Clark's colleagues likes him a lot.

In many languages verbs take inflectional morphemes that convey some information about one or more participants in the event or state that the sentence is about. One way to think about this is in terms of the agreement between the verb and those participants on a small number of abstract properties. On the one extreme are languages like Mandarin Chinese and Japanese, which have no morphology of this type (though sometimes the choice of a verb in Japanese is governed by some properties of the participants). In what follows, I'll briefly discuss verb agreement in four languages that have some form of it. Notice that since agreement morphology conveys abstract properties of participants, that is, things, this topic overlaps with the topic of the last section.

English is a language with limited verb agreement morphology, the vestiges of what was a full-blown agreement system in Old English. Consider these sentences.

Clark plays golf.

Lois and Clark play tennis.

I play croquet.

Clark played 18 holes yesterday.

Clark likes team sports.

In English – s is plural when it appears on nouns but singular when it appears on verbs.

Notice that the form of the verb play differs in sentence 9 and 10. In sentence 9 the subject of the sentence, Clark, is 3rd person (that is, including neither the Speaker nor the Hearer) and singular, and the verb takes the suffix – s to indicate this. When the same verb is used with a subject that has any other combination of person and number, as in sentences 10 and 11, the verb takes no suffix. Notice also that an agreement suffix is only added to verbs in the simple present tense, that is, the tense category used in sentences 9, 10, and 11. Sentence 12 is in the simple past tense, and no distinction is made on the basis of person and number. Finally, notice that it is the participant in the syntactic role of subject, rather than any particular semantic role, that the verb agrees with. So in sentence 13, the verb again takes the – s even though the subject in this case refers to an experience rather than an agent, as in sentence 9.³

With the verb be, there are three forms rather than two in the simple present, and rather than suffixes, completely unrelated forms are used: am (1st person singular), is (3rd person singular), and are (other person-number combinations). The verb be also has two forms in the simple past tense, was and were.

Thus English subject-verb agreement is limited both in terms of the number of different forms and the situations in which it must apply. However, it behaves just like the other examples of grammatical morphology we've been considering. It is often redundant, but it is obligatory even when it is. So in standard English dialects, at least, it is ungrammatical to say Clark like Lois, even though the missing – s would convey no new information.

So does the – s in play in sentences 9 and 13 mean anything? Yes, it means that the subject of that verb is 3rd person singular. In addition, because this suffix only occurs on verbs in the simple present tense, it also marks that tense category. Under most circumstances, this information would be obvious from the subject itself and from the context. But if the Hearer missed the subject for some reason, that – s could help sort things out. Also there are gray areas where Speakers may choose to use a verb in the 3rd person singular with a plural subject. Compare these two sentences.

A hundred students are in this course.

A hundred students is more than this room can hold.

In sentence 15, the subject is viewed as an individual quantity rather than a collection of individual things, so the verb is singular.

2.2 American Sign Language

The grammars of sign languages may be just as complex as those of spoken languages.

Finally let's consider agreement morphology on verbs in a sign language. We have already seen one example of this in the discussion of mutation morphology. ASL has a category of verbs that sign linguists call «directional verbs». These are verbs designating transfer events, or information transfer events, or other events viewed as having a direction. These verbs have a basic handshake and a position on the body, but their direction has to agree with the source and the goal (often the recipient) of the event. The agreement is with what corresponds to person in ASL, the position in signing space of the participants. 1st and 2nd person have the position of the signer and the sign interpreter, and other participants are «placed» in signing space by the signer as they come up.

For example, to produce the sign for 'give' in ASL when the source/agent is neither the signer nor the sign interpreter and the recipient is the signer, the signer uses the basic handshake for 'give', moving one hand from the position of the giver in signing space to the signer's own chest. The direction would be the opposite if the roles were reversed.

Another form of agreement in ASL makes use of classifiers. Classifiers in ASL take the form of particular handshakes that represent general properties of things. For example, an index finger pointing upward represents a standing person, a cupped hand represents a container, and the extended thumb and first two finger represents a vehicle One use of classifiers is as morphemes agreeing with the subjects of verbs designating move events and be at states. In this case the agreement is the opposite of what happens with verbs of giving and telling. It is the handshake that represents the agreement morpheme and the movement of the hand(s) that represents the content of the verb. For example, to sign a sentence meaning 'the car is here', the signer would make the sign for 'car', then with the 'vehicle' classifier handshake sign 'be here', that is, move the hand downward in front of the body.

How is verb agreement in ASL like the verb agreement in the spoken languages we have considered? At least in many cases agreement in ASL is obligatory, as it is in spoken languages. It may also be redundant, as in the 'vehicle' example.

Agreement in ASL, in fact morphology in sign languages generally, is strikingly different from spoken language morphology in one way. It is invariably iconic; all of these examples we have seen «make sense». With respect to form alone, sign language grammatical morphology differs in another way from most spoken language grammatical morphology in that it occurs simultaneously with the root morpheme. Of course this derives from the potential in sign languages to maintain a particular handshake while a movement is executed.

One point of this section has been to show how much languages can vary in terms of what information gets represented on their verbs. It is on verbs that we see how different languages can get. Within our set of languages, we have seen a range of possibilities, but we still are not close to the extreme of some American Indian and Eskimo languages, like Inuktitut, where verbs frequently include more than ten morphemes. However, those words usually include morphemes that go beyond the functions we've discussed in this chapter. Such languages excel at creating new words from a small number of roots and extensive productive morphology. How this sort of process works is the topic of the next chapter.

2.3 The category of voice

In English as in many other languages, the passive voice is the form of a transitive verb whose grammatical subject serves as the patient, receiving the action of the verb. The passive voice is typically contrasted with the active voice, which is the form of a transitive verb whose subject serves as the agent, performing the action of the verb. The subject of a verb in the passive voice corresponds to the object of the same verb in the active voice. English's passive voice is periphrastic; that is, it does not have a one-word form. Rather, it is formed using a form of the auxiliary verb be together with a verb's past participle.

Canonical passives

Passive constructions have a range of meanings and uses. The canonical use to map a clause with a direct object to a corresponding clause where the direct object has become the subject. For example:

John threw the ball.⁴

Here, threw is a transitive verb with John as its subject and the ball as its direct object. If we recast the verb in the passive voice (was thrown), then the ball becomes the subject (it is promoted to the subject position) and John disappears:

The ball was thrown.

The original subject can typically be re-inserted using the preposition by:

The ball was thrown by John.

Promotion of other objects

One non-canonical use of English's passive is to promote an object other than a direct object. It is usually possible in English to promote indirect objects as well. For example:

John gave Mary a book. → Mary was given a book.

In the active form, gave is the verb; John is its subject, Mary its indirect object, and a book its direct object; in the passive form, the indirect object has been promoted and the direct object has been left in place. (In this respect, English resembles dechticaetiative languages.)

It is also possible, in some cases, to promote the object of a preposition:

They talked about the problem. → The problem was talked about.

In the passive form here, the preposition is «stranded»; that is, it is not followed by an object. (See Preposition stranding.) Indeed, in some sense it doesn't have an object, since «the problem» is actually the subject of the sentence.

Promotion of content clauses

It is possible to promote a content clause that serves as a direct object. In this case, however, it typically does not change its position in the sentence, and an expletive it takes the normal subject position:

They say that he left. → It is said that he left.

Stative passives

The passives described so far have all been eventive (or dynamic) passives. There exist also stative (or static, or resultative) passives; rather than describing an action, they describe the result of an action. English does not usually distinguish between the two. For example:

The door was locked.

This sentence has two meanings, roughly the following:

[Someone] locked the door.

The door was in the locked state. (Presumably, someone had locked it.)

The former meaning represents the canonical, eventive passive; the latter, the stative passive. (The terms eventive and stative/resultative refer to the tendencies of these forms to describe events and resultant states, respectively. The terms can be misleading, however, as the canonical passive of a stative verb is not a stative passive, even though it describes a state.)

Some verbs do not form stative passives. In some cases, this is because distinct adjectives exist for this purpose, such as with the verb open:

The door was opened. → [Someone] opened the door.

The door was open. → The door was in the open state.

Adjectival passives

Adjectival passives are not true passives; they occur when a participial adjective (an adjective derived from a participle) is used predicatively For example:

She was relieved to find her car undamaged.⁵

Here, relieved is an ordinary adjective, though it derives from the past participle of relieve In some cases, the line between an adjectival passive and a stative passive may be unclear.

Passives without active counterparts

In a few cases, passive constructions retain all the sense of the passive voice, but do not have immediate active counterparts. For example:

He was rumored to be a war veteran. ← *[Someone] rumored him to be a war veteran.

(The asterisk here denotes an ungrammatical construction.) Similarly:

It was rumored that he was a war veteran. ← *[Someone] rumored that he was a war veteran.

In both of these examples, the active counterpart was once possible, but has fallen out of use.

Double passives

It is possible for a verb in the passive voice – especially an object-raising verb – to take an infinitive complement that is also in the passive voice:

The project is expected to be completed in the next year.

Commonly, either or both verbs may be moved into the active voice:

[Someone] expects the project to be completed in the next year.

[Someone] is expected to complete the project in the next year.

[Someone] expects [someone] to complete the project in the next year.

In some cases, a similar construction may occur with a verb that is not object-raising in the active voice:

The project will be attempted to be completed in the next year. ← *[Someone] will attempt the project to be completed in the next year. ← [Someone] will attempt to complete the project in the next year.

(The question mark here denotes a questionably-grammatical construction.) In this example, the object of the infinitive has been promoted to the subject of the main verb, and both the infinitive and the main verb have been moved to the passive voice. The American Heritage Book of English Usage declares this unacceptable but it is nonetheless attested in a variety of contexts

Other passive constructions

Past participle alone

A past participle alone usually carries passive force; the form of be can therefore be omitted in certain circumstances, such as newspaper headlines and reduced relative clauses: