A letter of motivation in the efforts to discover Ururogōri

Hi Lung'afa,

I wanted to engage you in a little (somewhat one-sided) discussion of your goals and assumptions as regards the Logoori group. You have, rather unusually, come to terms with a basic fact about language that most people never understand, regarding the bits and pieces that make up a language. I'd like to encourage you but also to try to guide you away from potential mis-statements. I think the most important thing for me to understand is what your goals and assumptions are.

Based on what I've seen on the group, you seem to be focused on finding that form of the language that is "more original", perhaps representing "The language as spoken by Mulogooli". Of course there is a certain element of conjecture involved in saying what that is since that was many centuries ago. But, your surmises about the alternation between [gu....] and [gw...], [k...] and [ch...] and so on, are basically correct. Now let me comment on how we as linguists come to for hypotheses regarding the earlier state of languages.

The basic idea is to compare words in two contexts. The first is what you are doing, looking at the current language and seeing that a pronunciation [nd] can come from /nt, nd, nl/ and some other things. The other is to compare systematically with related languages, therefore looking at Tiriki, Isukha, Bukusu, Wanga, Saamia and also Gusii, Kikuyu, Luganda, Haya, Chaga, Pare, Shona, and Lingala (there are about 500 languages to compare with). We can reconstruct what earlier stages of languages were like using that method, with a reasonable degree of accuracy. This second "archaeological" approach gives you a much deeper history of a language, going back 2500 years or more – we can also correlate linguistic findings with bones-and-stones archaeology to get some real-time dating (though with many centuries worth of uncertainty).

There is a different non-historical concept of technical linguistics, the "underlying form", which is an assumed representation that speakers of a language have figured out, which gives each morpheme a uniform representation and feeds into the rules of pronunciation of the language. It is, basically, what every speaker of Logoori knows about their language. One important thing to know about this underlying form is that is it not necessarily the same for every person. It is what that individual knows about the language, and not everybody (not even anybody) knows everything about the language. For example, the root "grind" is /sy/ for some people and /sh/ for some people, it just depends on the person. It was historically /sy/. Individuals differ in whether they have /vodogan/ or /vodong'an/ – historically, I can't say that it is clear what the original form was.

You seem to be more interested in the historical aspect of the language, so you're not necessarily interested in the technical linguistic analysis, you seem to be more interested in the earlier historical form. Correct me if that is wrong. The method of extrapolating from patterns in the modern language to the earlier state gives you a pretty good idea about earlier states of the language, but it cannot tie into actual history, for that you have to do a comparison between languages. My finding is that your assumptions (such as reflected in the recent discussions about nlakola versus ndakola and so on) are mostly historically correct, the only question is at what stage of history are they correct? To answer that question, we have to compare languages, to see "What other languages do this; what languages do not do this?". Logoori has four sources for [nda]: original nla, nta, nda, nya. You have future tense -la- which undergoes the rule /l/→[d]/n__. There is the negative -ta- which has the rule /t/→[d]/n__. There are some cases of [d] which are simply from d. Then finally the case of n+y shows up in the past tense e.g. ndaadeeka, ndaakadeeka and so on, compare kwaadeeka, kwaakadeeka.

The change /l/→[d]/n__ is very old in Bantu, and goes back to the original language 2500 or more years ago. All Bantu languages had this and almost all of them kept this rule. The change /t/→[d]/n__ is more recent, but it is part of what defines Luyia, and distinguishes Luyia from Gusii or Luganda. It is a bit of a puzzle for us to understand how this change came into some Bantu languages – it is in the Kamba-Kikuyu group, Chaga and Taita, the Southern Tanzanian language, but not the Swahili group, not Luganda and the other languages around Lake Victoria, not in the more western languages like Luba, Lingala, Kongo. We suspect that is has to do with contact between Nilotic languages like Kalenjin and Bantu, where Nilotic pronunciation patterns filtered into some Bantu languages. The case of past tense ndaa is a fairly specific puzzle – it is only about this past tense prefix -a-, which seems to have come from original ya, and somehow the combination n+y became [nd] (not a general rule, see inyama, nzambuchi). Finally, the case where [nd] is simply nd is because there are still some words that have d distinct from l.

So we can see that /t/→[d]/n__ and /l/→[d]/n__ are actually very old additions to the precursor to Logoori: Mulogooli would not have said "nlakola", "lwa ntakola". he would have said "ndakola", "lwa ndakola". But the "mental precursor", the underlying form, would have been /nlakola, lwa ntakola/. My general point here is that there are different questions, each with its own answer. What was the historical pronunciation? What was the historical underlying form? What is the current pronunciation? What is the current underlying form?

I feel that you are implicitly asking these questions, but it is not always clear to me which question you asking at a given time. But as I say, I suspect that your main interest is in the historical. I think you can probably grasp this abstract concept of an underlying form, where a speaker "distills out" the essence of a morpheme across realizations and arrives at a uniform form, but many people in the group probably do not get that concept. It is similar to the historical analysis, but not the same. The specific thing that distinguishes the discovery of the underlying form and the discovery of the historical earlier form is that the latter requires studying and comparing many Bantu languages, but the underlying form is just based on how the language is pronounced – today, by specific people.

Now by way of suggestion. I think that there is some confusion in the group that related to these different goals and their associated perspectives. This is compounded by the fact that orthography is artificially grafted onto language, therefore if we agree on what the roots and prefixes are and how things are pronounced, we still don't necessarily know how words should be spelled. The segments of ndaakola, nlakola and ntakola are the same in pronunciation, different when you decompose the words. Should writing reflect underlying forms, or should it reflect how people actually speak? If you write ndaakola, nlakola and ntakola, the meaning is clearer from the spelling. But it is harder to figure out what the spelling should be – you have to engage in linguistic analysis and think about whether the future marker has /l/ or /t/, or whether the negative has /l/ or /t/ (or d, or no consonant in the case of the past). This is the nature of language, there is always some difficulty.

Your tendency to insist on clarity of meaning (entailing a more complex analysis of word-composition) is interesting, and I do not think it is wrong, it is just not the traditional way of thinking about language. All prior experience with written Logoori says "writing reflects pronunciation" (various Bible translations and published books), but still one might persuade people to re-think the spelling system if there is some advantage. Your discussions have most productively focused on clarity of meaning, and that is good. There have also been some implications that people are corroding the language or creating confusion by writing the way they actually talk, which does not personally persuade me. The big question that I have is, can participants actually successfully write the way you advocate. Are people easily able to write the present 1st person singular as ndaha, ndaangooya, ntuma, ntaaga, nlaanga, nleeta or will they write ndaha, ndaangooya, nduma, ndaaga, naanga, ndeeta? If the goal is to get people to write in a way that spelling does not change within a morpheme, then first you may want to focus on explaining that goal and then to introduce practical methods of discovering what that form is (how do they know to write d, t or l?).

Anyway, that's enough linguistics for the year. ^D.O.

Facebook emojis of Liking/Loving a Posting.

...

Learn Lulogooli on YouTube