The past is continuously slipping out of living memory. At this moment in time (2007), there are only half a dozen or so living persons who experienced combat during World War One. In another few years, World War One will no longer be in anyone’s memory. Our knowledge of the war will not disappear, but it will undergo a change of status. As long as a memory exists in even a single person, humanity as a whole can be said to have direct perceptual contact with the remembered event. When the last living witness to the event goes, our knowledge becomes indirect.
Indirect knowledge of the past is based on inferences drawn from an enormous number and variety of artefacts—old newspapers, memoirs and diaries, ledgers and receipts, correspondence, clothing, tools, kitchen implements, skeletal remains, etc. These inferences are often sufficient to fill the gaps in our historical knowledge created by the loss of living memory. Even when there’s no one left to say "I was there, I saw it" of World War One, we’ll still have more than enough indirect evidence to continue to believe that the war really happened. But there’s no guarantee that inference will always be able to do the job of substituting for living memory. Suppose that someone has a brilliant philosophical insight and then immediately drops dead. For the brief moment between insight and death, the fact that this insight was obtained was part of the collective knowledge of humanity. Upon the death of the insightee, the fact drops out of living memory. In this case, however, it's difficult to imagine how the slippage from living memory can be recovered by inferential processes. The death of the insightee thus produces an irremediable decrease in the sum of human knowledge. This state of affairs is by no means rare. Most of us have private knowledge about our own life that will be erased from the big ledger book of all human knowledge when we die.
There are three broad categories of this ephemeral variety of knowledge that I find particularly interesting. What makes them interesting is that they weren’t merely the possession of isolated individuals or small groups. These are items of information that were known by every normal member of societies that in some cases ran into the millions. Yet this universal knowledge has been lost completely.
First, we don’t really know what the speech of the past sounded like. The written literature of the time informs us that the English spoken in, say, 1350 followed the same grammatical rules as contemporary English. It also tells us how written words were spelled. But it doesn’t tell us how the words were pronounced. Students of Middle English are taught schemes for its pronunciation, but these schemes are in large measure free inventions, unburdened by the need to satisfy any evidential constraints. This is true to an even larger extent of ancient Latin. When it comes to Egyptian hieroglyphs, one might as well pick their pronunciation out of a hat. Yet their correct pronunciation was once, in the time of ancient Egypt, known to millions.
We know even less about the intonation with which the speech of the past was spoken. Our habit of reading Chaucerian or even Shakespearean interrogative sentences with the contemporary interrogative rising lilt is based on nothing more than a wild guess. For all we know, speakers of Chaucerean English might have used the same intonation for their interrogatives as for their declaratives. Alternatively, their question-asking might sound like a threat to us, or a complaint, or a seductive advance. It might also fail to resemble any intonation in our expressive repertoire, so that, were we to hear it, we would have no idea whether we were being asked a question, threatened, seduced, etc. It's also possible that speakers of Middle English customarily spoke much more loudly than we do, or much more softly—or that they spoke in falsetto.
Secondly, we don’t know how people of the past moved. The way in which we humans walk, sit, and gesticulate is only partially determined by our biology. Nature impels us all to locomote by putting one foot in front of the other, but it leaves a lot of the details open to individual variation. Different cultures and subcultures foster or discourage different variations. To see this quickly, compare the movements of almost any contemporary Afro-American urban teenager with those of a middle-aged Swiss banker. Better yet, imagine the banker moving like the teenager and vice versa. Combine this observation with the fact that different historical eras represent different cultures, and you get the result that people of the past may very well have moved differently. They may have moved more rapidly or more slowly, more rigidly or more flowingly. They may have engaged in more gesticulation or less. This is once again knowledge that was once universally distributed but which was unreplaceable when it left living memory.
Thirdly, we don’t know what the food of the past tasted like. We're in possession of old recipes that give us some idea of the taste of past food. But this source of indirect knowledge gets thinner and thinner as we go back in time. There’s also some evidence that the taste of basic ingredients has been changing with time: the older generation is always complaining that the strawberries and tomatoes of their youth were much tastier than their present-day descendants. This degradation may have been going on for centuries at a pace so slow that it was imperceptible in any one individual’s lifetime. There might very well be a perceptible difference in taste after several lifetimes, but there would be no one in a position to make that comparison. The taste of the earlier produce would be known to the earlier generations, but this knowledge would be lost to the later generations. For all we know, the produce of the 18th century might taste so much better than the contemporary stuff that one bite would make us swoon.
Technological advances in the past hundred or so years have profoundly altered our relationship to the past. The invention of sound recording and cinematographic devices has provided us with evidence of how people walked and talked at the time of recording that’s almost as good as direct perception. But the evidence of direct perception carries evidential weight for only as long as the perceiver lives, whereas the evidential weight of a film or a CD is for all practical purposes forever. A thousand years from now, humanity will still know how people moved and talked in the 1920’s. Unfortunately, this technological access to the past isn't retroactive—we still don’t know how people moved and talked in the 1820’s. Perhaps a future technology based on presently unimagineable principles will be able to build a device that permits the operator to view and record scenes from the past.
What about the taste of past food? There is as yet no known way of recording tastes for future playback. But there’s no need for such a device—the food of the past can literally be tasted in the present by the application of a familiar and unglamorous technology: freezing. With the advent of frozen food, it becomes possible to know what the food of the past tasted like with the certainy of direct perception. A thousand years from now, the discovery of an intact TV dinner will be major archeological news. The find will be thawed, heated, and consumed under controlled laboratory conditions by trained paleogastronomists.