2

I am always curious about it-cleft sentences. How do you draw trees using binary branches and how is it different from normal structures: "It is .... that...."

Is the difference interpreted in the semantics rather than the tree? Or it could be explained in the tree?

Compare these two sentences:

(1) It was he that saw Mary. (not she)

(2) It was him that saw Mary. (not her)

How do we analyze the difference between these two sentences then?

This is a picture from Wikipedia

https://en.wikipedia.org/wiki/Cleft_sentence

enter image description here

Improve this question
23
  • 2
    Xia and @AlexB. There's a substantial problem with the tree. The problem is that according the tree, "John that Mary saw" is a constituent. However, it isn't "John" and the relative clause are two completely different constituents (part of the reason for considering the structure 'cleft' in two). This is quite different therefore from when the relative clause is a modifier within the DP/NP.
    – Araucaria - him
    Commented Dec 4, 2022 at 21:15
  • 1
    Xia, there are lots of ways of doing syntax, hence my question, which theory of syntax did you have in mind? That is why you don’t understand what BillJ is saying here, because he likes CGEL a lot and not Adger or Pesetsky. In Government and Binding we have the case filter, in the MP it’s all about feature checking.
    – Alex B.
    Commented Dec 4, 2022 at 22:15
  • 1
    @Araucaria-him you're right of course, I didn't look at the tree carefully enough. John that Mary saw is a not a constituent, so it should be slightly revised. Thanks!
    – Alex B.
    Commented Dec 4, 2022 at 23:06
  • 1
    Interesting exchange and commentary here! The tree is indeed bad. The string John that Mary saw is shown as a constituent, which it clearly shouldn't be based on various tests for constituents (proform substitution, answer fragments, clefting itself). The problem, though, is one that afflicts much of mainstream syntactic theory, namely all those theories that assume strict binarity of branching. N-ary branching is necessary if one intends to produce a plausible rendition of the constituent structure of cleft sentences and many other types of sentences.
    – Tim Osborne
    Commented Dec 5, 2022 at 5:18
  • 1
    @AlexB. I am thinking about mainstream syntax "GB" but also I am open to other fields as well. Yes I found that "John that Mary saw" is not a constituent.
    – Ellie Xia
    Commented Dec 5, 2022 at 15:30

2 Answers 2

Reset to default
0

enter image description here

For the sake of simplicity, I have treated "that" as a relative pronoun. Strictly speaking, it's actually a subordinator.

Improve this answer
7
  • 2
    The answer is terribly brief; it's simply a tree with no explanation. And its correctness is debatable because it shows was John as a constituent. What evidence is there that was John should be viewed as a constituent? Tests for constituents do not identify was John as a constituent.
    – Tim Osborne
    Commented Dec 5, 2022 at 14:56
  • Thanks, @BillJ although I am not quite familiar with the framework you have provided. It strikes me very much that there are two clauses (one is head clause and the other is the dependent clause.) I have found it very different and I am curious about the relatively flat structures compared to other frameworks, i.e., Government and Binding Theory.
    – Ellie Xia
    Commented Dec 5, 2022 at 15:40
  • 2
    What evidence is there was John is a constituent? That string cannot be questioned, it cannot be elided, it cannot be replaced by a proform substitute. Your tree is still too layered. It needs to be yet flatter in such a manner that John and that Mary saw are constituents, but not John that Mary saw and not was John.
    – Tim Osborne
    Commented Dec 5, 2022 at 15:56
  • @TimOsborne In any clause, the subject and predicate are the immediate constituents. I don't see any reason why the head clause here should be any different.
    – BillJ
    Commented Dec 5, 2022 at 16:00
  • 1
    The predicate of an it-cleft sentence likely does not include the focus, but rather the focus is most certainly the/an argument of the predicate. In a sentence such as It is John that is happy, the property of being happy is being predicated of John, making John "the subject" in the relevant sense.
    – Tim Osborne
    Commented Dec 6, 2022 at 1:31
1

In the MP, omitting some irrelevant details, one way of doing it is this (based on Belletti 2009 and Haegeman et al. 2015):

![enter image description here

see section 2.1. The Middlefield analysis for further details in Haegeman, Liliane, André Meinunger, and Aleksandra Vercauteren, 'The Syntax of It-clefts and the Left Periphery of the Clause', Beyond Functional Sequence: The Cartography of Syntactic Structures, Volume 10, Oxford Studies in Comparative Syntax (New York, 2015)

If you're familiar with the basics of the MP, it should be pretty clear to you (e.g. vP, strikethrough for deleted copies etc.), nothing unusual there. It is still binary branching - consistent with the theory - and it captures the constituency tests pretty well too. I tried to use the standard MP notation found e.g. in Adger 2003

Improve this answer
12
  • The answer stays within the parameters set out in the question insofar as strict binarity of branching should be maintained. In this regard, it is good. The answer also claims, however, that the analysis "captures the constituency tests pretty well". I disagree. It actually does worse in this area than the Wikipedia tree in the question.
    – Tim Osborne
    Commented Dec 6, 2022 at 2:44
  • The analysis has the same problem as that Wikipedia tree: it shows John that Mary saw as a constituent. Furthermore, it also shows It was as a constituent. Neither of those strings is verifiable as a constituent by diagnostics for sentence structure. The analysis has to allow n-ary branching if it is going to capture what tests for constituents reveal. An analysis is needed that shows John and that Mary saw as constituents, and It was, was John, and John that Mary saw as non-constituents.
    – Tim Osborne
    Commented Dec 6, 2022 at 2:45
  • A phrase structure analysis that accomplishes the job is as follows: [It [was [John] [that Mary saw]]]. Note, though, that even this analysis has a weakness: it shows was John that Mary saw as a constituent despite the fact that there is no evidence supporting that string as a constituent (Note that BillJ's tree does not have this weakness). In my view, then, the best phrase structure analysis of the sentence would be as follows: [[it] [was] [John] [that Mary saw]]. The structure is indeed hence quite flat.
    – Tim Osborne
    Commented Dec 6, 2022 at 2:48
  • 1
    Thanks for your time @AlexB this tree is much easier to understand compared to the former complex tree. It is interesting to see that "John" is moved from the complement of "saw" to spec FocP.
    – Ellie Xia
    Commented Dec 6, 2022 at 3:22
  • 1
    @BillJ I’m afraid you didn’t understand the notation here, was John that Mary saw is not a constituent.
    – Alex B.
    Commented Dec 6, 2022 at 14:30

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.