This is very cool and brand new: Heindio Uesugi’s Old Avestan Dictionary is available as a free pdf from this page (in Japanese; just click the download button immediately below the image of the book cover). It looks very well done; here’s the beginning of the Preface, which explains its history:
The Old Avestan Dictionary (OAD) is an attempt at a lexicographic synthesis of Old Avestan studies since the Altiranisches Wörterbuch (1904) by Christian Bartholomae (1855-1925) with a particular focus on aiding the elucidation of the Gāthās based on the line of analysis laid down by Helmut Humbach (1921-2017). The dictionary is accompanied by a new annotated translation of the Gāthās to further facilitate the general reader in discerning the sense behind the respective terms and passages when reading, reciting, or studying the original Avestan texts.
This dictionary is the fruit of a seed planted by Karl Hoffmann (1915-1996). In the years following the war, Hoffmann resumed his research on Vedic texts, which had been interrupted by his military service, and in particular on the Vedic injunctive verbal mood, which would later produce his habilitation thesis Der Injunktiv im Veda (1951, but not published until 1967). He had the idea of completing this work by citing the Avestan parallels.
While he worked on the Vedas, he asked his mother to collect all the Avestan verbal forms along with their translations from Bartholomae’s dictionary into a card file so that he could analyze them. In going through the Avestan lexicon based on what he had discovered in Vedic, he noticed that Bartholomae (along with other scholars of his time) seemed to lack a clear idea of the single forms and of the verbal system of the Avesta, and in particular of the Gāthās.
Hoffmann soon realized that the Bartholomaean view of the Avestan verbal system had to be thoroughly revised, a task that seemed to exceed what he had anticipated as part of his research on the Vedic injunctive. He turned to Helmut Humbach to take on the task. In October 1949 or 1950, Hoffmann gave him the stack of cards containing the Avestan verbs and told him: “Schau mal, was du damit machen kannst.”
Subsequently, Professor Humbach went on to publish three translations of the Gāthās with commentaries (and one minor edition (in parentheses below)) that would help shift the paradigm in Gāthic studies […]. The project of creating a dictionary became explicit during the preparation of the aforementioned 1991 translation.
However, various circumstances prevented him from completing this dictionary project. Considering his advanced age, he in turn decided to pass on his stack of materials to me, to whom he entrusted the task of completing the work. Although the skeletal framework of the dictionary was already in place, when looking at the details and through our discussions it quickly became apparent that each entry would require a thorough (re)assessment, and that the project would necessitate considerable time.
Until shortly before his passing, racing against time, we worked relentlessly on the dictionary project, going over my revisions, and spending hours debating each detail. […] After his passing, I was left with a promise to fulfill, the promise to complete the work as best I can, and to publish it in the best way possible.
We used Bartholomae back when I was studying Avestan with Stanley Insler half a century ago, and even then it was way out of date. This is a great resource, which I found at Nelson Goering’s Facebook post. There a commenter said “The website has only 33 views. How did you end up being among first 30 in the world to know about this?” Nelson responded “saw the announcement in the Indo-Iranian Linguistics group here on FB. Only 159 people in that group to start with, apparently.” So you’re getting in on the ground floor.
he asked his mother to collect all the Avestan verbal forms
My mother would never have done that. I now realise that I had a deprived childhood.
The dictionary does indeed look very good.
The “Acknowledgments” section ends
I’m going to have to work harder on my Acknowledgments sections …
While he worked on the Vedas, he asked his mother to collect all the Avestan verbal forms along with their translations
Obviously a native of Old Avesta.
I recall reading somewhere that Parsi priests were careful to chant correctly the traditional liturgy, many words of which have meanings they no longer understood.
If that is so, it might be interesting to hear now from Zoroastrian priests.
From what I think I’ve read, the pronunciation has been preserved – in India at least – much less exactly than that of the Vedas, with lots of substitutions of sounds from the local languages, so historical linguists have to rely on the spellings.
I might as well quote this bit from my 2009 post on Beckwith’s Empires of the Silk Road:
(I might add that the thread is enormously interesting and educational.)
I know nothing about Indo-Aryan, but was interested to discover recently that Pahlavi does the past-tenses-have-ergative-alignment thing, like Middle and Modern Indic languages. Similar origin, in replacing the original finite past tenses with periphrases using past participles, too.
Not sure what to make of that. I suppose it could be pure coincidence; after all, the Maya languages and Sumerian all do something rather similar with ergative alignment, so fancy Sprachbund explanations may well be superfluous.
Georgian, too. Ergative aorist, accusative otherwise.
Past tense ergativity is also Kurmanji Kurdish, and even a couple of Neo-Aramaic varieties – very much an areal feature. Geoffrey Haig and Eleanor Coghill are two of many good names to search.
Presumably the Neo-Aramaic could be plausibly attributed to Iranian influence, in which case perhaps this could be taken as a specifically Iranian-Indic areal phenomenon after all.
“Avestan looks less like an Iranian language than like a phonologically Iranized Indic language”
That’s a bold claim, but not a very likely one. For one thing, you’d have to assume that the “Indic” language in question was something more archaic in various respects than Vedic. Avestan preserves inherited words like aēšəmō “wrath” (cf. Greek οἶμα “attack (of a predator)”), which are not found in Vedic (nor, as far as I know, anywhere else in Indic). On the other hand, this is found in some other Iranic languages: we have, for example, Middle Persian xešm “wrath”. (Martin Kümmel lists this as a possible example of a PIE laryngeal reflected as Persian x-.) There are also morphological archaisms, like the Old Avestan nominative/vocative masculine plural vīspā̊ŋhō “all”, which metrically can’t be replaced by vīspe, the more innovative form found in both Vedic and Young Avestan (see 32.3a: at̰ yūš daēuuā vīspā̊ŋhō “but all you daiwas”).
There are also things like vāstrəm “pasture” that look like innovations in Old Avestan (to an inherited base), and certainly have no Vedic counterparts. This also has cognates in later Iranic. Same for Old Avestan spādā (dual to spāda- “army”), and probably at least a few more in the Old Avestan corpus (which isn’t huge), if you’d keep looking. Lexicon aside, I’m not sure how you’d explain morphological innovations like the genitive singular xratə̄uš vs. Vedic krátvas except by the operation of analogy (but not one found in Vedic for this word!) in a living language. (Young Avestan has xratβō, equivalent to the Vedic form, incidentally.) Old Avestan also shows distinctive Iranic semantic innovations, like vaēnahī “you see”, with *√wain- taking on the general semantic function of “see” that is very common in later Iranic. Vedic has the root, but still in the older sense “look for”. We do also have √spas- = Vedic √paś-, √spaś- still in Avestan, but the point is that the broader semantics of *√wain- already seem to be present.
A slightly different type of problem concerns close cognates that fit differently into the metrical patterns of Old Avestan just because of different sound changes in the two branches. In 45.4d, at̰ hōi dugədā “and his daughter”, disyllabic [dugdɑː] is required metrically, not trisyllabic Vedic duhitā.
I’m not a specialist in Indo-Iranic myself, and I’m sure more (and probably better) examples could be gathered. It’s certainly a claim that immediately rang false, even to a mere dabbler like myself. It sounds like he heard this one comment as a student, which is the sort of thing you say to get people excited, took it very literally, and built some grand claims off the idea without bothering to do even a little double-checking.
I’m also completely puzzled as to how the (rather larger) Young Avestan corpus, which is more conspicuously Iranian (in the normal view, because it’s younger and has had more time to accumulate innovations), is meant to fit into this. I guess you could propose that it’s an artificial attempt to emulate the Iranicized-Indic hymns of Old Avestan? It’s true that Old Avestan casts a long shadow over Young Avestan, but I have a hard time seeing how you’d make this work, if you drilled into the level actual language rather than glib overgeneralizations. (The quotation from Beckwith doesn’t actually give the impression that he’s aware of the difference between Old and Young Avestan. At any rate, his comments about lexical and morphological correspondence, even though they’re slightly exaggerated, only even sort of apply to Old Avestan, so I’ve read his “Avestan” with that adjective in front of it.)
“This allows us to get rid of the separate Indo-Iranian branch entirely”
What a strange thing to say. Even if you did want to explain Old Avestan this way, and ignored Young Avestan entirely, what about Old Persian? I imagine you’d probably even be able to establish Indo-Iranic well enough just on the basis of Middle Iranic, though it would obviously be more complicated. But some innovations, like the merger of *e, a, o > *a (both long and short!), and the relative chronology of palatalization first then vowel merger (e.g. PIE *-kʷe “and (enclitic)” > Vedic ca, Old Persian ca) is also there. I actually wonder how many of the major shared Indo-Iranic phonological or lexical innovations you *couldn’t* recover without Avestan.
Thanks for the bucket of cold water, which is convincing — I confess it got me excited and I was suckered!
I mean, it would be pretty exciting, if true! And you’d hope that anyone making a claim like that in a published book might have something to back it up with.
I’m afraid Beckwith is one of those brilliant scholars who’s too reliant on his own brilliance to bother with tedious fact-checking.
Another striking Iranicism already in Old Avestan struck me this morning: the various forms of gə̄uš.ā “ears”. The word does exist in Vedic, but only in the meaning “noise, din, sound”. Cognates of this still mean “ear” in various middle and modern Iranic languages (e.g. Sogdian ɣwš, Persian guš, Kurmanji guh, Baluchi goš, Pashto ɣwaǧ — the last with the extremely common derivative suffix *-ka- added on).
(I said “various forms”, but actually the word occurs just twice in the Gathas.)
Avestan also has straightforward phonological archaisms compared to Vedic. The cognate of Vedic medhá- is not *maida- (**maēda-, I guess, but the exact outcome would depend on the stress or something), but mazda-. Vedic (and AFAIK Indic generally) has a (completely regular) merger here that Avestan (and AFAIK Iranic generally) lacks, and clusters like *zdh were sufficiently rare that anyone faced with a Vedic e would guess *ai.
(This is part of the Indic elimination of voiced fricatives; it may be blamed on Dravidian – like the reinterpretation of certain affricates as palatal plosives.)
I’m afraid Beckwith is one of those brilliant scholars who’s too reliant on his own brilliance to bother with tedious fact-checking
I thought something was nagging at my memory: yes – he’s that Beckwith.
https://en.wikipedia.org/wiki/Goguryeo_language
So: Yes indeed. I actually own his Koguryo book. It’s a real object lesson in how not to do it. Here’s an absolutely justified takedown:
https://hal.science/hal-00194111/document
I think he basically doesn’t really understand how comparative linguistics works (I was just talking the other day about otherwise competent linguists who just don’t get comparative linguistics, and, worse, don’t realise that they don’t get it.) The fact that he seems to self-identify as a “philologist” may be a bit of a red flag in this regard.
@David Eddyshaw: No, no, this was that Beckwith.
Wikipedia offers dozens of Beckwiths, many of whom sound vaguely crackpottish in some ways but still non-murderous, e.g. https://en.wikipedia.org/wiki/Emma_Beckwith
Now, she sounds like my kind of crackpot.
It does seem that the name may be cursed in some way, though,
Indeed, it is possible that the curse has an even wider scope:
https://en.wikipedia.org/wiki/Vathek
I think he basically doesn’t really understand how comparative linguistics works
Yes, that’s my memory of his work as well.
I’ve read two of his non-linguistic books.Warriors of the Cloisters was an enjoyable, if somewhat overstretched, argument that universities are basically a Central Asian Buddhist invention; Empires of the Silk Road was a (well-informed) polemical rewriting of Eurasian history from the perspective of the steppe nomads being the good guys. (I say “good guys”, but honestly the book’s whole aesthetic felt worryingly fascist-adjacent. Possibly the biggest living fan of the comitatus.)
Beckwith’s views on Indo-European seem to be basically crank stuff, and certainly betray a lack of understanding of even the basic principles of historical linguistics; see the comments on Appendix A here:
https://web.archive.org/web/20131226105908/http://www.ynlc.ca/ynlc/staff/hitch/review_of_Beckwith.pdf
It seems to me that his opinions on historical linguistics are essentially worthless.
He also wrote a comparative syntax thing called Phoronyms; the snippets of this on Google Books look very dicey. (He disses Alexandra Aikhenvald’s opinions on these matters in one footnote, which frankly does not inspire confidence.) It looks like a survey based on questionable analyses of a few of his own pet languages, while lumping together several quite different phenomena under a supposedly unifying label.
I get the impression that he’s very keen to picture himself as an “iconoclast.” And to present himself that way.
I have actually read _Phoronyms_, maybe ten or twelve years ago, and thought it interesting although I did not dig deep enough into surrounding literature to assess with any reliability how plausible/implausible his attempt to draw connections between features of different languages (none IE as best as I can remember) typically treated as distinct by splitter-taxonomists was.
OTOH when I now go back to the online catalog of the New York Public Library (whose copy I read) I cannot seem to find it. Perhaps it has been deaccessioned because They are trying to suppress the Uncomfortable Truths he discovered? They do however seem to have two non-circulating copies of his Koguryo book (a 2004 printing and a 2007 one that may be a revised edition?) in the research collection.
Admittedly, judging a book on the basis of Google snippets is hardly fair play …
I’d have no trouble with him finding typological similarities between genetically unrelated languages*; though he seems to be prone to the schoolboy error of supposing that stray typological similarities can actually demonstrate genetic relatedness. (But, mysteriously, not in the case of Japanese and Korean, or of “Altaic” generally.) That wouldn’t necessarily undermine the actual comparative-syntax side of the book, though.
* Waama and French share (among many other features): contrastive vowel nasalisation; grammatical gender; default noun-before-adjective ordering; basic SVO clause constituent order, but with personal-pronoun direct and indirect objects preceding the verb … one is driven to the conclusion that either Waama is Romance rather than Oti-Volta, or that the ancestors of the speakers of Waama spent a long period in contact with the speakers of Old French prior to migrating to Benin (or Dahomey, as it then was, of course …)
A gushing review of B’s work on the Scythians tells me that he says that “much of the Arthurian legends of Britain come from the Ossetians.” Another, that he describes himself therein as a “practitioner of what he calls ‘classical comparative-historical linguistics’.” And I’m an eminent nuclear physicist …
This emperor has no clothes.
“clusters like *zdh were sufficiently rare that anyone faced with a Vedic e would guess *ai”
Also a good point! Now that you mention it, the whole phonological conversion process seems a bit vaguely conceived. How would the reciters would have known to convert Indic h so consistently to either j or z (han-/jan- “smite”, but hasta-/zasta- “hand”), depending on a word’s etymology? Either there already was a spoken language pretty close to Old Avestan anyway so that the reciters could automatically substitute sounds in recognized cognates (and remarkably consistently so — no Old Saxon tins-effects* here, as far as I know), or reciters were extraordinarily lucky in how they happened to treat h in every instance. Or they were “Iranicizing” a form of Indic so archaic that it was practically Proto-Indo-Iranic.
*(This means “tax, tribute” is clearly adapted from OHG cins, itself from Latin census. The initial affricate of OHG has been replaced by its usual correspondent in OS: [t] for [ts]. An overgeneralization of the pattern from tin/cin “tin”, or timbar/cimbar “construction”, etc.)
Found a pdf of the phoronyms book. It does reference a good bit of other work on classifiers, in fact, contrary to what I too rashly assumed. And he doesn’t here commit any sins of deducing language-relatedness from typological similarities – quite the contrary, if anything.
However, he does indeed lump together e.g. the “pack” element of both “pack of wolves” and “wolfpack” with classifiers of the kind seen in Mandarin – and more besides. It even has a section on “Russian classifiers” … (which turn out to words of the “pack” and “herd” type, in fact. “Terms of venery” …)
It describes English constructions like “pack of wolves” and “cup ot tea” as “pseudopartitive” (not a term B originated himself – he hasn’t simply invented it for the purpose) and attributes their distinctiveness to the syntax rather than the semantics in fine early-Chomskyan style, which makes it all seem more classifier-like. He asserts that these constructions differ formally from partitive genitive construnctions “in all languages examined.”
He should broaden his sample. Kusaal has formally identical possessor and partitive constructions, and the “pack of wolves” type is the same as the partitive/possessive. There is a difference between two distinct constructions of the noun nam “a plurality (of)”, which is used to pluralise any nominal which for whatever reason can’t make its plural by flexion (the standard orthography misleadingly writes nam as a suffix):
daam “beer”
daannam “beers”
na’ab “chief”
na’anam “chiefs”
But this difference is just in line with a much more general principle in Kusaal, that count nouns as generic dependents preceding another noun form compounds with them, whereas mass nouns do not. This has nothing to do with pseudopartitives versus partitives.
Beckwith does in fact go through quite a number of languages, in the process actually showing that treating all these constructions as aspects of one “phoronym” category is a pretty Procrustean process, even within his chosen language subset.
There’s a lot of stuff talking about the semantics of classification in general, but I can’t see that all that really helps his case that these are all manifestations of one single “phoronym” thing. All it shows is that classifier systems (properly so called) have strong semantic correlations. That doesn’t make “pack” a classifier in “pack of wolves.”
There’s something in all this, I think, and something quite interesting, too; but he’s seized on a partial insight and pushed it far beyond what the facts really support. Rather like what he’s done with “some Old Koguryo words rather look as if they are related to Japanese” or “Central Asia was much more civilised and significant culturally than has often been alleged.”
Japanese is described here as a “Koguryo-Japonic” language, incidentally, en passant, as if this were just an accepted fact rather than a fringe theory.
[That should be daamnam “beers”, sorry. In the orthography I use in my grammar, which, unlike the standard, has consistent criteria for word division, “beers” is daam nám and “chiefs” is nà’-nàm, where nà’ is a bound word, rather than a mere word fragment.]
I figured there would be some entertaining choice words by the late Sasha Vovin on something by Beckwith. There are.
Articles subtitled “A response to …” are often entertaining. Vovin, naturally, does not disappoint.
I was recently reading “Why Nations Fail”, and felt in need of a catchy term for that feeling you get when clearly eminent and knowledgeable authors treat an issue that you happen to know much more about than they do, and you find that their presentation is riddled with errors of fact and displays shocking unfamiliarity with basic sources; while, at the same time, they have a reputation within their own broader discipline of being particularly expert in that very subdomain, a reputation which they leverage into a key part of why people unfamiliar with that subdomain find their overall analyses convincing.
African history, in the case of WNF; historical comparative linguistics, in the case of Beckwith.
The editorial postscript to that Vovin piece is also quite something. I particularly liked the casual “du moins dans le monde francophone”.
Yeah, you can’t really expect decorum from les rosbifs.
There was a long and bitter exchange of articles in Y Cymmrodor
https://en.m.wikipedia.org/wiki/Y_Cymmrodor
between Gwenogvryn Evans and John Morris-Jones, who truly loathed each other personally. I recall reading an editorial note which basically amounted to “bring popcorn.”
(The whole thing is available online.)
Is this from the same Arabic word that has been loaned and corrupted (via Persian and Hindi) into the English term “nabob”?
I doubt it.
The stem is na’a-, and the -b part is a noun-class suffix. It’s actually the human-plural suffix (cognate with the ba- part of “Bantu”), which appears as a singular suffix in a handful of nouns referring to older relations and other important people.
Mampruli, which has distinctly more words of ultimate Arabic origin than Kusaal, actually inflects the cognate word regularly: naa “chief”, plural naaba.
The etymon is, however, found only in Western Oti-Volta and Buli/Konni; it seems on first principles to be a plausible candidate to have originated from the unknown mother-tongue of the invaders who actually created the chieftaincy system of the Mossi-Dagomba states in the thirteenth century. They may have been influenced by Islam, though they were probably not actual Muslims (an awful lot of their descendants would have had to have become apostates if they were, which seems unlikely.) So I suppose you could speculate that proto-WOV *nàʔbá was the form of an original loan, and that the -bá was subsequently taken as a noun class suffix by analogy, with Kusaal (like Mooré and Farefare) showing a more conservative stage of the process than Mampruli and Dagbani, in still construing the suffix as singular.
Such a reanalysis is not in itself implausible. In the modern Oti-Volta languages, loan nouns are in fact usually fitted into the class system by analysing the final syllable as a class suffix where possible; as with e.g. Kusaal malif “gun”, plural mali, (cf mɔlif, plural mɔli “kob”), which is ultmately from the Arabic مدفع midfaʿ.
But the language of these invaders is entirely unknown, and although Western Oti-Volta shows a good bit of unique vocabulary compared to the rest of Oti-Volta, none of the relevant words look very Arabic; moreover, they include words like “water”, which do not seem like very likely candidates for borrowing from the language of foreign conquerors.
(When I first started on comparative Oti-Volta work, I overestimated the degree to which WOV is lexically divergent; eventually, I realised that this was an artefact of my knowing WOV languages much better than other Oti-Volta languages, so that I noticed the divergences more. When I started to look at actual numbers, I unsurprisingly found that every major OV branch has a good number of distinctive etyma of its own.)
“Chief” is not reconstructable to proto-Oti-Volta.
“Chief” is reconstructable in proto-Gurma, however: Gulmancema bádō, Moba bád̀, Ncam ūbɔ́tī, Gangam ūbɛ́r, Konkomba ubɔr.This may share its root with Mooré bédré “great, big, fat” and Nawdm bed- “grow, get fat”: so perhaps originally “big man.” Lugal …
Now I think of it, Kusaal na’ab is the only case of this pluralis majestatis use of the -b(a) class suffix which isn’t a word for an older family member. Given the hostility of Kusaasi culture to imposed hierarchy in general, this might not mean much, but I think it’s the same in Mooré and Farefare, and Moorë even uses plural 2nd and 3rd person pronouns as respect singular forms. If “chief” is the sole word in WOV with a human-plural suffix which is not a family relationship term, that might bolster the hypothesis that the final *ba element was not actually a class suffix at all originally, and that proto-WOV *nàʔbá was a loanword.
There seems to be a semantic difficulty, in that even the Supreme Leader/King/Emperor is a na’ab, not just the local chiefs, who could reasonably be thought of as “deputies.” Even the first Mamprussi king himself is called Naa Gbewaa in Mampruli.
On the other hand, I suppose there are plenty of precedents for words originally meaning “deputy” to get above themselves.
Regarding the semantic issue — one might speculate that even a Supreme Leader/King/Emperor might find it politically expedient to be thought of as a deputy or representative appointed by God/tutelary spirit(s)/ancestral spirit(s).
True. Examples do spring to mind …
Though in the Mossi-Dagomba realms, it’s actually the tendaannam, so-called “earth-priests”, but literally “landowners”, who have the religious backing; their approval is necessary for all land transactions, and to be eligible to be a tendaan you need to be able to assert a claim to be a descendant of the first indigenous settler. This is so entrenched as a principle that even the conquest by the Red Hunter’s condottieri and the subsequent creation of the chieftaincies has never altered it. Proverbially, the chief owns the people, but he doesn’t own the land.
(The resulting political tensions are very much a live issue to this day. We had an instance of that not long ago here on LH, regarding the politics of Tamale, in the Dagomba realm.)
Anyhow, “divine kingship” is very much not a thing in these cultures. Good old-fashioned right of conquest is more like it.
(Afrocentrists seem to be keen on divine kingship, as part of their ahistorical fixation on ancient Egypt. I expect the Golden Bough has got mixed into the stew somewhere along the line, too.)
On the other hand, I think there has been more Muslim influence than is often recognised in quite a few West African cultures that are not themselves Muslim, and Islamic notions of rulership do lend themselves quite readily to the ruler-as-deputy idea. And a piece of originally-Islamic political terminology could easily have been adopted without keeping all its specifically Muslim connotations.
I’d be keener on this idea if there were other cases in West Africa of words for “chief, king” that looked as if they might go back to the Arabic نائب; I can’t think of any, but there are literally hundreds of West African languages of which I know nothing at all …
One for the Afrocentrists: Coptic ⲛⲏⲃ, Middle Egyptian 𓎟𓀀 nb “lord.”
Coincidence! Is there anything it can’t do?
There is also a Kusaal verb na’as “honour, respect”; this would do for a cognate, as -s can make causative/factitive derivatives. I’ve recorded it myself as having the right (low) tone to be cognate with na’ab, though both the Toende Kusaal and the Farefare dictionaries disobligingly show cognates with root high tone.
More to the point, there is a derived abstract noun na’am “chieftaincy”, with exact cognates everywhere in Western Oti-Volta, and Buli too.
It’s not just a stem na’a- with the mass/liquid/abstract noun class suffix -m: ringing the changes on noun-class membership as a derivation strategy is very common in Oti-Volta (as in Bantu) but na’am actually has a stem-final -m in all these languages.
Moreover, it has an unexpected tone pattern, with the Buli tones showing an exactly parallel irregular tone relationship to “chief” even though the actual Buli tones differ from WOV, with each word individually having the expected tone correspondences to WOV to be cognates, suggesting that the formation is older than proto-WOV itself. Though there’s a lot of guesswork involved, I think that the protolanguage underlying proto-WOV and proto-Buli-Konni pretty much has to be put at a time depth of well over a millennium, and probably nearer two (or more), so certainly antedating the formation of the Mossi-Dagomba kingdoms.
So I think the WOV “chief” word is probably “honourable one” rather than “deputy” after all, and the *ba part really is just the human-plural suffix used as singular honorifically. Simple answers are often the best …
Thanks for the interesting question!
Simple answers are often the best …
Often they’re also the hardest to find. But let’s not get too real about this. You take what works for you at the moment, and repent later.
Is etymological nativization an option? (It would complicate everything most beautifully.)
Yes, possibly.
The tone thing in Buli can possibly be finessed: the tone system of Buli differs from that of WOV, but matching up the correspondences is almost trivially easy from a comparative point of view, and the systems may well have been even more similar – even superficially – relatively recently.
The fact that the cognates of na’as “honour” outside Agolle Kusaal have the “wrong” tones to be related to na’ab may mean that this is just a chance resemblance, too. And I may have got its tones wrong: I’d back my own notes over the Toende Kusaal dictionary, but the Farefare dictionary is pretty reliable on tone.
The tonal peculiarity of na’am could be something of a two-edged sword, too. This peculiarity is specifically that it is the only case in my Kusaal materials of a word with Tone Pattern A (mid tones throughout, except for combining forms, which have all low tones) being derived from a noun* with Tone Pattern L (initial low tone.) If the word was a loan which has been reanalysed in terms of WOV morphophonemics, an imperfect reanalysis might underlie the apparent tonal inconsistency.
To complicate matters even further, it’s not out of the question that the proto-WOV stem was *naʔ-b-m- rather than naʔ-m-. Underlying *bm may get broken up by the insertion of an epenthetic vowel when it arises in flexion, but in derivation it appears always to have assimilated to *mm or just *m. I can’t see any way that this would help with the odd tone developments, though.
* There are a lot of Tone Pattern A nouns derived from low-initial verbs, but verbs do not distinguish between the A and L patterns at all, and it is not clear whether the verbal A/L pattern is the equivalent of noun Pattern A or of Pattern L, or represents a historical merger of the two. Synchronically it resembles the noun Pattern A most closely (I call it Pattern A in my grammar) but this may be misleading historically.
See now this Log post, with some very useful examples (images) of Humbach’s work.