| { | |
| "paper_id": "W98-0104", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T06:03:08.992140Z" | |
| }, | |
| "title": "Motion Verbsand Semantic Features in TAG*", | |
| "authors": [ | |
| { | |
| "first": "Tonia", | |
| "middle": [], | |
| "last": "Bleam", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Cognitive Science University of Pennsylvania 400A", | |
| "location": { | |
| "postCode": "3401, 19104", | |
| "settlement": "\\i\\Talnut St. Philadelphia", | |
| "region": "PA" | |
| } | |
| }, | |
| "email": "[email protected]" | |
| }, | |
| { | |
| "first": "Martha", | |
| "middle": [], | |
| "last": "Palmer", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Cognitive Science University of Pennsylvania 400A", | |
| "location": { | |
| "postCode": "3401, 19104", | |
| "settlement": "\\i\\Talnut St. Philadelphia", | |
| "region": "PA" | |
| } | |
| }, | |
| "email": "[email protected]" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Vijay-Shanker", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Cognitive Science University of Pennsylvania 400A", | |
| "location": { | |
| "postCode": "3401, 19104", | |
| "settlement": "\\i\\Talnut St. Philadelphia", | |
| "region": "PA" | |
| } | |
| }, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "", | |
| "pdf_parse": { | |
| "paper_id": "W98-0104", | |
| "_pdf_hash": "", | |
| "abstract": [], | |
| "body_text": [ | |
| { | |
| "text": "We show how the domain of locality in a TAG elementary tree, (Frank 1992) , can be extended through adjunction to include optional arguments for a dass of motion verbs and how the adjunctions can be restricted appropriately through the use of semantic features. Same examples of motion verbs we consider are shown in Table 1 , which categorizes the verbs according to Levin classes (Levin 1993) . Note that we are using a broader definition of \"motion\" verbs than Levin's These verbs are cla.ssified according to their syntactic behavior, which is taken to be a reftection of their underlying semantic properties. Motion verbs are able to occur with path phrases, where the term \"path\" is used as a cover term for source, goal, via and directional modifiers (PPs and adverbs), along the lines of Jackendoff (1976 Jackendoff ( , 1990 . Examples of these are given in (1-4).", | |
| "cite_spans": [ | |
| { | |
| "start": 61, | |
| "end": 73, | |
| "text": "(Frank 1992)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 382, | |
| "end": 394, | |
| "text": "(Levin 1993)", | |
| "ref_id": "BIBREF4" | |
| }, | |
| { | |
| "start": 464, | |
| "end": 471, | |
| "text": "Levin's", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 796, | |
| "end": 812, | |
| "text": "Jackendoff (1976", | |
| "ref_id": "BIBREF1" | |
| }, | |
| { | |
| "start": 813, | |
| "end": 832, | |
| "text": "Jackendoff ( , 1990", | |
| "ref_id": "BIBREF2" | |
| } | |
| ], | |
| "ref_spans": [ | |
| { | |
| "start": 317, | |
| "end": 324, | |
| "text": "Table 1", | |
| "ref_id": "TABREF0" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "2 Manner of motion verbs: (Run and Roll classes)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "I ran to the store. (goal)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "\u2022 We would like to thank Hoa. Trang Da.ng, Christy Dorau, Aravind Joshi, Tony Kroch, Jelf Lidz, Joseph Rozenzweig, Matthew Stone, aud two anouymous reviewers for helpful discussion and/or participation in this researcli.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "(2) (3) . The path features can take the values +, 0, or NONE. 1 A \"+\" value means that the feature has been specified. A \"O\" value means that it has not yet been specified, but that it is appropriate for this feature to have a value. The feature [path: goal:OJ or (path: source:OJ occurs in the foot node of adjoining trees that represent source or goal, to ensnre that an element with that value has not already been adjoined. An example is shown in Figure 1 above. The value \"NONE\", on the other hand, means that it is not appropriate to specify this value. EventType features are also atomically valued, taking the values + or -. Having the feature [eventType: motion:-] means that the event is unable to be interpreted as a motion event and entails that path phrases cannot adjoin on. An example of a verb with this feature might be eat. On the other hand, nonspecification of the (eventType: motion] feature entails that path phrases can adjoin. If a path phrase does adjoin, the event becomes a motion event. Sound emission verbs are of this sort.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 452, | |
| "end": 460, | |
| "text": "Figure 1", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Introduction", | |
| "sec_num": "1" | |
| }, | |
| { | |
| "text": "The class of verbs of inherently directed motion (VIDMs) have a path component built into the meaning of the verb. Usually the verb specifies a source, as in leave and exit, or a goal, as in enter, arrive. One interesting property of VIDMs is that they have a more limited ability to take path PPs even though they are motion verbs. For example, arrive does not take a prototypical goal PP (with the preposition to), but instead takes a locative PP which represents the goal of motion .", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Verbs of Inherently Directed Motion", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "(5) 6a.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Verbs of Inherently Directed Motion", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "Mary arrived at the station.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Verbs of Inherently Directed Motion", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "b. * Mary arrived to the station. arrive = (GO (TO X]J (where X=location)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Verbs of Inherently Directed Motion", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "Following J ackendoff 1990, we analyze the goal function \"TO\" as being incorporated in the LCS of arrive, shown in (6). The PP slot in 1 \\Ve usc atomically valued features for source and goal rather than putting in the actual value of the goal (i.e. the referent of the goal) because the simple presence or absence of these features is what affects the derivation. That is, having a goal present means that another goal cannot adjoin on (but see footnote 3). For this purpose, the referent of the goal does not need to be represented. the subcategorization frame is coindexed with the location argument slot X. Therefore, the PP that represents the goal must be a location. In TAG terms, we assume that the part of the path inherently specified in the verb semantics constitutes an (optional) argument. In order to constrain what kind of preposition can instantiate the goal, we will need to define a dass of locative prepositions and impose a constraint on the P node so that only this dass is allowed to occur there. For now, we show the feature (locative:+ J on the P node of the elementary tree for arrive in Figure 2 . In addition, many VIDMs (like arrive, enter, exit) are achievement verbs; that is, they have no durative aspect. Because of this, they cannot take a path phrase that modifies durative motion.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 1113, | |
| "end": 1121, | |
| "text": "Figure 2", | |
| "ref_id": "FIGREF2" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "Verbs of Inherently Directed Motion", | |
| "sec_num": "4" | |
| }, | |
| { | |
| "text": "John arrived (\"'around the lake) at Mary's house.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(7)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The [via:NONE] and (direction:NONEJ features in the VP node in Figure 2 represent a non-durative path. While via and direction PPs cannot occur with arrive, a source can be specified, as shown in (8), because this does not conflict with the lack of durativity of the event.", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 63, | |
| "end": 71, | |
| "text": "Figure 2", | |
| "ref_id": "FIGREF2" | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "(7)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "John arrived in Chicago from Philadelphia.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(8)", | |
| "sec_num": "5" | |
| }, | |
| { | |
| "text": "Path phrases can adjoin to a VP node which is unspecified fcr motion. Even verbs that are not inherently motion verbs can be modified by path phrases, augmenting their semantic representatinn to include explicit motion. For instance, verbs of sound emission such as whistle and roar can convey directed motion when they appear with path phrases, as in (9) and (10).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Regular sense extensions", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The train whistled into the station.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Regular sense extensions", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The truck roared past the weigh station, Additionally, we see other cases where the syntactic frame in which a verb occurs determines the senses that a verb can have. For example, push can have the senses shown in (11) (12) (13) (14) . (See Dang et al. 1998 for discussion) . 11Mary pushed the chair. {force:+, contact:+] 12Mary pushed the cart to the store. [motion:+, path:+J", | |
| "cite_spans": [ | |
| { | |
| "start": 214, | |
| "end": 218, | |
| "text": "(11)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 219, | |
| "end": 223, | |
| "text": "(12)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 224, | |
| "end": 228, | |
| "text": "(13)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 229, | |
| "end": 233, | |
| "text": "(14)", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 241, | |
| "end": 273, | |
| "text": "Dang et al. 1998 for discussion)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "{10)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Mary pushed the branches apart. (motion:+, separation:+J", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "{10)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Mary pushed at the boulder. {motion:-]", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "{10)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The transitive sentences (11), (12), and (13) will all be generated from a transitive elementary tree where the VP node has the features [force:+} and [contact:+], but is unspecified for {motion]. Adjoining in the modifiers to the store and apart will introduce the additional features listed in (12) and 13, respectively.", | |
| "cite_spans": [ | |
| { | |
| "start": 296, | |
| "end": 300, | |
| "text": "(12)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "{10)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The conative construction (illustrated in (14)) is represented by the elementary tree given in Figure 3 . ", | |
| "cite_spans": [], | |
| "ref_spans": [ | |
| { | |
| "start": 95, | |
| "end": 103, | |
| "text": "Figure 3", | |
| "ref_id": null | |
| } | |
| ], | |
| "eq_spans": [], | |
| "section": "{10)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "The first case is represented by the following example, where at the station represents the goal that is implicit in the meaning of arrive.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Optional arguments of Verbs of Inherently Directed Motion", | |
| "sec_num": "6.1" | |
| }, | |
| { | |
| "text": "The train arrived at the station.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(15)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Note that the meaning of (15) is not compositional since at the station by itself or combined with a motion verb like run can only mean a location of the event. 2 lt cannot represent the goal of motion in these cases. 16The athlete ran at the gym. 171 saw Mary at the station.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(15)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "lt is only with a verb whose meaning includes [goal:+}, that an at-PP or any other locative PP can represent a goal. Thus, in this example, it is the head verb w hich determines the role that the PP phrase ha.~ in the clause. This kind of idiosyncracy is evidence that a constituent is an argument rather than an adjunct (see e.g., Pollard and Sag 1987; Marantz 1984) . By this criterion, then, the PP representing an inherent .role of a VIDM should be considered an argument, and thus, should be present in the elementary tree.", | |
| "cite_spans": [ | |
| { | |
| "start": 332, | |
| "end": 353, | |
| "text": "Pollard and Sag 1987;", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 354, | |
| "end": 367, | |
| "text": "Marantz 1984)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(15)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "lt has been noted that all source and goal PPs simultaneously show both argument and adjunct properties. Larson (1988) discusses the argument status of the source and goal phrases in sentences like (18) and (19).", | |
| "cite_spans": [ | |
| { | |
| "start": 105, | |
| "end": 118, | |
| "text": "Larson (1988)", | |
| "ref_id": "BIBREF3" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(15)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "John walked to the store. ", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(18)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "*Max got a letter from Felix from his friend.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(18)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Jackendoff {1976) takes motion verbs to contain the abstract predicate GO which is a threeplace relation, taking the arguments (x,y,z), where x is an element that moves from y (source) to z (goal).", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(18)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "For current purposes, however, we do not take all sources and goals to be present in the elementary tree. Only PPs whose meaning is implicit in the meaning of the verb itself are present in the elementary tree, whereas all other PPs are adjoined. This is in contrast with the analysis provided by Levin and Rappaport Hovav (1995) in which all sources and goals are treated as arguments as a result of a lexical rule that applies to verbs of motion.", | |
| "cite_spans": [ | |
| { | |
| "start": 297, | |
| "end": 329, | |
| "text": "Levin and Rappaport Hovav (1995)", | |
| "ref_id": "BIBREF5" | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "(18)", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "Elementary Trees The other case to consider is the conative at construction, shown in (22). 22The child hit at the ball.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The Conative Construction and", | |
| "sec_num": "6.2" | |
| }, | |
| { | |
| "text": "We assume that the conative at PP is present in the elementary tree. If we took the at PP to be adjoined in, then an intranstive elementary tree for hit is required. However, hit can only occur transitively, and so we would need additional mechanisms for blocking the intransitive tree from ever occurring outside of the conative construction.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The Conative Construction and", | |
| "sec_num": "6.2" | |
| }, | |
| { | |
| "text": "On the other hand, we could take the position that the noun phrase ( the ball in (22)) is an argument of the verb, and at adjoins in. However, it is not possible for a PP to adjoin at this point. 4 The conative is properly analyzed as a lexical process of object demotion -an operation that applies to the lexical representation of the verb, affecting its argument-structure. lt demotes a direct object to be an oblique element with the effect that the object is interpreted a.s not affected by the action of the verb. However, in TAG, there is no level of representation independent of the elementary trees in which demotion operations of this sort could take place. Therefore, the best TAG analysis of the conative treats the PP a.s an argument, and hence, present in the elementary tree.", | |
| "cite_spans": [ | |
| { | |
| "start": 196, | |
| "end": 197, | |
| "text": "4", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "The Conative Construction and", | |
| "sec_num": "6.2" | |
| }, | |
| { | |
| "text": "The goal of our work is to capture lexical semantic properties that we hope will be helpful in reducing the search space in parsing, as well as aid in generation (SPUD; see Stone and Doran 1997; Stone and Webber 1998) and machine translation (in the transfer of lexical semantic properties) (see Palmer, et al. (to appear) ).", | |
| "cite_spans": [ | |
| { | |
| "start": 173, | |
| "end": 194, | |
| "text": "Stone and Doran 1997;", | |
| "ref_id": null | |
| }, | |
| { | |
| "start": 195, | |
| "end": 217, | |
| "text": "Stone and Webber 1998)", | |
| "ref_id": "BIBREF8" | |
| }, | |
| { | |
| "start": 296, | |
| "end": 322, | |
| "text": "Palmer, et al. (to appear)", | |
| "ref_id": null | |
| } | |
| ], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusion", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "We have examined several subclasses of motion verbs, and posited features to capture their semantic properties. These features not only allow us to place restrictions on the verbs to constrain possible derivations, but also allow us to account for regular sense extensions through the underspecification of certain features and by having modifiers introduce these features in the course of the derivation.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Conclusion", | |
| "sec_num": "7" | |
| }, | |
| { | |
| "text": "It can also have the meaning of toward3.3 Note that (20) isn't that bad lf the second PP is interpreted as a further specification of the goal location.(1)? Iran to Philadelphia to IRCS.We do not yet have an account of this phenomenon, but we do not take it as counterevidence to the generalization that only one goal may be given per event. This is unlike true modifiers like PPs of location, of whic!: more than one can be given without any restriction:(2)15 I hid in the building on the third floor in a classroom under a desk.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| }, | |
| { | |
| "text": "It would only be possible for an NP to adjoin, requiring an analysis of the at PP as an NP, which is linguistically unmotivated.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| } | |
| ], | |
| "back_matter": [], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "Syntactic locality and '.lree Adjoining Grnmmar: grammatical, acquisition and processing perspectives", | |
| "authors": [ | |
| { | |
| "first": "H", | |
| "middle": [], | |
| "last": "Dang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Trang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Kipper", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Palmer", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Rosenzweig", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1992, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Dang, H. Trang, K. Kipper, M. Palmer, J. Rosen- zweig. (1998). Investigating regular sense exten- sions based on intersective Levin classes. ACL Frank, R\" (1992). Syntactic locality and '.lree Ad- joining Grnmmar: grammatical, acquisition and processing perspectives. PhD Thesis, University of Pennsylvania, Technical Report IRCS-92-47.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Toward an Explanatory Semantic Representation", | |
| "authors": [ | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Jackendoff", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1976, | |
| "venue": "Linguistic lnquiry", | |
| "volume": "7", | |
| "issue": "", | |
| "pages": "89--150", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Jackendoff, R. (1976). Toward an Explanatory Se- mantic Representation. Linguistic lnquiry 7, 89- 150.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "Semantic Strv.ctures", | |
| "authors": [ | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Jackendoff", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1990, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Jackendoff, R. (1990). Semantic Strv.ctures. MIT Press.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Implicit Arguments in Situation Semantics", | |
| "authors": [ | |
| { | |
| "first": "R", | |
| "middle": [], | |
| "last": "Larson", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1988, | |
| "venue": "Linguistics and Philosophy", | |
| "volume": "11", | |
| "issue": "", | |
| "pages": "169--201", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Larson, R. (1988). Implicit Arguments in Situation Semantics. Linguistics and Philosophy 11:169-201.", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "English Verb Classes and Altemations", | |
| "authors": [ | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Levin", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1993, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Levin, B. (1993). English Verb Classes and Altema- tions. University of Chicago Press.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "Unaccusativity: At the Syntax-Lexical Semantics Interface", | |
| "authors": [ | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Levin", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Rappaport Hovav", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1995, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Levin, B., and M. Rappaport Hovav. (1995). Un- accusativity: At the Syntax-Lexical Semantics In- terface. MIT Press.", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "Capturing Motion Verb Generalizations with Synchronous TAGs", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Palmer", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Rosenzweig", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "W", | |
| "middle": [], | |
| "last": "Schuler", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1998, | |
| "venue": "St. Dizier, Patrick, Predicative Forms in NLP", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Palmer, M., J. Rosenzweig, W. Schuler. (to appear, 1998). Capturing Motion Verb Generalizations with Synchronous TAGs. in St. Dizier, Patrick, Predicative Forms in NLP, Kluwer Press.", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "Sentence planning as description using tree-adjoining grammar", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Stone", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Daran", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1997, | |
| "venue": "Proceedings of ACL", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "198--205", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Stone, M. and C. Daran. (1997). Sentence planning as description using tree-adjoining grammar. Pro- ceedings of ACL, 198-205.", | |
| "links": null | |
| }, | |
| "BIBREF8": { | |
| "ref_id": "b8", | |
| "title": "Textual economy through close coupling of syntax and semantics. International Natural Language Generation Workshop", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Stone", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Webber", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 1998, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Stone, M. and B. Webber. (1998). Textual econ- omy through close coupling of syntax and seman- tics. International Natural Language Generation Workshop", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "FIGREF0": { | |
| "uris": null, | |
| "num": null, | |
| "text": ":motion:-t; palh:goal:+.", | |
| "type_str": "figure" | |
| }, | |
| "FIGREF2": { | |
| "uris": null, | |
| "num": null, | |
| "text": "Elementary Tree for arrive", | |
| "type_str": "figure" | |
| }, | |
| "FIGREF3": { | |
| "uris": null, | |
| "num": null, | |
| "text": "discussion of VIDMs and regular sense extensions above is the assumption that some PPs are arguments of the verbs they occur with, and hence are present in the verb's elementary tree. The cases in question are (1) the PP which reoresents the inherentlv soecified path of a VIDM; and (2) the at PP of the conative construction.", | |
| "type_str": "figure" | |
| }, | |
| "FIGREF4": { | |
| "uris": null, | |
| "num": null, | |
| "text": "the hause. They act like adjuncts in being optional, but like arguments in being non-iterable. (The following examples are Larson's.) 3 (20) *John flew to New York to Kennedy lnt'l Airport.", | |
| "type_str": "figure" | |
| }, | |
| "TABREF0": { | |
| "num": null, | |
| "content": "<table><tr><td colspan=\"2\">VIDMs Roll</td><td>Run</td><td colspan=\"2\">Force Carry</td></tr><tr><td>arrive</td><td>ftoat</td><td colspan=\"3\">jump press carry</td></tr><tr><td>enter</td><td>roll</td><td>run</td><td>pull</td><td>lug</td></tr><tr><td>escape</td><td>slide</td><td colspan=\"3\">slide push pull</td></tr><tr><td>exit</td><td colspan=\"2\">rotate walk</td><td/><td>push</td></tr><tr><td/><td>turn</td><td/><td/><td/></tr></table>", | |
| "html": null, | |
| "type_str": "table", | |
| "text": "dass 51. Levin Classes of Verbs Involving Motion" | |
| } | |
| } | |
| } | |
| } |