In this book, Dagmar Haumann makes an important contribution to the field of adverbial syntax, presenting interesting ways of understanding and analysing the empirical facts at hand. Concentrating on English, she investigates the licensing of adverbs in the entire clause, along with the relationship between adverb licensing and clause structure. Haumann's analysis is set within a Chomskyan framework, although she does not take into consideration all of its more recent developments. More specifically, Haumann does not discuss phases or the whole theoretical apparatus connected to this concept. To the best of my knowledge, however, there is nothing in Haumann's analysis that hinges on being set in a ‘pre-phases framework’.
Haumann's book contains six chapters, which I will briefly list to give the reader an idea of the organization of the book, before turning to the parts of the book that I would like to highlight. In chapter 1, ‘Introduction and overview’, the reader is introduced to the problems and issues in the field of adverbial syntax and semantics. Chapter 2, ‘Proposals regarding the structural integration and licensing of adverbs’, is a theoretical chapter in which various approaches to adverbial syntax are compared, discussed, and critically evaluated in an illuminating presentation. Once the scene has been set, we get the three main chapters of the book, namely chapter 3, ‘Adverbs within the lexical layer’, chapter 4, ‘Adverbs within the inflectional layer, mostly’, and chapter 5, ‘Adverbs within the Complementizer layer, mostly’. Each of these three chapters deals (mostly) with adverbs in one of the three domains of the clause. Finally, chapter 6, ‘General conclusion’, provides a good summary of what Haumann has argued for in her book.
During the 1990s, the adjunction analysis of adverbs started to be questioned by, for instance, Alexiadou (Reference Alexiadou1997) and Cinque (Reference Cinque1999), in the light of Kayne's (Reference Kayne1994) Linear Correspondence Axiom. Proponents of the adjunction analysis reacted to these works, the reaction ultimately resulting in the publication of Ernst (Reference Ernst2002). While Cinque (Reference Cinque1999) outlines a syntactic theory of adverb distribution, according to which sentential adverbs are licensed through specifier–head agreement with hierarchically ordered functional heads, Ernst (Reference Ernst2002) advocates a traditional, semantic view of adverb distribution and licensing, according to which an adverb can be adjoined anywhere you like, as long as this does not violate any important principles. Furthermore, Ernst's theory allows right-adjunction and rightward movement, whereas Cinque's proposal conforms to Kayne's theory of antisymmetry, which dispenses with rightward movement. (Haumann devotes sections 2.2.1, ‘Licensing adverbs as unique specifiers’, and 2.3, ‘Semantic scope’, to Cinque's and Ernst's proposals; for a more detailed overview of these proposals, see Beijer Reference Beijer2005.) Haumann's book constitutes an attempt to expand on Cinque's theory of adverb licensing by making the approach less vulnerable to various types of criticism and addressing the licensing of adverbs within the higher and lower layers of the clause, i.e. the domain of the complementizer (C) and the verb (V).
Let me now turn to Haumann's arguments designed to counter Bobaljik's (Reference Bobaljik1999) important criticism of Cinque's theory. Bobaljik argued that if sentential adverbs are merged in unique positions in the inflectional domain, the fact that non-finite auxiliaries may occur to the left of, say, frequency adverbs cannot be accounted for without violating the Head Movement Constraint (Travis Reference Travis1984). He demonstrated that it is possible to use Cinque's own logic to argue that it is the auxiliaries, rather than the adverbs, that have fixed positions and that it is in fact the adverbs that undergo syntactic movement. Moreover, in a Cinque-type analysis, we are forced to assume that finite auxiliaries move to the Tense Phrase (TP) only optionally (236).
The possible orders in which auxiliaries may occur in relation to each other and in relation to the frequency adverb often are illustrated in table 1 (based on Haumann's example (144), 236): rows (a)–(d) demonstrate the optional movement of the finite auxiliary, while rows (e)–(j) show that non-finite auxiliaries may also occur to the left of the frequency adverb and enumerate the additional landing sites that would have to be postulated between TP and often to account for the attested word orders.
Table 1 Deriving alternative verb-adverb word orders with only one merge site for often
One way out of Bobaljik's paradox is to pursue an adjunction analysis. Haumann chooses another path, however, since an adjunction analysis would require additional filters to rule out impossible orders and ensure that the system does not overgenerate. She proposes that a certain adverb type can have more than one merge site, as long as the universal adverb hierarchy is adhered to (205, 236). Table 2 (which is based on Haumann's example (145), 236) summarizes Haumann's proposal for the adverb class represented by often, omitting X-bar positions that do not host overt material or traces in the examples under consideration.
Table 2 Deriving alternative verb-adverb word orders with three merge sites for often
As table 2 shows, if we assume three possible merge positions for often, we can derive all the possible relative orders between auxiliaries and the frequency adverb in (1) (adapted from Haumann's example (143), 235), without violating the Head Movement Constraint.
- (1)
(a) They will often have been described in the literature.
(b) They often will have been born in this country.
(c) I have often been thinking of you.
(d) They often have been doing volunteer work.
(e) Diagnosis will have often been confirmed by a Doctor.
(f) Boris's son could be often seen dancing.
(g) The team could have been often led by a pensioner.
(h) ?These towels could have been being often used to wipe up all kinds of liquids.
(i) ?He had been often being mistaken for his twin brother.
(j) Nepali workers are being often deprived of their rights.
The question is whether we really want to assume multiple merge sites for one and the same adverb type. While Haumann's approach provides a solution to Bobaljik's paradox, it postulates additional adverb positions simply in order to derive the attested word orders. Perhaps we should instead follow Bobaljik (Reference Bobaljik1999) and others and assume adverbs to be hierarchically ordered in another dimension. Such an approach would explain how it is possible that different adverbs and verbs occur in a fixed order within their respective group.
Concerning the C-domain, Haumann advocates the ‘central idea … that preposed adverbs occupy designated functional projections in the left periphery of the clause’ (327), but she also argues in favour of an analysis according to which illocutionary, evaluative and evidential adverbs ‘are [externally] merged as specifiers of designated functional projections in the left periphery’ (335). Haumann thus questions the general assumption that the highest adverb-hosting projections in Cinque's (Reference Cinque1999) hierarchy are part of the I-domain. This makes sense, and I find Haumann's arguments regarding the C-domain generally convincing.
Regarding the V-domain, Haumann argues that VP-adverbs also have designated and ordered positions (for similar claims, see Alexiadou Reference Alexiadou1997, to which Haumann frequently refers). While the author makes a number of interesting observations regarding the V-domain, the ways in which she employs the theoretical machinery to accommodate the VP-data are sometimes too complex and imaginative to be intuitively appealing.
Haumann adopts a version of the Split VP Hypothesis, according to which the verb's external argument originates in the specifier of vP, while the internal arguments originate within VP proper (118). According to the phrase structure assumed by Haumann, TelP (telicity), AgrIOP and AgrOP occur between vP and VP (126). Other important parts of Haumann's proposal are (i) that active and passive verbal elements project lexical layers that differ in their internal makeup (passives have no vP), and (ii) that the movement properties of active and passive verbal elements differ. There is no place in Haumann's analysis for ‘the notorious VoiceP’ (138). Instead, passive sentences contain Pass(ive)P, while actives do not. Haumann further demonstrates (i) that different types of VP-adverbs are ordered in relation to each other, (ii) that we must distinguish between ‘means-domain adverbs’ and ‘pure domain adverbs’, since the former are VP-bound (This was still a sizeable text to process manually; 133) while the latter may occur freely in sentence-initial position (Psychologically, England were right; 164), and (iii) that scene-setting spatial expressions in the C-domain do not originate in the V-domain but are merged in the C-domain between Rizzi's (Reference Rizzi and Liliane1997) ForceP and FinP.
Haumann's book is very comprehensive, detailed, and impressive in its scope, but it is too long, and it would have benefited considerably from more editorial guidance (which would also have reduced the number of errors in the text). Furthermore, the extensive use of footnotes listed at the end of each chapter occasionally makes the reading of Haumann's book quite a frustrating experience, especially since so much interesting information is contained in the footnotes.
One of Haumann's main contributions is her theoretically interesting solution to Bobaljik's (Reference Bobaljik1999) paradox. Haumann also provides us with an empirically and theoretically sound division between ‘subject attitude adverbs’ and ‘subject-oriented adverbs’ – a division not made by Cinque (Reference Cinque1999) or Beijer (Reference Beijer2005), even though nothing in their empirical data argues against such a division.
On the negative side, it should be mentioned that there is no explicit information in Haumann's book that helps the reader to position the book in its correct context. The book was published in 2007, but it is impossible not to get the impression that it was written much earlier than that. For instance, works more recent than 2002 are rarely referred to and never used as main sources. There are probably good reasons for these facts, but they should have been mentioned in a preface. Moreover, reference is sometimes made to unpublished versions of sources that were actually published several years prior to 2007. Finally, it is not obvious which of Haumann's claims about English are intended to be regarded as universally valid – a point on which Haumann could (and should) have been much more explicit.
All in all, Adverb licensing and clause structure in English is an important book that anyone interested in clause structure and in the syntax and distribution of various adverb types (in English) should try to find the time to read.
