On Wednesday, August 30, 2017 at 7:55:23 PM UTC-7, wjhonson wrote: > Y-DNA cannot properly support a paper trail because there is almost no > way to confirm where someone goes in that tree. Autosomal DNA *can* > properly support a paper trail. You keep saying that, and it keeps being wrong. > And Todd I do know that *on average* each generation inherits one half, > however if every generation has multiple descendants, than you have a > good chance that *the vast majority* of the DNA is *preserved* across > the entire population. And all it takes is one generation of having a single child, and you lose half the information. Not every family has a dozen children, each with a dozen children, . . . . The effect of this is cumulative information loss, no matter which way you trace it. > You keep dividing, without multiplying. You have to do both. Loss is loss, it is progressive, and it doesn't matter how many descendants there are that don't have the lost material. taf

On Wednesday, August 30, 2017 at 8:45:08 PM UTC-7, taf wrote: > > You keep dividing, without multiplying. You have to do both. > > Loss is loss, it is progressive, and it doesn't matter how many > descendants there are that don't have the lost material. By the way, it isn't good enough for the DNA to be preserved in some line, it has to be identifiable to be any good to the analysis. There is an amount actually, literally, lost but a further amount that is lost to the analysis because its presence can not be linked to the common descent. You act like if the DNA is anywhere, among any of the descendants, then it is useful in a reconstruction/confirmation, but it has to be shared by multiple descendants and it has to be known that it is shared because of the specific common descent being evaluated. If either of these conditions fail, then the sequence in question is not literally lost but is lost to the analysis nonetheless. In the other post, I already showed how dramatic this loss can be if you ever have a generation with a small number of siblings. I can remember someone telling me that with a powerful enough superdupercomputer it should be possible to recover the voice of George Washington, because we can track individual molecules in the air to determine their trajectories and velocities, then extrapolate them back into the sound waves that dissipated to leave them in their current state, and thereby reconstruct progressive sets of sound waves, back in time, until you get to George Washington. Like with what you propose, the theory is sound (if simplistic), but in the real world any attempt would fail abysmally. taf

On 31-Aug-17 2:08 PM, taf wrote: > On Wednesday, August 30, 2017 at 8:45:08 PM UTC-7, taf wrote: > >>> You keep dividing, without multiplying. You have to do both. >> Loss is loss, it is progressive, and it doesn't matter how many >> descendants there are that don't have the lost material. > By the way, it isn't good enough for the DNA to be preserved in some line, it has to be identifiable to be any good to the analysis. There is an amount actually, literally, lost but a further amount that is lost to the analysis because its presence can not be linked to the common descent. You act like if the DNA is anywhere, among any of the descendants, then it is useful in a reconstruction/confirmation, but it has to be shared by multiple descendants and it has to be known that it is shared because of the specific common descent being evaluated. If either of these conditions fail, then the sequence in question is not literally lost but is lost to the analysis nonetheless. In the other post, I already showed how dramatic this loss can be if you ever have a generation with a small number of siblings. > > I can remember someone telling me that with a powerful enough superdupercomputer it should be possible to recover the voice of George Washington, because we can track individual molecules in the air to determine their trajectories and velocities, then extrapolate them back into the sound waves that dissipated to leave them in their current state, and thereby reconstruct progressive sets of sound waves, back in time, until you get to George Washington. Like with what you propose, the theory is sound (if simplistic), but in the real world any attempt would fail abysmally. I think this is an apt comparison not just as to practicability, but also as to futility - curiosity value may be vastly higher than potential usefulness. Some years ago a sound recording was found that is thought to have captured Queen Victoria speaking in the background. The only slight value of this, apart from the novelty of hearing what may be a voice from history, is that the woman seems not have had much of a German accent (that is a canard often repeated about her family). It would be about equally useful to find out that William the Conqueror did not carry the genes for a receding chin - and before anyone chimes in that we don't yet know how to make such a linkage anyway, there is at least one major project, reported on TV news in Australia a few days ago, to identify genes for such facial features. Peter Stewart