Return to home page

Circe’s loom

We could see her working on her great web all day long, but at night she would unpick the stitches again by torchlight. Odyssey, Book 2.

They could hear Circe within, singing most beautifully as she worked at her loom, making a web so fine, so soft, and of such dazzling colours as no one but a goddess could weave. Ibid, Book 10.

What is truth? said jesting Pilate, and would not stay for an answer. Francis Bacon1.


In his report on the 1999 Athens seminar on history teacher training in Southeast Europe2, the Rapporteur, Dr. Robert Stradling, includes a short exchange with another speaker, Professor Kremidas, on a quite fundamental question: what is historical truth? This Historikerstreit is no mere scholarly quibble. The controversy goes to the heart of the Council’s struggle against the manipulation of history and education. There are still many in power who think that politics has the right to impose the teaching of history as heroic national myth, and it is vital to defend the intellectual autonomy of the discipline. It is an important matter if our spokesmen give the enemies of history unnecessary comfort. While the history reform project has no official norms, Stradling’s view clearly represents not only a widespread current of thought among professional teachers of history, but an unofficial doctrine put forward on behalf of the Council of Europe3. If I take him to task, it is not to impugn the very great merits of a comrade-in-arms, but because the work itself must be held to the standards of rigour it proclaims.

Kremidas’ position is the straightforward one that (in Stradling’s words) “it is the job of the historian to find out what really happened” and therefore that it is “the job of the history teacher to tell his or her students what happened (and why)”. Fine, we applaud: let’s have the past wie es eigentlich gewesen in Ranke’s famous phrase. Who could possibly disagree with this?

One person does: the Council’s leading expert for advising post-communist countries on the reform of their history curricula. While Stradling recognises that truth is the touchstone of historical inquiry, he emphasises that history teaching should be cautious in its claims to have found it. Instead teaching should be multiperspective, “introducing students to different interpretations and explanations of particular historical events and developments”. He adduces two main arguments for this:

  1. historians deal with historical evidence not historical truth”;

  2. more often than not the evidence is incomplete and provisional”.

From these propositions he infers:

  1. there may not be one correct version of events;

  2. it may be that the same piece of evidence can be interpreted differently by different historians depending on their perspective, the time when they were writing and the connections which they seek to make between that piece of evidence and other evidence;

  3. it is certainly the case that many sources of historical evidence (official documents, statistics, archives, eye witness accounts) are far from objective and impartial;

  4. and, above all, it is obvious that the same apparently clear and unambiguous statement can still have very different meanings for people depending on their particular situation.”

(Numbering added).

Most of this is plainly good sense. History cannot be presented as a monolithic body of proven fact, like much of the natural sciences. It is surely a good thing to encourage pupils to test historical assertions against the evidence, and to introduce them to the methods as well as the results of historical inquiry. Stradling’s position is far from complete relativism. However his emphasis on the conjectural, the dubitative, the subjective aspects of history encourages a wide scepticism about claims to historical truth. In my view, it goes too far, for several reasons. I also think that it obscures the true case for multiple perspectives. In short, I wish to argue that history is not woven and unmade endlessly on the loom of all too human Penelope, but once only – although it shimmers with her dangerous magic - on that of the Olympian Circe.


First, the question of evidence. The statement (b.) that “more often than not the evidence is incomplete and provisional” is only true if we take a holistic approach to history, as an attempt to grasp the totality of events. But “understanding the eighteenth century” is a project as impossible as it is pointless. We do not have access to a mental videotape of our own experience, let alone that of others. In the words of Sir Thomas Browne, “oblivion shares with memory, a great part even of our living beings”4. Nor would we know what to do with such zigabytes of data if we had them. Most of the past is simply not accessible to us now; but we are not interested in reconstructing most of it. History is indeed “what one age finds of interest in another”; it is an attempt to answer questions, typically about the trends and cruces that brought us to where we stand today. Like the police, historians have to let a lot of suspects go free because they cannot make a case to stand up in court.

A general assertion that the evidence is inadequate is therefore baseless; it depends on the question. Important events are usually known to be so at the time, and literate people record what is going on5. The American and French Revolutions, or the build-up to the First World War, are abundantly documented. There is no real doubt over the course of events and the motivations of the actors. Problems arise with conspiratorial movements (like the Bolsheviks6 and especially the Nazis), with the clandestine activities of states – fertile ground for real and imaginary conspiracies -, and with distant events like the campaigns of Attila or Genghis Khan. The Armenian genocide of 1915 is documented fact, but we don’t know why it was committed.

Historical researchers, like any other scholars, are interested in difficult problems: eg when did Hitler decide to exterminate the Jews – 1920, 1938, 1941? Given Hitler’s paranoia and extreme deviousness, we will never know the answer to this question with any certainty. However, research interest is a poor guide to historical importance, which must be the criterion for history teaching. We can be absolutely certain that the Holocaust took place, and overwhelmingly sure that it was by Hitler’s will7.

Bias in evidence (proposition (e.)) is quite true, but is not grounds for general scepticism any more than an astigmatic person cannot trust her vision, or an astronomer must distrust ground-based telescopes because the atmosphere absorbs certain wavelengths. Bias creates systemic problems for the researcher, who learns tools for dealing with them. The record of a show trial for heresy or treason is pretty much worthless as evidence of the alleged crimes, but excellent evidence of the concerns of the authorities, of changes in legal procedures or language, or perhaps of sexual mores8. The most tainted sources – political propaganda and sensational journalism – can be safely used a contrario: if their enemies did not accuse John Major or V.I. Lenin of personal improprieties, that strongly indicates they did not commit them. The sources of bias in official statistics are well understood by statisticians9. Incidentally archaeological data also suffer from biased sampling – pottery and weapons survive better than food and clothing, religious art more than secular10. Perhaps only DNA never lies.

It is quite true, as Stradling points out, that historical evidence is incomplete: always so in the unimportant sense that we do not and cannot have the whole story; often so in that the evidence is too thin to rule decisively for or against our hypotheses. The right conclusion from this state of affairs is surely not to bewail that we are lost in a cloud of unknowing, but to accept restraint and suspend judgement. Historians are particularly subject to the human tendency to paint in the gaps, to guess where we cannot know. Mathematicians try to classify problems as soluble by efficient algorithms, soluble but computationally complex, and insoluble. Historians should take a leaf from their book; perhaps they should classify events as known and explained, known and inexplicable on the evidence, and unknowable.


Such epistemic self-discipline may be found among primary researchers in archives, but is so far as I can see atypical of historians who write with a broader brush, and of course history teachers in schools and their pupils. The subjective position of such students of history, confronted with a complex episode or period, is typically: how do I make sense of these events? It is natural to use the term interpretation for the process and the result of the inquiry. Cannot historians of different tendencies rally round this language? But such a consensus would be fool’s gold, bought only by inclusive ambiguity.

There are several distinct issues around this term. The first is brought out by Stradling’s most radical proposition (c.) on interpretation: “there may not be one correct version of events”. (This relativism is incidentally incompatible with his “Copenhagen theory” (a.) that historians deal with evidence not truth, since there is only one set of evidence.) The main problem is making any sense of the proposition itself. If it only means that we often do not or even cannot know the truth, that is self-evidently correct (see above). To cover all possibilities, let us take Stradling at face value here: in which case the proposition is absurd. The idea of multiple pasts, or alternative causalities, can be entertained in a mental sandbox or fable by Borges as a literary or theological speculation11: but taken seriously, it falls to Kant’s refutation of all systematic scepticism. Multiple pasts and causalities violate the presuppositions of all reasoned enquiry and the language in which it is pursued; their proponents saw off the very branch from which they swing their airy mental trapezes. Kremidas and Ranke have to be right: it is an article of faith, a presupposition of historical enquiry, that there was one and only one course of events12, held together on a single (if infinitely complex) chain of causation.

The evidence may be insufficient to differentiate between several equally plausible versions of events, and in such cases we do not know which is true. This is a common situation in all the sciences and in daily life, not a philosophical conundrum. Contrary to the physics envy that afflicts many social scientists, few natural systems are algorithmically predictable (or in reverse, algorithmically explicable): the weather can be predicted statistically, but chaotic causation, as in the collapse of a supercritical state, may make it impossible in principle to predict or explain the course of a thunderstorm raindrop by raindrop. Gould has argued persuasively 13 that in evolution randomness and inherent unpredictability are features not only of mutation, but also of natural selection: many more body plans of animals were created in the Cambrian explosion than we see today, and the survival of a mere 4% of species through the great Permian extinction depended on incidental factors which had nothing to do with their adaptation to a more normal environment.

In one sense, the existence of different interpretations, highlighted in proposition (d.), is then the normal reflection of “work in progress”; like other sciences, history grapples disputatiously with its current agenda, settles things and moves on. Much of the research published in scholarly journals in all disciplines turns out to be wrong, but only a little (one hopes) of what reaches textbooks. In 1860 creationism was a tenable alternative view to evolution, but the debate is now settled as a matter of science. But it seems Stradling means more than this: that the interpretation is a matter of individual self-expression14, that in some sense the different interpretations are equally valid: as two stagings of Hamlet may be equally good. Here we are faced with a clear choice between history as a form of literature and as a social science. This is connected to two distinct senses of interpretation.

A first type of interpretation is the expansion of an inherently ambiguous or multipotent template produced by human culture: a musical score, the script of a play, a formal axiomatic or mathematical model, a set of laws, a sacred text. When text is mapped on to text, there is scope for an infinite regress of exegesis: the Talmud piles Mishna on to Torah, Gemara on to Mishna, and that is but the start of rabbinical commentary. The degree of ambiguity in art varies greatly (compare Racine to Beckett or Chekhov, Dickens to Kafka, Raphael to Rembrandt), but some is invariably present. In every such case, there is no single right interpretation, though not all are of equal value. The reason is that the aim is not to establish causation, but meaning: and meaning is not an intrinsic property of an artefact, but the outcome of a complex transaction between the originators and audiences of messages. We can say with perfect truth of a piece of experimental music, “it may mean a lot to you, but it means nothing to me”.

The second use covers instances of perceptual inference: the passer-by interprets the girl’s smile as amiability, the dog’s snarl as a threat; the intelligence analyst interprets the image in a reconnaissance photograph as a missile launcher; the diplomat interprets the note verbale as a shift in the partner’s negotiating position; the visual cortex interprets the boundaries of brightness detected by the receptors of the eye as the edges of 3-D objects in the visual field. There is nothing inherently doubtful about these interpretations. They are subject to error, but only in defined ways: the smile may be insincere, but it cannot express threat or grief; the supposed missile launcher may be a dummy or a drilling rig, but not a school or balloon. More important, in all cases there is a single right answer. The soothsayer’s interpretation of auguries may be wholly unwarranted, but even so his predictions are either right or wrong. This object of this type of interpretation is the natural and social world, and its target is causation.

When we speak of the interpretation of a complex event like the French Revolution, of which type of interpretation are we speaking? If history claims to be a form of science - and if does not, should it not be replaced in the curriculum by something useful like carpentry? - the answer is the former. The attempt to read intrinsic meaning into history is in this view a nonscientific procedure, and one of the aims of historical education is to undermine its fascination. Its results are either folly or of the domain of theology or poetry.

There is one more way in which the siren term “interpretation” may lead us astray: in promoting a confusion between explanation, empathic recreation, and moral evaluation. It is best to approach these issues through the question of the reception of history.


I have some difficulty in grasping the exact point of Stradling’s observation (f.) that facts have a different meaning for people in different situations. At first sight, so what? Why should the historian as scholar be concerned with the reception of her work? The chips fall where they may. This would be too narrow a reading of the scholar’s responsibility. The place and style of presentation of sensitive findings does matter. Without going quite so far as the personally conservative Darwin, who delayed publication of his revolutionary and disturbing theory of natural selection for ten years until he was about to be scooped by Alfred Russel Wallace, his example of prudence, accumulation of thorough proof, and a judicious, unsensational tone is a good one. But cultural sensitivities are in the end no grounds for withholding or altering the results of scholarship.

These principles apply even more to history educators, whose wider object is the whole mental development of their pupils. Some history, especially that of the great iniquities of genocide, war, slavery, and oppression, is hard to bear and disruptive of the tribal myths of heroic virtue that may have been learnt at one’s mother’s knee or in bad earlier schooling. Reception is naturally different for those who identify with oppressors or with victims, and will modify this identification. History education should take account of these moral effects. This seems to me a distinct issue from that of historical truth, and is wrongly conflated with it by the misleading term of “meaning”.

There is a great psychological difference in reading natural history and human history. In the former, we do not naturally identify with other living species, even the higher ones that feel pain. John Hick is a very unusual theologian in considering the pain of animals from the beginnings of sentience as a problem for theodicy15. But any story, fictional or true, that involves fellow humans, automatically evokes what the Enlightenment called “moral sentiments”: sympathy, indignation, admiration, and so on. It takes extreme technical skill by contrarian writers like Borgès, Brecht, and Beckett, not to mention the less convincing legion of advanced critics, to cancel these reactions and shift us to the same critical distance we have to say wounded mammoths.

History teaching cannot evade the duty of educating these responses. This is partly a matter of setting the record straight; Richard Coeur-de-Lion was no hero, but a particularly brutal warrior and neglectful ruler; Thomas Jefferson was a slaveowner to the end of his life; Saint George never existed at all (at any rate as a Christian reincarnation of Perseus); most Frenchmen supported Vichy until 1943; Hitler was popular, and Lenin not. These corrections are important but not the heart of the matter.

Understanding history quite generally involves both empathy and alienation: overcoming barriers to seeing people with strange customs – Mongols, cave dwellers, Aztecs - as human, and also grasping the distance between our values and other cultural equipment and those of our predecessors. This process is corrosive of the simple projection into the past of current group identities. It would now be silly for Englishmen still to take sides, as Victorians did, between the Saxons and Normans who fought at Hastings: they had much more in common with each other than with us, and both groups are the genetic and cultural ancestors of most of the current population of England. The case is psychologically harder where one culture was clearly damaged in the conflict, as in the long decline of the Celtic languages and subjugation of their speakers, or the dreadful story of African slavery. But here too, almost all the relevant modern peoples or groups have a double heritage; Celtic revivalism is preached to Scottish Highlanders through the concepts, institutions, and technology introduced by Lowlanders and Sassenachs; African slaves turned into Afro-Americans, in part by absorbing and redefining their owners’ Christianity, and claiming the rights these had declared only for white men. A good history education will lead schoolchildren to a broader and more nuanced set of moral judgments on all the actors in these dramas.


The real epistemological problems of history are then overstated in Stradling’s text. Does this weaken the case for multiple perspectives? Not at all; for it rests on different foundations. “Multiperspectivity” covers two distinct propositions. The first is the technical truth that different factors were at work in the course of events and that these can be analysed using different sources and techniques (political, economic, social, women’s history, etc). The only issue of principle raised by this truism is that it is easy in presenting different strands to lose sight of the fact that the past is Circe’s single seamless web, and to encourage the haphazard accumulation of disconnected insights into a useless magpies’s nest. The history of the emancipation of American blacks and of American women, for example, is intertwined. This leads to a methodological presumption in favour of ordinary narrative history as a common pedagogic frame.

The second and more challenging sense of multiple perspectives is that of empathy. The moral dimension of history, and especially history education, presupposes an informed empathy with the actors of the past. As with education in general, the learning of history should be a process of channelling and taming the libido. We cannot achieve empathy with everybody, and in many cases we simply have no information that would allow this (say into the daily life of a Hun warrior). But with ingenuity, modern historians have shown how to offer a soundly based imaginative recreation of the experience of ordinary people16 as well as leaders, and a basis for training in the wide-ranging exercise of moral judgement. Where history is non-empathic (say that of early man based on DNA, linguistic and archaeological evidence) the problem does not arise. But if we attempt empathy, this must not be with one group alone. This duty of equal respect clashes with the natural bias and selective concerns that fuel our curiosity in the past; there is here a pedagogical challenge to which I cannot contribute. But I insist that the principle that we should seek empathy with multiple actors is ethically obligatory: the dead have no moral claims on us except to justice in our memory. Voltaire put it better : On doit des égards aux vivants, on ne doit aux morts que la vérité.

James Wimberley

May 2001

1Essay I: Of Truth, Essays, 1625. The opening paragraph is a notable assault on the relativists of his time.

2DGIV/EDU/HIST (2000) 07

3A slightly different exposition of the same position is given in the foreword to Teaching 20th -century European history, Council of Europe/DG IV, Strasbourg 2001. For simplicity I will concentrate on the Athens version.

4Sir Thomas Browne, Hydrotaphia - Urne Buriall, 1658, Ch. V.

5The classic problem of the fate of the “Princes in the Tower” (England, 1483) is illuminated by the contemporary public belief that Richard III had disposed of them.

6For lack of a documentary “smoking gun”, Lenin’s personal responsibility for murder of the Tsar and his family is disputed, though it is highly probable from other evidence and the general character of the régime. See Richard Pipes, The Russian Revolution 1899-1919, 1990, p.770.

7A small minority of qualified specialists hold that Hitler never ordered the Holocaust as such, but provided its ideological driving force and went along with a snowballing collective process of increasingly radical actions conceived by underlings. The other pole of legitimate interpretation is occupied by Lucy Davidowicz, who argued that Hitler decided on the extermination in 1918 and entered politics to carry it out. Cf. Ron Rosenbaum, Explaining Hitler, 1998. The David Irving position, that Hitler was a bystander, is of course specious apologetic.

8Cf. the use by Emanuel Le Roy Ladurie of Inquisition records in Montaillou.

9For example, base-year price indices understate growth in real income through technical progress; the number of registered unemployed understates the total of those who would like to work.

10Jan Huizinga, in The waning of the Middle Ages, suggests that our perception of the period is distorted by this highly selective survival of its art.

11One Gnostic heretic is supposed to have held that at the birth of Christ, the whole world was created together with its entire previous history.

12In the macroscopic world, we need not worry about quantum superpositions, which collapse when the quantum state interacts with the environment. In any case, a quantum superposed state is a fact like any other, and quantum physics allows deterministic calculations about it to a staggering degree of precision. The famous puzzle of Schrödinger’s cat has apparently been resolved without recourse to a magic role for the observer, or the heroic multiple universes of Everett – not, remember, a gratuitous speculation, but an attempt to deal with a mathematical inconsistency in an otherwise spectacularly successful and predictively reliable physical theory. Does anyone argue that Wallenberg was neither alive nor dead till his death was proved? See Max Tegmark and John A. Wheeler,"100 Years of Quantum Mysteries", Scientific American, February 2001.

13S.J. Gould, Wonderful Life, 1989, Ch. V.

14Cf. Bacon, op. cit.: “Certainly there be that delight in giddiness, and affect it a bondage to fix a belief; affecting free-will in thinking, as well as in acting.”

15John Hick, Evil and the God of Love, 1968, p.252

16For a fine example, see John Keegan’s The Face of Battle.