In The noun phrase in English: Past and present, editors Alex Ho-Cheong Leung and Wim van der Wurff bring together a collection of six high-quality papers discussing topics connected to the English noun phrase. The topics are – to the general linguistics community perhaps surprisingly – wide-ranging. In a short introduction, the editors situate this collection within the thriving research on the English noun phrase from synchronic and diachronic perspectives in the last ten to fifteen years. Within this active field, they identify the aim of this collection as sustaining and further developing noun phrase-related work. As they point out, this collection moves attention from global descriptions and hypotheses to strongly focused studies addressing specific grammatical patterns and constructions within the noun phrase complex and interactions of the noun phrase with larger grammatical constructions. The six papers in this collection share a commitment to usage-based methods, and they all use data from corpora or the Web to support their analyses. The authors, who all have extensive expertise in describing and modelling the noun phrase, work within different theoretical and methodological traditions. As a result, this collection not only expands and improves our understanding of the noun phrase, but also makes a wider contribution to, amongst others, the study of the lexis–grammar interface, the study of grammatical structure as (motivated) form–meaning relations, the study of the loss of grammatical items and of written-based grammatical change, and the variationist study of morphosyntax.

The first paper in this collection is ‘Complex NPs with third-order entity clauses: Towards a grammatical description and semantic typology’, in which Kristin Davidse proposes a comprehensive lexico-grammatical description of noun phrases (NPs) such as the question whether he …, the fact that he …, his belief that she …, etc. which ‘depict propositions and processes as abstract entities’ (p. 12) or third-order entities. Two different syntactic analyses of the relation between noun and clause are found in the literature: noun and clause are argued to stand in apposition or the clause is argued to be a complement of the noun. Davidse starts her paper by dissecting both analyses identifying their limitations and their explanatory power, which leads to the conclusion that neither provides a comprehensive account of the phenomenon. In the course of the discussion, the appositive relation, in particular, is delineated more clearly and sharpened up with reference to recent studies and re-conceptualised as a modifier–head relation as defined in Langacker's Cognitive Grammar, thereby preparing the way for Davidse's own grammatical analysis. This detailed cognitive–functional analysis assumes a motivated relation between structure and meaning: form codes meaning and differences in form reflect differences in meaning. Grammatical alternations, evidenced with examples from the Wordbanks Online corpus and the Web, are used as a tool to provide a principled grammatical analysis and tease apart fine-grained semantic distinctions. The result is a sophisticated three-level analysis rich in detail and careful in argumentation. At the highest level these noun phrases are argued to have a common denominator, not only in containing third-order entities, but in presenting them as given information through the typical use of definite determination. On the second level, two distinct grammatical constructions are distinguished defined by the different functional and structural relations between noun and clause. On the third level, further subtypes of the two constructions are proposed, in which different nominal semantics is accompanied by and motivates distinct grammatical and syntactic behaviour. In the process of developing this comprehensive analysis, Davidse contributes to a better description and understanding of theoretical concepts such as apposition, modification versus complementation, and factivity.

Victorina González-Díaz's paper ‘Adjective stacking in Early Modern English: Some stylistic considerations’ is concerned with noun phrases that contain multiple characterising adjectives (e.g. adjectives whose function is to describe or characterise the referent as in a small pretty blue box). Co-occurrence of multiple such adjectives can be managed by means of different syntactic structures involving stacking (AAN) and coordination (A coordinator AN; AN coordinator A). It has been widely acknowledged that stacking only develops in Late Middle English and Early Modern English. Using data from the Penn Parsed Corpus of Early Modern English, González-Díaz provides the first systematic in-depth analysis of noun phrases with two characterising adjectives in this period of change, in addition to general observations on noun phrases with more than two adjectives. The corpus study, which considers both semantic-structural and socio-stylistic factors, is carefully embedded in literature on adjectival modification and on Early Modern English genres and style. The first part of the study investigates the structural realisation of three semantic modification relations: descriptive modification, in which the two adjectives independently modify the noun (a small blue box); indirect modification, in which the first adjective is evaluative and the second adjective provides a justification of the evaluation (a good fast car in which the car is evaluated as good because it is fast); and unitary modification, in which the two adjectives form a single unit expressing high degree of the characterising feature (a tiny little bird) or affective stance (good old John). González-Díaz finds that indirect modification and affective unitary modification are restricted to stacked adjectives. She proposes that the coordinator precludes the evaluative adjective to have scope over the second adjective and makes reference to Feist's (Reference Feist2012) limited diachronic analysis of the development of the premodifying zone. This is not wholly satisfying as an explanation. The relation between stacking and these two types of modification, which, as noted by González-Díaz, both involve a first adjective with a subjective meaning, is likely to be more complex. Diachronic studies such as Adamson (Reference Adamson, Fischer, Rosenbach and Stein2000) and Vandewinkel & Davidse (Reference Vandewinkel and Davidse2008), to name but two, have shown that stacking provides the necessary structural context for subjectification of the first adjective and hence the development of indirect modification. The relation is perhaps not so much precluded as not developed in its most typical form; and whether it has a different structural realisation in earlier periods can only be answered conclusively through a study of evaluative adjectives in earlier data. The most novel and exciting part of the paper is the socio-stylistic analysis, and the subtitle of the paper ‘some stylistic considerations’ is rather too modest. González-Díaz convincingly connects increase of tokens and types of stacked adjectives with the establishment of new, written genres, the waning high-style, and individual developments in travel writing and medical writing. The analysis adds a further example to a growing number of cases presenting written-based grammatical innovation in the history of English.

In the paper ‘The rich, the poor, the obvious: Arguing for an ellipsis analysis of “adjectives used as nouns”’, Christine Günther joins the long-standing debate on the analysis of determiner + adjective combinations with generic or abstract reference such as the young, the impossible, in English and German. The debate centres round two opposing analyses of these noun phrases as containing nominalised adjectives or as involving an absent, ellipted noun. It is the latter view which Günther defends by means of a detailed discussion of the similarities of these constructions with ‘ordinary’ elliptical noun phrases, where the antecedent can be anaphorically retrieved from the context. Evidence, for English, consists of data from the British National Corpus, the Corpus of Contemporary American English, as well as examples from the Web. Günther not only strengthens the case for an ellipsis analysis empirically, but also, in a compelling way, draws attention to lexical constraints – predicative-only adjectives are not found in the constructions – and to parallels with ellipted noun phrases with quantifiers and noun phrases with substitutive one. Occasionally, however, one gets the impression that she is trying perhaps too hard to find parallels and make everything fit. The use of corpus data as illustrations of what is possible rather than what is common plays into this. One case in point is the discussion of semantic equivalence between constructions with one denoting humans (the poor ones) and combinations such as the poor. The fact that examples can be found in which the former have generic reference and the latter concrete reference does not dismiss the fact that the former prototypically have concrete reference and the latter prototypically have generic reference. The contrast between English and German that is at the forefront of the analysis is interesting and enriching, but it occasionally suffers from a similar rigidity. Different constraints in English and German are explained by morphological divergence, i.e. the fact that German adjectives carry inflection, but the possibility that English also diverged in other ways is not really entertained. The fact that the discussion mentions, with references to other work, that other less related/unrelated languages show the same trade-off between constraints and adjectival morphology strengthens the contrastive claims considerably. However, as a native speaker of Dutch, which structurally often occupies the middle ground between German and English, a key piece of evidence sat uneasily: Dutch adjectives have a much-reduced inflection, only -e, but the equivalent of the poor and the rich receive the nominal plural marker -n, e.g. de armen, de rijken, resulting in a contrast between de rijken eten caviaar ‘the rich eat caviar’ and the clearly elliptical de arme mensen eten soep en de rijke caviaar ‘the poor people eat soup and the rich ones caviar’. There is confusion in actual usage, but the grammatical rules point to a degree of nominalisation or lexicalisation impossible to establish for either German or English in this way. All in all, Günther provides a very rich description and intriguing argumentation, which by allowing for some variation and acknowledging change will only be more convincing.

In the paper ‘Variable article usage with institutional nouns: An “oddment” of English?’, Marianne Hundt investigates the occurrence of the definite article with the nouns church, hospital and university in be at X and go to X strings, that is, the variation between be at university / be at the university, between go to hospital / go to the hospital, etc. Hundt aims to find out whether the conventional wisdom that differences are largely a matter of regional variation, with American English speakers preferring the string with the article and British English speakers the one without, holds up in a statistical, multivariate analysis. Hundt uses data from the British National Corpus and the Corpus of Contemporary American English, and includes a range of factors in addition to regional variety, most importantly, lexical item (church versus hospital versus university), verb + preposition (be at versus go to), presence of postmodification (postmodification versus no modification/premodification). The variable rule analysis identifies all these factors as significant, but attributes the greatest effect to lexical item, followed by presence of postmodification, regional variety and, finally, verb + preposition. (Hundt also discusses a semantic factor, whether the referent is general – leading to a stereotypical activity interpretation – or specific, but rightly decides to exclude it as it could not be accurately operationalised.) In the discussion section, Hundt explores possible motivations for the lexical differences. She suggests diverse historical origins of the three nouns – only church is attested in Old English when article usage, in particular in prepositional phrases, was not systematic yet – and frequency effects might play a role. Hundt touches on the topic of the lexis–grammar continuum, pointing out that her study provides evidence for a lexical base to parts of grammar. It would have been interesting to connect back to and contextualise this within the Construction Grammar framework that is adopted at the start of the paper but that does not feature in the discussion or conclusion. Throughout the paper, the rigour and transparency of the analysis and the clarity with which the analysis is explained are exemplary. The application of variationist methods in corpus studies of morphosyntactic variation is becoming more common and a study such as this one provides both an example for other scholars of what can be done and how it should be done.

Alex Ho-Cheong Leung and Wim van der Wurff's paper ‘Anaphoric reference in Early Modern English: The case of said and same’ is concerned with two phrases, the said N and the same used on its own with a pronominal function as in they ought to preyse and loue the chirche and the commaundements of the same ‘the church and its commandments’ (p. 152, example from Caxton's Aesop's Fables). Both phrases were frequently used as anaphoric expressions in sixteenth-century English, but then dramatically decreased in frequency. In present-day English, the two expressions are infrequent, largely limited to legal texts, and both have a variant without the definite article. The shared general function, overlapping general timing of decline and similar distribution in present-day English prompt the question whether similar factors were responsible for their demise. Leung and van der Wurff investigate this question based on thoughtfully sampled sets of historical data drawn from the Early English Books Online corpus. The analysis and discussion are carefully framed in the context of existing analyses of the said N and the same and studies of their historical development. This background knowledge enables them to interpret the decline of the anaphoric uses in relation to other, non-anaphoric uses of the same phrases. Leung and van der Wurff use these case studies to address more general questions about the causes behind the decline of grammatical items, which, in contrast to causes driving innovations, have not received much scholarly attention. They propose that the decline of the two expressions, which they show to be less similar than initially assumed, can be motivated by individual explanations at the local lexical level. They argue that the different explanations have a common ground in that they all involve the lack of or breakdown of a proper fit between form and function. For example, for the same, they invoke the close similarity in form and distribution to a different, more frequent use of the same and the resulting high-processing cost in distinguishing its correct meaning and function in context as one potential factor, and the same's more cumbersome phonological shape compared to pronouns with a similar function such as it as a second factor. Leung and van der Wurff point out that Postma (Reference Postma, Lucas, Watts, Breitbarth and Willis2010) identified the suboptimal fit between form and function as a motivation in his study of ‘failed changes’ too. This provides an intriguing hypothesis for further studies of loss and decline, which is an area of growing interest in diachronic linguistics.

The final paper in the volume is ‘That-complementiser omission in N + BE + that clauses: Register variation or constructional change?’ by Annette Mantlik and Hans-Jörg Schmid. This paper takes us away from the noun phrase as such to investigate a case of clausal complementation. The omission of complementiser that after lexical verbs (I think that he has arrived versus I think he has arrived) has been studied intensively, and Mantlik and Schmid take this body of work as starting point for their investigation of that-omission in a related context, following a shell noun + is/was, e.g. the truth is (that), the problem is (that). They seek to test two opposing hypotheses: (1) omission is a matter of style and genre; (2) omission is one feature of the development of a new construction. The data they use are mainly derived from two sources: pre-1810, they use a self-compiled corpus called the Classics corpus which they probe qualitatively; post-1810, they use the Corpus of Historical American English to conduct a quantitative study. The hypotheses are tested indirectly, that is, no comparison of different genres/registers/individual authors is done. Instead the null hypothesis is that if extra-linguistic factors determined the presence/absence of that, the two variants will covary in a systematic way in terms of quantitative development, collocational range, semantic range and pragmatic restrictions. The further lack of delineation or discussion of what is meant by stylistic factors, as illustrated amongst others by the variety of terms used (style, genre, register, medium), leaves the impression that it is the hypothesis of constructionalisation/constructional change that the paper is genuinely investigating. With regard to this hypothesis, Mantlik and Schmid find that there is indeed evidence for constructional change. They argue that there is a more specific form ‘the + N + is + zero’ associated with a particular pragmatic function, namely, an argumentative function, found with a semantically definable subset of nouns. In the construction, the clause following the copula is upgraded to main clause status, and there are examples in which form is further reduced to ‘N is, …’. An early instance of this emerging construction is found with the noun truth from the seventeenth century onwards. The construction expanded to other nouns such as fact, thing, point and trouble, with the most recent decades of the COHA corpus (1980–2000) seeing a drastic increase in frequency of zero and expansion to other nouns. The findings reveal interesting parallels with epistemic parentheticals and other discourse markers. They also raise intriguing questions about the role of the truth as a specific catalyst for the development of a new construction. The origin of the truth is can likely be explained independently with reference to a well-documented process of grammaticalisation giving rise to a semantic set of discourse markers, e.g. to be honest, truth be told, but how precisely it – and other early common combinations without that – influenced the development of the new construction through analogy and generalisation is an issue that remains to be fleshed out.

Overall, this is a high-quality collection of papers that is a must-read for researchers working on the noun phrase, but that is also of wider interest to researchers working on English morphosyntax, especially in a diachronic perspective. The papers share a commitment to empirical accountability and a beneficial willingness to look beyond grammar for explanations. Two important trends in this respect are the accumulating evidence for the importance of lexical items in the development and functioning of grammatical structures (papers by Davidse, Hundt, Leung and van der Wurff, Mantlik and Schmid, and probably also Günther) and the cross-overs into sociolinguistics, with the inclusion of register, genre and stylistics (González-Díaz, Leung and van der Wurff, Mantlik and Schmid) and the application of methods of variationist sociolinguistics (Hundt). The collection as a whole showcases the noun phrase as a fruitful subject for the testing and formation of hypotheses around the modelling of grammatical constructions and grammatical change.