the Gell-Mann Amnesia effect

Murray Gell-Mann won the 1969 Nobel Prize in Physics for his work on the theory of elementary particles; naively put, he invented quarks - one of the handful (embarassingly so) of the smallest (as of yet) building blocks of matter. The Gell-Mann Amnesia Effect was coined and descibed by Michael Crichton (of the Jurassic Park fame) in his 2002 talk at the International Leadership Forum criticizing the prevelance of speculation in journalism, evoking the name of his famous physicist friend because “I once discussed it with Murray Gell-Mann, and by dropping a famous name I imply greater importance to myself, and to the effect, than it would otherwise have.” Refreshingly honest.

Chrichton describes the effect in his talk as follows:

You open the newspaper to an article on some subject you know well. In Murray’s case, physics. In mine, show business. You read the article and see the journalist has absolutely no understanding of either the facts or the issues. Often, the article is so wrong it actually presents the story backward – reversing cause and effect. I call these the “wet streets cause rain” stories. Paper’s full of them. In any case, you read with exasperation or amusement the multiple errors in a story, and then turn the page to national or international affairs, and read as if the rest of the newspaper was somehow more accurate about Palestine than the baloney you just read. You turn the page, and forget what you know.

I know nothing of the business of journalism except consuming its products in a daily ritual like any other bored skeptic. Therefore, often when I remember the Gell-Mann Amnesia Effect, it is not in the context it was originally presented in but rather in relation to my everyday work of inter-disciplinary research. I am a computer scientist by training and I’ve spent the past half-decade working on applied machine learning research for the super niche scientific field of gravitational-wave astronomy as a PhD student. Exactly how I found myself in this position is a story for another time but the AI for Science movement is in full swing and this space is typically populated by domain scientists applying ML to the problems in their specific domains. So it follows that I, as a non-domain scientist, often find myself reading ML publications on topics in astrophysics. I read these research papers, often written by physicists who are not from the ML world, and I find myself spotting certain, umm, deficiencies in the information presented. Often it is the presentation style, rigor, ease of reproducibility, computation costs of the ML analysis, a sufficient justification for even using ML in the first place and so on. I tend to chart it up to the fact that the authors are not ML experts, take the information presented with a pinch of salt and move on. But in the same publication, when I read the physics-specific part of the paper, I find myself taking that information as gospel. Because…I am not a physicist and surely the physicist who wrote the paper must certinly be an expert in, well, physics! So, sometimes, I find myself thinking of the physicist who is likely treating the ML portions of the same publication as gospel. You know the ones that made me furrow my eyebrows?

As a young student, I had assumed that the institutional systems of academia would be, by design, a bulwark against the Gell-Mann Amnesia Effect amongst academics. We are skeptics by nature and we are meant to talk to each other, provide feedback and rebuttals via a robust peer-review process and thus distill the signal from noise at the frontiers of research. But, alas, the silos of specialized disciplines populate the landscape and provide a fertile breeding ground for the Gell-Mann Amnesia to take root… even inside the ivory towers.