Can we trust science from China?

The metascientific implications of authoritarian censorship

SARS-CoV-2 emerged in China. There is an unusual lack of zoonotic evidence that contrasts with a mountain of evidence suggesting a laboratory origin. If SARS-CoV-2 emerged in a lab, it emerged in a lab funded by the Chinese government, creating a major conflict of interest.

Were the Chinese government a democratic one that safeguarded the freedom of speech, then scientists in China would be free to speak about evidence suggesting a lab origin. They could openly debate both sides of this important scientific issue, present the strongest evidence regardless which theory it supported, and broadly I would say the science from China could be trusted.

Of course, the contemporary Chinese government is not a democratic one and it does not safeguard the freedom of speech. On the contrary, the Chinese government has extensively suppressed speech on COVID-19 epidemiology and the origins of the virus, and the Chinese government passed laws specifically forbidding any speech that “undermines social stability” or criticizes the central government. So, what happens when scientists need to investigate the socially destabilizing possibility the central government supported risky research that created a pandemic, killing perhaps 1 million people in China?

Scientists in China are governed by these same laws. It is entirely consistent with the behavior of the Chinese government and similar authoritarian states to presume scientists are forbidden from publishing work that counters the narratives and interests of the central government, even if it’s true that the central government narratives are false. I don’t fault the scientists themselves absent some other information suggesting they’ve forgotten how to distinguish between what is true and what is false, but as a scientist I feel we need to consider the uncomfortable possibility that science emerging from authoritarian nations pollutes our global, democratic common pool of scientific information.

Consider a hypothetical scenario where an authoritarian leader implements a policy at great reputational risk (e.g. a lockdown or, for imaginative purposes, mandatory cigarettes for kids). If that government also has the power to fund scientists and censor scientific publications, they may only permit studies showing that the Dear Leader’s policy is enormously effective. They may even be able to get sufficiently connected in western peer-reviewed journal systems to publish their results, pointing the papers to favorable editors with the same national origin or loyalty and recommending reviewers with similar dispositions or coercibility.

While the rest of the world publishes results irrespective of which theory they support, this hypothetical authoritarian regime would only allow publication of results supporting the Dear Leader’s theory. Anyone running a meta-analysis would be obligated to include all studies that appear in their search, and that would include a perhaps large number of studies funded & allowed by the Dear Leader suggesting their policy is effective at moving key performance indicators in the preferred direction. Such a systematic bias of scientific literature would tilt our innocent scientific analyses towards the narratives of authoritarian who sought to pollute our scientific information ecosystem.

Authoritarians like Xi Jinping and his communist party, or Vladimir Putin and his Russian oligarchs, prioritize their geopolitical ambitions and domestic support above all else, including things we lovers of the free world hold dear like truth, science, accountability, and impartial deliberation. They are well known to wage misinformation and disinformation campaigns in social media and mainstream media outlets - it would be naive to assume these ambitious adversarial authoritarians believe our scientific journals and scientific information ecosystems are off-limits. On the contrary, if the Chinese government knew SARS-CoV-2 originated from a lab, the truth would be so contrary to the national security interests of the Chinese government that one ought to expect multi-pronged efforts to deflect attention and disinform the world across social, mainstream, and scientific media.

We like to think that science cannot be polluted because it is based on reproducibility. If someone runs an experiment and claims to find X, but others can’t replicate it, then the irreproducible theory is thrown into the intellectual garbage bin. Reproducibility does indeed buffer against scientific disinformation, but there are some fields of science and some scientific questions that require irreproducible pieces of evidence.

For example, suppose members of the Chinese Academy of Sciences were to publish a serosurvey of blood donors in Wuhan in the fall of 2019. There is no way anyone outside of the jurisdiction of the Chinese government could replicate that finding. In fact, that study was released earlier this year and claimed to have found no evidence of widespread infections in Wuhan in the fall of 2019. Incidentally, that claim supports the idea that there was not an earlier and massive outbreak in Wuhan, and the idea that there was not an earlier and massive outbreak in Wuhan is a key line of defense against a possible lab origin of SARS-CoV-2. In other words, an irreproducible serosurvey from a Chinese state-affiliated academy published a serosurvey, and the rest of us are now obligated by informal scientific norms to cite and consider that paper, often without the informal liberty of being able to exclude the study by saying “I don’t trust irreproducible findings published through the Chinese government”.

The serosurvey that made it through the Chinese government’s filter contradicts serosurveys finding evidence of earlier infections outside of China, findings supporting the idea of an earlier and more widespread outbreak within Wuhan in the fall of 2019, an idea lending support to the theory that SARS-CoV-2 came from a lab, a theory that threatens the national security of the Chinese government.

Again, I don’t fault scientists in China insofar as they retain the ability to tell truth from falsehood and maintain a commitment to ethical research. Many other groups of people have their biases, including groups delineated not by national identity but paradigmatic belief. We expect string theorists to keep publishing on string theory, zoonotic theorists to keep publishing on zoonotic theory, and so on. However, science is intended to be a competition between hypotheses, paradigms, and theories, with scientists granted the freedom to change their minds and publish their updated opinions. That freedom - to change their mind and publish contradictory work - does not exist in China, and this provides the Chinese government the power to pollute science with bias by censoring scientists within China. We have to consider the possibility that trusting science from China is providing the Chinese government undue access to influence our global scientific system by coercing or filtering scientists within its jurisdiction.

Insofar as there are scientific theories threatening the power and reputation of authoritarian regimes, there will be efforts by the authoritarian regimes to interfere in the competition between scientific theories. Scientists rarely consider the geopolitical vulnerabilities of our community - we are collegial philosophers lounging around the academy, often without a care in the world except for our next paper, grant, or response to reviewer 2. Yet, the world depends on our scientific ecosystem, our scientific ecosystem depends on our ideological freedoms and research ethics, and authoritarian regimes may not care about ideological freedoms or research ethics. We scientists may be philosophical Dodo’s on an island invaded by authoritarians who don’t care about the balance of our system built by and for liberal democracies.

Can we trust science from China? If it is reproducible then yes, but the beauty of reproducibility is that it is trustless - I don’t need to trust anyone when I can run an experiment for myself. Fields of science, like evolutionary biology and epidemiology currently grappling with SARS-CoV-2 origins, historically rely on trust of irreproducible findings. Now, as our scientific efforts cross the line to threaten the national security of an authoritarian regime, we must consider the obvious possibility that authoritarian nations, and the science they publish, are not to be trusted.