Current Progress in Chadic Linguistics: Proceedings of the International Symposium on Chadic Linguistics, Boulder, Colorado, 1-2 May, 1987.

Wolff, H. Ekkehard

This volume contains fifteen papers from the third (not fourth, as the editor maintains in his introduction) of its kind Chadic linguistics meeting in Boulder, Colorado, in 1987. The topics of the papers tend to show the increasing attractiveness of Chadic language structures for general linguistic discussions in recent theoretical frameworks on syntax. Yet the major issue remains the diachronic perspective in which many authors place their contributions, thereby showing how Chadic linguistics has come of age.

Syntactic questions are dealt with in six contributions. L. Tuller ("Variation in Focus Constructions," pp. 9-33) establishes three types of FOCUS (=focalization and question) constructions in Chadic: in situ, pre-sentential, postverbal. Her paper focuses on the postverbal type which displays some interesting intra- and interlinguistic variation. The issue of a general postverbal FOCUS position is the more fascinating when we remember that Chadic contains languages with both SVO and VSO neutral word order. T. argues within the Government and Binding framework. For postverbal FOCUS she dismisses both ideas that +FOCUS could be either assigned to a sentence-final /SPEC,CP/ position or that +FOCUS is assigned by the verb. Rather, she proposes that INFL (I) functions as a FOCUS assigner when it contains a raised verb. This cleverly accounts for phonological intra-language differences found between V+DO and V+FOC-DO constructions! Inter-language variation of V+FOC+DO vs. V+DO+FOC is then explained in terms of Case Theory. As interesting as the paper may be because of its theoretical implications, it should be pointed out that it is also extremely interesting from descriptive and comparative points of view within Chadic. It is, to the best of my knowledge, the best contrastive overview of such constructions that has been published hitherto for this language family.

Working in the same theoretical paradigm as Tuller, B. J. Johnson ("Case Assignment in Hausa, Kanakuru, and Ngizim," pp. 35-54) uses data from three Chadic languages to take issue with some notions within Case Theory. She rejects inherent case, assumes that governors assign case only once, and claims that NPs may surface case-free, thus rejecting the Case Filter as proposed within the received case framework. The arguments are based on the observation that dependent ("accusative case") pronouns are replaced under certain conditions by ("case-free") independent pronouns. J. relates this to the fact that in all cases studied the (independent) pronoun is no longer governed by the verb and therefore cannot be assigned accusative case (semantically, however, it still functions as the direct object of the verb!). Independent pronouns are also used in non-verbal sentences: Again, "nominative" case cannot be assigned (since there is no verb present). J. then tries to stretch her analysis to cover the continuous ("imperfective") paradigm which--among other things--tends to be characterized by the nominalization of the verb. At least in the case of Hausa J. runs into serious trouble here because she bases her arguments on oversimplified descriptions, not taking into account the far-from-clear distinction between "primary" and "secondary" verbal nouns.

R. C. Marquis ("Word Orders in Gude and the VSO Parameter," pp. 55-86) sets out to explain why Central Chadic Gude allows two word orders: VSO and SVO, and tries to do so within the Government and Binding paradigm. Like most papers of this kind (including the two just reviewed) the reader is first of all told which analyses are wrong for whichever reasons within the chosen theoretical framework before he is allowed a glimpse of the truth, i.e., the author's solution to the dilemma (at times self-inflicted, as it will appear). M. lengthily reports on two different theories relating VSO order to underlying SVO order (i.e., Sproat 1985, Koopman 1985) before he gets around to presenting his own alternative solution which does not make crucial use of case theory. His solution introduces the notion of Propositional Government, a well-formedness condition parameterized for the SVO/VSO distinction which could be also of interest to linguists studying V-second phenomena in Germanic. As for Gude itself, the Chadic language under analysis, we learn rather little.

G. J. Dimmendaal offers a taxonomic account of "Complementizers in Hausa". He discusses syntactic and semantic characteristics of sentential complementation in Hausa, focusing on the use of complementizers. After examining the matrix clauses which take complementation, D. identifies three complementation types: VP-type, S-type, and S-bar type. He then illustrates the following complementizers: ceewaa, wai, koo, kada, and don. In the final and very stimulating section of his account, D. comments on the interaction of matrix clauses, subordinate clauses, and complementizers. Verbs differ and/or allow doublets of "form" (Parsons' A-, C-, D-forms) before sentential complements, and the latter may vary in type. Here semantic compatibility comes in, and D. therefore tentatively proposes a semantic analysis of the five complementizers under review. D. raises some highly relevant questions that go far beyond Hausa and pertain to Chadic in general--if not to universal grammar. The notion of "degree of tension" between a verb and what follows (already hinted at by Parsons 1960) may prove to be of particular relevance in explaining certain formal properties of the (matrix) verb in Chadic.

K. Williams suggests "An Alternative Model of Word Order in Proto-Chadic". He challenges Frajzyngier's (1983) reconstruction of Proto-Chadic (PC) as being of VSO type. He suggests that PC had two basic word orders: VS (for intransitive constructions) and SVO (for transitive constructions), i.e., nominal arguments with the thematic role -AGENT would follow the verb, whereas those with +AGENT would precede it. This ingenious idea is very attractive: it allows us to have our cake and eat it! But what about SV intransitive and VSO transitive constructions in Chadic? W. has another magic device up his sleeve: analogical leveling! Since this overwhelmingly powerful "alternative model" hardly leaves any questions unanswered, W. feels compelled himself to direct our attention to what "looks to be a profitable area of exploration": "the link between word order and aspect." The reader is grateful for this suggestion but remains a trifle irritated: didn't the reviewer's Grammar of the Lamang Language (1983: 188ff.), among other things, establish just that? (But W. is obviously not aware of all the relevant literature on his topic.)

M. M. Garba takes "A New Look at the NP+naa+NP Constructions". The nature and history of the morpheme naa has indeed induced linguists of all kinds to come up with the most creative and at times astonishing hypotheses. Garba's surely is the most fanciful I have come across lately. Some Hausaists may agree to view naa as a (in Hausa usually pre-verbal) "tense/aspect marker"--but where is the verb in NP+naa+NP? Garba argues ex nihilo: There used to be one, but now it is gone. And which would have been the verb? Simple answer: nan--an "existential verb" as in examples like Laadi ta naa nan "Ladi exists." Would you believe it?

Three papers deal with the morphology of individual languages. B. Caron ("The Verbal System of Ader Hausa," pp. 131-69) introduces the reader to the spread of some dialectal differences in the conjugational paradigms of northwestern Hausa verbs in a contrastive manner, using "Standard Hausa" as a point of reference. In the second part of the paper, C. points out some peculiarities of Ader Hausa in the derivative subsystem ("verbal grades" in traditional terms, "verb classes" in C.'s terms). Some of the "grade forms" in this dialect are likely to upset some ideas dear to Hausaists who base their work solely on Standard Hausa. This is particularly so for the notoriously "difficult" grades 2 and 5. Unlike Standard Hausa, Aderanci has a productive directional suffix -k- which combines with both grade 4 and grade 5. Also, the use of the verbal noun suffix -waa in Ader Hausa shows some very interesting idiosyncrasies, both tonally as well as syntactically (at least with speakers of a particular sub-dialect).

R. G. Schuh ("Gender and Number in Miya," pp. 171-81) provides us with a minute account of how gender and number marking and agreement function in West Chadic Miya. Interestingly, the semantic feature animate (mainly designating humans and most domestic animals, also some large wild animals--thereby correlating with sex in the real world) vs. inanimate plays a very important role in gender and number agreement. The noun itself carries no overt gender markers; plural, of course, is morphologically marked. As for number, Miya animate nouns are obligatorily marked for plural in the appropriate semantic environment, whereas for inanimate nouns plural marking would be optional in the same environment. In terms of gender, however, animate nouns, when marked for plural, show number agreement (and gender plays no role here), whereas inanimate nouns, even if morphologically plural, show lexical gender agreement (but no number agreement!).

V. de Colombel ("Origine de l'extension verbale (|epsilon~)r(|epsilon~) instrumentale et connecteur, en ouldeme, synchronie dynamique et diachronie," pp. 183-97) illustrates what she takes to be a single derivative morpheme in a Central Chadic language (Udlam) which apparently has a wide range of meaning nuances (instrumental, focus, subordination with different nuances) which, as she claims, is unique in Chadic. She attempts to relate it to a morpheme of similar shape in Ancient Egyptian (but of rather different meaning: future, directional). Her claim is untimely and rather far fetched, because, on the basis of the data and analysis presented, it cannot be considered an established fact that we are dealing with only one morpheme in Udlam, and--in the absence of any systematic comparative study on derivational markers in Chadic--no attempt is made to establish sound and meaning correspondences between Udlam -r- and other possible reflexes in Chadic (simultaneously allowing for semantic changes, of course). The author adheres to vague notions of an implicit idiosyncratic method of "synchronie dynamique" which does not make the reading any easier.

Three more papers deal with phonology. W. R. Leben ("Intonation in Chadic: An Overview," pp. 199-217) reports on experimental studies concerning question intonation contours in several Chadic languages measuring FO values. L. distinguishes (a) phonetic phenomena, such as "Global Raising" in Hausa questions, and (b) phonological phenomena of two subtypes: (i) changes on the lexical tonal tier, such as the addition of a low tone morpheme at the end of yes|is similar to~no questions in Hausa, and (ii) addition of register high tones (with, as it will appear, different extensions to the left from sentence-final position in the different languages). L. is content to see that intonation appears to be rule-governed in these languages and thus constitutes part of an acquired system of some complexity.

C. Hodge ("Hausa and the Prothetic Alif," pp. 219-32) provides more data to support his (old) diachronic idea that word-initial CC-clusters are "resolved" by different strategies: vowel insertion, 'V- prothesis, or simply loss of the initial C. He allows also for "secondary developments," among them the existence of an N (nasalization) affix (cf. the paper by Frajzyngier and Koops). As a result, prothetic material "disguises" the original shape of the root in Hausa. Once we strip away such prothetic material, it will be possible to discover more possible cognates elsewhere in Afroasiatic (and even beyond, for instance in H.'s "Lislakh" macro-phylum).

Z. Frajzyngier and R. Koops ("Double Epenthesis and N-Class in Chadic," pp. 233-50) draw our attention once more to what looks like a "lexical class" of nouns which is characteristically marked by a prefix N- and designates some animals and body parts, yet by no means all. As far as animals are concerned, similar observations can be made for Kanuri and Songhay--and, of course, everybody immediately thinks of Class 9 in Bantu. F. and K. consider early borrowing from non-Chadic sources the most likely explanation. At the same time, the paper takes issue with ideas developed by P. Newman (1976) and C. T. Hodge (1986, 1987, present volume) about the relative chronology and phonological status of initial (prothetic) vowels and--later and in addition--consonant prothesis (h or ') which result in initial non-lexical CV-sequences, all motivated by the tendency to reduce initial CC-sequences of the stem. In the "worst" case, a Chadic noun may contain up to three non-lexical segments: C-V-N-. This will give comparatists something to chew on! In a wider comparative perspective, of course, one wonders whether the "class prefix" n (of highly doubtful semantic content) and the "prothetic" postvelar C could have anything to do with the ubiquitous and notorious "N/T/K" morphological elements of Nilo-Saharan, and how all this relates to the Proto-Chadic determines *n/*t/*k etc. (reconstructed in Schuh 1983).

Two papers address diachronic issues of relative chronology within Chadic and Afroasiatic. H. Jungraithmayr ("Is Hausa an Early or Late Stage Chadic Language?" pp. 251-66) calls on Diakonoff's (1965, 2nd ed. 1988) slightly dusty topological distinction between "Ancient, Middle, and New Stage" languages within Afroasiatic and relates it to his own ideas about "four stages" in the historical development of Chadic languages in terms of verbal morphology (first presented 1974 where J. proposed certain morphological criteria for classifying verbal systems into chronological "stages": an ancient Stage I and progressing corruptions of that system, i.e., Stages II-IV). However, reading Diakonoff (1988: 17) diligently, one will note that he also says that "we shall assign to the Ancient Stage those languages ... which are closest ... to ... structures which can be reconstructed by the comparative method." As for Chadic, alas, we cannot claim that any structures have been reconstructed by the comparative method which are beyond dispute. So, how relevant can any typology based on "stages" be when we cannot be sure what the "Ancient Stage" really looked like in Proto-Chadic? (Who said it had to look anything like Proto-Semitic?) Regarding verb morphology, J.'s historical analysis of Hausa verbal nouns (nowhere does J. ever justify his equation of synchronic verbal nouns in Hausa with a hypothetical Proto-Chadic "imperfective" verb stem) leaves much to be desired. He assumes "innovation of a suffix -waa, e.g., kaamaa:kaamaa-waa," yet he illustrates his case with the verb "to die" (mutu:mutuwaa) failing to recognize that the final syllable of mutuwaa does not contain the same nominalizing suffix--'waa (the different tonal pattern alone is indicative enough). The paper culminates in a rather unhappy polemic against Newman (1977a, b), which is hardly supported by the contents of the paper under review; incidentally, J. also fails to mention, let alone discuss, several relevant publications on the issue (some by the present reviewer).

R. M. Voigt ("Verbal Conjugation in Proto-Chadic," pp. 267-84), attempts to point out what he considers to be a typological correspondence between some East Chadic (plus Hausa) conjugational verb paradigms and Cushitic verbal morphology (he goes as far as postulating a Chado-Cushitic branch within his "Semitohamitic"!). Not very convincingly in terms of the arguments and data presented, both lexical and inflectionally conditioned verb-final vowel qualities in synchronic Chadic systems are reanalyzed as reflecting older suffixed "auxiliary" (or: "indirect") conjugational devices--similar to those in Cushitic. Even if this eventually turned out to be so, I would say that the present state of comparative Chadic verb morphology simply does not allow statements beyond Chadic proper, since even within this language family (or its branches) there is little agreement between specialists as to the lexical, derivative and/or inflectional nature of some (if not all) of the verb-final vowels. Should we not first reconstruct a proto-system on Chadic evidence alone before venturing into the notoriously controversial discussion of the relative chronology of Afroasiatic conjugational verb paradigms? At present, I am afraid, little or nothing can be gained for solid Chadic-internal reconstructions from perpetually peeping across the fence into other families within the Afroasiatic phylum.

In the final paper, S. Baldi writes "On Semantics of Arabic Loan Words in Hausa". He presents a list of 124 Arabic loans in Hausa in order to whet our appetite for more: this particular list represents but 10% of the bulk of loans which the author has compiled from the available dictionaries. Rather vaguely, six semantic domains (nature, man as physical being, man as spiritual being, man as social being, social organization and politics, natural laws) and one syntactic class (interjections and particles) are mentioned, instances of semantic changes are pointed out, and some interesting phonetic and phonological points are illustrated. Finally, the list of examples is ordered according to the word class in which Arabic nouns, verbs, or particles are found in Hausa: noun:noun (0), noun:verb (10), noun:interjection (2), noun:particle (12), verb:verb (67), verb:noun (28), verb:interjection (1), verb:particle (2), particle:noun (2). As it is presented, Baldi's list is little more than anecdotal--we look forward to the full inventory and its systematic presentation.