I had a test and the question was to test the underlined part for constituency; however, I got confused because I used the test of deletion, but I found the sentence ungrammatical, so my answer was that the underlined part is not a constituent. My friends applied the pseudo-clefting test and found that the sentence is grammatically fine.

The sentence is:

My mother told me to pick up the dry cleaning on my way home.

The part in italic is what we were required to test.

This is my answer after I applied the deletion test:

*My mother told me - on my way home.

My friends' answer with pseudo-clefting applied:

What my mother told me is to pick up the dry cleaning.

So I want to ask the following:

  1. Could a specific structure or sting of words be a constituent when we fail to apply one of the tests on?
  2. Is my answer wrong?
  3. Are there specific tests for each structure depending on its type or category? For example can we always apply It-cleft to test the subject?

2 Answers 2


To quote Adger's syntax textbook:

It is important to note that when these tests are successful then what you have is an argument that a particular sequence forms a constituent, rather than proof that it does. It is even more important to notice that when one of these tests fails, we cannot conclude that the sequence is not a constituent, we can only conclude that there is no evidence that it is.

The problem is, there's no single test that's both necessary and sufficient to show that something is a constituent. For example, it's usually taken as an axiom that single words (or rather lexical items) are always constituents. But they can easily fail all sorts of tests:

  • The cat is hungry.
  • *The is hungry.
  • *Cat is what the is hungry.
  • *What the is hungry? Cat.
  • *The it is hungry.

(In particular, this shows that you can't generally move or delete an NP out of a DP, and pronouns replace DPs, not NPs. These are separate rules of the language.)

The advice I generally give is to apply several different constituency tests, rather than just one. Try at least one movement test and one substitution test and see what happens.

  • Hmm. Devil's advocate: You nearly always can delete the 'NP' in a 'DP'. You more or less only can't if the D is the, a, every (i.e. if it's an article). For all the very many other D's you can! Devil's devil's advocate: aren't pronouns intransitive D's? Commented Jun 6, 2022 at 16:41
  • @Araucaria-him For the first one, true, but I tend to think of articles as the most common determiners. For the second one—yes, but thus aren't they also DPs, since they don't have any unmet c-selections?
    – Draconis
    Commented Jun 6, 2022 at 16:47
  • 1
    Yes, I think every pro-X must be an XP too. Maybe by definition? [With 'pronoun' being a misnomer as they are really pro-D's]. Commented Jun 6, 2022 at 16:55
  • Only if you go to a certain type of church. The rest of us still believe that NPs exist and DPs are superstition.
    – jlawler
    Commented Jun 6, 2022 at 17:13
  • 2
    @jlawler Very true, but then there's the other type of church that thinks dependency is more important than constituency and so on and so forth. I don't want to go too in-depth on the options since it sounds like OP is a beginner.
    – Draconis
    Commented Jun 6, 2022 at 17:18

The test question (sentence) may be problematic. The string "My mother told me to pick up the dry cleaning on my way home" has two analyses, and two semantic interpretations. The interpretations are probably more obvious. One is that we don't know when or where you mother told you this, but that thing that you are supposed to do is "pick up the dry cleaning on (your) way home". The other is that we don't know when she wants you to pick it up, but it was when you were on your way home that she told you this. This may or may not relate to different constituencies (though I don't know of a theory that bothers to have a semantics, and that only allows one of the two syntactic structures). Generally speaking, constituency tests only work if the semantic interpretation is held constant.

Your answer is wrong by dint of the fact that you assert that "*My mother told me-on my way home" is ungrammatical. But it is grammatical, and you can use various substitutions to convince yourself of that ("My mother spoke to me on my way home; my mother saw me on my way home; a cat bit me on my way home"). I assume that the instructor had in mind only the interpretation that the picking up would take place when you are on your way home, and did no consider the interpretation that the telling was on the way home.

There are two constituencies to consider:

My mother told me [to pick up the dry cleaning on my way home]

My mother told me [to pick up the dry cleaning] on my way home

The output "What my mother told me is to pick up the dry cleaning" does not directly correspond to either, also I find that output to be a bit sketchy though for purposes of the exam one might claim that it is grammatical ("...told me to do..."). What you want is

What my mother told me (to do) on my way home is [to pick up the dry cleaning]

What my mother told me (to do) is [to pick up the dry cleaning on my way home]

If only one of these two options were grammatical (using both tests), you would have evidence for a single constituency. Instead, you have contradictions – evidence for two constituencies, i.e. syntactic ambiguity.

It is possible that the correct answer per the instructor is "there's a contradiction, therefore there are two constituencies". I guess you'll find out when you get the results.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.