Syntax is an important subsystem of language system, and quantitative linguistics is a sub-discipline of linguistics addressing, by virtue of real texts and precise mathematical means, the laws constraining language structure and evolution. However, quantitative approaches are now rarely used in “mainstream” syntactic studies, which is a big disadvantage for the attempts to empirically unveil the operating syntactic mechanism of human language or statistically parse the syntactic structure. This book is the first book that tries to bridge gap between quantitative linguistics and syntactic theories.
In the first chapter, the Introduction, the author emphasizes the importance of syntax to human language, but at the same time particularly makes a point that “this book is not based on any a priori statements about properties of language that will not immediately be tested on empirical data” (p. 2). The author believes that a text, regardless of its length, is an embodiment and representation of a complex cognitive system with multiple structures. Though such an embodiment or representation may be affected by various competing factors, it is eventually shaped by speakers' need for the least effort in production and the hearers' need for the convenience in extracting conceptual structures from the linguistic input. This is the author's rationale, which underlies the endeavor to discover, through mining real texts, the structures, laws, and principles of human cognition and languages. In this sense, language is not a totally autonomous symbolic system independent of human beings, but a dynamic one serving people in information communication and knowledge transmission.
In the second chapter titled “The quantitative analysis of language and text,” the author defines the concern of quantitative linguistics as “the multitude of quantitative properties which are essential for the description and understanding of the development and the functioning of linguistic systems and their components” (p. 9). In other words, the objective of quantitative linguistics is no different from that of other linguistic disciplines. What is different is the methodology: Quantitative linguistics distinctively attaches great importance to the application of scientific and quantitative means for promoting precision of linguistic researches. In this chapter, the author also discusses some fundamental issues concerning quantitative linguistics, such as methodological grounds, epistemological aspects, etc., which are important to establish “quantitative linguistics as a scientific discipline,” and expounds, in terms of philosophy of science, the relations among theory, laws, and explanation. This chapter introduces to readers the importance and necessity of quantitative means and models to linguistics and study of texts, proving that, as a scientific discipline, linguistics must be solidly based on quantitative methodology.
At the level of syntax, “Empirical analysis and mathematical modeling,” which is the title of Chapter 3, can never do without a clear definition of basic syntactic units and corresponding properties. The author gives a long list of syntactic units and properties that are of significant value to linguistic study (pp. 28–29). This list, though not exhausting for all possible syntactic units, covers all the important ones in terms of both phrase structure (PS) syntax and dependency syntax, and hence cannot only reflect but also well serve the present syntactic studies. It is worthy to note that in this list, some units, such as semantic roles, are not, strictly speaking, syntactic units or properties, though the semantic valency of a verb is easily associated with semantic roles. Once the basic units of quantitative syntactic research are determined, another importance issue is how to quantify and measure these units. In this regard, the author holds that the quantification and measurement basically consist in taking count of these units in texts, for linguistics is overwhelmingly concerned with discrete units.
Quantitative syntactic research, different from those at the lexical level, involves two specific problems. One is the choice of grammatical system, and the other is the syntactic resources needed for research. Once a certain syntax (annotating scheme) is chosen, the next step is to build a treebank. In short, quantitative syntactic studies are based not only on syntactic theories but also on available languages resources. In recent years, the mushrooming syntactically annotated corpora (treebanks) guarantee a fruitful perspective of quantitative syntactic researches.
Section 3.3, in which the author discusses the acquisition of data from linguistic corpora, introduces some forms and formats commonly used in syntactically annotated corpus, such as tagged text, treebanks, column structure, and feature value pairs. Section 3.4, which is about syntactic phenomena and mathematical models, deals with topics such as sentence length and relevant models, word classes and their distributions, frequency spectrum and rank-frequency distribution, Frumkina's law on the syntactic level, type token ratio, information content, motifs, and Gödel Numbering. This section is one of the most important sections of this book (pp. 42–135), covering almost all important works in the field of quantitative syntactic study. The author elaborates on nearly all these important studies with abundant data, models, graphs, and diagrams. This detailed review objectively reports both achievements and limitations of quantitative syntactic studies. The majority of the works introduced in this section are based on PS syntax. Dependency syntax, compared with the PS syntax, has greater universality, reflects more faithfully the psychological mechanism of language processing, and is more convenient and suitable for transition from the syntactic level to semantic level. Therefore, quantitative syntactic researches based on the dependency syntax are of great value to linguistic study. To make up for this deficiency, the author particularly devotes Subsection 3.4.9, entitled “Dependency grammar and valency,” to the introduction of some works within the theoretic framework of dependency syntax, although some recent interesting works on dependency and valency are not included in it.
In terms of philosophy of science, the syntactically related distributions and mathematical models introduced in Chapter 3 are just preparations for the construction of a scientific linguistic theory. That is why Chapter 4 is entitled “Hypotheses, laws, and theories”: these are the three necessary stages along which a scientific discipline develops.
In Subsection 4.1.4, which is entitled “Distributions of syntactic properties,” the author reports the distributions of some syntactic properties, such as complexity, depth, length, and position. The title of Section 4.2 is “Structure, function, and processes,” which reveals that quantitative linguistics holds a dynamic and systematic view of language. The author argues that theories are systems composed of universal laws, without which it is impossible to explain the phenomena observed. An axiom of synergetic linguistics is that language is a self-organizing and self-regulating system – a special form of dynamic system with special properties (p. 169). After a brief introduction of the foundation and general principles of synergetic linguistics, the author proceeds to build a synergetic linguistic model in Subsection 4.2.7.
In this model, the syntactic units were syntactic constructions defined in terms of phrase structure syntax. At the syntactic level, eight properties of syntactic constructions and four inventories were selected to build into the model (pp. 186–187). The hypotheses assumed in this model are verified with one English corpus, the Susanne corpus, and one German corpus, the NEGRA corpus. The specific procedures of building a subsystem of synergetic linguistics are as follows setting up axioms; searching for functional equivalents that can meet the requirements and determining on their effects on other system variables; and setting up the empirical testing of hypotheses.
The syntactic synergetic subsystem resulting from the above procedures is presented in Figure 4.3.3 (p. 201). In addition, the author particularly mentions a few fields where the approaches recommended in this book may play significant roles: typology, grammar, and pragmatics/discourse analysis.
The author hypothesizes that the quantitative researches based on corpora of other languages will arrive at similar results as those obtained with the English corpus. He was sure that the theoretical studies as reported in this book will bring significant benefits to some fields in applied linguistics, such as language learning and teaching, compiling minimal vocabularies, determining the text coverage of dictionaries, building better syntactic parsers, etc.
In the concluding part of this book, the author especially points to the fact that the numerous hypotheses advanced in this book are still waiting to be tested on typologically different languages and new hypotheses are yet needed to consummate extant subsystems. An even bigger challenge is the integration of extant quantitative models with psycholinguistics, neurolinguistics, sociolinguistics, phonetics, etc.
Undoubtedly, this book will play significantly positive role in the following topics of research that are relevant to quantitative syntactic studies: (1) the mechanism and laws that governs the combination of words into sentences; (2) a syntactic parsing model, which can process the real texts of various languages; (3) the interrelations among various levels of language with syntax at the center; (4) further explorations into the laws that shape the structure and the evolution of human languages; (5) the application of discoveries obtained in quantitative researches to the syntactic parsing of real-world language usage; (6) the relations between syntactic quantitative discoveries and human cognition; (7) the quest for a synergetic language model that can tally better with discoveries from psychological and cognitive experiments; and (8) the construction of a lexical–syntactic synergetic model with a stronger explanatory power.
What is more, extensive references and a complete index of subjects and authors appended at the end of the book are of great help for readers to further go into the issues discussed in the book.
In conclusion, this book opens a door to the scientific syntactic investigation. There are certainly still many challenges and unanswered questions, which, however, are just disguised opportunities for those who endeavor to uncover the laws shaping the syntactic structure of human language. It is a valuable resource for those developing syntactic parsing systems or searching a statistical usage-oriented syntactic theory.
