"Fools rush in where angels fear to tread."

For angels let us read logicians, and for fools substitute that class of persons who of all others indulge most, and succeed least, in ratiocination-persons who dip into history for the first and only time with a determination that it shall supply some final solution, or, worse still, that it shall confirm their own preconceived solution of some exciting problem or controversy, be it the truth of their religion or the character of Mary Queen of Scots.

And why is such an enterprise almost certain to fail, and worthy therefore to be denounced as foolish? Because no one could imagine beforehand, no one without wide and unprejudiced research could believe, the weakness and foolishness of us mortals in furnishing and in using the materials of history.

Yet we must needs expect something of the kind, far as our expectations are surpassed by the reality. Historical evidence, in matters of detail, is merely legal evidence of a very weak sort. Was Mary an accomplice in the murder of Darnley? Did the Persian Xerxes cut a canal through the isthmus of Mount Athos? These and very many other historical problems

could have been investigated at the time, as some of them were investigated, in courts of law. The lapse of years cannot render the conditions of certainty or probability any less stringent. We cannot acquit or condemn Mary now on evidence that made "not proven" the only just verdict three centuries ago. The same narrative is not any more credible in itself as "History of the Crimean Campaign" than as "Letters to the Times from the Seat of War."

A professed historian is a person whose capacities and motives for telling the truth we have not yet discussed. We may provisionally assume that all men tell the truth if they can, if they take the trouble, and if they have no interest the other way. As the basis of the labours of the historian, we have to deal with certain kinds of "real" evidence, as coins and inscriptions; and long unquestioned notoriety claims a considerable weight with practical men. We ask no other proof that Mahomet fled from Mecca to Medina A.D. 622; and let it not be thought that there is here any contradiction of the protest just launched against conniving at weaker evidence on account of remoteness of date. We refer here to broad, general facts, and not to details, such as the question whether the Prophet was saved from his pursuers by the spider's web and the pigeon's nest in his

cave.

Moreover, the great principle to be borne in mind while we proceed to distinguish historical problems from law suits and criminals trials is this: Usually these problems are merely speculative. It makes no difference to our peace of mind or our manner of life whether there were seven kings of Rome or none at all. It is for this reason that we are prone so readily to adopt either the traditional account of the matter in hand, however slender its supports, or else the hypothesis suggested by the best and latest authorities, however slight the preponderance of evidence in its favour. But, assail the truth of our religion or the honour of our country, and our attitude is wholly different. We insist on maintaining any theory we like, unless it is upset by more than the full amount of proof which the ideal court of law would require in a case that rested merely on hearsay. We are as assured as if we possessed documentary and "real" evidence, and held a dispensation to ignore the priceless privilege of oral crossexamination.

And here the student may once more be reminded that there are two tests of truth of fact which supersede all others: verificationi.e., the verdict, when it is attainable, of our own (undeluded) bodily senses-seeing is believing; and consistency, wanting which, we are at least certain that one of the two conflicting accounts is, in this item, false.

Nor should he for a moment forget that, while inconsistency is conclusive as to the presence of error, consistency is only a condition, not a guarantee, of truth. An inference deduced from admitted premises is irresistible only when the assumption of its contradictory leads to a reductio ad absurdum. No induction is correct unless we can retrace our steps deductively

(by the rules of the Syllogism), and thus arrive again at all the particulars from which we started, and under similar conditions (if any) at similar particulars; while under no conditions must we thus infallibly reach any wrong conclusion. Our hypothesis must be adequate, must (if possible) predict, and must on no account prove too much. And, further, no rival hypothesis must do equally well.

We must expect, then, to find great difficulty in establishing Canons of Historical Credibility, more lax than those of Law Courts, yet so axiomatic that we may denounce as unreasonable and illogical anyone who refuses to credit a historical event which satisfies those

canons.

They must rest, of course, upon experience. They will not be selfevident. A denial of them can never involve a contradiction in terms, a refusal to think on the subject, like a denial of the canons of causation. At best our ground for asserting that they never will fail will be the fact that they never have failed, and even this involves a kind of argumentum ad hominem i.e., if you do not feel obliged to believe this narrative, you cannot feel obliged to believe any similar narrative, for the evidence is as complete as (in such a case) historical evidence can be.

What, then, does experience tell us as to the separate and combined value of the various sources of historical information-i.e., notoriety, books of history (including biographies, autobiographies, and historical allusions in letters, treatises, and other compositions), monuments (including coins, inscriptions, and ceremonies), institutions, language, and undisputed subsequent events?

It tells us this: Every one of them that involves any conscious effort to be historical may lead us

into the grossest error. The last three give unconscious, and therefore unerring, testimony; but they say little, and that little hard to interpret aright. Taken all to

gether, they are outweighed by a single grain of verifiable fact or demonstrable truth. The mention of an eclipse in ancient history enables modern astronomers to fix a date infallibly, though in the teeth, it may be, of all received chronology. "Mons. Thiers himself and a host of French historians may repeat the anecdote of Le Vengeur refusing to strike her flag in the action of June 1, 1794, and going down into the depths of the ocean, while her crew shouted 'Vive la Republique.' But Admiral Griffiths saw Le Vengeur taken possession of by the boats of the Culloden; saw the Frenchmen trying to save themselves; heard their outcries, which were merely those of horror and despair."*

Nor would any impartial person hesitate to treat great intrinsic improbability and inconsistency with all analogies drawn from what is verifiable, as counteracting even seemingly flawless testimony. For instance (though neither the narrative nor the objection is unexceptionable), belief in the Seven Kings of Rome (B.c. 754-510) till A.D. 1624 was as universal as belief in the Twelve Cæsars (B.c. 50 to A.D. 100); but is almost overthrown by the one circumstance that an average duration of thirty-five years to a reign in such times is unheard of and perhaps morally impossible.

As the chief object of this paper is to impress upon the student of logic that in this field of controversy, as in all others, "because and therefore are edged tools not

to be played with except by experts," we will exhibit some further difficulties to be found in arriving at historical truth.

There is of course a broad line of demarcation in this respect, nay, an immeasurable gap, between the long ages before the birth of historical criticism and the two centuries that have since elapsed. No so-called history is now received as such unless it satisfies all that we have learnt to demand in the way of care and research; while universal education, and all our modern means for the publication, transmission, and correction of news are sufficient guarantees that no broad or general historical facts of these and future times will ever be involved in the mists of doubt.

As to details, on the other hand, errare est humanum must always hold good. Lord Russell was unable to keep his " Recollections" free from such a gross blunder as transposing the chief diplomatic events of two consecutive years, and thus "inverting the true relations of the persons most cerned." All Mr. Kinglake's labo-riousness could not produce a really accurate account of the battle of Inkermann.

con

But enough has been said about details already. The following are some instances of signal failure in the various kinds of historical evidence with regard to general facts. It is admitted that all Roman History before the war with Pyrrhus (ie., for the first 472 years from the era of the foundation of the city) must always be mere guesswork. And the records of every ancient nation include a period of mythology and vague tradition, a period so obscure that,