A statistical approach to African personal pronouns
Guillaume Segerer & Martine Vanhove - CNRS LLACAN, France

7. Other patterns in African systems

We checked other combinations involving m, n or T for 1sg-2sg pronominal subsets.

  • The most striking feature for Africa is the frequent presence of m-m languages in the Niger-Congo stock (19%).

  • Although 8 of the 11 major branches are concerned (Adamawa, Atlantic, Benue-Congo, Gur, Kru, Kwa, Mande, Ubangi), the feature is really salient for only 4 of them (Kru, Adamawa, Ubangi, Atlantic), and to a lesser extent in Gur.

  • The only Nilo-Saharan groups showing this feature are Maban and Central Sudanic, the latter of which is already well-known for its long-standing contact with Ubangi languages of the Niger-Congo stock (Cloarec-Heiss 1995).

  • But: Statistically based predictions:
      65% of the languages have an m for 1st person in at least one paradigm.
      19% of the languages have an m for 2nd person in at least one paradigm.
    In the 738-lgs sample, if the presence of m in 1sg and m in 2sg are independant from each other, there should be 738*0.65*0.19 = 93 languages showing the m-m pattern, which is very close to the 109 m-m languages found.

  • Other salient types are:
    - m-N: 62 languages (8%)
    - N-T: 60 languages (8%)
  •  <  ALT8 (8th Biennial Conference of the Association for Linguistic Typology), University of California, Berkeley, July 23-26, 2009  >