Knowledge, and the past
Our knowledge of the past, particularly in the paleontological sciences, geology and phylogenetics, is problematic in large part because of the way in which the modern approach to epistemology has been framed. If we require that knowledge is immediate, direct, testable experimentally, and so forth, then knowing what happened when those conditions cannot be assured becomes difficult, if not impossible. And all these presentist criteria for knowledge have been used. Moore, responding to extreme skepticism, asserts that he does know that he has a hand. If "to know" means anything, he knows that.
The issues in paleontological science is less severe than the problems Moore considered and to which Wittgenstein responded (for him, knowledge is the outcome of a language game in which sentences like "I know" are foundational parts of that game). All they want to know is what the ancestors of a species at a later timeframe were, or what the processes that generated some observable case, such as a basalt plain. But the information appears to be lost to some; to others it is just harder to get.
Consider that there is, for any domain of possible knowledge, a state space of all possible knowledge claims, now, in the past, or in the future:
We have, for any domain of investigation, a state space of all possible cases that are consistent with that domain. For example, if the domain is physics, then the state space is all physically realizable states. Of any physical system, say, an orbital system, we want to ask of this entire space what was actually realized. Where was the satellite in the past? Where will it be in the future? What is the trajectory is was, is and will be following?
Take Pluto's orbit for example. It might have once been almost anywhere in the outer regions of the solar system. It might have been a part of the Oort Cloud, and got sent flying into the inner solar system. Its past for a short period is known, but because of the lack of knowledge about what has affected or will affect it, go back or forwards far enough and we cannot predict or retrodict where it was and what it did or where it will be (in a surprisingly short time). So in the past we have an increasing cone of uncertainty. Likewise in the future. The observer can locate and represent the physical situation at a moment in time with increasing degrees of accuracy, but the knowledge of the future and the past is increasingly uncertain.
This uncertainty plays havoc with our knowledge. As the title of Wittgenstein's book indicates, certainty and knowledge have been closely connected in western philosophy, at least since Descartes, but even earlier. Philosophically respectable knowledge is that which is certain, or nearly so. And that slight vagueness at the end is what gives us our problem. We are almost never certain, not even with Moore and his hand. Perhaps he doesn't have a hand, but only dreams it, and so forth...
Of course, science, and in particular the historical sciences like paleontology or phylogeny, have to make do with a much less strict, but richer, kind of knowledge. In this case, relative certainty is sufficient. But as we see, information about the past states is lost to us increasingly over time. Sober has a nice illustration, as I said before - a ball at the bottom of a smooth spherical bowl must have rolled from one part of the lip, but we do not know which part just from the observation that it has come to rest at the centre of the bottom. That information has been lost.
When Wittgenstein and Waismann objected to Darwin, along with Moore and Russell and the other opponents of evolution in philosophy, they did so in part because of the requirement for pristine and pure certainty in our knowledge. But if science is understood to be simultaneously the best form of knowledge that we have, and also falliblistic (that is, knowledge that can be mistaken) we have no such objection to evolutionary, and historical in general, accounts of what it is to know things. And so we must retrun to the philosophical questions that were for so long rejected, from Goethe to the Pattern Cladists. Is it knowledge that we have of the past. Can we know ancestors, for example? We shall see, I hope.