The feminine anaphoric gender gram, incipient gender marking, maturity, and extracting anaphoric gender markers from parallel texts

The aim of this paper is to carry out a typological study of feminine anaphoric gen-
der grams (such as English she/her) in a large world-wide convenience sample of
816 languages based on a strictly procedural definition. The investigation pursues a
radically functional approach where the functional equivalence of the forms under
study is assured by exploring an identical search space in parallel texts (transla-
tions of the New Testament) in all languages of the sample. This is the first large
scale typological study of grammatical gender based on parallel texts, and a large
part of the paper is devoted to methodological aspects. The study shows that gen-
der has a functional core like any other grammatical category, and that it can at
least partly be studied without resort to the notions of noun class, agreement and
system. The results show that a large number of languages possess simple forms of
gender, often representing incipient gender from a grammaticalization perspective.
The paper discusses how simple gender differs from more mature and genealogi-
cally more stable forms of anaphoric gender. Finally the feminine anaphoric gram
type is considered in its wider context, reconciling it to the traditional global ap-
proach focusing on the notions of system, noun class and agreement.