything, will work.

The thing about these periods is that while they’re about the single most frustrating and unrewarding part of the scientific process, they are also a ridiculously effective way to develop new skills. By the time you’ve tried 5 different methods of fitting a straight line to some data points, you’ve picked up a heck of a lot about statistics (or, in my case, about solving a network of partial differential equations where the time dependence of three quantities depends on the spatial variation of 1 or more of the others. Ick.).

The good news is that as you get older, you spend less and less time in such states, because you’ve got an enormous swiss army knife of tools that you carry around in your back pocket. (“Hmmm, broken axle? I recognize that particular brand of axle from Abromowitz & Stegun, and I believe I might happen to have some IDL code right here that can fix it…”) The bad news is that you can forget how much those periods, well, suck.

(Unsolicited but Related Advice for Junior Scientists: When giving talks, avoid dwelling on the Doldrums, even if you spent 90% of the project drifting around them. Focus on the science, even if it was only a small fraction of your research effort.)

  • Arun M

    Thank you for this post… I need it for the week :)

  • ollie

    Something similar can happen in mathematics; I once spent 2 years trying to prove a result. I couldn’t because….the result was false.

    I did publish the counterexample though. :)

    But yeah, it can be discouraging, especially to those who are teaching at 12 semester hour load institutions.

  • Julianne

    Exactly Ollie! When your time to work on something is incredibly limited, you start weighing sunk costs pretty heavily.

  • Arun

    Most real life adventures have enormous stretches of dull parts. The parts that make it to the news, movies or biographies are the few high points. So it is with scientific work.

    But “no one can tell you the answer. There‚Äôs no hint in the back of the book, or 24-hour help line” is also a motivation.

  • astropixie

    if its any consolation, we had a seminar talk this week where the (well known) speaker mentioned your (and peter’s) exciting result of finding a galaxy whose thick disk potentially counter rotates with respect to the thin disk! the atmosphere in the audience was one of tense excitement!

    slowly weighing down my swiss army knife…….

  • Ben

    It could be worse. You could publish a highly cited paper about 5 different methods of fitting a straight line to data, and come to a dangerous conclusion, like Isobe et al. 1990. That paper is not universally applicable and is often wrong. Do not use bisector fits!

    Really. I mean it. If the data has significant intrinsic scatter, and especially if it has a selection limit on one variable, DO NOT USE BISECTOR FITS.

  • Counterfly

    Actually, sometimes there is a help line, or a back-of-the-book, and it’s simply a colleague who is much smarter than you.

    *Those* moments, where someone solves immediately a problem that’s been bugging you for months, is brutal in its assessment of your character: you should be HAPPY, and have no EGO. Woo!

  • Aurelie

    Great reminder, especially in those days of science-oriented or geeky TV shows where there’s no downtime and the mystery is solved in 45 minutes.

  • Julianne

    Counterfly — but often it’s that the other person just has a different set of tools. It has nothing to do with your character. It took me a week to get through solving the PDE’s (analytically intractable and numerically unstable bastards that they were), but, one of my N-body/hydro colleagues could probably have gotten it working in a few hours. It’s just not a skill I had bothered to develop during the past 1.5 decades.

  • Pingback: Some links science-y and academic « Entertaining Research()

  • Phillip Helbig

    When I read the title of this post, I immediately thought of the Oblique Strategies.

  • Lab Lemming

    Good Luck with it. Hopefully you’ll meet an axle-maker at your next conference looking for a car to test drive their new drivetrain on (I’ve seen it happen).

  • Fermi-Walker Public Transport

    Agree with Counterfly, those people who can do in hours what would take me weeks or months I politely ask to become a collaborator.

  • somebody

    I would be *happy* if there was someone out there who could solve my problem in a flash, so I can get on with my work instead of being stuck. I just want to solve my problem by hook or crook – the ego aspect is very secondary. But most of the time, when you are stuck, NO ONE can help you. You just have to thrash around in the darkness like Julianne says, until you stumble upon an out.

  • Phelipe

    I learned the hard way the main difference between having a senior professor or a young assistant professor as PhD supervisor. The first one still won’t know the answer you are looking for, but he will most likely know the things that won’t work, and will let you know if you care to tell him you will try it. The younger one won’t know it doesn’t work (he probably never tried it himself, and is too busy writing a funding proposal to actually do it now), so you will have to try it and then convince him you are not doing anything wrong, and that it actually doesn’t work. This can go on for quite sometime, until you accidentally step on a possible solution…

  • Ryan

    (or, in my case, about solving a network of partial differential equations where the time dependence of three quantities depends on the spatial variation of 1 or more of the others. Ick.)

    Regrettably, reality is usually best described by these sorts of equations. The upside is that it keeps anybody who has to do fluids (for example) employed

  • John R Ramsden

    Worse than the “muttermutter” stage is the “AHA! (leave overnight, and the next morning ..) BLAST!” roller coaster.

    The famous maths collaborators G H Hardy and J E Littlewood had differing opinions on how to leave things after the initial “AHA!”. Hardy liked to sleep on it, with a glow of satisfaction that his idea or result was probably right even if his hopes might be dashed the next day, whereas Littlewood preferred double checking immediately to be absolutely sure.

  • Julianne

    John — There’s also the opposite “BLAST! (leave overnight, and the next morning…) AHA!” experience. That’s a lot more fun.

  • thales

    I’m no professional scientist – just a wannabe – but I wonder why there isn’t some kind of Craigslist-style classifieds for solving thorny scientific or math problems. As Tyler Cowen often points out at Marginal Revolution, there are “markets in everything.”

    Example: you’ve spent a frustrating month trying to solve a network of partial differential equations. So you go to, say, and post your problem, promising the first person who can solve it to be credited as a collaborator. Maybe you promise a small cash payment as well.

    My hunch is that such a system hasn’t already been put in place because there’s too much pride at stake for the researcher/scientist to risk. Or perhaps there is such a system and I am simply unaware of it.

  • Charon

    thales: usually the people who could solve your problem have problems of their own to work on, and however interesting you think the problem is, they have different priorities.

    However, such collaborations do sometimes happen. Example I’m aware of: an astronomer working with a computer scientist, because both skill sets are needed to solve the problem.

  • Richard

    I’ve been working for a long time in my spare time on a long paper, and I recently became suspicious that a theorem in it was not passing the sniff test, so I did some research in old papers and in fact found a few counterexamples. I had already spent substantial time building some carefully constructed additional theory on top of that result. The problem originated from a mistake in a proof of a proposition on page 122, so the entire paper was not a house of cards, and not everything that followed depended on it, but nevertheless I felt a little shocked, fried and paralyzed for a week or so. Now I’m starting to work on it again, first salvaging whatever I can and then moving on. I was shocked because my memory of those old papers had failed me, and my proofs are normally carefully constructed with no “it’s clear that …” that actually requires a page of proof, but here in this one proposition I got careless and did just that, and did not catch the problem on a second reading either.

    This reminds me of someone else’s even much worse experience. A graduate student once told me that he had gotten half way through typing up his thesis (yes, I come from pre-computer days, and typing math was painfully tedious back then) and discovered that a lemma that he and his advisor had assumed was obvious was in fact false. He had to start a thesis all over again from scratch.

  • Stephen Serjeant

    I heard a story of a thesis examiner in astronomy who made a grad student repeat the viva when it turned out that the student couldn’t put names to images of local group galaxies. Yikes. Perhaps not as bad though as the story I heard of a PhD viva on numerical simulations, which started out with the examiners and student talking about the introductory chapter, when one examiner said “Of course you do realise equation 2 doesn’t conserve energy?” It had been taken from a text book which had an error. Ashen-faced student seen rushing out of the viva to print out her source code… It turned out there was a bug which cancelled the error. Having been on both sides of the fence, I’m sure the examiners were almost as keen as the student to find that bug.

    Getting back to the subject in hand, I think there are interesting possibilities in using social networking software in science. There have been conferences where attendees tweet, and I’ve heard of technical questions coming up in grant panel discussions which have been answered in 10 seconds through Twitter. At some conferences, attendees have done real-time ADS searches to address questions from the floor or even refute a speaker. Peer review in seconds! Wow, buzzy. But astronomy, at least as I’ve experienced it, thrives on collaboration. Can we do more with web 2.0? (Academia reclaiming the web! :) For me, the main obstacle isn’t pride, it’s trust. I like interacting with other researchers and being free with ideas, but (mentioning no names) there have been a couple of times when I’ve felt an idea of mine has been taken without attribution or collaboration – nothing earth-shattering, but enough to feel a bit burnt and cross. You can be careful who you talk to in a corridor, but Craigslist talks to anyone and everyone. So I think people would need to be a bit guarded on a Craigslist-type scientific match-up listing. Still, I think thales has an interesting point that we could make more of web 2.0 social networking in science. What do other people think?

  • Yvette

    I once wrote a sci-fi story that was in the ‘finding aliens’ genre that turned out to be quite challenging because, well, I couldn’t get past the fact that most of my protaganist’s work would be pretty boring. And I think regardless of any hype on the issue I’m still right on that point.

  • coolstar

    I think thales and SS are on to something here. There’s really no reason (other than ego) to spend months working on a problem that others you know can solve in hours. Swiss-army knife? It’s silly to add another tool that you may never use again. If you really need that skill, you can certainly learn it quicker with help than without. Of course, like SS, I know of horror stories in astronomy, such as ideas being stolen off of grant or observing proposals. There are also enough people forced out of their field due to overproduction of phds to solve LOTS of difficult problems……

  • Anonymous

    I saw a case that was much worse than having to repeat a viva. A student built an entire dissertation around a particular graph theoretic theorem but didn’t bother to even pre-flight the idea with the very famous graph theorist down the hall much less invite him to be on his committee. The student wasn’t the sharpest tool in the shed in any case.

    Then came the public defense. The student started by stating his theorem. The afore-mentioned famous graph theorist raised his hand and pointed out a simple counter-example that extinguished in one stroke both the theorem and all of the consequences that filled out the dissertation.

    The sad thing is that the student’s committee and adviser invented a face-saving way to “amend” the dissertation and pass the student. Essentially the result wound up saying “This theorem is provably false, but if it were true then we could do all of these things”. The ultimate result was that the student substantially decreased the value of the paper he printed his dissertation on by the simple act of printing it. This guy actually got an academic job and then proceeded to terrorize students for many years with unanswerable questions collected into non-sensical exams.

    The moral for students is to ask your peers and professors for feedback. And the moral for committees and advisers is to force your students to ask for feedback. And the moral for all of us is that we shouldn’t give PhD’s to people just to get them to leave.


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Cosmic Variance

Random samplings from a universe of ideas.

See More

Collapse bottom bar