1The book Ikpana interrogatives by Jason Kandybowicz, Bertille Baron, Philip T. Duncan and Hironori Katsuda is a comprehensive study of interrogative constructions in Ikpana (previously better known as Logba), a Ghana-Togo Mountain (henceforth GTM) Kwa language of the Na-Togo group. Based on targeted elicited data supplemented both by recorded texts and by experimental data, the authors analyze the morphosyntax and prosody of polar and wh-questions. The authors follow the Chomskyan generative grammar paradigm and propose formal analyses of Ikpana data in this framework. At the same time, the book also serves as a typollogically informed description of Ikpana data.
2The book consists of six chapters.
3Chapter 1, “Overview”, is an extended introduction into the Ikpana language and the general setting of the study.
4Chapter 2, “Polar questions”, presents a description and an analysis of yes/no-questions. The main strategy used by Ikpana consists in final vowel lengthening, which is analyzed by the authors as a segmental interrogative marker. Except for yes/no-questions involving disjunction, vowel lengthening occurs in the last word of the clause. Different particles can be used to express additional pragmatic functions.
- 1 The following abbreviations are used in the cited examples (reproducing the glossing of the book un (...)
5Chapter 3, “Wh-questions”, is a very detailed account of questions about specific constituents. Like other GTM languages and, more widely, like many other Kwa languages, Ikpana allows for wh-constituents to occur in situ or ex situ. However, intriguingly, the latter option is not applied uniformly but realized in several different constructions. First, bare wh-movement with no additional marking is possible. Second, an ex situ wh-constituent can be followed by an overt focus marker. Third, an ex situ wh-constituent can be followed by a relative clause marker, which is analyzed by the authors as a cleft construction. Relative clause markers are widely attested in ex situ wh-questions, and in particular in GTM languages (cf. for example Allan 1973: 456 for Lelemi), but explicit analyses of such ex situ wh-questions as cleft focus constructions are rare (but cf. Rose et al. 2014 for Moro, a Heibanic language of Sudan). As a cleft analysis looks convincing for both Moro and Ikpana, one could hypothesize that this may also hold for other languages with wh-constructions involving relative clause markers. Fourth, the wh-constituent in the cleft construction can also combine with an overt focus marker (which is, in this case, followed by the relative clause marker). Further detailed research into ex situ wh-constituents in Ikpana leads the authors to two cross-linguistically remarkable observations. The authors note that the co-called superiority effect is not relevant for Ikpana’s multiple wh-questions – i.e. any wh-constituent can be placed ex situ or in situ independently of their syntactic position in the clause, and in the case that both are ex situ, they can be placed in any linear order, see (1a-b).1
(1)
|
a.
|
ɔ̀-mɔ̀
|
ɔ̀-mɔ̀
|
ɔ̀-tá
|
mɛ́
|
|
|
cm-who
|
cm-who
|
sm-give.pst
|
what
|
|
|
‘Who gave who what?’
|
|
|
|
|
|
|
|
b.
|
mɛ́
|
ɔ̀-mɔ̀
|
ɔ̀-tá
|
ɔ̀-mɔ̀
|
|
|
what
|
cm-who
|
sm-give.pst
|
cm-who
|
|
|
‘Who gave who what?’ (p. 76)
|
6While, as the authors mention, the absence of the superiority effect is not unique to Ikpana, in particular, it is known to exist in some other West African languages (cf. Saah 1994; Adesola 2006), it is still reported rather rarely. The authors hypothesize that the superiority effect is not relevant only for languages with cleft or focus wh-constructions, while those with only bare wh-movement are expected to have it. So far, we do not have a big enough language sample to falsify the claim, but it is a good illustration of why such detailed studies in underdescribed languages are useful. Importantly, different wh-expressions in Ikpana have different syntactic properties: ‘how’ questions do not allow pied-piping, as shown in (2a-b), while ‘why’ questions do (3).
(2)
|
a.
|
mɛ́
|
Kofí
|
ɔ̀-zá
|
a-zaì
|
ɛ-ta
|
|
|
what
|
Kofi
|
sm-cook.pst
|
cm-beans
|
cm-means
|
|
|
‘How did Kofi cook beans?’
|
|
b.
|
* mɛ́
|
ɛ-ta
|
ka
|
Kofí
|
ɔ̀-zá
|
a-zaì
|
|
|
what
|
cm-means
|
foc
|
Kofi
|
sm-cook.pst
|
cm-beans
|
|
|
Intended: ‘How did Kofi cook beans?’ (p. 66)
|
(3)
|
mɛ́
|
ɔ-kplɛ̀
|
ka
|
Fafa
|
ò-kpló
|
i-dzɔ́
|
|
what
|
cm-reason
|
foc
|
Fafa
|
sm-fry.pst
|
cm-yam
|
|
a-fàn
|
nu
|
u-dântʃì-e
|
|
|
|
|
cm-home
|
in
|
cm-morning-det
|
|
|
|
|
‘Why did Fafa fry yams at home this morning?’ (p. 55)
|
7Chapter 4, “Interrogative intonation”, contains the results of instrumental studies on Ikpana intonation – not only in interrogative but also in declarative sentences, in order to contrast the two. For both types, the authors start with informal observations done on isolated elicited examples, but still based on sample pitch tracks and not just on acoustic perception. They then proceed to controlled experiments with several Ikpana speakers. Since Ikpana, like other Kwa languages (and most West African languages), has lexical tones (three level tones, which is typical for GTM), an additional task is to take into account the interaction of the two different functions simultaneously expressed by the pitch level. One can conclude from the study that Ikpana follows the typical cross-linguistic pattern whereby interrogatives have a higher pitch level at the end than declaratives. However, the details are rather intriguing. In declarative sentences, one can observe a tone reversal of high and mid tones: the former is downstepped, while the latter is upstepped; the low tone is downstepped. In interrogative sentences, there are no downstep and upstep effects, which means that both high and low tones are produced at a higher level than in declarative sentences. Moreover, in yes/no-questions, a low final tone is raised. One more specific feature of yes/no-questions is the final vowel lengthening already considered by the authors in Chapter 2, but which is here analyzed as part of the interrogative intonation.
8Chapter 5, “Interrogative prosodic structure”, deals with vowel hiatus resolution in different types of constituent ordering. Here again, the authors deal with both declarative and interrogative sentences, and again start with informal observations done on isolated examples and then proceed to controlled experiments with statistically valid results (the authors mention that they had no opportunity to conduct all planned experiments because of Covid restrictions). Different strategies of hiatus resolution are evidence for different levels of syntactic integrity between the constituents. In particular, the authors show that so-called heterosyllabification (which means, no resolution at all, both vowels are kept) is most frequently observed on the boundary between a subject and a verb, while elision (where one of the vowels is simply omitted) is most common on the boundary between a verb and an object. In both cases, these strategies are several times more frequent than others, which means that this difference is not just statistically significant, but even self-evident. At the same time, it is clearly not by chance that the vowel elision strategy is observed between a phrase head and its complement, while the heterosyllabification strategy occurs between the two items that are more distant from another in terms of the constituent structure. Having done this preliminary work on core types of Ikpana constituents, the authors then analyze the prosodic structure of wh-questions in the same logic. Their research question concerns the degree of syntactic integrity between wh-constituents and the rest of the interrogative sentence, and the result is that it differs for in situ and ex situ wh-expressions. Internal wh-expressions constitute a single prosodic domain with the rest of the sentence and exhibit elision like verb and object sequences, as is the case of the wh-word ɔ̀mɔ̀ ‘who’ in (4), which loses its final vowel. In contrast, ex situ wh-constituents are prosodically independent from the rest of the sentence, and their usual pattern is heterosyllabification where a vowel hiatus remains unresolved; this can be illustrated by the item mɛ́ ‘what’ in (4), which keeps its vowel and is separated by a short pause.
(4)
|
mɛ́
|
ɔ̀-mɔ̀
|
ɔ̀-zá
|
[mɛ́...ɔ̀mɔ̀za]
|
|
what
|
cm-who
|
sm-cook.pst
|
|
|
‘Who cooked what?’ (p. 155)
|
9Finally, Chapter 6, “Ikpana interrogatives from a GTM perspective” provides a report of what is so far known about interrogative constructions in GTM languages in general. In the beginning of the chapter, the authors give an overview of topics treated in publications on GTM, and it is no surprise that researchers start with basic grammatical topics, such as phonology, nominal and verbal morphology and main syntactic constructions. Still, the available information makes it possible to see that, on the one hand, questions in Ikpana follow general GTM patterns, but that, on the other hand, GTM languages are not at all uniform. Final vowel lengthening in yes/no-questions is typical for the Na-Togo group Ikpana belongs to, but not for the Ka-Togo group. Featuring both in situ and ex situ strategies of wh-questions is reported for all but one GTM language for which accounts of interrogative constructions are available; Fiedler & Schwarz (2009) have shown that only in situ wh-questions exist in Lelemi. As for different kinds of ex situ constructions, GTM languages appear not to be uniform (and not described well enough), but, importantly, many Ikpana phenomena reported and analyzed by the authors have parallels in other GTM languages. Detailed information on interrogative intonation and other phenomena related to prosodic structure is most often absent in the existing descriptions; but if present, phenomena similar to those attested for Ikpana are reported.
10The book under review is a very important contribution both to the theory and typology of interrogative constructions and to the description of Ikpana, of GTM languages, Kwa languages and, even more broadly, of West African languages. It is an exemplary study that fits perfectly into the series Oxford Studies of Endangered Languages. First, the book provides detailed systematic data on Ikpana, and we can be absolutely sure that there is no other GTM language where interrogative constructions have been studied in greater depth, as the authors have shown themselves in the last chapter of the book. Second, it is a wonderful example of theoretically informed research in a specific domain of a language system. For any fact reported for Ikpana, the reader is given an explanation what exactly this fact adds to the typology of interrogative constructions and to syntactic theory. At the same time, it is the authors’ theoretical background that allows them to collect as many relevant facts and to answer as many specific questions about interrogatives in Ikpana as contemporary linguistics can ask. The authors repeat many times, starting from the introduction up to last section, that their theoretically based approach leads to fuller and more in-depth results than the limited accounts of interrogative constructions provided in descriptive grammars. In particular, they stress that their book reports more constructions and especially more facts about them than Dorvlo’s (2008) grammar of Ikpana. The authors’ frequent mention of the advantages of their study may be exhausting, but indeed, the book is a brilliant example of a focused, theoretically informed study.
11It is surely true that in order to make real progress in linguistic theory, we need such detailed studies of specific domains for languages that are hitherto underinvestigated. But before such studies are possible, progress in descriptive linguistics is still required – especially, in West African languages, where we are still very far from a situation in which every language has a good comprehensive grammar. The study under review shows that even a good descriptive grammar cannot report all the relevant details – and thus typological studies based on secondary data cannot provide a full picture either. At the same time, this study would never have been possible in a language that was not previously described to a fair degree. The authors stress their descriptive discoveries that were not reported by Dorvlo (2008), but it is also important to underline that they were only able to collect and analyze their focused data because they already had a good account of Ikpana’s grammatical system in their hands. Moreover, while a comparison of Ikpana interrogative constructions with those in other GTM languages is inevitably preliminary, as no study of comparable depth exists in any other GTM language, the authors still had access to the basic information about interrogative constructions provided in the corresponding grammar chapters. They say that they hope their book inspires comparable detailed studies of interrogative constructions in other GTM languages, and indeed, it would be great to have such studies for more languages. But again, it is maybe even more important that some GTM languages have got thorough descriptions that allow researchers to focus on specific parts of the grammar at a later point and lay the foundation for comparative studies. Here I would like to mention that while only some GTM languages had fairly detailed grammars by the end of the 20th century (such as Allan 1973), the situation has radically changed in the course of the latest two decades, in which a handful of comprehensive grammars appeared, such as Essegbey (2019).
12Given the serious endangerment of linguistic diversity in general, we have to put even more effort into collecting and analyzing data of the languages that remain. While the traditional descriptive approach has been complemented by documentary methods (starting from Himmelmann 1998), the book under review nicely exemplifies one more general approach – a theoretically informed focused study including experiments. Surely, this approach should complement, not substitute, the former two and, as mentioned, it is not possible on its own. At the same time, description and documentation should not be regarded as the only things that can be applied to “small” languages. Still, in my opinion, it is misleading when the authors use the term “documentation” for their study and their approach: studies of this type cannot be an alternative to documentary work aimed at collecting and annotating large amount of primarily naturalistic data, which can be used further for studies on an unlimited number of linguistic topics. While the authors of the book mention that they recorded and analyzed several traditional songs and tales, it is clear that their purely documentary activities were small as compared to elicitation and experiments. This does not make their study less significant, but checking the results against a representative text corpus would be an important task for the future. And since it is not a realistic perspective to have comparable studies for all imaginable grammar topics, not even of Ikpana, they cannot be a hypothetical substitute for language documentation.
13One can also formulate some other criticisms about the book under review. The most important one concerns the fact that intonation and other prosodic phenomena are analyzed separately from (morpho)syntactic ones, and thus there is no analysis provided for any kind of interrogative construction as a whole. In particular, the authors’ interpretation of the final vowel lengthening in Ikpana yes/no-questions remains ambiguous: while in Chapter 2, it is discussed as a segmental morphological unit, in Chapter 4, it is considered as part of the intonation pattern of polar interrogative sentences. Indeed, choosing between these two options for this kind of phenomenon can be difficult, but this is exactly why a joint account of morphosyntactic and prosodic phenomena would help, and one could at least say explicitly that arguments for both segmental and suprasegmental interpretations can be brought forward. It would also be interesting to treat syntax and prosody together for many other issues. For example, in subsection 3.3.2 “Wh- in situ in embedded clauses”, the authors discuss sentences that are ambiguous at the morphosyntactic level. Here, it would be worth to check whether different meanings of such sentences can always have different intonation patterns. Another question about Ikpana interrogatives that remains mostly unanswered concerns possible semantic, pragmatic or any other distinctions between constructions that seem, at first sight, synonymous. In particular, as mentioned, Ikpana has five options in total for wh-questions: wh-expression in situ, bare wh-movement, focused wh-movement, cleft construction and focused cleft construction. It is probable that the choice between them is not absolutely random; but so far, we are still unaware of the factors that could be held responsible for it. More topics related to interrogative constructions are outside the scope of the book, e.g. the authors never address Ikpana echo questions and do not report whether Ikpana has tag questions. Although the authors are surely competent in the area of African linguistics, sometimes their approach to language classification looks somewhat negligent – for instance, on p. 168, Nupe (currently: < Nupoid < Benue-Congo) is cited as a Kwa language, while it is not a member of the “new” Kwa language family since Bennett & Sterk’s (1977) classification. Moreover, the authors take for granted that GTM languages form a genetic group and do not give enough details about them (such as their older label “Togo-Remnant languages”). While Na-Togo and Ka-Togo are commonly accepted to constitute genealogical units, it remains disputable if the GTM group that they form together is a genealogical or just an areal unit inside the Kwa language family (see Kropp Dakubu 2017; Blench 2009).
14These and other possible criticisms do not change the fact that the book by Kandybowicz and colleagues is a very important and currently unique contribution to the study of interrogative constructions in GTM languages (and, more generally, in West Africa) as well as from a typological and theoretical perspective. It is also a good example of a study with a well elaborated methodology, combining a theoretically and typologically informed approach with a solid empirical basis including experiments. The authors stress that their study is a step towards understanding the Language Faculty (presumably to be understood in the Chomskyan sense). Still, one of the greatest advantages of the book is that it is easily readable by any linguist independent of their own framework, be it any formal or functional theory, a purely descriptivist approach or a future paradigm that does not even exist today. Currently, Ikpana is reported to be stable in terms of language vitality, but it is a language with only several thousands of speakers, which is an order of magnitude less than most other GTM languages, and so we cannot be sure that a study like the one presented in the book will still be possible several decades from now. Against this background, there is no doubt that it is exactly studies of this type that theoretically oriented linguists should conduct now (and funding bodies should support).