Victor Manfredi — writing
The selected texts, at various degrees of publication, are posted in .pdf format unless otherwise noted. Critiques are invited, especially data corrections, in addition to the ones already posted here in the supplementary comments.
Philological boilerplate plus Unicode crib sheetÀjàyí Crowther's orthographic subdot, adopted in 1851 to signify the systematic phonetic feature [+ narrow pharynx] — better known after Chomsky & Halle (1968) as [‑ATR] — is an essential piece of tech for tens of millions of 9ja literates. In principle, it can be digitally rendered as the Unicode glyph U+0329 ("combining vertical line below") = HTML "& # 8 0 9 ;" (without the wordspaces), or alternatively as U+0323 ("combining dot below") = HTML "& # 8 0 3 ;" (without the wordspaces), but in practice this patch leaves a lot to be desired. For one thing, not all browsers display the subdot, so some readers of this page will see nothing under the first two vowels of É̩hu̩gbò. For another, the kerning (horizontal alignment) of this kludge is chaotic, especially in book‑quality faces in sizes above 8 pt. Even in certain typefaces wishfully called "Unicode compliant", legibility is woeful as can be verified by inspection of the enclosed test page, which was prepared at random. (Real published examples can be much worse.) Polite inquiries to The Unicode Consortium about this unsatisfactory status quo were replied (email of M.D., 25 September 2008, available on request) with a smooth blend of bureaucratic indifference ("The disclaimer you mentionned [sic] in your email *is* absolute, and these combinations will not be encoded, so you should not waste your time making a proposal for them") and lazy palming off of the problem to apocalyptic and politically radioactive SIL/WBT missionaries, who are assumed to 'own' subdotted roman scripts, maybe based on an implied metaphysical 'ownership' of the speakers of the respective languages ("Those look fine with a font like Doulos SIL that can handle them"). (And anyone who thinks that pentacostal interventions are harmless has not observed the condition of civil society in evangelized West Africa of recent decades.) This is a nice illustration of how priorities are set, who benefits and so on, under neoliberalism. The ruling elite in Abuja could choose to take a benevolent interest in the problem, but the opposite has been the case: in late 2006 while preoccupied with funding the PDP's 2007 auto‑succession campaign, General "Kánkpé" went so far as to cancel the parastatal autonomy of the National Institute for Nigerian Languages in a fit of pique at its second executive director, who did not understand the plain meaning of the expression o̩mo̩lúwàbí percentage (cf. Guardian [Lagos], 25 July 2008 and emails of A.A. and O.E., 19 January and 3 September 2007).[NOTE: Technically savvier discussion of these points, plus more doctrinal pushback from Unicode defenders, appeared on Language Log shortly after my own hapless collision with the planet's typographic authorities. Maybe we need to update Max Weinreich's immortal wisecrack about glossopolitics to something like "A language is a dialect with precomposed diacritics."]Thanks to the clout of Vietnamese and romanized Indic languages in corporate software's higher echelons, an integrated (precomposed), therefore correctly aligned and (we can hope) unambiguously searchable subdot has been made available for upper and lower roman vowels and s. This resource accidentally represents real progress for any users of 9ja orthographies who choose not to mark tone, or even for those who are prepared to play the lottery of combining tonemark alignment. Here are the Unicode and HTML (remove wordspaces) bit addresses for the subdotted letters most commonly used in 9ja orthographies, and for the nonintegrated but nonspacing (i.e. possibly "combining") acute and grave accents as well as the less important macron, plus a more fully descriptive link for each:U + 1 E A 1 = & # 7 8 4 1 ; = LATIN SMALL LETTER A WITH DOT BELOW
U + 1 E B 9 = & # 7 8 6 5 ; = LATIN SMALL LETTER E WITH DOT BELOW
U + 1 E C B = & # 7 8 8 3 ; = LATIN SMALL LETTER I WITH DOT BELOW
U + 1 E C D = & # 7 8 8 5 ; = LATIN SMALL LETTER O WITH DOT BELOW
U + 1 E E 5 = & # 7 9 0 9 ; = LATIN SMALL LETTER U WITH DOT BELOW
U + 1 E 6 3 = & # 7 7 7 9 ; = LATIN SMALL LETTER S WITH DOT BELOW
U + 1 E A 0 = & # 7 8 4 0 ; = LATIN CAPITAL LETTER A WITH DOT BELOW
U + 1 E B 8 = & # 7 8 6 4 ; = LATIN CAPITAL LETTER E WITH DOT BELOW
U + 1 E C A = & # 7 8 8 2 ; = LATIN CAPITAL LETTER I WITH DOT BELOW
U + 1 E C C = & # 7 8 8 4 ; = LATIN CAPITAL LETTER O WITH DOT BELOW
U + 1 E E 4 = & # 7 9 0 8 ; = LATIN CAPITAL LETTER U WITH DOT BELOW
U + 1 E 6 2 = & # 7 7 7 8 ; = LATIN CAPITAL LETTER S WITH DOT BELOW
U + 0 3 0 1 = & # 7 6 9 ; = COMBINING ACUTE ACCENT
U + 0 3 0 0 = & # 7 6 8 ; = COMBINING GRAVE ACCENT
U + 0 3 0 4 = & # 7 7 2 ; = COMBINING MACRON
Unfortunately the "combining" superscript accents are just as blunt an instrument as the "combining" subdot, so it would still be more reasonable for the wise bosses of Unicode to kindly provide precomposed subdots for all of the five roman vowels upper and lower case EVEN WHEN THESE VOWELS ALSO BEAR PRECOMPOSED ACUTE AND GRAVE ACCENTS. But until that glory day arrives, southern 9ja literates will be forced to choose their poison: either to tweak the alignment of the combining subdot, or that of the acute and grave accents. At least, thank goddisses, there's no need to drink both poisons at the same time; instead, the prudent 9ja typist will ensure that either the subdot or the accent is used in the composed or integrated (not the "combining") form. A fortiori, the "combining" superscript accents should emphatically not be used with plain (non‑subdotted) vowels or tonebearing nasals for which composed/integrated tonemarks exist (see complete list below). Some text editors are programmed to automatically substitute the respective composed/integrated character for the sequence of letter plus combining diacritic, but it would be naive to trust this to happen on any given day. Apple's tablet‑like OS10.8 (enervatingly named after yet another cat species — can Civettictis civetta be next?) is the worst of both worlds: all the illegibility of combining characters, wrapped in a pseudo‑composed display format so that diacritics can't be tweaked.U + 0 0 E 1 = & # 2 2 5 ; = LATIN SMALL LETTER A WITH ACUTE
U + 0 0 E 0 = & # 2 2 4 ; = LATIN SMALL LETTER A WITH GRAVE
U + 0 0 E 9 = & # 2 3 3 ; = LATIN SMALL LETTER E WITH ACUTE
U + 0 0 E 8 = & # 2 3 2 ; = LATIN SMALL LETTER E WITH GRAVE
U + 0 0 E D = & # 2 3 7 ; = LATIN SMALL LETTER I WITH ACUTE
U + 0 0 E C = & # 2 3 6 ; = LATIN SMALL LETTER I WITH GRAVE
U + 0 0 F 3 = & # 2 4 3 ; = LATIN SMALL LETTER O WITH ACUTE
U + 0 0 F 2 = & # 2 4 2 ; = LATIN SMALL LETTER O WITH GRAVE
U + 0 0 F A = & # 2 5 0 ; = LATIN SMALL LETTER U WITH ACUTE
U + 0 0 F 9 = & # 2 4 9 ; = LATIN SMALL LETTER U WITH GRAVE
U + 0 0 C 1 = & # 1 9 3 ; = LATIN CAPITAL LETTER A WITH ACUTE
U + 0 0 C 0 = & # 1 9 2 ; = LATIN CAPITAL LETTER A WITH GRAVE
U + 0 0 C 9 = & # 2 0 1 ; = LATIN CAPITAL LETTER E WITH ACUTE
U + 0 0 C 8 = & # 2 0 0 ; = LATIN CAPITAL LETTER E WITH GRAVE
U + 0 0 C D = & # 2 0 5 ; = LATIN CAPITAL LETTER I WITH ACUTE
U + 0 0 C C = & # 2 0 4 ; = LATIN CAPITAL LETTER I WITH GRAVE
U + 0 0 D 3 = & # 2 1 1 ; = LATIN CAPITAL LETTER O WITH ACUTE
U + 0 0 D 2 = & # 2 1 0 ; = LATIN CAPITAL LETTER O WITH GRAVE
U + 0 0 D A = & # 2 1 8 ; = LATIN CAPITAL LETTER U WITH ACUTE
U + 0 0 D 9 = & # 2 1 7 ; = LATIN CAPITAL LETTER U WITH GRAVE
U + 1 E 3 F = & # 7 7 4 3 ; = LATIN SMALL LETTER M WITH ACUTE
U + 0 1 4 4 = & # 3 2 4 ; = LATIN SMALL LETTER N WITH ACUTE
U + 0 1 F 9 = & # 5 0 5 ; = LATIN SMALL LETTER N WITH GRAVE
U + 0 E 3 E = & # 7 7 4 2 ; = LATIN CAPITAL LETTER M WITH ACUTE
U + 0 1 4 3 = & # 3 2 3 ; = LATIN CAPITAL LETTER N WITH ACUTE
U + 0 1 F 8 = & # 5 0 4 ; = LATIN CAPITAL LETTER N WITH GRAVEBy whatever means tones manage to be typed, the tonemarks given here for any of the Benue‑Kwa (BK) languages follow a dual convention, consistent with best practice known to me although not necessarily matching current popular or official usage — fully explicit tonemarking being rare outside of Yorùbá. Throughout BK, [ ́ ] = high, [ ̀ ] = low, but marking differs in a principled way between the BK1 and BK2 subgroups correlated to prosodic type as discussed in Manfredi (2009a). For clarity, the pitch labels H, M, L and ! (downstep juncture) are added parenthetically. In BK2 (comprising the Gbè, Yorùbá, Nupe and Ìdọmà macro‑clusters) with 3 lexical tones, marking economy is paradigmatic i.e. syllable‑by‑syllable, thus no mark = mid, but in BK1 (the historic remnant including Àkan, Ẹ̀dó, Ìgbo, Tiv, "Bantu" &c.) with only 2 lexical tones, marking economy is syntagmatic, thus no mark = same as preceding and a sequence of two high marks = downstep starting on the second (Welmers & Welmers 1968, cf. Christaller 1875), e.g. Ẹ̀dó Ólokún (HH!H) '[tutelary supernatural, from Yorùbá]'. Furthermore, following Bám̄gbóṣé (1966) and Ámayo (1976), downstep preceding nonhigh is marked by a word‑internal period, e.g. Yorùbá Oló.kun (MH!M) 'possessor/epitomē/personfication of òkun (LM) [the ocean]' vs. Olókun (MHM) 'possessor/epitomē/personfication of okun (MM) [energy]' and the same expedient conveniently generalizes to a non‑spreading juncture between high and a following low, e.g. Yorùbá oló.dù (MH!L) 'possessor/epitomē/personfication of an òdù (LL) [clay cauldron]' vs. olódù (MHL) 'possessor/epitomē/personfication of an odù (ML) [8‑bit oracle sign]', cf. also Ẹ̀dó nó.dè̩ (H!L) 'yesterday'.
1991My dissertation, chaired by S.J. Tambiah, with Ken Hale as the Harvard‑external committee member, and Susumu Kuno and Sally Falk Moore as the other Harvard members. Compact edition [400 pp. 8.5 x 11 inches].
No photos were included, but a few representative ones from the fieldwork years (1976‑77, 1980‑84) are posted on the research page — where can also be found links to the 90 minutes of audio originally distributed together with the thesis as a cassette tape.
Four sections of this work were published in essentially identical form (not separately posted on this page, but offprints are available on request).
Pp. 30‑33 together with 46‑61 had previously appeared as "Igboid"
The Niger‑Congo Languages, edited by J. Bendor‑Samuel, 337‑58. University Press of America, Lanham, Maryland, 1989. ISBN 0819173762.
Pp. 61‑93 appeared as "Spreading and downstep; prosodic government in tone languages"
The Phonology of Tone; the Representation of Tonal Register, edited by H. v.d. Hulst & K. Snider, 133‑84. De Gruyter, Berlin, 1993. ISBN 3110136058.
Pp. 188‑220 appeared as "Verb focus in the typology of Kwa/Kru and Haitian"
Focus & Grammatical Relations in Creole Languages, edited by F. Byrne & D. Winford, 3‑51. Benjamins, Amsterdam, 1993. ISBN 1556191669.
Pp. 220‑41 had previously appeared as "Antilogophoricity as domain extension in Ìgbo and Yorùbá"
Niger‑Congo Syntax & Semantics 1, 97‑113. Boston University African Studies Center, 1987.JolanPanNigerian, the 20 year old PostScript face created in order to accommodate Nigerian orthographies (see p. 396 for details), looks grizzly on the screen but prints out smoother. (Not only people get grizzled with age — a consoling thought!)
As to content, an apologia is called for. Much of the document de travail has been superceded in recent years, especially due to two steps which I took in response to the intractable mysteries faced by more standard kinds of analysis: (i) a shift from the autosegmental‑metrical hybridity of Chapter 1, to toneme‑free metrics (also here); and (ii) a move from the projected lexical conceptual structures of Chapter 2, to compositional lexical syntax (also here and here). In the interim I've also (iii) backed off of the strong WYSIWYG morphology conjecture of Chapter 3, to settle for an indispensable minimum of overt structural cueing, consistent with parametric, derivational phases (here) and irreducible to Distributed Morphology — a style of hyper‑ or meta‑phonology cum recapitulation of syntactic "debris" (Halle & Calabrese p.c.). All these changes were prefigured in the 1991 text, but not so for (iv) "scopophobia", a principle of syntax‑semantics mapping — Chapter 3 scrupulously avoided QR but offered nothing in its place apart from appeals to intelligent pragmatics. My current interest is (v) to combine the foregoing claims in order to treat focus phenomena without cartographic/criterial feature checking (here and here), and in order to capture subject/object and argument/adjunct asymmetry in wh‑dependencies.
My best luck came at the start of the journey, to compare Ìgbo and Yorùbá while learning about each from eminent southern Nigerian scholars and traditional verbal artists. That happenstance was somehow inevitable for an alárìnso̩ (the efficient translation into Yorùbá, applied to me by my ASUU sister Dr. Yétúndé Olúwafisàn, of the concept of 'peripatetic linguist with theatrical and mendicant airs') working in the then‑thriving federal university system at a time of cheap and safe public transportation. Under such conditions, the comparative window opened wide enough to show one nontrivial difference in the relation between sound and meaning. The agenda of topics could have been shorter, though my committee kindly refrained from pointing out this fact. My revered colleague Dr. Frannie Oyèbádé gave me friendly grief about the book's polymathic title, but from my outsider's standpoint I can't imagine engaging with these languages to the needed extent without facing what Chomsky calls Orwell's Problem. Maybe a grammarian can untangle puzzles of ideology and demography in the cognitive science vein (cf. Wallerstein quoted here).
My worst typo is the omission of Kay Williamson's name from the Acknowledgements (p. 13), although her generous mentorship is writ throughout. She was too polite to mention my mistake; when a third party pointed it out in 1994, she gracefully accepted that no vendetta was behind the lapse.Tone correction: the language name conventionally spelled "Igala" probably does not have the LHL tones printed on pp. 24 (fn. 15) and 31, following Banfield (1914, 178) and Armstrong (1965, 78). It appears with MHL in Ìlò̩rí (2009), but two independent citations give initial H (Étù & Mìáchî 1988‑92, Ọmachonu 2012). Initial H is consistent with reported prohibition of initial M in "noun class words" (Ọmachonu 2012, 22), presumably including the ethnonym with its proclitic vowel, and loanword change HHL>MHL is plausible given that Yorùbá prohibits initial H in the corresponding slot (Ward 1952, 37). The quality of the medial vowel needs more investigation.Armstrong, R. [1965]. Comparative wordlists of two dialects of Yoruba with Igala. Journal of West African Languages 2, 51‑78.
Banfield, A. [1914]. Dictionary of the Nupe Language. Niger Press, Shonga.
Étù, Y. & T. Mìáchî. [1988‑92]. Ígílâ ékọ́chẹ 1‑4. Heinemann, Ìbàdàn. OCLC 38436098. [Not personally consulted.]
Ìlò̩rí, J. [2009]. Noun‑plural formation in Igálà. Current Perspectives in Phonosyntax & Dialectology, edited by G. Adika & al., 1‑15. Department of Gur Gonja, University of Education, Winneba, Ghana.
Ọmachonu, G. [2012]. Ígálà language studies and development. Slides from 12th Igala Education Summit, Kogi State University, Anyigba, 28-29 December.
Ward, I. [1952]. Introduction to the Yorùbá Language. Heffer, Cambridge.
1992MIT Working Papers in Linguistics 17 (= Proceedings of the Kwa Comparative Syntax Workshop), edited by C. Collins & V. Manfredi, 205‑17.IRCS Report 92‑37, edited by M. Liberman & C. Maclemore, 103‑15. Institute for Research in Cognitive Science, University of Pennsylvania, Philadelphia.The complete list of 81 examples transcribed in the Appendix (pp. 110‑12) can be heard in the same order here. Here's a foto of the speaker, Julius Ògbú, and me.
Significant correction: Ex. (18a) as cited on p. 109 is observationally correct in some prominent Ìgbo varieties — e.g. it occurs as such in Ḿbàisén and adjacent areas — but not in Standard Ìgbo, moreover the form as given on p. 109 is seriously misleading with respect to considerations of clause‑typing. In Standard Ìgbo, the verb root plus suffix in a subject question of this inflectional type bears L, and not downstepped H:
Ònyé hù̩‑ru̩ Ézè? 'Who saw Ézè?' = [LH LL HL].
But Standard Ìgbo does maintain the downstepped H in subject relative clauses like (18a) = [H!H !HH H!H]. Accordingly, one might well claim that T‑to‑C (or Infl‑to‑C) does occur in (18a), just as indicated in the paper, but there is no prosodic evidence for parallel treatment of (18b), contra the little upward arrow to the left of ‑rV in the tree structure in (18b). Incidentally, the identification of Ìgbo finite ‑rV as a morphological instantiation of the Infl or Tense node of the clausal Middle Field, assumed in 1992 the paper, must be abandoned for independent reasons of temporal and aspectual quantification. Specifically, the longstanding controversy in the Ìgbo literature regarding the temporal content of ‑rV inflection dissolves, once ‑rV is recognized to have the distribution of an argument‑type clitic.Otherwise, the general point of the paper remains unaffected: so‑called "upstep" is epiphenomenal, and is not an upside‑down counterpart of "downstep" as pretended in taxonomic ("tagmemic") work by E. Pike and maintained by many Africanists with the help of enriched autosegmental‑metrical notation. Rather, "upstep" is no more than the phonetic realization of antidownstep (downstep reset), a phenomenon itself constrained by syntactic phrasing. Several elementary and general observations support this conclusion, quite apart from the particular analysis of Ágbò̩ presented in my paper. First, there is no antidownstep without a preceding downstep in the relevant prosodic domain. Secondly — and this point was admitted anecdotally during the roundtable discussion at the 1992 UPenn workshop by the two invited SIL Mayanists — Mayan languages do not show upstep cumulation, contra E. Pike's published descriptions that started the whole upstep goosechase. Naturally, upstep retains support as a strictly phonetic category, my only purpose here being to show that it plays no role in the statement of linguistically significant generalizations. In sum, phonological use of the term upstep is, on current knowledge, nothing more than a hypostasis or in other words a mystification. SIL/Wycliffe may possess excellent soteriological reasons to look "upward" as much as "downward", but natural languages including Ìgbo and apparently also the Mayan family are quite innocent of this skyward temptation, and should not be blamed for it.
[Update 6 December 2012] Another telling indication that syntax is not 'different' (in the sense of Bromberger & Halle 1989) is that not only prosodic footing, but also syntactic agreement, has now been enthusiastically offered as grist for an upward‑looking parameter of structural variation across natural languages (Baker 2008). Automatic sacrifice of restrictiveness is always expedient in the short term, but always too a bad idea for constructing testable theory‑space (Martin & Osherson 1998) and at least as far as Niger‑Congo languages are concerned, always a recipe for exoticism.
Baker, M. [2008]. The Syntax of Agreement & Concord. Cambridge University Press.
Bromberger, S. & M. Halle. [1989]. Why phonology is different. Linguistic Inquiry 20, 51‑70.
Martin, E. & D. Osherson. [1998]. Elements of Scientific Inquiry. MIT Press, Cambridge Mass.
Journal of African Languages & Linguistics 14, 108‑16. [ISSN 0167‑6164]
Published typos hand‑corrected in this offprint.
1993[Spreading and downstep; prosodic government in tone languages]**
The Phonology of Tone; the Representation of Tonal Register, edited by H. v.d. Hulst & K. Snider, 133‑84. De Gruyter, Berlin. [ISBN 3110136058]**Note: archival offprint available on request; all the content is folded into my dissertation as pp. 61‑93.
[Verb focus in the typology of Kwa/Kru and Haitian]**
Focus & Grammatical Relations in Creole Languages, edited by F. Byrne & D. Winford, 3‑51. Benjamins, Amsterdam. [ISBN 1556191669]**Note: archival offprint available on request; all the content is folded into my dissertation as pp. 188‑220.Significant empirical wrinkle: In Yorùbá, clefted nominalized focus of the full VP, as in exx. (46b, 48, 57b) on pp. 20‑22 of the published version [= exx. (80b, 82, 91b) on pp. 203f. of the dissertation] exists independently in the literature (Awóyalé 1985, 78; Ajíbóyè & al. 2004, 33) but is rejected by other speakers. The conditions of this disagreement remain obscure to me!
Ajíbóyè, O̩. & al. [2004] "On the syntax of nominalization in È̩dó and Yorùbá" (Kinyĩra Njĩra! Step Firmly on the Pathway! Trends in African Linguistics 5 = Selected Papers from ACAL 31 at Boston University, edited by C. Githiora & al., 23‑42. Africa World Press, Trenton N.J.)
Awóyalé, 'Y. [1985] "Focus as an unbounded movement rule in Yorùbá" (Journal of the Linguistic Association of Nigeria 3, 75‑83)
1994Joint paper with R.‑M. Déchaine.
Natural Language & Linguistic Theory 12, 203‑57. [ISSN 0167‑806X]
1995Joint paper with K. Hale & Ụ. Íhìọ́nụ́, based on the Spring, 1994 field methods class at MIT.
Theoretical Approaches to African Linguistics [= papers from ACAL 25, Rutgers University], edited by A. Akinlabí, 83‑107. Africa World Press, Trenton, N.J. [ISSN 1080‑5478]Langues et Grammaire, Actes du Premier Colloque, édités par L. Nash et G. Tsoulas, 237‑52.
Département des Sciences du Langage, Université de Paris‑8, Vincennes à Saint‑Denis.Niger‑Congo Syntax & Semantics 6, edited by V. Manfredi & K. Reynolds, 91‑112. Boston University African Studies Center.Significant empirical wrinkle: contra the standard description exemplified in ex. (5b) on p. 92, Adéṣọlá (2005, 190f.) observes two examples of ōrātio oblīqua in which the possible reference of an embedded 3sg. nominative clitic can include the main clause subject. Specifically, ó may optionally share the referent of Olú in the following examples:Granting that these data entail a breakdown of referential complementarity between the independent (òun) and clitic (ó) 3sg pronominal expressions, such breakdown is not unusual for morphological competition among pronominals across the binding literature as a whole (as reviewed here). The more relevant question is whether Adéṣọlá's observations disprove domain extension and by implication any syntactic account of antilogophoricity, motivating instead an a‑syntactic mechanism for logophoric effects, such as a de se dreamtime operator (Anand 2006, cf. Lakoff 1970).(i) Olú ti gbà [kí ó má .jẹ ìrẹṣì mọ́].
O. AUX receive COMP 3SG AUX eat rice any.more
'Olú agreed that s/he (= Olú or someone else) should not eat rice again thereafter'
(ii) Olú ti kéde [pé ó ń bọ̀ l'ọ́la].
O. AUX cry.proclamation COMP 3SG AUX come tomorrow
'Olú announced that s/he (= Olú or someone else) is coming tomorrow'Note that the dreamtime account doesn't come free of charge: it must pay the incalculable price of insulating semantic interpretation from syntax in principle, via the general possibility of "overwriting semantic parameters" (Anand 2006, 64, cf. Chierchia 1998), besides incurring the more limited tax of treating as strictly accidental the uncontested absence of phenomena like (i) and (ii) in cognate languages where "logophoricity" has been studied, such as Gbè (BK2), Ìgbo and Abe (BK1).Before jumping to unfalsifiable conclusions demanding unlimited bailouts from the Semantical Central Bank, it's worthwhile to check the more economical possibility, that some independent syntactic factor explains the nominative clitic's lack of antilogophoric behavior in Adéṣọlá's data. Pending systematic investigation, it jumps out from both examples that the indirect discourse containing the anomalously interpreted nominative clitic is presumptively not a syntactic complement, but rather a paratactic adjunct. If so, then domain extension (antilogophoricity) would not be expected to occur, so the data don't falsify the null, syntactic theory pace Anand.To a casual observer, English translation may conceal what is generally accepted in Yorùbá literature, namely (i) that the surface L tone of gbà 'receive' diagnoses surface intransitivity and marks the subsequent clause as an adverbial adjunct (Déchaine 2001, cf. Rosenbaum 1965), and (ii) that the predicate kéde 'announce' decomposes transparently as an unergative expression with immediate constituents ké 'cry' plus òde 'proclamation' (Abraham 1958, 361), similarly entailing that an immediately following clause necessarily occupies a noncomplement position. Parallel considerations of philological adequacy refuse Schlencker's philosophical "plea for monsters" (2003).Abraham, R. [1958]. Dictionary of Modern Yorùbá. University of London Press.
Adéṣọlá, O. [2005]. Pronouns & null operators; A‑bar dependencies & relations in Yorùbá. Dissertation, Rutgers University, New Jersey.
Anand, P. [2006]. De de se. Dissertation, M.I.T., Cambridge Mass.
Chierchia, G. [1998]. Reference to kinds across languages. Natural Language Semantics 6, 339‑405.
Déchaine, R.‑M. [2001]. On the left edge of Yorùbá complements. Lingua 111, 81‑130.
Lakoff, G. [1970]. Linguistics and natural logic. Synthése 21, 151‑271.
Rosenbaum, P. [1965]. The grammar of English predicate complement constructions. Dissertation, M.I.T., Cambridge Mass.
Schlenker, P. [2003]. A plea for monsters. Linguistics & Philosophy 26, 29‑120.
Niger‑Congo Syntax & Semantics 6, edited by V. Manfredi & K. Reynolds, 171‑82. Boston University African Studies Center.
1996International Journal of African Historical Studies 29, 557‑82. [ISSN 0361‑7882]
"Corrected republication" of subeditor's typographic botch in same journal vol. 29 pp. 95‑121.
1997Object Positions in Benue‑Kwa; papers from a workshop at Leiden University, June 1994, edited by R.‑M. Déchaine & V. Manfredi, 87‑122. Holland Institute of Generative Linguistics/Holland Academic Graphics, The Hague. [ISBN 9055690317], publisher closed.Significant correction: On p. 104, the gloss given for Yorùbá the left‑hand example in (53a) i.e. without the pluralizer àwọn should not include the definite plural 'the dogs'. The excluded reading is in fact available in a similar example of Mandarin, as reported on p. 29 of R. Yang [2001] Common nouns, classifiers & quantification in Chinese (Dissertation, Rutgers University, New Jersey), but the two examples are nonparallel: the Mandarin crucially lacks a sortal classifier which arguably corresponds to little n, and in the Yorùbá example the notional noun has a closed‑class prefix which again arguably corresponds to little n. The interpretive difference disproves any structural equation between the bare noun root of Chinese and the minimal free form of Yorùbá; the latter is structurally a bare singular as discussed below, and allows a plural reading if either indefinite or inanimate, as preciently noticed by Welmers [1973] African Language Structures (University of California Press, Berlekey), p. 220.Cahiers d'études africaines 145, 157‑211. [ISSN 0008‑0055]This article fills in a placeholder blurb in my dissertation (§4.5, pp. 304f.). The É̩hu̩gbò text on pp. 195‑201, copied verbatim from the dissertation pp. 343‑48, had been originally transcribed with more attention to content than to dialect‑particular inflectional morphology and consistency of orthographic word division — both issues in need of reconsideration.
Significant correction: On pp. 177f., I proposed in passing that the Ìgbo nouns for 'reincarnating spirit double' and 'day(light)', both of which are pronounced chí, share a common root which is still productive in the language as a so‑called 'verb extension' meaning either 'repeat' or 'return'. This idea is however disproved by one simple fact which I should have checked at the time, namely that the predicative root in question has an aspirated onset consonant in all Ìgbo dialects which include this phonological feature, whereas the two nouns do not (cf. pp. 110 and 119 of Rev. Ígwè's 1999 Ìgbo‑English Dictionary). Hence it is impossible for that predicative root to be historically connected to either of the two nouns, and so an alternative explanation is required if the nouns are etymologically linked to each other at all. (Of course the null hypothesis is also possible, that 'day(light)' and 'reincarnating spirit double' are accidental homophones in Ìgbo, but as astute a cultural commentator as Òdógwu Chínù̩á Àchebé ventures otherwise, in his famous 1975 essay "Chí in Ìgbo cosmology".) Moreover there can be no doubt that the 'day(light)' meaning of chí is ancient, indeed it is reconstructed by Mukarovsky (1976, 146, 152) all the way back to a stage close to the ancestor of the entire Niger‑Congo language family, therefore it would be highly unlikely for an etymology of this word to be still recoverable Ìgbo‑internally. Accordingly, a new proposal is made in the final section of this manuscript, taking into account a cosmological comment in Melzian's È̩dó dictionary for the analogous noun è̩hi: "It is believed to be 'with a man all the day' " (1937, 51). Based on this remark (which is probably attributable to Melzian's primary consultant, Mr. H.G. Amadasu), and on supporting evidence cited there, I suggest that the Ìgbo noun for 'reincarnating spirit double' is historically identical to the noun for 'day(light)' thanks to a trivial metonymy of a thing to its canonical context, motivated by an assumption which was originally made by Northcote Thomas (1914, 19) and which can scarcely be doubted by anyone who has compared the two neighboring civilizations, to the effect that the cosmological pragmatics of È̩dó è̩hi and Ìgbo chí are remarkably parallel. The analogy is further demonstrated at Ágbò̩ ("Agbor"), the intermediate border kingdom with strongly bicultural Ìgbo and È̩dó leanings, where è̩hi appears in countless personal names in the same 'slot' in which chí is found in the eastern Ìgbo counterparts, to all appearances synonymously, e.g. È̩hi edú = Chí nà‑edú 'the reincarnating spirit‑double leads'.
1998Joint paper with R.‑M. Déchaine.
Recherches Linguistiques de Vincennes 27, 71‑94. [ISSN 0986‑6124]Significant correction: A referential plural gloss is systematically and erroneously attributed to cased, bare animate nPs in Ìgbo. The interpretation 'rats' should be deleted wherever it appears in examples (18b), (19b,c), (20a), (22a,c), (25a), (28), (29), (36a), (40a). The mistake is not directly relevant to the argument of the paper.
2001Encyclopedia article, prepared jointly with E. Ézè.
Facts about the World's Major Languages; an encyclopedia of the world's major languages, past & present, edited by J. Garry & C. Rubino, 322‑30. H.W. Wilson, Bronx, New York [ISBN 0824209702].Tone correction: the language name conventionally spelled "Igala" probably does not have the LHL tones printed on p. 322 following Banfield (1914, 178) and Armstrong (1965, 78). It appears with MHL in Ìlò̩rí (2009), but two independent citations give initial H (Étù & Mìáchî 1988‑92, Ọmachonu 2012). The latter is consistent with reported prohibition of initial M in "noun class words" (Ọmachonu 2012, 22), presumably including the ethnonym with its proclitic vowel, and loanword change HHL>MHL is plausible given that Yorùbá prohibits initial H in the corresponding slot (Ward 1952, 37). The quality of the medial vowel needs more investigation.Armstrong, R. [1965]. Comparative wordlists of two dialects of Yoruba with Igala. Journal of West African Languages 2, 51‑78.
Banfield, A. [1914]. Dictionary of the Nupe Language. Niger Press, Shonga.
Étù, Y. & T. Mìáchî. [1988‑92]. Ígílâ ékọ́chẹ 1‑4. Heinemann, Ìbàdàn. OCLC 38436098. [Not personally consulted.]
Ìlò̩rí, J. [2009]. Noun‑plural formation in Igálà. Current Perspectives in Phonosyntax & Dialectology, edited by G. Adika & al., 1‑15. Department of Gur Gonja, University of Education, Winneba, Ghana.
Ọmachonu, G. [2012]. Ígálà language studies and development. Slides from 12th Igala Education Summit, Kogi State University, Anyigba, 28-29 December.
Ward, I. [1952]. Introduction to the Yorùbá Language. Heffer, Cambridge.
2003Typologie des langues d'Afrique et universaux de la grammaire, vol. 2: Benue‑Kwa, Soninke, Wolof [publisher page], edited by P. Sauzet & A. Zribi‑Hertz, 127‑62. Presses Universitaires de Vincennes/Éditions de l'Harmattan, Paris. [ISBN 2747554872; manually corrected proofs]
2004Contours; a journal of the African diaspora 2.2, 239‑87. [ISSN 1543‑902X], journal deceased.Tone typo: The second word in fn. 27 (p. 270) should be Ékpè [HL] and not Èkpe [LL]; the LL item refers to a dance genre (i.e. a subtype of the large category of performances called in Ìgbo égwu [HH]) rather than to the title society under discussion which happens to deploy its own proprietary masked dance representations. The latter institution is denoted by the HL item wherever it is found, including the southern and eastern Ìgbo‑speaking area (cf. Ígwè, Ìgbo‑English Dictionary, University Press Ltd., Ìbàdàn 1985‑1999, p. 155) and specifically in Árụ̀ (= colonial "Arochukwu", cf. video interview with Ézè Árụ̀ by I. Miller, 2008) where the title society has big political clout. Two observations show that the respective forms with HL and LL represent one single word which traveled distinct historical paths, as opposed to being random near‑homophones. First, festival songs of "Èkpe" [LL] in Ọ́màáhyá (= colonial "Ụmụahi̩a") include numerous emblematic references to ágụ́ [H!H] 'leopard' (cf. E. Íkeokwú long essay, Department of Linguistics & Nigerian Languages, University of Nigeria, Ǹsụ́ka 1984, pp. 18f.). Second, the Ẹ̀dó word for 'leopard' is è̩kpe̩n (cf. Melzian, Concise Dictionary of the Bìní Language of Southern Nigeria, London, 1937, p. 53), with the same LL as in the Ìgbo word for the dance genre. (The nasalization of the root in the Ẹ̀dó reflex is probably related to the syllabic nasal which constitutes the noun prefix, sometimes singular and sometimes plural, of 'leopard/panthère' in many of the languages sampled in the Benue‑Congo Comparative Wordlist, Ìbàdàn 1968, pp. 222‑25.) In sum, the existence of the form ékpè [HL] in Ìgbo‑speaking communities reflects the borrowing from Èfịk of the historic word for 'leopard' along with its Èfịk pitch pattern of [HL], used as the proper name for the leopard‑themed title society which flourished in the catchment area of the Èfịk trade empire known to European merchants and colonists as "Old Calabar" (cf. Talbot, Peoples of Southern Nigeria, 1926, pp. 183f.). This borrowing endowed many southern and eastern Ìgbo‑speaking communities with a secondary lexical item separate from the primary/inherited Ìgbo form of the word with LL, which had meanwhile lost in Ìgbo the literal denotation of 'leopard' (having been supplanted by the hypokoristic ágú̩, literally 'the ravenous one', cf. águ̩ú̩ or águ̩ó̩ [HH!H] 'hunger') and become semantically opaque in the specialized context of phrases like í̩gbá èkpe [H!H LL] referring to heroically vigorous dance‑play — so‑called from its emulation of what the 1986 Nobel laureate in literature might like to call the leopard's "tigritude" or more prosaically, "a type of dance requiring much energetic action and so restricted to able‑bodied young men" (Ígwè p. 155).Missing from references:
Crabb, D. [1965]. Ekoid Languages of Ogoja, Eastern Nigeria. Cambridge University Press.
Jones, G. [1956]. Report of the position, status & influence of chiefs & natural rulers in the Eastern Region of Nigeria. Government Printer, Énugwú.
Significant empirical wrinkle: The cartographic errorism — what would today be called unavoidable and regrettable targeting error — responsible for sticking the name "Calabar" on the Efịk capital is much less complex in legendary inspiration, and slightly less farfetched in geographic distance, than modern historians have supposed. As quoted in the paper (p. 254), Jones wondered whether "the European attribution of the name Old Calabar to the Efịk people could be a reflection of the Korome myth of origin [… about] the place which the Opukoroye line of Kalaḅarị kings claimed as their original home" (1965, 159). More plausibly and prosaically, Ejituwu suggests that the intended referent of the "old" term in this myth was not the Efịk‑speaking village group on the "Rio da Cruz", but instead a Kalaḅarị‑speaking settlement (subsequently abandoned) on a branch of the "Rio Real" estuary labeled "Old Calabar River" by Barbot's 1699 "New Correct Mapp of Calbar River" (reprinted by Barbot 1732, 462 and Ejituwu 1998, 137). Further ambiguity (as if any were required) is supplied by the fact that "when New Calabar itself segmented from 1879 to 1885 […] [t]he Kalaḅarị in Bakana, Abonema and Buguma continued to regard New Calabar as Elem Kalaḅarị, which means 'Old Calabar', and the latter continued to appear in official documents till 1931" (Ejituwu 1998, 142). None of this confusion is surprising, given that adnominal modifiers like old and new are indexical "shifters" whereas map terms ideally aren't. But unlike Jones' frankly speculative account of the Efịk mistaken identity for "Calabar", Ejituwu's explanation of the mishap has independent documentary support and is moreover simpler: no need to assume that Dutch mappers of 9ja's eastern coast had even indirect access to Kalaḅarị dynastic tales, if what happened is that they ploddingly reproduced some coastal traders' casual misplacement of the older of the "Old" Calabars — i.e. the one so designated by locals in the 17th century — by a few hundred miles, to a different slaving depot a few estuaries further along to the east, albeit in a very different linguistic territory.Barbot, J. [1732]. A description of the coasts of north and south‑Guinea […]. Churchill, London.
Ejituwu, N. [1998]. Old Calabar rediscovered. The Multidisciplinary Approach to African History; essays in honor of Ebiegberi Joe Alagoa, edited by N. Ejituwu, 133‑50. Hisis Press, Port Harcourt, for University of Port Harcourt Press.
2007Focus Strategies in African Languages; the interaction of focus & grammar in Niger‑Congo & Afroasiatic [publisher page], edited by E. Aboh & al., 15‑54. DeGruyter, Berlin. [ISBN 9783110195934, ISSN 1861‑4302]. Final manuscript [17 pp. A4]. Offprints available by request.UPDATE 2 NOVEMBER 2010: Encouraging agreement with this analysis in many empirical as well as theoretical respects (though with significantly different formal implementation) can now be consulted in the following publication:Zubizarreta, M.‑L. [2010]. The syntax and prosody of focus; the Bantu‑Italian connection. Iberia; An International Journal of Theoretical Linguistics 2, 131‑68.
2009Historical Syntax & Linguistic Theory [publisher page] (= DIGS 9, the Ninth Diachronic Generative Syntax Conference, Trieste 2006), edited by P. Crisma & G. Longobardi, 329‑43. Oxford University Press. [ISBN 0199560544]. Prefinal proofs [15 pp.] plus bibliographic references for this chapter in manuscript form [4 pp. A4]. (Omnibus references for the volume appear on pp. 375‑412 of the actual publication.)
ABSTRACT: Niger‑Congo's Kwa and Benue‑Congo zones, jointly covering most of tropical Africa, run between isolating and agglutinative types. Historical phonology finds few shared innovations above the local cluster, but assuming the phase theory of generative syntax, a sharp division emerges based on the timing (early/VP vs. late/TP) of PF‑spellout.
UPDATE 18 July 2007: A letter in the current issue of Nature reports that paleontological and genetic data converge in reconstructing human origins to an area which — as I can't help noticing — happens to coincide closely with the current Benue‑Kwa speaking area. See this screenshot of Fig. 2. Likely origin of anatomically modern humans from p. 347 of A. Manica & al., "The effect of ancient population bottlenecks on human phenotypic variation" (Nature 448, 346‑48, 19 July 2007). Of course the observation is anachronistic, because the physiological reference point is >50K years old while BK's collective identity is presumably much younger, but the areal coincidence is still striking. A more neutral paraphrase: present BK‑speaking populations include the greatest phenotypic as well as genetic diversity of any large‑scale human aggregate. This result cannot lessen the general interest of comparative BK studies, but I predict that someone soon will conveniently forget the "K" (Kwa) part of "BK" (Benue‑Kwa) and try to interpret these maps in a Bantuist vein!
UPDATE 15 April 2011: It's disappointing that anyone would be counting "phonemes" nearly a century after Jakobson, Karcevsky & Trubetskoy discovered binary distinctive features and implicational universals (la regularité des rapports de corrélations), cf. "Quelles sont les méthodes les mieux appropriées à un exposé complet et pratique de la grammaire d'une langue quelconque?" (Actes du premier congrès international de linguistes à La Haye, du 10‑15 Avril 1928, 32‑36; reprinted in Roman Jakobson Selected Writings 1; Phonological Studies, 3‑6. Mouton, The Hague) and 25 years after Kaye, Lowenstamm & Vergnaud refined this idea as a sub‑syntax of unary/privative phonological elements, cf. "The internal structure of phonological elements; a theory of charm and government" (Phonology Yearbook 2 [1985], 305‑28). Nevertheless it's still probably significant that the demography of taxonomic (pre‑Jakobsonian) lexical contrast units — roughly, phonetic phenotypes — converges on the same originating population as does the demography of human anatomical genotypes: once again, the presently‑existing historically‑defined linguistic unit most closely corresponding to the set of languages in question is Benue‑Kwa. Twice lucky? See the top half of " Fig. 2. Likely area of language origin" from p. 347 of Q. Atkinson, "Phonemic diversity supports a serial founder effect model of language expansion from Africa" (Science 332, 346‑49, 15 April 2011) or this screenshot.
UPDATE 15 September 2012: Some of the same issues are covered in Hyman (2004), a learned survey enlarging Westermann's (1927) broad observation of east‑to‑west decline in morphosyntactic complexity and in the maximum size of predicate roots. In the absence of theory, of course, there can be no expectation of quantal restructuring. Instead, Hyman suggests (i) that the innovations under investigation "modified the proto system… in an areal fashion" (p. 71) and (ii) that the process was not unidirectional, since Ìgbo is cited as a language in which "doubtless… extensions have arisen via renewals" (p. 86). But how far can a theoretically untrammelled picture of reversible Wellen be maintained while still admitting a large‑scale structural shift towards isolating syntax and monosyllabic roots? Can such a dramatic diachronic asymmetry, expressed across vast stretches of time and space, be more than a statistical fluke? Part of the problem may be that Hyman bravely assigns himself the whole of "Niger‑Congo" (the world's oldest and most complex language family) as the canvas for his illustrations, whereas Westermann's original observations were restricted to the more tractable — but still impressively big — zone which came to be known as "Benue‑Kwa" (e.g. Elugbe & Williamson 1977). It's also surprising that someone who went on to write Hyman (2011) would not follow up his own observation that "almost all Bantu languages show stembound phonological prosodies" (p. 85) by mentioning the markedly richer tonemic inventories at the "Kwa" end of the spectrum. (Either that's another accident, or else tone is really "different" after all. Not both.) Aligning his three parameters of change, Hyman notes that "at least relics of the original morphology survive beyond the syntactic and phonological restructurings" (p. 86), entailing that "syntax" changed autonomously to some extent, and hinting at no mechanism connecting syntax to the shape of roots. Coincidence upon coincidence.Elugbe, B. & K. Williamson. [1977]. Reconstructing nasals in Proto-Benue Kwa. Linguistic Studies Offered to Joseph Greenberg, edited by A. Juillard, 339‑63. Anma Libri, Saratoga.
Hyman, L. [2004]. How to become a 'Kwa' verb. Journal of West African Languages 30, 69‑88. [free prepress version]
———. [2011]. Tone; is it different? Handbook of Phonological Theory, 2nd edition, edited by J. Goldsmith & al., 197‑239. Blackwell, Oxford. [free prepress version]
Westermann, D. [1927]. Die westlichen Sudansprachen und ihre Beziehungen zum Bantu. [=MSOS 29 Beiheft]. DeGruyter, Berlin.
UPDATE 16 May 2013: BK2's birth being the hypothetical result of deletion (specifically, the erosion of finite inflection), or so I say, risks the irrelevance of a linear branching model of substantial inheritance — as opposed to disinheritance. It makes you wonder what share of i‑language speciations are of this general type, and whether identification of more such events could reduce apparent radiation/multibranching in archaic Stammbäume — a matter which continues to disquiet Indoeuropeanists, most of whom understandably were trained to study e‑language (e.g. Garrett 1999). Genetic reduction is a live topic in cellular phylogeny (Bapteste & Gribaldo 2003) though not as hot as incongruence/lateral transfer (Leigh & al. 2011).Bapteste, E. & S. Gribaldo. [2003]. The genome reduction hypothesis and the phylogeny of eukaryotes. Trends in Genetics [ISSN 0168‑9525] 19, 696‑700.
Garrett, A. [1999]. A new model of Indo‑European subgrouping and dispersal. BLS [ISSN 0363‑2946] 25, 146‑56.
Leigh, J. & al. [2011]. Evaluating phylogenetic congruence in the post-genomic era. GBE [ISSN 1759‑6653] 3, 571‑87.
2010Lingua 120, 1327‑32. [ISSN 0024‑3841]. Final manuscript [5 pp. A4]. Offprints available by request.
2013Radical Egalitarianism; local realities, global relations [publisher page], edited by F. Aulino & al., 119‑36 with consolidated volume endnotes and references (Fordham University Press, Bronx, New York, 2013, ISBN 0823241904). Revisiting §4.4 of my dissertation and applying Tambiah's idea of "galactic polity" to the Niger Delta. [cover page plus 17pp. prefinal proofs of main text plus 4pp. A4 of individual endnotes and references, last modified 18 February 2012.]ABSTRACT: If the Southeast Asian "galactic polity" is defined by centripetal‑centrifugal "pulsation between modalities" of power (Tambiah 1976), something similar is true of West Africa's 'Asiatic' social formations — indigenous tributary monopolies that flourished during the transatlantic slave trade before becoming "encapsulated" in British rule (Otite 1975). Structural ambivalence explains how Fortes & Evans‑Pritchard (1940) could lump "stateless" Ìgbo and nomadic Nuer together into the "acephalous" political type. This paper reviews three southern Nigerian examples of evanescent hierarchy. In April 1979, Íkenchúku the youthful ruler of Ágbọ̀ Kingdom died from nocturnal gunfire during a land case and an electoral campaign, as an elderly chief was rewriting Ágbọ̀ history to push back the pendulum from absolutist primogeniture towards collective lineage rights. In November 1995, the writer Ken Saro‑Wiwa was eliminated by "judicial murder" ostensibly because he had denounced a list of ten Ogoni chiefs — four of whom were subsequently lynched — as having sold out to Shell Oil Corporation and General Abacha. In November 2000, an Ìgbo anthropologist was roughly rebuked by an Ìgbo historian for refusing to parse the slogan Ìgbo énwé ezè 'Ìgbo has no paramount ruler' as synonymous to colonial cliches like "Biafran society is traditionally egalitarian" (Òjúkwu 1969 cf. Meek 1937). Instead of treating these perturbations as unrelated events, Tambiah's framework suggests they are causally linked in a long‑term social formation, within which the entrenchment of local political brokerage is complemented by the steady collapse of nationalism.BIBLIOGRAPHIC NOTE: The main descriptive source of this study is Chief Augustin È̩gwabó̩ Ìdúùwe̩'s History of Greater Ágbò̩[r]. That irreplaceable text remains unpublished today, three decades after the author entrusted the manuscript to me, so I belatedly and apologetically post it here in its unfinished form of circa 1982.Correction: as now debunked, the "Possible image of a salon in General Babangida's château" presented in an earlier draft of this paper as §5.6 actually shows a California mansion. Nevertheless Babangida's immense personal wealth has no legal explanation.UPDATE: The issues reviewed in this paper receive a less political take in a 25‑minute TV documentary, "George Oṣòdì: Kings of Nigeria", eliciting my response: "Déin in Abachan drag?" [3 pp. A4, 9 May 2013].
and somewhere in the pipeline…For some of these essays, the relevant tube of the pipeline can be regarded as clogged, and the items in question merely blogged. To steal a cute metaphor from my teacher Jochem Schindler, "modern linguistic publications have a very short half‑life" — the time it takes for "half of their claims to be proven wrong". If things are so bad for printed books and articles, then for these mere electrons it's a real race against exponential decay before any scientific weight converges to the infinitesimal. So catch them while you can!Social Anthropology Colloquium, Department of Anthropology, Harvard University, 27 February 1995.
[handout, 5 pp. 8.5x11]For a more detailed and definitive answer to my rhetorical question, see now:Harbour, D. [2011]. Mythomania? Methods and morals from 'The myth of language universals'. Lingua 121, 1820‑30.
Département de Linguistique, Université Paris‑7, 1 December 2004.
[handout, 6 pp. A4]Ms., Boston University, 25 November 2012.
[3 pp. A4, last modified 1 December 2012]ABSTRACT: Yorùbá ò̩gá 'boss' (< ‑gá 'stand aloof') is not cognate to Fọ̀n‑Gbè gán 'chief, patron' (< ‑gán 'support'), pace Herskovits (1956, 156). By standard criteria of form and meaning, the closest Westafrican correspondent of the Afrobrazilian ritual title ogan (Carneiro 1940, 274, cf. Landes 1947) is Gùn‑Gbè ògán, its functional counterpart in the vodun of Xọ̀gbónù alias "Porto‑Novo" (Rouget 2001, 97). This philological finding is demographically significant because the most vodum‑oriented candomblé tendency is called Jeje (also spelled Gêge), while Jeje (also spelled Djédjé) is an exonym of the Gùn‑Gbè population (Capone 1999, 15 fn 6).Akoha, A. [2010. Syntaxe & lexicologie du Fò̩n‑Gbè, Bénin. Harmattan, Paris.
Awóyalé, 'Y. [2007]. Yorùbá lexical database. Linguistic Data Consortium, University of Pennsylvania, Philadelphia.
Capone, S. [1999]. La quête de l'Afrique dans le candomblé; pouvoir & tradition au Brésil. Karthala, Paris.
Carneiro, E. [1940]. The structure of African cults in Bahia. Journal of American Folklore 53, 271‑78.
Herskovits, M. [1956]. The social organization of the Afrobrazilian candomblé. Phylon 17, 147‑66.
Landes, R. [1947]. The City of Women. Macmillan, New York.
Rouget, G. [2001]. Initatique vôdun; images du rituel. Éditions Sépia, Saint‑Maur.
TOPICAL ADDENDUM 3 April 2013: In recent weeks the Yorùbá term ò̩gá has achieved epidemic renown — helpfully illustrating the connotation of obsequiously‑regarded "bigmanity" referenced in the foregoing discussion — in a viral video "My ò̩gá at the top" whose context is explained here.Response to the "fourth dimension" constitutional strategy in the 9ja area, 1 March 2005. [original link: Nigeria World]
[web post, 2 pp. A4]The prosodic infrastructure of focus in Benue-KwaU Venezia 25/11/2004; U Genève 30/11/2004; U Amsterdam 3/12/2004; U Leiden 5/12/2004; U College, London 8/12/2004.
[handout, 9 pp. A4]Plentiful typos aside, I also managed to reverse the meaning of the mathematical term "diagonalization", which Wagner had applied sarcastically to the mainstream literature of syntax‑free prosodic analysis. In fact the exact aim of this talk was to apply Wagner's syntactically non‑diagonalizing idea to Benue‑Kwa!Ineffable tenses in Benue-Kwa and RomanceResearch document, 12 September 2012.
[handout, 3 pp. A4]ABSTRACT: Some natural languages can't distinguish past from present perfect without recourse to periphrasis or context. This expressive gap is unexpected if tenses are autonomous meaning postulates (Reichenbach 1947, Hornstein 1990, Giorgi & Pianesi 1997), but is grist for theories where tense interpretations emerge from syntax-pragmatics interaction indirectly (McCawley 1971, Vikner 1985, Enç 1987). English has an unambiguous present perfect (*I have danced yesterday, Klein 1992) but French doesn't (J'ai dansé hier taking over Je dansai hier), and the atrophy of the simple preterite in northern Romance coincides with several other developments related to case, agreement and aspect (Zamboni 2000, 87 cited by Ledgeway 2012, 314). In the Benue‑Kwa (BK) subgroup of Niger‑Congo, an unambiguous present perfect occurs in most of the major clusters, including Àkan, È̩dó, Ìgbo, Cross, Plateau and Bantoid, but does not exist in a substantial, contiguous subset — call it BK2 — comprising Gbè, Yorùbá, Nupe and Ìdo̩mà, where any finite affirmative episodic predicate in principle allows either present perfect or simple past construal. This bifurcation of BK arguably followed on the erosional loss of finite affixation in BK2, and introduced other correlated changes of a quantal nature in its wake (Manfredi 2005a, 2009). One reason UG doesn't need to stipulate a tense system is that time reference is partly redundant with default interpretation of lexical Aktionsart (Green & Ígwè 1963, 53; Welmers & Welmers 1968, 76; Dowty 1986, Déchaine 1991, Sorace 2000). But the crosslinguistic parallelism is incomplete: in southern Romance, present perfect morphology automatically codes for recent past, but in BK1 these two traits are orthogonal (Welmers 1973).Tense parameters and serial verbs[2nd draft, last modified 12 July 2005, 34 pp. A4]Invited in 2002 to a project originally entitled Studies in the Syntax of Kwa; a generative perspective, edited by E. Aboh & J. Essegbey. Second draft, much revised in response to substantive comments from the editors, was withdrawn 9 May 2007 after the (technically anonymous) external reviewer refused to even read it on the creative excuse that it's "dense" and "flowery" — if he meant "al dente" and "floury" he may be pardoned on orthodontic and dietary grounds although he should nevertheless have been ethical enough to disqualify himself instead of shirking the job while accepting the reviewer's honorarium! Less excusable was the decision of the editors not to insist that the publisher obtain an actual review of the chapter, in the absence of which I was unfairly placed in the position of having to guess how much of the reviewer's problem was due to his delicate stylistic sensibilities and how much was an unwillingness to read substantive criticism of his own work and that of his close cumpari. (No empirical or theoretical errors were indicated in the non‑review which he provided.) Another debilitating confusion on the part of the editors was their concept of "Kwa" whose descriptive coverage doesn't coincide with any proposed historical use of this term (e.g. Greenberg 1963; Williamson 1989), conforming instead more or less to Westermann's lexico‑typological sense of those "Sudanic" languages which tend to monosyllabic roots (1927, 20). In this way the project lost its coherence as a comparative syntax handbook, and the volume which eventually appeared (chez Springer, with the title Topics in Kwa Syntax, ISBN 978‑90‑481‑3188‑4) would have been more transparently titled Syntactic studies in some of the more isolating Benue‑Kwa clusters, namely Gbè, Àkan & Yorùbá, with special reference to Gùn‑Gbè (cf. review by M. Dakubu, Studies in Language 34, 442‑52).BK1 alias 'Proto‑Potou‑Akanic‑Bantu' — a study of diachronic syntaxDepartment of Linguistics, University of Ghana, Legon, 6 October 2010.
[handout, 9 pp. A4]ABSTRACT: John Stewart's many studies of Àkan yielded general findings about Niger‑Congo, including synchronic diagnostics of serial constructions (1963) and historic soundshifts (2002). Read together, his oeuvre confronts a more abstract issue: how phonology and syntax interact i.e. whether they share formal properties (KLV 1985; Bromberger & Halle 1989; Scheer 2008). In principle, language‑specific information could be transmitted on multiple, parallel paths; in practice, "external evidence" (Kiparsky 1971) decides. Since Westermann (1927) it's been unknown how to reconcile the strong diversity of the area later called Kwa and Benue‑Congo (Greenberg 1963; Williamson 1989) with weak lexical reflexes of historical discontinuity; three proposals are compared. Large‑scale dialect continua may well exist in general, but some restructuring must separate the agglutinating and isolating extremes subtended by Stewart's "Potou‑Akanic‑Bantu" (PAB); the task is to find causal chains and see whether various components of grammar shifted in synch. As it happens, PAB's reconstructed 4‑way stop contrast was reduced by half in BK2, and this 'catastrophic' phonetic change is matched by a fourfold mutation of grammar, with two inaudible interpretive traits, one piece of audible morphology and something traditionally but mistakenly considered as narrow phonology. The trigger for the lot, thus for bifurcation of PAB=BK, was arguably suffix erosion, a language‑external process (Manfredi 2005, 2009).
ReferencesAboh, E. & J. Essegbey eds. [2010]. Topics in Kwa Syntax. Springer, Dordrecht.
Bromberger, S. & M. Halle. [1989]. Why phonology is different. Linguistic Inquiry 20, 51‑70.
Greenberg, J. [1963]. The Languages of Africa. Mouton, the Hague.
Kaye, J., J. Lowenstamm & J.‑R. Vergnaud [=KLV]. [1985]. The internal structure of phonological elements; a theory of charm and government. Phonology Yearbook 2, 305‑28.
Kiparsky, P. [1971]. Historical linguistics. A Survey of Linguistic Science, edited by W. Dingwall, 576‑649. UMaryland, College Park.
Manfredi, V. [2005]. Tense parameters and serial verbs. Manuscript invited to, then disinvited from, the collection which eventually became Aboh & Essegbey eds. (2010).
———. [2009]. Morphosyntactic parameters and the internal classification of Benue‑Kwa. Historical Syntax & Linguistic Theory, edited by P. Crisma & G. Longobardi, 329‑43. Oxford University Press.
Scheer, T. [2008]. A Lateral Theory of Phonology Vol. 2; How morpho‑syntax talks to phonology; a survey of extra‑phonological information in phonology since Trubetszkoy's Grenzsignale. Ms., Université de Nice.
Stewart, J. [1963]. Some restrictions on objects in Twì. Journal of African Languages 2, 145‑49.
———. [2002]. The potential of Proto‑Potou‑Akanic‑Bantu as a pilot Proto‑Niger‑Congo, and the reconstructions updated. Journal of African Languages & Linguistics 23, 197‑224.
Westermann, D. [1927]. Die westlichen Sudansprachen und ihre Beziehungen zum Bantu. [=MSOS 29 Beiheft]. De Gruyter, Berlin.
Williamson, K. [1989]. Niger-Congo/Benue-Congo overview. The Niger-Congo Languages, edited by J. Bendor-Samuel, 3‑45/247‑74. American Universities Press, Lanham, Md.
Aspect versus the serialization parameterInstitute for African Studies, Universität Leipzig, 12 October 2005.
[handout, 8 pp. A4]Caveat lector: data (22), (24) and (26) in the handout were garbled in haste, and will be corrected as soon as possible.Ìgbo transitivity in a derivational frameworkPresented in absentia at Ìgbo Language Symposium 3, Zik University, Ó̩ka ("Awka"), Ànámbàra ("Anambra"= Ò̩mámbàla) State, 9ja ("Nigeria"), 28 November 2011.
[6 pp. A4, last modified 18 January 2012]The transitivity of focus in western Benue‑Kwa (Niger‑Congo)BU Linguistics Research Sharing Forum, 15 November 2006; earlier versions given at SFB 632 "Information Structure", Humboldt‑Universität Berlin, 7 October 2005; Leiden University Centre for Linguistics, 18 October 2005; Dipartimento di Scienze del Linguaggio, Università di Venezia Cà Foscari, 13 June 2006; Workshop on "Minority languages: lexical aspects & discourse grammar", Dipartimento di Linguistica, Università degli Studi Roma‑3, 17 June 2006.
[handout, 7 pp. 8.5 x 11 inches]Phonosemantic subordinationDepartment of Linguistics, Harvard University, 12 March 2007
[handout, 8 pp. A4, last modified 25 March 2007]Instrumental evidence for Nuclear Stress and destressing effects in Kinande/Luyiira (JD 40 superceding Guthrie D 42)HLT Research Group, Meraka Institute, Pretoria/Tshwane, 11 October 2007
[handout, 10 pp. A4]Supporting pitch tracks (audio on request) can be consulted at the following links:
15 May 2007 (12 pp. A4)
22 May 2007 (14 pp. A4)
29 May 2007 (13 pp. A4)
13 June 2007 (17 pp. A4)
26 June 2007 (10 pp. A4)
Áfa, the Ǹri‑Igbo counterpart of IfáConference on Ifá divination in Africa & the Diaspora, Harvard University, 14 March 2008. Grupo de Estudios Africanos e Afrobrasileiros em Línguas e Culturas, Universidade do Estado da Bahia, 15 April 2009 ("Os caminhos de Ifá"). Abridged version submitted to conference proceedings.
[19 pp. A4, last modified 6 May 2013]ABSTRACT: Áfa and Ifá are two of many localizations of a farflung West African 'oracle' (information retrieval system) whose digital processor keys natural language text to 256 ordered pairs of 4‑bit arrays (ordered binary sets). While being ported south and west from the Benue rivervalley some 500 years ago, the oracle's own name and those of its proprietary terms underwent sound change in the receiving languages. These phonetic shifts, plus paralinguistic mutations, remain behind as footprints on the transmission routes. The texts themselves also evolved along their branching path as emergent élites turned from ancestral legitimation rooted underground, to novel authority dangling from the sky.BIBLIOGRAPHIC NOTE: Two files of supplementary data included in earlier versions of this manuscript are now posted separately:
Comparison of 4‑bit array names and associated information from oracle localizations across 5 historical zones
[1 p. 8.5 x 14 inches, last modified 8 April 2013]
8‑bit semantic key — comparison of 7 oracle localizations
[5 pp. A4, last modified 8 April 2013]Part two of this research can be found in the companion manuscript Before Wazobi̩a; Òminigbo̩n and polyglot culture in medieval 9ja, posted immediately below.Before Wazobi̩a; Òminigbo̩n and polyglot culture in medieval 9ja[37 pp. A4, last modified 16 April 2013]
Walter Rodney Seminar, African Studies Center, Boston University, 13 February 2012. Handout [4pp. A4, last modified 8 March 2012]ABSTRACT: A scholarly reincarnation of Egharhevba's Ìha Ominigbo̩n (1936) is long overdue. This Ẹ̀dó classic has intrinsic value as cultural description but is virtually unobtainable today. It is also relevant to wider historical relationships — even if not in the Ifè̩‑centric way intended by its author — and requires critically‑informed translation for both academic and popular access.
Preliminary comparisons with Yorùbá Ifá and Ìgbo Áfa show that Òminigbo̩n breaks the presentist mold of an encompassing "Yorùbá-Ẹ̀dó world system" (Ògúndìran 2003, 57, cf. Burton 1863, 222). Ìgbo etymologies exist for core Ìha terms ògwé̩è̩ga and n'áàbe, as well as for "Ogbẹide" — the second appellation of the person who introduced the system to Ẹ̀dó (Egharhevba 1936, 3, no tone indicated). Respectively, the Ìgbo sources are òkwé è̩ja 'oracle seeds', n'áàbo̩ 'double' and ò‑gbú ìre '[performing/possessing] very effective [sacrifices/medicine]'. An Ìgbo source for Òminigbo̩n is independently plausible from comparison of the casting method (Emọvọn 1984) with Ǹri‑Igbo procedures (Ọ́nwụejìọ́gwụ̀ 1978) in contrast to those of Ifá (Abím̄bọ́lá 1976). Transmission from Ẹ̀dó to Ìgbo is also consistent with how the oracle crossed the southern 9ja area in medieval times as reconstructed by evidence of loanword phonology, and with matches between the 256‑part semantic keys of Áfa and Ìha on the order of 35% — well above the level of chance similarity — whereas analogous array‑matching with Ifá is on present knowledge undefined (Manfredi 2009a).
These observations and a long list of others support the view that Yorùbá oral history has "confused Odùduwà with Ọ̀rúnmìlà" (Erediauwa 2004, 206) by telescoping several distinct eras together in the service of Ifẹ̀-centric narratives motivated by the modern politics of 9ja, the Nàìjá area, colonial "Nigeria" (cf. Ryder 1965; Vansina 1971, 457; Law 1973; Ọbáyẹmí 1979). The codification of Ifá shows the mnemonic bias of several arguably modular cognitive domains — number, folk biology, folk sociology, theory of mind — plus the nonmodular but eminently memorable "genius of paganism" (Augé 1982) and its "superstimuli" (Sperber & Hirschfield 2004, 45).UPDATE 12 January 2013: On Saturday 5 January 2013, the boss of Ifẹ̀ Central Local Government joined General Akínrìnádé and the Ọọ̀ni at "the commissioning of the gigantic Orí Oló.kun edifice […] at the Mayfair Roundabout" (Adéṣìnà 2013). This is the very spot where 9ja police (alias Sorrow, Tears & Blood) fatally stampeded a peaceful anti‑Ṣíjúwadé protest on Sunday 7 June 1981 (cf. §3.5 of my paper). Public doubling down on the Frobenius myth shows the force of Wazobi̩an consciousness; in the same vein, Olúpọ̀nà (2011).Adéṣìnà, B. [2013]. At Orí Oló.kun launch, Yorùbás canvass cultural renaissance. Guardian [Lagos], 11 January.Olúpọ̀nà, J. [2011]. City of 201 Gods; Ilé‑Ifẹ̀ in time, space & the imagination. University of California Press, Berkeley.BIBLIOGRAPHIC NOTE: Part one of this research can be found in the companion manuscript Áfa, the Ǹri‑Igbo counterpart of Ifá, posted immediately above.Bailey‑bridge to oil doom; Kay Williamson reveals SIL's official role in the maladministration of the post‑Biafran Niger Delta[6 pp. 8.5 x 11 inches, last modified 2 July 2007][APIP, early spellout and displacement in Benue‑Kwa (Niger‑Congo)]**
Presented at Comparing prosodies grammatically, Department of Linguistics, Harvard University, 20 September 2008; Conference on functional projections in analytic languages (the Sino‑Kwa project), RU Leiden, 1 November 2008 and Groupe de recherche sur les Grammaires Créoles (CNRS UMR 7023), Université Paris‑8, 3 November 2008.
[handout, 13 pp. 8.5 x 11 inches]**Note: original manuscript available on request; most of the content is recycled in Phase bifurcation — causes and consequences in Benue‑Kwa (Niger‑Congo) or in §2 of The referential prosody of bare arguments.The referential prosody of bare argumentsRevision of "APIP, early spellout and displacement in Benue-Kwa (Niger-Congo)". Presented at the workshop Bare nouns: syntactic projections and their interpretation, UParis‑7, 27 November 2009; Département de Linguistique, UOttawa, 9 March 2010.
[27 pp. A4, last modified 21 June 2012]ABSTRACT: Argument nP freely accesses unlexicalized referential D in certain languages (Cantonese, Wenzhou, Yorùbá, Gbè) but not in others (Mandarin, Taiwanese, Ìgbo, Haitian respectively) which are closely similar but whose lexical roots fall below a ternary threshold of prosodic complexity (§1). Consequences of this unexpected generalization are independently necessary.
At PF (§2), if 'more tones' cue 'more structure' then lexical tone is not phonology (pace Pike 1948; Goldsmith 1976; Ladd 1996) and/or phonology is not 'different' (pace Bromberger & Halle 1989). Instead, lexical tone is the pitch effect of metrical accent with phrasal distribution (Idsardi & Purnell 1997; Déchaine 2001) i.e. lexemes are phrases (Hale & Keyser 1993), and ternary tone makes available an 'extra' level of branching beyond what is required to host segmental material.
At LF (§3), the minimal free form in Yorùbá, e.g. a notional noun like ajá 'dog', is at least nP (cf. Welmers 1973a, 189f.). But argument ajá is not a bare singular: it can refer to indefinite plural individuals, which an Ìgbo animate nP cannot (Welmers 1973b, 220). In Mandarin, nP (sortal classifier plus root) in object position can only be indefinite singular, whereas root N (without SORT) is definite singular or (in)definite plural but not indefinite singular (Yang 2001, 29). The differences follow if so‑called 'little n' of an Ìgbo animate count noun like òké 'rat' is featurally singular, whereas Yorùbá little n is defective, being the pure spellout of an empty phrasal node but still with the status of a typed category, and Mandarin root N exploits pseudoincorporation and object pro drop. Yorùbá allows generic readings in episodic contexts with phonetic elision of semantics‑less little n as a pseudoincorporation cue, just as in Mandarin a similar interpretive effect is blocked by an overt classifier (Huang 2005). Ìgbo little n is not elidable under any circumstances, but Ìgbo makes episodic generics in a different way: in caseless configurations, nP is pseudoincorporated and not indefinite singular (Déchaine & Manfredi 1997). In addition to the foregoing, an unadorned nP can be definite in the ternary languages mentioned in §1.
As to learnability (§4), the varied conditions for free null D resist encoding as an abstract invariance/isomorphism of the computational system, but such failure is no loss, because the factorization of syntactic differences as (micro or macro) I‑language parameters is unfeasible on general grounds (Keenan & Stabler 1994, 2003). In Sinitic and Benue‑Kwa, the limits of possible mapping between referential and prosodic type in argument expressions can be modeled in the left‑peripheral prosodic template known as "the initial site" (Lowenstamm 1999), consistent with "inter‑modular argumentation" (Scheer 2010).UPDATE: Literature on category determination at spellout, consistent with the architecture of argument‑type expressions ("substantives") developed in this paper, according to which the semanticist's notional "nouns" are not lexical heads, now includes:Adger, D. [2012]. A Syntax of Substance. ISBN 0262518309 [publisher page], MIT Press, Cambridge Mass. [→free 10 pp. summary]
Oppose the nomination of Dr. Susan Rice as representative to the United NationsAn open letter to Senator John Kerry, Foreign Relations committee of the U.S. Senate, 7 December 2008.
Abacha's Quid-Pro-Clinton. Update to Open Letter, with a minor date correction, 20 December 2008.
[4 pp. 8.5 x 11 inches; p. 1 was originally cross-posted at www.saharareports.com/oppose_susanrice.php, but that link is currently dead]UPDATE 12 December 2012: Recent grandstanding by a couple of southern white male Republican senators about snafus in Obama's "Arab spring" catchup game in Libya has inadvertently renewed public attention on Susan Rice's official as well as covert benevolence to client African dictators throughout her ascent of executive‑branch appointments in the past 20 years (interspersed with lucrative corporate lobbying). The most discussed examples are special favors to civilian‑clothed avatars of Generals Meles and Kagame, whereas similar indulgences to Generals Abacha and O̩básanjó̩ have scarcely elicited a mainstream peep, despite U.S. planners' "failed state" scenario of Nigeria "dragging down a large part of the West African region". Whether Clintonesque financial "conflicts of interest" have been in the last analysis either more or less determinant than power intoxication of the kind displayed on TV by her mentor Madeleine Albright, in explaining Rice's repeated record of framing/implementing shameful U.S. sellouts of the populations enclosed within the nominal boundaries of Nigeria, Ethiopia and Rwanda/Congo, is less certain than that the present scandal provides a teachable moment for anyone who wants to know why Africa "still dey suffer today".Phase bifurcation — causes and consequences in Benue‑Kwa (Niger‑Congo)Departamento de Lingüística, Universidade de São Paulo, 22 April 2009
[handout, 8 pp. A4, last modified 2 July 2009]ABSTRACT: Benue‑Kwa (BK), the main branch of the Niger‑Congo family, is an 'epigenetic' (or non‑parametric) typological space (Gianollo & al. 2008; cf. Keenan & Stabler 1994; Newmeyer 2004) created by interaction of the internal and external dimensions of the human language faculty (I‑ and E‑language, Chomsky 1986). §1 reviews a timing shift in cyclic spellout — the derivation‑internal branching to PF — apparently triggered by external factors (Manfredi 2005, 2009a). In the remnant area (BK1) the lowest spellout domain is TP, including finite inflection aligned on or after the predicate root, but in the innovating area (BK2, including Gbè, Yorùbá‑Igálà, Nupe‑Ebira and Ìdọmà) the lowest domain contains only open‑class predicators, thus is no bigger than VP, leaving closed‑class clausal superstructure to a separate, subsequent phase. §2 considers the mechanism of this bifurcation and the status of prosody therein as either determinant or determined. The initial state, today represented collectively by BK1, allows at most binary 'tone' (lexical discrimination of roots by pitch), whereas this contrast is at least ternary throughout BK2. Some BK1 languages (Ẹ̀dó, Western Ìgbo) maintained late spellout even after obligatory finite suffixation had eroded to zero; short of appealing to analogy with optional suffixes, the only explanation is externalist and accidental: acquisition in BK2 of early spellout was triggered by a ternary lexical pitch contrast, most likely pushed by phonation effects which are still transparent in Gbè (Stahlke 1971). Such an inference is ruled out by standard views of prosody as "phonology" (Nespor & Vogel 1986; Ladd 1996) and of phonology as "different" (Bromberger & Halle 1989), but is supported by syntagmatic/metrical analysis of 'tone' (Bamba 1992; Manfredi 1995a,b; Dilley 2005a,b,) and also on statistical grounds both in BK and in Sinitic (Manfredi 2009b). §3 claims that other cross‑BK differences are incidental to phasing: not just the pragmatic (narrow‑semantic) contrasts discussed in §1, but also certain options of overt morphosyntax: in‑situ variables, affixal anaphors and marked accusatives in BK1; modals, inverse copulas and marked nominatives in BK2. All these would standardly be treated as independent parameters, but within the BK typological space all are arguably contingent on prosody as understood in §2.Introductionto "Some notes on nominal phrases in Haitian Creole and Gùngbè; a trans‑Atlantic Sprachbund perspective" by O̩. Aboh & M. DeGraff.
5th conference on the Semantics of Under‑represented Languages in the Americas, Harvard University & M.I.T., 16 May 2009
[handout, 1 p. 8.5 x 11 inches]Cyclic accentuation in Yorùbá [abstract][1 p. A4, last modified 30 August 2012]
UPDATE 9 January 2013: The phrasal accentuation pattern described here has been missed by decades' worth of "tonal" analysis, though it's been hiding in plain sight, all the alternations having been separately observed in mainstream literature. Anyone who finds this generalization to be linguistically significant has at hand a ready reply to the rhetorical provocations in the title of Hyman (2010) — a paper that makes easy hay out of the traditional, translation-based distinction between word and phrase, and that elides effortlessly between phonetic "stress" and PF "accent" — all this Sturm und Stress just to save the exotic Pikean toneme from extinction. Similarly, Yorùbá as analyzed here is a clear counterexample to Hyman's brave claim that "no language MUST be analyzed with a third device called pitch-accent. A tonal and/or stress interpretation is always available" (2011, 221, emphatic capitals in original).Hyman, L. [2010]. Do all languages have word accent, or What's so great about being universal? 2010 Annual Report, 297‑315. Phonology Lab, University of California, Berkeley.
———. [2011]. In defense of prosodic typology; a response to Beckman & Venditti. 2010 Annual Report, 200‑35. Phonology Lab, University of California, Berkeley.
Pike, K. [1948]. Tone Languages; a technique for determining the number & type of pitch contrasts in a language, with studies in tonemic substitution & fusion. University of Michigan Press, Ann Arbor.
ABSTRACT: Lexically spurious H tone is a core phenomenon of Yorùbá. Awóbùlúyì (1975) and Bám̄gbóṣé (1983) debated whether it is syntactico‑semantic or morpho‑phonological in origin, but the boundaries of phonology and syntax have shifted meanwhile. Traditionally, lexically spurious tones have two possible sources: tone‑bearing segmental morphemes that suffer vowel deletion in phonology, or morphemes whose PF is strictly tonal. Minimalist grammatical architecture adds a third option: prosodic effects of cyclic linearization (Dobashi 2004; cf. Bresnan 1971), and only the last view offers a general solution, because the varied contexts of spurious H lack a representational common denominator which could be encoded as a single morpheme. Of course the different instances of spurious H in Yorùbá could be genuinely unrelated, but that would leave a double mystery: why no toneme other than H is inserted, and why the set of insertion contexts is syntactically well‑defined.
The Yorùbá facts conform to the claim that the phase head — that syntactic category whose sister XP maps to phonology as a unit (Chomsky 2001) — is metrically strong, and that VP is strong by default (nuclear stress, Chomsky & Halle 1968; Cinque 1993). This procedure requires an independent way to determine the set of phase heads in Yorùbá, plus a metrical theory of the toneme capable of representing syntagmatic strength. Both ingredients are at hand, and their conjunction yields the observed effects. The phase heads of Yorùbá are C (universally) and T (parametrically), and H tone is the structural head of the metrical foot [s w] (Manfredi 1995, 2009). One lexically spurious H appears per cyclic spellout domain, as evidenced in the examples below once per pair of brackets. The effect is signaled by boldface H in the tone transcription line. Italic, lowercase m occurs in the tone line of (2) to indicate a separate effect: lexical L can't be parsed before a branching complement (Déchaine 2001).In Standard Yorùbá, spurious H predictably fails to occur before the set of modals — descriptively, the negative and future auxes — and this behavior is predicted because these items are transparently defective as phase heads: they take non-nominative subject clitics and 3S pro‑drop (Oyèláràn 1982; Déchaine 1992).(1) [TP Iṣú [VP wù wọ́n ] ].
yam.T please 3P.ACC
'(The) yam appetizes them'
[MH [L H] ](2) [TP Iṣú wu ọmọ [CP ọ́n jẹ _ ] ].
yam.T please child C eat
'(The) yam appetizes a/the child'
[MH m MM [H M] ](3) iṣu [CP tí [TP wọ́n rà _ ] ].
yam.NOM C 3P.T buy
'the yam(s) which they bought'
MM [H [H L] ](4) iṣu [DP rí‑ rà _ ]
yam D‑buy
'yam‑buying'
MM [H L].
Exceptions exist, but not counterexamples. Spurious H is "optional" after kí, the subjunctive complementizer, and before á and yóò, the two future modals (Bám̄gbóṣé 1966, 1983). Optional subjunctive is no surprise (e.g. Quirk & al. 1985, 155), unlike the option of spurious H before á and yóò: these items can't be 'optionally defective' phase heads because they never take nominative subject clitics. Instead, the sporadic ability of á and yóò to trigger spurious H falls out from the conjunction of two derivation‑external (E‑language) properties: (i) the pragmatic ambiguity of the future with respect to the realis/irrealis distinction — recalling the inconsistent specialization of will and shall in Modern English (Fries 1925); and (ii) the phonetic fact that both á and yóò begin with lexical H tone, providing a target for phonetic anticipation (a noise factor). The non rule‑governed occurrence of spurious H before these two modals can therefore be treated as surface economy (Reinhart 1997), in other words as an optimization of the extra‑grammatical 'outerfaces' (Déchaine & Manfredi 1995). This OT‑like wrinkle contrasts with, and does not threaten, the lawlike character of spurious H tracking the phased spellout of syntax. ReferencesAwóbùlúyì, Ọ. [1975]. On "the subject concord prefix" in Yorùbá. Studies in African Linguistics 6, 215‑38.
Awóyalé, 'Y. [1983]. On the development of the verb‑infinitive phrase in Yorùbá. Studies in African Linguistics 14, 71‑102.
Bám̄gbóṣé [1966]. A Grammar of Yorùbá. Cambridge University Press.
———. [1971]. The verb-infinitive phrase in Yorùbá. Journal of West African Languages 8, 37‑52.
———. [1983]. On timeless sentences in Yorùbá. Journal of Nigerian Languages 1, 1‑16.
Bresnan, J. [1971]. Sentence stress and syntactic transformations. Language 47, 257‑81.
Chomsky, N. [2001]. Derivation by phase. Ken Hale; a life in language, edited by M. Kenstowicz, 1‑52. MIT Press, Cambridge, Mass.
Chomsky, N. & M. Halle. [1968]. The Sound Pattern of English. Harper, New York.
Cinque, G. [1993]. A null theory of phrase and compound stress. Linguistic Inquiry 24, 239‑98.
Déchaine, R.‑M. [1992]. Inflection in Ìgbo and Yorùbá. MIT Working Papers in Linguistics 17, 95‑120.
———. [2001]. On the left edge of Yorùbá complements. Lingua 111, 81‑130.
Déchaine, R.‑M. & V. Manfredi. [1995]. Cohabitations of syntax and phonology. HILP 2, Universiteit v. Amsterdam, 26 January.
Dobashi, Y. [2004]. Multiple Spell‑Out, label‑free syntax and PF‑interface. Explorations in English Linguistics 19, 1‑47.
Fries, C. [1925]. The periphrastic future with shall and will in Modern English. Publications of the Modern Language Association of America 40, 963‑1024.
Manfredi, V. [1995]. Tonally branching s in Yorùbá is [LH]. Niger‑Congo Syntax & Semantics 6, 171‑82. African Studies Center, Boston University. http://people.bu.edu/manfredi/NCSS6b.pdf.
———. [2009]. Morphosyntactic parameters and the internal classification of Benue‑Kwa. Historical Syntax & Linguistic Theory, edited by P. Crisma & G. Longobardi, 329-43. Oxford University Press.
Oyèláràn, Ọ. [1982/1992]. The category AUX in Yorùbá phrase structure. 15th West African Languages Congress, Port Harcourt/Research in Yorùbá Language & Literature 3, 59‑86.
Quirk, R. & al. [1985]. A Comprehensive Grammar of the English Language. Longman, London.
Reinhart, T. [1997]. Interface economy; focus and markedness. Studia Grammatica 40, 146‑69.
the location of this page is http://people.bu.edu/manfredi/writing.html
last updated 17 May 2013
