1Lopit (or Lodongie) is spoken by some 50,000 people in the Lopit Mountains in the Eastern Equatoria province of South Sudan, near Torit, not far from the border with Uganda. Along with relatives such as Maasai and Turkana, it belongs to Eastern Nilotic, and at a much greater remove to Eastern Sudanic (commonly classified, following Greenberg [1963], as part of the Nilo-Saharan macrophylum). Like the rest of South Sudan, the Lopit people have suffered the effects of many decades of war, forcing many speakers to become refugees in countries ranging from Uganda to Canada. This comprehensive grammar, the first ever written of Lopit, was accordingly produced entirely through work with refugee speakers, mainly around Melbourne in Australia, providing a way around the difficulty of travel in South Sudan. It describes a head-marking marked-nominative VSO language with complex verb morphology and pervasive ATR harmony. Like the rest of Eastern Nilotic, but unlike most Nilo-Saharan languages, it classifies all nouns into two grammatical genders, masculine and feminine, as shown by agreement marking; of these, the feminine gender is more common, with the masculine serving primarily for males and diminutives. It has a tripartite number marking system with singulative, plurative, and replacive marking, characteristic for much of Nilo-Saharan.
2The book opens with an introduction situating the Lopit people, briefly outlining their history and culture, and describing relevant patterns of bilingualism, notably involving Otuho (a close relative spoken on the plains) and Juba Arabic (an Arabic-based creole which is the lingua franca of much of South Sudan), as well as a minority of Tennet (Surmic, Eastern Sudanic) speakers in the north. The language is considered by speakers to have six dialects, forming a linear chain along the mountains; the primary focus of this work is the northerly Dorik dialect. The introduction then discusses its classification, gives an overview of relevant previous publications (among Lopit’s closest relatives, only Otuho has been described in any detail), presents a typological sketch of the language (see above), and outlines the research’s motivations and methodology.
3Chapter 2 describes the phonology in considerable detail, identifying an inventory of 27 consonants and 9 monophthong vowels. Voiced plosives occur word-finally only in specific verb inflections; historical linguists will wonder whether they have been restored analogically. Notable features include geminate approximants, a phonemic glottal stop realised only in prepausal word-final position, a contrast between /p/ and /f/, and a robust ATR harmony system with leftwards spreading of +ATR only. The treatment of tone is somewhat tentative, as the authors commendably acknowledge (p. 88). There are three phonemic tones in their analysis, High, Low, and Falling; tone can be shown to be lexically contrastive, though minimal pairs are rare, but seems to play a more important role in establishing grammatical contrasts, notably case marking and verbal aspect. The syllable structure is (C)(G)V(C), where G is a glide; geminates are possible word-initially as well as medially. No official orthography exists, but a working orthography, not marking tone, is used in storybooks printed for community members.
4Chapter 3 outlines the word classes of Lopit as analysed here. Only nouns and verbs are analysed as open word classes; others include pronouns, prepositions, conjunctions, adjectives, adverbs, ideophones, and attitude markers (also called discourse markers). There are very few true adjectives (e.g. máɾwànì ‘old’, ŋὲɟʊ̀k ‘new’, rʊ̀xʊ́lànɪ̀ ‘clever’); the productive strategy for expressing properties is verbal, with attribution normally handled using relative clauses. Five of the eleven conjunctions found are borrowed from Juba Arabic, while three of the remainder are grammaticalised verb forms. Demonstrative, relative, and sometimes interrogative pronouns are marked for gender, while personal pronouns are not.
5Chapter 4 examines the noun phrase, whose default order, in conformity with Greenberg’s Universal 20, is head – possessive/demonstrative/numeral – adjective – relative clause – prepositional phrase (though numerals and adjectives may precede the head noun in some usages). Unlike most Eastern Nilotic languages, Lopit does not mark gender overtly on most nouns, and has feminine as the default gender. A normally feminine noun can take masculine agreement to indicate small size; this too is unusual for Eastern Nilotic, where masculine is usually used for large size. The number-marking system is of extraordinary morphological complexity, and will undoubtedly require further analysis. Like many Nilo-Saharan languages, Lopit has a tripartite marking system of singulative (e.g. xàlá-tí ‘tooth’ – xàlá ‘teeth’), plurative (xɪ́rɪ́ ‘waterhole’ – xɪ̀rɪ̀-ɟà ‘waterholes’), and replacive (xúŋ-ú ‘knee’ – xʊ̀ŋ-à ‘knees’), depending on the noun; a wide range of markers are found for each, with some variation across speakers for certain nouns. An attempt to identify default plurative marking strategies using loanwords and nonce words (p. 129) revealed some correlation between the phonological shape of the singular and the choice of plural marker; however, I wonder how much of a role the similarity between English -s and Lopit -sɪ/-sɪn played in influencing suffix choice for the loanwords. A typologically unusual feature is the use of morphologically singular forms in place of the plural to indicate an exceptionally large number, labelled the “greater singular”. Case (absolutive vs. nominative) is marked by tone change, most often specifically tone inversion but often seemingly unpredictable and lexically specified. Nouns are very productively formed from verbs using a variety of circumfixes. Numerals 1-4 are marked for gender (optionally except in the case of 1); the numeral system is decimal but retains traces of base 5. ‘One’ can also be used as “a kind of” indefinite article (probably specific). The spatial demonstratives are person-oriented; a tripartite paradigm of temporal demonstratives is also identified, although these seem to be transparently formed from the genitive/relative markers plus appropriate temporal adverbs. The possessive pronoun paradigm seems oddly reminiscent of North Omotic, with 1Sg ìtí and 2sg ìnó.
6Chapter 5 presents the verb, focusing on morphology. Like other Eastern Nilotic languages, Lopit has two verb classes, one consonant-initial (Class I), one starting with a synchronically uninterpretable ɪ- (Class II). Verb roots appear to be maximally disyllabic (including the ɪ-, though the authors do not rule out as yet unattested *ɪCVCV). The verb can be marked more or less agglutinatively for person, number, aspect, mode, voice, directionality, and applicative derivation, with the relevant morphemes assigned specific positions in a seven-part template. Infinitives are formed using five main strategies, depending on the verb’s class, the presence of derivational suffixes, and whether it is a property verb. Three motion verbs have suppletive forms for plural subjects (labelled “pluractional”), while partial reduplication of verb roots is productively used for repeated or habitual actions (confusingly not labelled “pluractional”, despite matching the term’s original definition at least equally well; cf. Newman [1990]). Subjects and non-3rd person objects are obligatorily marked by prefixes on the verb, with portmanteau morphemes for combinations of 1st/2nd person subject and object. Apart from causatives (which are marked by prefixes), verbal derivation is handled by suffixation, almost entirely non-recursive; it serves to change valency or to express direction. For the latter, ventive and itive markers are also used to express associated motion and several other semantic extensions.
7Chapter 6 covers aspect and modality (distinct, in the terminology adopted, from mood). Tense is not grammaticalised, and neutral aspect can refer to events in the past, present, or future. Aspect is marked affixally and tonally, with allomorphy depending on the verb class; the primary contrast is neutral, imperfective, and perfective, with other aspects including continuative, habitual, and persistive. Inchoative (change of state) is marked on property stative verbs through any of four distinct strategies, two of which involve directionality suffixes. Only two result state verbs are attested with inchoatives; the authors commendably note “More research is required to understand this”. Modality includes irrealis, potential, conditional, and obligative, all marked prefixally; there are also several lexical modal verbs, including the Arabic loan ixeder ‘be able’. Imperatives and hortatives are marked with portmanteau prefixes, including pronominal marking, and can take aspect marking; for some verb classes, singular and plural imperatives are distinguished only by tone.
8Chapter 7 addresses basic sentence structure. The unmarked word order is VSO, as usual for Eastern Nilotic. VOS order is said to be usual when the object is higher in prominence than the subject, though it is striking that the examples given each involve a subject phonologically longer than the object. Fronting is also found in pragmatically appropriate cases, yielding SVO/OSV. Prepositions are used to mark various obliques or adjuncts. Valence-increasing affixes on the verb — causative, dative, and instrumental — are highly productive; valency-decreasing ones include detransitivisation and middle voice, also marked tonally. There is no passive construction. Reflexives and reciprocals are typically formed with the grammaticalised noun kwan ‘body’. Predicate nominal and existential constructions use copular verbs ra ‘be’ and wːɔn ‘exist’, although verbless existential and locative predicates are possible with demonstratives. Predicate possession is typically expressed using existential wːɔn and comitative xo. Interrogatives –polar or content– add a verbal prefix x- (replaced by subordinate l- if the question word is clause-initial). Standard negation is handled with an auxiliary verb ɲa, with the negated verb following the subject and marked with subordinating l-. This ɲa does not inflect for person unless a modal marker or subordinator is prefixed; it seems to be a reduced form of the negative existential íɲà. It alternates with a different “persistent” negative auxiliary ɲei ‘no longer’. Negative imperatives (prohibitives) use a different construction with a verb idek ‘leave, abandon’ and a prefix xaɪ-.
9Chapter 8 is dedicated to the ways in which property concepts and adverbial notions are expressed. In addition to the primary strategy of using stative verbs, and the less prevalent strategy of nominal adjectives discussed earlier, there are three other minor strategies for property concepts: “relativised adjectives” (forms that take the relative marker in attributive function without displaying other defining traits of relative clauses), “gender-coded adjectives” (used attributively with the gender prefixes, always in reference to people), and colour adjectives apparently borrowed from Otuho and unrelated to the semantically corresponding stative verbs. Comparative constructions (as is typologically usual outside of western Eurasia) mark the standard of comparison (with táxɔ̀ ‘from’) but not the property word, while superlatives use a nominal form. Gradation or similarity (‘-ish’) is handled with inchoative strategies: the itive suffix for colours, and the inchoative perfective ɪ- for other property concepts. A limited number of underived adverbs are used to encode temporal reference, location, manner, and degree, also encoded (unsurprisingly) as prepositional phrases. A distinction is proposed between ideophones proper and ideophonic adverbials, whereby the former are usually utterance-final and the latter occur in a wider range of syntactic positions. “Adverbial concepts” (in context, this seems to mean concepts of temporal reference) may also be encoded as verbs.
10The lengthy final chapter, 9, is devoted to clause combining constructions. While simple juxtaposition can fulfil a variety of functions, more usually a non-clause-initial verb takes a prefix: sequential x- to indicate continuity with a previous event, persistive lV- for ‘while’, subordinate l- in extraction contexts such as relatives, contrastive focus, and clause-initial WH-forms, as well as conditionals. Adverbials or other words can also be used for clause combining; for instance, coordination of clauses can be handled by xɔɟɔ ‘and; then’ < ‘and they say’, while adversative coordination tends to use (aɾa) xatɪ (Arabic asan is sometimes used for ‘because’). Comitative constructions are unexpectedly also addressed in detail in this chapter, presumably by analogy with coordination; they include both “normal” and inclusory ones, with the latter corresponding better to nominal coordination proper. Complement clauses are then described, often using a complementiser (xɪɟɔ) transparently derived from the verb ‘say’, but also using juxtaposition, non-finite forms, or verbal nouns. Logophoric subject markers may be used in complement clauses, taking the same form as the 1st person subject markers (their status can be demonstrated by the fact that they can co-reference 3rd person pronoun subjects within the same complement clause). Relative clauses, which follow their head noun, are introduced by a relative pronoun agreeing in gender and number with the head, and marked for whether or not the conversation participants have a shared knowledge of the referent. Relativisation is possible on subjects, objects, obliques, and possessives, sometimes with resumptive pronouns in the latter two cases. Adverbial clauses often just use the subordinating prefix l-, but may also be formed with various adverbial subordinators, such as nafa/ifa ‘when’, or may be formed as relative clauses with an appropriate head noun such as ‘place’. Purpose clauses use xɪ̀ɟɔ́ ‘so that’, with distinct imperative-like verb forms in the 1st and 2nd persons. Secondary predication is handled with adverbial constructions. Conditional clauses are formed in three main ways: non-counterfactual ones may use subordinating l- or the conjunction lɔɟɔ ‘if’, while counterfactuals use specific modal forms without either of the previous. Conditionals are also sometimes found with no specific marking. The book ends with appendices including a 207-word Swadesh list, a couple of glossed narratives, and a list of data citations for chapter 2, followed by references and a short index.
11This well-constructed grammar will undoubtedly be an indispensable reference for students of Nilotic or Eastern Sudanic languages, and a useful resource for typologists of any stripe. It makes effective use of cross-linguistic typological work to put the observed phenomena in a wider perspective and provide consistent, easily cross-referenced terminology, though the occasionally conspicuous influence of the typological literature on the analyses themselves creates a slight risk of circularity. Its numerous comparative notes on Eastern Nilotic are particularly helpful in giving a sense of how the language might have developed over time. Eastern Nilotic historical linguistics is already relatively advanced compared to most supposed branches of Nilo-Saharan, as exemplified by the reconstructions of Vossen (1982), or by Heine & Claudi’s (1986) studies in grammaticalisation; such comparative notes will facilitate the process of updating historical work to take account of this extensive new data. All audio and video data collected for the project has been archived on PARADISEC (Billington & Moodie 2018), allowing in principle for eventual better testing of the proposed analyses. Elicitation has evidently played a key role in the production of this grammar, but more naturalistic text types have also been used extensively, and a number of different speakers were consulted. An in-situ documentation of the language in South Sudan would in principle have produced more representative results, but seems unlikely to be feasible for the foreseeable future; the authors are rather to be commended for having successfully produced this work despite the obstacles to fieldwork.
12This book does, however, suffer from one serious drawback: its eye-watering price, which, though standard for the series, appears difficult to justify in this day and age even for a well-bound and well-edited hardback. €170 is apparently almost half a year’s salary for a teacher in South Sudan (though one hopes that alternative publishing arrangements will be made available there); it is a high enough figure to exclude most would-be buyers even in Europe or America. Such a price effectively guarantees that only well-funded institutional libraries with a particular interest in linguistic typology or African studies will consider purchasing the book –and, ultimately, that it will not be read (at least not legally) outside a tiny circle of people concentrated in places like Leiden and Melbourne. All this at a time when the entire text could have been distributed online for free (as the two theses on which it is based indeed already are), and when grammars of comparable length, such as Friesen et al. (2017), can be printed on demand for €50. Fifty years ago, to publish a book, at almost any price, was to make it more widely and easily available than if it had been left in manuscript form. But today, with prices like this in an age of widespread cheap broadband, an outside observer might be forgiven for supposing that the purpose of academic publishing was to prevent specialist works from falling into the hands of laymen.