Subject prefixes in Bantu languages

Question

Background

I'll be up front: this question arises from trying to crack a Lingala song. There is this polyphonic Lingala religious hymn, Salelaka Mokonzi, which is very well-known among members of this Catholic religious movement I belong to. Recently, I tried cracking it, that is, armed myself with several versions of the lyrics, a couple of translations, a couple of dictionaries, and a lot of patience, and tried to analyse the lyrics word by word to figure out what they meant (I usually distrust ready-made translations, especially when they and the lyrics oscillate between different sources as in this case). I had some successes and some problems.

One problem I had was subject prefixes, and it arises from the following cognitive dissonance:

• On one side, I know Zulu and Swahili, both Bantu like Lingala, use a system of subject prefixes which mandates agreement with the class of the subject noun, plus prefixes for first and second person; a couple of passages in the song suggest this could be the case in Lingala as well, that is, there seem to be a couple verb forms that use noun class prefixes as subject prefixes, and one isn't analyzable otherwise and the other seems to mean something apparently extraneous to the sentence of the song it's found in;
• However, this grammar seems to give full conjugations, and only has three third person prefixes: a (s)he, an it, and a they; also, this dictionary analyses forms like aye ((s)he came), but loye isn't analysed, whereas I'd see it as similar to aye but with the lo- prefix.

So…

Question

Are there Bantu languages where, unlike Zulu and Swahili, the subject prefix system has only three third person pronouns, two singular and one plural, or anyway far fewer prefixes than noun classes? In particular, did this system collapse happen in Lingala, or does Lingala use noun class prefixes as subject prefixes?

PS

Naturally I strongly encourage any Lingala speaker who sees this to either answer the Quora question about the song or open a chat room to discuss it. Lingala seems to be rather unpopular on Quora, with only 12 people having answered questions in it, exactly one each, at the time of posting my question, and then a couple more answers were posted to questions of mine.

UPDATE

Two things: one, I'll clarify that «exactly one each» in the P.S. meant each user had posted one answer, not that each user had answered a distinct question; there are at least 12 questions though, excluding mine; two, I created the chat room.

Answer 1

Excellent question! It just so happens that this is exactly what my group is researching at the moment!

Are there Bantu languages where, unlike Zulu and Swahili, the subject prefix system has only three third person pronouns, two singular and one plural, or anyway far fewer prefixes than noun classes?

If by "pronouns" you mean "subject agreement markers", there are indeed!

In particular, did this system collapse happen in Lingala, or does Lingala use noun class prefixes as subject prefixes?

Depends on the dialect! And this is probably the source of your confusion. It's not something that's documented nearly as well as it should be.

Short version: some dialects show this system collapse down to three classes, while others don't.

Long version:

"Literary" or "High" or "Catholic" Lingála

For most classes, the subject marking on the verb resembles the noun class prefix. But there are a few where the two diverge. The same happens in Swahili sometimes: ndege inaenda, ndege zinaenda.

Cl N   VS  VO
 0  ∅-  a- mo-
 R         mi-
1S     na-  N-
2S      o- ko-
 1 mo-  a- mo-
1P     to- lo-
2P     lo- bo-
 2 ba- ba- ba-
 3 mo- mo- mo-
 4 mi- mi- mi-
 5 li- li- li-
 6 ma- ma- ma-
 7  e-  e-  e-
 8 bi- bi- bi-
 9  N-  e-  e-
10  N-  i-  i-
11 lo- lo- lo-
14 bo- bo- bo-
15 ko- ko- ko-

This is the prestige dialect used for works of literature and such. You can recognize it by the wide variety of class prefixes and by the use of "agreeing attributives" (moto wa, bato ba, etc, like you see in Swahili).

Kinshasa Lingála

Cl N   VS
 0  ∅-  a-
1S     na-
2S      o-
 1 mo-  a-
1P     to-
2P     lo-
 2 ba- ba-
 3 mo-  e-
 4 mi- ba-
 5 li-  e-
 6 ma- ba-
 7  e-  e-
 8 bi- ba-
 9  N-  e-
10  N- ba-
11 lo-  e-
14 bo-  e-
15 ko-  e-

As you can see, this dialect has reduced the noun class system drastically! Nouns still tend to show their class prefixes, but third-person agreement has been reduced to three classes: "animate singular" (originally class 1), "inanimate singular" (originally class 7), and "plural" (originally class 2). In some places the distinction is still eroding further, with e- taking over some of the inanimate plurals, but this doesn't seem to be universal yet.

This is probably the most famous dialect, and it's generally associated with the military. A lot of popular music is written in this dialect, and as you might expect, it's the main dialect spoken in Kinshasa. The easiest way to recognize it is the simplified verb prefixes, the use of na for the attributive regardless of class (moto na, bato na, etc), and the loss of prenasalized voiceless consonants.

"Spoken" or "Low" or "Protestant" Lingála

This is sometimes called "the third dialect" (separate from Literary and Kinshasa), but it's not so much a single dialect as it is a pretty huge continuum. It tends to fall somewhere in between the two extremes, with some classes getting simplified and some remaining.

A big part of our current work involves examining and categorizing all the different types of "Spoken Lingála" out there in the world, and I'll probably update this with a link once we publish. But suffice to say, there's a whole lot of variation.

Sources

Much of this answer is based on Bokamba, Eyamba G. 2012. A Polylectal Grammar of Lingála and Its Theoretical Implications. In Selected Proceedings of the 42nd Annual Conference on African Linguistics, ed. Michael R. Marlo et al., 291-307. Somerville, MA: Cascadilla Proceedings Project. www.lingref.com, document #2778.

The rest of it is based on a corpus collected by the same author, as yet unpublished. The corpus analysis is my own work, as yet unpublished.

Notes

• "N" is nouns, "VS" is verb subjects, "VO" is verb objects
• "Class 0" is also known as "class 1A" or "class 25"; it's used mostly for kinship terms and (according to some analyses) personal names, and acts just like class 1 except with a zero prefix in the nouns. You'll also recognize this one from Swahili: mama, baba, etc.
• "R" is a special object marker that indicates that the object is the same as the subject; it can be used with single subjects, unlike the reciprocal -an-
• "1S, 2S, 1P, 2P" are first and second person singular and plural
• The prefix "N-" means "a nasal without a defined place of articulation" (same as Swahili m-bwa, n-dege, ny-oko)
• Classes 12 and 13 no longer exist in any dialect I know of, nor do any of the locative classes

Answer 2

For Lingala grammars, you might want to read this paper, which contains references to grammars (not well-meaninged class projects labeled "grammar") and some discussion of the dialect problem. Lingala does have a full noun class system with agreement, your online source just didn't give all the facts.

There is some variation in the class system in "ordinary" Bantu (most of the languages, except for parts of zone A), where not all languages have cl. 20 gʊ and most don't have cl. 19 pi or the classes above 20. The classes 1-11 plus 14 are the most resilient, and even then there are minor reducing tendencies such as merger of cl. 11 lʊ and cl. 14 vʊ. There is also a dialectal merger of cl, 9, 10 in Swahili because you can't tell from the noun for what class ratili is in (again, phonological merger of the distinctive prefixes). Kituba is a good example of a reduced class system, but it still has most of the original class distinctions. There are frequent additional reduction patterns where semantic properties override formal agreement (all humans inducing agreement in cl. 1-2 even when such nouns can appear in other classes). I don't know of any Bantu language that reduces noun classes or agreement patterns to just two or three 3rd persons.