The Hamburg Classification of North America

     At the University of Hamburg a project extending over the last ten years investigated the native languages of North America. A new classification of these languages was worked out, called Hamburg Classification. This page is intended to give some information about this research.

     The native languages of North America include the Eskimo-Aleut languages and a vast number of Indian languages, and how to classify these languages from a genealogical point of view has been the matter of a long history of research - as is well known.
     In 1997 Lyle Campbell published his book "American Indian languages - The historical linguistics of Native America". It contains an enormous amount of information about the languages, methods to investigate their relationships, theories about these topics and more. For North America the book presents a classification with 58 units. Now it should be possible to reduce this number. Indeed there have of course been theories which do so for more than a century; one just has to think of the work of Dixon / Kroeber, Sapir, Haas, etc. None of the complete classifications published earlier, however, was without shortcomings.
     Co-workers and I took a fresh look at the matter in the years after the publication of Lyle's book, evaluating the vast literature, comparing data etc. It turned out that Penutian indeed exists. The same is valid for Hokan, to which Zuni and Keres could be added, as well as Yukian. There is a small Gulf consisting of Muskogean plus Natchez and Timucua. In fact the literature often already contains the evidence that proves a point, though sometimes it is hidden away among wrong or doubtful statements.
     Time did not stand still elsewhere, of course; John Enrico made a good new effort for Na-Dene. My 2005 book "Einführung in die eskimo-aleutischen Sprachen" contains a detailed proof for Morris Swadesh's old hunch that Eskimo-Aleut and Wakashan are related. A slightly enlarged Tunican exists. Finally, and quite surprisingly, Kutenai and Kiowa-Tanoan turned out to be related.
     In the numbering of Lyle Campbell's book, the following classification emerges, consisting of 14 language families:

Eskimo-Wakashan 1, 6
Na-Dene 2 - 4
Salish 8
Algonquian-Ritwan 57, 58
Siouan 39
Uto-Aztecan 35
Penutian 5, 10 - 18, 30 - 33
Hokan 19 - 29, 34, 36, 38, 42 - 48
Iroquoian 56
Caddoan 40, 41
Chimakuan 7
Tunican 49 - 51, 55
Gulf 52 - 54
Kutenai-Kiowan 9, 37

     This classification is extremely reliabe. It seems that it is correct, i. e. that it meets the historical facts, in all aspects. A computer was used for work on it, too.
     The Hamburg Classification has been published in two articles:

Holst, Jan Henrik (2008a):
Indianersprachen: Ein Forschungsgebiet stellt sich vor. P. 10 - 24 in: Amerindian Research 3 / 1, No. 7.
(Published February 2008.)

Holst, Jan Henrik (2008b):
Geschichte der Klassifikationen der nordamerikanischen Indianersprachen. P. 226 - 237 in: Amerindian Research 3 / 4, No. 10.
(Published November 2008.)
(ISSN of the journal: 1862-3867.)

     The first article is actually an introduction to American Indian languages. However, it already uses the new classification as its basis and reports something about how it came into being.
     The second article is a history of the classifications of North America. It lays special emphasis on how the new classification emerges from the insights that had already been achieved in the past, but it also covers what new paths were taken.
     It is interesting that there already appeared pieces of evidence from non-linguistic disciplines that confirm the Hamburg Classification. As I was informed only recently, the Kiowa have a tradition which says that they came from considerably further north, from a place near the Black Hills. Now this is very close to where Kutenai is located. I did not have this knowledge about Kiowa culture at my disposal when I investigated the languages and set up Kutenai-Kiowan from purely linguistic evidence. This is a nice instance of how things converge. It also shows that oral literature often has kernels of truth, or more, in it and must be considered seriously.
     Hopefully these results will be an incentive for further research on the languages of North America. Moreover, it will be possible to direct future classificatory work to other areas, e. g. Middle America, and Africa.

back to main page