Tue Oct 13 06:21:14 UTC 1998

A somewhat similar development of new forms of pronouns, sometimes based

upon phrases such as "my soul, your soul etc." occurs in several
Ethiosemitic and Cushitic (e.g., Beja) languages, but usually only in
and 2.nd persons, while 3.rd person pronouns are innovated through
demonstratives as in many other language families. 2.nd and 3.rd plural
forms are also innovated in Southern Ethiosemitic and Highland East
Cushitic because the old forms specialise semantically for polite address,
and new forms are created for plain 2.nd or 3.rd pl. Specific literature on
this, however, I don't know. The single cases are discussed in grammatical
studies focused on the individual languages.

Now the question you ask, i.e., "on what systematic and objective grounds
opposed to tradition and/or fuzzy intuitions) are we able to distinguish
between "primary / underlying / basic" meanings of pronouns and
"secondary / derivative / extended" ones". One criterion would be
etymology, i.e., what can be reconstructed historically as the original
pronoun forms. In a language with an agreement system, an additional kind
of distinguishing criterion would be agreement patterns.

A parallel can be found in several East and Southern Cushitic languages
such as Bayso, Arbore, or Iraqw, where the inherited occurrence of pluralia
tantum and of non-plural agreement for some semantically plural forms, as
well as the development of singulative forms developed into a very complex
system of agreement classes where formal agreement patterns of nouns do not
match the morphological meanings of unit reference, plural reference and
(if existing) paucal reference. This is described in:

Hayward (1984) The Arbore language. Hamburg: Buske.

Corbett and Hayward (1987) Gender and number in Bayso. Lingua 73:1-28.

References to this kind of phenomena are also found in Corbett, Gender.

For Iraqw I discussed this in my review of Mous's Grammar of Iraqw, on JALL
18 (1997):95-106, especially pp. 98 ff.


Giorgio Banti
IUO - Naples

More information about the Lingtyp mailing list