There is no single syntactic or distributional property of nouns which is sufficient to guarantee the inclusion of an item in the word class.
For most modern grammarians there is a major subcategory of the noun class, the PRONOUN, the members of which show slightly different properties and distributions. The remainder of the class can be subdivided into the subcategories of COMMON NOUN and PROPER NOUN. Again, proper nouns have certain properties which mark them out from common nouns. For the purposes of this answer, I am concentrating solely on common nouns. Common nouns may be seen as the central, prototypical members of the noun class.
1. Four core syntactic functions/ grammatical relations
Huddleston & Pullum, 2005, describe four syntactic functions that may be carried out by phrases headed by nouns (including phrases consisting of just one word). Within the clause they function freely as Subjects, Objects and Predicative Complements. Within Preposition Phrase structure they occur freely as Complements. This fact is one aspect of nouns that is shared by all the subcategories of noun. Here are some examples with the common noun dog:
- Dogs bite. (Subject)
- I like dogs. (Direct Object)
- We are dogs. (Predicative Complement)
- I'm not scared of dogs. (Complement of Preposition)
If the word being investigated has one of these functions, or is the head of a phrase carrying out one of these functions then there is a good chance that it is a noun. Notice though, that all of these functions or syntactic roles may be carried out by other words or types of phrase. So having one of these functions is in no way a guarantee of nounship. For example, verbs, finite and non-finite clauses also frequently occur as Subjects:
- Smoking is bad for you. (Verb as subject)
- To err is human. (Verb phrase as Subject)
- That he was continuously late was a problem for us. (Finite clause as Subject)
- For him to do that would be inconvenient. (Non-finite clause as Subject)
A good question at this point might be: why does smoking count as a verb and not a noun here? The answer is that although it is in Subject function, it still retains the syntactic properties of being (a non-finite form of) a verb. For example, this verb can still take a Direct Object:
- Smoking cigarettes is bad for you.
It can still take a preposition phrase as a Modifier or Adjunct (read adverbial) like other verbs:
- Smoking in the morning is bad for you.
It can take an adverb as modifier:
- Smoking quickly is bad for you.
It is clear then that if we are to bother to have parts of speech these need to be separated out from the syntactic role or function that a word may have in a phrase or sentence. (We'll do a more thorough investigation of a case study at the end of the post.)
The syntactic role that a word has may be more or less important depending on what categories are contenders for the word. So for example, adjectives commonly function as Predicative Complements, so if the contending word is in Predicative Complement function, and we are wondering if it's an adjective or a noun, this is not much use to us. However, if we're wondering whether an item is an adverb or a noun, then the fact that the word is a Predicative Complement would be fairly decisive in showing it is a noun. Adverbs generally can't function as Predicative Complements:
- *They are worriedly. (adverb as PC - ungrammatical)
- They are worriers. (noun phrase as PC)
2. Other syntactic roles of common nouns
The prototypical functions described above apply to the whole class of nouns. Common nouns however have other syntactic roles that they often take. In particular nouns in English, as well as appearing as the Heads of Noun Phrases, can also be Modifiers within the Noun Phrase structure.
To illustrate, take any item that may have a wrapper, for example biscuits. We can freely generate the compound noun biscuit wrapper. Now if some mad people start collecting these, they will be biscuit wrapper collectors, where the compound noun biscuit wrapper is modifying collectors. If there were enough of them, these people might form a biscuit wrapper collectors association where you could be a Biscuit Wrapper Collectors Association member, a disgruntled one of whom might leave and become a Biscuit Wrapper Collectors Association member assassinator. You get the picture. Just in the same way that both adjectives and nouns can be Predicative Complements, they both freely function as Modifiers within Noun Phrases. If a word cannot be freely used to modify a common noun, then this word is almost definitely not a common noun (pronouns do not freely modify other nouns, nor are they freely modifiable by other nouns).
Nouns and noun phrases also occur much less freely in other types of syntactic roles. So for example, in the following sentence, the noun phrase every three weeks functions as Adjunct of the verb phrase (I use Huddleston & Pullum's terminology here, where ADJUNCT is a special term for a modifier of a verb phrase. Other grammars use the term Adjunct in the same way that the term MODIFIER is used here, namely for an item which is syntactically extra within any given phrase structure):
- Bob gets his back waxed every three weeks.
The phrase every three weeks here is a Noun Phrase. The head of this phrase is the noun weeks. Although this phrase is an Adjunct, there are no grounds for considering it an Adverb Phrase, nor for considering weeks an adverb. The word weeks here retains all of the properties of nouns discussed elsewhere in this piece. It is this fact that makes it a noun. So, the point here is that although the occurrence of nouns as Adjuncts in Verb Phrases is quite restricted, being an Adjunct does not prevent a word from being a noun, although it may make it statistically less likely to be one.
There are many other functions that nouns can have. For example they can be Modifiers in preposition phrases:
- They were five minutes into the match.
I will not attempt to give a definitive list here - if such a list even exists. Suffice it to say, that if a word has a syntactic role that is known to be performed by nouns, then in order to determine the word class of the item, we need to compare the ability of rival word classes to fulfill that role , and look at other criteria too.
Nouns are prototypically modified by adjectives, but as has been shown they are also very commonly modified by other nouns. However, nouns are rarely if ever modified by adverbs. In the very rare instances where they are, the adverb must necessarily come after the noun and not before it:
- The arrival recently of the plague was going to have disastrous consequences for the farmers.
- *The recently arrival of the plague was going to have disastrous consequences for the farmers. (pre-modification of noun by an adverb - ungrammatical)
If a word is modifiable by an adjective, and not by an adverb then it is very, very likely to be a noun. Adverbs can modify many other types of words and phrase, but they are excluded from pre-modifying nouns.
Nouns are frequently post-modified by other phrases, notably preposition phrases, relative clauses and adjective phrases.
Many nouns take various complements, such as finite clauses and preposition phrases:
- The knowledge that this will not be done easily.
- My dislike of it.
However, Huddleston & Pullum note that:
... nouns differ from verbs and prepositions in that they do not take objects: I dislike it, but not * my dislike it. (p. 326)
This is one of the factors that rules out smoking from being a verb in the examples further above.
Certain words occur nearly exclusively in Determinative function in noun phrases. These external dependents of the noun are a good indication of a word being a noun. Here are some examples:
- this patient
- every opportunity
The determiner this in the example above shows that this is the noun patient and not the homonymous adjective. The word every here is a near cast iron guarantee that the word opportunity is a noun. However, not all determiners are as good as others for this job. In particular, the can occur in Determinative function in phrases that have no noun:
- I'll take the blue one and the green.
- the good the bad and the ugly
- the bigger the better
Here we see the occurring with adjectives. Notice that every cannot be used in these ways.
6. Inflection for number
Common nouns usually inflect for number and for case. Typically, plurally inflected nouns take /s/ or /z/ as a suffix represented in the writing by the letter 'S'. If we see a word which seems to have an 'S ending because it is plural, this is a strong indication that the item is a noun. If a word doesn't inflect for number then this may be a sign that it is not a noun. However, we need to be a bit careful. There are many irregular plurals in English:
There are also nouns which are the same in the singular or plural:
- a fish, five hundred fish
- a sheep, two sheep
There are also nouns that end in an 'S' whether they are singular or plural:
- crossroads, means, kennels
There are also nouns that only occur in the plural:
- police, cattle, minutae, odds, doldrums, clothes
There are nouns that only occur in the singular, or are uncountable:
- crockery, luggage, equipment.
What this shows is that the ability for a word to show singular and plural forms with 'S' is useful, but a lack of inflection, or of regular inflection, is not sufficient reason to exclude an item from being a noun.
7. Inflection for case
Common nouns inflect for case. They have two forms. The uninflected form is often called PLAIN CASE. The other case is GENITIVE. The genitive form involves the suffix/clitic /s/ or /z/, represented in the writing as 'S' used in conjunction with an apostrophe:
- people, people's
- woman, woman's
- women, women's
- baboon, baboon's
We need to be a little careful here. First of all, this clitic actually gets appended to noun phrases, not just nouns:
- The woman you like's boyfriend
Here we see this clitic appearing after the noun phrase the woman you like. We should not infer from the writing that like is a noun here! Secondly some phrases without nouns can also take this clitic:
- The blind's access to braille versions of periodicals ...
Here we see the clitic after the phrase the blind. Some writers like Huddleston & Pullum regard the blind here as a noun phrase without a noun. Here we see this clitic attached to an adjective.
8. Negative attributes
Of course, when considering whether a phrase is a noun or some other type of word, it is the properties that another class of words have and that nouns don't, which will provided the decisive data. Importantly, nouns can't function as the head of a main clause. In other words they can't assume the syntactic role of Predicator (the function carried out by the verb).
Another feature of nouns is that they are not gradable, whereas many other adjectives are. Equally it is the property of not being modifiable by adverbs which is crucial for demonstrating that a noun modifying other nouns is still a noun and not an adjective.
There are many other negative criteria we can use to demonstrate that a given word is likely to be or definitely isn't a noun.
9. The pronoun substitution test
It is sometimes posited that a good test to see if an item is a noun is to see whether it can be grammatically replaced with a pronoun. In actual fact, although nouns will often be replaceable with pronouns, more often than not in declarative sentences so will any other phrase in Subject or Object function, or functioning as the Complement of a Preposition:
- 1 a. To err is human.
- 1 b. It is human.
- 2 a. I don't like that you always take the window seat
- 2 b. I don't like it
- 3 a. I'm worried about him eating all the strawberry ones when I'm not looking.
- 3 b. I'm worried about it
None of the items that have been replaced by the pronoun it here is a noun. Notice as well, that pronouns don't 'replace' nouns if the noun has dependents. When we substitute a phrase with a pronoun, the entire noun phrase goes, determiners, modifiers and complements too. In addition we cannnot use this test when a noun has the function of Modifier within a phrase including when they are Adjuncts (- modifiers in verb phrases). The nouns and noun phrases in the following examples cannot be replaced by pronouns:
- three miles away from here
- miles wide
- two days late
- the writer James Joyce
- we went there the day before yesterday
It seems that the pronoun substitution test is of limited value. The item being replaced may not be a noun. Furthermore nouns can often not be replaced at all depending on their syntactic role. If we have to determine the syntactic role before we do the test there is no point doing it. The reason is that the syntactic role will already tell us how likely an item is to be a noun with the same reliability as the pronoun test. This is because in declarative sentences any item in Subject function will be able to be replaced by a pronoun, whether it's a noun or not. So the chances of the word being replaceable by a pronoun are the same as the chances of the word being a Subject. Neither of these is a guarantee of nounship. The syntactic role that an item has in a given sentence is very rarely a definitive indication of the category of a word. The reason is, of course, that all syntactic roles apart from Predicator may be carried out by more than one type of word or phrase.
A trivial point, a red herring
There are of course some words which sound and are spelt the same. This includes words whose meaning is radically different, as well as words with very similar meanings which belong to different parts of speech. It is common in linguistics to talk about lexemes. We require the notion of a lexeme to indicate that words like liked and liking are versions of the same entity, namely they are inflections of the verb like. So liked and liking are instantiations of the same lexeme.
However, we would not say that man in man of the match and man the decks are instantiations of the same lexeme, even though they may look and sound identical. The reason is that the first is an instance of the noun man whereas the second is an instance of the verb. Because we consider them as instantiations of different lexemes with the same sound and the same spelling, these words are considered homonyms. Although they have some semantic similarities, they represent different types of concept. Also syntactically, the noun man belongs with other instantiations of the noun such as man's, men and men's, whereas the verb belongs with mans, manning, manned and the present tense form man.
So what does this have to do with the question here? Well, it is often said that the category of a word depends on how it used in a sentence. Now if "word" here means a particular series of sounds or a group of letters, then this is trivially, but maybe interestingly, true. We can only understand whether a group of letters represents the noun lexeme man or the verb lexeme man, for example by seeing this group of letters in context.
However, there's another, very harmful and unprincipled idea which sounds very similar, which is that a word should be assigned its part of speech according to what syntactic role, or grammatical function it has in a particular sentence. People who have this kind of idea say things like:
- this word is modifying a noun therefore it is an adjective.
The problem with this type of theory is that it works on the schoolkid assumption that syntactic roles belong to specific word categories. This is because, sadly, the teaching of grammar in schools is of particularly poor quality. Most people are taught at school that words that describe nouns are adjectives, words that describe verbs are adverbs, nouns represent things and verbs represent actions. None of this of course is true.
Here is what Geoffrey Pullum has to say about this problem. The quote is taken from LEXICAL CATEGORIZATION IN ENGLISH DICTIONARIES AND TRADITIONAL GRAMMARS 2009:
Most of the deepest blunders in English grammar as traditionally presented over the past two or three centuries stem from a single long-standing confusion between (i) grammatical categories or word classes; (ii) syntactic functions or grammatical relations; and (iii) semantic and discourse-related notions.
It is surprising to see the tenacity of this confusion. It does not appear in other domains. People do not confuse butter knives with screwdrivers, even though occasionally someone who cannot find a screwdriver may use a butter knife to turn a screw. Yet in grammar people just cannot keep syntactically relevant categories or classes of words separate from the relational properties they have when used in particular constructions, and cannot keep either separate from meaning. They insist on trying to define the first of these in terms of the other two, and they have done so since the very earliest attempts to write grammars of English.
What this means is that we should be able to distinguish between the fact that something modifies a noun and the fact that it is a particular part of speech. Additionally, ideas such as "this word tells us when something happened, so it is an adverb" or even worse "this word tells us when something happened, so it is being used as an adverb" are severely misguided. Neither the semantic content of an item, nor its syntactic role are defining characteristics of its part of speech.
Probably the easiest way to understand how we might run various tests to check the noun status of an item, is to look at these tests in action. There are (or will be shortly) two case studies of particular words in a different post here.