The part of speech of a word, also called its syntactic or lexical category, is a classification of its behaviour. Some examples of parts of speech include nouns, verbs, adjectives, adverbs, prepositions, and so on. You likely learned a little bit about these categories in school, with the definitions described in Table 1. These traditional definitions are also sometimes called semantic definitions, since they classify part of speech based on the meaning of words. But, as you will see shortly, the traditional semantic definitions have some problems. In linguistics, we don’t use the traditional definitions to determine part of speech. Instead, we use distributional criteria, as will be discussed below.

Problems with the traditional approach

The same word can belong to multiple different categories

One big problem with the traditional definitions of Table 1 is that the same word can belong to multiple different categories. For example, in (1a), the word work is a verb; in (1b), it is a noun; and in (1c), it is an adjective.

All three instances of the word work in (1) have related meaning—they are all associated with some kind of labour. However, they have very different functions and behaviours in each sentence. Therefore, when we are determining the part of speech of a word, it is these functions and behaviours that we depend on to classify words.

Abstract nouns and stative verbs

Not all words are easy to classify according to the traditional semantic definitions, either. What part of speech is idea? You might know it is a noun, but how do you know that? It’s definitely not a person or a place. It is debatable whether it’s a thing. Or what about the noun destruction? That’s an action, so we might think it’s a verb! The destruction of the city by Godzilla describes an event—it’s clearly not a person, a place, or a thing. However, the behaviour and function of destruction is that of a noun.

Likewise, not all verbs are actions, and some verbs stretch the notion of state. The sentenceTasha owns a car contains the verbs owns.  Is owning an action or a state? Only under a very loose definition. What about the verb seems in Abigail seems tired. Is seem an action or a state?

Words can change category

Another piece of evidence that part of speech should not be determined by semantic definitions is that a word can change syntactic category over time without changing its meaning. One example of this is the word fun. For most speakers, fun is a noun. Compare the sentences in (2) and (3) and notice how fun has the same behaviour as other nouns, such as water. It does not share the same behaviour as adjectives, such as hard.

For younger speakers and in informal speech, though, fun can also behave like an adjective. Like other adjectives, fun can take comparative and superlative inflection for these speakers, forming the words funner and funnest. For these younger speakers, fun can behave the same way as adjectives like hard, as shown in (4). Nouns, on the other hand, cannot appear in this kind of structure, as shown for water in (4b).

Different categories in different languages

Another piece of evidence that we cannot use meaning to define syntactic category comes when we compare languages. For instance, if you’ve studied a Romance language, you’ve learned that certain concepts which are expressed in English using adjectives are expressed in Romance using verbs and nouns. For example, if you want to express hunger in English, you would say I am hungry, using the adjective hungry. In contrast, in many Romance languages and in Swahili, you would say the equivalent of I have hunger, using a noun, as shown in (5).

If we (as English speakers) were to use meaning to define word category, we might be tempted to call words like hambre, faim, and njaa adjectives meaning ‘hungry.’ But this would be wrong, because they actually behave as nouns, as the literal translations in (5) suggest. Meaning doesn’t help us to define categories within a language, and it doesn’t help us define across categories across languages.

People can identify the category of made-up words

Perhaps the most fascinating piece of evidence that we cannot use semantic criteria to classify part of speech comes from the fact that people can reliably identify the category of made-up words, even when they don’t know their meanings.

Consider below an excerpt from the poem Jabberwocky from Through the Looking-Glass by Lewis Carroll. These lines contain five made-up words, which are bolded below.

(6) And as in uffish thought he stood,
The Jabberwock, with eyes of flame,
Came whiffling through the tulgey wood,
And burbled as it came!

Even though we don’t know what the words mean, we can tell that uffish and tulgey are likely adjectives, a Jabberwock is likely a noun, and whiffling and burbled are likely nouns. We cannot be using semantic criteria, since we don’t even know what these words mean, so we must be using some other kind of evidence to determine the part of speech of these words.

Distributional criteria

If we cannot use semantic criteria, then what do we use? We use three criteria for determining part of speech:

• Syntactic distribution, which is the words that it can appear adjacent to.
• Inflectional morphology, which is how the word can be inflected.
• Derivational morphology, which are the derivational morphemes that it contains.

Let’s look again at the stanza from the Jabberwock to see these criteria in action.

(7) And as in uffish thought he stood,
The Jabberwock, with eyes of flame,
Came whiffling through the tulgey wood,
And burbled as it came!

We can tell that Jabberwock is a noun because it appears after the word the, which is an example of syntactic distribution. For an example of inflectional morphology, consider the word burbled. It takes the past tense marker -ed, which is a property of verbs. Finally, uffish uses the suffix -ish, which is a derivational morpheme that appears on adjectives.

At this point you may be wondering if this is circular reasoning. We have decided to call things nouns because they behave like nouns. And we have decided how nouns behave by looking at nouns. The observation here is that there are groups of words that have similar behaviour. The names of the categories are essentially arbitrary: we could call them blickershnitzels, and the observation would still be true and useful. Table, dog, teacher, brick, honesty, and so on, all behave in a similar fashion in some respects, and we give this group of words the name noun. Lick, sit, own, jump, describe, believe, and so on, also all behave in a similar fashion in some respects, and we give this group of words the name verb.  When we say that something is a noun and has category N, all we’re saying is that that thing acts like the other noun-words. Likewise, when we say that something is a verb, we’re simply saying that that thing acts like the other verb-words. 

In the next few sections, we will go through the common syntactic, inflectional, and derivational properties of the different parts of speech. You can use these different behaviours to classify words by part of speech. However, as someone who  is able to use words in different languages, you already know these properties, at least subconsciously. We just need to learn to be consciously aware of these patterns. We can also look at data from a language we don’t speak to find patterns of behaviour of different words and identify parts of speech in these other languages.

Of course, different languages have different word order, as well as different morphological patterns, and so the criteria won’t be exactly the same across languages. Some linguists question whether all languages have the same categories at all! But there does seem to be some common patterns across languages. Many of the criteria we use in one language are parallel to similar criteria in other languages.

Categorizing phrases

Individual words are not the only things that have part of speech. Groups of words that behave as units—called phrases—also have distributional patterns that we can use to categorize them. Plus, there’s a pattern. Phrases have a head—the main word of the phrase—and the part of the speech of the head of the phrase will predict the distributional patterns of the phrase. For example, a phrase headed by a noun is called an NP (short for noun phrase). Any NP can be the subject of a sentence, as shown in (8), even if we add more words to the phrase, or if we switch the head noun for a a different head noun.

(8) a. [the apple] is red

b. [the sweet apple] is red

c. [the apple on the ground] is red

d. [the apple that fell off the tree] is red

e. [some paint] is red

f. [my bicycle] is red

Likewise, any phrase headed by a verb is a VP, any phrase headed by a preposition is a PP, and so forth.

Check yourself!

References and further resources

Attribution

Portions of this section are adapted from the following CC BY NC SA source:

↪️ Gluckman, John. n.d. Chapter 1: Syntactic categories. The science of syntax. https://pressbooks.pub/syntax/chapter/chapter-2-parts-of-speech/