Rederived correct generalizations
Yüklə 94,42 Kb.
|
- Bu səhifədəki naviqasiya:
- Table I: predictions about the distribution of stative passives
- References
|
Restitutive re- and the first phase syntax/semantics of the VP Alec Marantz, marantz@nyu.edu 0. Restitutive re- (1) re- prefixation creates the restitutive of the verb, not the repetitive. a. The walls in the house were green long before they purchased it; they plan to re-paint them (white) as soon as they can. b. The door of the cabinet was built open, and John closed it for the first time when he brought it home. Mary then re-opened the door. (2) As has been observed by many, re-Verb doesn’t refer to the repetition of the activity described by the whole VP but rather describes the re-occurrence of the state within the VP. That’s why re-, unlike “again,” is incompatible with simple activity predicates: a. *John re-smiled b. John smiled again.
c. This door was built open and hasn’t been touched since. d. I just closed the door and re-opened it. e. I.e., I opened it for the first time, but restored it to the open state it was built in. Lieber: “[re-] also does not attach to verbs which imply a result which cannot be reversed. For example, it is impossible to *reeat the apple…” f. “The other day i had an incredible stomach disorder which made me vomit and then in order to get over the problem, my doctor told me to re-eat the vomit while being beaten over the head with a frozen fish by a german bodybuider named Helga.” Note: “the vomit” hadn’t been eaten in the first place – eating restores it to its previous state. Lieber’s analysis fails because she treats re- prefixation as a lexical process that must affect the meaning “denoted by the verb,” rather than a construction at the vP level where the prefix targets a constituent below the verb. (3) a. “I used the carrot cake recipe which includes instant vanilla pudding. I have just realized that I used "Cook and Serve" vanilla pudding by mistake. This cake is for a groom's cake this weekend. Do I need to rebake the cake or will it really make any difference?” (Although here the agent of the activity is the same in both the baking and the re-baking (“I”), the meaning of re-bake is compatible with different agents – the crucial sentence could be, “Do I need to ask someone to rebake the cake [i.e., cause “the cake” – another token of the same type – to come into existence via baking]…?”) b. “I want to suggest, though, that one lens through which we can reasonably view Al Quie's approach to education these past eighty years is that of a person trying to apply the doctrine of subsidiarity to that multi-tiered marble cake. Putting it differently, Al has patiently and assiduously tried to re-bake the cake with neater layers and less marbling.” (wants to cause the cake to come into existence again, only better) d. re-build the 18th century cathedral e. re-create the current situation
b. I reopened the door, repainted the house, rearranged the furniture… c. ??I resmoked, relaughed, resang. d. ??I rekicked the wall, rereached the top,.. (5) Horn’s generalization: re- requires object (transitive object or underlying object of unaccusative) (6) As explained by von Stechow in a number of publications, the meaning of restitution depends on identifying the end state of a change of state: a. For apparent de-adjectival verbs and others which participate in inchoative alternations the end state is named by the verb root: i. The door re-opened, I opened the door. (end state: open) ii. The ice-cream re-melted. (end state: melted)
i. I re-built the house. (end state: a house) ii. I re-constructed the argument. (end state: an argument)
i. I re-arranged the furniture. (end state: arranged) ii. I re-painted the wall. (end state: painted)
state s is filled in by NP, in the case of verbs of creation, by the verbal root as modifier of event-1, in the case of verbs like paint by the verbal root as modifier of event-2, in the case of verbs like open (8) Although von Stechow describes a restitutive reading for again and achievement verbs such as “catch” (where the end state is “prisoner”), re- resists attachment to achievements in which the direct object is unaffected, preferring verbs for which the underlying object is the measure of the caused event: a. ??John re-reached the top of the mountain. b. ??John re-left (the room). (9) Two distributional issues: a. Too many environments. Why should re- attach in cases where the verb root names the restored state (reopen) and in cases where the direct object names the restored state (rebuild (e.g., the house)) and in cases where the causing activity names the restored state (repaint (e.g., painted)? b. Too few environments. Although restitutive again is fine with obligatory double complement constructions (double object constructions (Beck and Johnson), resultatives with non-thematic objects, put verbs), re- obeys the “Sole Complement Generalization” and won’t attach in these constructions. i. John entered the room for the very first time and saw a folder hanging dangerously over the edge of a table. The folder fell, and John put the folder on the table again/*re-put the folder on the table. ii. The first discussion group emptied the tea kettle but re-filled it before they left the room. The new discussants entered and drank the teapot dry again/*redrank the teapot dry. iii. John threw out his U of Stuttgart t-shirt, but Mary, seeing it in the trash and not knowing that it had belonged to John, washed it and handed John the t-shirt again/*re-handed John the t-shirt for his birthday.
b. The walls remained painted throughout the centuries. c. *They re-put the book on the table. d. *The book seemed put on the table. e. *They re-gave John the award. f. *John seemed given the award. (11) Approach: Simplest Syntax = Minimalist Program + Distributed Morphology a. Semantic interpretation interprets the syntax, phase by phase b. Phonological/morphological interpretation interprets the syntax, phase by phase c. Consistent semantic interpretation of structures in terms of causation or change of state does not by itself motivate a syntactic head for cause or become.
b. and restitutive again may appear with vPs that disallow re- c. the solution to (7) cannot be a simple semantic/syntactic scope issue with respect to the constituent to which again/re- attaches; i.e., re- should in principle parallel restitutive again from a purely semantic perspective.
e. Or, since re- is an affix on verbs, it may be prohibited from appearing in a position not sufficiently local or otherwise accessible to the v head f. If either d. or e. is the correct solution, then re- could provide evidence for the correct syntactic/semantic analysis of double complement constructions 1. vP structure in Simplest Syntax (13) vP structure: subevents are circled for emphasis a. Activity verbs, mono-eventive – eventive little v merged with root John hurried John jumped 
 
v jump b. Verbs with direct objects interpreted as change of state (incremental theme verbs, verbs of creation, most accomplishments) – built from (13a) by merging a subevent Claim: causative interpretation of relation between the subevents comes from the structure without a mediating head Claim: inchoative interpretation of the embedded subevent comes from the structure without a mediating head
build a house clean the apple
v the cake DP coerced into a change of state event   involving the cake as “measure” of the event
v bake
NOTE: activity verbs in (13a) can become incremental theme verbs as in (13b) via addition of a cognate object: John danced (a dance/a waltz) John sang (a song/an aria)
John baked me a nice cake/ John danced me a nice dance “Someone who will sing me to sleep. Dance me a waltz when i wake up. Longwalks at 5pm, hotchoco at 5am.” "Tell Me a Story, Dance Me a Dance" John re-baked the cake/ re-danced the dance the carefully baked cake/ nicely danced dance c. Verbs where the root names the end state of a change of state on the direct object – these easily allow transitive/anticausative alternations depending on whether or not the causing event is an agentive activity So, assume also a voice head that relates an agent to the activity little v, when there’s an agent open the door spin the wheel (eventive root) flatten the pancake (-en realizes little v as event head) freeze the ice cream  
 v
the door open DP coerced into a change of state event where the end state is named by the root d. Small clause structures with a complex inner event (built from (13a) by merging a small clause event as complement to v) send [the book to John] put [the book on the table] drink [the teapot dry] shake [the machine to pieces] paint [the wall red]
v put the book on the table NOTE: It is an open question whether the embedded event for these small clause causative constructions should include the direct object (=the small clause subject).
i. is there an additional mediating head creating the small clause, marking the relation between the subject and predicate of the small clause, as in e.g. den Dikken (2006) ii. is the direct object the “measure” of the event, or does the change of state unfold over the “path” transversed by the direct object
give [John a book] hand [John an apple]
f. “High applicative” (Pylkkänen (2000), here following a line of work from Marantz (1993)) double object constructions with benefactive readings (built from (13b) with an applicative head that relates an event to an individual) bake John a cake open John a beer make John a sandwich
(14) For the discussion of re-, keep in mind: a. the treatment in (13b) of objects as change of state events
a. Thematic approaches, where explanations are based on generalizations over theta grids or similar lexical argument structures (see Levin & Rapport) b. Templatic approaches, where generalizations are based on complementary distributions predicted by phrase-structure type restrictions (includes Hale & Keyser, Keyser & Roeper, Ramchand and a long tradition)
a. complement to seem and remain John seems/remained angry/blessed/pissed off *John remained given a book. *The teapot remained drunk dry. b. input to un- prefixation unexamined, unprepared, unloved *an ungiven person, *an undrunk dry teapot (18) Restrictions on stative passives, the first obviously paralleling restrictions on re- a. must take thematic object i. It seemed *(to have been) uttered that John should leave. ii. Tabs seemed *(to have been) kept on Jane Fonda. iii. The teapot seemed *(to have been) drunk dry. b. Sole Complement Generalization: only arguments that can stand alone with the verb in the active VP can be the subject of a stative passive i. rules out first object of double object constructions ii. rules out some verbs of spray/load type iii. rules out some verbs of change of location
b. hand a book to John/slip a message to the spy c. *hand John, *hand a message/*slip the spy, *slip a message d. *the unhanded person/book/ *the unslipped spy/message e. *The book seemed/remained handed to John. f. *The message seemed/remained slipped to the spy. g. cram food into the freezer, cram the freezer with food h. *They crammed food yesterday/ They crammed the freezer
with food. j. pile the books on the shelf/ pile the shelf with books k. They piled books yesterday/ *They piled the shelf l. The books remained piled on the shelves/ *The shelves remained piled with books. m. place books on the table/ put books on the table n. ??They placed the books yesterday/ *They put the books yesterday o. *the book seemed/remained placed/put on the table
b. *John seemed given a book yesterday. (bad because adjectives can’t “project” the obligatory second object) (22) However, this account doesn’t cover the data in (19e, f, j, l, o), all falling under the Sole Complement Generalization. (23) Syntactic theory of stative passive construction (Embick (2004) and Kratzer (2004), although there are important differences between their approaches that will not be relevant here) a. Type One, root attaching: the stative functional head attaches to a stative head (for Embick, a root) (what Kratzer calls “target state” passives) b. Type Two: the stative head attaches higher and the structure includes a causing event (so describes a completed change of state, rather than simply a state (an end state)) (=a subset of what Kratzer calls “resultant state”) (24) Today: Type two merges with a DP interpreted as a change of state, i.e., lower subevent in (13b) above, then joins the verb. (25) a. Type One (no little v)  
stative close   b. stative (could be adj) Type Two (eventive, little v)
  v
 
stative PRO (or variable) v bake The reading of the sub-event expressed by a DP, here PRO or a variable, is that the DP undergoes a change of state compatible with the causing event (here, the activity of “baking”) and with the nature of the DP itself. The stative head picks out the end state of the change of state. The entire participle phrase in (25b), then, is a predicate describing an end-state computed in the change of state sub-event and caused by the activity of “baking.” I’ve switched the linear order of the subevents here to indicate that the end state is the main subevent – and heads the construction – while the causing activity is the subordinated event. This structure would probably rule out the presence of a voice head, given other observations about how voice heads close out domains (see Pylkkänen 2007). (26) Embick on Type One: Root-attaching affixes can have stem-dependent allomorphs a. The door seems closed/open/*close/??opened b. hung jury, fallen tree, rotten fruit (apple was rotted/ *rotten by the sun) Crucially, Type One statives pattern with simple adjectives (the stative is the first category-determining head to attach to the root).
(27) How does this analysis derive the Sole Complement Generalization? To put it differently, why should the optionality of the second complement in the examples in (19) matter for the formation of the stative passives?
i. ??The food remained crammed in the freezer/ The freezer remained crammed with food.
a. re-spray the wall with water (the water could be used only for the second spraying, meaning that there’s a sub-event in the VP that excludes the water and that “with water” can be adverbial, modifying the subevent of change of state caused by spraying) cf. re-cram the freezer with meat (could have been crammed with ice-cream before) b. repaint the wall red (the wall could have been painted blue the first time, so “red” can stand outside the event of changing state via painting and can be an adverbial modifier) (30) If the second complement is optional, then, structures with both complements are potentially ambiguous between a structure in which the second complement is the head of a small clause and a structure in which the second complement functions as an adverbial modifier of a subevent: a. paint the wall b. paint [the wall blue] c. paint [ [the wall] blue]
a. re-paint the wall red, re- scopes over the change of state of the wall, not the becoming red b. re-stuff the refrigerator with beer, re- scopes over the change of state of the refrigerator, not the becoming “with beer,” so refrigerator may have been stuffed with food prior to being re-stuffed with beer
a. stative won’t attach to a small clause b. an obligatory second complement necessarily implies a small clause structure c. an optional second complement implies possible adverbial interpretation for the second complement d. presence of stative passive with second complement (painted blue) disambiguates the second complement to the adverbial reading, just as re- prefixation does
although re- requires a direct object, it will not co-occur with certain additional complements, nor will it attach to transitive verbs that exhibit certain argument structure alternations. (34) Restrictions follow from the “Abstract Clitic Hypothesis” if re- either occupies or implicates a clitic position also required by other constructions. Then re- prefixation and these other constructions will be in complementary distribution (templatic, phrase-structural approach to restrictions and co-occurrences) (35) Some of Keyser and Roeper’s generalizations and some of their data are wrong. We’ll extract the insights from their paper and only explicitly point out errors when they’re important for present purposes. (36) Re- prefixation generally obeys the Sole Complement Generalization: a. *John re-handed (Bill) (a book). b. *John re-crammed the food. c. *John re-piled the shelves. d, *John re-put (the books) (on the table).
a. Why are there so many unrepainted Scanias? b. That's the real thing.. unedited, unrecreated. c. VIA also purchased 7 more mid-train domes built for US roads, all stored unrebuilt. (38) Moreover, and more crucially, Keyser and Roeper predict that re-, like stative passives, should not co-occur with double object constructions, even benefactive double object constructions. But this prediction is false: a. Oh, and it's "not mission-critical" so they will rebuild me a computer "in a day or two." b. I went to get a mocha, iced, and they gave me a moca frappuccino... so they remade me a mocha and let me keep the frappuccino. c. He agreed that the scoop did have some flaws and remade me a new insert within a matter of days. d. When I think about the human disaster in the wake of Hurricane Katrina, there are two moments that stand out in my mind. The first is George W. Bush's press conference in Mississippi on September 2, during which he bounced uneasily from foot to foot like he couldn't wait to get out of there, looking sullen and furrowed, observing with tense jocularity that Trent Lott's house had been lost, too, and that "we" were going to rebuild him "a fantastic house" and that he, our President, was looking forward to rocking on the porch when that day came to pass.
re- the door open the door re-opened all our freezing rain and sleet melted and refroze solid   b. Type Two: DP or adjectival passive name the end state
    v
v bake re- the cake Table II: predictions about the distribution of re- prefixation
(40) Keyser & Roeper: verb classes that do not allow re- are characterized by their potential to occur in the double object construction (rather than the obligatoriness of a second complement, as in the Sole Complement Generalization) a. John bought me a car. b. *John purchased me a car. c. *John rebought the car. d. John repurchased the car. [False data: e. I rebought the first two albums and they still sound really good.] (41) Really, really false generalization. Consider “rent” a. The landlord rented the tenant an apartment. b. *The tenant rented the landlord an apartment (meaning from the landlord) c. The landlord re-rented the apartment. d. My tenant signed a one-year lease in February and was transferred to California in July. Does she forfeit her security deposit? I re-rented the apartment for Aug. 1. e. The landlord re-rented the apartment on April 2, 1996. f. Bob discovered that he had re-rented me the broken truck. (couldn’t find any examples on Google of re-rent where the tenant was the subject) (42) Contrast between stative passives and re- a. re- requires change of state and thus bi-eventive structure, while passive is OK with state and thus no higher event i. open = stative passive ii. loved, bearded, envied, ?*re-love, *re-envy
ii. unrepainted, with stative having scope over re- (43) re- reveals the parallelisms among the various change of state vPs shown below. In particular, restitutive re- targets a DP undergoing a change of state, where the target endpoint of the change may be named by the DP itself or by the verb root, or it may be filled in by the identity of the causing event. Note that the identity of the restored state must be established independently of the adverbial state modifiers in parentheses in these structures. 
   
  (wide)
 
v paint (red) v open
  re- the door re- the door
    
(red)
v build (tall) stative re- PRO
re- the house v paint (red)-re-painted house (44) If re- selects for the kind of object that the change of state DP is, semantically, its distribution would follow, given the analysis of resultative vPs argued for here. However, since restitutive again with the meaning of re- is compatible with small clauses and other double complement constructions, the question arises of whether we can predict the distribution of re- without needing to stipulate these selectional properties. Would something go wrong for a re- in the position of restitutive again attaching to, e.g., a small clause complement? 4. Theta and templatic theories (45) Against a theta approach: There is no relevant difference in behavior with respect to the stative passive or re- prefixation between “put the book on the table” “give John the book,” on the one hand, and “laugh John out the door” “drink the teapot dry,” on the other. (46) Against a templatic approach: Compositionality is determined by selection and semantic compatibility, not by syntactic or morphological slots. Where the semantic compositionality allows recursive combination, you get recursive combination, as in stative passive over re- or high applicative construction over re-. 5. Issue: (47) If the transitivity requirement on stative passives follows from semantic selection, where does the transitivity requirement on verbal passives come from, since verbal passives seem to have a formal requirement for a DP, not a semantic requirement (so non-thematic DPs including idiom chunks, expletives, etc. are OK)? 6. Williams (2006), “Telic too late” (48) Williams (2006), working within a Ramchand framework, notes the connection between the restored state of a re- prefixed verb and the part of the event structure that the verb root names. He proposes that re- attaches to the verb stem in the lexicon and thus that its scope is related to the position of the verb stem in the Ramchand first phase syntax of the VP. 17. re- attaches to whatever the verb gives content to: all the verbs are Proc-Result verbs but differ in what the verb gives content to: 1. Proc paint 2. Res>CompRes whiten, inflate 4. Proc>Res>CompRes exterminate [3 is missing in original]
why there is no ambiguity wrt Existential closure why no Proc modifying adverb can be in the scope of re- why incorporated Results must be in the scope of re-, but unincorporated Results not. why in general re- has in its scope whatever part of Fseq that the verb gives content to.
John re-built the house (the house needn’t have been “built” before; it just needs to have been a house) John re-constructed the argument A pile can go into scatter mode either by using the Scatter Pile item in the right-click/control-click menu… In scatter mode, all the files in a pile are scaled and placed around the screen so that they can all be seen at once… Clicking on any file will reform the pile on the desktop with that file on the top of the pile. (the pile of files wasn’t “formed” before; it becomes a pile again via forming when you click) “Walsh is pleased the state chose not to rebuild the Old Man of the Mountain” (a naturally occurring rock formation never built in the first place) b. In the case of Williams’ Proc attaching re-‘s, there’s no way to prevent small clause complements ReferencesArad, M. 2005. Roots and Patterns: Hebrew Morpho-syntax. NLLT Monograph 63. Dordrecht: Springer. Beck, S. and K. Johnson 2004. "Double Objects Again." Linguistic Inquiry 35(1): 97-124. Bobaljik, J. 2002. “A-Chains At The Pf-Interface: Copies And Covert Movement,” NLLT, 20.2: 197-267. Borer, H. 2005. Structuring Sense. Oxford University Press. Cuervo, C. 2003.Datives at Large. MIT PhD Dissertation. Dikken, M. den. 2006. Relators and Linkers: The Syntax of Predication, Predicate Inversion, and Copulas, Cambridge, MIT Press. Embick, D. and R. Noyer. 2001. “Movement operations after syntax.” Linguistic Inquiry 32: 555-595. Embick, D. 2004. “On the Structure of resultative participles in English.” Linguistic Inquiry, 35.3:355-392. Hale, K. and SJ Keyser. 1993. “On argument structure and the lexical expression of syntactic relations.” In K. Hale and SJ Keyser, eds., The View from Building 20, Cambridge: The MIT Press, 53-109. Hale, K. & S.J. Keyser. 2002. Prolegomenon to a Theory of Argument Structure. Cambridge, MIT Press. Halle, Morris and Alec Marantz 1993. “Distributed Morphology and the Pieces of Inflection,” in K. Hale and S.J. Keyser, eds., The View From Building 20, Cambridge, Mass.: MIT Press, 111-176. Harley, H. 2004. “Merge, conflation, and head movement: The First Sister Principle revisited,” Proceedings of NELS 34, edited by Keir Moulton and Matthew Wolf, UMass Amherst, GLSA Horn, L. 1980. “Affixation and the Unaccusative Hypothesis.” CLS 16, 134-146. Keyser, J. & T. Roeper 1992. "Re: The Abstract Clitic Hypothesis." Linguistic Inquiry, 23(1), 89-125. Kratzer, A. 1996. “Severing the external argument from its verb.” In J. Rooryck and L. Zaring, eds., Phrase structure and the lexicon, Dordrecht: Kluwer Academic Publishers, 109-137. Kratzer, A. 2004. “Building Resultatives.” To appear in C. Maienborm & A. Wöllstein-Leisten, eds., Event Arguments in Syntax, Semantics, and Discourse, Tübingen, Niemeyer. Kratzer, A. 2004. “Building Statives.” BLS.26. Levin, B. and M. Rappaport 1986. “The Formation of Adjectival Passives.” Linguistic Inquiry 17.4:623-661. Levin, B. 1999. “Objecthood: An event structure perspective.” CLS 35, volume 1. Lieber, R. 2004. Morphology and Lexical Semantics. Cambridge: Cambridge University Press. Lin, J. 2004. Event Structure and the Encoding of Arguments: The Syntax of the Mandarin and English Verb Phrase. MIT PhD dissertation, Department of Electrical Engineering and Computer Science. Marantz, A. 1984. On the Nature of Grammatical Relations. Cambridge: MIT Press. Marantz, A. 1992. The way Construction and the Semantics of Direct Arguments in English. In E. Wehrli and T. Stowell (eds.), Syntax and the Lexicon, Syntax and Semantics, Vol 20., New York: Academic Press, 179-188. Marantz, A. 1993. “Implications of Asymmetries in Double Object Constructions.” In Sam A. Mchombo, ed., Theoretical Aspects of Bantu Grammar 1. CSLI Publications, Stanford, CA, 113-151. Marantz, A. 1997. "No Escape from Syntax: Don't Try Morphological Analysis in the Privacy of Your Own Lexicon," in A. Dimitriadis, L. Siegel, et al., eds., University of Pennsylvania Working Papers in Linguistics, Vol. 4.2, Proceedings of the 21st Annual Penn Linguistics Colloquium, pp. 201-225. Pylkkänen, L. 1999. “Causation and External Arguments.” In L. Pylkkänen, A. van Hout & H. Harley (eds.), Papers from the Second Penn/MIT Roundtable on Argument Structure and the Lexicon, MITWPL 35. Pylkkänen, L. 2000. “What Applicative Heads Apply To.” The 24th Penn Linguistics Colloquium, University of Pennsylvania, Philadelphia. Pylkkänen, L. 2000. “Deriving Adversity.” WCCFL XIX, UCLA, Los Angeles. Pylkkänen, L., 2007. Introducing Arguments. Cambridge, MIT Press. Ramchand, G. 2003. “First Phase Syntax.” Oxford University ms. Stechow, A. von 1996. "The Different Readings of Wieder'Again': A Structural Account." Journal of Semantics 13(2): 87. Williams, E. 2006. “Telic too Late.” Harvard University talk handout, November, 2006. Yüklə 94,42 Kb. Dostları ilə paylaş:
|
v stative eventuality, interpreted as change
v
v
