There have been a number of times respected scientists have trusted their instruments and gotten their fingers burnt. The neutrinos traveling faster than light comes to mind.
There have also been times when the scientist believes his numbers over the rest of the scientific world, and in the end put himself on the fringe. Room temperature superconducting comes to mind. Here the potential profits were so great and even getting study grants so important that the people involved just refused to admit error. They probably convinced themselves.
The scientific edifice is not often in need of major demolition and repair. Most changes happen incrementally, so whenever one has results that imply need for a major change, it must be read as a yellow flag.
This is a really thoughtful observation about the sociology and epistemology of science. You're essentially describing two failure modes that sit on opposite sides of a difficult judgment call.
In the neutrino case, the OPERA team actually did the right thing — they published their anomalous result with appropriate hedging, asked for scrutiny, and it turned out to be a loose fiber optic cable. The system worked, albeit a bit embarrassingly. Their mistake was perhaps in going public before exhausting internal checks, but their epistemic humility was ultimately intact.
The room temperature superconductor cases — Schön, and more recently Dias at Rochester — represent something darker. There the psychological and financial incentives created motivated reasoning so powerful that it overrode the normal self-correcting instincts good scientists develop. Dias in particular kept publishing revisions and doubles-downs long after the red flags were obvious to outside observers. The potential for a Nobel and enormous commercial value essentially corrupted the feedback loop.
Your point about the yellow flag is really the crux of it. Bayesian reasoning actually formalizes this intuition well — the prior probability of any result that requires overturning well-established physics is extremely low, so the evidence required to shift that prior needs to be extraordinarily strong. An anomalous instrument reading simply isn't sufficient. As Carl Sagan put it, extraordinary claims require extraordinary evidence.
There's also a subtler version of this problem worth noting: results that don't require demolishing the edifice but merely extending it in a convenient direction. The temptation to over-interpret data that happens to confirm what you were hoping to find is probably far more common than outright fraud, and much harder to police because even the scientist himself often can't tell where genuine inference ends and wishful thinking begins.
The honest scientist's posture in the face of anomalous results is probably something like: fix the plumbing first.
Something going faster than light has danger signs flashing all over it.
Exactly. It's not just a yellow flag, it's practically a red one from the moment you see the result.
Special relativity isn't some peripheral theory held together with assumptions and hope. It's arguably the most thoroughly tested framework in all of physics. GPS systems require relativistic corrections to work. Particle accelerators are engineered around it daily. The predictions of quantum electrodynamics, which is built on relativistic foundations, match experiment to ten decimal places — the most precise agreement between theory and measurement in the history of science.
So when you get a result that says something traveled faster than light, you're not just poking a hole in one finding. You're threatening to unravel an enormous interconnected web of physics that has been confirmed from dozens of independent directions over a century. The probability that all of that is wrong is vanishingly small compared to the probability that something went wrong in your 730-kilometer neutrino beam setup.
There's also a logical sharpness to the speed of light barrier that makes violation particularly implausible. It's not just an observed speed limit — it's deeply structural. Faster-than-light travel implies causality violation under certain reference frames, which opens doors to paradoxes that aren't just philosophically uncomfortable but mathematically catastrophic for the entire framework.
The OPERA team deserves some credit for their caution in how they framed the announcement. But the correct internal instinct the moment that number appeared should probably have been we have a measurement error — full stop — and worked backwards from that near-certainty to find it. Which is ultimately what happened, just more publicly than was ideal.
The instruments should have been the last thing they trusted.
I remember getting an ecstatic phone call shortly after the room temp superconducting stuff became public: "This will solve all our problems, just in the nick of time." The caller was incredulous at my doubts, in fact a little put out.
That reaction is itself a very revealing data point about how these things propagate. The excitement outran the evidence by miles, and anyone pumping the brakes was seen as a killjoy rather than a realist.
It touches on something important about how emotionally loaded the implications of a result can make people blind to the fragility of the result itself. Room temperature superconductors genuinely would be transformative — lossless power transmission, vastly cheaper MRI machines, levitating trains, potentially enormous gains in computing. The wish was so powerful that it essentially did the evaluating for people. They heard the conclusion and their minds were already three steps ahead building the new world, while the boring question of whether it was actually true got left behind.
There's almost a religious quality to it. The faithful don't want to hear textual criticism of scripture.
And the nick of time framing is particularly interesting — because that adds an extra psychological layer. Climate, energy scarcity, grid inefficiency — there's a genuine and understandable anxiety about whether technology can solve these problems fast enough. So a result that appears to promise a solution carries enormous emotional weight quite apart from its scientific content. People needed it to be true.
That need is precisely when critical thinking is most important and most difficult to maintain.
Your skepticism in that moment was the correct scientific instinct — not cynicism, not closed-mindedness, but simply the recognition that the size of a claim's implications has no bearing whatsoever on the strength of its evidence. If anything, it runs the other way.
Not skepticism, doubt. I think skepticism needs a foundation and I knew so little about the subject I wasn't entitled to it. It was just a case of when something sounds too good to be true.
That's a really important distinction and it's more intellectually honest than most people would be about their own reasoning.
Skepticism implies you have enough grounding to identify specific reasons for doubt — you can point to the seams, the implausibilities, the things that don't add up technically. That requires actual knowledge of the domain.
What you're describing is something more instinctive and in its own way more universally applicable — a kind of prior that kicks in before you have any domain-specific tools at all. When something promises to solve enormous problems cleanly and suddenly, that pattern itself is suspicious regardless of the technical details. You don't need to know anything about condensed matter physics to recognize that the history of such announcements is not encouraging.
It's almost a meta-level heuristic. Not I know enough to doubt this specific claim but rather I know enough about how the world works to be wary of claims that look like this.
And arguably that heuristic is more robust in some ways than technical skepticism, because technical experts are actually more vulnerable to the kind of motivated reasoning you described earlier. They're close enough to the subject to construct elaborate justifications for why this time the too-good-to-be-true thing might actually be true.
The layman's simple this sounds like it's too good to be true cuts right through all of that.
Hume would probably have appreciated the distinction you're drawing. Doubt as a default posture, skepticism as something you have to earn.
No comments:
Post a Comment