On July 5, 2021, a Correspondence was published in The Lancet called “Science, not speculation, is essential to determine how SARS-CoV-2 reached humans”.1 The letter recapitulates the arguments of an earlier letter (published in February, 2020) by the same authors,2 which claimed overwhelming support for the hypothesis that the novel coronavirus causing the COVID-19 pandemic originated in wildlife. The authors associated any alternative view with conspiracy theories by stating: “We stand together to strongly condemn conspiracy theories suggesting that COVID-19 does not have a natural origin”. The statement has imparted a silencing effect on the wider scientific debate, including among science journalists.3 The 2021 letter did not repeat the proposition that scientists open to alternative hypotheses were conspiracy theorists, but did state: “We believe the strongest clue from new, credible, and peer-reviewed evidence in the scientific literature is that the virus evolved in nature, while suggestions of a laboratory leak source of the pandemic remain without scientifically validated evidence that directly supports it in peer-reviewed scientific journals”. In fact, this argument could literally be reversed. As will be shown below, there is no direct support for the natural origin of SARS-CoV-2, and a laboratory-related accident is plausible.There is so far no scientifically validated evidence that directly supports a natural origin. Among the references cited in the two letters by Calisher and colleagues,1, 2 all but one simply show that SARS-CoV-2 is phylogenetically related to other betacoronaviruses. The fact that the causative agent of COVID-19 descends from a natural virus is widely accepted, but this does not explain how it came to infect humans. The question of the proximal origin of SARS-CoV-2—ie, the final virus and host before passage to humans—was expressly addressed in only one highly cited opinion piece, which supports the natural origin hypothesis,4 but suffers from a logical fallacy:5 it opposes two hypotheses—laboratory engineering versus zoonosis—wrongly implying that there are no other possible scenarios. The article then provides arguments against the laboratory engineering hypothesis, which are not conclusive for the following reasons. First, it assumes that the optimisation of the receptor binding domain for human ACE2 requires prior knowledge of the adaptive mutations, whereas selection in cell culture or animal models would lead to the same effect. Second, the absence of traces of reverse-engineering systems does not preclude genome editing, which is performed with so-called seamless techniques.6, 7 Finally, the absence of a previously known backbone is not a proof, since researchers can work for several years on viruses before publishing their full genome (this was the case for RaTG13, the closest known virus, which was collected in 2013 and published in 2020).8 Based on these indirect and questionable arguments, the authors conclude in favour of a natural proximal origin. In the last part of the article, they briefly evoke selection during passage (ie, experiments aiming to test the capacity of a virus to infect cell cultures or model animals) and acknowledge the documented cases of laboratory escapes of SARS-CoV, but they dismiss this scenario, based on the argument that the strong similarity between receptor binding domains of SARS-CoV-2 and pangolins provides a more parsimonious explanation of the specific mutations. However, the pangolin hypothesis has since been abandoned,9, 10, 11, 12 so the whole reasoning should be re-evaluated.