sentence (n.)

The largest STRUCTURAL UNIT in terms of which the GRAMMAR of a LANGUAGE is organized. Innumerable definitions of sentence exist, ranging from the vague characterizations of TRADITIONAL grammar (such as ‘the expression of a complete thought’) to the detailed structural descriptions of contemporary LINGUISTIC analysis. Most linguistic definitions of the sentence show the influence of Leonard Bloomfield, who pointed to the structural autonomy, or independence, of the notion of sentence: it is ‘not included by virtue of any grammatical construction in any larger linguistic form’. Research has also attempted to discover larger grammatical units (of DISCOURSE, or TEXT), but so far little has been uncovered comparable to the sentence, whose constituent structure is stateable in FORMAL, DISTRIBUTIONAL terms.

Linguistic discussion of the sentence has focused on problems of identification, classification and generation. Identifying sentences is relatively straightforward in the written language, but is often problematic in speech, where INTONATION and PAUSE may give uncertain clues as to whether a sentence boundary exists. Classification of sentence structure proceeds along many different lines, e.g. the binary constituent procedures of IMMEDIATE-CONSTITUENT analysis, or the HIERARCHICAL analyses of HALLIDAYAN and other grammars (sentences being seen as composites of CLAUSES, which in turn are analyzed into PHRASES, etc.). In GENERATIVE grammar, likewise, there are several models of analysis for sentence structure, with competing views as to the direction in which a sentence DERIVATION should proceed. Certain analytic problems are shared by all approaches, e.g. how to handle ELLIPTICAL sentences (or ‘sentence fragments’), such as To town (in answer to Where are you going?); how to handle cross-reference between sentences, such as She’s writing (‘sentence CONNECTIVITY’); and how to handle the MINOR, non-PRODUCTIVE sentence types in a language (e.g. Yes, Please, How do you do?).

Most analysts agree on the need to recognize a FUNCTIONAL classification of sentences into STATEMENT, QUESTION, COMMAND and EXCLAMATORY types. There is also widespread recognition (albeit with varying terminology) of a formal classification into DECLARATIVE, INTERROGATIVE, IMPERATIVE and EXCLAMATIVE types.

Most analyses also recognize some such classification of ‘sentence patterns’ into simple v. complex or compound types, i.e. consisting of one SUBJECT–PREDICATE unit, as opposed to more than one. Whether one calls this subject–predicate unit a CLAUSE or a ‘simple’ sentence, or uses some other term depends on one’s model of analysis – but something analogous to this unit emerges in all theories, e.g. NP + VP, ACTOR–ACTION–GOAL, Subject–Verb–Object. Likewise, the number of formal sentence types recognized, and how they are best defined, has been and remains controversial. Several linguists insist on making a systematic distinction between sentence (a theoretical unit, defined by a grammar) and UTTERANCE (a physical unit, a matter of speech production or PERFORMANCE): in this view, utterances can be analyzed in terms of sentences, but utterances do not ‘consist of’ sentences.