I will nail my colours on the mast from the outset: this is a most remarkable book in a number of respects; it is careful, it is exhaustive, it is bold, and it combines theoretical vision with descriptive élan. Glyn Hicks succeeds to a large extent in recasting Principles A and B of the Binding Theory (Chomsky Reference Chomsky1981) in Minimalist terms (Chomsky Reference Chomsky and Kenstowicz2001). However, the volume under review is considerably more than an exercise in adapting to changing notations or, even, shifting viewpoints: it is a fine example of how rigor in theory may open the way to broader empirical coverage in generative linguistics, contrary to what may occasionally be assumed.
The book opens with a detailed ‘Introduction’ to Binding Theory, its history and the theoretical challenges it presents, a topic upon which Chapter 2, ‘Binding Theory and the Minimalist programme’, expands, with emphasis on the relevant challenges presented by current Minimalist theorizing. These first two chapters, besides laying the groundwork for what follows, make a fine introduction to Binding, which is better and more comprehensive than those in many textbooks. Chapter 3, ‘The Binding Theory does not apply at LF’, makes the case for Binding Principles A and B being derivable as grammar-internal operations and not as interface conditions, against the consensus among the majority of researchers. Chapter 4, ‘Eliminating Condition A, recasts Principle A’ as an instance of the operation Agree and sets its domain to the LF-Phase. Chapter 5, ‘Eliminating Condition B’, reanalyses Principle B as an instance of a general constraint on Merge, viz. Maximise Featural Economy (which is a syntactic version of the Obligatory Contour Principle), and sets its domain to the PF-Phase. Chapter 6, ‘Extensions to other Germanic languages’, applies and adapts the theory to evidence from Germanic. Chapter 7, ‘Conclusions’, concludes the book.
The task Hicks sets himself is considerable. To start with, summarizing the work carried out on Binding is an enterprise in itself: anaphoric relations are among the most extensively debated issues in grammatical theory. Second, the Binding Principles are one of Chomsky's major achievements and, far more significantly, a success story of the Principles and Parameters framework, as its alternative name ‘Government and Binding Theory’ indicates. Third, exactly because the Binding Principles are expressed in a framework that avails itself to multiple modules of grammar interacting with each other (such as S-structure and D-structure levels of representation), translating these mechanisms into Minimalist terms is anything but straightforward. Finally, extending the empirical basis of (what used to be) the Binding Principles is a yet more demanding, almost daunting affair.
Hicks successfully carries out these tasks and, in the process, resolves a number of issues, including the irrelevance of phi-features for establishing referential dependencies (107–112), the structure of English ‘logophoric’ reflexives such as your good self (141–145), binding into Preposition Phrases (175–179), the correlation between [VAR(iable)] features and person features (209–212).
In this book, Hicks concentrates on Principle A (which requires anaphors to be bound within their local domain) and Principle B (according to which pronouns must not be bound within their local domain). To my mind, the exclusion of Principle C (‘referential expressions must be free’) from the discussion is not justified in the detail and precision that characterize the rest of the argumentation – however I leave this matter aside. Proceeding from the work of others who attempt to translate Principles A and B into more Minimalist terms, the book's core proposals are as follows:
(a) Binding Principles A and B are not interface conditions, rather they are grammar-internal operations (67–75). In this Hicks agrees with Lebeaux (Reference Lebeaux2009), published virtually simultaneously with the book under review and hence not considered here (Hicks does, however, discuss Lebeaux Reference Lebeaux1998). Note that there is independent evidence against Principle B as a principle of Universal Grammar: numerous acquisition studies have shown that it has to be learned and that it is learned quite late, making it an unlikely candidate for an interface condition.
(b) As already observed in the literature, Principles A and B are not in real complementarity, despite what their received formulations suggest. Hicks argues that this is due to their applying within different domains: while Principle A applies within the LF-Phase, Principle B applies within the PF-Phase.
(c) Anaphoric relations derive from principled syntactic interactions among well-defined formal features; Hicks posits [VAR] (for ‘semantic variable’) and [REFLEXIVE] (which is relevant for analyzing S(imple)E(xpression) anaphors).
Armed with the above assumptions, Hicks offers a Minimalist alternative to classical Binding Theory while, crucially, extending its empirical basis. In the rest of this review I will concentrate on some specific issues arising from the theory outlined above.
Binding Theory was a success partly because it reduced the surface complexity of anaphoric relations to conceptually simple and elegant theoretical principles. Minimalism can be broadly understood as the drive to ascribe simplicity and elegance to the object of enquiry itself, namely the Language Faculty in the narrow sense. This forces us to reduce anaphoric relations to the following bare essentials:
(a) features
(b) structural relations derivable via successive Merge
(c) interface conditions
(d) locality–cyclicity – although not ‘virtually conceptually necessary’
In other words, if narrow syntax itself is a ‘dumb machine’ (Grohmann Reference Grohmann2008: 10, echoing Chomsky Reference Chomsky and Kenstowicz2001: 27), then we expect the handling of anaphoric relations to be a matter of interaction of features, as related by syntax and read off at the interfaces, constrained by cyclicity. One way to achieve this is the one proposed by Hicks: anaphoric relations are due to features related by grammatical relations within phases.
Now, on a more personal note, I am not averse to Minimalism, but I am hardly prepared for a wholesale sacrifice of empirical coverage to theoretical parsimony (see Newmeyer's Reference Newmeyer2008 review of Boeckx for a synopsis of this thesis). The reason I opened this review praising Hicks's book is exactly because it constitutes a most welcome piece of news for those sharing this programmatic desire, as the author demonstrates that honest-to-god Minimalism can actually achieve an empirical coverage equal, if not superior, to that of classical approaches to Binding. This is nothing short of an achievement.
An example of this is how Hicks explains (away) cases of non-complementarity between Principle A and Principle B, as for instance illustrated by the availability of both an anaphor and a pronoun in the complement of certain prepositions in English (e.g. John found a snake near him i/himself). If the domain within which Principle A applies is the LF-Phase and the domain within which Principle B applies is the PF-Phase, and if the two are distinct, as has been claimed in the literature and as Hicks himself extensively argues, then we expect non-complementarity to obtain in environments like (1).
(1) [LF-Phase … DPi … [PF-phase … anaphori/pronouni …]] (182, ex. (39))
In other words, non-complementarity is possible in John found a snake near him i/himself because Preposition Phrases are not LF-Phases but may be PF-Phases. Thus, in the above configuration Principle A (recast as an Agree relation) can freely obtain between the D(eterminer)P(hrase) and the anaphor. At the same time, when the Preposition Phrase is a PF-Phase, then the pronoun can be free within it, as per the reformulated Principle B (whereby only elements with non-identical features of a particular kind may merge in order to form a PF-interpretable syntactic object, a PF-Phase).
Interestingly, non-complementarity is part of a more complex and intriguing picture and Chapter 5 of the book considers the empirical foundations of a PF phase in greater detail, providing a principled explanation for the puzzling data in (2) (first noted in Fiengo & Higginbotham Reference Fiengo and Higginbotham1981).
- (2)
(a) %Johni read [books about himi].
(b) Johni read [books about HIMi].
(c) *Johni read [books about ’imi].
Recall that PF-Phases are syntactic objects that are readable by the articulatory–perceptual systems. While the bracketed constituent in (2b) is such a (prosodic) unit, and the pronoun is free within it, this situation does not obtain in (2c). The bracketed constituent in (2a) can, at least for some speakers, constitute a PF-Phase and hence it is possible for the pronoun to be free (182–189). As an aside, the relevance of prosody to syntax is one area where further work is likely to yield exciting results and derive many grammatical phenomena as either PF-conditions (cf., for example, Grohmann's (Reference Grohmann2003) ‘Anti-Locality’ and Kotzoglou's (to appear) ‘Restriction on Copy Reduction’) or as constraints on Merge discernible at PF (like Hicks's ‘Maximise Featural Economy’, which subsumes Principle B).
At this point, I have two main criticisms. The first concerns the discussion about DPs and their PF-phasehood. The second regards Chapter 6, and its discussion of SE anaphors – which will bring us to a broader issue with much of current Minimalist theorizing.
Given the way Hicks's theory works, the question posed in section 4.5.2, ‘Are DPs (LF-)Phases?’, is crucial, even though its discussion does not match the standard established in the rest of the book. Thus, the author summarily makes the case for DP-internal subjects sitting in SpecNumP, a projection below DP (151–153), which is a convenient stipulation. Hicks here fails to provide counterarguments to the established theses that DP-internal subjects are in SpecDP and that SpecDP may serve as an escape hatch for A′-movement out of DP (e.g. Szabolcsi Reference Szabolcsi1983, Horrocks & Stavrou Reference Horrocks and Stavrou1987). Within the same section, we encounter a discussion of the phasehood of nP (153–155). Again, Hicks's treatment does not include crucial references (e.g. the line of work stemming from Marantz Reference Marantz2000), with only a footnote pointing to work by Chomsky, which cites the claim that nP is a phase as an ‘idea attributed to Alec Marantz’ (154). The sketchy exposition has an effect on the discussion in the rest of the subsection. Thus, the author claims that subjectless nPs are not LF-phases, just like defective/unaccusative vPs (154). However, this claim entails that n here is not understood as the ‘nominalizer’ of roots but, rather, as the equivalent of a VoiceP – which would be quite a different thing. To be fair, terminological profligacy and confusion – with, for example, v indiscriminately being used for ‘root verbalizer’, ‘transitivizer’, and ‘Voice’ – is not a problem exclusive to Hicks.
My second criticism is similarly not exclusive to Hicks's approach, but concerns much research carried out in the Minimalist framework. Reducing syntactic mechanisms to ‘dumb’ operations like Merge and Agree (however the former is constrained and however the latter is construed) entails the following: in order to understand syntactic variation and the subtle workings of syntax we need, besides a highly articulated theory of interfaces and the conditions they impose on syntactic derivations, a theory of formal features (see also Adger & Svenonius Reference Adger and Svenonius2009). One of the reasons why this is necessary involves the propagation of purported formal features that characterizes much Minimalism-oriented research, especially when it aims for substantial (or even simply non-trivial) empirical coverage. More of our research effort should be dedicated towards the particularly difficult task of compiling an inventory of UG formal features. Chapters 4 and 6 of the book adumbrate this. In Chapter 4, the very interesting section 4.2.4, ‘Features involved in binding’ (107–124), provides motivation for [VAR], the variable feature involved in the interpretation of pronouns and anaphors and underlying both anaphora and coreference. Essentially, [VAR] subsumes the referential indices of old without being a mere notational variant thereof, and the section is one of the many success stories of the book.
In Chapter 6, however, in order to capture the behaviour of the perennially troublesome SE anaphors (Dutch zich, Scandinavian seg/sig and the like), Hicks in addition introduces a [REFLEXIVE] feature. Leaving details aside, this entails that there are at least two sources of reflexivity: an uninterpretable [VAR] feature on SELF anaphors, and a [REFLEXIVE] feature for specific interpretations of SE anaphors – not to mention the conundrum of anaphoric expressions like Greek o eaftos tou ‘the self his’ and Basque bere burua ‘his head’, alluded to in the context of Dutch possessive reflexives (222). My point here is not against positing features: if they are necessary, they will have to be posited and – at the end of the day – it is more parsimonious and reasonable to argue for more features than for more ad-hoc grammatical principles. However, features, if indeed elements of the Language Faculty, are certainly not without cost to the grammar and can definitely not be posited in a pretheoretical fashion.
From the above discussion I hope that three things have become clear: (i) the book under review achieves significant empirical coverage with a minimum of methodological and theoretical resources in a broad and thorny field; (ii) it is wide in its scope and, also, detailed enough so as to constitute – to my mind – a milestone in our understanding of anaphoric relations; (iii) it gives new perspective and – I hope – new impetus to the research into Binding, refreshing the field and placing the much-debated issue on a new and more principled basis.
