Kamis, 21 Februari 2013
Phrases, Clauses, Sentences
In previous summaries we have slowly worked our way upwards from smaller to ever larger building blocks of human language. Initially, we discovered that there are meaning-distinguishing units of sound (phonemes), followed by meaning-bearing parts of words (morphemes). What, then, is the next larger unit of measurement in linguistic analysis? The following list gives an indication:
1. sentences contain one or several
2. clauses contain one or several
3. phrases contain one or several
4. words contain one or several
5. morphemes
While some languages blur the boundaries between words and longer expressions to some extent due to their morphology, English allows a fairly clear segmentation into phrases, clauses and sentences. The structural relations of these units with one another fall into the domain of syntax. Just like morphology, syntax is not concerned with what a sentence means, in the sense of what it tells us about the world, but with the internal structure of units and their relations to one another. In other words, syntax asks which sentences are in accord with the grammatical rules imposed by a particular language and which aren’t.
When talking about sentences as units in grammar, it is important to recognize that we idealize their status to some extent. Spoken language often consists of incomplete utterances and seemingly disjointed pieces, but this does not make it ‘less grammatical’.
Two simple sentences demonstrate aptly what sort of relations are covered by syntax:
John likes pie
*John pie likes
The words used in both sentences are identical and common expressions in English. But clearly there is a problem with John pie likes. While in Persian such a sentence structure would be acceptable, it cannot be considered grammatically well-formed in English, because it does not conform with the canonical word order of English (Subject – Verb – Object, or SVO). Clearly the words themselves would also be different in Persian, but what counts in the context of syntax is that what is grammatical in one language may well be ungrammatical in another and that this dimension is detached from meaning as we frequently understand it.
A famous example sentence helps to exemplify this last aspect of language:
Colorless green ideas sleep furiously
The sentence may not ‘make sense’ in that ideas cannot be colorless, do not have the ability to sleep and are not able to do so furiously. But grammatically the sentence is perfectly acceptable, because every word is in a place where it can potentially be, something that is not the case with *John pie likes or *Furiously sleep ideas green colorless.
Two central questions from the vantage point of syntax are therefore ‘What are the building blocks of a sentence?’ and ‘How do they interact with one another in a particular language?’
Constituency
Those elements in a sentence that form structural units are called constituents. From a functional perspective (in other words, when asking what the constituents in a sentence do), it is possible to distinguish between two basic building blocks that exist in any complete sentence: a referring expression and a predication.
referring expression predication
Mary ate an apple
I don’t like Mondays
The car crashed into the wall with a bang
It rained steadily all day long
The referring expression is essentially something (a person, thing, emotion, state or abstract concept) that we make a statement about (for example, that it went somewhere, did something, has a certain quality and so forth).
Phrases
If we turn from function to grammar we can make another observation: the referring expression is always a noun phrase (NP), while the predication is a verb phrase (VP). A complete English sentence will always contain these components.
NP VP
Mike likes pizza
Many people decided not to vote
These two examples demonstrate why we use the terms noun phrase and verb phrase, instead of just speaking of nouns and verbs: Mike and Many people fill the same syntactic slot, while likes pizza and decided not to vote fill another (they are constituents). Phrases can vary a great deal in terms of length and complexity and they can themselves contain other phrases. For example, the verb phrase likes pizza contains the noun phrase pizza.
Constituency tests
It is possible to test whether part of a sentence is a constituent via several relatively simple tests. Three of the most common tests are substitution, movement and question-forming.
When the fire broke out, the girl on the roof cried for help
In the example above, replacing the noun phrase the girl on the roof with the pronoun she yields a grammatical sentence.
When the fire broke out, she cried for help
If we replaced only the girl with she, the result would not be a grammatical sentence: *When the fire broke out, she on the roof cried for help. The fact that we can substitute a pronoun for the girl on the roof proves its status as a constituent of the sentence.
Another possibility is to move the assumed constituent to the front of the sentence:
When the fire broke out, the girl on the roof cried for help
The girl on the roof cried for help when the fire broke out
Moving only parts of the noun phrase would not result in a well-formed sentence because it would destroy its structural integrity (*The girl when the fire broke out on the roof cried for help).
Forming questions that ask specifically for the constituent is another approach:
Q: Who cried for help when the fire broke out?
A: The girl on the roof
Headedness
The element that gives a phrase its name (a noun phrase, verb phrase, prepositional phrase) is commonly called the head. To give an example, in a noun phrase the head noun may be preceded by a determiner, adjective or another noun (e.g. the crowded football field) and followed by a prepositional phrase or relative clause (the book on the table; the girl who called the police). If we replace a noun phrase with with something else, it must be another noun or a pronoun – all other words in the noun phrase are optional.
All sorts of people love pizza
Sue loves pizza
I love pizza
*The loves pizza
*Green loves pizza
*But loves pizza
While this kind of endocentric headedness generally applies to noun phrases and verb phrases, prepositional phrases frequently behave in a different way. For example, the prepositional phrase in The keys are on the table cannot be replaced with only on or the table – *The keys are on and *The keys are the table are both ungrammatical.
If we examine the constituency of sentences (in other words, their phrase structure) we find that frequently units are grammatically ‘packaged’ inside other units, producing a hierarchical structure. One way of expressing said structure is by using brackets (note that in the example S stands for sentence, not subject):
The old man’s cat slept
= (S (NP (NP The old man ’s) cat) (VP slept) )
Another, very popular method of expressing a phrase structure is the use of tree diagrams, as in the example below
One good approach when looking at the phrase structure of a sentence is to identify the main clause’s noun phrase and verb phrase and then break down the sentence into smaller units, one constituent at a time.
Verb arguments
Two terms related to grammar that you have probably already encountered in school are subject and object. While the constituents of a sentence are its components, subject and object are specific syntactic roles that define the relationship of constituents to the verb (in other words, they are the verb’s arguments). In the example above, John lost his pants, John is a noun phrase that fills the role of subject in relation to lost which is the main verb of the sentence, while his pants is the direct object. We use the term transitivity to describe what arguments a particular verb assigns.
Transitivity
Verbs assign specific argument slots to constituents according to the predication they express.
Sue yawned
S V
Sue likes cookies
S V O(d)
Sue gave John a beer
S V O(i) O(d)
In the first example, the verb yawn is intransitive – it does not permit an object (*Sue yawned John is not grammatical). By contrast, a direct object is required in the second sentence (*Sue likes does not work), making the verb like monotransitive. Finally, the third sentence has a so-called double object construction. The subject is followed by two objects, the indirect one inserted before the direct one. Because of this, we refer to give as a ditransitive verb. Note that many verbs permit multiple argument configurations: John bought a beer (montransitive) and John bought Sue a beer (ditransitive) both work.
Two additional syntactic roles, which are in turn associated with subjects and objects, are those of complement and adverbial.
Complements
Complements are associated with either subjects (subject complements) or objects (object complements). They provide more information about the thing they are associated with (they predicate the subject or object) and are often required in order for the clause they appear in to be grammatically well-formed. Several examples help to illustrate this behavior:
Subject complements:
The cookies taste delicious
S V C(s)
John is a teacher
S V C(s)
Ruth seemed tired
S V C(s)
The role of complement is closely tied to specific verbs, sometimes called copula or linking verbs, because they link the subject or object and its complement. The notable difference between a subject-complement construction and a subject-object construction is that the complement ‘completes’ the subject. A teacher specifies something about John, just as delicious is how the cookies taste and tired is the state that Ruth is in. By contrast, in Jane pushed her sister, her sister does not complement Jane but is the direct object of pushed.
Object complements function very much like subject complements:
They found the movie disappointing
S V O C(o)
I consider Susie a genius
S V O C(o)
The party made him prime minister
S V O C(o)
Just as subject complements predicate subjects, object complements predicate objects. Complements are generally not optional but required – *They found the movie, *I consider Susie and *The party made him are each incomplete without the object complement (the same thing holds true for the subject examples above). Typically complements are noun phrases, predicative adjectives, or participles that behave similar to predicative adjectives (They found the movie disappointing).
Adverbials
In contrast to complements, adverbials can generally be described as predicating either the verb or the entire clause. They are usually adverb phrases, temporal noun phrases or prepositional phrases.
He ate the cake slowly
S V O A
She drops the eggs every time
S V O A
The keys are on the table
S V A
While in the first two examples the adverbial is optional (He ate the cake and She dropped the eggs are grammatical), the last example looks a little like a subject complement at first sight. However, looking more closely reveals that on the table answers the question of where the keys are. It modifies the entire clause and not just the keys.
It is important to point out is this context that terms in syntax in general and the terms complement and adverbial in particular are used with a variety of meanings by different linguists with different theoretical backgrounds. The relatively traditional terminology (which we use) poses certain problems, especially in cases that are fuzzy.
Clauses
Looking beyond phrases, the next larger structural unit we encounter are clauses. A clause is generally defined as consisting of a referring expression and a predicate (or NP + VP), which makes it possible to use the terms clause and sentence synonymously when dealing with simple sentences.
John likes pizza
John likes pizza and Mary likes pasta
John likes pizza because it tastes awesome
The first is an example for a simple sentence, the second for a compound sentence and the third for a complex sentence.
A simple sentence contains a single independent clause. Note that the clause may be quite long and contain a number of phrases, i.e. The old miner’s fantastically rich cousin frequently traveled to South Africa many years ago is still a simple sentence.
Compound sentences contain multiple clauses that are strung together via coordinating conjunctions (and, but, or) or parataxis (connection without a conjunction).
Coordination:
John likes pizza and Mary likes pasta
Parataxis:
John likes pizza, Mary likes pasta
Finally, complex sentences combine an independent clause with one or more dependent (or subordinate) clauses. Dependent clauses are generally signaled by subordinating conjunctions such as because, since, after, while, although or when, or by relative pronouns such as who, which or that. While they always contain a referring expression and a predication, they often don’t make much sense on their own. Many textbook definitions therefore state that independent clauses represent ‘complete thoughts’, while dependent clauses do not.
After he had called Mary, John picked up the pizza that he had ordered earlier
The example sentence begins with the dependent clause After he had called Mary, followed by the independent clause John picked up the pizza which is in turn followed by the relative clause that he had ordered earlier. Only the independent clause sounds right on its own, while the dependent clauses seem incomplete by themselves.
Relative clauses may be either restrictive or non-restrictive and provide more information about the subject or object of the main clause.
Restrictive relative clause:
The man who had lost his wallet decided to call the police
Non-restrictive relative clause:
The man, who had lost his wallet, decided to call the police
The difference between the two examples in one of meaning. The first sentence implies that a specific man (the one who lost his wallet) decided to call the police. By contrast, the second sentence is about a man who called the police – the information that he has also lost his wallet is given on the side.
Clauses may be part of phrases. For example, restrictive relative clauses like the one above (who had lost hist wallet) are always part of a noun phrase. This embedding of clauses is also the reason why we can always assume the basic structure NP -> VP: the verb phrase will often contain other phrases and clauses.
Finite vs. non-finite clauses
Dependent clauses exist in two basic varieties: finite and non-finite. In a non-finite clause the verb shows no inflectional agreement with the subject, while in the finite variant it does.
When he saw the mess in the kitchen, John took a deep breath (finite dependent clause)
When seeing the mess in the kitchen, John took a deep breath (non-finite dependent clause)
Verb inflection
English has three basic varieties of non-inflected verbals: participles, gerunds and infinitives. While participles act similar to adjectives, gerunds behave like nouns:
Staring at the empty box, John took a deep breath (participle)
Arguing will not help (gerund)
To-infinitives can fill subject and object roles (i.e. behave like nouns), or modify existing subjects or objects:
To study is the smart thing to do before an exam (to-infinitive as subject)
John was asked to leave (to-infinitive as object)
He started to talk (to-infinitive as adverb)
Key terms
grammatical – ungrammatical
referring expression – predication
constituents
constituency tests
substitution
movement
question-forming
phrases
noun phrase
verb phrase
prepositional phrase
headedness
phrase structure
clause
dependent – independent
finite – non-finite
relative clause
restrictive – non-restrictive
coordination – subordination
sentences
simple – compound – complex
transitivity
verb arguments
syntactic roles
subject
direct object
indirect object
complement
adverbial
http://introling.ynada.com/session-9-phrases-clauses-sentences
Langganan:
Posting Komentar (Atom)
Tidak ada komentar:
Posting Komentar