How Is Syntax Studied?



Constituency Tests

Syntactic Tools part 1: Constituency Tests

Syntax can be defined as the study of how words are combined into sentences and how sentences are linked to each other, giving shape to what is known as sentence structure. Generally speaking, words are at first combined to create phrases, namely strings of words, which behave the same way linguistically; phrases are then bound to form sentences. Certain combinations thus occur before others, and syntax and constituency aim at unveiling precisely what the order must be.

Let us look at the following example, where square brackets have been used to highlight when syntactic links are formed:

(1) [The [car]].

Without worrying about what the labels are for such combinations, it is hopefully evident from the example above that [car] constitutes a phrase on its own and is subsequently linked to ‘the’. The order in which combinations similar to the above occur is typical of numerous strings in language, and can be applied to longer ones, as in (2) below:

(2) [With [the [car]]].

As in (1), in (2) ‘car’ is first linked to ‘the’; ‘the car’ is then linked to ‘with’.

Let us now introduce the labels usually used by syntacticians to refer to the previous strings:

  • ‘Man’,’oranges’ as well as any other common nouns form a Noun Phrase. Therefore [man] or [oranges] would be labelled as NP.
  • When such nouns as ‘man’ or ‘oranges’ are preceded by ‘the’ or ‘a/an’ (generally known as determinate and indeterminate articles), the resulting phrase is called a Determiner Phrase. Therefore, [the [man]] or [the[oranges]] would be labelled ‘DP’. The same would apply if an NP was preceded by a ‘this/that/these/those’ (a.k.a. ‘Demonstratives’) or any other words which would, so to say, make the NP more specific, or more determined.
  • When DPs such as [the car] are preceded by ‘with’ or ‘in’ or ‘for’, the resulting phrase is called a Prepositional Phrase, or PP.
    It is now possible to rewrite (1) and (2) as in (3):

(3) PP[in DP[the NP[car]]].

From (3) it is possible to generalise the following syntactic rules

  • an NP will always contain a common noun;
  • that a DP will have a determiner preceding an NP;
  • prepositions will be at the head of a DP.

In order to look at how syntax is studied, it is useful to introduce another concept, namely the concept of constituency. What is a constituent? A ‘sentence [or phrase] consists of words; alternatively, words are the constituents of a sentence’ (Thomas 1993:2). If a word, or a string of words, is a constituent, we can manipulate it as a syntactic unit of the sentence.

The way to prove the correctness of (3) is by applying so-called constituency tests.

A very useful constituency test is substitution. It enables to determine the type of phrase by replacing the relevant word (or string of words) with another one which belongs to the same category [*].

Therefore, to prove that car is an NP in (3) we replace it with another common noun; if the substitution yields a string that is similar in meaning, the constituency test has worked. It is immediately noticeable that in saying PP[in DP[the NP[van]]] or PP[in DP[the NP[bike]]], the meaning is roughly the same. This first test has therefore been successful.

What syntacticians usually do to test for DPs is they replace the string with a pronoun. Again, if the substitution leaves the meaning roughly unchanged, then the constituency test has worked. We notice that PP[in DP[the NP[car]]] is quite similar to PP[in DP[it]], only there is less information in the second, due to the presence of the pronoun. This second test has also worked.

Using substitution to test for PPs usually entails replacing the PP with ‘there’. If the meaning is roughly the same, we can be quite sure that the string is indeed a PP. This seems to be the case in (3), if, without worrying too much for now about the labels for what comes before the PP, we say:

(10) The man PP[in DP[the NP[car]]] → The man PP[there]


Lexical Entries

Syntactic Tools part 2: Lexical Entries

Once we have learned (using constituency tests) which parts of the sentence constitute each phrase, we can then assign a head. If we know the name of the phrase that we have identified, then finding the head is easy. For example, the head of a DP is the determiner, the head of a VP would be the verb. For a complete guide with examples please see the table below:


Each head of a phrase has its own lexical entry, which is stored in our brains in something linguists call a lexicon (like a dictionary). This tells us certain features about the head including where it can be positioned in a sentence. The lexical entry also shows what can/must precede and follow the head (syntacticians call these requirements selection and c-selection, respectively).

If we look at the future tense marker will (this would be the head of the TP), then we can make a lexical entry for it, as follows:

will (future): T, selects DP subject, c-selects VP

So let’s break it down.

  • The T tells you that will is a type of T(ense)
  • ‘selects DP subject’ shows that the head must be preceded by a DP (the subject of the sentence in this case)
  • ‘c-selects VP’ shows that the unit following the T head will be a full VP.
  • Notice that the VP is itself a phrasal unit. Heads often select for phrasal complements rather than simple heads because syntax is hierarchical, rather than linear! (See the Constituency section for discussion of this).

Car: N
Blue: A
The: D, c-selects NP
Wash: V, selects DP subject, c-selects DP
Quickly: Adv

Why do you think that the noun car or the adjective blue don’t select or c-select for anything? Intuitively, what is different about them versus a determiner like the or a verb like wash?

A global reputation

Sheffield is a research university with a global reputation for excellence. We're a member of the Russell Group: one of the 24 leading UK universities for research and teaching.