What is the difference in a constituent and a phrase?

Question

From Wikipedia:

In syntactic analysis, a constituent is a word or a group of words that functions as a single unit within a hierarchical structure.

A phrase is a sequence of one or more words (in some theories two or more) built around a head lexical item and working as a unit within a sentence.

There doesn't seem to be much of difference between them. Could someone please help clarify (using examples, possibly using phrase markers) what the difference between them is?

Thanks.

Answer 1

A constituent is easly identified by a series of possible tests. Among them, the replacement of a phrase for a pronoun, thus working as a single unit, for instance:

• The good man drinks the water

It could be:

• He drinks it

It means that both "the man" and "the water" are constituents. We can also keep the "the" an change the nouns following it, meaning that the NP inside the DP is also a constituent:

• The child drinks the soda

But notice that if we try to replace the whole VP for something like a pronoun or breaks:

• *The man it

Of course it wouldn't work, so another way of testing it is using "did so"

• The man did so

So, as you can see, there is not MUCH a difference, we use a constituent test to help us build the trees, to test what goes with what, and a phrase is the stucture itself. When we realise that certain words are indeed constituents we know that it forms a phrase. I don't know if it was clear, but if you have any questions, please comment down here :D

Answer 2

Contrary to the suggestive nature of the question, the answer already produced by Ergative Man, and a couple of the comments, there is an important difference between the constituent and the phrase. A phrase is a sub-type of constituent, which means every phrase is a constituent but many constituents are not phrases.

For orientation, I think many syntacticians would agree with the following definition of the constituent unit when considering tree analyses:

Constituent: A node N plus everything that N dominates.

With this definition in mind, consider the next tree analysis, which I have taken from the Wikipedia article on the constituent unit cited in the question:

While one can certainly disagree with the structural analysis shown here, its validity is taken for granted for the sake of illustration. According to the definition above and this tree, there are nine distinct constituents present:

1. Drunks,

2. could,

3. put,

4. off,

5. the,

6. customers,

7. the customers,

8. put off the customers, and

9. could put of the customers.

But according to the labels given in the tree, there are only three distinct phrases present:

1. the customers,

2. put off the customers, and

3. could put off the customers.

The difference, then, lies in particular with what one considers a phrase to be. I think many syntacticians would disagree with the tree analysis above concerning the subject Drunks; they would view it as a phrase because it could take dependents but simply does not do so in this particular case. Few syntacticians would ever choose to view a particle like off or the definite article the as phrases, though, because these words rarely if ever take dependents.

Consider the following tree analysis of the same sentence next; it is the dependency grammar (DG) analysis and is also taken from the Wikipedia article that the question cites:

Applying the definition of the constituent above to this analysis, the following words and word combinations are constituents; there are just five of them:

1. Drunks,

2. off,

3. the,

4. the customers, and

5. put off the customers.

While the labels used do not identify phrases, I think some DG people would define the phrase as two or more words that form a complete subtree. On this definition, the tree shows just two phrases:

1. the customers, and

2. put off the customers.

The examples just discussed illustrate that a given phrase is always a constituent, but many constituents are not phrases. The constituent is a more inclusive unit than the phrase. I think most grammarians would agree about this. When one looks further, though, to which constituents should be construed as phrases, opinions vary, sometimes drastically so.

Answer 3

While a constituent is any proper subpart of a sentence (a morpheme, a word, a phrase, or even a clause), a phrase is typically a sequence of words built around a word class (the head) and existing as a unit of structure and function in a sentence. A constituent may or may not be a phrase, but a phrase, if it indeed is a phrase, is always a constituent. Here, it should be pointed out that the same sequence of words may or may not function as a constituent, depending upon the sentence the sequence is in.

Ex. Although he exercised daily, old Sam was in poor shape. (The sequence of words "old Sam" is a constituent)

Ex. Though he was old, Sam exercised daily. ("old Sam" is not a constituent)