The course I'm taking is based on Radford's Analysing English Sentences (2009). I'm having a lot of trouble analysing sentences within the split VP framework. There are very few examples in the book (and on the Internet). I'm not sure I'm doing right. I decided to start with very basic sentences. How would you analyse the following sentences?

Links bring to images.

  1. He arrived (unaccusative)
    It should be "He arrived home" but I don't really know where to put "home" in this tree.

  2. She works hard. (unergative)

  3. The ship sank. (ergative)

  4. They sank the ship. (ergative)

The definition of "ergative" verbs in my book reads: This term originally applied to languages like Basque in which the complement of a transitive verb and the subject of an intransitive verb are assigned the same morphological case. However, by extension, it has come to be used to denote verbs like "break" which occur in structures like "Someone broke the window" and in structures like "The window broke", where "the window" seems to play the same semantic/thematic role in both types of sentences, in spite of being the complement of "broke" in one sentence and the subject of "broke" in the other.

  • 1
    The ship sank is surely not ergative?
    – Colin Fine
    Dec 2, 2012 at 15:39
  • I think your examples are mixed up. The ship sank is unaccusative. Dec 2, 2012 at 16:27
  • These examples make no sense anyway, since "unaccusative", "ergative", etc. refer to verbs and not to constructions. (That is, in generative tradition. In typology and functionalism, "ergative" means something different altogether)
    – Fryie
    Dec 2, 2012 at 18:22
  • 1
    No, Veronica is correct, in a way. While some (most?) generativists talk about unaccusative and unergative verbs, others -like Haegeman and Radford- talk about unaccusative and ergative verbs. Anyways, Radford does have trees in the book you mention. However, I personally prefer David Adger's textbook.
    – Alex B.
    Dec 2, 2012 at 18:37
  • 2
    @AlexB. It's only some generativists who talk about unergative and unaccusative, and AFAICT none of them mean the same thing by the terms. The distinction between syntactic and lexical status is one reason; another is that the choice of terminology is really, really unfortunate. Using a term like unergative when talking to people who are at minimum very unclear about what ergative means in the first place, then telling them to imagine the opposite, is just not helpful.
    – jlawler
    Dec 2, 2012 at 19:36

4 Answers 4


As people have already pointed out in the comments there are some confusions in the question. The problem is that ergative and accusative, which are normally used to talk about case/agreement patterns across transitive and intransitive verbs, have been extended to talk about contrasts across different types of intransitive verbs, namely unergatives and unaccusatives.

A language manifests ergativity if the subject of intransitive verbs behaves like the object of transitive verbs to the exclusion of the subject of transitive verbs. Languages can manifest ergativity in their case, their agreement, their syntax, etc. Mayan languages, for instance, have ergative agreement.


X    *e*     ki    tz'ët
ASP  *O3p*   S3p   see
*They* saw them.

X    *e'*    el
ASP  *S3p*   left
*They* left.

Note above that the object agreement in the transitive verb is the same as the same as the subject agreement in the intransitive verb, namely e. This makes it ergative. In contrast, look at the English pronouns. The two third person subject pronouns are the same, but the object pronoun is different. This means that case in English is patterning accusatively.

Where does unaccusative and unergative come from then? Well it was noticed that some subjects of intransitive verbs seem more agent-y while other subjects of intransitive verbs seem more patient-y. Moreover, in some languages, there are clear syntactic and morphosyntactic differences between these two classes of intransitive predicates. The logic behind the names goes something like this: Well, the agent-y subjects of certain intransitive verbs are getting treated special. That reminds me of how agents of transitive verbs in ergative languages are treated special. Let's call those special intransitive subject unergative. In contrast, the patient-y subjects of those other intransitive verbs are being treated special. That reminds me of how patients of transitive verbs in accusative languages are treated special. Let's call those guys unaccusative.

In English its hard to talk about unaccusativity and unegativity because we don't have the best diagnostics. Other languages, do though. A classic case is perfect forms of intransitives in Italian. Different intransitive verbs take different forms of the copula.

  • È arrivato.
    He has arrived

  • Ha telefonato.
    He has called.

When you look at more and more of these intransitives it looks like the agent-y, or unergative ones, use ha, while the patient-y, or unaccusative ones, use È.

As for drawing trees. People usually assume that it's the unaccusative subjects that are treated differently. In a derivational theory, they are assumed to start out as objects to the intransitive verb, which then move to subject position. One piece of evidence for this, looking back at Italian, is that the passive form of the transitive verbs use the È you see with unaccusatives. In derivational theories of the passive, the subject starts out in object position and only moves to subject position later.

È stato ucciso.
He was killed.

So long story short, your example (4) is a normal transitive clause. It has no bearing on unaccusativity or unergativity. You examples (1) and (3) are unaccusative, but to convince me you would need to give me some diagnostics. Once you have those, you could try to argue that the subjects of (1) and (3) are underlying objects. (2) might in fact be unergative. Once again, marshall some arguments. If it turns out to behave differently than other patient-y intransitive subejects, you might be able to argue that these subjects originate in subject position or something.

  • 2
    Why for godsake is it the "agent-y" ones that are unergative? Ergative means "agent-y"; unergative shouldn't mean the same thing. Non-compositional terminology is less than ideal.
    – jlawler
    Dec 3, 2012 at 17:10
  • I discussed this phenomenon a while back (though without the dubious help of unergativity) in Austronesian languages, Acehnese and Fijian. It's clear that a system like that works here, but there's no real evidence, as rmh points out, that it does in English, where there are only small fragments of ergativity.
    – jlawler
    Dec 3, 2012 at 17:18
  • Yeah, the terminology is very disturbing. I think the idea is something like. In unaccusatives, the thing that you might expect to get accusative case in a transitive is not (because the subject of an intransitive for gods sake). Since it's NOT getting accusative, it's unaccusative. For unergatives, the argument is the argument you might expect to get ergative case in a transitive clause, but it's not getting ergative case because it's an intransitive argument, so it's an unergative.
    – rmh
    Dec 3, 2012 at 17:18
  • 4
    Any theoretical termimology that confuses more people than it helps should be abandoned. I think of it as phlogisticated terminology.
    – jlawler
    Dec 3, 2012 at 23:57
  • I agree with that. However, from the question it seems clear that the topic wasn't "ergative languages" (in the typological sense), but some kind of however ill-defined notion about verb categories (maybe akin to what others would call "ambitransitives")
    – Fryie
    Dec 4, 2012 at 15:14


  • 1 intransitives

    • a) unergatives: run
    • b) ergatives

      • i unpaired ergatives (unaccusatives)
      • (subject is logical object. object raises to subject position): arrive
      • ii paired ergatives
      • (subject is logical object. object raises to subject position. also transitive).

      • a) Inchoatives (anticausatives) The doar opened.

      • b) Middles - (agent is implied, refers to inherited property) This window cleans easily

    2 transitives (includes pseudo-transitives, which can be used with and without internal argument)

    • I ate in the cafe. I ate an apple in the cafe.

unergatives have both vP projection and VP which consists of just the verb unaccusatives have just VP projection. They do not have vP projection Transitive verbs have vP and VP which consists of V and Object

  • So you have to be able to make a distinction between vP projection and VP projection in order to distinguish unergatives from unaccusatives? What are the tests for those?
    – jlawler
    Dec 4, 2012 at 19:44
  • That what we hav seen in syntactic theory graduate course.
    – Dariya
    Dec 5, 2012 at 10:57
  • 1
    Um, that's the Official Theory, of course, but Occam's Razor militates against positing undetectable Angels to dance on the Pin'. I.e, not everybody believes this is anywhere close to correct, and therefore it's kind of like a matter of Faith.
    – jlawler
    Dec 5, 2012 at 20:33
  • 1
    Sure thing. But that means you have to consult your teacher or mullah to get the correct theology; heretics abound (especially on the Web).
    – jlawler
    Dec 5, 2012 at 21:01
  • 1
    You are totally right. This analysis is accepted by my teachers but İ can't say anything about the author of the question.
    – Dariya
    Dec 5, 2012 at 21:18

Posting a picture from Adger 2003 now - will write a full answer a bit later.

enter image description here

  • 1
    There's a lot presupposed by that notation, whatever the explanation might be.
    – jlawler
    Dec 5, 2012 at 4:29

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.