Some forty years ago a revered and internationally known scholar gave it as his opinion to the writer that within a few years all interest in prehistory would be dead. “It has shot its bolt, it is already on the decline,” he said. Nothing could have been further from the truth. At a rough guess, between fifty and one hundred books on prehistory are now being published for every one that appeared forty years ago. But if this is an indication of public interest the question of scientific quality is another matter. As anyone who has watched the style of recent publications must be aware, there has been a marked shift in character and content, especially noticeable over the past decade. The book under review is no exception.

It may be sobering for prehistorians to reflect on the cognate (but somewhat more mature) field of geology where the spoofery and science fiction of, say, a Velikovsky can sell a hundred times more than seminal and penetrating products of genuine research. In the field of human prehistory “aquatic man” and more recently “submarine woman” may well win a large popular audience, while a truly shattering discovery like that which has just shown that man is two and a half times as old as we previously thought, with all that that implies for the concept of human evolution, may be barely noticed except by a dedicated few. But perhaps in the end such nonsense as the recent theories about our watery ancestors does less harm in debasing the coinage than does the derivative journalism of writers like Robert Ardrey or Jaquetta Hawkes who ladle out half-understood results in which they have had no hand, with a pontifical air of “I know all the secrets”—an attitude which in fact is the very antithesis of genuine scientific discussion.

It is accordingly with a wary, if not weary, attitude of mind that many professional anthropologists will open this work by a writer who frankly admits that he is a “scientific journalist” innocent of formal training in prehistory, or for that matter of firsthand experience of scientific inquiry of any kind. As if this were not enough, The Roots of Civilization is written in a style that seems based on the assumption that obscurity equals profundity. To top it all off, the author appears to have a distinctly defective grasp of the English language. As many as two or three times on a single page we get the absurd phrase “time factored time factoring” where “dating,” “sequence,” or “concept of time” is all that is needed, or the still more irritating “rendition” for “rendering” or simply “drawing.”

How comes it then that The Roots of Civilization has already attracted a remarkable diversity of comment and not been simply ignored by competent critics, as is the case of many hundreds of other books of an apparently comparable kind? Neither the subject matter nor the presentation is at first sight wildly original. For Marshack the main path toward an understanding of human cultural origins lies in the study of art in the Paleolithic or Old Stone Age, which began over two million years ago and ended in various places between 4,000 and 10,000 years ago. The study of this art is certainly a well-worn subject giving rise to at least a score of books a year, most of which are forgotten before the next batch appears, in spite of elaborate formats and illustrations quite as abundant as those in Marshack’s book.

Yet the reaction to The Roots of Civilization is something out of the ordinary. Why? Even if we eliminate the expected reactions of those reviewers and archaeological hangers-on for whom any novelty assumes the proportion of a “break-through,” the reaction of some scholars of repute, sometimes positively enthusiastic, is surprising, and all the more so in view of the book’s manifest weaknesses.

There are I think a number of contributory factors. One is perhaps to be sought in certain changes of attitude that seem to be affecting studies of early man. One such, arguably of some considerable value, is the realization, by scientists and journalists alike, that the accruing results of animal ethology need somehow to be more directly related to our studies of early man. Another of a more general kind is a growing enthusiasm for more inductive and theoretical approaches to the past, as opposed to an earlier emphasis on data collection followed by relatively superficial and particularist explanation.

How far these and other new directions of research, sometimes collectively labeled the “New Archaeology,” are productive of real insight, and how far they merely create the illusion of such insight, is, of course, another matter. It certainly seems that Marshack by his general approach is to be ranked with the “new archaeologists,” for what that is worth. What is probably much more significant however, and certainly very much rarer, is that he has sought to combine his “new” approach with a genuine and energetic attempt to open up a new factual field.


As most people with some interest in prehistory will be aware, the artistic expressions of the Old Stone Age survive in two distinct forms: the famous cave-wall decorations, generally referred to as Parietal Art, and small decorated objects—fragments of bones, tools, etc.—collectively referred to as Mobiliary Art. In style and content the two are closely comparable and display a basically similar range of ideas and concepts. Yet it is remarkable that despite the wealth of attention lavished by scholars on the wall decorations, relatively little really systematic work has been devoted to the mobiliary class.

It is on this neglected field that Marshack has sought to make his main factual contribution, and in my view it is not negligible. It is also upon this that he rests virtually all of his elaborate structure of theory and surmise. As a result he has come up with some genuinely original suggestions, although some are more and some very much less securely based on fact.

Lumping all his conclusions together, from the probable through the merely possible to the almost wholly speculative, Marshack attempts what amounts to a general discussion of the part played by timing and consciousness of time in the emergence of human conceptual thought and behavior generally. The undertaking is to say the least an ambitious one. But even if it were wholly convincing could it really be described as an inquiry into the “roots of civilization”? One aspect, surely, not more, one might have supposed. (But let us be fair, perhaps this is really all the author meant anyway.)

His starting point is an observation which, although not strictly unknown before, is certainly new in the emphasis that he gives it. Many fragments of bone, ivory, and so forth carry, in addition to or instead of representations, series of marks—dots, strokes, and the like—which cannot reasonably be interpreted as decoration. They tend to be arranged in groups or alignments that often show no obvious relation to the general shape of the object or to any other representation or design it may bear. A reasonable deduction is that these marks are some kind of tally or counting system. But of what?

A further point, which is less easily checked from the micro-photographs, but which there seem to be at least some grounds for, is that the dots were made a few at a time. Groups or alignments of, say, seven or eight, two or three, and ten or fourteen seem to recur. Often these are separate groups with minute differences detectable under the microscope, suggesting the use of a separate point or at least stroke so that the whole series is unlikely to have been made at one time. Finally there appear to be some indications that the groups themselves were made one after another so as to form a sequence. When subsets are added together it seems that they not infrequently add up to twenty-nine or slightly more, or multiples of this figure, sometimes plus or minus seven or fourteen.

Considering this apparent patterning Marshack hit on the idea that the tally might be one of days arranged in lunar periods or phases of the moon. The idea is not in fact as startling as it might at first appear, since it seems that among known groups of preliterate hunters of the recent past some in fact counted time in lunar months and even observed minor phases of the moon. Whether they also recorded time in any permanent way and afterward made some use of their records is, unfortunately for the theory, not yet known. The most that Marshack can offer is a recent ethnographic specimen which seems to show the same sort of grouping of marks. Its real significance is not in fact independently documented. But suppose for the sake of argument that we grant the author his first point and agree that he has a strong or even unbreakable case for lunar time-keeping in the Upper Paleolithic of Europe and lasting until much later in other territories, what follows? Where precisely is the deep significance for the “roots of civilization”?

Marshack goes on to couple his supposed evidence of time estimation with indications of seasonality in the art subjects themselves. He cites symbols of salmon (representing perhaps the spring or fall run) and, associated with them, grass snakes believed to belong to a species that emerges from hibernation in the spring, growing plants, molting musk ox, etc. There are also other symbols possibly indicating late summer and fall (rutting behavior of stages and bison). The time estimations might be taken as aids to more efficient seasonal planning of the activities of the hunters and could be seen as an indication of a relatively high degree of purposeful conceptual thought in time beyond that usually attributed to our early ancestors of 30,000 years ago.


Pressing back even further he seeks the origins of such conceptualization in the much earlier planned activities of a creature like Homo erectus three-quarters of a million or more years before. The earliest actual example of a tally is much later, a scratched bone said to have come from the Middle Paleolithic Mousterian zone at La Ferrassie at the time of Neanderthal man about 40,000 years ago. However it is noticeable that such specimens are elsewhere absent from the huge masses of bone work reliably dated to that period, and the same may be said of all trace of representational or decorative art. This is generally unknown before the Upper Paleolithic in Europe, during which varieties of Homo sapiens sapiens such as Cro-Magnon man appeared.

Building further yet on his assumptions Marshack thinks that he can infer the existence of sequential patterns of thought or “stories,” ritual, sacrifice, and other forms of symbolic behavior, which are of course typical of early civilization at later stages. By this point in the argument it must be admitted that positive evidence is getting pretty thin on the ground; yet one observation he puts forward, if correct, is certainly of some significance on its own account. Many well-known cases of redrawing or over-drawing animal representations show signs of being done separately at intervals of time, in a few cases of substantial duration. In one case he claims that a scene showing some kind of bird is overlaid by what is apparently a nest with eggs, and finally covered in turn by a sketch of a snake. Since at least two if not all of these separate symbols were made with different tools he draws the conclusion that “the reconstruction implied a sequential composition in time, apparently representing a series of seasonal events: the arrival of ducks, after this the building of nests and the laying of eggs, and finally the theft of eggs by the snake.” All this, he feels, adds up to a “story” if not a “ritual.”

It must be admitted that this example is characteristic of Marshack’s thought. Even if we agree (which we are by no means obliged to do) that his interpretation has a certain beguiling plausibility, it is at most no more than one out of a near infinity of equally possible suggestions. If we examine the figure closely it must be allowed that the crudely rendered “duck” is by no means certainly a duck or even a bird, nor is the nest a nest, while even the reading of “snake” is no more than probable. As for the implied sequence of thought, even this admits of quite different interpretations, a pure doodle for instance, or a sequence of drawing no more than accidental.

Other examples of overstrained arguments are to be found scattered throughout his text, some less and some even more arbitrary and fanciful. So the question arises in the end, must we write off nearly all of Marshack’s argument and regard the whole affair as mere fanciful speculation? There is no doubt that many will adopt this view, indeed already have done so, to judge from recent published comments. My contention is otherwise. One of his suggestions, to take an example that I find among the more convincing, is the interpretation he offers, backed up by beautifully clear close-up photographs, of the famous Montgaudier baton. The combination of salmon, seal, grass snakes, and opening bud, almost certainly an intentional grouping, has indeed all the properties of a seasonal symbol of spring. At least, whatever the real significance may be (and there are of course other possibilities), Marshack’s reading makes excellent sense as compared, say, to the highly fanciful “sexual” interpretation offered by André Leroi-Gourhan. Other works susceptible to the same degree of seasonal connotation are not lacking: the molting musk ox already referred to, or the Teyjat baton, with mare and young foal, swans, and the grass snakes again intentionally surcharged with a doe’s head and a series of enigmatic humanoid figures with chamois masks.

Although we cannot, and possibly never will, understand the full meaning of such elements as the last named in isolated cases, it may well be that comparative study of repeated occurences based on the impressively precise methods of Marshack will in the end produce further insight. I would certainly not like to write off the possibility. In my opinion his methods have already shown their worth by demonstrating (as opposed to asserting) the repeated nature of specific associations in the Magdalenian phase of the European Upper Paleolithic.

At this point I should like to turn aside to take up a point which I feel Marshack might have pursued to advantage. The modern methods of dating now available to us do not merely segregate his finds into a rough sequence as older methods did; they provide actual time estimates of a precision undreamed of a generation ago. Few students of Paleolithic art have begun to take full advantage of the opportunity now offered. Many indeed seem willfully to ignore it. To be more specific; the scenes just referred to and many others like them belong almost exclusively to the period 14,000-12,000 years ago. Others again of a noticeably different style belong to the Solutrean phase, approximately 18,000 to 15,000 BC, or to the earlier stages of the Magdalenian, while yet others can be securely placed in the Gravettian phase for instance at 26,000-20,000 BC, (such as the great frieze at Pair-non-Pair, the sculptures at Laussel, or the paintings at Sergeac). While Marshack does indeed make some mention of these phases and makes a start on features of style and content which appear to be zoned in time, there remains a considerable body of issues which deserve to be more closely explored in this general field.

It is true that out of the great mass of suggestions that he puts forward in the course of his somewhat rambling and confused account, many cannot possibly be ranked as scientific deductions in the ordinary sense. For this reason it is to be expected that they will be rejected out of hand by many trained in more exact disciplines such as for instance astronomy. It should however be remembered that the habit of mind in such fields has little in common with the inductive approach which is after all common to all historical research in the broad sense.

In seeking to offer a convincing explanatory model for some particular historical event we are in a sense making a hypothesis of testable predictive power in so far as it may be infirmed or confirmed by fresh evidence as it comes to light. But much of the process of thought involved is inevitably inductive. We not merely ask ourselves what we feel as individuals to be reasonable, but also draw on our fund of purely empirical knowledge of human behavior as we see it in action around us. While most scholars would surely admit that very little of all this shares the validity of the experimental sciences, it is also true that when the attempt is made to apply reasoning of the latter kind to purely human problems, the results are not infrequently unconvincing—witness the speculations based on the Pyramid inch or, to take a contemporary and much discussed example, the all too precise and elaborate deductions of Professor Thom on megalithic monuments.

Some element of “What does it feel like to be part of another culture?” is essential to the equipment of prehistorian and historian alike. Without it they are almost certain to fall a prey to the Thom syndrome; equally if they carry it to excess the result is a sort of inductio ad absurdum. The threshold between the two is all too narrow, and Marshack as I feel comes dangerously near to it. His assumption that men of the Upper Paleolithic age from 32,000 to 10,000 years ago thought and reasoned much as men in comparable societies do today is wholly reasonable. But this last proviso is the operative one; it is all too easy to project the special features of twentieth-century industrial thought into the human situation of technologically primitive hunters, with wholly misleading results.

As I write I think of a case in point from contemporary New Guinea: a slat of wood with a roughly worked handle bearing all Marshack’s principal categories of signs—groups of grooves in fives and sixes, “V” marks in series, worn and scratched over with groups of hatching. Its purpose? To act as a gauge, the ethnographers tell us, for making the correct mesh on a cassowary snaring-net. Its makers were wholly illiterate and innumerate, yet they could keep count, measure, and maintain a record; and all for a very specific purpose which a future prehistorian would be hard put to identify! A different situation is offered by the well-known birch-bark mnemonic scenes from North America, which Marshack perhaps has in mind when he suggests that some of the associated symbols he has identified may be evidence for “stories.” His suggestion is not unreasonable, only very difficult to prove. There is in fact little structurally in common between the paleolithic scenes and the birch-bark designs. On the other hand it may well be that careful scrutiny of the kind offered by Marshack, if coupled with repeated analysis, might indeed take us further in our understanding.

I feel that Marshack has made a real contribution by calling attention to evidence for patterned symbolism connected specifically with seasonality and linked to a practice of counting which seems to have spread virtually throughout the Upper Paleolithic of Europe. I do not feel that his numerical evidence is yet strong enough to demonstrate positively time-reckoning in lunar units, or indeed to show that time-reckoning of any kind was the main purpose. His suggestion is interesting and not altogether implausible, but it will require considerably greater and more systematic mathematical analysis before it can be regarded as more. The alternative suggestion that the time tallies were not related to a lunar calendar but were simply a record of elapsed time between seasons seems worth further investigation. It is easy to see how such a practice could be of value to hunters operating in an environment marked by highly differentiated seasons, such as Europe in the Ice Age.

It is less easy to see how to test such a hypothesis. But if it can be successfully tested I feel sure that Marshack has taken the essential first step by showing what can be achieved by precise methods applied with remarkable patience and stamina. Nothing quite like this has been attempted since the original observations of the Abbé Breuil on Paleolithic art.* It is to be hoped that Marshack will pursue his researches undeterred by irrelevant criticism (of which he has certainly had his share) on the one hand, and avoiding the pitfalls offered by over-enthusiastic admirers on the other. So in the end his observations may yet lead to the first real break-through in many years in our understanding of Paleolithic art and to further understanding of its meaning for history.

This Issue

September 20, 1973