Rebuttal of David Karger’s Defense of Cancel Culture

Pedro Domingos
5 min readJan 3, 2021


Cancel culture has descended on computer science, as on many other fields, making it harder and harder to conduct research — or even voice any opinion — that does not conform to a particular left-wing ideology. In response, a group of us has written a letter that, among other things, calls on the community to reaffirm its core principles as a scientific field. In an effort to prevent more computer scientists from signing this letter, the cancelers have conducted a campaign of intimidation about which the less said the better. But David Karger has taken the time to write a reasoned rebuttal of some of our points. I appreciate this, because reasoned debate — rather than insults, taunts and threats — is what we need more of. Nevertheless, I have to say that I am shocked and dismayed by the contents of his piece. Below I summarize the principles we spoke of, Karger’s rebuttals, and my responses to them.

1. Scientific work should be judged on the basis of scientific merit, independent of the researcher’s identity or personal views.

Karger says that this is absolutely true, but not in need of defense because blind peer review takes care of it. But of course it doesn’t: blind peer review does nothing to combat researchers being personally attacked for what they publish after it’s published, being pressured to not do non-conforming research before it’s done, or indeed being harassed and silenced at every step of the research process. Cancel culture is pervasive: it is felt at universities and companies, in research labs and in classrooms, in conferences, informal gatherings and mailing lists. People who defy it risk being fired, ostracized, antagonized, and more. None of this is obviated by blind peer review, and Karger’s argument reveals a shocking lack of awareness of cancel’s culture true nature and extent.

Karger’s point is also directly at odds with his claim later in the piece that the community is plagued by pervasive racism and sexism. If that is the case, then how is the principle that scientific work should be judged independent of the researcher’s identity not in need of defense?

Karger then goes on to speculate about what our other motives for raising this point might be. He says an attack on the scientist is not an attack on the science, but this is beside the point — the problem occurs when the science is attacked because of the scientist. He says that, for example, the recent retraction of a Nature Communications paper whose conclusions were at odds with the prevailing ideology is OK because ultimately the retraction was based on the papers’ claimed scientific flaws. Again, this is shockingly naïve: why was another paper by the same authors and with similar methodology, but that happened to agree with the ideology, not attacked and retracted? Or thousands of other such papers? Is it OK to have one standard for the papers that agree with our ideology and a different one for the ones that don’t? What does that do to the balance, objectivity and trustworthiness of the resulting body of research?

Karger then argues that, even if a scientist has done research that should be published by a conference, it’s OK to not allow her to attend the conference to present it if she has unrelated views that Karger objects to. But, even if you agree with it (I don’t), this is obviously letting objections to the person’s views affect the reception of the scientific work. So, after saying that Principle 1 does not need to be defended, he himself provides another example of why it needs to be.

Finally, Karger says that the subtext of Principle 1 is that when the science is good it is our obligation to celebrate the scientist who did the work. But there is no such subtext. Rather, the whole point of Principle 1 is to separate judgments of the science from judgments of the scientist.

2. Discussion and debate in the scientific community must be free of prior restraint as to topic or viewpoint.

Karger says that this principle is vacuously satisfied because there is no prior restraint on topics or viewpoints. Again, this reveals a shocking naïveté about the social pressures on scientists, which I already addressed a propos of the previous principle. Karger claims that (for example) NeurIPS’s requirement of broader impact statements and ethics reviews is not really a restraint on what research can be published, because it’s no different from listing the topics a conference is interested in. This despite the fact that the NeurIPS call for papers explicitly says that “‘regardless of scientific quality or contribution . . . a submission may be rejected for . . . including methods, applications, or data that create or reinforce unfair bias.” How is this not an explicit restraint on what can be published? And, in particular, an ideologically-driven one?

3. No individual should suffer harassment or attack based on their personal or political views, religion, nationality, race, gender, or sexual orientation.

Karger says that he agrees with this, except that he strongly approves of attacks based on personal and political views. He says that the attacks should not be ad hominem — but that is precisely what we’ve seen far too much of. (Is he not aware of any of it?) And, despite that statement, he seems to have trouble distinguishing attacks on the person from attacks on the views, specifically saying that we “should attack the individual on the grounds of the views they hold.” From here to a full-throated endorsement of cancel culture is only a step.

And he takes that step in his final remarks. He says: “I support the increased pushback over speech that is occurring in our field.” He fails to clarify which forms of pushback, if any, he does not support. (Does he support bullying, harassment, threats and retaliation? Where does he draw the line?) He argues that we need to push back against racism and sexism, and that it’s OK to have some false positives when detecting them because we currently have too many false negatives. On the contrary: the scales, in academia in particular, are currently tilted strongly in the direction of seeing racism and sexism everywhere, magnifying this misperception to absurd extremes, and victimizing anyone who falls afoul of it. I speak from experience: after questioning NeurIPS’s requirement of ethics reviews and broader impact statements, I was called, among other things, racist, sexist, a misogynist and a bigot. Is this the kind of false positive Karger is comfortable with? Or is he pointedly illustrating the obliviousness of many to cancel culture that is precisely what our letter is trying to combat?



Pedro Domingos

Professor of computer science at U. Washington and author of “The Master Algorithm”.