Page:Language and the Study of Language.djvu/267

VI.] one: it waits for the demonstration in each separate case. The claim made in our third lecture, that, in the history of linguistic changes, any given sound may pass over into any other, any given meaning become modified to its opposite, or to something apparently totally unconnected with it, may seem to take away from etymology all reliable basis; but it is not so; for the same researches which establish this claim show also the difference between those facile changes which may be looked for everywhere, and the exceptional ones which only direct and convincing evidence can force us to accept as actual in any language; they teach us to study the laws of transition of each separate language as part of its idiosyncrasy, and to refrain from applying remote and doubtful analogies in the settlement of difficult questions.

In short, the modern science of language imposes upon all who pursue it thoroughness and caution. It requires that every case be examined in all its bearings. It refuses to accept results not founded on an exhaustive treatment of all the attainable evidence. It furnishes no instruments of research which may not be turned to false uses, and made to yield false results, in careless and unskilful hands. It supplies nothing which can take the place of sound learning and critical judgment. Even those who are most familiar with its methods may make lamentable failures when they come to apply them to a language of which they have only superficial knowledge, or which they compare directly with some distant tongue, regardless of its relations in its own family, and of its history as determined by comparison with these. A scholar profoundly versed in the comparative philology of some special group of languages, and whom we gladly suffer to instruct us as to their development, may have nothing to say that is worth our listening to, when he would fain trace their remoter connections with groups of which he knows little or nothing. Notwithstanding the