This monograph, based on the author's doctoral dissertation, is a comparative investigation of the lack of multiple wh-questions in four unrelated languages, Somali, Berber, Italian and Irish. It aims at providing a typological specification of languages that do not license multiple wh-questions within the typological system developed for languages that do. In this aim, the investigation starts from an overview of the strategies employed by multiple wh-question languages (i.e. multiple wh-fronting, multiple wh-in-situ, a mixture of both) and relates them to the strategy of question formation employed by non-multiple wh-question languages, viz. single wh-fronting. The questions addressed are the following:
1. Why do languages without multiple wh-questions not allow for a mixed system, e.g. fronting one wh-element and leaving the other one in situ?
2. Why is multiple wh-fronting not an option either?
The analysis that the author provides is based on the parallel behaviour displayed by wh-questions and focusing constructions: in both cases, the fronted element is adjacent to a head with specific properties. This requirement is captured by the ‘Head-Adjacency Generalisation’; together with the ‘Uniqueness Hypothesis’, which states, roughly, that there is a unique position in which both wh- and focus phrases are licensed, it accounts for the behaviour of non-multiple wh-question languages. Let us now turn to the organisation of the monograph.
The aim and structure of the book are presented in chapter 1, ‘Introduction’, which lays out the research questions, discusses their relevance and interest for current research, and devotes a brief section to highlighting salient points of the theoretical programme adopted, Chomsky's Minimalism. This is followed by an evaluation of three generative accounts of wh-questions, which make use of (i) clausal typing (Cheng Reference Calabrese, Guerssel and Hale1997), (ii) the parallel behaviour of wh-questions and focus constructions (e.g. Rizzi Reference Lecarne, Rebuschi and Tuller1991), and (iii) the need for wh-questions to satisfy both clausal typing and focus requirements (Sabel Reference Rizzi and Haegeman2006). Each of these analyses falls short when applied to multiple wh-question languages, either by failing to provide an exhaustive explanation for them, or by making predictions that are not borne out. Moreover, from a typological point of view, a number of questions about the nature of languages that do not allow multiple wh-questions remain unanswered, such as, for example, what prohibits the possibility of multiple questions in languages as typologically diverse as Irish, Italian, Berber and Somali. The chapter concludes with a few words on the theory of focus, introducing the two types identified in the literature, information focus (IF) and contrastive focus (CF), as well as Rizzi's (Reference Rizzi1997) split Complementiser Phrase (CP) hypothesis.
Chapter 2, ‘Previous analyses of the ungrammaticality of multiple wh-questions’, sketches three analyses that have been proposed for Italian and Irish (Adams Reference Adams1984), Italian and Berber (Calabrese Reference Calabrese, Geest and Putseys1987) and Somali (Lecarne Reference Cheng1999), chosen for their relevance to the four languages under investigation, and because they attempt a comparative explanation of the impossibility of multiple wh-questions in these languages. However, Adams's proposal does not satisfactorily account for the variability and full range of wh-constructions in Italian and Irish; the analysis put forward by Calabrese can produce the right outcome only with a number of stipulations; and Lecarne's ideas are problematic in that they result in contradictory typological properties for Somali.
After these two introductory chapters, the volume proceeds to the ‘core’ – in terms of both length and content – of the discussion in chapter 3, ‘The overview: What is possible in which language?’. The chapter presents and discusses the facts in each of the four languages under investigation, addressing focus constructions as well as wh-questions. Stoyanova poses five questions that guide her enquiry (167):
Q1: Is the lack of multiple wh-questions related to any other syntactic phenomena?
Q2: How do these phenomena interact with one another?
Q3: Which properties block multiple wh-question formation?
Q4: Are there any other relevant similarities between the four languages under investigation?
Q5: Is it possible to achieve a uniform analysis of the phenomenon?
The section on Italian (3.1) establishes two generalisations. The first one concerns IF and CF, which Stoyanova claims are structurally non-distinct and occupy the same position, viz. the specifier position of the Focus Phrase (SpecFocP). The differences between the two types of focus are accounted for by assuming remnant movement of the background part of the sentence: in IF but not in CF constructions, the remaining Inflection Phrase (IP) must undergo remnant movement after the focalised phrase has moved to the left periphery. Wh-elements also occupy SpecFocP; it follows that only one focused or wh-element is permitted in a sentence, hence ruling out the possibility of multiple wh-questions. The second generalisation is that there is an adjacency requirement between a focused or wh-element and a [+Focus] head, which can be either lexically filled or left empty. These generalisations also hold for Somali (in which the relation between wh-elements and focus is tangibly represented by the former being expressed through the use of a focus marker particle), and for Berber and Irish (in which wh-questions and focus constructions are expressed as clefts). In all four languages, wh-questions and their answers (i.e. focus constructions) display the same structural properties, and they are analysed as syntactically identical. The chapter concludes by providing the following answers to the five questions above (126–127):
A1: Both wh-questions and focusing constructions in languages without multiple wh-questions require the fronting (or base generation) of one element into a designated positions, identified as Spec, Foc, and adjacency to a [+Focus] head;
A2: The two phenomena are the same, syntactically speaking;
A3: The existence of a unique Focus position, which allows neither multiple wh-questions nor multiple foci;
A4: The four languages display the same properties with respect to focus and question formation;
A5: Given the parallels between the four languages, only a unified analysis can make sense.
A third phenomenon that is investigated alongside wh-questions and focus constructions is the anti-agreement effect: Somali, Berber, Irish and some Italian dialects, but not Standard Italian, require (or allow) lack of agreement in a variety of constructions, mainly (short distance) A′-dependencies. In Somali and Berber, the anti-agreement phenomenon is found in both wh-questions and focus constructions, while in Irish and the Italian dialects there does not appear to be a direct relation between the two. Anti-agreement phenomena have been explained on the basis of the properties and distribution of different empty categories. Stoyanova concludes that (i) it is unlikely that anti-agreement effects can be accounted for in a unified way, (ii) anti-agreement effects are not among the properties of a language that block multiple wh-questions, and (iii) the relation between the two phenomena is indirect.
Chapter 4, ‘Analysis’, investigates the realisation of the three parameters that make up the Uniqueness Hypothesis – lack of in-situ focus, lack of multiple specifiers and recursion of FocP – and briefly considers their theoretical implications. In Stoyanova's analysis, the triggering force behind movement of the wh-element to a designated position is its uninterpretable, strong focus feature, a proposal already put forward in Sabel (Reference Rizzi and Haegeman2006).
Chapter 5, ‘Conclusion’, closes the book with a lucid summary of the claims made. It reflects on the possible potential combinations of wh-question formation and focusing strategies, which may or may not involve movement, and presents these combinations (and the languages that instantiate them) in tabular form.
Stoyanova's book is a valuable contribution to our understanding of focus and wh-questions, both in its integration of typological observations and theoretical syntax and in its comparative approach to languages that do not allow multiple wh-questions, however unrelated to one another they may be. The study contributes to current debate in generative linguistics by combining newly awakened interest in Information Structure phenomena with traditional research on wh-questions. Stoyanova's argumentation is overall clear and systematic, and the data presented, although mostly not collected by the author herself, provide a useful reference for anybody interested in a comparative approach to focus and wh-constructions.
A couple of more specific comments: one of the fundamental assumptions of Stoyanova's account of the impossibility of having multiple wh-questions in Italian is the syntactic identity of IF and CF. This is needed in order to reconcile Rizzi's (Reference Lecarne, Rebuschi and Tuller1991) and Calabrese's (Reference Calabrese, Geest and Putseys1987) analyses, which equate wh-questions with, respectively, CF and IF constructions. The two types of focus constructions are shown to share the same structural properties, and the suggestion of a higher CF projection adjacent to a lower IF projection is rejected on the basis that CF and wh-questions cannot co-occur. Hence, Stoyanova concludes that CF and wh-elements must occupy the same syntactic position. However, there is evidence that suggests that CF, unlike IF, does not need to be adjacent to the finite verb, as illustrated in (1)–(2). This is particularly apparent in Sicilian (cf. the asymmetry with respect to verb-adjacency between IF a virità ‘the truth’ and CF na littra ‘a letter’).
(1) (Standard) Italian (Rizzi Reference Rizzi1997: 296, ex. (37e))
Credo che QUESTO, a Gianni, domani, gli dovremmo direI-think that THIS, to Gianni, tomorrow, to-him we-should say‘I think that we should say THIS to Gianni tomorrow.’
(2) Sicilian (Bentley Reference Bentley, Bentley and Ledgeway2007: 53, ex. (7a–b))
(a) Chi cci ricisti a tò niputi? A VIRITÀ (*a mè niputi) cci rissiwhat to-him you-said to your nephew? the truth (*to my nephew) to-him I-said‘What did you say to your nephew? I told (*my nephew) him the truth.’
(b) NA LITTRA, a Pina, cci scrissi, no un pizzinua letter, to Pina, to-her I-wrote, not a card‘It was a letter I wrote to Pina, not a card.’
This clear difference between CF, on the one hand, and wh-constituents and IF, on the other, suggests that IF and CF cannot be understood as syntactically identical: the incompatibility of CF and wh-questions may not be due to the two competing for the same position. Furthermore, interpreting the fact that both IF and CF display properties of moved elements as an indication that they move to the same slot is too hasty a conclusion.
Stoyanova makes reference to Italian marginally allowing multiple questions, alongside multiple CF constructions, but discards the evidence on the basis that it clashes with the data put forward in Calabrese (1984) and Rizzi (Reference Rizzi1997). Given the well-known diatopic variation of Italian, it would be interesting to investigate this claim in more depth, obtaining more data and systematically testing for the connection between multiple wh-questions and multiple focus.
At the start of this book, the author sets out the characteristics of question formation strategies employed by multiple wh-question languages, and proceeds to discuss how these relate to what is found in languages that do not allow multiple wh-questions. While this is a typologically relevant and interesting question that aims at gathering evidence for language profiling, it nevertheless runs the risk of turning into a contradiction in terms. Occasionally, the reader cannot help but suspect that non-multiple wh-question languages are being defined on the basis of a typology styled on multiple wh-question languages, almost as if the former constituted part of the latter. This suspicion is confirmed when Stoyanova presents the typology of multiple wh-question languages in a table that subsumes both languages with and languages without multiple wh-questions (163). Logically, it is not possible to define [not A] as part of [A].
Although the Minimalist analysis provided is attractive and elegant, it does not come without assumptions and postulations that may not necessarily be plausible. The most valuable parts of Stoyanova's book are, in my view, her generalisations about languages without multiple wh-questions and the typological classification derived.
The book will certainly be of interest to any scholar working on wh-questions, who will appreciate the breadth of the data and the ideas presented. The book constitutes a stimulating incentive to find answers to Stoyanova's insightful five questions, answers which will contribute a further piece to the puzzle presented by the typology of wh-questions.
