If this be the spirit in which we must scrutinize documentary evidence, with what eyes must we look upon traditions--traditions wherein the record, instead of being permanently registered, is transmitted from mouth to mouth, from father to son, from the old man to the young, from generation to generation? The mere etymological import of the word will mislead us. It is not enough for a thing to have been _handed down_ from father to son. A relic may be so transmitted; indeed, written papers and printed books are traditions of this kind. Heirlooms of any sort--whether belonging to a nation or an individual--are such traditions as these. In a true tradition we must consider the _form_ and the _origin_. A narrative which has taken a definite shape, either as a formula or a poem, can scarcely be called a tradition. It is a specimen of composition handed down by tradition, but not a tradition itself. It is an unwritten record--as much a record in form and nature as a written document, but differing from a written document in the manner of its transmission to posterity. Many a good judge believes that the Homeric poems are older than the art of writing, and, consequently, that they were handed down to posterity orally. Yet no one would say that the Iliad and Odyssey were Greek traditions. The fact of a narrative having taken a permanent form, inasmuch as that permanent form both facilitates its transmission, and ensures its integrity, distinguishes an unwritten record from a tradition. A true account of a real event transmitted from father to son in no set form of words, but told in a way that a nursery tale is told to children, or the way in which a piece of evidence is given in a court of justice, constitutes a tradition; for in this form only is it liable to those elements of uncertainty which distinguish tradition from history--elements which we must recognize, if we wish to be precise in our language. Such is its _form_, or rather its _want of form_. But this is not enough. A tradition, to be anything at all, must have a basis in fact, and represent a real action, either accurately described or but moderately misrepresented. I say _moderately misrepresented_, because the absolute transmission of anything beyond a mere list of names, and dates, without addition, omission, or embellishment, is a practical impossibility. Hence we must allow for some inaccuracy; just as in mechanics we must allow for friction. But, allowing for this, we must still remember that the event and the account of it, are correlative terms. An opinion--an account of an account--only takes the appearance of a tradition. It is a _tradition_ so far as it is _handed down_ to posterity, but it is no tradition with corresponding facts as a basis.