Summary
From the article:
HIV is defined as “various unrelated viruses that cause AIDS.” And AIDS is defined as “various unrelated diseases that are caused by HIV.” Substituting equals for equals, “HIV is the cause of AIDS” means:
Various unrelated viruses are the causes of various unrelated diseases.
[...]
The situation is highly unsatisfactory logically, but it works perfectly in practice—for specific reasons I’ll explain in this post.
[...]
The HIV viruses arose through convergent evolution from various viruses that infect various monkeys and apes. Those are called “simian immunodeficiency viruses,” SIV. The term “HIV” lumps twelve monophyletic groups. Each contains all the descendants of a single SIV that mutated so it was able jump from some other species to the first human infected. Two HIV groups descend from two different chimpanzee SIVs; two from different gorilla SIVs; and eight from SIVs that infect various monkey species.
[...]
To count as HIV, a virus has to cause AIDS. That is because we don’t care (much) about monkey diseases, and we do care (a lot) about fatal human pandemics. “HIV” is an inherently purpose-laden category. That is why some biologists defend polyphyletic categories: because the members function similarly in ways we care about.
“Being HIV” is not a property of a virus by itself. It is a functional, contextual property, based on the virus’s behavior in humans.
[...]
Unlike HIV, AIDS has a specific, official definition, issued by the U.S. Center for Disease Control. You don’t need to understand its details to follow the rest of this discussion! However, the CDC’s definition is that a patient has AIDS if:
They are infected with HIV, and
They have either: (a) any from a list of 27 other infectious diseases, or (b) a severe deficiency of T helper immune cells, defined in turn as either: i. a CD4+ T-cell count below 200 cells/μL, or ii. a CD4+ T-cell percentage of total lymphocytes of less than 14%.
[...]
This makes “HIV is the cause of AIDS” a clear example of “the social construction of scientific knowledge.” That’s anathema for some rationalists. Rightly so, in many cases! “Social construction” is often code for “powerful groups can force you to treat false claims as true.” However, in this case, and many others, it just means that many people collaborated to construct pragmatically useful categories. Those constructions don’t affect the virus or the disease. They are nevertheless consequential: they do affect how we treat them, clinically and in public health projects.
[...]
Including HIV in the definition of AIDS makes the whole thing circular, because HIV is implicitly defined as viruses that cause AIDS. For rationalism, circular definitions are not OK. Definitions are supposed to tell you what something is. If you didn’t already know what AIDS and HIV are, defining them in terms of each other wouldn’t help.
[...]
The circularity is unproblematic because marginal cases can be dealt with using meta-rational judgement in particular situations. If a patient is HIV-infected, and somewhat immunodeficient, and develops an infection that is not on the CDC’s list, a doctor could say “well, maybe this isn’t officially AIDS, but who cares—I’m counting it as AIDS for treatment purposes!”
[...]
In the next few sections, we’ll apply each of several mainstream general theories of causality to “HIV is the cause of AIDS,” plus one theory specialized to infectious diseases:
Necessary and sufficient conditions
Contributing causes, enabling and background conditions
Non-causal field models
Probabilistic theories
Mechanistic theories
Interventional theories
The germ theory
None of these works as a general theory of causality. None of them even works, on its own, as an explanation of what “causes” means in the case of HIV and AIDS!
There is a better alternative: to admit that “causality” is not a single, clearly-definable phenomenon, but a nebulous collection of vaguely-similar notions. Each “theory” tries to capture a class of informal uses of “cause.” Each ignores others, and also fails even to nail that class.
[...]
Now that we have a better understanding of what “cause” means in this case, we can make a better reformulation:
A variety of unrelated diseases can usually be prevented or treated with this set of interventions, which are usually effective against any of a particular set of viruses. Therefore, it’s often not necessary to identify which virus a patient is infected with before beginning treatment, nor to predict which of the diseases they might otherwise suffer.