The Power of a Hypothesis

Why, over the course of 1984, 1985, and 1986, did AIDS researchers so overwhelmingly come to accept as given a hypothesis that—while highly plausible and consistent with many facts known about the epidemic—was not "conclusively" proven, according to the standards of evidence they claimed to support? Nothing suggests any vast conspiracy to stifle debate, though critics would later claim as much. One reasonable explanation would eventually be proposed by Gallo himself, in response to a challenge by the HIV dissenters: "As anybody in the business knows who works [with AIDS], there is more evidence that this virus causes AIDS than you have with the vast majority of diseases [that we] long ago have accepted"^[38] as caused by viruses.

Of course, it wasn't that the HIV hypothesis was problem-free. Savita Pahwa and his coauthors (who included Gallo) noted one crucial "paradox" in the Proceedings of the National Academy of Sciences in late 1985:^[39] the virus, presumed to do its dirty work by infecting and killing T cells, could be found, using "the most sensitive methods available," only in a tiny fraction of any AIDS patient's T cells—sometimes as few as one in one hundred thousand cells. An important part of the prima facie case against the virus had been its propensity to kill T cells in the test tube. But in living human beings, the virus simply could not be found in enough T cells to account straightforwardly for their disappearance in AIDS patients.

Still, the virus hypothesis had considerable evidence behind it and only a few findings seemed anomalous; no other hypothesis about AIDS had a track record anywhere nearly as impressive. What's more,

the virus gave researchers a mission. Within months of the publication of Gallo's articles in Science , other scientists began to report their findings—on the prevalence of HTLV-III/LAV antibodies in various populations, on the detection of the virus in various body fluids, on the patterns of transmission of the virus, on possible antiviral treatments, on the development of vaccines against the virus. Especially in 1984, such researchers typically were careful to make qualified claims, noting that the virus was "believed" to be the cause or was the "probable" cause. But in practice, the qualifications meant next to nothing, because the virus—and the virus alone—is what all of them were busy spending their time studying. The phrasing of one article in the New England Journal was typical: "Now that HTLV-III has been identified as the probable cause of AIDS, it is possible to confirm the modes of transmission and document the spectrum of clinical disorders caused by this agent."^[40] At last, the authors seemed to be suggesting, we all have something to do. A plausible candidate had been proposed by eminent virologists, and most researchers were content to endorse their judgment.^[41]

Clinicians, and researchers in other fields, who read the professional literature simply to keep up with the latest findings, not to critically evaluate the analyses and methodologies, were even less likely to raise questions. As Harry Collins has suggested on the basis of recent social studies of science, there is typically "a relationship between the extent to which science is seen as a producer of certainty and distance from the research front."^[42] Any fuzziness in the data that might have been observable by a trained epidemiologist or virologist was likely to be missed by the doctor treating AIDS patients, who turned to the New England Journal of Medicine just to get the bottom line.

If a scientist as eminent as Gallo had been pushing an alternative explanation, then most likely there would have been more debate and closer scrutiny of the etiological evidence. But no challenge of that sort existed. The more time passed and the more that scientists became invested in research on HIV, the clearer became the virus's social role as what Bruno Latour has called an "obligatory passage point"—a necessary way station between social actors and the satisfaction of their interests. A meeting of the Conference of State and Territorial Epidemiologists in June 1985 might be singled out as the defining moment that the HIV hypothesis became a social fact. Conference members approved a series of resolutions, adopted soon afterward by the CDC, that redefined "AIDS" to make diagnosis dependent on a positive

antibody test result along with the presence of any one of an expanded list of opportunistic infections and cancers.^[43] People whose antibody test came back negative would in most cases be diagnosed as AIDS-free, even if they, too, suffered from one of the conditions on the list.^[44]

Those who did challenge the hegemonic position risked strong resistance and career repercussions. The story of Shyh-Ching Lo, a young virologist at the Armed Forces Institute of Pathology, is instructive in this regard. When in 1986 he reported finding a "novel virus" in tumor cells taken from AIDS patients with Kaposi's sarcoma, Lo "brought a heap of abuse down on his head," in the words of a reporter for Science .^[45] Lawrence Altman treated the story as important, and the Times gave it front-page treatment.^[46] But skeptical AIDS researchers, including Gallo, lashed out at Lo, claiming he had failed to justify his claims. Lo pressed on, discovering the agent in various body tissues of AIDS patients; when he injected it into four silverleaf monkeys, they all died within nine months. Furthermore, Lo was able to isolate the agent from six HIV-negative patients who had symptoms similar to those of AIDS. Yet his articles were rejected by more than half a dozen journals before finally seeing the light of day in the American Journal of Tropical Medicine and Hygiene in 1989.^[47]

Eventually, it became apparent that the agent Lo had isolated was not a virus but a primitive organism called a mycoplasma.^[48] Since mycoplasmas are common contaminants of tissue cultures, it remained debatable whether the mycoplasma had really come from the AIDS patients. But only in 1990, when Montagnier independently began talking about a possible role for mycoplasmas as a cofactor in AIDS, did researchers belatedly begin to take Lo seriously. Until that time, Lo—despite his credentials as a virologist—could not make his voice heard over the chorus supporting the reigning theory.