Grammatical Categories and Inflection
Removing the grammatical morphemes in bold in the following sentences makes them ungrammatical, but how does it affect their interpretability? That is, would any information be lost?
- Jimmy wrote the letters.
- Jimmy wrote two letters.
- Clark unbuttoned his shirt.
- Lois reread the chapter.
In the last section, we saw that morphemes can be divided into those with relatively specific meanings and belonging to large, open-ended classes — lexical morphemes — and those with very abstract meanings and belonging to small, closed classes — grammatical morphemes. In this section and the next, we'll look more closely at some of the meanings and functions that grammatical morphemes have. Grammatical morphemes are always associated with a particular lexical morpheme. They may be combined with the lexical morpheme to form a single word, as in applesor walked, or they may form a separate word that belongs to the same phrase as the lexical morpheme, as in the apple or is walking.
Grammatical morphemes have two basic kinds of functions distinguished from one another in terms of how the morphemes relate to the lexical morpheme that they combine with. One function, the subject of the next chapter, is the creation of a new concept based on the meaning of the lexical morpheme. For example, in shorten, the -en takes the meaning of the adjective short and turns it into a change of state along the dimension of length. In the process -en makes a verb out of the adjective. This function of grammatical morphemes is called derivation.
The other function of grammatical morphemes, the subject of the rest of this chapter, is similar to modification; the grammatical morpheme specifies some very abstract feature of the category that is the meaning of the lexical morpheme. In other words, its meaning is a very abstract grammatical category. For example, in walked, the -ed specifies that the walking took place before the time of speaking; it assigns the feature past to the event. In other words, past, contrasting with present and future, is a grammatical category in English. The combination of a grammatical morpheme with a lexical morpheme to form a word, as in walked, is called inflection. As we'll see, though, grammatical categories can also be defined by grammatical morphemes that are separate words.
Languages differ quite strikingly in terms of which grammatical categories are built into their morphology. In this section I'll describe a few of the kinds of grammatical categories that play a role in noun phrases. In the next section, I'll describe some grammatical categories that are marked on verbs.
People not only have the capacity to recognize individual objects in their environment and categorize them as apples, stones, people, etc. They have the ability to recognize sets of objects that share a category, for example, sets of apples, stones, or people. Though an individual and a set seem to be very different things, the categorization process for the individual and for the elements of the set must be similar. This is reflected in an apparently universal property of human language: the same morpheme is used for individual objects belonging to a category and for sets of objects whose members belong to that category. In English, the morpheme apple is applied both to individual apples and to sets of apples.
People have a further ability; they can assign a cardinality to a set, that is, they can tell (or estimate) how many elements are in the set. And apparently all languages have systems of numerals such as two and eight. Each numeral is a label for a category of set, independent of what kinds of members the set has. For example, eight labels the category of sets consisting of eight elements.
Now let's imagine two tribes of Grammies. One uses common nouns like apple and tiger and numerals like two and eight, as well as adjectives like many, to talk about individuals and sets and finds that these forms suffice. They say things like give me apple whether they want one or several, and when it matters, they say things like give me two apple or give me several apple.
In another tribe, for one reason or another, a subgroup of members begins to explicitly mention whenever they are talking about a set rather than an individual. So they say things like give me apple, some whenever they want more than one and give me apple when they want exactly one. But they leave out the some when there is a numeral because the numeral makes it clear that more than one is intended. This practice catches on, and eventually two things happen. First, because the somedoesn't convey very much information, it gets pronounced more and more quickly and carelessly, and eventually all that's left of it is the s at the beginning. This s is pronounced as if it were part of the noun that it follows, and it even assimilates to the voicing of the last phone in the noun, so it is pronounced /z/ in apples. Second, the members of the tribe find it weird to say apple whenever they mean more than one, even when the context makes it clear that they do. So now they say things like give me two apples.
Even though this story is completely fictitious, it illustrates what has apparently happened in two kinds of modern languages. English is a language of the second type. It is ungrammatical in English to say apple when more than one apple is referred to. It is of course equally ungrammatical to say apples when only one apple is referred to. English grammar makes a two-way distinction in the way objects are referred to: individual objects and sets of objects are referred to differently. That is, English has the grammatical dimension number with two values or grammatical categories, singular and plural. English nouns are inflected for number, and number inflection is obligatory. Thus three apple and lots of person are ungrammatical in English.
The Grammar of a Language May "Force" its Speakers to Use Certain Morphemes in Certain Contexts, Even When They Seem to Contribute Nothing to the Meaning
Notice that in the case of the phrase three apples, the plural morpheme, -s, doesn't really carry any information; the numeral already makes it clear that more than apples is being referred to. That is, the grammatical morpheme is redundant. Redundancy is a frequent property of grammatical morphemes. Because they are obligatory, Speakers in a sense do not ask themselves whether they are necessary when producing sentences; they insert them in any case. It may seem odd that language would allow redundancy, but it is probably helpful to Hearers. Redundancy permits Hearers to understand the message even when they miss some part of it.
Japanese is a language of the other kind. Japanese does have a morpheme for plural, -tachi, but it can only be suffixed to nouns referring to people or animals, and it seems never to be obligatory. (The same morpheme is used for plural personal pronouns, meaning 'we' and 'you plural', and here it is obligatory.) Apparently number is not a dimension in Japanese grammar, except for the limited case of personal pronouns. Here are some examples that should make this clearer.
|'I want some/an apple(s).'|
|'I want two apples.'|
Singular and Plural are Part of the Grammar of English; They are Not Part of the Grammar of Japanese
In the first sentence, there is no indication of whether one apple or a group of apples is desired. It's perfectly grammatical in Japanese to leave this unspecified. The Hearer might be expected to figure out which is intended if it matters at all. In the second sentence it is clear that more than one apple is desired, but note that, unlike in English, there is no morpheme that explicitly indicates plurality. (The second sentence also contains a morpheme which I've indicated with class; more on this later.)
What does this mean about English and Japanese speakers? It certainly does not mean that Japanese speakers are incapable of understanding the difference between individuals and sets. Speakers of all languages not only understand this distinction but have ways of expressing it in their languages. What it means is that English grammar forces its speakers to make the distinction in places where other languages do not and in many cases cannot. In fact, in Japanese it's not only acceptable to leave the plural unmarked; it is impossible to mark plural on the noun for 'apple'. That noun, ringo, has no plural form.
Number Appears in English Grammar in Multiple Places
When a dimension such as number is part of the grammar of a language, it often turns up in more than one place. This is true for number in English. Consider the following sentences.
- An apple is on the table.
- Some apples are on the table.
Apple and apples are preceded by the words an and some. These words are called indefinite articles; both function roughly to say that the thing referred to is not already known to the hearer. But they differ in another way: an (or a) is used only before singular nouns, while some is used before plural nouns (and also before some singular nouns; more about this below). That is, these words also distinguish singular from plural. The verbs in the two sentences are also different. Is is appropriate only when the subject is singular, whereas are is used when the subject is plural. Again, the distinction between singular and plural matters somewhere in the grammar of the language. In Japanese, on the other hand, there is no distinction like that between an and some or between is and are.
Another thing to notice about sentences 3 and 4 is the degree of redundancy. Sentence 4 indicates in three places that multiple apples are being referred to, in the use of some rather than an, in the presence of the -s, and in the form of the verb are.
In English we can say lots of milk, lots of sand, and lots of salt, but not normally lots of milks, lots of sands, and lots of salts. On the other hand, we can say lots of girls, lots of trees, and lots of rivers, but not normally lots of girl, lots of tree, and lots of river. What do you think is the difference between these two kinds of nouns?
There is another grammatical dimension with two values in English that is tied up with the use of plural and the distinction between a(n) and some. Consider these sentences.
- Some rice is on the table.
- Two piles of rice are on the table.
Notice that in sentence 5, rice is singular, and the verb is also in the singular form is. However, instead of a, the noun is preceded by some, the form used with a plural noun in sentence 4. In fact no matter how much there is on the table, we still won't say some rices. If the Speaker wants to mention the amount of rice, they have to use another noun such as pile or bowl or cup, putting that noun in the plural, as in sentence 6.
In the English Lexicon, Peas are like Beans and Potatoes; Rice is like Sugar and Vinegar
Apparently English has two kinds of nouns. One kind, count nouns, is used mainly for objects (and for abstract things that are construed as object-like). In the singular these nouns may be preceded by the article a(n), and they are always pluralized when more than one of the objects is referred to. The other kind, mass nouns, is used mainly for masses (and for abstract things that are construed as mass-like). These nouns are always singular except in the special sense of 'multiple kinds' (for example, wines referring to different brands or varieties of wine), and they may be preceded by the article some. Of course there is a gray area between clear cases of objects and clear cases of masses, and in this area, a noun can go either way. Thus rice, as we have seen, is a mass noun. But pea, which designates something that, like rice, consists of small objects usually gathered together in a group, is a count noun. (In fact pea, in the form pease, used to be a mass noun like rice.)
So English has the dimension of countability built into its grammar. But note that it appears in the language in two places, in the grammatical forms that go with one or the other category (a(n) with singular, some with plural for count; some with singular and no plural for mass) and in the lexicon, where most nouns belong to one or the other type. That is, there is a strong tendency in English for the count grammatical patterns to go with certain nouns (such as apple and house) and the mass grammatical patterns to go with other nouns (such as rice and milk).
Unlike English Nouns, Spanish Nouns Don't Have Inherent Countability
Spanish differs from English in an interesting way. Like English, Spanish has both number and countability in its grammar. The count and mass grammatical patterns are similar to those in English, except that some usually corresponds to no article at all in Spanish. But Spanish differs in that its nouns do not strictly belong to one category or another. Apparently most nouns can be used with either mass or count morphology depending on what meaning is intended. Here are examples with the nouns madera 'wood' and papel 'paper'.
- madera '(some) wood'
- una madera 'a board'
- papel '(some) paper'
- un papel 'a sheet of paper'
Actually the difference between English and Spanish is a matter of degree. English also has some nouns that can appear with either mass or count morphology. For example, when the noun chicken designates the bird, it is treated as a count noun (a chicken, some chickens), but when it designates the meat of the bird, it is treated as a mass noun (some chicken). But Spanish goes a lot further in this respect; many more nouns can be treated in both ways.
The main point of this example is that a dimension like countability can be both a way in which things in the world are divided up conceptually by speakers and a way in which the words in the lexicon are divided up. Countability in English is both conceptual and lexical, whereas in Spanish the conceptual aspect predominates.
Spanish grammar makes another division that is even more lexically oriented than countability is in English. All Spanish nouns have a gender, belonging either to the masculine or the feminine grammatical category. A noun's gender affects several aspects of Spanish morphology, including the form of adjectives that modify the noun and the pronouns that replace a noun. Though gender seems to have had its beginnings as a categorization of the things in the world on some conceptual basis, and it is still true that nouns for female animals and people tend to be feminine while those for male animals and people tend to be masculine, for the most part the membership of nouns in one category or the other seems fairly arbitrary now. That is, Spanish gender is an example of a grammatical dimension which is much more lexical than it is conceptual.
We have seen three ways in which languages may divide the things that speakers talk about into two very general categories, on the basis of whether they are individuals or sets, on the basis of whether they are masses or objects, and on the basis of a single conceptual property (biological gender) that is extended more or less arbitrarily to cover all labeled categories of things. Another possibility, found in many languages, is a somewhat finer-grained grouping into a larger set of categories, each of which is still more general than the kind of category represented by a noun such as apple, baby, or paper. Each of these abstract categories is represented by a grammatical morpheme called a classifier. The most common basis for the classification of things appears to be shape, but it may also be based on orientation, animacy, function, or cardinality (for sets).
When Objects are Counted in Japanese, Their Shape Must Often Be Taken into Account
In many languages with classifiers, including Mandarin Chinese, Japanese, and Tzeltal, classifiers are used along with numerals when objects are counted. Many classifiers in these languages are shape-based. Here are some Japanese examples. The classifiers are indicated by the label class in the second line of each example.
|'two sheets of paper'|
|'two (roundish) eggplants'|
|'two (long, narrow) eggplants'||'two (long, narrow) eggplants'||'two (long, narrow) eggplants'|
The three classifiers illustrated here, -ko, -hon, and -mai, represent categories based on the gross shape of the object referred to. -ko is used when the object has roughly the same extent in all three spatial dimensions, for example, when it has the shape of a sphere or a cube. -hon is used for long, narrow objects like pencils, legs, and cucumbers. -mai is used for flat objects like sheets of paper, boards, and pancakes. These morphemes are obligatory in the phrases above; leaving them out, or using one that doesn't agree with the shape of the referent, would result in an ungrammatical phrase.
As with other grammatical morphemes, classifiers are often redundant. This is true in sentences 11, 12, and 13, where the noun itself makes the gross shape clear. However, there are cases where more than one shape is consistent with a particular noun, for example, the noun eggplant. Eggplants can be roughly pear-shaped (the usual European variety) or long and narrow (the usual Asian variety). With a noun like this, classifiers can be informative, as they are in sentences 14 and 15.
Note that English has words which behave something like classifiers in Japanese. When we count masses in English, we must use words like cups or pieces after the numerals. In fact one way to think about the way languages like Japanese, Chinese, and Tzeltal work is to see them as treating all nouns as mass nouns. You can't count apples directly using the Japanese noun ringo because it really means something more like 'apple stuff' than 'apple individual'.
ASL also uses classifiers. They take the form of handshapes that signify general properties of things, much like those represented by classifiers in Japanese, Chinese, and Tzeltal. But rather than being combined with numerals in NPs, ASL classifiers are combined with verbs, so they are discussed in the next section along with verb inflection.