Download a copy of the final exam study guide (PDF).
Things to Know for the Final Exam (24.902)
Disclaimer: This list is representative, but not necessarily exhaustive.
I. Basic Stuff [Lec #2-3]
- Principles vs. parameters, Innate vs. learned.
- Basic arguments for the existence of phrases and movement, e.g. sentence fragment test, pronominalization.
- X-bar theory: head, complement, adjunct/modifier, specifier.
- Correlation between attachment site of modifiers and their scope (English alleged baron vs. alleged English baron).
- Ambiguities: possible left- vs. right-attachment of modifiers (priceless broken vase vs. broken priceless vase etc.); complement vs. modifier (one and do-so pronominalization tests; ambiguity of Mary jumped on the table in English, how Dutch disambiguates [from problem set], etc.)
II. Subcategorization [Lec #4]
- Locality condition on subcategorization: head of sister. [Know what Chomsky said in 1965 and how we revised it.
- Using the locality condition to diagnose head of CP and IP.
- Tip: Be sure you can distinguish preposition for from complementizer for. [Lec #5]
- Japanese: the uniform head-final property gives us a cross-check on our use of the locality restriction on subcategorization to identify heads.
III. Arguments for Head Movement and Phrasal Movement
- Scrambling in Japanese: numeral quantifier stranding argues that it is movement.
- V-to-I movement in French: separation of V from complement by adverbs etc. argues for movement. [Lec #6]
- V-to-I movement in Irish: separation V from complement by the subject argues for movement.
- Tip: Irish also argued that the external argument is first merged as specifier of VP. The reading from the Carnie textbook was important here.
- Tip: Pay attention to the discussion of specifiers in section 2 of this summary, bearing in mind that when we eventually got to Bulgarian multiple questions right before Thanksgiving we saw a C that takes multiple specifiers.
- An argument that the position of the verb in German verb-second is due to head movement: stranding the separable prefix.
- An argument that the position to which the verb moves in German verb-second is C: complementary distribution with the complementizer.
- An argument that the element in front of the verb has moved to Spec, CP: you only get one!
IV. English Auxiliary System [Lec #7]
- Necessity of a template: modal>have>be>V.
- Subcategorization of each verb for the morphology of the next verb down.
- Main verbs do not move to I, but highest auxiliary verb does.
V. Case Theory [Lec #8]
- Case morphology: nominative, accusative
- Tip: Know what ergative/absolutive languages are. Bonus: be able to name a few such languages.
- Case assigners (V, P) vs. non-case assigners (N, A)
- Phrases that need case (NP) vs. phrases that do not need case (others).
- Adjacemcy condition on case assignment
- Minimal c-command condition on case assignment Tip: Be able to produce a working case assignment rule for nominative or accusative case.
- ECM (Exceptional Case Marking), including case assignment by complementizer for.
- Small Clauses! Tip: We didn't do this in class too much, but learn it anyway.
- From PS #5: The argument from numeral quantifier stranding that passive sentences in Japanese involve movement of the subcategorized complement to subject position.
- The analysis of passive in English. Know what distinguishes a passive form of a verb from its active counterpart.
- This is where we introduced the notion EPP (for I) for the first time.
- PS #6: Quirky case in Icelandic is a morphological form of NP required in certain instances by the verb that assigns it its θ-role. The NP moves around just as if it were in search of nominative or accusative case, however. This argues that quirky case in this language is just a kind of morphological "paint" that obscures the fact that the NP needs "abstract" nominative or accusative just like an NP in English.
VII θ-roles [Lec #9]
- Semantics of Agent, Causer, Patient, Recipient, Goal, Experiencer, Theme
- Correlation of θ-role with syntactic position. Tip: Learn this. Later on in the class, we called this the Uniformity of Theta Assignment Hypothesis (UTAH), using terminology of Mark Baker.
- Internal vs. external argument
VIII. Raising vs. Control. ECM
- Existence of verbs with no external argument and a CP or IP internal argument. [In the next section, we call these a species of unaccuative.] Arguments that this is so.
- Raising to Subject (a.k.a. "Raising")
- Similarity between long-distance passive (Mary was believed to speak French.) and Raising (Mary seems to speak French).
- Tip: Review tests for Raising, e.g. idioms, dummy it. Do not merely apply them blindly. Make sure you understand what they show, which is simple: the θ-role of the subject of a predicate like seems is actually being assigned by the lower verb.
- Contrast with Subject Control, which fails tests for Raising.
- PRO. NB: The term "Subject control" refers to control of PRO by the subject of the higher clause.
- Object Control: Control of PRO by the object of the higher clause. Just as Raising is Subject Control's evil twin, so ECM is Object Control's evil twin.
- Tip: Make sure you can apart ECM vs. Object Control, Raising vs. Subject Control.
- Tip: It wouldn't hurt to just learn that seem is Raising, try is Subject Control, believe is ECM, and persuade is Object Control. For one thing, these are poster children for their respective constructions. For another thing, you can always remind yourself what one of these verbs looks like, if you have to figure out the status of some other verb.
- Other types of control: control from the object of P, partial control, etc.
IX. Unaccusativity [Lec #10-11]
- Unaccusative: be able to define.
- Why if UTAH is correct, unaccusative verbs definitely exist.
- Impossibility of passive. The closet was slid into by... etc. The impossibility of taking a subject theta-role away from a verb that doesn't have one in the first place entails the impossibility of passivizing unaccusative verbs.
- Tip: Know an unaccusative verb when you see one, especially if it's easy — e.g. drop in The ball dropped.
- PS #8: Test for unaccusativity. 1. Japanese numeral quantifier float; 2. Crude British Isles English fuck-all. 3. Russian genitive of negation [Neg may assign case to an object DP of an unaccusative or passive verb and permit the DP to remain in object position]. Also, from class: Italian ne cliticization.
- Burzio's Generalization: No subject theta-role —> no object case. Applies to unaccusative/raising verbs and to passive verbs.
- Negative Tip: I didn't stress the interesting material from Beth Levin's work on the semantics of unaccusativity. It's very important, but won't be on the final.
X. DP Hypothesis [Lec #12-14]
- The structure of English possessives, and how they fare under the DP hypothesis.
- How to reformulate Case theory under the DP hypothesis.
XI. Binding Theory
- The notion bind. [be able to define it: c-command plus coindexation]
- Basic binding theory: the data and Binding Conditions A, B and C. Be able to produce data that show these conditions at work.
- The notion "Binding domain" as we finally fixed it in stone on the last page of the summary.
- Problems posed in English by Mary likes her friends and similar examples, and how a discussion of PRO as Spec,DP bears on this question.
- Tip: Don't forget why nominative anaphors are bad, and why nominative pronouns may be bound from the next clause up. Many people forgot about this on the problem set!
- PS #9 part 2. Binding Theory may evaluate a DP in a position it occupies before movement/
- PS #10: Binding Theory teaches us that the first object in a double-object construction asymmetrically c-commands the second object. [Interestingly, get comes out as an unaccusative of give.]
- PS #10: How Binding Theory might need to be rethought in light of the disjoint reference/overlapping reference facts.
- Evidence for wh-movement from gaps, stranded all in Irish English and interactions with Binding Theory
- Analysis of questions in English: movement to Spec,CP; Doubly-filled COMP filter, pied piping.
- Important Tip: Learn the analysis of relative clauses in English. Why there are three types (introduced by an overt wh-phrase, by that, and by ø).
XIII. The Model
- How movement and agreement works: uninterpretable features, probe, goal, Agree, EPP, move
- How the model works: Merge, Move
- An older model: Deep structure (D-structure), Surface Structure (S-structure)
- Attract closest, Shortest move as evidenced in Bulgarian multiple questions
XIV. Successive Cyclicity & Islands
- Evidence that wh-movement may land in Spec, CP of declarative as well as interrogative clauses and may move to its final landing place in multiple steps (successive cyclicity). Evidence for this from all-stranding in W.Ulster English and from binding phenomena
- Complex NP Condition, WH-island condition: how these might derive from Subjacency.
- How subjacency works. The notion Bounding Node. Tip: Know why successive cyclicity is important, if subjacency turns out to be correct.
- Condition on Extraction Domains (CED): how it shows up with Incorporation as well as wh-movement.
- Coordinate Structure Constraint (CSC)
- Tip: Be able to identify these conditions when violated, and produce examples that exemplify the violation.
XV. Covert Movement
- What it is: pronounce the trace rather than the higher position, after movement.
- Covert wh-movement in Japanese: evidence from islands
- Covert vs. overt movement:
- Standard view: All overt Move before all covert Move. The level after all overt Move and before all covert Move was called Surface Structure (1980s: S-structure). Spell-out (interface with PF ("phonetic form" ) was at Surface Structure. The final representation, after all covert movement, is called LF (for "Logical Form").
- Other possibilities: Overt and covert move interspersed. Spell-out either derivation-final or interspersed with Move and Merge in some fashion.
- Does English have covert movement:
- Obedience of wide-scope quantifiers to islands suggests covert movement as a means of scope-assignment.
- Binding facts suggest that in multiple questions, after overt wh-movement, other wh-hrases move much as in Bulgarian — only covertly!