[setr]

> But there is no mechanism that guarantees the inverse — that improvements in infectiousness will be accompanied by a decrease in lethality.

I’m not clear; the argument effectively defines infectiousness as how-not-lethal the virus is. Thus, the virus cannot be made more infectious without decreasing lethality.

There is however an interim period, as evolutionary arguments are long-term arguments, where the virus can be both more infectious and more deadly, before settling into its stationary state.

The molecular perspective isn’t too useful against an evolutionary argument as well — evolution is always a sum-of-the-parts system. The question is simply: is the virus ultimately more capable of spreading? And at least in the extreme, an instantly killed host is unlikely to do an real spreading, no matter how trivially the virus could spread (had the host persisted). And in the other extreme, a completely non-lethal virus can survive and spread forever, especially in dense populations, especially if it doesn’t incapacitate the host significantly (acting as a simple parasite), allowing it to continue meeting new hosts.

Of course, a virus that doesn’t need the victim to be the host is a different story; eg malaria persists and spreads through mosquitos, so it could instantly kill humans and yet receive no evolutionary pressure against it.

[nevinera]

One of the things that has made this particular virus so successful is that it does the larger part of its spreading before there are any major symptoms. In this case, the only thing about lethality that will reduce infection rate is the degree to which it makes everyone more careful (or makes more people at all careful).