If we are to believe veteran anthropologist Walter Goldschmidt, "Linguistics did not become a separate discipline until some ten years after World War II", that "Boas taught his students how to learn native languages and to work out grammars by appropriate questioning of native speakers", and that "[t]his is where the scientific study of language began and out of which the early linguists came" (Goldschmidt 2000:801). If we were to connect the scientific study of language with the establishment of separate linguistics departments in the United States, Goldschmidt may not have been that far off the mark. It is true that despite the founding of the Linguistics Society of America in 1924, there were few universities that could boast a department of linguistics (in contrast with traditional Indo-European philology units which frequently were housed within departments of classics) before the 1950s or even the 1960s. However, using formal institutionalization as the criterion for marking the beginning of linguistics as a scientific discipline would mean taking a rather narrow view. In fact, it would suggest that 19th-century historical-comparative Indo-European linguistics and, more importantly in the present context, the great work of Sapir and Bloomfield of the first half of the 20th century was not yet 'scientific', because their work was done within departments of anthropology and Germanic languages, respectively.
However, as Julie Andresen has shown in her book Linguistics in America 1769-1924, we could safely go back at least to the work of William Dwight Whitney (1827-1894) when talking about what she calls 'the arc of development of "linguistic science'" (Andresen 1990:180ff.). As a matter of fact, she sketches the 'institutionalization of American linguistics' from 1842 onwards, with the founding of the American Ethnological Society, in which Albert Gallatin (1761-1849), a former Swiss financier, played a seminal role (Andresen 1990:120). For Robert E. Bieder, in his Science Encounters the Indian, 1820-1880 (1986), the scientific study of American Indian languages would have begun even earlier and, as I intend to show in this chapter, there may be good reasons for putting the date of the beginning of linguistic analyses of languages without written tradition much further back than the early 19th century.