Considering that the human genome project found several *thousand* endogenous retroviruses...and that endogenous retroviruses are only the retroviruses that happened to infect a germ cell which then became an embryo that survived to reproductive age, a very rare set of circumstances...it seems pretty clear that infection of humans and human ancestors by retroviruses has been a very common event throughout our evolution.
This is a very interesting argument, which I avoided. Part of the problem is counting. We have at least three difficulties: 1) RNA viruses are notoriously unstable, mutating often for no apparent reason; 2) defenses against retrovirus include hypermutation of inserted provirus; 3) after a sequence has been inserted in the germ line, it is subject to an unknown number of later transcriptional errors; since it is not necessary for host survival, these are generally not corrected. Data on ERVs in the genome are noisy. My own feeling is that we are looking at a much smaller number of extremely rare events, complicated by the factors above. A second possible criticism would be that these events were spaced out over millions of years, making the chance of finding one in a century very small. I've already seen too many arguments over extremely rare events go astray. People just aren't built to think about things beyond the limited scope of daily life.
If you accept the reconstructed sequences of ERVs as actual viral genomes, you do have an excellent argument that people were infected by a virus 95% homologous to XRMV at some time in the distant past. This is slightly closer than the 94% homology with MLV, but the difference is not very significant.
What do you suppose the odds are that we currently have more than two or three retroviruses currently in existence that are able to infect humans, considering the *millions* we must have been infected with over the course of our evolution in order to allow for several thousand to make it into our genome? The modern world has done almost nothing to prevent viruses from jumping species from other animals into humans - if anything, it's only made it easier for them to trot the globe and quickly hit a lot of people in urban areas, once they've gotten into humans.
The real kicker on this argument is the sheer number of human beings, and possible interactions with other species. As you point out, these are not isolated human beings, we have demonstrated we are all one family in any number of ways, including the transmission of diseases. Once a virus jumps species it has no natural barrier to spreading in this recently unified population.
During the course of a discussion over possible sources, one person said "well, people don't eat mice". Unfortunately, I know a counterexample, a place where roasted mice are a popular snack. When I suggested domestic cats as an intermediate vector, the "people don't eat cats" response came up. (Unfortunately, while this is largely true, there are, once again, exceptions.) People do have many interactions with domestic cats which could transmit virus found, for example, in saliva. Even though most of us on-line no longer know our dinner personally, there are large numbers of domestic animals raised exclusively for human consumption. Testing livestock for signs of slow infections by viruses unknown to researchers is simply not being done.
Sheer quantity can have a quality all its own. The scale of human interactions with many other species is simply hard to imagine. This is one reason I personally don't place much weight on the idea of things escaping from laboratory experiments and becoming widespread.
(Yes things have escaped. In some cases, the general population appears to have 'dodged a bullet' by chance, but consider those other sources. There are simply too many 'natural experiments' constantly taking place with no control and no safeguards whatsoever. Most of us have to ignore such ideas simply to avoid turning into paranoid recluses. Right now, in a few minutes, I could take you to several different little old ladies keeping cats alive which are clearly infected with a variety of mysterious things. They have no idea they are conducting experiments in virology.)
I'm hoping some of us can agree to disagree on some questions above without missing the essential point of urbantravels' post -- there are strong reasons to be looking for currently undetected retroviral infections in humans. I see no reason to rule out gamma retroviruses, and strong reasons to suspect them.