This similarity is not to be looked for in likeness between words, but in the inner structural development of tongues. To ascertain and estimate such identities is a far more delicate undertaking than to compare columns of words in vocabularies; but it is proportionately more valuable.
Nor should we expect it to be absolute. The example of the Basque in a pure white nation in Western Europe warns us that there are exceptions which, though they may find a historic explanation, forbid us all dogmatic assertion. They are so few, however, that I quote Dr. Winkler’s words as the correct expression of the latest linguistic science, and I wish that some investigator would make it the motto of his study of American tongues.
The task—no light one—which such an investigator would have, would be, first to ascertain what structural traits form the ground plan or plans (if there are more than one) of the languages of the New World. Upon this ground-plan he would find very different edifices have been erected, which, nevertheless, can be classified into groups, each group marked by traits common to every member of it. These traits and groups he must carefully define. Then would come the separate question as to whether this community of traits has a genetic explanation or not. If the decision were affirmative, we might expect conclusions that would carry us much further than etymological comparisons, and might form a scientific basis for the classification of American nations.
Possibly some one or two features might be discovered which though not peculiar to American tongues, nor fully present in every one of them, yet would extend an influence over them all, and impart to them in the aggregate a certain aspect which could fairly be called distinctive. Such features are claimed to have been found in the grammatic processes of _polysynthesis_ and _incorporation_.
Peter Stephen Duponceau, at one time President of the American Philosophical Society, was the first to assert that there was a prevailing unity of grammatic schemes in American tongues. His first published utterance was in 1819, when he distinguished, though not with desirable lucidity, between the two varieties of synthetic construction, the one (incorporation) applicable to verbal forms of expression, the other (polysynthesis) to nominal expressions. His words are—