SENTENCE The largest structural unit normally treated in GRAMMAR. The sentence is notoriously difficult to define; numerous definitions have been offered and found wanting. The classical definition, that a sentence expresses a complete thought, dates from the first Western treatise on grammar, by Dionysius Thrax (
c.100 BC), whose interest lay primarily in analysing, using, and teaching written Greek. This traditional notional definition, however, only solves the problem by transferring it: how does one define a complete thought? Linguists and anthropologists in the 19–20c, trained in the Greco-Latin grammatical tradition and faced with the analysis of previously unwritten Amerindian languages, have often noted how difficult it is to establish the boundaries between what might be words and what might be sentences in some of those languages. Because of this, the sentence as understood in the Western linguistic tradition has not yet been unequivocally established as a universal of language.
The syntactic sentence
Formal definitions usually refer to the structural independence of the sentence: that it is not included in a lager structural unit by such devices as coordination and subordination. However, dependence is relative: what are generally recognized as sentences may be dependent to some extent on other sentences through such devices as pronoun substitution and connective
ADVERBIALS:
therefore,
however,
yet. Elliptical sentences such as
Tomorrow (in answer to
When is your birthday?) are clearly dependent in some sense on linguistic context. There are also problems in deciding the status of formulaic utterances such as
Yes or
Good morning, which in dialogue are complete in themselves. Formal definitions may also refer to the internal structures of sentences. Indeed, it is possible to recognize as canonical sentences those that conform in their structure to the normal clause patterns, such as subject-verb-direct object. Other constructions would then be considered irregular or minor, and some (such as
Yes) perhaps not sentences at all.
The orthographic sentence
In written language, sentence status is signalled by punctuation, primarily through the
PERIOD (especially AmE) or
full stop (especially BrE), but the orthographic sentence is not necessarily identical with the syntactic sentence:
CLAUSES separated by
SEMICOLONS or
COLONS might well be analysed as independent sentences, the punctuation reflecting the writer's feeling that the sentences to linked are closer semantically than the surrounding sentences. In addition, in the prose of publicists and advertisers, traditional conventions for the organization of written sentences are routinely abandoned, in order to highlight certain points:
Have a little pick-me-up before you get back to work. Iberia's Business Class always welcomes you with a glass of sherry. A taste of Spanish sunshine to whet your appetite for the delicious meal ahead. And afterwards relax and take advantage of our unique, multilingual, on-board library. Efficient and professional but warm and hospitable. That's how we think business should be (advertisement, 1990).
The phonological sentence
Contemporary linguists tend not to worry over the definition of a sentence. They assume that they can recognize sentences, implicitly relying on their familiarity with their orthographic forms. The spoken language, however, does not signal sentence boundaries. The syntax of speech, particularly in spontaneous conversation, differs considerably from the regularities of the written (particularly printed) language in ways that have yet to be fully investigated. Everyday conversation exhibits abundant hesitations, shifts in sentence construction, apparently incomplete structures, and interconnections that are odd by the norms of the written language. Most grammarians focus on structures in the written language (and their analogues in more or less formal speech) for the data they use in constructing their grammars, and if they turn their attention at all to samples of speech tend to derive spoken structures from what they consider to be fuller forms normal in written texts. What constitutes a sentence in the language should be (but does not appear to be) of particular concern to generative grammarians who view the goal of their grammar as accounting for all and only the sentences of the language.
Sentence structures
If the sentence is to constitute a grammatical unit for the language as a whole, then the orthographic sentence cannot serve as that unit: speech signals do not correspond to sentence punctuation, and punctuation only crudely signals some elements of speech, such as possible points for pausing or changing one's tone. Instead, reference can be made to the relative independence of the unit and its internal structure. A sentence may be viewed as a clause complex, in which the parts are clauses linked to each other by coordination and subordination. From this viewpoint, the traditional
SIMPLE SENTENCE is indeed simple, because it consists of only one clause, as in:
The governor of the prison negotiated with the prisoners throughout the day. A
COMPOUND SENTENCE involves the coordination of two or more
main clauses (each of which could constitute a simple sentence), linked by the coordinators
and,
or,
but, as in:
The governor of the prison negotiated with the prisoners throughout the day and talks were continued into the night. A
COMPLEX SENTENCE consists of one main clause within which there are one or more subordinate clauses:
The governor of the prison negotiated with the prisoners after police had seized control of the kitchen and food-store area. The subordinate clause, here introduced by the subordinator
after, can be moved to the front of the sentence, a typical property of subordinate clauses.
Further complexities are quite usual. A
COMPOUND-COMPLEX SENTENCE resembles the simple sentence in having more than one main clause, but in addition one or more of the main clauses contains one or more subordinate clauses:
A police officer said that the prison authorities could not confirm that there were bodies inside the prison,
but he believed that there had been some deaths. Here, the two main clauses are coordinated by
but, the first main clause has a
that-clause within which is embedded another
that-clause, and the second main clause also contains a
that-clause. It is further possible to recognize a
complex-compound sentence (though the term is not often used), in which the one main clause contains two or more subordinate clauses that are coordinated:
The Home Secretary said that nine prisoners had been forcibly injected with drugs and that eight others had taken drugs voluntarily. In this example, the two subordinate
that-clauses are coordinated by
and. The subordinate clause may be embedded in a
PHRASE rather than directly in another clause, as in this example of a relative clause (introduced by
who) that is embedded in a noun phrase:
Twelve prison officers who received minor injuries during the riot have all been discharged from hospital.
All the examples of subordinate clauses given above have been
finite clauses, but subordination may be effected through
non-finite clauses and
verbless clauses. In the next example, two coordinated participle clauses (which follow the comma) are subordinate to the main clause in this complex sentence:
The rioters have destroyed most of the ten wings of the prison,
systematically smashing cells and setting fire to buildings. In the example that follows, the
if-clause is verbless:
If possible,
prisoners will be moved to other prisons.
Constituents of clauses
This structured account of sentences assumes the recognition of clauses. A clause consists of central and peripheral constituents, the first usually obligatory, the second optional. The central constituents are the
subject (though generally omitted in imperative sentences) and the
verb, as in
Nobody moved. Other central constituents are
complements of the verb: that is, elements that complete its meaning, such as the direct object
my typewriter in
Somebody has taken my typewriter, the indirect object
me in
Derek gave me some books, the subject complement
hungry in
Jane is hungry, the object complement
strong in
I like my tea strong, the adverbial complement
in the garage in
The car is in the garage and in
I put the car in the garage. The peripheral or marginal constituents are mainly adverbials such as
incidentally,
also, and
last week in
Incidentally, Derek also gave me some books last week, and vocatives such as
Natalie in
I like my tea strong,
Natalie.
Irregular structures
Some constructions are irregular in some respect, but are generally considered sentences or (if attached to a sentence) parts of sentences: (1) Certain types of subordinate clauses constitute independent exclamations:
That we should come to this! To think that I once helped him! If it isn't my old friend Jeremy! If only you had listened to me! (2) Questions in which the phrases or subordinate clauses are introduced by interrogative words:
How about a kiss? Why all the fuss? What if they don't come? How come you're not ready yet? (3) Such headings as
How to get help in an emergency;
Where you should eat in Paris. (4) Elliptical constructions such as
Serves you right and
Never fails,
does it?, and elements of dialogue such as
A:
Where are you? B:
In the kitchen. (5) Problematic sequences that cannot easily be analysed into clausal constituents appear in such contexts as labels, titles, warnings, and greetings:
Baked beans;
The Department of English;
Good morning;
The police!;
Thanks;
Yes. Conversations often contain such sequences as
That one.
The big one.
No,
the one over there.
Higher up.
Yes.
Beside the green jug, which might as easily be written
That one—the big one—no—the one over there—higher up—yes—beside the green jug, orthographically sidestepping the problem of deciding the status of the phrases in question. Such uncertain sequences are often referred to as
sentence fragments.
Functional and syntactic categories
Sentences were categorized above by degree of internal complexity:
simple,
compound,
complex,
compound-complex,
complex-compound. They are also commonly classified according to dominant function in discourse, as
declarative,
interrogative,
imperative, and
exclamatory. These functions are reflected in the four corresponding sentence types
statement,
question,
directive (or
command), and
exclamation. In addition, sentences can be classified according to syntactic features that affect the sentence as a whole:
mood (indicative, imperative, subjunctive),
voice (active, passive), and
polarity (positive, negative). Finally, they can be classified by the patterns of the central or kernel constituents that they exhibit: for example,
Subject–Verb–Direct Object (
SVDO);
Subject–Verb–Subject Complement (
SVSC).
See
COMMA,
COMPARATIVE SENTENCE,
COMPLEMENT,
EXISTENTIAL SENTENCE,
OBJECT,
PARTICIPLE,
PART OF SPEECH,
PERIODIC SENTENCE,
PREDICATE,
PUNCTUATION,
SENTENCE WORD,
STRESS,
STYLE,
SUBJECT,
TAG,
VERB,
VERBLESS SENTENCE,
VOICE, WRITING.
