What languages have extraction markers?

Question

I'm looking for languages that have extraction markers or wh-agreement markers like those cited in Chung (1994) in Chamorro below:

The example shows wh-agreement morphology on the verb when the questioned constituent is extracted. The constituent moves successive-cyclically and triggers wh-agreement on each verb of each clause it passes through.

I don't have an example of a general extraction marker or displacement marker (caused by focus movement, topicalisation, or any kind of movement in general) but, if I'm not mistaken, some Bantu languages have these. If you know any languages that have either or both of them, feel free to comment or answer with a list.

Answer 1

As I understand it, you're asking about a specific extraction marker that occurs always and only when a nominal is extracted, and in clause chains it appears in each clause. As far as I know, this does not exist in Bantu.

There are languages which show subject and object agreement reflecting the pattern of movement, for example in Kipsigis (Jake & Odden 1979 "Raising in Kipsigis") when a 1st or 2nd person is raised out of an object complement it controls object agreement and when raised out of an subject complement it controls subject agreement, and each intervening clause shows the appropriate agreement. Bantu languages do likewise with subject-to-subject raising. But in both cases, this involves marking that is not specific to extractions.

What many Bantu languages do have is a special verb form employed in relative clauses, which can obviously be related to extraction. However, this is probably a subset of extractions. The same verb forms are widely used in subordinate clauses like "having build a house, the farmer took a nap", and also "when he took a nap, the dog barked". A distinction is often made for these languages between main clause affirmative forms, versus all else (subordinate, or negative). It's not obvious that those clauses all involve extraction, but perhaps that is a contemporary analysis.

It is also generally the case in Bantu that an immediately pre-verbal wh-NP causes the verb to shift to the aforementioned relative clause form, thus *"Which child cried?", but okay "What child rel-cried?". Wh-extractions in Tharaka are reported by Muriungi ("WH-questions in Kitharaka") to exhibit a non-extraction-sensitive allomorphy in the present tense marker, which is -ku- ordinarily, but -ri- in case the clause contains an in situ WH object (so, if the WH object is extracted, -ku- is used). Selection of -ri- is not actually governed by extraction though it superficially correlates with it (in a strange way): it correlates with in-situ focus. You also use -ri- in the present on object-complement verbs (no extraction).