The approach followed

Having established the theoretical framework for this study—in terms of semantic types, semantic roles, their mapping onto syntactic relations, and so on—I worked inductively, examining the semantic and syntactic properties of a large number of individual verbs, and gradually inducing generalizations from these.

I began with a list of the 2,000 most commonly used words in English (in West 1953) and looked in detail at all those which can function as verbs (about 900 in all). Each verb was taken separately, and its semantic and syntactic characteristics investigated. The verbs were grouped into types— on the basis of semantic and syntactic similarities—and the semantic and syntactic profiles of each type were then studied. In this way—proceeding from the particular to the general.

The present volume should be regarded as the first attempt to follow through a new approach to grammatical description and explanation. It is essentially programmatic, providing a broad outline of the semantic types, and the ways in which their meanings condition their syntactic properties.

The reader will not find a fully articulated grammar of English from a semantic viewpoint, with the meaning of every important verb discussed in appropriate detail, and each syntactic construction dealt with exhaustively. Such a study would fill a dozen or more volumes of this size. Rather, I try to provide the parameters in terms of which more detailed studies—of individual semantic types, and of individual constructions—may be carried out. This book aims to lay a foundation, upon which elegant edifices of semantico-syntactic description and explanation may be constructed.

Studying syntax in close conjunction with semantics, and in an inductive manner, differs from the approach followed by many modern linguists. It is most common to begin with syntax (looking for ‘semantic interpretation’ at a late stage, if at all) and also to begin with putative generalizations, later looking to see if there might be any counter-examples to them.

I noted above that there is a many-to-one mapping between semantic types and grammatical word classes, and also between semantic roles and syntactic relations. Textbooks of grammar will typically note that both hope and believe accept a THAT complement in the O slot (Susan hopes that she will win the race, Susan believes that she will win the race) and then express surprise that only hope takes a to complement clause in which the subject is not stated but is understood to be the same as the subject of the main verb (Susan hopes to win the race, but not *Susan believes to win the race). They appear to begin with the premiss that if two words share some grammatical properties then they might well be expected to share them all. Linguists who argue in this way generally pay only perfunctory attention to meaning.

A more rewarding approach is to commence with consideration of semantic types. Believe belongs to the THINKING type, together with think, reflect, wonder, doubt, all of which take THAT complements, but not TO complements of the type illustrated here. Hope can be semantically grouped (in the WANTING type) with dread, desire and wish, all of which take both THAT complements and also to complements (with omission of a subject that is identical to the subject of the main verb).

Although there are important differences—as just illustrated—my approach does of course have many points of similarity with the work of other linguists. I have tried to build on all previous work (and add to it) rather than to ignore the insights of other scholars and strike off in some idiosyncratic direction of my own.

One idea that has been taken from the Greek tradition (being also used in the early ‘transformational theory’ of Chomsky), and adapted to the needs of the present study, is the usefulness of recognizing ‘underlying forms’, and then general conventions for omitting or rearranging parts of them in specifiable circumstances. (In fact, I go a good deal further than many linguists along this path, and am consequently able to explain things that others have dismissed as perverse irregularities.)

To mention one example, there are a number of verbs in English which must take a preposition and a following NP, e.g. decide on, rely on, hope for, refer to, object to. This NP behaves like a direct object (e.g. it may become passive subject). I suggest that decide on, hope for, and the like are each a transitive verb, involving an inherent preposition. There then appears to be a general rule of English syntax stating that a preposition must be omitted when it is immediately followed by one of the complementisers that, to and for. Compare (1)–(2) with (3)–(4):

(1) Everyone in the office hoped for an English victory

(2) They decided on the order of precedence

Here there is an NP in O slot, and the prepositions for and on are retained.

(3) Everyone in the office hoped that England would win

(4) They decided that Mary should lead the parade

Here the O slot is filled by a THAT complement clause, before which for and on are omitted. The fact that there is an underlying preposition in (3) and (4) is shown under passivization. The THAT clause, as object, is moved to the front of the sentence to become passive subject and the preposition again appears as the last part of the verb:

(5) That England would win was hoped for

(6) That Mary should lead the parade was decided (on)

But note that a THAT complement clause in subject position is typically extraposed to the end of the sentence, with it then occupying the subject slot. When this happens the THAT clause again follows the inherent preposition of the verb, which is omitted:

(7) It was hoped that England would win

(8) It was decided that Mary should lead the parade

There are fashions and fads in linguistic explanation. At one time it was all the rage to talk of underlying forms and deep structures and ways in which surface forms and structures could be derived from these. Nowadays some scholars are reluctant to work in such terms. My explanations in terms of underlying forms and structures could perfectly well be restated as alternations between two possibilities—saying that hope for is used in certain environments and hope in other, complementary environments, for instance, without suggesting that hope is derived by prepositional omission from hope for; and similarly in other cases. This is essentially a matter of ‘terminology’, carrying no differences in descriptive or explanatory power. The approach I adopt involves shorter statements and seems pedagogically more effective; but nothing else hangs on it.

There is one respect in which I differ from the early practitioners of Chomsky’s ‘transformational grammar’. They might say that I believed him to be mad has a ‘deep structure’ something like [I believed [it [he is mad]]] with the third person singular (3sg) pronoun as subject of be mad, but after a ‘raising’ transformation has been applied the 3sg pronoun is now the object of believe. I suggest that the 3sg pronoun bears two simultaneous syntactic relations, as object of believe and as subject of be mad (even though it is morphologically encoded as the unmarked and object form him).

I have tried to make this book consistent and self-sufficient. In particular, I have not ventured to recapitulate every previous attempt to deal with the syntactic questions I consider, and to criticize aspects of these before presenting my own solution (which is, in most cases, partly based on earlier work and partly original). To have done this would have made the book two or three times as long and much less easy to read.

Finally, let it be stressed that I am describing educated British English— essentially my own dialect of it (which is based on what I learnt as a child in Nottingham, slightly modified by several decades of residence in Australia). I am fully aware that other dialects, such as American English, differ markedly—more so concerning the topics discussed here than concerning the topics dealt with in traditional grammars. These differences should not affect the broad sweep of conclusions reached in this book, only their detailed articulation. It would be an interesting and rewarding task to investigate dialect differences in terms of the framework adopted here; this remains a job for the future.