Word Classes

In the previous section, we’ve tried to narrow down the exact meaning of the concept word, and seen that it’s by no means trivial to define this properly, so that we probably need to content ourselves with a rough and possibly incomplete definition. What we’ve also seen is how to mark up and indicate what the individual parts of a more complex word are from a morphological perspective, and roughly how we may be able to determine the order in which the different morphemes have been joined together. One of the main objectives of morphology is to identify potential regularities in this production process for new words and word forms. Another is how to explain these, based on the properties of specific bases, possibly also including their origin. This is dealt with under the heading of word-formation, which we’ll cover on the next page. Before we can explore the individual properties and functions of the different types of morphemes in this process, though, we first need to develop an understanding of what the major word classes are and how they ‘behave’ morphologically. For a more detailed discussion, you can consult a grammar, such as the Longman Grammar (see reference at the bottom of the page).

While some of the examples you’ll find below have just been invented for the sake of convenience, I’m also including samples from the COCA (Corpus of Contemporary American English) corpus and the BNC (British National Corpus), two large and representative collections of electronic text, as well as the SPAADIA corpus, a smaller corpus containing transactional dialogues.

Each word, according to its function in context (i.e. morpho-syntactically) can usually be assigned to a particular word class. The list of word classes we tend to use these days in linguistics is still predominantly based on the word classes as defined by scholars in ancient Greece or Rome, though, and hence tends to reflect the types of word classes – or parts of speech (often abbreviated PoS), as the Romans called them – of the classical Indo-Germanic languages. This, however, may make it at least partly unusable, even for another Indo-European language like English, so we should spend some time to consider what they might mean in our case.

An important distinction we can draw, even before looking at the exact nature of each word class, is that between open and closed classes. Words that have a predominantly grammatical function and relatively little meaning themselves generally belong to the latter category, e.g.:

As you can hopefully see from the above examples, this group of words is pretty much fixed, and belongs to the absolute core of the vocabulary. In other words, even though the words themselves may not have much meaning on their own, we need them so frequently to link together different parts of a sentence, or to create a sense of continuity (cohesion) between different parts of a text, that those are among the most important words to learn (and understand). In contrast, now look at the following examples:

All the above words carry a high degree of meaning, and new members of this group can easily be created and added if new things are created or new concepts developed, so that the vocabulary has to be expanded to accommodate these items.

Before we look at the individual word classes more closely, take a brief look at the above examples again to see whether you can already identify some tentative morphological properties for them.

Determiners

Determiners represent a kind of special word class that does not exist in the same way in many languages around the world, which is why often their proper use is difficult to learn. The most prototypical member of this word class is the article, which, in English, can take three forms, a definite (the), an indefinite one (a/an), and a negative one (no). Other members of this group are demonstrative (e.g. this person, that house) and possessive determiners (e.g. my car, your idea). Definite articles, demonstrative determiners, and possessive determiners generally have the function of marking any nouns they ‘accompany’ (i.e. occur in front of) as being specific, unique, or easily identifiable, while the indefinite article has more or less the opposite effect, i.e. either indicating that something is unspecific or only one out a number of potentially ‘limitless’ options.

Numerals & Quantifiers

Numerals (i.e. numbers) are similar to the definite types of determiners in that they allow us to provide a very precise specification of the amount or position of an item under discussion. Cardinal numbers (e.g. 1, 2, 5, 10, etc.) do the former by providing a definite count, while ordinal numbers (e.g. 1st, 2nd, 5th, 10th; first, second, fifth, tenth, etc.) indicate a relative position within some form of temporal or local sequence, or ordering/ranking/hierarchy if used with nouns, in which case they also need to be preceded by a determiner. In contrast to determiners, though, which can almost only go with noun-structures, numbers – especially ordinals – can also have textual function, indicating a sequence within a text to describe specific steps in a process or set of instructions, or develop an argument progressively.

Quantifiers allow us to provide a rough, but inexact specification of how many items are being talked about. We can most easily ‘visualise’ them as items on a range, starting from a state that describes complete absence, and moving towards one of complete ‘presence’ or inclusiveness, i.e. no → few/some → many → most → all. Please note that these examples are only considered quantifiers if they co-occur with nouns, but would otherwise be considered indefinite pronouns (see below).

Try to identify the different functions of the numerals, quantifiers, and determiners in the examples below, based on the descriptions above. Also compare them to one another and see whether you can identify any special cases of use the above descriptions may not cover. If so, how could you possibly explain them?

Pronouns

Pronouns are cohesive (‘linking’) or interpersonal devices. As their name implies, they are commonly used as a possible ‘replacement’ for nouns – or noun phrases, to be more precise. At least this true for the 3rd person personal pronouns he, she, it, and they, the possessive pronouns mine, yours (sing./pl.), his, hers, its, ours, theirs as well as the demonstrative ones this & that, and these & those. Of course, we here need to distinguish some of the possessive forms from their use as determiners, where they don’t replace nouns, but co-occur with and specify them.

The situation is different for the 1st & 2nd person non-possessive personal pronouns I, we, and you. These indicate discourse referents, i.e. the interactants in any form of communication, who are referring to one another, which thus makes these pronouns purely deictic (‘pointing’) or indexical (‘indicating’) in nature. It would simply be strange to assume that they replace nouns – even if the interactants themselves obviously have names –, especially because the names would constantly be shifting, depending on who’s currently speaking, i.e. happens to be the ‘I’ at any given point in the interaction. The (direct & indirect) object versions of the personal pronouns I, he & she, me, him & her, are further special cases, as they represent remnants of the Old English case (marking) system, and thus some of the few remaining features of inflection Modern English still retains, along with their reflexive counterparts, myself, himself/herself, as well as itself. The same applies to the object form of the interrogative & relative pronouns, whom, which complements the basic system of subject interrogative & relative pronouns, who, what & which. Standard English, however, unlike some dialects, does not allow what as relative pronoun and replaces it with the the demonstrative form that.

The indefinite pronouns no one/nothing; none → someone/something; some/few → many → most → everything/everyone; all are the nominal counterparts to the quantifiers we saw above, only that they do nor co-occur with any nouns, but ‘represent’ them instead. Note that this range also includes a possible distinction between inanimate objects and people, as well as individuals (sing.) and groups (pl.).

Prepositions/Particles

Prepositions are devices that generally help to turn the noun phrases they precede – and thus form a prepositional phrase with – into locative or temporal indicators, i.e. place or time adverbials that describe specific circumstances of a process, event or object. Thus, in in In the evening, they went to a nice restaurant., in conjunction with the temporal noun evening, produces a temporal specification of an event, whereas in The food is in the kitchen., it helps to specify a location. Both locative and temporal prepositions can have more static/stative or more dynamic uses. For locative prepositions, we can distinguish between a use that indicates a location (e.g. The neighbour was in his house.) or a direction (e.g. The neighbour went in(to) his house.), and for temporal ones, those that specify a fixed point or period in time (e.g. They met at noon), as opposed to those that indicate a period/duration of activity (e.g. They played tennis for the whole afternoon). Whether the static or dynamic meaning is invoked depends on the nature of the verb.

If prepositional phrases don’t indicate a location or time, they may specify a particular manner, as in e.g. She explained the matter in a very concise way.. In this way, they may be similar to adverbs, which we’ll discuss further below.

Prepositions can also be complex when multiple prepositions occur in a row to form a complex meaning, e.g. up to or down in. We already saw something similar above in the example that indicated directionality, where we can use into instead of in to make it more obvious that we’re talking about a direction, rather than a location, and where we can probably assume that the two prepositions in and to simply became ‘fused’ at some point in time.

The prepositions of & to have a special status because they, apart from just indicating basic relationality, they can also signify possession and mark an infinitive, respectively. And, although many prepositions can be multi-functional, these two functions are not covered by any of the others.

Sometimes, we also find forms that look like prepositions following verbs whose meaning they modify. In this case, we can refer to them as particles, as in e.g. He turned up/down the volume. or He turned/switched on/off the sound.. Often, this occurs with relatively basic, non-specific, verbs, such as take, give, bring, etc., and helps to create more specific and diverse meanings from them.

Another important particle we shouldn’t forget to mention here is the negation particle not and its contracted counterpart n’t.

Conjunctions

Conjunctions, just like prepositions, allow us to specify associations or dissociations, either between different elements in one and the same clause, or between multiple clauses. Those conjunctions that create an association between items are referred to as coordinators, and those that have a dissociating effect as ‘subordinators’.

For coordinators, English, unlike some other languages, such as Chinese, not only allows phrase-, but also clause-level coordination. The main coordinators are and, but, and or. And basically has additive meaning on both levels, either grouping together and creating a relation between nouns/noun phrases, e.g. the dog and the cat, or doing the same for clauses, as in e.g. They stayed at home all day and watched television.. On the clause level, and may also have an extended meaning of ‘and then’, indicating a specific order in a sequence, as in She got up and (then) had breakfast., where the first clause expresses an event/action that is completed prior to the second one.

But always has a contrastive meaning, as in the influence of communication technologies on the interaction of citizens and their governments show small but significant differences across technologies. (COCA), A computer fault put the line out of action temporarily, but it is now up and running again. (BNC). In our two examples, the comma before the clause-level clearly marks the dissociation between the two clauses, while the absence of a comma in the phrase-level coordination marks a kind of association between the two properties of the noun differences. Indicating the difference in this way in writing is good practice, but unfortunately many native-speaker authors tend to ignore this option. In spoken language, this distinction is also present in that there is generally a distinct pause before the second clause, whereas at phrase-level, there is either no pause or a minimal one for skilled speakers.

Or almost always expresses an alternative, and its usage is relatively straightforward. There may be a distinction between the phrase and the clause level in usage, though, at least in alternative yes-no questions. In spoken language, we can (relatively) easily make a distinction between what I call a ‘true’ and a ‘false’ alternative, the true one being where the respondent needs to make a decision between two items in a list, indicated by a rising tone on the first item, a pause, and a falling one on the second, Would you like ↗ tea, or ↘ coffee?, while, in the false one, indicated by a rise that starts on the first item and continues until the end of the question, Would you like some ↗ tea or coffee?, the choice is not between having either tea or coffee, but between whether to have one of the two or not. In my examples, the first one is more similar to the earlier clause-level coordination we saw for but, and we could actually expand this into Would you like tea, or would you prefer coffee instead?. Please note that, in writing, we again have to option to let the comma indicate the ‘dissociation’, although, as before, you’ll probably rarely find native speakers making this distinction. To establish two options more clearly, we can also use the complex coordinators, either ... or, or their negative counterparts, neither ... nor.

‘Subordinators’, such as because, if, whether, while, (al)though, that, link clauses, where the second clause is often referred to as being ‘dependent’ on the first. This, however, may be a slight miscategorisation in traditional grammars because, frequently, the so-called ‘dependent clause’, i.e. the one that begins with the subordinating conjunction, may either occur before or after the other clause, which then changes the focus or emphasis. To illustrate this, let’s take a look at two alternative versions of two combined clauses that express more or less the same thing, where the first one is slightly adapted from the SPAADIA corpus:

  1. If you miss the service I’ve reserved you on, you’re able to get the next available train.
  2. You’re able to get the next available train if you miss the service I’ve reserved you on.

Example a) here focusses on the condition marked by the if, while b) focusses more on the possibility, adding the condition in a less marked position and with less emphasis. This can again be seen in the prosody, where in a), there would normally be a clear pause between both clauses, while in b), the pause is likely to be very short, if not completely absent. Due to this exchangeability and the fact that both clauses clearly belong together, it would probably be best to use the term ‘co-complementiser’. This is because ‘subordinator’ somehow seems to imply that the clause introduced by the conjunction is less ‘important’ than the other, while both are somehow co-dependent, and neither fully makes sense without the other, even if the one without the conjunction can of course occur independently.

Nouns

The first major distinction we need to make in the category of nouns is that common and proper nouns, where the first describes general (more or less) concrete objects, people, events (e.g. house, car, water, weather, temperature, builder, actor, wedding, play) or abstract notions or categories (thought, feeling, concept, design, plan, mankind), and the latter what we commonly refer to simply as (proper) names, i.e. the names of people, places, or specific – usually unique – objects that have a specific cultural value (e.g. John, Mary, Britain, the US, the Tower of London, the Eiffel Tower).

At another level, we can distinguish between count (e.g. day, picture, book, property) vs. non-count (e.g. love, water, sand, fruit, software, research, feedback), and plural (only) nouns/noun forms (e.g. people, arms, children, oxen, works). The first category are the most common and frequent ones that describe basic, countable entities which always need to be marked for their plural if they occur more than once. The second group are less frequent, and tend to represent either masses or abstract concepts or processes, where the common element is that their exact extent is not clearly defined according to their size or other dimension, such as e.g. time. Finally, the third type, proper nouns, is again fairly frequent, due to the large number of possible names in different languages. But whereas most common nouns have a relatively clearly defined meaning, proper nouns usually do not, even if their origin can often be traced to more meaningful words or characters, depending on the language. One very clear indication of their inconcrete nature is that different people may have the same name, and every time we refer to one of these namesakes, we either need to indicate clearly who exactly we’re talking about or have to be absolutely sure that the context disambiguates the person. In this way, names are very similar to the indexical pronouns we saw above.

Go through the above example words and try to understand their individual characteristics, including the use of articles with them, to the best of your ability.

Apart from the general ability to occur with determiners and numerals – at least for singular forms of count nouns –, most nouns can, or often need to, be modified by adjectives (attributive pre-modification), or relative clauses and prepositional phrases (post-modification) in order to reach their full meaning potential. Thus, the ability to occur with determiners, quantifiers, numerals, and/or modifiers is often seen as distinguishing criterion for nouns, especially if they look somewhat verb-like. For instance, if we take the complex noun phrase the many delightful works of John Donne, the fact the plural noun works, which looks exactly like the 3rd person singular verb form of work, can only be a noun because it is both preceded by a determiner, a quantifier, an adjectival modifier (adjective), and followed by a post-modifiying prepositional phrase.

Another interesting feature of non-count nouns is that they can often be made countable in different ways, which may, however, also entail a change in meaning. So, for example, if we pluralise a noun like fruit, we effectively end up not talking about many items of exactly the same type, but of different kinds. At other times, we can make more abstract non-count words countable by referring to either individual instances of them or parts of the whole, essentially using a classifier system, similar to that in Chinese or other Asian languages, as e.g. pieces of research/software, or grains of sand/salt. For the former type, we also have another option, where we can use the noun in order to modify another noun that is countable, for instance in referring to research projects or software programs, where, of course, the latter is a tautology ;-)

Verbs

As with nouns, we can again sub-categorise verbs according to different criteria. One of them is the distinction between the use as a main/lexical vs. an auxiliary verb. Main verbs (e.g. go, like, dream) can occur on their own and essentially carry full semantic meaning, while auxiliaries (e.g. will, may) ‘behave’ like function words in that they support other verbs to create tensed complex forms or express some form of modality. Auxiliaries can therefore never occur on their own, unless they occur in elliptical constructions where the main verb is implicit, such as in elliptical answers to questions like Can you please help me with this? Yes, I can.

Amongst the auxiliaries, we can again sub-divide between primary (be, have, do) and modal ones (can, could, may, might, will, would, should). The primary ones are special case here because they can also act as full verbs, as in e.g. I’m a teacher., They have two children., They’re doing business with a large company.. Be as a full verb is also referred to as a copula verb, a verb that expresses some form of relationship between the subject and a noun or adjective complement.

Lexical verbs can express physical (e.g. go, swim, build), verbal (e.g. say, declare, state), & mental actions (e.g. think, plan), physical (e.g. be, exist, feel (cold)) & mental/emotional states (feel (sad), wish, want, hate), changes in state (e.g. cause, force, finish), etc. Here, we can roughly distinguish between verbs that are stative, i.e. express a state, vs. dynamic, i.e. reflect an ongoing action, although much finer distinctions are still possible.

As a PoS category, verbs constitute the only category in English that is still strongly inflected, even more than the pronouns we discussed earlier, and thus retains some of the characteristics we already encountered in Old English. In the use of verbal inflections, English distinguishes between four basic properties of the verb, the ability to express tense and aspect, person & number, which can co-occur, but in limited combinations. The base or infinitive form of a verb (e.g. walk, talk, eat, do, make) in most cases also represents the non-3rd person singular present tense forms, while the 3rd person form generally just adds an -{s} morpheme, although minor adjustments (e.g. do→does, with <e> insertion) may occur in the process. When we add a morpheme to the end of a word like this, this morpheme is normally referred to as a suffix.

Adding an -{ing} suffix to the base of the verb, again with possible minor modifications (e.g. give→giving, with <e> deletion in orthography) changes the aspect of the verb to a progressive (i.e. dynamic) one that indicates continuity, rather than a potentially fixed state. This is often traditionally also referred to as a present participle, but it’s probably best to avoid this term as this for does not only occur in the present tense, and therefore a better – more neutral – term to use would be ing-form.

In the same vein, it’s advisable to use the term ed-form to refer to what’s commonly been called a ‘past participle’. This term covers both the allomorphs -{ed} and the older, irregular, form -{en}, another remnant from Old English times, just like the irregular (strong) past tense forms of some verbs that exhibit a change in the vowel. Perhaps one of the best examples for this is the verb give, where the paradigm includes both a change in the vowel for the simple past tense form gave, as well as the perfect ed-form given. In contrast to the ing-form, which indicates something ongoing and unfinished, the ed-form signals the aspect completedness/completion. Another reason for dispreferring the term past participle is the that this verb form is also a part of the passive construction.

In some linguistic theories, verbs are seen as the most important components of a syntactic unit, and this is certainly true to some extent, as can easily be seen in the influence they exert on other components, at least as far as their number is concerned. Thus, it depends on the nature of the verb whether one or more noun phrases are required to create a complete (non-elliptical) syntactic unit. A verb like rain, for instance, only requires a single (subject) noun phrase (it) to occur with it, although it is possible to add other (object) noun phrases, as e.g. in the idiom It’s raining cats and dogs., or if used metaphorically in other ways that don’t actually involve rain, e.g. The rough floorboards shook overhead, raining dust and splinters. (BNC). Verbs like this are generally referred to as intransitive. Verbs that call for one single object are called mono-transitive, e.g. Most people won't see [the results of doing crunches], so why do it? (COCA), while those that need need two objects to convey their full meaning are di-transitive, e.g. Some analysts believe [X] is in a strong enough position to give [shareholders] [a marginal increase in their dividends]. (BNC). That the verb cannot be assumed to be in complete control of the whole syntactic unit, however, can be seen in the fact that the subject of a sentence, depending on its number, forces a verb into subject–verb agreement in the 3rd person singular present, apart from when the verb is a modal. As we’ve also seen before, some verbs require particles to go with them in order to acquire their full meaning potential.

Adjectives

Adjectives are essentially nouns modifiers, i.e. they somehow change the nature of the nouns they occur with. We can distinguish between two different types of use of adjectives, an attributive one, where the adjective precedes the noun (a green light), vs. a predicative one, where it follows it, linked via a copula verb (e.g. be, appear, seem), a verb that indicates a change of state (e.g. The light is/turned green.), or a referring verb, as in Did you just call them silly?, etc.

In terms of inflectional morphology, most adjectives can be inflected to allow comparison, i.e. to create a comparative form by adding -{er}, or a superlative form appending -{est}, possibly again with minor modifications (e.g. hot→hotter→hottest, with reduplication of <t> orthographically). Not all adjectives allow such inflectional adjustments, though, and when this isn’t possible, we need to resort to grading via quantification through more & most (e.g. more unlikely & most unlikely).

As we’ve seen before, some adjectives actually look like ed-forms of verbs (e.g. scented in scented oil) and may originally be related to passive constructions, expressing a result of some operation/process (i.e. The oil was scented.), while others look like ing-forms, where they generally retain their character of indicating an ongoing process or habit (e.g. dancing in dancing girl). Please note, though, that in both the previous examples, there has been a further process involved that later turned the respective adjective + noun combinations into a compound nouns, where prosodically, only a single (primary) stress remains on the word.

Adverbs & Discourse Markers

Essentially, adverbs are modifiers, just like adjectives, only that they can modify a wider range of types on the phrase level, apart from also having the potential to modify whole clauses. On the phrase level, they can either modify the meaning of the verb – which is where the original name came from – as in She ran quickly., or that of an adjective (phrase), e.g. very big, while on the clause level, they affect the whole proposition of the clause, as in Unfortunately, I couldn’t finish my work on time..

Although many adverbs simply look like adjectives that have had the suffix -{ly} attached to them (e.g. quickly, strongly), of course this doesn’t always have to be the case, and some adverbs actually look exactly like their corresponding counterparts (e.g. hard, fast), even if, misleadingly, there appears to be a corresponding -{ly} form, as in hardly, which usually means something like ‘almost not’, rather than ‘in a hard manner’, as learners of English often assume. A fair number of adverbs related to times and places, such as e.g. here, there, now, then, tomorrow, yesterday, look nothing like any adjectives, perhaps because they perform a particular deictic (indexical) function that only works on the clause level, where again, the exact reference is difficult to determine and only becomes clear in a particular context. This is perhaps most apparent in the well-known example reported to have been found on an office door as a note saying I’ll be back in 20 minutes, where, without knowing when the 20-minute time period actually started, we never know exactly when to expect the person who wrote it to be back, although we always know that this should never be more then 20 minutes form the time that we read it, which would be ‘now’ for us.

Another interesting feature of certain English adverb forms is that they may in certain contexts loose (at least part of) their modifying function, and become discourse markers instead. This is, for instance, the case with now & well, where, in certain positions in a text/dialogue, they indicate particular stages or directions in the discourse. For example, now at the beginning of a speaker turn, i.e. a more or less uninterrupted period of speaking by one particular person, frequently loses almost all of its temporal deictic meaning, but simply comes to mean something like ‘at this stage in the dialogue’ or ‘moving on to another topic’, while the meaning of well in the same position is nothing like the general adverb form of good, but instead indicates that the speaker is somewhat reluctant to make a commitment, i.e. to fully agree or disagree with what has just been said/suggested/asked.

Sources & Further Reading:

Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman Grammar of Spoken and Written English. London: Longman.

Davies, M. (2008-) The Corpus of Contemporary American English: 450 million words, 1990-present. Available online at http://corpus.byu.edu/coca/.

Leech, G. (1992). 100 million words of English: the British National Corpus (BNC). Language Research 28, 1 (1992). 1-13.

Leech, G. & Weisser, M. (2013). The SPAADIA Annotation Scheme.

Payne, T. (1997). Describing Morphosyntax: a Guide for Field Linguists.. Cambridge: CUP.

© 2014 Martin Weisser; last edited: 05-Jul-2014 09:53:11