logo  ELT Concourse teacher training
Concourse 2



I spotted the man with a telescope

This guide concerns how we unravel what is meant by that sentences.
We cannot be sure whether this means:
    I used a telescope to spot the man
    I spotted the man who had a telescope

If you have not yet followed the guide to phrases on this site, now's a good time to do that.  If, however, you are comfortable with identifying Noun phrases, Adverb phrases, Verb phrases and Prepositional phrases, read on.


What makes a constituent?

The issue here is often one of ambiguity.  Take, for example, an apparently simple sentence such as


The teacher shouted at the child in room 13

A moment's thought will reveal that there are two possible ways to understand this sentence.
Click here when you have figured out what they are.

We can only understand which interpretation is the correct one by identifying the constituent parts of each clause, like this.

For interpretation 1 (the object of the shouting is the boy in room 13):

Constituent 1 Constituent 2 Constituent 3
The teacher shouted at the boy in room 13
noun phrase verb phrase noun phrase

For interpretation 2 (the shouting happened in room 13):

Constituent 1 Constituent 2 Constituent 3 Constituent 4
The teacher shouted at the boy in room 13
noun phrase verb phrase noun phrase prepositional phrase

The trick is to decide, in this case, whether the prepositional phrase, in room 13, is a constituent in its own right, in which case Interpretation 2 applies, or whether it forms a part of the constituent noun phrase, the boy in room 13, in which case Interpretation 1 applies.

If you prefer a diagram, it looks like this.  The subject noun phrase is in green, the verb phrase in blue and the object noun phrase in red.  The question to answer is
Does the prepositional phrase tell us about the verb (in which case it is an independent constituent) or does it modify the noun (in which case it forms part of the constituent noun phrase)?

If you yearn for a neat technical term for this kind of phenomenon, it's called syntactic ambiguity.


Levels of ambiguity

There are many times, of course, when no, or very little, ambiguity exists and in these cases we can normally rely on our intuition to decide what qualifies as an independent constituent.  For many learners, intuition alone is not always enough so we need to be clear in our presentation.
For example:

It is not only noun phrases which need to be disambiguated.  The same phenomenon of trying to decide what belongs where can occur with verb phrases:  For example, the sentence:
    The people who came quickly got lunch
has two interpretations:

  1. The verb phrase is came quickly in which case we have:
  2. The verb phrase is quickly got in which case we have:

The only way that can properly be disambiguated is by pausing in speech after quickly (and signalling sense 1.) or after came (and signalling sense 2.).
In writing, only re-phrasing the sentences will completely remove the ambiguity although inserting commas (after quickly and came respectively) may help.  Moving the adverb settles the matter but for many, quickly came sounds clumsy.  We might re-phrase, then, as:

  1. The people who were quick to arrive got lunch
  2. The people who came got lunch quickly.

As you can see, the situation is not at all obvious so we need to have some tests to determine which phrases in a clause actually form independent constituents.


Testing for constituents

There are a number of tests we can apply to see if our intuitions about what is really an independent constituent and what is a phrase forming part of a constituent are correct.
Authorities vary on how many of these tests there are and whether they are all valid but we will content ourselves with six well known ones.


Test 1: joining another constituent (the coordination test)

Depending on the function of the phrase, we have to use a different sort of language if we want to add to the phrase in question.  For our example:

In the first case, we have to add another noun phrase (all his friends) to perform the same function and in the second case we have to add another prepositional phrase (in the corridor) to perform the same function.


Test 2: making a passive

When we make a passive, we keep constituents together.  For our example, two passive sentences are possible and they reveal which interpretation is the right one:


Test 3: making a cleft

There are a number of ways to make a cleft sentence (there's a guide on this site) but one example will do to show how the results differ depending on the nature of the phrase.  This is called an it-cleft, incidentally.


Test 4: questions

We can, of course, form questions from our example sentence but, depending on the nature of the constituents, the questions and answers will differ:


Test 5: pro forms

Pro-forms include many pronouns but the one we select will depend on how we have understood the constituent parts of the sentence.  We can have, therefore:


Test 6: fronting

Certain types of phrases allow fronting (i.e., moving to the start of the sentence).  We can have, therefore:


Limitations and problems


Not all of the six tests can always be applied because English resists certain types of structure.  For example, using who in cleft sentences is not normally allowed (*Who the teacher shouted at was the boy in room 13).
However, if we can apply two or more of these six tests and come up with grammatically acceptable forms, then we have, almost certainly, successfully identified the real constituent phrases.


Constituent analysis works very neatly when the constituents know their places and keep to them.  For our example sentence, we can draw two pretty tree diagrams to show the interpretations we can put on the clause.  Like this:

constituents interpretation 1 constituents interpretation 2

On the left, you'll note, the Object noun phrase is a single constituent and on the right, it is made up of two independent constituents.
We can refine the analysis by going down a step and analysing the Verb phrase as Verb + Past marker + Particle and analysing the object noun phrases more carefully as Determiner + Noun etc. but for our purposes here, we won't.
Unfortunately, constituents don't always know their places and move around as discrete units.  Take for example:

  1. Did the teacher shout at the boy in room 13?
  2. The teacher woke the boy in room 13 up
  3. The teacher woke the boy up in room 13

In these case, the verb phrases have become separated by the noun phrases and the tree diagrams now look very messy if we try to draw them so we are reduced to having something like:


because traditional branching tree diagrams just won't work.

For more concerning this and other objections to constituent analysis, see the guide to Chomsky and transformational functional grammar.



In spoken English, constituents can often be identified by the fact that speakers tend, unless they are speaking very quickly, to insert a slight pause between phrases they perceive as constituents and to create a nuclear stress on each sense or tone unit.
We can compare our two original interpretations of The teacher shouted at the boy in room 13 and split the clause in two possible ways:

  1. || The teacher | shouted at | the boy in room 13 ||
  2. || The teacher | shouted at | the boy | in room 13 ||

In a., a speaker will tend to keep the rhythm of the sentence intact with the voice tone falling towards the end.
In this case, the sentence forms a single tone unit with the nucleus on boy and an unstressed tail, in room 13.
In b., by contrast, most speakers will insert a very slight pause between the boy and in room 13 and add an additional stress on 13.
In this case there are two tone units, the first with a nucleus on boy and the second with a nucleus on 13.
This is by no means easy for learners to hear without a good deal of practice (and is not consistent across all speakers) but it can be illustrated with something like:

a.  1
b.  2
In the case of b., there is a stress on 13 but that is absent in case a.
For more, see the guide to sentence stress.


Teaching implications

The first thing to note is that languages will have different ways of ordering the constituents of sentences and will even have different tests for identifying constituents so it's not the case that learners can simply transfer their intuitions from their first languages to English.  That's one reason we get errors such as:
    *In the room with energy worked the people
    *The working in the factory people
    *I like reading and play tennis
    *He looked the children after

The ability to recognise what is and is not a constituent phrase allows learners to develop a sense for what can qualify as the subjects, verbs and objects in sentences and set up mental templates for how acceptable sentences are constructed.  This may help them avoid, e.g.:

Perversely (to some), one simple way of doing this and getting your learners to recognise what the constituents of a clause actually are is to use precisely the sorts of messy diagrams we saw above.  For example, using:
example constituents
can alert people visually to the fact that the verb-phrase constituent comes in two parts and the object noun-phrase constituent is not just the boy but the boy in room 13.
Contrasting that diagram with:
example constituents
can alert people to the fact that now the prepositional phrase is an independent constituent which has a separate meaning (and can be moved around to give In room 13, the teacher woke the boy up).
The obvious follow-up exercise is to get your learners to draw their own block diagrams to show which bits of the clause do what.  How complex and interesting these are is dependent on the level of the learners, of course, but a few suitable candidates are:

The sentence stress issue mentioned above is probably not something with which you will wish to trouble learners at lower levels but with more advanced learners, the ability to hear the slight pause gives valuable information concerning the speaker's understanding of what is and is not a constituent of the clause.
For more, see the guide to sentence stress.

The final important issue is that doing a little constituent analysis on the target language we have in mind to teach will often pay dividends and prevent us from falling into some obvious traps and actually inducing errors.

Related guides
cleft sentences explaining how we get from, e.g., She liked the hotel to What she liked was the hotel
sentence stress for more on how constituents are stressed
phrases for a general guide to phrase structures
intonation this guide considers how intonation may be used to disambiguate the constituents of clauses
Chomsky for some consideration of objections to this kind of analysis