Well, having spent a LOT of time looking into it, I feel comfortable saying: Nope! There is no plausible theory of how cold fusion could possibly work. Fitting cold fusion into the known laws of physics is like fitting a square peg into a round hole. Over the past 30 years, people have tried hammering that peg from every angle, some with incredible ingenuity and persistence. But none of the proposals to date can actually explain cold fusion, and there does not seem to be any way to fix any of them.

There is an elegant explanation here: Cold fusion cannot be explained by known laws of physics because there’s no such thing as cold fusion! Something is wrong with the experiments—more likely, many things, with different experiments having different problems.

[UPDATE—I just wrote a whole separate blog post making the case that a reasonable person can (and should) doubt the body of cold fusion experimental results, even ignoring the theoretical aspects: The case against cold fusion experiments]

Readers may object: Maybe cold fusion cannot be explained by known laws of physics because the known laws of physics are wrong? I can see the appeal of this argument, but there’s much less to it than meets the eye. I’ll copy some text from my first blog post:

[According to this argument], cold fusion is an experimental observation, and experiments are the ultimate arbiter of truth in physics, and if theorists cannot explain an experiment, then they should get to work finding a better theory.

This sounds very nice. It sounds like The Scientific Method like we all learned in school and read about in Karl Popper. Who could object to The Scientific Method?

It sounds nice, but it’s wrong! It is rational to give experiments a complete veto over theory only if experimental results are always correct. That’s not the case! Sensors can be calibrated incorrectly. Procedures can be followed incorrectly. Results can be described and interpreted incorrectly. Experiments can be wrong for reasons that are extraordinarily subtle, reasons that are not understood for months, or years, or ever. This is not a nitpicking hypothetical, it is one of the most basic facts of life for everyone in experimental science and engineering.

Therefore it is not only extremely common in practice, but also entirely correct, to use theoretical physics to inform our guesses about which experimental results are trustworthy. In other words, we are doing a Bayesian analysis of what to believe, and both theoretical and experimental knowledge are legitimate inputs into this analysis.

(Example: Here is a link to a meta-analysis in support of parapsychology. Oh, you still don’t believe in parapsychology? Did you meticulously read that article and judge its methodological soundness on its own merits? Or did you rule it out based on prior expectations derived from theoretical physics?)

We have a theory, let’s call it “Standard Model Quantum Field Theory With Perturbative Gravity” (I wish it was more widely known and had a catchier name), which, as far as we can tell, applies absolutely everywhere in the solar system. We have tested this theory an enormous amount, including in a wide variety of extremely specific and targeted ways, and it passes every test with extraordinary, ridiculous accuracy. The theory does not apply everywhere in the universe, because it cannot offer any predictions about certain exotic situations like microscopic exploding black holes or the Big Bang. That’s what’s spurring ongoing work in string theory and quantum gravity. I’m glad people are working on that stuff, but we shouldn’t lose sight of what we already have: a “theory of everything” sufficient to explain the microscopic goings-on in every laboratory experiment that we expect humans to ever be able to perform.

It’s not impossible that our current understanding of the fundamental laws of physics will be overturned by future experimental evidence. But that evidence needs to be exceptionally good and exceptionally specific, far beyond cold fusion experiments. Remember, the meta-analysis linked above shows that parapsychology is a reliably-reproducible experimental result. Experiments are hard.

Finally, you might say, maybe there’s a way to explain cold fusion in terms of the known laws of physics, and we just haven’t thought of it yet? Well, I sincerely tried over the past 4 years, and I failed. Julian Schwinger, one of the greatest nuclear physicists of the 20th century, tried and failed. Peter Hagelstein, a good enough physicist to get tenure at MIT, has been trying for 15 years, and failed. […in my opinion … I haven’t written it up, but I’ve grown much more confident in the criticisms hinted at here.] Notice any pattern?

And no, this isn’t like the mystery of high-temperature superconductivity, an example physics phenomenon which has legitimately taken decades to understand. In high-temperature superconductivity, physicists understood almost immediately why it can happen in broad outline, and we are just having endless trouble narrowing down the details of the “pairing mechanism” out of a few plausible candidates. Cold fusion is not like that at all. In fact it’s the opposite: We understand why cold fusion canNOT happen, in broad outline, and then all this theoretical work goes into fighting against that broad understanding, hunting for loopholes, or even just special pleading.

I can’t think of any analogous situation in modern physics, where an experimental result has turned out to be real, against such strong, longstanding, and carefully-considered theoretical reasons to disbelieve it. It’s easy enough to find a story of some well-regarded theoretical physicist claiming that, for example, lasers cannot work, days before they’re first demonstrated. Haha, stupid know-it-all theorists, you say. But there’s a world of difference between a theorist breezily dismissing something off-hand, and (say) this blog, where I am dismissing cold fusion after spending countless hours thinking about it and reading all the best ideas from decades of brainstorming by sympathetic proponents.

So, readers, if you want to keep hunting for theories of cold fusion, whatever, it’s your life, you can do what you like. Maybe I’ll even keep posting sporadically myself. But I feel totally satisfied that I got the right answer.