In Section 7.2, we set out to analyse the way in which words are combined into sentences. In order to do so, we introduced the idea of constituents and discussed different types of constituents and their structure in some detail. We also used the word sentence and the word subject quite a bit, without really discussing them in their own right. Let us remedy this now!
The meaning of the word sentence is rather vague in everyday language, which is one of the reasons why linguists often try to avoid it, using the term clause instead. There are two types of clauses: independent clauses (also called main clauses), that can stand on their own, and dependent clauses (also called subordinate clauses), that needs to be attached to an independent clause. In this section, we will focus on independent clauses and return to dependent clauses in Chapter 10.
An (independent) clause in a phrase that contains a verb and all of its complements, including the subject. Since all of the complements except for the subject are included in the verb phrase, the phrase-structure rule for English sentences is very simple:
Note that we use the letter S to label main clauses — this stands for “sentence”, of course — the word sentence is used in linguistics as a synonym for main clause.
All independent clauses in English have the structure in (1), as long as they are declarative clauses — roughly speaking, clauses that express a statement (we will see other types of clauses in Section 7.7, that require additional rules). If you don’t believe us, you can check this by looking at all examples in this chapter.
In English, the phrase structure in (1) allows us to identify two types of complement that are special in a number of ways and that have their own name: the subject (a word we have already used) and the object. Let us first see how to identify them and then talk about why they are special. The subject of a main clause is the NP that is attached directly to the S node, the object is the NP that is attached directly to a VP node which does not contain another VP. This sounds much more complicated than it is — if we represent the phrase structure rule in (1) as a tree structure with a VP containing an NP, as shown in Figure 7.6.1, we see that it is actually very simple.
In the following sentences, the subjects are shown in red, the objects are shown in blue — compare them to the tree structure in Figure 7.6.1 to see that they follow the definitions we have just given:
As you can see, all sentences in English have subjects, but not all of them have objects. If a verb phrase contains two NPs that meet the definition for objects, as in (2d), then both of them are objects. So why aren’t the NPs in the VPs in (2e) and (2f) objects? For (2e), the answer is simple: the NP is part of a PP, so it is not attached to the VP node and does not meet the definition. For (2f), the answer is more complicated — cases like this are the reason why the definition of an object contains the caveat that an object is an NP attached to a VP that does not contain another VP. The tree structure of (2f) is shown in Figure 7.6.2
A replacement test shows that slept forms a VP by itself, which is part of a larger VP slept all day. In (3a), did the same replaces slept, in (3b) it replaces slept all night:
At this point, you might be wondering why we are making life so difficult for all of us — would it not be easier to define subject and object in the way many of you probably remember from your language classes, i.e., something like “The subject is the doer of the action, the object is acted upon”? Well, the examples in (2) show that this definition is wrong: In (2a) and (2f), the subject is not a “doer” but a Theme — stumbling or sleeping are not “actions” that anyone “does” — stumbling is a process that happens to someone, and sleeping is a state in which someone is. In (2c) and (2d), the subject is not a “doer” but an Experiencer — liking and believing are not “actions” that anyone “does”, they are experiences that someone has. Likewise, in (2c), the bagels are not “acted” upon — they are simply the Stimulus that triggers Zoe’s positive feeling towards them. And in (2d), the glass of water is indeed acted upon by Noah, but Aylin is not — she is a Recipient.
So, why are subject and object special enough to deserve their own names, while all other complements are lumped into a category that is sometimes called “adverbials” (a very bad name, because it suggests that they are related to adverbs), “obliques” (nice if you have a classical education that includes Latin, but otherwise a bit, well, oblique) or that remains nameless?
There are many reasons, but there are two that stand out: subjects and objects play a special role in structuring the information in sentences (we will return to this issue briefly at the end of this section and in Chapter 10), and subjects and objects play a special role in grammar.
Both sentences use the verb bake, which has the valency pattern \(\text{[NP}{\tiny{\text{Agent}}} ~ \text{\_\_ NP}{\tiny{\text{Patient}}}\text{]}\text{,}\) and both sentences mean roughly the same thing. But in (4a), Aylin is the subject and these bagels is the object, while in (4b), these bagels is the subject and Aylin occurs as an NP within a PP.
What is going on here? In English, as in many other languages, verbs with the valency pattern [ NP __ NP ] can be used in two different ways: in what is called the active voice (shown in 4a) and in what is called the passive voice (shown in (4b). These two kinds of sentences are related in a systematic way: The NP that is the object in the active sentence is always the subject of the passive sentence, and the NP that is the subject in the active sentence can (but does not have to) be an NP in a PP headed by the preposition by. An additional change concerns the verb: in the passive sentence, the auxiliary be occurs in the same tense as the lexical verb in the active sentence, while the lexical verb occurs as a past participle.
This systematic relationship requires us to distinguish subject and object, so that we can state the relation between the object of the active sentence and the subject of the passive sentence, and it requires us to distinguish between objects and other NPs that follow a verb, so that we can state which active sentences have a passive counterpart.
If a sentence has two objects, for example, the relationship between active-voice objects and passive-voice subjects holds for both of them — either one of the objects can be related to the subject of a passive sentence:
How can we express this relationship? In language-teaching (especially in foreign-language teaching) it is often expressed as a derivational relationship — language learners will be asked to “turn active sentences into passive sentences”. And in older theories of language, such a derivational relationship was also assumed — there was a particular type of grammatical rule called a “transformation” that described, in a formal way, the changes necessary to turn an active sentence into a passive one.
This derivational approach is no longer very widespread, for a range of reasons. For example, there are many cases like (7) where a passive sentence sounds much more natural than the corresponding active one, so that it is weird to derive the passive from the active:
Instead, the relationship is often shown in a way similar to the one we introduced for lexical alternations in the preceding section. It could look something like this:
The structure of the passive sentence looks quite complex, so if you want to understand it in detail, apply constituent tests to the three verb phrases in the tree diagram in Figure 7.6.4 — you will see that we need all three of them!
If you don’t want to worry about the details right now, just ignore the complexity and focus on the basic point: The rule above simply states the correspondences between active and passive sentences without taking one of them to be the more basic form. Note that both active and passive sentences conform to the phrase structure rule that we gave at the beginning of this section.
So, why do passives even exist? What is the point of having two ways of saying the same thing? The reason can be found in a special property of the subject. In declarative sentences, the subject of a sentence refers to the entity that we are talking about (the topic) and the verb phrase contains the information that we want to communicate about that topic. With transitive verbs, there are two entities that could be the topic of a sentence. Take the verb bake again, which has an Agent (for example, Aylin) and a Patient (for example, bagels). In some situations, we might want to talk about Aylin and what she bakes, in other situations, we might want to talk about bagels and who baked them. Having the active and passive voice at our disposal allows us to put either Aylin or the bagels in the subject position:
By the way, note that passive sentences do not require an Agent and a Patient — (almost) any valency pattern with a subject and an object can be used in the passive, as the following examples show:
Before we conclude, a brief word about subjects and objects and word order. English is a so-called configurational language, which means that it has a relatively fixed word order, so language users can rely on the position of constituents in the sentence to identify subjects and objects. Because of the phrase structure rule in (1), subjects always come directly before the verb, and the object normally directly follows it (there are some circumstances where it does not, that we will talk about in the next section).
Other languages have a more flexible word order (they are sometimes called non-configurational). For example, all Germanic languages other than English have a phrase structure rule for main clauses that could be represented very roughly as shown in (11) (ignoring some differences in detail between the different Germanic languages):
What this rule says is that, in these languages, main clauses consist of one phrase of any kind, followed by the verb, followed by more phrases. Speakers are completely free in which phrase they put into the position before the verb, as the following examples from German show — most often, it is the subject (as in 12a), but it can also be the object (as in 12b), a prepositional complement (as in 12c) or even an adjunct (as in 12d):
So, how do speakers of German identify the subject in such sentences? They use a mix of strategies, but one of these is case marking: subjects and objects have different cases, which you can see by looking at the article: der is nominative (for masculine nouns like Hund and Nachbar), which indicates the subject, and den is accusative, which indicates the object.
English also has case, but only for pronouns, where, for example, she is for subjects and her is for objects (as in 13a); for proper names (as in 13b) or regular noun phrases (as in 13c), there is no difference between subjects and objects except for the word order.
