I was using a common label to avoid a long circumlocution. With 24 identified families, existence of HERVs (pl.) is not exactly a wild claim.
Here is a paper on a reconstructed, functional HERV-K If, as repeatedly stated, XMRV is nearly homologous with some ancient retrovirus now integrated into the human genome, it is entirely possible it could reactivate one. In this case, the missing "viral payload" genes might already be in humans. This would explain how a virus which doesn't appear to do anything except reproduce the elements needed to reproduce (gag, pol, env) could have wildly divergent consequences in different people.
I agree the PBMC should have been activated. However, if the level of virus were high enough, there would still be a chance of detection. This negative result could help to set bounds on the number of infected cells.
I am not arguing in favor of the Oxford definition for CFS, it is so vague as to be useless. I think they might detect XMRV in a cohort which meets DSM-IV TR criteria for clinical depression. This could be a game changer for much more than CFS/ME.
I'm still wondering about your statement of 0.2 ml samples. This doesn't seem like a typo.