Why did Greenberg reject classifications based only on typological similarities?

Created from Britannica articles using AI. AI can make mistakes, so verify using Britannica articles.

Thematic map of Africa showing the distribution of Niger-Congo language branches. Distinct colors represent language groups such as Atlantic, Mande, Gur, Kwa, Benue-Congo, and Bantu, with Bantu languages covering most of central, eastern, and southern Africa. A numbered key lists prominent languages in each area, and country borders are marked for reference. — Niger-Congo languages Map showing the branches and distribution of the Niger-Congo languages.

Joseph H. Greenberg rejected classifications based solely on general typological features—such as the presence of noun classes—because he believed they were insufficient to prove a genetic relationship between languages.1

He argued that such classifications were only valid if they were substantiated by a detailed comparison of the actual forms by which those systems were realized. To ensure accuracy, Greenberg insisted on comparing specific grammatical morphemes across languages to determine if they shared both similar forms and functions. This rigorous approach was intended to move beyond superficial structural similarities, which can occur for reasons other than common ancestry, and instead focus on evidence that could demonstrate a shared genetic origin.12