The term connective is usually used as a cover term for adverbs, prepositions and conjunctions (complementizers). In a wide variety of languages and certainly in the languages of Europe, elements of these three word classes are often closely related (e.g. German nach, nachdem, nachher, English for, therefore) or even identical in form (e. g. English before, since), so that various linguists have proposed to subsume them under the major word class of preposition and to describe the differences in distribution in terms of subcategorization (cf. Emonds Reference Emonds1976).

The chapters compiled and published in this volume are revised versions of papers presented at a workshop ‘Clausal Connectives in the History of English’ at the 13th Conference on English Historical Linguistics in 2004 (ICEHL) and discuss the rise, development and partly also the disappearance of connectives in the history of English. The book comprises twelve chapters, a word index and a subject index in addition to an introduction by the two editors. The list of connectives analysed and discussed includes the following expressions: lest, to, till, any/each/every time, since, albeit, while, whereas, if as well as various forms of relative pronouns with sentential antecedents. In each of the twelve chapters one or at most two connectives are discussed, with the exception of the chapter contributed by Ursula Lenker, whose scope includes the complete lexical field of causal connectives, and the chapter contributed by Anneli Meurman-Solin, who discusses the complete set of relatives relating to sentential antecedents. The last chapter (Thomas Kohnen, ‘“Connective profiles” in the history of English texts’) is quite different from the rest, focusing as it does on the distribution, frequency and proportion of connectives in sermons and statutes, in order to uncover patterns of use in these two genres over the centuries. This chapter is meant to make some contribution to the study of registers and text types, but does not have anything to say on connectives as such.

All analyses are based on authentic data taken from various corpora, to whose compilation many of the authors have made significant contributions, as well as on relevant typological studies (e.g. Raible Reference Raible1992; Kortmann Reference Kortmann1997) or language-specific analyses (e.g. Ford Reference Ford1993), on studies of grammaticalization and lexicalization (e.g. Brinton & Traugott Reference Brinton and Traugott2005) as well as on studies of language contact. Most of the chapters follow a general pattern: changes in the syntax and in the meaning of connectives – which also include the possibility of total loss − are pursued through the history of English on the basis of authentic (qualitative and quantitative) data.

In order to give the reader a clearer idea of their content, I will now summarize the chapters.

In the first one (‘Adverbial connectives within and beyond adverbial subordination’), María José López-Couso shows that the connective lest developed into a complementizer in contexts of verbs of ‘fear’ (‘for fear that’), starting out as an adverbial connective expressing negative purpose (Lest we forget; ‘so that . . . not . . . ’). These two uses are also the ones listed in the Dictionary of Contemporary English (Langenscheidt 2005): (i) ‘in order to make sure that something will not happen’ and (ii) ‘used to show that someone is afraid or that a particular thing might happen’. One thing that is not addressed in this discussion is the question whether the newly developed complementizer (subordinating conjunction) can also occur without the context of a predicate denoting ‘fear’, ‘anxiety’, or ‘worry’. The following authentic sentence, used by a headmaster in an announcement of an excursion, suggests that this is possible, even if the situation described is still evaluated as something undesirable:

1. (1) Boys are enjoined to bring raincoats along, lest the weather should prove inclement.

In her highly original, thorough and convincing chapter on the history of to as a connective in English, Bettelou Los argues that the to-infinitive is already verbal and hence clausal in Old English (OE), rather than a prepositional phrase as is often assumed. Matti Rissanen discusses the origin (borrowing from Old Norse), short history and replacement (by till, until) of the adverbial subordinator oþ in Old English and Early Middle English (EME), showing that the replacement of oþ by till takes place as early as the thirteenth century. True to its Old Norse origin, til(l) was used as an allative (> dative) marker before its meaning was restricted to a purely temporal use.

In an extremely interesting, if potentially controversial, contribution (‘Rise of the adverbial conjunctions {any, each, every} time’), Laurel J. Brinton discusses the relatively recent use of any time/every time and each time as adverbial conjunctions with the meaning ‘whenever’ in sentences like the following:

1. (2) Each time they smoke they give Nell a cigarette, too.

In contrast to what could be called ‘the standard analysis’, she rejects the view that each time (also increasingly written as eachtime, parallel to anytime, everytime) introduces a relative clause, claiming instead that it is an adverbial subordinator developing along the well-known path of grammaticalization: Prepositional Phrase > Adverbial > Conjunction (at any time (we meet) > anytime > any time we meet). Brinton shows quite convincingly that in her data the use of at any time as a prepositional phrase predates the adverbial use. What is not clearly established by the data given, however, is the status of the clause that may follow the prepositional phrases (or noun phrases), as for instance in the following example (see (10b), p. 83):

1. (3) Any time you have a chance to get pictures of the children I wish you would have them taken and send them to me.

Analogous problems of analysis also arise in connection with adverbial clauses of place and of manner (Any way you want it . . ., any place you go) and it is not established in the chapter beyond reasonable doubt that all these noun phrases with quantifiers and abstract nouns like time, place, manner are of exactly the same category as whenever, wherever or however. This controversial quality, however, together with the novelty of the observations makes the chapter much more interesting than some of the other more solid but less controversial documentations. There is one more problem that needs mentioning: even though the author does not say very much about the meaning of the three new additions to the class of adverbial conjunctions, she describes them as expressing ‘simultaneous overlap’ and as being synonymous with whenever. By implication this means that the three complex conjunctions are synonymous among themselves. This may apply to each time and every time in all contexts and to any time and the two others in some contexts, but I doubt if the usual exclusion of any from episodic contexts (* Anyone came to my party) is invalid for the cases under analysis. In the example (14e) on p. 85 and (15f) on p. 86 – repeated here as (4) and (5) − each time does not seem to be replaceable by anytime:

1. (4) He rally'd thrice his Troops, each time repulsed.

2. (5) ‘Well?’ said she again; and each time she said it, she opened her lovely eyes the wider.

The categorical versatility (adverb, preposition and conjunction) of since (< siþþan) could already be observed in earlier stages of English, as is shown in the chapter by Rafal Molencki (‘The evolution of since in medieval English’). Semantically, he describes a development from a basic meaning ‘later’ to a posterior meaning ‘after’ (misleadingly called ‘anteriority’) and further to a meaning ‘after and up to some point of orientation’, which replaced the anterior meaning in ME. When exactly the causal interpretation was added seems to be difficult to establish. What the author neglects in his discussion is the question of bridge contexts and reasons for these semantic changes. As far as the causal meaning is concerned it is clearly based on the conversational implicature (or fallacy) post hoc ergo propter hoc, which statements of temporal sequence typically give rise to (see Traugott & König Reference Traugott, König, Traugott and Heine1991):

1. (6) Since she left me, I have been feeling miserable.

As far as bridge contexts are concerned, purely stative contexts must have played an important role. If the clause introduced by since does not specify a beginning point through an event, as it does in (6), the temporal interpretation is excluded, the causal one being the only option:

1. (7) Since I am an old-age pensioner, I can spend a lot of time in museums.

A similar categorical versatility can be observed in the career of the fused form of the original sentential connective albeit, which is discussed in a thoroughly documented contribution by Elina Sorva (‘Grammaticalization and syntactic polyfunctionality: the case of albeit’). The author shows that the connective phrase all be it emerged in the fourteenth century − probably as a calque of Old French tout soit il/ce que − was subjected to univerbation and became very frequent in EME. After that period its frequency declined only to rise again in the nineteenth century. Since the twentieth century albeit is almost exclusively used in verbless clauses. To call this current use simply ‘concessive’ is not very illuminating, however, given that this expression cannot be replaced by other concessive connectives like although, even though, nevertheless, still, etc. The meaning in typical examples like (8) is best described as excluding inferences that would be justified on the basis of the main clause alone, by adding an additional (elliptical) proposition.

1. (8) Hitler was driven by a political vision, albeit a perverse one. (p. 134)

Here, as in many other cases, a comparative perspective would have been useful: albeit is very different from related phrasal expressions like ne serait-ce que in French or und sei es nur in German, but is very similar in meaning to allerdings in German or joskin in Finnish.

In a study focusing primarily on the semantic development of the conjunction while from ME to Present-day English (‘On the subjectification of adverbial clause connectives’), Ana I. González-Cruz demonstrates convincingly that in addition to the concessive reading frequently discussed in the literature there is also a contrastive, a causal and an additive reading to be considered. These additional meanings can be augmentations (enrichments) of a basic temporal meaning, but they may also constitute the primary content after the loss of the temporal element.

Another article with a clear semantic focus and the requisite sophistication in the analysis of meaning is the contribution by Carsten Breul (‘A relevance-theoretic view on issues in the history of clausal connectives’), whose empirical focus is the semantic development of where and whereas. The major theoretical problem, by contrast, concerns the question of how an original conversational inference (an ‘implicature’ or ‘invited inference’) becomes part of conventional semantic content. Breul's answer to the question is that both the principle of informativeness and the principle of relevance play an important role in this development.

In a highly informative chapter on the historical developments of causal connectives (‘Forwhi “because”: shifting deictics in the history of English causal connection’), Ursula Lenker looks at the overall changes in the systems of causal connectives in the history of English. In addition to this comprehensive perspective it is also the comparisons with German, French and Latin which make the chapter very interesting. The diachronic phenomena discussed are mainly changes in the formal make-up of connectives as well as their use (or non-use) in correlative constructions. The chapter will no doubt provide a comprehensive and important source of information on causal connectives in the history of English from now on. My only objection to the analyses and results presented concerns the terminology used for the subclassification of connectives on the basis of their formal make-up. First of all, Lenker fails to distinguish between ‘deictic’ and ‘anaphoric’, using the term ‘discourse deixis’ for the latter manifestation of referential dependence instead. Similarly unfortunate is the distinction between ‘pronominal deixis’ (OE forþæm, ME for that), ‘time/space deixis’ (therefore, hence, thence) and other deixis (OE swa, ModE so, thus) or the distinction between ‘deictics’ and ‘lexical means’. The latter distinction is unfortunate because the so-called ‘deictics’ also contain lexical components as a part of their formal make-up (for, German -halb/en), and the former is inconsistent in its use of criteria, mixing aspects of referential dependence (+/− anaphoric) with criteria of content. So and thus originally expressed manner deixis (‘in this/that way’) and are used anaphorically as causal connectives, and the difference between for that and therefore is primarily that between an originally local anaphor and a semantically unmarked anaphor.

Conditionals are one of the semantically most complex, and therefore most frequently discussed, adverbial clause types. The task of providing a comprehensive and coherent summary of previous analyses alone is a major challenge for any semanticist. It should therefore not come as a surprise that there is not much new in Claudia Claridge's discussion of the functions of conditionals in EME texts; at least not much that one could comment on. One of the most recent and probably the best studies of conditionals in English (Declerck & Reed Reference Declerck and Reed2001) is not even mentioned in the list of references and the conclusions about the functions of if-clauses in EME discourse (‘conditionals were both used for argumentative and interactive purposes, which sometimes interlink in one and the same instance’, p. 252) is not highly surprising.

Anneli Meurman-Solin (‘Relatives as sentence-level connectives’), by contrast, has a lot of interesting things to say about relatives with sentential antecedents in epistolary (Scottish) prose and in the history of English in general. That which rather than that or who(se) is used in such contexts is well known, but the range of other possible markers discussed, from which as part of a prepositional phrase (e.g. for which purpose) to anaphoric expressions introduced by where (e.g. whereupon), will provide valuable information for most readers. In the context of this discussion a lot of attention is given to supplementary relatives.

To summarize, the book under review is an important and very informative contribution to the study of the historical development of connectives in English and in European languages in general, and will provide another point of orientation for many years to come in addition to books like Raible (Reference Raible1992), Ford (Reference Ford1993), Kortmann (Reference Kortmann1997) and Pasch et al. (Reference Pasch2003). The vast majority of contributions to this collective monograph are well informed as far as their theoretical basis is concerned, thorough in their description and their documentation of authentic data and highly inspiring for further work. With a few exceptions, the quality of the semantic analyses does not quite match that of those concerning spelling, distribution and syntax. Many of the analyses might also have gained from a comparative perspective here and there. It is not so much a book for one's private bookshelves, but I can certainly recommend it to any library with a section on English language and linguistics. Given its wide scope, the book will provide useful information to readers interested in the history of English, in textual cohesion, in connectives and in linguistics in general.