1Chadic is a highly diverse family of nearly 200 languages spoken in West Central Africa, with a distribution centred on Lake Chad. It constitutes a primary branch of Afroasiatic, along with Semitic (e.g. Arabic, Hebrew), Berber (Amazigh), Egyptian, Cushitic (e.g. Oromo, Somali), and arguably Omotic in southwestern Ethiopia. The best-known Chadic language is Hausa, a widespread, rapidly expanding, and well-studied lingua franca spoken by probably more people than any other African language, and enjoying a prominent literary tradition. All other Chadic languages, however, are spoken by relatively small minority groups. Despite significant progress over the past fifty years, most are still poorly documented, and sometimes entirely undocumented; many are threatened with extinction. Deeper research on Chadic is essential if any serious progress is to be made on the difficult problem of Afroasiatic reconstruction, and — as highlighted by the primary focus of this book — appears to hold a special interest for the study of phonological typology.
2Chadic is divided into four branches: West Chadic in northern Nigeria (including Hausa); Central Chadic from the shores of Lake Chad to the Mandara Mountains; Masa, further south of Lake Chad; and East Chadic, in central and southern Chad. The book reviewed here addresses the historical phonology of the Central Chadic branch, remarkable in particular for its minimal vowel phonologies and complex prosodies. While the author has extensive experience with Lamang in particular, the book rests “exclusively” (2.4) on the rich database prepared by Richard Gravina (2015) as part of a PhD. thesis (2014) on the reconstruction of proto-Central Chadic, covering 66 of this family’s 79 known members. This database, freely available online, supplements published work with SIL survey data and various linguists’ unpublished fieldnotes.
3However, while using the same data as Gravina and adopting his consonant reconstruction almost wholesale, Wolff takes a notably different approach to vocalic reconstruction, for reasons first summarised in Chapter 1 (“Introduction”). The reconstruction of vowels — called “at least doubtful, if not entirely impracticable” by Jungraithmayr & Shimizu (1981: 45), and “the most difficult or even an unsolvable task” by Stolbova (2016: 45) — has posed persistent difficulties for Chadic studies due to an appallingly irregular-seeming plethora of surface-level correspondences (1.1). The gradual recognition of the important role of suprasegmental prosody, in particular the long-distance spread of rounding and fronting from consonants to vowels, has provided Chadicists with better tools for addressing this problem. In Wolff’s view, as I understand it, the sometimes tentative vowel reconstructions proposed by Gravina do not take this far enough, applying the insights of prosody theory where unavoidable but not following them through to their ultimate implications.
4This book is accordingly largely dedicated to working out the consequences of one big idea about Central Chadic vocalic phonology, summarised in Section 2.5 (after necessary extralinguistic, typological, and methodological background, 2.1‑2.4) under four headings. Epenthesis theory is the hypothesis that schwa — along with, of course, all its myriad synchronic allophones — was entirely epenthetic and predictable in proto-Central Chadic, as it is today in languages such as Lamang, and as such should not be reconstructed phonologically. Weak radical theory extends the observation that underlying /y/ and /w/ often surface as [i] and [u], allowing apparent vowels to be reinterpreted as “weak” consonants, much as in traditional analyses of Semitic. Vocalisation theory is the assumption that roots are consonantal, and any given root consonant except the last may variably be followed either by /Ø/ or by /a/; this implies that a single triliteral root may yield up to four distinct reflexes: CCCa, CaCCa, CCaCa, and CaCaCa. Prosody theory extrapolates from the synchronic observation that across Central Chadic rounding and fronting behave as floating prosodies, spreading leftwards from specific consonants to affect multiple segments, to account diachronically for otherwise problematic cases of rounding and fronting as the result of lost segments (often suffixes, producing “augmented” roots). Combining these approaches allows proto-Central Chadic to be reanalysed as a language with only one phonemic vowel, /a/ (2.6) — whose distribution is however often variable within a single root in seemingly unpredictable ways — and a complex system of “augmental” morphology discussed in more detail in Chapter 3. In accordance with this focus, Wolff reconstructs proto-forms directly to proto-Central Chadic rather than, like Gravina, first reconstructing for its subfamilies (2.7). No attempt is made to reconstruct tone (2.6.3).
5Chapter 3 addresses “Proto-Central Chadic diachronic phonology and morphophonology”, focusing primarily on how the latter can be used to explain the former. Variation in the presence or absence of /a/ in a given position in a given root — posing serious problems for any rigid application of the comparative method — is accounted for as reflecting fossilised root-and-pattern morphology (3.1.1), comparable to that found in other Afroasiatic languages and especially in Semitic. Seemingly irregular variation in prosody and in initial or final consonants is similarly accounted for as reflecting “petrified root-augmentation material” — synchronically non-productive affixes (or reduplicants) to the left and right of a reconstructible root (3.1.2). This allows quite striking levels of variation; ‘work’ is thus reconstructed as *(a‑, ka‑, ma‑)ɬ(a)na(‑t, ‑y, ‑k, ‑kʷ, ‑n, ‑l; final vowel), with only two variants of the root proper but four attested choices of prefix (including null) and seven attested choices of (often composite) suffix, some of which could probably be eliminated by better synchronic data on the source language’s morphology. No attempt is made to identify any original grammatical functions for such fossilised morphology, except insofar as some of the suffixes are taken to derive from proto-Chadic determiners (3.3). “Root augmentation” allows originally shorter roots to become longer in descendant languages, while segment loss in turn allows originally longer roots to become shorter (3.1.2). Metathesis, which “occurs quite frequently in the data” (3.1.3), also complicates the application of the comparative method. Roots are assumed to end in *a by default, despite some synchronic variation interpreted as reflecting apocope (3.4). Loanwords, mainly from Kanuri, are given proto-Central Chadic “pseudo-reconstructions” despite not dating back to proto-Central Chadic, in order to illustrate their adaptation to Central Chadic phonology (3.5); along with the rejection of subfamily reconstruction, this implies that the reconstruction of a given root here should not be used as evidence for its age.
6Chapter 4, “Diachronic processes in Central Chadic language evolution”, zooms in on the ways in which originally allophonic variation has been variably phonemicised: through schwa epenthesis and prothesis (4.1), through prosodic “colouring” of the resulting schwa and of *a (4.2‑4.3), through syllabification of underlying semivowels or monophthongisation of diphthongs involving them (4.4), and through fusion of originally distinct adjacent segments (4.5). Epenthetic schwa may in turn combine with semivowels to yield synchronic long vowels (4.8). These processes may result in remarkable surface difference not only across the family but even across dialects of the same language (4.6). Consonantal changes are briefly summarised in 4.9, largely following Gravina, but are not a primary focus of the present work.
7In Chapter 5, “Central Chadic languages and the Neogrammarian hypothesis”, the author highlights the ways in which his apparently justifiable assumptions about (fossilised) pattern morphology and suprasegmental prosodic phenomena combine with the relatively limited size of the dataset to make assumptions about Neogrammarian regularity of sound change “practically impossible to ascertain”. 5.1 illustrates this across two closely related (probably mutually comprehensible) varieties particularly familiar to the author: Lamang and Hdi. Seemingly irregular correspondences between the two are pervasive; however, the assumptions previously outlined about fossilised morphology and its relict prosodic affects allow these to be accounted for, not regularly but at least systematically. A (perhaps slightly caricatural) mechanical application of Neogrammarian methods, looking only at synchronically present surface segments in synchronically similar adjacent contexts, would create a proliferation of misleading phantom correspondences, with the problem getting worse as more distantly related languages are compared (5.2).
8The remainder of the book is dedicated to “Full lexical reconstructions” (Chapter 6). It opens with a more detailed examination of the ways in which his reconstruction of the vowel system diverges from Gravina’s: in explicitly rejecting phonemic *ə and the idea of vowels moving between slots (6.1) and in rejecting phonemic *i, whose low frequency in Gravina’s reconstructions he takes to be suspicious (6.2). Section 6.3 illustrates the consequences of these differences with a variety of specific examples, as well as examining the problem of where synchronic glottalisation and /ʔ/ come from. 6.4, finally, dives into the details of 64 selected lexical reconstructions, covering about a quarter of the items in Gravina’s database and specifying, item by item, how observed variation going beyond regular correspondences can be explained in terms of Wolff’s model. This section is best treated as an appendix to be referred to, rather than as a text to be read sequentially. The remaining lexical reconstructions will apparently be addressed in a successor volume with the same publisher: Lexical reconstruction in Central Chadic: A comparative study of vowels, consonants and prosodies.
9This work poses interesting challenges for historical linguistics. All the phenomena proposed here as complicating the identification of regular correspondences appear individually plausible. Rounding and fronting prosodies are synchronically pervasive in these languages, while non-concatenative pattern morphology remains pervasive across much of Afroasiatic (including a few Chadic languages), and fossilised morphology (affixal and otherwise) is a recurrent phenomenon worldwide, well-known from the history of Chinese for instance. In the absence of any good account of the original functions of this fossilised morphology, however, these combine with sporadic metathesis and apocope to allow a degree of freedom in comparison that should give any historical linguist pause. Under such circumstances, how sure can we be that any given reconstruction does not reflect chance resemblance, even with the laudably narrow semantic latitude adopted here and by Gravina? If these hypotheses are accepted, narrowing down the original functions of these “augments” and “patterns” necessarily becomes a vital task for Chadic historical linguistics; only thus can comparisons be made more restrictive. Yet the bitter experience of work on Semitic root extensions suggests that, even in a much better documented family whose written history goes back nearly five millennia, seemingly obvious fossilised morphology can persistently defy systematic analysis.
10Whether or not such a goal turns out to be feasible, Gravina’s 250 lexical sets — let alone the 64 which Wolff selects here — are clearly not enough to allow its full attainment. Far more extensive Central Chadic lexical data will be needed to allow the reconstruction of more proto-roots, and detailed descriptions of synchronic morphology — still unavailable for most Central Chadic languages — might well reveal that some of Wolff’s proposed affixes remain productive somewhere. Better documentation of these languages is thus all the more desirable for the purposes of historical linguistics, as well as in its own right. Ideally, the two should be mutually complementary; a linguist who, like Gravina or Wolff, has done fieldwork on Chadic languages starts out in a much better position to understand their historical development. However, the difficulty of undertaking fieldwork on these languages under present circumstances should not prevent linguists from continuing to seek a deeper understanding of their diachrony, based on recorded materials and ex situ elicitation if necessary. As this work illustrates, such an understanding may have broader consequences for the question of how a proto-language can ultimately become unreconstructible even without lexical replacements.