SIL Electronic Book Reviews 2005-006

Theoretical notations

As a student myself, and afterwards as a researcher and teacher of syntactic phenomena, I have been frustrated with a seemingly endless series of syntactic “theories” whose chief goal appeared to be the construction of a representation language for syntactic description. The result has been a continuing kaleidoscope of notations which have made even five-year-old journal articles—and many reference grammars—difficult to decipher. (p. xiii)

Instead of developing theoretical notations for representing syntactic constructions, Croft has adopted a SEMANTIC MAP model. He states that one goal of RCG “is to represent universals of human language in a way that is plausible as a model of the knowledge of an individual speaker of a particular language.” He goes on to say that the Semantic Map model has been recently embraced by typologists for the representation of language-particular grammatical knowledge in the context of universal patterns of variation. This representation model is one in which the distributional patterns of language-particular categories are mapped onto conceptual space with some structural properties that are hypothesized to be universal.

The relationship between proposed analyses of languages and empirical reality

Croft cites three types of research articles that he believes give evidence for the lack of defining a clear relationship between a given syntactic model, proposed analyses, and empirical reality. This is what he says about these research articles:

A large class of research articles…both “formalist” and “functionalist” is devoted to claiming that phenomenon X in language Y really is a passive, or really is not a passive. Such research discounted or ignored the opposing evidence…. It also missed the point, which was that phenomenon X was interesting and challenging precisely because it sort of was a passive but sort of wasn’t at the same time; both its passivelike and its unpassivelike syntactic properties were equally important.” (pp. xiii–xiv)

Methodology and argumentation

In Chapter 1, Croft describes the general characteristics of Radical Construction Grammar by placing the theory in the context of syntactic argumentation. He poses two questions:

1. Theoretical. What is the nature of the grammatical knowledge that the speaker has in his/her head, and how should it be represented?

2. Methodological. Is there a general language universal method for justifying the existence in a particular language of the syntactic elements—categories and relations—that are the basic units of syntactic theory?

Throughout the remaining chapters of the book, Croft discusses the answers other syntactic theories give to the first question but claims that the second question is largely ignored. He covers the main categories of grammar and proposes analytical methodologies and analyses of data from the RCG perspective. Titles of chapters indicate the range of issues discussed:

• Chapter 2: Parts of Speech
• Chapter 3: Syntactic Categories and Semantic Relativity
• Chapter 4: Clausal Syntactic Roles (Grammatical Relations)
• Chapter 5: Dependency, Constituency, and Linear Order
• Chapter 6: A Radical Approach to Syntactic Relations
• Chapter 7: Heads, Arguments, and Adjuncts
• Chapter 8: The Voice Continuum
• Chapter 9: The Coordination-Subordination Continuum

Rather than attempt to summarize each chapter, I cannot do better than to quote excerpts from Chapter 10 where Croft summarizes RCG in five points. He introduces the five points like this: “Radical Construction Grammar is disarmingly simple—it is a genuinely minimalist model of syntactic representation” (p. 362).

1. Primitive grammatical units in the model. “The only type of primitive grammatical units are CONSTRUCTIONS—pairings of form and meaning which may be atomic or complex, schematic or substantive” (p.362).

2. Primitive syntactic relations in the model. “The only type of syntactic relations within a construction are the MERONOMIC (part-whole) relations between the construction on the one hand and the ELEMENTS that fill the ROLES of the construction. Since syntactic roles are defined relative to constructions, there is no universal finite inventory of syntactic role types” (p. 363).

3. The relation between form and meaning. “A construction joins its form and meaning via SYMBOLIC relations. Symbolic relations hold between the construction’s formal structure as a whole and its meaning as a whole. Symbolic relations also hold between the elements of a construction’s formal structure and COMPONENTS of its semantic structure” (p. 363).

4. Generalizations within and across languages. “One type of generalization within languages are categorizing relations…A second type of generalization is found across constructions and across languages. These are systematic patterns of variation, such as PROTOTYPES and IMPLICATIONAL HIERARCHIES that characterize cross-constructional and cross-linguistic diversity and constrain the distribution and even the form of constructions used for particular functions” (p. 363).

5. Explanations of linguistic generalizations. “The patterns described above are accounted for by the hypothesis of a largely universal CONCEPTUAL SPACE which represents structures in the human mind. Distribution patterns are overlapping SEMANTIC MAPS on conceptual space. Language universals are accounted for by constraints on the topography of conceptual space which predict what sort of patterns can be found within and across languages, and what sorts of diachronic grammatical changes may occur” (p. 363).

In general, this book confronts us with the reality that neither formal nor functional syntactic theories have been able to adequately account for the diversity of the syntactic facts of a single language or the syntactic diversity of the languages of the world. A fundamental principle seems to underlie RCG: the syntactic facts of a single language must not be forced to fit a theory; instead a theory must be adjusted to fit the facts of single languages. RCG has been developed to provide a conceptual and semantic framework for accounting for the syntactic categories and structures of single languages, as well as across languages.

The book contains a total of 105 graphic representations of the following:

• Theoretical generalizations, contrastive models of syntactic analyses
• General conceptual space and semantic maps
• Language-particular conceptual space representations

Of the 105 representations, 71 are figures and 34 are tables. Most of the captions clearly indicate the content of each figure and table. The following examples show the broad range of grammatical categories and analytical issues covered in the book.

• “The relation between syntax, semantics, and conceptualization” (Figure 3.3)
• “Conceptual space for participant roles in transitive and intransitive situations” (Figure 4.2)
• “Semantic map for Yurok Bipersonal, Unipersonal, and Passive constructions” (Figure 8.8)
• “Distribution of arguments across tests for Direct Object status” (Table 1.4)
• “Semantic properties of prototypical parts of speech” (Table 2.2)
• “Distribution of case marking in Yuwaalaraay” (Table 4.1)

The figures and tables are scattered throughout nine of the ten chapters, along with data from 200+ languages. RCG analyses of the data are described and contrasted with the analyses based on other syntactic theories or models.

Croft claims that RCG is “disarmingly simple,” but reading this book is not. However, the issues he raises are extremely important to syntactic theory, and for that reason, his claims and arguments ought to be processed and discussed with open minds by linguists with differing theoretical persuasions.