How easy would it be to dig yourself out of this hole?

i-f898296c593875af5b5a6c85dfbbd850-Cantor.jpg
We just hit the point in the semester where my "Ethics in Science" class discusses the novel Cantor's Dilemma by Carl Djerassi. For those who inhabit the world of scientific research -- and for those who don't but are hungry for an insight to how human relationships and scientific activities are entwined -- it's a nice little novel. (Indeed, I've discussed it already in a couple other posts.)

What I'm going to discuss in this post is a situation that's pretty much at the end of Cantor's Dilemma, a situation where my view of what was most likely to happen after the last page (in Novel-land, where the fictional characters go on with their lives after we close the book and put it back on the shelf) turns out to be very, very different from my students' views of how things would probably go for those characters. I'm curious to know whose reading of the likely outcomes seems most reasonable.

But, to lay that out, I need to give you details about where things are at the end of Cantor's Dilemma.

If you have not yet read Cantor's Dilemma, and if there is even a remote possibility that you might read Cantor's Dilemma at some point in the future, and if knowing how the novel ends has any non-zero probability of taking the fun out of your future reading of this novel (as I imagine it would for me), then for goodness sake do not read any further in this post! This post will be loaded with spoilers. Not just minor spoilers, either. To really explain the situation at the end of the novel about which my students and I disagree, I need to spoil most of what there is to spoil.

I've warned you. Choose carefully.

OK, here's the situation at the end of the novel:

Cantor, the title character, has just won the Nobel Prize for a revolutionary tumorigenesis theory. He shares the prize with his former post-doc, Jerry Stafford, an experimentalist with serious hand-skills who conducted a successful experiment supporting the theory.

Probably.

About halfway through the novel, events push Cantor to question whether he can trust Jerry's experimental work. One of those events is word from another scientist who is kind of a friend and definitely a competitor, Krauss, that attempts to reproduce the experimental results in the Krauss lab have not been successful.

Cantor gets squirrelly and devises another experimental test, conducting this experiment himself. Jerry is aware that he's being frozen out by Cantor, so he bails, heading off to start another post-doc -- with Krauss. The word that Cantor and Jerry have been awarded a shared Nobel Prize freaks Cantor out, since he's confident about his theory (and his own experimental test of it, not yet published) but scared that the experiment Jerry did to "support" it might not hold up to scrutiny.

By the end, Cantor and Jerry have actually dealt with their issues. Jerry has essentially cleared up all the stuff that gave Cantor reason to doubt his work. But in the process, Krauss has figured out that Cantor had reason to doubt the bit of Jerry's work that was the basis of the shared Nobel Prize he accepted.

Krauss uses this information to blackmail Cantor into nominating Krauss for a Nobel Prize, annually, until he wins one. Here are some of the key bits from his letter to Cantor:

Was there something not quite kosher about that first Nobel Prize experiment? And if there was, who did that experiment? And who was in my lab when Ohashi [Krauss's post-doc] finally succeeded on the third try? Admittedly, Ohashi himself told me about the scintillation counter calibration, but we all know only too well the save face syndrome. Maybe he just made that up to explain why he failed the first two times in Stafford's absence. ...

Jerry Stafford announced in his Nobel lecture that we did confirm his experiment but, fortunately for everyone concerned, our article has not yet appeared. I trust you will not be too shocked to learn that I have withdrawn Ohashi's paper from publication. You need not worry -- it was done in a very low-key fashion by just indicating to the editor that we wanted to check a few outstanding points. After all, it never hurts to be overly cautious ...

On rereading this letter, I just realized that I neglected to write about the confirmation of your experiment [by Jerry Stafford in Krauss's lab]. Given the faint cloud hovering over Stafford (which, incidentally, you could dispel instantly), I suggest that we let that one dangle as well. There really isn't any hurry -- after all, you, like Stafford, have already reported your work in a Nobel lecture. (pp. 226-227)

My students seemed pretty unified in the view that it would be quite easy for Cantor to get out from under Krauss's thumb. Just get some other lab to confirm the result (not that there would likely be a publication for them for a mere confirmation of someone else's finding) and the theory is vindicated. What could Krauss do then?

But, I asked, what about the fact that Krauss had at least good circumstantial evidence that Cantor was in doubt about Jerry Stafford's work and didn't make any moves to investigate it, adopting a "see no evil" approach instead? Heck, the guy flew to Stockholm to accept a Nobel Prize awarded on the basis of a paper which privately he doubted -- a lot. Wouldn't that piece of information, if Krauss let it out, badly hurt Cantor's own credibility as a scientist?

Eh, said the students, why should anyone believe Krauss? Now that Cantor and Jerry have talked through the misunderstanding, they just need to get their story straight about how the doubt was resolved ... and, probably, agree upon a date by which it was resolved well before the Nobel Prize announcement.

Agree on a lie?. I asked. (This is an ethics class, after all.) Not a very important one, they replied.

Now, I'll allow as how Krauss's reputation might well be shot if he were revealed as a blackmailer of his fellow scientist. But I'm not convinced that all Cantor needs here to maintain his good reputation is empirical proof that his theory holds up. Surely some part of his scientific reputation rides on acting on doubts he may have about the work of his collaborators (where "acting" is something more rigorous than sticking his head in the sand). And to my mind, Krauss must be counting on Cantor's concern to maintain his own scientific reputation, which is why he doesn't seem worried that Cantor will expose him as a blackmailer.

But, it's entirely possible that I'm overthinking this. What do you think?

More like this

I think we see this a lot in how long it takes for papers to get retracted to actually get retracted. I think the bar for retraction should be lower, or if there are mistakes, or what have you, some outlet for retracting and republishing experiments. It will never happen, but I think it should. I denounce all of the characters!

You are overthinking it, I believe. Getting the jitters about some work you didn't do yourself when you suddenly realize a major career event is hanging in the balance is hardly abnormal, and he did do a follow up himself just to ease his own mind. No need to even fabricate a story afterwards.

Or is there some unwritten rule I missed that forbids anybody from having doubts and further questions about your own work, only others'?

Scientists are as subject to doubts and fears as the rest of humanity. We have all woken up at might sweating over: Is that paper (or report) as thorough as it should be? Have I covered everything? These can be accentuated by personal insecurity, fear of failure, fear of humiliation in front of peers etc. Perhaps Cantor's doubts were just manifestations of that? In fact, he might subscribe to this explanation as a post-hoc rationalization. Is the novelist exploring the difference between psychological doubt that undermines our personal wholeness, and scientific doubt, the type that is supposed to be the "beginning of wisdom"?

By Toby Joyce (not verified) on 02 May 2007 #permalink

Isn't the whole point of the nobel prize that it is only awarded to work so important and fundamental that the cononical experiments are being done routinely in undergraduate labs by the time the prise is awarded?

Then entire premise seems hokey.

Can you imagine people puzzling over whether or not DNA is really a double helix?

Whether PCR really mass-reproduces DNA?

Whether radium and polonium can really be extracted from uranium ore?

If the question is whether or not the experiment is RIGHT, the question is absurd. As for whether or not the discovery was dishonest, then just look to the Watson / Crick / Franklind situation for a historical analogue.

Maybe something wasn't entirely on the up-and-up, but they got the answer by hook or crook. And ultimately, in science results are all that matters.

Only my second post but a great topic. I agree with Janne about the jitters. It seems to me like Cantor was really played. First Krauss injects doubts about the process then uses blackmail after he is able to reproduce the results. Cantor should have said "published and be damned" and tell the truth. Krauss would come off in a worse light and people (scientists are people) are willing to forgive actions like Cantor's. He may lose the Nobel but would retain his reputation.

I agree with Janne. Doubts are pretty normal - both inside and outside the scientific community.

I also feel like I'm missing something here. I haven't read the book, but from what I read of your summary, Cantor DID set-up a second experimental test to prove to himself that his theory was right regardless of the post-doc he had working on the experiment.

He may not have published the results, but he confirmed them - this wasn't a public confirmation of public doubts, it was a private confirmation of private doubts, and to me it shouldn't matter to Krauss or the rest of the scientific community whether or not Cantor had reason to doubt if his private doubts were laid to rest.

Little things like this happen all the time in labs during experiments. "Hey, no I think you're doing it wrong. Oh, no, you got it right." These things can be self-corrected and don't require intervention from the outside world.

Fabricating a story at that point would be a Very Bad Idea (tm). There was nothing unethical about their work to begin with, so doing something unethical to quickly resolve the issue would be career suicide if it were discovered. Its not a risk worth taking given the circumstances.

By Brian Thompson (not verified) on 02 May 2007 #permalink

On a tangent, I have a hard time swallowing a premise in which the notoriously conservative Nobel committee awards a prize to someone whose work has not been thoroughly replicated in numerous other labs.

By Ambitwistor (not verified) on 02 May 2007 #permalink

Let's swap the stakes: Nobel for human lives. To be consistent, wouldn't Cantor be expected to uphold the same standards in either case? (I'm assuming it obvious that he would be acting unethically if his hesitation to raise doubts could have endangered lives.)
As for Krauss, he too acted in a calculating and self-serving manner.

Awarding a Nobel prize for work that would be undermined by finding a single study to be flawed? I don't think so.

While I tend to agree with Janne and those who took her position, it is Krauss who bothered me much more. We must recognize that our vocation is a human activity and, as the case is for any other human activities, we have our own scumbags who are willing to ignore ethical rules in their attempts to gain fame, riches and/or power. Cantor's should expose Krauss for what he is and not worry about the effect the story of his doubts may have on the Nobel committee.

By S. Rivlin (not verified) on 02 May 2007 #permalink

I concur, the premise pegs the Bogo-meter (My own invention measureing how bogus a story is.) The time lag for Nobel recognition is measured in years. The Nobel committee is fundamentally conservative, and there would be independant conformation. Thirdly, there were private doubts which were resolved. No need to make up a story. Finally, independent researchers doing this kind of work have egoes the size of Chicago. Think Linus Pauling--blackmail, he'd tell 'em to go pound sand!

By William the Coroner (not verified) on 02 May 2007 #permalink

Yes, the premise of such a quick turnaround between discovery and Nobel Prize is unrealistic. But given that this is a novel, I'm prepared to suspend my disbelief. I understand that the ability to suspend disbelief about various premises is a personal matter, though.

As for the status of Cantor's doubts, let me give a little more detail about my interpretation of the situation (after which, of course, you are welcome to tell me again that it's no big thing).

Jerry Stafford was entrusted with an experimental test of Cantor's theory (and asked to get it done, ASAP, while keeping what he was doing secret even from the rest of the Cantor lab).

Jerry reported the results Cantor hoped to see.

Cantor wrote it up and sent it off to Nature, expecting it would make a very big splash in his scientific community.

Krauss told Cantor his post-doc was having a hard time reproducing the results; could they give some guidance (maybe photocopies of the relevant notebook pages)?

Cantor asked Jerry to run the experiment again, with Cantor by his side, so Cantor himslef could be sure it really worked. It did, but ...

Cantor received an anonymous note indicating that Jerry was in Cantor's private lab alone during the space of time when he was doing the experiment "with Cantor". In other words, maybe Jerry had to rig the results!

After this, Cantor devised a different experiment to test his theory, and conducted that experiment all by himself so he wouldn't have to worry about anyone screwing with it.

However, he didn't:

  • ask Jerry about the note, or what he was doing in Cantor's private lab alone if, in fact, he was there alone
  • try to run that first experiment himself
  • ask anyone else in his lab to try to run that first experiment
  • make any efforts to contact the editors at Nature to ask that they hold off on reviewing the manuscript until he had made sure of one little outstanding detail.

In other words, Cantor behaved an awful lot like someone who didn't want to know what the situation was with Jerry, or with that first experiment. The doubt, to my mind, isn't strictly private when failing to take steps to resolve it may mean that a scientific publication with your name on it -- something you're presenting to the rest of the community as knowledge -- is much less certain to you than the rest of the community would have reason to believe.

Are we really completely comfortable counting on the scientific testimony of someone who pulls this sort of evasive maneuver when faced with an inconvenient possibility about his collaborator?

An "evasive maneuver" is a human frailty. I think our dependence on Cantor's scientific testimony depends ultimately on whether the experiment was truly reproducible, not on whether he is a straight guy, or whether he had/ had not the moral courage to confront Jerry.

From what we now know about Cantor, we probably would not go into business with him, but I think we can accept him as a good scientist with human weaknesses.

By Toby Joyce (not verified) on 02 May 2007 #permalink

OK, it looks like Cantor used fairly bad judgement, which is, unfortunately, a very human trait. I think his best course of action, given the initial bad decisions about how to treat his doubts, would be to simply say to Krauss (as quoted from Arthur Wellesley): publish and be damned. Then tell the truth, that he had doubts but cleared them up. Admit to a perhaps ill-advised reluctance to sink what he personally believed to be an important finding. Sum it up: I was convinced of the correctness but had some doubts about the experiment confirming it, so I checked and rechecked until I was satisfied.

What disturbs me is the reaction of your class. How could they possibly suggest an unethical course of action in an ethics course? Even if they might have done it themselves, aren't they smart enough to recognize the fundamental stupidity of making that suggestion in an ethics class? Or are you teaching a class in an elementary school?

A scientist's bias is unavoidable. Cantor is biased in favor of his hypothesis. His experiments are designed with that bias inherited in them. That is why it is so important that others try to reproduce those experiments. Krauss appears to be the right scientist to reproduce the experiments since if he has any bias, he is probably hoping to refute Cantor's hypothesis. His experiment (performed with the help of Jerry) did corroborate Cantor's hypothesis. This is good enough for science.

I still find the blackmailing by Krauss the most egregious when compared to Cantor's or Stafford's shortcomings.

By S. Rivlin (not verified) on 02 May 2007 #permalink

In order to play along I'm going to have to pretend it's some other prize since, as others have pointed out, the Nobel premise beggars belief. That said --

Are we really completely comfortable counting on the scientific testimony of someone who pulls this sort of evasive maneuver when faced with an inconvenient possibility about his collaborator?

We'd better be, since there are nearly no scientists of my acquaintance who could be counted on to do better. In fact, I wouldn't bet serious money on any working scientist to do better (ethically) than Cantor did.

I agree that Cantor showed himself less than trustworthy -- I'd never do business OR science with him. His behaviour showed pretty clearly that he valued personal prestige above the production of reliable information about the world; I think that's despicable. But I'm an outlier and, if his theory held up, all would be forgiven by the scientific "community" and Krauss would be the bad guy for trying to blackmail the "successful scientist".

Winner takes all, and the ends justify the means -- that's the system, and if we don't like it we'd better start changing the way science is taught and research is done.

I'm willing to suspend disbelief when I start reading a novel. For example, I'm quite willing to grant the author that vampires exist. "OK, we've got the night-walking immortal bloodsuckers — let's start the story!" However, I'd probably get at least a little exasperated if the author then declared that vampires are also unicorns. I mean, that's just exceeding your credit rating.

Furthermore, the more disbelief I have to suspend, the less likely I am to extract any ethical lessons which I would even think of applying in my own life. As Neil Gaiman says, vampires are a way to talk about death, and fear of death, sex, and fear of sex — and what else is there? I could easily imagine a scenario in a vampire story which would serve as a "rehearsal" for a real-life dilemma. But when you bring in the unicorns, adding new story elements which are bizarre even by your original, fictional standard (and perhaps logically inconsistent to boot), you're not "rehearsing" for anything.

In this case, I'd first suspend disbelief when the "revolutionary tumorgenesis theory" is introduced. OK, we've got our plot device; it's credible enough within the fiction. Maybe it's a MacGuffin, and the internal details don't matter. (We don't need to know the details of whatever theorem Matt Damon is proving in Good Will Hunting; all that matters is that the characters react in sensible ways to one another given the fact that the math is esoteric and revolutionary.) That sets the story's credit rating. Exceed it, and no lessons can be learned.

Furthermore, what fictional analog of the Nobel Prize could have a prestige remotely comparable to the real thing if it did not have the real Nobel's standards? One could imagine a rather bleak and cynical novel in which a prize brings fame and fortune even though the work it honors is often dreck, but that's a novel set in Hollywood, not Stockholm.

One premise is flawed. In a great deal of modern scientific operations the PI doesn't really have to technical know how, reagents and hands to actually replicate experiments him/herself.

but here's a thought. are you really this naive about the current state of "hot" science? what the supposed pinnacle Science/Nature/Cell paper really is all about? Take a look at the number of retractions and corrections that appear in top journals. " a placeholder figure was left in" (who *does this* honestly). "it was the fault of postdoc X who we can't locate or won't sign the retraction" (please).

Spouse-of-Drugmonkey is in one of those labs. Everyone knows who the data fakers are, who never has an experiment come out other than as hypothesized. If the PI doesn't know it is because s/he doesn't WANT to know.

I have not read the book, but I assume the entire reason for the book is the ethical issue the main character faces. Given that, I think the criticisms about how real-life it is are beside the point. Such shortcomings would matter a lot more if the science were the point of the book, but it's not. So I think it's reasonable to give a pass to both the science and the compression of the Nobel process. The compression of the Nobel process is the type of thing that's done in many, many dramatic treatments of real-life situations. Accepting them as dramatic conventions should be no harder for us than watching a movie in which the camera switches between full-face shots of characters supposedly engaged in face-to-face conversation.

In other words, the story-telling weaknesses are not what this discussion is about.

Cantor seems to be the least sleazy person of the three, by a country mile. From your description, Krauss is a manipulative bastard, Jerry's a mercenary, and Cantor is caught in the middle. People get things right by accident in science all the time; I can't see him being guilty of anything other than that. Failure to be perfect is not a crime, it is reality.

You're also forgetting the human aspect. All he has to say is something like

"Yes, the stress of the nomination may have caused me to doubt, temporarily, that the results were reproducible. But while I had doubts, I had nothing that even came close to being actual evidence, and destroying the career of a colleague requires at least some proof."

And everyone is going to have flashbacks to that difficult partner/supervisor/student they had, and be grateful that people like Cantor exist, people who don't throw you to the wolves just because they can.

If he has the guts for it, he could even spike Kauss' guns (since prizewinners tend to get interviews, and I'm sure at least one is going to go after the human interest angle).

"The amount of stress you're under between the nomination and the when they actually hand you the award is incredible. I can't say ... You get a bit crazy. I started to wonder if the results were real. Maybe I'd dreamed it all, maybe somehow we'd messed up and it was all going to fall apart. I even ran the experiments again, myself, just to be sure."

Never tell a lie when a partial truth will do.

(Of course, I'm in high tech, so that may be more targetted to my ilk than yours, but I suspect human nature is pretty universal.)

Are you familiar with the LabLit ezine? The editor worked in my friend's lab before moving into the publishing biz. I know that she welcomes new contributors, and it seems like you and your readers might be interested in the concept, if not the site.