Years ago when starting work with the Koro Ashɛ translation team (whose language is called Ishɛ) the translators asked me about an issue they had. There were too many Hausa words in it. In particular the word ‘amma‘. It’s a common conjunction rather like English ‘but‘ and you do actually hear it frequently in conversation. Many Nigerian languages surrounded by Hausa end up picking it up and appropriating it for their own uses. So that gives us quite a challenge: how should we decide what is too much and when it should go? Some teams end up creating a ‘pure language’ equivalent. But that introduces further issues. Essentially none of us could properly answer the question without looking carefully into how this word ‘amma’ is currently used in Ishɛ storytelling, teaching and conversation.
That’s what we have done for the last 2 years, and the results were rather interesting.
Where did amma show up in the natural texts?
The Hausa ‘conjunction’ amma did appear in our 8 natural texts 14 times… but almost never in the voice of the narrator. Normally it was used within speech, and especially by one younger bilingual speaker.
Clearly it can be used, but it’s not normally needed. My take was that it was used to contradict an expectation or to make an emphatic point, but it never seemed necessary for these purposes, since other Ishɛ means could be used. This gives us evidence and confidence to remove amma or at least use it sparingly in conversation or debate.
But there’s more to it than that
However, simply removing the amma everywhere is only part of the story. It’s helpful to examine why the original translator may have been tempted to put it there in the first place so that we can see if there’s a better strategy that should be used in the future.
Often the Ishɛ translator wrote amma whenever he saw amma in the Hausa translation or but in English translations. These are common translations of the Greek δε or sometimes αλλα. Particularly in Luke-Acts narrative, δε is used to mark:
- Conversational turns (after an articular pronoun)
- New paragraphs (episodes)
- Intense action near the narrative peak
While English has rather complicated rules about clause linkage markers and whether ‘and’ or ‘but’ should be used, Ishɛ generally uses no clause linkage markers at all and certainly doesn’t have the complicated system that English has for connecting clauses.
Ishɛ narrators mark new paragraphs using a pronoun or noun phrase where an index would suffice (over-encoding participants), or by using a head-tail linkage where the new paragraph repeats the final action of the previous paragraph, with a background marker prefix. So in many cases when a new paragraph began with amma, it’s better to replace this with the appropriate pronoun or noun phrase. Often we found both strategies had been used.
Ishɛ narrators mark conversational turns in one of 4 ways depending on whether the conversation is open or closed, and whether the response is expected (resolving move) or a surprise (countering move). This is a distinction that the translators have to deduce from the content of what is said; Greek does not necessarily mark it and English certainly doesn’t. The presence of amma between conversational turns can simply prompt us to be careful to use the correct participant encoding and predicates, but amma itself should never be used by the narrator in this situation.
We haven’t yet studied the characteristics of the narrative peak enough to know what Ishɛ narrators do there, except that we know they do not use amma the same way that Greek uses δε.
What did we change then?
So here’s an example of what’s changed. You can see in Luke 9:59 there are 2 Greek δε which turned in NIV into one but which turned into one Ishɛ Amma. Below the Greek writing in the before-after comparison, the red coloured text is what was removed and the green text is what was added as a replacement. Apart from the amma issue we also found that speech introducers were being misused and our study of how Ishɛ narrators report speech has helped us use the appropriate introducers correctly. You’ll see ∅ markers (which vanish in the final translation) to indicate where confusing redundant words have been replaced with implicit information depending on whether the response is surprising or not. But that’s another story!
The ‘interlineariser’ on the bottom is mostly helpful, but I just noticed ‘nehe’ really is ‘allow’ not ‘shouldn’t’.