Clause types

When we introduced the phrase-structure rule for English sentences in the preceding section, we noted that all declarative sentences follow this rule, which implies that there are other types of sentences that do not follow it. And indeed, there are. Take a look at the following sentences:

All of these sentences use the verb submit, which has the valency pattern [ NP_Agent ___ NP_Patient ] and all of them could be uttered in the same situation, but they differ in their structure and in the intention with which they would typically be used:

• Sentences like that in (1) are called declaratives. Such sentences follow the phrase-structure rule [_S NP VP] that we introduced in Section 7.6 and would typically be used as statements.

• Sentences like that in (2) are called polar interrogatives. They always start with the auxiliary, followed by the subject NP, followed by the verb phrase. They would typically be used to ask a question about the truth or falsity of a situation (a typical answer would be yes or no).

• Sentences like those in (3) are called wh-interrogatives. Their phrase structure is more variable and more difficult to describe — they always start with an interrogative pronoun (most of which start with wh– in English, which is where the name of the sentence type comes from), followed by a verb or an auxiliary, followed by what looks like a verb phrase which, in some cases, is missing one of its complements. Such sentences would typically be used to ask a question aimed at identifying an unknown aspect of a situation (a typical answer would consist of the constituent that would correspond to the interrogative pronoun if the sentence were declarative).

• Sentences like those in (4) are called imperatives. Their phrase structure can be described as [_S VP] for now, i.e., they seem to consist of a verb phrase without a subject. They would typically be used to present a possible course of action to someone, for example, as a request or as advice.

• Finally, sentences like those in (5) are called exclamatives. Their phrase structure may resemble that of wh-interrogatives at first sight, as in the case of (5a), or it may consist of a noun phrase with an additional word in front of the determiner, as in the case of (5b, c). They are used to express surprise, often coupled with a strong positive or negative evaluation.

Sentence types are interesting in (at least) three ways. First, in terms of their functions, which we characterized very briefly in the list above and will look at more closely in Chapter 9. Second, in terms of their structure, which we will discuss next, and third, in terms of their relation to each other, which we will come back to later.

The structure of different clause types

Let us begin with the simplest case, imperatives. They have the phrase structure shown in (6), i.e., the subject is optional (and typically omitted), so that they often consist of nothing but a verb phrase; note that the verb must always be in the infinitive:

Imperatives are the only exception to the requirement that all English sentences must have a subject — the reason presumably being that imperatives are used to instruct a person to do something, so the subject would always correspond to the person that is being addressed. However, note that you can include a subject in imperatives if you want to put emphasis on it or if you are addressing more than one person and want to avoid ambiguity:

Polar interrogatives are already quite a bit more complex. Their structure seems simple at first glance:

But there are two complications. First, as (8) indicates, they always contain an auxiliary. In (2) above, this is not a problem — it uses the future construction with will, so we have an auxiliary anyway (the corresponding declarative also contains one). But if you look at the declaratives in (8), you may wonder what the corresponding polar interrogatives would be:

As you know, they look like this:

So where does the auxiliary do come from, all of a sudden? It turns out that English does not allow lexical verbs to have tense in polar interrogatives. This is not a problem in cases like that in (2) above, where the tense is expressed by the modal will, but it is a problem in sentences like the ones in (7a, b), because the tense has to be expressed somewhere. Speakers of English solve this problem by inserting the auxiliary do.

English is quite odd in this way — there are no other languages that have obligatory do-support in interrogatives. In other languages, if the declarative clause does not have an auxiliary, neither does the polar interrogative, as these examples from Danish (10) and Polish (11) show:

This fact that English polar interrogatives always contain an auxiliary is represented in (8), but the fact that that auxiliary is do unless the corresponding declarative contains some other auxiliary is not represented, and it is difficult to see how we would do so. In other words, phrase structure rules are not enough to describe the grammar of a language.

And, as we hinted at earlier, there is a second problem: we decided in Section 7.4 to include the auxiliary in the verb phrase, but in (11), it is clearly outside the verb phrase.

So, were we wrong? If the auxiliary had its own branch, it seems that we would not have a problem — we could simply reverse the order of the auxiliary and the subject, as shown in Figure 7.7.1 But we cannot have it in the verb phrase but occurring before the subject, as Figure 7.7.2 shows — we would have to give up the idea, which we have so far taken for granted, that the elements in a constituent are ordered, which would mean that we could no longer account for word order.

Figure 7.7.1: An interrogative with the auxiliary outside of the VP

Figure 7.7.2: An impossible interrogative with the auxiliary outside of the VP

There are several potential solutions to this problem, which also address the question how different clause types are structurally related.

Structural relations between clause types

First, we could limit the application of phrase-structure rules to declarative sentences and have a separate type of rule that turns these declarative sentences into imperatives, polar interrogatives, etc. This is an idea that we mentioned in Section 7.6 in connection with passives, and there are linguistic theories that have used such rules. They were called transformational rules, and they consisted of two parts: a structural description of a declarative clause, and a description of changes that would assign numbers to each element in the structure in sequence and then state how these elements would have to be reordered. For the polar interrogative, this would have looked something like this:

Since this rule can only apply to clauses that have an auxiliary verb, there was a transformational rule called “do support” that inserts the auxiliary do into clauses that do not have an auxiliary:

So, in order to describe the structure of a polar interrogative, one would describe the corresponding declarative in terms of phrase-structure rules, then apply do-support, and then apply the rule in (12).

A second solution would be to use the format we used for passives, and simply state the correspondence between declaratives and polar interrogatives. This would require an additional mechanism that would allow us to state default values for certain parts of the structure, that can be replaced by something else. For example, we would have to state that the auxiliary in the polar interrogative is do unless a different auxiliary is present in the declarative. This could look as in (14):

Finally, we could give up the idea of constituency and replace it by two separate levels of analysis: one that captures constituency in the form of an unordered set, and one that specifies the order of constituents under different circumstances. For example, the constituency rule for main clauses could be something like (15) and that for verb phrases could be something like (16) (we are using curly braces to indicate that these are sets without order:

We would then have ordering rules like those in (18), where > means “occurs earlier in the clause”:

Rule (18a) says that subjects occur earlier than verbs, this is true of declaratives and interrogatives. Rule (18b) says that auxiliaries occur earlier than verbs, again, this is true of declaratives and interrogatives. Rule (18c) says that, in declaratives, subjects occur earlier than auxiliaries, rule (18d) says that, in interrogatives, one auxiliary occurs earlier than the subject, which, in turn, occur earlier than other auxiliaries if there is more than one. This ensures that in sentences that contain more than one auxiliary, only one of them precedes the subject, as in (19b), not two, as in (19c), or even more, as in (19d):

If you look at the clauses in (19a) and (19b), you can see how the ordering rules work: rule (18a) says that Aylin must occur earlier than submitted — this is true in both sentences; rule (18b) says that will must occur earlier than submitted — again, this is true in both cases. Rule (18c) says that in a declarative, Aylin must occur before will — this is possible while also following rules (18a, b). Rule (18d) says that in an interrogative, will must occur before Aylin — again, this is possible while following rules (18a, b).

The point of showing you these three potential ways of dealing with the structure of different sentence types and their relation to each other was not to teach you how to perform such analyses, but to raise your awareness for the fact that we can come up with many different ways of describing grammatical structure with precision. Which of these we choose ultimately depends on two considerations: first, does it allow us to describe all rules in all languages? And second, does it allow us to describe these rules in a way that fits the way humans use language. Transformational rules, for example, are a very clear and powerful mechanism for capturing relationships between different types of sentences and clauses, but it does not fit what we know about language use: users of English clearly do not formulate declarative sentences in their head and then turn them into interrogatives.

If you are really interested in understanding the mechanisms illustrated above, try to apply them to wh-interrogatives, which are, no doubt, grammatically the most complex sentence type in English. It is very likely that you will not be able to arrive at a complete description, but it could be fun nevertheless!

← Previous section Next chapter →

CC-BY-NC-SA 4.0, Written by Anatol Stefanowitsch