Using narrative data from u̱t‑Maꞌin [gel], a Kainji language of northwestern Nigeria, I first present two main strategies for speech reporting: the prolific SAY‑construction and the SPEECH VP + ɘ̄zɘ̄ SAYING‑construction. Second, I discuss uses of the SAYING‑construction that have extended beyond the reported speech domain.
The SAY‑construction uses a form of the verb zɘ̄, which can occur with the full range of verbal suffixes. The reported speaker (RS) occurs in the preverbal subject position. An optional reported listener (RL) occurs in the postverbal primary object (O₁) position, typically used for recipients in ditransitive clauses. The speech report (Report) occurs after any optional reported listener. This is parallel to the location of the secondary object (O₂) , typically used for theme arguments in ditransitive clauses. Prosodically, certain speakers pause or adjust pitch immediately before the Report. Pause and pitch adjustment are not features of other ditransitive clauses which highlights questions of constituency.
See Example 1 below
In the SAYING‑construction, the form ɘ̄zɘ̄, which is transparently related to the verb zɘ̄ ‘say’ and yet never takes verbal inflection, occurs between the optional RL and the Report, as in (2). The main verb is often a speech related verb like ‘tell’, ‘ask’, or ‘shout’. Any prosodic pause or shift in pitch follows ɘ̄zɘ̄, and immediately precedes the Report as in the SAY‑construction.
See Example 2 below
In the final section, I present a broad range of uses of the complementizer ɘ̄zɘ̄ and the various monoclausal and biclausal constructions in which it is found with predicates of perception and cognition like ‘hear’, ‘know’, or ‘think’, as well as clauses expressing result, reason, hypothetical condition, and definition. Evidence for the grammaticalization of ɘ̄zɘ̄ comes from change in meaning, function, and adjustments to prosodic boundaries.
Note: Originally scheduled for in person presentations and discussion but the event was held online due to Covid-19.