Ce

nous arriverons à cette triste conclusion, qu'il ne s'est rien passé depuis la création du monde. n'est pas que je méprise la critique; au contraire, et je la préfere de beaucoup à l'éloquence chez ceux qui me racontent le passé, mais encore faut-il qu'ils me racontent. Je demande à l'histoire de rentrer dans le genre narratif."

What then can we do towards constructing the wished-for canons of historical credibility?

We begin, of course, with careful definitions.

Historical events, as distinguished from personal details, are such as concern a whole nation, and (except secret treaties and the like) are notorious throughout the country, and usually also in neighbouring countries.

Credible means probable, i.e., resting on evidence which in the majority of similar cases has been verified, or at least is accepted as

true.

It must be noted that, as in matters of "probability," so here also, antecedent likelihood must be allowed to strengthen or weaken the chances made out by evidence; and moreover, when our conviction that a certain alleged event is true would drive us into a change of conduct or of sentiment, we require the odds in its favour to be very great. The foundations must correspond to the superstructure.

Next to definitions come axioms; only, in this inexact science, we must use the less forcible Latinderived equivalent,-assumptions. In historical as in legal investigations we assume.

To resume, men tell the truth if they know it, and take the trouble, and have no interest the other way. Events admitted to have happened must have been caused by previous events; and if history (not mythology) mentions a vera causa, i.e., previous events such as might very

well have happened, any other hypothesis is gratuitous.

of

The undesigned consensus independent authorities is convincing and conclusive as far as it goes (of this proviso more hereafter), whether the authorities (i.e., persons who knew) are authors of books, inscriptions, coins, &c., in which case facts may be established "out of the mouth of two or three witnesses," or are merely "folks," in which case they must be very numerous, for only so is the " populi vox Dei"- only so may constat be translated "it is well known."

VOX

It is interesting to note how our definitions bear upon our assumptions. The public importance of historical events leads us to expect that many contemporaries did know the truth about them, so that what was uncontradicted at the time ought not to be lightly discredited, while what was omitted at the time ought not to be easily accepted.

The magnitude of historical events forbids us to put aside alleged causes of admitted effects on the plea (which is, however, a sceptical man's only safeguard in the face of well- substantiated ghost stories and the like) that we cannot be expected to account for everything."

66

The great number of circumstances, important or unimportant, surrounding a historical event, offer additional scope for undesigned coincidences, while it is also impossible to suppose-as may happen in the subjects of lawsuits and trials-that a false account is due to a conspiracy either of either of historians or of multitudinous

rumour-mongers.

Next, to examine more closely some expressions that have just been employed.

What kind of persons "know the truth" about historical events?

Experience must furnish the answer. Sometimes it is experience of the particular authority we are valuing. If we have found him evidently as well-informed as he professes to be on all points which we can test by other authorities, we shall esteem him as well-informed as he professes to be on other similar points.

Thus, Herodotus's statement that Xerxes made a canal on the isthmus behind Mount Athos having been verified by the discovery of its channel, we have a reasonable presumption that he has correctly described the whole route of the Persian fleet and army.

But, unfortunately, it too often happens that we have only one authority for certain events, and that it relates no other events, or at least no other similar events, which we can test, and so estimate its trustworthiness.

Thus it is only in Herodotus that we can find a record of, or even an allusion to, the remarkable conversations which he tells us Xerxes held with Demaratus, the exiled king of Sparta. In such cases our resource is to consider what proportion of such conversations are usually reported so eagerly and faithfully as to reach a foreigner fifty years afterwards; and, when other writers give similar conversations which can be verified, what proportion of them are found to be substantially correct.

66

In like manner we calculate the probability that a writer has taken the trouble" to ascertain the truth, and that the evidence before us is not biassed by personal prejudice or national vanity.

It is needless to point out how great is the credibility of, say, a democratical writer who records democratical excesses and failures, and of national chronicles or monuments which publish national crimes or disasters.

So far we have been speaking of writers (including authors of inscriptions, monuments, and superscriptions on coins) who could have known the truth, the only question being whether they have the requisite discrimination and research, and were free from bias - a question decided partly by comparison with other authorities, partly by the writer's own statements about the means he employed to ascertain the truth, partly by the internal evidence of style and sentiments.

By the expression "the question is decided" is meant that, after examination, we can reckon our authority among those which are always, or usually, or as often as not, or sometimes, or very seldom, or never, to be trusted for such and such a kind of event. We can take the evidence as worth, say, 20, 28, 28, 25, 25, or 6, the simple fact being (suppose) that fifteen out of twenty contemporary inscriptions record historical events as they really happened.

19

15

As to persons who could not have known the truth there is a great controversy. Is contemporary evidence indispensable? i.e., evidence derived from what was reported or recorded in some way by persons who had arrived at years of discretion when the event took place?

We must beware of taking as an axiom either this favourite assumption (so dear to Sir G. Cornwall Lewis), or another antagonistic to it, namely, "There must be some foundation in fact for the current account, even though quite mythological, when any other account is confessedly mere guesswork." For, on the one hand, the accurate preservation of historical events by tradition is not only quite conceivable, but has been proved in some instances by the unexpected discovery of confirmatory records.

Thus the traditional site of a battle has been verified by excavations.

On the other hand, though the existence of any current account must have had a definite cause, that cause may easily be conceived to be for ever undiscoverable; nor can we recount any instance in which a history has been successfully built up out of the ruins of a 'pure and simple" mythology.

66

It is easy enough to omit or explain away the miraculous or superhuman details when they do not interfere with the sequence of events. Thus the credibility of Livy's history after the war with Pyrrhus is not at all impaired by a paragraph here and there about a rain of drops of blood, or an ox speaking with human voice. Thus, too, we may quite believe Xenophon's assertion that he won such and such a battle after waiting several hours for favourable omens, without agreeing with him as to the prophetic value of the entrails inspected by his priests. But when discarding the "marvellous element" obliges us to reconstruct the whole narrative, as we must do when we seek the foundation in fact for a genuine myth, then the result has no claims to be accepted. So it is with the Trojan war, and with what happened when "good King Arthur ruled this land."

On the whole it is established by experience that wherever the historical faculty of a nation was too undeveloped to produce among them authentic books, monuments, coins, or other records of public events, their unhistorical instinct was so active, or their "talent for silence" so great, that their traditions, historically speaking, are worth little or nothing at all. Nor must there be any appeal to the assumption of undesigned consensus. Such tales are not supported by the undesigned coinci

dence of independent authorities. People who repeat a myth about what took place "a long time ago" are not authorities, since they had no means themselves of ascertaining the truth; and they are not independent, because the narrative probably sprang first from some professed "maker" (poet) in prose or verse, instead of flowing from a great number of different sources like the authentic tradition of a national migration or an important battle.

Investigators may therefore take this as a canon of historical credibility:

A current narrative not founded on contemporary evidence cannot be maintained against objections from intrinsic improbability, or from discrepancies and variations. But if its items are such as might very well have happened, and are an adequate explanation of subsequent facts, and if it cannot be traced to individual invention, but seems to have grown up naturally as a genuine tradition; and, further, if it stands alone without a rival narrative, then scepticism is uncalled for.

Next let us suppose ourselves dealing with authorities who were within reach of the truth, and were independent of one another, but liable, of course, to error.

What is the precise importance of coincidence between them, or between different passages in the same authority? And what is the importance of discrepancy, whether in the way of variation or omission?

The favourite and well-deserved epithet for coincidences is striking. Here is a phenomenon which, once noticed, cannot be neglected. A cause for it must exist, and is readily found in the reality of the event recorded by one writer, but merely alluded to or implied by the others.

Is this the only cause? We must be on our guard here.

Certainly the phenomenon is one that cannot be accidental and meaningless, and its importance has been fully recognised ever since the publication of Paley's Hora Pauline. Only one inference, however, is quite safe the rule will cover the rare cases of its introduction into fictitious narratives by ingenious authors-namely, that the thing which forms the subject of the "undesigned coincidence was very deeply impressed upon the mind of each writer who "undesignedly" alludes to it or implies it-so deeply as to influence his thoughts and language whenever he touched upon it. More briefly, we cannot doubt that he had got the thing thoroughly into his head. "How he could have got it into his head if it is not true" is, therefore, the point to be cleared up by those who wish to deny it.

We have to note one comprehensive explanation and one caution.

The character of a hero, and the date and locality of his chief (supposed) adventures, become exceeding familiar to his countrymen and admirers. Hence we find very subtle coincidences in early Greek authors concerning the famous Hercules and Achilles.

The caution is, to limit the proving power of coincidences to the one fact in which the authorities concur, with, of course, its necessary antecedents and consequents. We must not pledge ourselves to any separable surroundings or avoidable inferences. Here is an instance :

The details of the visit paid to Spires by St. Bernard of Clairvaux, while he was preaching up the Second Crusade, are related in two independent narratives, both composed by persons who were there at the time and had the fullest means of information. By a sudden

enthusiastic appeal he had prevailed upon the reluctant Emperor Conrad to become a Crusader. One of the writers, Godfrey, in his "Life of Bernard," tells us that while the saint was being escorted by Conrad to his lodging, the people brought him a lame boy, whom he cured instantaneously, and that while he was at mass in the chapel near the canon's apartments he restored sight to a blind woman. Our other authority, one Philip (in a diary kept by himself and nine others for the express purpose of recording the exact truth), states that Bernard, in his own presence, healed a lame boy, on the same day as he made his speech to the Emperor, while at mass in the bishop's chapel, and afterwards restored sight to a blind man. Both historians lay much stress on the cure of the lame boy; and as Godfrey wrote his biography some ten years later-so that mistakes in details do not impair his authority-we cannot but be struck by the coincidences of the two narratives in this material point, and we are bound to believe that St. Bernard went through some public performance with a (professedly) lame boy, in or near a chapel, on the day that Conrad took the cross.

Next as to discrepancies.

First from Omission. No omission is of any importance unless the author must have been acquainted with the event if it had really happened; unless also the design of his work made it absolutely necessary for him to mention such an event if he did know of it; and, further, unless the omission cannot possibly be considered a mere slip or accident.

Thus Livy does not tell us of the commercial treaty made between

Rome and Carthage B.c. 509; but he may not have had access to the archives in which Polybius discovered it; while, on the other hand, if Belisarius had ever really become a blind beggar, Procopius, his companion and a contemporary historian, must have known it; and, having known it, Procopius must have mentioned it in his account of that famous general's life and misfortunes, or else must be condemned as a feeble, stupid biographer. Though a historian of the war between the Northern and Southern American States might know, but need not inform the world, that one of the most famous of the Northern generals is now manager of a drygoods store.

It was by a mere slip that Marco Polo omitted from his narrative of travel all notice of the Chinese Wall, the design of the work requiring its insertion; but the silence of Josephus with regard to the Christian Church warrants the strongest suspicion that, like the Pharisees with regard to the baptism of John, he could not and would not tell "whether it was from heaven or of men." The omission is a proof of disingenuousness; while a silence on the same subject in the philosophical works of Seneca, who flourished in the reign of Nero, proves him strangely unacquainted with the noblest moral system of his time-the omission is a proof of ignorance.

Omissions are highly significant if they occur in the earlier as compared with the later versions of a story intrinsically improbable. Montaigne (quoted by Bentham) shrewdly observes: "When men repeat an improbable story, they find out its weak points, and supplement it by inventions," neces