Page:Language and the Study of Language.djvu/343

VIII.] same tongues have been brought out in sufficient number, or of a sufficiently unequivocal character, to constitute, along with these correspondences of form, such an argument in favour of the unity of the family as may be deemed satisfactory and accepted, is also a matter for doubt. It is safest to regard the classification at present as a provisional one, and to leave to future researches its establishment or its overthrow. The separate investigation and mutual comparison of many of the dialects is as yet only very imperfectly made, or even hardly commenced: farther and more penetrating study may strengthen and render indissoluble the tie that is already claimed to bind together the eastern and western branches; but it may also show their connection to be merely imaginary.