Specification
What the observations made in the preceding section (How does a language acquire gender markers?) suggest is, first, that it is not demonstratives but rather erstwhile personal pronouns that appear to be responsible for the spread of gender marking and the rise of gender agreement in the Khoe languages. Second, it is not a syntactic structure, like [head — modifier], that is involved but rather a pragmatic pattern according to which a constituent of the noun phrase (= the specified) is further specified by means of a personal pronoun that provides more information on the personal deixis, gender, and number of that constituent (= the specifier). The specifier has the following properties or part thereof (cf. Heine 1997:3.2):
It refers to the same general entity as the specified.
It further characterizes the specified by narrowing down the range of possible referents that may qualify as the specified. This can be achieved by providing more information on the specified, and/or by making clear which of the alternatives available qualifies as the specified.
It lays “emphasis” on the specified (e.g. by referring to it a second time by means of a different expression).
It makes the specified more easily recoverable in subsequent discourse.
Typically, the specifier follows the specified, although in some languages, the order of the two appears to be reversed.
All properties apply in some way to the pgn use described above. Perhaps the most immediately relevant property is that mentioned in (b). That pgn’s provide more information on the specified is perhaps most obvious in their derivative use: In all Khoe languages on which some grammatical information is available, pgn’s are used as derivative suffixes, where two kinds of functions can be distinguished. On the one hand, they serve “nominalization”, that is, they turn nominal modifiers into pronouns or verbs into nouns, as in (1).
(1)
On the other hand, pgn’s also turn semantically less specific nouns into more specific ones, as in example (2).
(2)
According to (c), specification may have to do with emphasis. A number of authors dealing with Khoe languages have in fact pointed out that pgn use can express “emphasis”. For example, we observed that the emphatic imperative in Kxoe requires a pgn to be placed between the verb and the vocative enclitic è (Köhler 1973:54), and Bleek (1928:55) refers to the xa marked third person pronouns in Naro as “the emphatic form” and, in a similar fashion, Barnard (1985:15ff.) calls the short form of the personal pronoun of Naro the “non-emphatic pronoun” and the full form the “emphatic pronoun”.
That the use of pgn’s has to do with “emphasis” can be shown perhaps most clearly with the following example from Kxoe. The third person pronouns of Kxoe consist of what is diachronically a demonstrative (in most cases xà-) plus the appropriate pgn, exemplified in (3a). Now, the same structural pattern exists for first and second person pronouns, except that in the latter two cases the combination [demonstrative + pgn] does not express a simple person pronoun but rather an “emphatic” one, as in (30b), while the “non emphatic” first or second person pronoun is identical with the pgn, as in (3c).
(3)
What these observations seem to suggest is that demonstratives are employed in Kxoe to “emphasize” personal pronouns, the result being a paradigm of “emphat ic pronouns”, exemplified in (3b). In the third person, “emphatic pronouns” have replaced the “non-emphatic” ones. In this way, they are now used as the unmarked third person pronouns. At the same time, they have also retained their earlier function, with the effect that they are ambiguous between their earlier (“emphatic”) and their later (“non-emphatic”) uses.
Note, however, that neither is the term “emphasis” reported to be present in all uses of pgn’s in Kxoe and other Khoe languages, nor does it become entirely clear what “emphasis” exactly means (cf. Hagman 1977:46 for a particularly idiosyncratic use of “.
