Preiously... Uniformity of Theta- Assignment Hypothesis, daughter of P = Agent, daughter of P = Theme PP, daughter of! = Goal P called P Chris P P books gae P to PP Chris Unaccusaties Transit i e s The ice, the boat, the door, all Themes: daughter of P. The ice melted. P The boat sank. The door closed. P melt Unaccusaties hae a relatiely inert, no causal meaning. There are two kinds of, the causal one that needs an (Agent), and a non-causal one. What if we pick the causal (and proide an Agent )? the ice Bill P melt P the ice Bill melted the ice. The causal adds an Agent. Bill was the agent/instigator of a melting that affected the ice. Unergaties Double object constructions Bill lied. That s got an Agent, and Agents must be daughter of. Bill P P lie gae a book to Chris. Agent: ; Theme: a book; Goal: to Chris gae Chris a book. Agent:, Theme:? a book?, Goal:? Chris? So, it would look like this. Don t these mean the same thing?
gae Chris a book, daughter of P = Agent PP, daughter of! = Goal, daughter of P = Theme The word order suggests this structure. UTAH (so far) doesn t tell us what theta role a book gets. And in what sense is Chris a Theme of a going? P P? Chris go gae abook Two kinds of giing The two forms of gie are not quite equialent, though: 1) gae a book to Chris. 2) gae Chris a book. 3) * gae a headache to Chris. 4) gae Chris a headache. Try paraphrasing 5) sent a letter to Chicago. 6) * sent Chicago a letter. 7) taught French to the students. 8) taught the students French., daughter of P = Agent, daughter of P = Theme PP, daughter of! = Goal, daughter of! = Possessee P P has abook To hae P P Chris hae gae abook On beyond Our trees hae now expanded beyond being mere Ps to being Ps. The Hierarchy of Projections: > Once you hae finished the P (uninterpretable selection fures are checked), if there s a on the, Merge it. The UTAH:, daughter of P: Agent, daughter of P: Theme PP, daughter of!: Goal, daughter of!: Possessee But this is only the beginning. 9 CAS LX 522 Syntax I TP, Agree, and our quickly growing tree (5.1-5.3) Auxiliaries and modals and erbs I ate. I could. I had en. I was ing. I had been ing. I could hae en. I could be ing. I could hae been ing. So: could, hae, be,. How do we determine what form each erb takes?
Auxiliaries and modals and erbs Auxiliaries and modals and erbs Hae: Perfectie (aspect) I hae en. I had en. Be: Progressie (aspect) I am ing. I was ing. Could: Modal I can. I could. I shall. I should. I may. I. I will. I would. I could hae been ing. *I could be haing en. *I was canning hae en. *I had cannen be ing. *I was haing cannen. *I had been canning. It looks like there s an order: Modal, Perf, Prog, erb. Auxiliaries and modals and erbs Suppose: Hae is of category Perf. Be is of category Prog. May,, can, could are of category M. They are heads from the lexicon, we will Merge them into the tree aboe P. Their order is captured by a new extended Hierarchy of Projections: Modal > Perf > Prog > > Except not eery sentence has these. So: (Modal) > (Perf) > (Prog) > > Negation Consider the following: I did not. I could not. I had not en. I was not ing. I had not been ing. I could not hae been ing. Suppose not is of category Neg. How do we describe where not occurs? How can we fit it into our Hierarchy of Projections? Where does Neg fit? Suppose that we can fit Neg in our Hierarchy of Projections. Just like the other things we just added. (Modal) > (Perf) > (Prog) > > Where would it go in the HoP, and how can we explain the word order patterns? I could not hae been ing. I had not been ing. I was not ing. I did not. Remember and how we explained where the erb is in gae a book to Chris? A-ha. Picture this: I?+ not <> hae been ing. I?+had not <had> been ing. I?+was not <was> ing. So what is?, then? He did not. He ate. He does not. He s. All that do seems to be doing there is proiding an indication of tense.
HoP reisited So, now we know where Neg goes. Aboe all the other things, but below tense (category T). T > (Neg) > (M) > (Perf) > (Prog) > > Just as moes to, so do Perf, Prog, and M moe to T. If Neg is there, you can see it happen. They T+shall not <shall> be ing. They T+shall <shall> be ing. What does do do? But what about when there s just a erb and Neg, but no M, Perf, or Prog? I ate. I did not. Eat clearly does not moe to T. But not gets in the way, so tense cannot see the erb. Instead, the meaningless erb do is pronounced, to support tense. Do-support We will return to the details in due course So, we hae T We e just added a category T, tense. The idea: The tense of a clause (past, present) is the information that T brings to the structure. T has fures like [T, past] or [T, pres] Or perhaps [T, past] or [T, nonpast]. These fures are interpretable on T. T is where tense lies. We see reflections of these tense fures on erbs (gie, gae, go, went) but they are just reflections. Agreement. The interpretable tense fures don t lie on erbs, they lie on T. [N, ] [, un, ] [N, ] We already know how this is supposed to work, to a point.. [, un, ] Merge and, checking the un fure of (and assigning a θ- role to, namely Theme this is daughter of P). [, un,... ]. [N, ] [, un, ] P [,...] [N, ] [, un, ] P [,...]. P P [,... ] We already know how this is supposed to work, to a point. Merge and, checking the un fure of (and assigning a θ- role to, namely Theme this is daughter of P). [, un,... ] We already know how this is supposed to work, to a point. Merge and, checking the un fure of (and assigning a θ- role to, namely Theme this is daughter of P).
[N, ] [, un, ] P [,...].. [N, ] P [, un, ] P Merge and the P, in conformance with the Hierarchy of Projections. projects, and still has a un fure. [, un,... ] P [,... ] Merge and the P, in conformance with the Hierarchy of Projections. projects, and still has a un fure. [, un,... ] P [N, ]. P [, un, ] [N, ]. P [, un, ] Merge and the P, in conformance with the Hierarchy of Projections. projects, and still has a un fure. P [, un,... ] P Moe the up to. P [, un,... ] + P < > [N, ]. P [, un, ] Merge with ʹ to check the un fure and assign a θ-role (Agent, this is daughter of P). P [, un,... ] + P < > P [, ]. Merge with ʹ to check the un fure and assign a θ-role (Agent, this is daughter of P). P [, un,... ] + P < >
.. P [, ] So, now what do we do with? 1) And shall. 2) What should do is. It kind of seems like it goes between the subject and the erb, but how? P [, un,... ] + P < > If we leae eerything as it is so far (UTAH, Hierarchy of Projections), the only option is to Merge with the P we just built. So, let s. P [, ] M [M,... ] P [,...] + P < >.. TP T [T, past] MP [M, ] Now, we hae one more thing on our (T) and the HoP says that once we finish with M, we Merge it with T. And so Merge T, we shall. MP M P + P < > Then, M moes up to T. Why? Because M, Perf, and Prog all moe up to T. For the same kind of reason that moes up to. Right now we hae no way to describe this in our system, except with this rule from the outside that stipulates that moes to, and {M/Perf/Prog} moes to T. T MP [past] M P + P < > Ok, that s all fine and good, except that the sentence is not Might How do we get out of this?. TP T+M MP < > P + P < > As preiewed earlier, the subject moes to this first position in the sentence, around the modal. Moing here means Merging a copy. TP T T+M MP < > P < > + P < >
. TP WARNING Gr. Why? Jumping ahead, we re going to say that this is a property of T- type things generally: T needs to hae an in its specifier. We can encode this as a (special type of) uninterpretable fure on T: [un*]. More on that later. T T+M MP < > P < > + P < > What we e done here is not quite the same as what is in the textbook. (But it s better). In the textbook, modals are not tred as their own category, but rather as a kind of T. The reision we made here will pay off soon. Keep this difference in mind as you reiew the textbook on this point. You will see no MPs in the book. But you should see them on the homeworks/tests you turn in. Side note: I s. T You may hae heard in the past that it tense should be of category I (for Inflection), rather than T (For Tense). Rest easy: T and I are (for current purposes) just two names for the same thing. Historically, this was called INFL, then I, and now usually called T. But these are just names. I s. T, Istanbul s. Constantinople; St. Petersburg s. Leningrad. ate Now that we hae T in the Hierarchy of Projections, we re stuck with it. Yet, where is T in ate or s? It looks like the tense marking is on the erb, we don t see anything between the subject and the erb where T ought to be. Now that we hae T, this is where tense fures belong. We take this to be the thing that determines the tense of the sentence, een if we sometimes see the marking on the erb. ate Since (most) erbs sound different when in the past and in the present tense, we suppose that there is a [past] or [present] fure on the erb. Howeer, to reiterate: tense belongs on T. The tense fures on the erbs are uninterpretable. Fure classes There are tense fures. Like past, like present. There are case fures. Like nom, like acc. There are person fures. Like 1st, like 2nd. There are gender fures. Like masculine, like feminine. So, we can think of this as a fure category or fure type that has a alue. [Gender: masculine]!!! [Person: 1st] [Tense: past]!!!!! [Case: nom]
Agree T nodes hae fures of the tense type. Maybe past, maybe present. Suppose that has an uninterpretable fure of the tense type, but unalued. What we re trying to model here is agreement. Agree In the configuration X[F: al] Y[uF: ] F checks and alues uf, resulting in X[F: al] Y[uF: al]. Unalued fures The idea is that a lexical item hae an unalued fure, which is uninterpretable as it stands and needs to be gien a alue in order to be interpretable. The statement of Agree on the preious slide is essentially saying just that, formally. This gies us two kinds of uninterpretable fures (unalued and regular-old uninterpretable fures), and two ways to check them (aluing for unalued fures, checking under sisterhood for the other kind). Unalued [uf: ]. Regular-old [uf]. To be continued... (Actually, it s unlikely we ll get to here anyway, so I suppose it will be continued immediately after you actually see the preious slide, next time.)