It is generally agreed that world languages can be divided into ‘classifier’ and ‘non-classifier’ types. English is usually considered a ‘non-classifier language’ in mainstream typological studies (Aikhenvald, Reference Aikhenvald2000; Allan, Reference Allan1977), and is set in opposition to a ‘classifier language’ like Chinese. English Quasi-Numeral Classifiers: A Corpus-Based Cognitive-Typological Study, the 226th volume in the book series Linguistic Insights: Studies in Language and Communication, is a re-examination of this traditional typological division from Croft’s (Reference Croft2003) “functional-typological” perspective (p. 15). In this volume, author Xu Zhang re-thinks the story of ‘non-classifier language’ and provides a novel notion of the ‘English Quasi-Numeral Classifier’. To this end, on the basis of a cognitive approach and a corpus-based study, she probes into the numeral classifiers (NCs) in English from a semantics-based and functional perspective, and investigates how far English Quasi-Numeral Classifier structures (QNCs) encode the function of categorization. In doing this, the author has provided new insights into the typological discussion on ‘classifier languages’ versus ‘non-classifier languages’.
This monograph is organized in seven substantial chapters, consisting of an introduction, two chapters of theoretical exploration, one on methodology, two chapters of case studies, and a conclusion. Chapter 1 presents the core purpose, perspectives, research questions, and the outline of the book. The author points out that, in a functional-typological background and taking a cognitive approach, this study mainly focuses on the categorization device of NCs, which has previously been exclusive to the so-called ‘non-classifier languages’, and the categories realized by the usage of Dimensionality-based QNCs (D-QNCs). Chapter 2 proposes the essential concept of English QNC, in the functional-typological background with the system of Chinese NCs as a reference point. It first provides a brief review of classifier systems, with a special focus on the prototypical type (Lucy, Reference Lucy and Senft2000, p. 329) of nominal classification, and a widely conceived normative and comprehensive system, i.e., the Chinese NCs system. NCs serve to quantify noun referents inherently, and also qualify or categorize them. NCs performing mainly a quantification function are classified into Mensural NCs (e.g., yī gōngjīn miànfěn ‘one kilogram of flour’), while those performing a mainly categorization function are regarded as Sortal NCs (e.g., yī háng shù ‘one row of trees’). With a discussion of the intermediary forms (such as yī pán shéngzi ‘a coil of rope’) and their semantic features, Zhang proposes a gradient relationship between Sortal and Mensural NCs, as shown in Figure 1. Through positioning Chinese mono-morphemic nominal NCs into the Mensural–Sortal NC dichotomy, and analyzing the semantic parameters of the Sortal ones, Zhang finds that most of the Chinese NC types have English counterparts, which paves the way for the investigation of English QNCs.
Sortal and Mensural NCs on a Quality–Quantity Continuum (p. 41)
The author’s next step is to adopt Croft’s (Reference Croft2003, p. 15) “cross-linguistic identification” of “functional equivalents” (p. 76), which leads to defining English NC-like words in the same way as Chinese NCs proper. Morphosyntactically, these words occur in the N1 slot of quantitative binominal phrases ‘Det+N1+of+N2’, which include partitive constructions (e.g., a bunch of the flowers) and pseudo-partitive constructions (e.g., a bunch of flowers). Similar to Chinese NCs, the category of English QNCs is fuzzy, intersecting with quantifiers at one end, and blending with nouns at the other, as shown in Figure 2.
QNC-Cs on a semantic and grammaticalization continuum (p. 75)
After the functional identification of English QNCs with NCs, Chapter 3 concentrates exclusively on the theoretical framework of the study, namely the cognitive categorization of English QNCs. In this chapter, the author first explores the categorization function of NCs from a cognitive-grammatical point of view (Langacker, Reference Langacker2008; Taylor, Reference Taylor2002). Then she critically reviews the Aristotelian model, prototype category model, and schema category model on how previous research identifies prototypes and extends prototypes to members. Arguing that NCs are schematic and NC category members share common features, the author proposes a modified feature-based schema category model to investigate the ways in which category members show the schematic feature designated by the NC, and to seek the underlying cognitive motivations such as salience, metaphor, and metonymy. As the spatial identity of entities is a key semantic parameter motivating NC/QNCs, the author takes Dimensionality construed by QNCs, especially one-dimensional and two-dimensional QNCs (1-D and 2-D QNCs), as the locus of the investigation. It is noted here that all salient dimensions of noun referents are construed in human conceptualization and are thus extremely subjective. Applying the modified category model, the categorization process of D-QNCs will be clarified in case studies from five aspects, viz. the schematic common features for QNC categories, category members, degrees of membership, motivations of categorizations, and frequency of members.
In Chapter 4, the author lays the methodological foundation for the empirical investigations presented in the following two chapters. A corpus-based approach is adopted in this study, with the British National Corpus (BNC) as the main data source for English QNC expressions. A substantial number of QNCs are also gathered from grammar books and related monographs, and are further classified into semantically based types. With 1-D and 2-D QNCs being the research focus, three of the most frequently used 1-D QNCs (thread, strip, column) and two of the most frequent 2-D QNCs (sheet, slice) are chosen for subsequent empirical studies. To represent the overall usage of the particular D-QNC, the concordances and collocations of these QNCs are extracted and analyzed by BNCweb with an exploitation of their occurrence patterns.
On the basis of the feature-based schema category model and the specific methodology outlined in Chapters 3 and 4, five empirical case studies are conducted in Chapters 5 and 6, investigating the categorization process of 1-D QNCs and 2-D QNCs, respectively.
Chapter 5 focuses on the case studies of three types of 1-D QNC categories: thread, strip, and column. According to the author, all THREAD, STRIP, and COLUMN categories categorize detached objects with radiating membership. In terms of categorization motivations, it turns out that members of the THREAD and COLUMN categories are significantly categorized by ontological metaphor and metonymy (e.g., a thread of forgiveness and columns of modified muscle cells), while the STRIP category is mainly based on salience (e.g., a dusty white strip of road), and therefore the STRIP category is less heterogeneous and diverse than the other two. The schematic features of QNC categories are numerous, in that these three categories manifest varied semantic features such as dimensionality, directionality, and size, although the degree is different. Moreover, concerning the potential to serve as a QNC, the author notes that strip is more likely to be involved in a quantitative construction, while column is subject to a common noun, and only half of the thread cases could be counted as QNCs.
Chapter 6 analyzes sheet and slice, two cases of 2-D QNC categories. The SHEET category embraces semantic features of ‘flatness’ and ‘largeness’, and more than half of the category members are paper-like objects (e.g., sheets of tissue). In the case of the SLICE category, however, over half of the members are food, and an action of cutting plays a significant role in its semantic features, which redresses Lehrer’s (Reference Lehrer1986, p. 117) claim that slice is “a completely conventionalized classifier, based on shape”. In the five case studies, first, all categories display a gradient structure and more D-QNCs take all salience, conceptual metonymy, and metaphor as their cognitive motivations; second, the high frequency of some category members is partly because their noun referents are frequently perceived as containing the corresponding dimensional feature in reality; third, the so-called Dimensionality-based QNC categories, except for the STRIP category, usually combine multiple semantic parameters which enable them to embody quite abstract members.
Chapter 7 concludes the findings on NCs mainly from two perspectives. From the cognitive perspective, D-QNC categories encompass both the SPACE domain and domains far beyond spatiality. The degree of spatiality of a member corresponds to its membership and the clarified schematic features. More spatial members have higher degrees of membership and more prominent qualitative functions. The reverse is true for members containing less spatial property. Concerning categorization motivations of D-QNCs, what is noteworthy is the metonymy of COMPONENT FOR THE COMPOSED and ontological metaphors of which D-QNCs can serve either as direct linguistic representations or as merely an entailment. From the functional-typological perspective, it is argued that the measurization devices typically represented by NCs proper “may exist universally across languages” (p. 259), and they form a new grammaticalization continuum, namely “Nouns–Sortal NCs–Mensural NCs–Quantifiers”, with English less restricted and grammaticalized than Chinese. As is shown in Figure 3, Chinese NCs appear to be located more on the Quantity/Quantification end, while more English QNCs are on the Quality/Categorization end. It is undeniable, however, that both the so-called ‘non-classifier language’ (English) and the ‘classifier language’ (Chinese) have measurization devices and can be situated on the same semantic map, sharing similar functions (categorization and quantification), conceptual space (quality and quantity), and an identical ordered nature.
A mini-semantic map of Chinese NCs and English QNCs (p. 261)
To sum up, this monograph has made significant contributions to NC research in several aspects. First, it provides an articulate outline of NCs, contributes to the understanding of NC functions in view of semantic parameters, and proposes a consistent and exhaustive functional Chinese NC taxonomy. Second, it has presented a novel effort to elucidate English NCs in a cognitive-grammatical and typological context, supported by an impressive body of corpus-based empirical research. Third, with a special focus on Dimensionality, a coherent set of investigations consistently show how English speakers spatially construe the world, which will also produce a better general recognition of human conceptualization. Last but not least, the analysis provides strong evidence for the presence of apparent linguistic quantification and categorization devices in ‘non-classifier’ English. Instead of setting Chinese and English in opposition, the establishment of the innovative notion of English QNCs puts them on a rather continued, comprehensive, and graded Quantity–Quality continuum. This not only helps to unveil the quantification and categorization processes in English, but also provides new insights into the interpretation of human quantification and categorization processes, and thus reveals generality across languages.
Classifiers and noun categorization systems have long been an object of linguistic investigation. A number of scholars (McEnery & Xiao, Reference McEnery and Xiao2012) have noticed the classifying or quantifying constructions in English that are somewhat equivalent to Chinese classifiers. What makes this monograph significantly different is its systematic functional-typological analyses of the distinction between Sortal and Mensural NCs/QNCs, both intra-linguistically and cross-linguistically, as well as between their quantifying and qualifying functions. As stated in Croft (Reference Croft2003, p. 2), “linguistic structure should be explained primarily in terms of linguistic function”. The identification of functional equivalents of NCs and the typological demonstration in a mini-semantic map will carry implications for cross-linguistic studies.
It could be added that this book also has certain restrictions. Regarding the proposed feature-based schema category model, for example, the judgment of features of NC/QNCs is mainly based on intuition and introspection with a dearth of evidence from experimental data, which appears to cause a slight degree of subjectivity. Taken as a whole, all human languages have linguistic devices to express quantity, even though different languages vary from each other in the quantification modes. The work distinguishes English QNCs from common nouns and indicates that the absence of a specialized category in one language does not necessarily lead to the absence of a similar category to play the same role. Due to its interdisciplinarity and cognitive-typological significance, this book can be expected to provide a useful reference for analogous studies in other languages and also provide new insights into English teaching or language pedagogy.
