Base_JSON/prefixS/json/S14/S14-1014.json

{
    "paper_id": "S14-1014",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T15:32:33.968888Z"
    },
    "title": "Generating Simulations of Motion Events from Verbal Descriptions",
    "authors": [
        {
            "first": "James",
            "middle": [],
            "last": "Pustejovsky",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "Brandeis University Waltham",
                "location": {
                    "region": "MA",
                    "country": "USA"
                }
            },
            "email": "jamesp@cs.brandeis.edu"
        },
        {
            "first": "Nikhil",
            "middle": [],
            "last": "Krishnaswamy",
            "suffix": "",
            "affiliation": {},
            "email": ""
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "In this paper, we describe a computational model for motion events in natural language that maps from linguistic expressions, through a dynamic event interpretation, into three-dimensional temporal simulations in a model. Starting with the model from (Pustejovsky and Moszkowicz, 2011), we analyze motion events using temporally-traced Labelled Transition Systems. We model the distinction between path-and manner-motion in an operational semantics, and further distinguish different types of manner-of-motion verbs in terms of the mereo-topological relations that hold throughout the process of movement. From these representations, we generate minimal models, which are realized as three-dimensional simulations in software developed with the game engine, Unity. The generated simulations act as a conceptual \"debugger\" for the semantics of different motion verbs: that is, by testing for consistency and informativeness in the model, simulations expose the presuppositions associated with linguistic expressions and their compositions. Because the model generation component is still incomplete, this paper focuses on an implementation which maps directly from linguistic interpretations into the Unity code snippets that create the simulations.",
    "pdf_parse": {
        "paper_id": "S14-1014",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "In this paper, we describe a computational model for motion events in natural language that maps from linguistic expressions, through a dynamic event interpretation, into three-dimensional temporal simulations in a model. Starting with the model from (Pustejovsky and Moszkowicz, 2011), we analyze motion events using temporally-traced Labelled Transition Systems. We model the distinction between path-and manner-motion in an operational semantics, and further distinguish different types of manner-of-motion verbs in terms of the mereo-topological relations that hold throughout the process of movement. From these representations, we generate minimal models, which are realized as three-dimensional simulations in software developed with the game engine, Unity. The generated simulations act as a conceptual \"debugger\" for the semantics of different motion verbs: that is, by testing for consistency and informativeness in the model, simulations expose the presuppositions associated with linguistic expressions and their compositions. Because the model generation component is still incomplete, this paper focuses on an implementation which maps directly from linguistic interpretations into the Unity code snippets that create the simulations.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "Semantic interpretation requires access to both knowledge about words and how they compose. As the linguistic phenomena associated with lexical semantics have become better understood, several assumptions have emerged across most models of word meaning. These include the following:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "(1) a. Lexical meaning involves some sort of \"componential analysis\", either through predicative primitives or a system of types. b. The selectional properties of predicators can be explained in terms of these components; c. An understanding of event semantics and the different role of event participants seems crucial for modeling linguistic utterances.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "As a starting point in lexical semantic analysis, a standard methodology in both theoretical and computational linguistics is to identify features in a corpus that differentiate the data in meaningful ways; meaningful in terms of prior theoretical assumptions or in terms of observably differentiated behaviors. Combining these strategies we might, for instance, take a theoretical constraint that we hope to justify through behavioral distinctions in the data. An example of this is the theoretical claim that motion verbs can be meaningfully divided into two classes: manner-and path-oriented predicates (Talmy, 1985; Jackendoff, 1983; Talmy, 2000) . These constructions can be viewed as encoding two aspects of meaning: how the movement is happening and where it is happening. The former strategy is illustrated in (2a) and the latter in (2b) (where m indicates a manner verb, and p indicates a path verb).",
                "cite_spans": [
                    {
                        "start": 606,
                        "end": 619,
                        "text": "(Talmy, 1985;",
                        "ref_id": "BIBREF48"
                    },
                    {
                        "start": 620,
                        "end": 637,
                        "text": "Jackendoff, 1983;",
                        "ref_id": "BIBREF23"
                    },
                    {
                        "start": 638,
                        "end": 650,
                        "text": "Talmy, 2000)",
                        "ref_id": "BIBREF49"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "(2) a. The ball rolled m . b. The ball crossed p the room.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "With both of the verb types, adjunction can make reference to the missing aspect of motion, by introducing a path (as in (3a)) or the manner of movement (in (3b)).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "(3) a. The ball rolled m across the room. b. The ball crossed p the room rolling.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Differences in syntactic distribution and grammatical behavior in large datasets, in fact, correlate fairly closely with the theoretical claims made by linguists using small introspective datasets. The path-manner classification is a case where there are data-derived distinctions that correlate nicely with theoretically inspired predictions. More often than not, however, lexical semantic distinctions are formal stipulations in a linguistic model, that often have no observable correlations to data. For example, an examination of the manner of movement class from Levin (1993) illustrates this point. The verbs below are all Levinclass manner of motion verbs:",
                "cite_spans": [
                    {
                        "start": 568,
                        "end": 580,
                        "text": "Levin (1993)",
                        "ref_id": "BIBREF28"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "(4) MANNER OF MOTION VERBS: drive, walk, run, crawl, fly, swim, drag, slide, hop, roll Assuming the two-way distinction between path and manner predication of motion mentioned above, these verbs do, in fact, tend to pattern according to the latter class in the corpus. Given that they are all manner of motion verbs, however, any data-derived distinctions that emerge within this class will have to be made in terms of additional syntactic or semantic dimensions. While it is most likely possible to differentiate, for example, the verbs slide from roll, or walk from hop in the corpus, given enough data, it is important to realize that conceptual and theoretical modeling is often necessary to reveal the factors that semantically distinguish such linguistic expressions, in the first place. We argue that this problem can be approached with the use of minimal model generation. As Blackburn and Bos (2008) point out, theorem proving (essentially type satisfaction of a verb in one class as opposed to another) provides a \"negative handle\" on the problem of determining consistency and informativeness for an utterance, while model building provides a \"positive handle\" on both. For our concerns, simulation construction provides a positive handle on whether two manner of motion processes are distinguished in the model. Further, the simulation must specify how they are distinguished, the analogue to informativeness.",
                "cite_spans": [
                    {
                        "start": 28,
                        "end": 86,
                        "text": "drive, walk, run, crawl, fly, swim, drag, slide, hop, roll",
                        "ref_id": null
                    },
                    {
                        "start": 884,
                        "end": 908,
                        "text": "Blackburn and Bos (2008)",
                        "ref_id": "BIBREF3"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "In this paper, we argue that traditional lexical modeling can benefit greatly from examining how semantic interpretations are contextually and conceptually grounded. We explore a dynamic interpretation of the lexical semantic model developed in Generative Lexicon Theory (Pustejovsky, 1995; . Specifically, we are interested in using model building (Blackburn and Bos, 2008; Konrad, 2004; Gardent and Konrad, 2000) and simulation generation (Coyne and Sproat, 2001; Siskind, 2011) to reveal the conceptual presuppositions inherent in natural language expressions. In this paper, we focus our attention on motion verbs, in order to distinguish between manner and path motion verbs, as well as to model mereotopological distinctions within the manner class.",
                "cite_spans": [
                    {
                        "start": 271,
                        "end": 290,
                        "text": "(Pustejovsky, 1995;",
                        "ref_id": "BIBREF39"
                    },
                    {
                        "start": 349,
                        "end": 374,
                        "text": "(Blackburn and Bos, 2008;",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 375,
                        "end": 388,
                        "text": "Konrad, 2004;",
                        "ref_id": "BIBREF25"
                    },
                    {
                        "start": 389,
                        "end": 414,
                        "text": "Gardent and Konrad, 2000)",
                        "ref_id": "BIBREF16"
                    },
                    {
                        "start": 441,
                        "end": 465,
                        "text": "(Coyne and Sproat, 2001;",
                        "ref_id": "BIBREF5"
                    },
                    {
                        "start": 466,
                        "end": 480,
                        "text": "Siskind, 2011)",
                        "ref_id": "BIBREF45"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "The interpretation of motion in language has been one of the most researched areas in linguistics and Artificial Intelligence (Kuipers, 2000; Freksa, 1992; Galton, 2000; Levinson, 2003; Mani and Pustejovsky, 2012) . Because of their grammatical and semantic import, linguistic interest in identifying where events happen has focused largely on motion verbs and the role played by paths. Jackendoff (1983) , for example, elaborates a semantics for motion verbs incorporating explicit reference to the path traversed by the mover, from source to destination (goal) locations. Talmy (1983) develops a similar conceptual template, where the path followed by the figure is integral to the conceptualization of the motion against a ground. Hence, the path can be identified as the central element in defining the location of the event (Talmy, 2000) . Related to this idea, both Zwarts (2005) and Pustejovsky and Moszkowicz (2011) develop mechanisms for dynamically creating the path traversed by a mover in a manner of motion predicate, such as run or drive. Starting with this approach, the localization of a motion event, therefore, is at least minimally associated with the path created by virtue of the activity.",
                "cite_spans": [
                    {
                        "start": 126,
                        "end": 141,
                        "text": "(Kuipers, 2000;",
                        "ref_id": "BIBREF27"
                    },
                    {
                        "start": 142,
                        "end": 155,
                        "text": "Freksa, 1992;",
                        "ref_id": "BIBREF13"
                    },
                    {
                        "start": 156,
                        "end": 169,
                        "text": "Galton, 2000;",
                        "ref_id": "BIBREF14"
                    },
                    {
                        "start": 170,
                        "end": 185,
                        "text": "Levinson, 2003;",
                        "ref_id": "BIBREF29"
                    },
                    {
                        "start": 186,
                        "end": 213,
                        "text": "Mani and Pustejovsky, 2012)",
                        "ref_id": "BIBREF30"
                    },
                    {
                        "start": 387,
                        "end": 404,
                        "text": "Jackendoff (1983)",
                        "ref_id": "BIBREF23"
                    },
                    {
                        "start": 574,
                        "end": 586,
                        "text": "Talmy (1983)",
                        "ref_id": "BIBREF47"
                    },
                    {
                        "start": 829,
                        "end": 842,
                        "text": "(Talmy, 2000)",
                        "ref_id": "BIBREF49"
                    },
                    {
                        "start": 872,
                        "end": 885,
                        "text": "Zwarts (2005)",
                        "ref_id": "BIBREF53"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Situating Motion in Space and Time",
                "sec_num": "2"
            },
            {
                "text": "In addition to capturing the spatial trace of the object in motion, several researchers have pointed out that identifying the shape of the path during motion is also critical for fully interpreting the semantics of movement. Eschenbach et al. (1999) discusses the orientation associated with the trajectory, something they refer to as oriented curves. Motivated more by linguistic considerations, Zwarts (2006) introduces the notion of an event shape, which is the trajectory associated with an event in space represented by a path. He defines a shape function, which is a partial function assigning unique paths to those events involving motion or extension in physical space. This work suggests that the localization of an event makes reference to orientational as well as configurational factors, a view that is pursued in Pustejovsky (2013b). This forces us to look at the various spatio-temporal regions associated with the event participants, and the interactions between them.",
                "cite_spans": [
                    {
                        "start": 225,
                        "end": 249,
                        "text": "Eschenbach et al. (1999)",
                        "ref_id": "BIBREF9"
                    },
                    {
                        "start": 397,
                        "end": 410,
                        "text": "Zwarts (2006)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Situating Motion in Space and Time",
                "sec_num": "2"
            },
            {
                "text": "These issues are relevant to our present concerns, because in order to construct a simulation, a motion event must be embedded within an appropriate minimal embedding space. This must sufficiently enclose the event localization, while optionally including room enough for a frame of reference visualization of the event (the viewer's perspective). We return to this issue later in the paper when constructing our simulation from the semantic interpretation associated with motion events.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Situating Motion in Space and Time",
                "sec_num": "2"
            },
            {
                "text": "The advantage of adopting a dynamic interpretation of motion is that we can directly distinguish path predication from manner of motion predication in an operational semantics (Miller and Charles, 1991; Miller and Johnson-Laird, 1976 ) that maps nicely to a simulation environment. Models of processes using updating typically make reference to the notion of a state transition (van Benthem, 1991; Harel, 1984) . This is done by distinguishing between formulae, \u03c6, and programs, \u03c0. A formula is interpreted as a classical propositional expression, with assignment of a truth value in a specific model. We will interpret specific models by reference to specific states. A state is a set of propositions with assignments to variables at a specific index. Atomic programs are input/output relations ( [[\u03c0] ] \u2286 S \u00d7 S), and compound programs are constructed from atomic ones following rules of dynamic logic (Harel et al., 2000) .",
                "cite_spans": [
                    {
                        "start": 176,
                        "end": 202,
                        "text": "(Miller and Charles, 1991;",
                        "ref_id": "BIBREF32"
                    },
                    {
                        "start": 203,
                        "end": 233,
                        "text": "Miller and Johnson-Laird, 1976",
                        "ref_id": "BIBREF33"
                    },
                    {
                        "start": 378,
                        "end": 397,
                        "text": "(van Benthem, 1991;",
                        "ref_id": null
                    },
                    {
                        "start": 398,
                        "end": 410,
                        "text": "Harel, 1984)",
                        "ref_id": "BIBREF22"
                    },
                    {
                        "start": 796,
                        "end": 802,
                        "text": "( [[\u03c0]",
                        "ref_id": null
                    },
                    {
                        "start": 903,
                        "end": 923,
                        "text": "(Harel et al., 2000)",
                        "ref_id": "BIBREF21"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Theoretical Assumptions",
                "sec_num": "3.1"
            },
            {
                "text": "For the present discussion, we represent the dynamics of actions in terms of Labeled Transition Systems (LTSs) (van Benthem, 1991) . 1 An LTS consists of a triple, S, Act, \u2192 , where: S is the set of states; Act is a set of actions; and \u2192 is a total transition relation: \u2192\u2286 S\u00d7Act\u00d7S. An action, \u03b1 \u2208 Act, provides the labeling on an arrow, making it explicit what brings about a state-to-state transition. As a shorthand for (e 1 , \u03b1, e 2 ) \u2208\u2192, we will also use e 1 \u03b1 \u2212\u2192 e 2 . If reference to the state content (rather than state name) is required for interpretation purposes (van Benthem et al., 1994) , then as shorthand for ({\u03c6} e 1 , \u03b1, {\u00ac\u03c6} e 2 ) \u2208\u2192, we use, \u03c6",
                "cite_spans": [
                    {
                        "start": 111,
                        "end": 130,
                        "text": "(van Benthem, 1991)",
                        "ref_id": null
                    },
                    {
                        "start": 133,
                        "end": 134,
                        "text": "1",
                        "ref_id": null
                    },
                    {
                        "start": 573,
                        "end": 599,
                        "text": "(van Benthem et al., 1994)",
                        "ref_id": "BIBREF50"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Theoretical Assumptions",
                "sec_num": "3.1"
            },
            {
                "text": "e 1 \u03b1 \u2212\u2192 \u00ac\u03c6 e 2",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Theoretical Assumptions",
                "sec_num": "3.1"
            },
            {
                "text": ". Finally, when referring to temporally-indexed states in the model, where e i @i indicates the state e i interpreted at time i, as shorthand for ({\u03c6} e 1 @i , \u03b1, {\u00ac\u03c6} e 2 @i+1 ) \u2208\u2192, we",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Theoretical Assumptions",
                "sec_num": "3.1"
            },
            {
                "text": "will use, \u03c6 i e 1 \u03b1 \u2212\u2192 \u00ac\u03c6 i+1 e 2",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Theoretical Assumptions",
                "sec_num": "3.1"
            },
            {
                "text": ", as described in Pustejovsky (2013).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Theoretical Assumptions",
                "sec_num": "3.1"
            },
            {
                "text": "We will assume that change of location of an object can be viewed as a special instance of a firstorder program, which we will refer to as \u03bd (Pustejovsky and Moszkowicz, 2011). 2 5x := y (\u03bd-transition, where loc(z) is value being updated) \"x assumes the value given to y in the next state.\"",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Distinguishing Path and Manner Motion",
                "sec_num": "3.2"
            },
            {
                "text": "M, (i, i + 1), (u, u[x/u(y)]) |= x := y iff M, i, u |= loc(z) = x \u2227 M, i + 1, u[x/u(y)] |= loc(z) = y",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Distinguishing Path and Manner Motion",
                "sec_num": "3.2"
            },
            {
                "text": "Given a simple transition, a process can be viewed as simply an iteration of \u03bd (Fernando, 2009) . However, as (Pustejovsky, 2013a) points out, since most manner motion verbs in language are actually directed processes, simple decompositions into change-of-location are inadequate. That is, they are guarded transitions where the test is not just non-coreference, but makes reference to values on a scale, C, and ensures that it continues in an order-preserving change through the iterations. When this test references the values on a scale, C, we call this a directed \u03bd-transition ( \u03bd), e.g., x y, x y:",
                "cite_spans": [
                    {
                        "start": 79,
                        "end": 95,
                        "text": "(Fernando, 2009)",
                        "ref_id": "BIBREF11"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Distinguishing Path and Manner Motion",
                "sec_num": "3.2"
            },
            {
                "text": "(6) \u03bd = df C? e i \u03bd \u2212\u2192 e i+1 . (7) loc(z) = x e 0 \u03bd \u2212\u2192 loc(z) = y 1 e 1 \u03bd \u2212\u2192 . . .",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Distinguishing Path and Manner Motion",
                "sec_num": "3.2"
            },
            {
                "text": "This now provides us with our dynamic interpretation of directed manner of motion verbs, such as slide, swim, roll, where we have an iteration of assignments of locations, undistinguished except that the values are order-preserving according to a scalar constraint. This is quite different from the dynamic interpretation of path predicates. Following (Galton, 2004; Pustejovsky and Moszkowicz, 2011) , path predicates such as arrive and leave make reference to a \"distinguished location\", not an arbitrary location. For example, the ball enters the room is satisified when the distinguished location, D, (the room) is successfully tested as the location for the moving object. That is, the location is tested against the current location for an object ((loc(x) = D)?), and retested until it is satisfied ((loc(x) = D)?).",
                "cite_spans": [
                    {
                        "start": 352,
                        "end": 366,
                        "text": "(Galton, 2004;",
                        "ref_id": "BIBREF15"
                    },
                    {
                        "start": 367,
                        "end": 400,
                        "text": "Pustejovsky and Moszkowicz, 2011)",
                        "ref_id": "BIBREF36"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "loc(z) = y n en",
                "sec_num": null
            },
            {
                "text": "(8) (loc(x) =D)? loc(z) = x e 0 \u03bd \u2212\u2192 (loc(x) =D)? loc(z) = y 1 e 1 \u03bd \u2212\u2192 . . . (loc(x)=D)? loc(z) = y n en",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "loc(z) = y n en",
                "sec_num": null
            },
            {
                "text": "While beyond the scope of the present discussion, it is worth noting that the model of event structure adopted here for motion verbs fits well with most of the major semantic and syntactic phenomena associated with event classes and Aktionsarten. 3",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "loc(z) = y n en",
                "sec_num": null
            },
            {
                "text": "Given the formal distinction between path and manner predicates as described above, let us examine how to differentiate meaning within the manner class. Levin (1993) differentiates this class in terms of argument alternation patterns, and identifies the following verb groupings: ROLL, RUN, EPONYMOUS VEHICLE, WALTZ, ACCOM-PANY, and CHASE verbs. While suggestive, these distinctions are only partially useful towards actually teasing apart the semantic dimensions along which we identify the contributing factors of manner.",
                "cite_spans": [
                    {
                        "start": 153,
                        "end": 165,
                        "text": "Levin (1993)",
                        "ref_id": "BIBREF28"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "Mani and Pustejovsky (2012) suggest a different strategy involving the identification of semantic parameters that clearly differentiate verb senses from each other within this class. One parameter exploited quite extensively within the motion class involves the mereotopological contraints that inhere throughout the movement of the object (Randell et al., 1992; Asher and Vieu, 1995; Galton, 2000) . Using this parameter, we are able to distinguish several of Levin's classes of manner 3 Cf. (Pustejovsky, 2013a) and (Krifka, 1992) .",
                "cite_spans": [
                    {
                        "start": 340,
                        "end": 362,
                        "text": "(Randell et al., 1992;",
                        "ref_id": "BIBREF42"
                    },
                    {
                        "start": 363,
                        "end": 384,
                        "text": "Asher and Vieu, 1995;",
                        "ref_id": "BIBREF0"
                    },
                    {
                        "start": 385,
                        "end": 398,
                        "text": "Galton, 2000)",
                        "ref_id": "BIBREF14"
                    },
                    {
                        "start": 487,
                        "end": 488,
                        "text": "3",
                        "ref_id": null
                    },
                    {
                        "start": 493,
                        "end": 513,
                        "text": "(Pustejovsky, 2013a)",
                        "ref_id": "BIBREF40"
                    },
                    {
                        "start": 518,
                        "end": 532,
                        "text": "(Krifka, 1992)",
                        "ref_id": "BIBREF26"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "as well as some novel ones, as described in (9), where a class is defined by the constraints that hold throughout the event (where EC is \"externally connected\", and DC is \"disconnected\"). For example, consider the semantic distinction between the verbs slide and hop or bounce. When the latter are employed in induced (directed) motion constructions (Levin, 1993; Jackendoff, 1996) , they take on the meaning of manner of motion verbs. Distinguishing between a sliding and hopping motion involves inspecting the next-state content in the motion n-gram: namely, there is a continuous satisfaction of EC(F,G) throughout the motion for slide and a toggling effect (on-off) for the predicates bounce and hop, as shown in (10). Distinguishing between a sliding motion and a rolling motion is also fairly straightforward. We have the entailments that result from each kind of motion, given a set of initial conditions, as in the following short sentence describing the motion of a ball relative to a floor (the domain for our event simulations).",
                "cite_spans": [
                    {
                        "start": 350,
                        "end": 363,
                        "text": "(Levin, 1993;",
                        "ref_id": "BIBREF28"
                    },
                    {
                        "start": 364,
                        "end": 381,
                        "text": "Jackendoff, 1996)",
                        "ref_id": "BIBREF24"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "\u2022 The ball slid.: At the termination of the action, object ball has moved relative to a surface in a manner that is [+translate] .",
                "cite_spans": [
                    {
                        "start": 116,
                        "end": 128,
                        "text": "[+translate]",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "That is, the movement is a translocation across absolute space, but other attributes (such as the ball's orientation) do not change.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "\u2022 The ball rolled.: At the termination of the action, object ball has moved relative to a surface in a manner that is [+translate]",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "and [+rotate] . Here, the translocation across space is preserved, with the addition of an orientation change.",
                "cite_spans": [
                    {
                        "start": 4,
                        "end": 13,
                        "text": "[+rotate]",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "We can further decompose these features, casting the [+translate] in terms of the translation's dimensionality. For both the ball slid and the ball rolled, it is required that the ball remain in the contact with the relevant surface, thus we can enforce a [-3-dimensional] constraint on the [+translate] feature. Thus, we arrive at the following differentiating semantic constraints for these verbs: (a) slide,",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "[+translate], [-3-dimensional]; (b) roll, [+translate], [-3-dimensional], [+rotate]",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": ". This is illustrated below over three states of execution. In our approach to conceptual modeling, we hypothesize that between the members of any pair of motion verbs, there exists at least one distinctive feature of physical motion that distinguishes the two predicates. While this may be too strong, it is helpful in our use of simulations for debugging the lexical semantics of linguistic expressions. 5 In order to quantify the qualitative distinctions between motion predicates and identify the precise primitive components of a motion verb, we build a real-time simulation, within which the individual features of a single motion verb can be defined and isolated in three-dimensional space.",
                "cite_spans": [
                    {
                        "start": 406,
                        "end": 407,
                        "text": "5",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "The idea of constructing simulations from linguistic utterances is, of course, not new. There are two groups of researchers who have developed related ideas quite extensively: simulation theorists, working in the philosophy of mind, such as Alvin Goldman and Robert Gordon; and cognitive scientists and linguists, such as Jerry Feldman, Ron Langacker, and Ben Bergen. According to Goldman (1989) , simulation provides a processdriven theory of mind and mental attribution, differing from the theory-driven models proposed by Churchland and others (Churchland, 1991) . From the cognitive linguistics tradition, simulation semantics has come to denote the mental instantiation of an interpretation of any linguistic utterance (Feldman, 2006; Bergen et al., 2007; Bergen, 2012) . While these communities do not seem to reference each other, it is clear from our perspective, that they are both pursuing similar programs, where distinct linguistic utterances correspond to generated models that have differentiated structures and behaviors (Narayanan, 1999; Siskind, 2011; Goldman, 2006) .",
                "cite_spans": [
                    {
                        "start": 381,
                        "end": 395,
                        "text": "Goldman (1989)",
                        "ref_id": "BIBREF17"
                    },
                    {
                        "start": 547,
                        "end": 565,
                        "text": "(Churchland, 1991)",
                        "ref_id": "BIBREF4"
                    },
                    {
                        "start": 724,
                        "end": 739,
                        "text": "(Feldman, 2006;",
                        "ref_id": "BIBREF10"
                    },
                    {
                        "start": 740,
                        "end": 760,
                        "text": "Bergen et al., 2007;",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 761,
                        "end": 774,
                        "text": "Bergen, 2012)",
                        "ref_id": "BIBREF2"
                    },
                    {
                        "start": 1036,
                        "end": 1053,
                        "text": "(Narayanan, 1999;",
                        "ref_id": "BIBREF34"
                    },
                    {
                        "start": 1054,
                        "end": 1068,
                        "text": "Siskind, 2011;",
                        "ref_id": "BIBREF45"
                    },
                    {
                        "start": 1069,
                        "end": 1083,
                        "text": "Goldman, 2006)",
                        "ref_id": "BIBREF18"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Mereotopological Distinctions in Manner",
                "sec_num": "3.3"
            },
            {
                "text": "The approach to simulation construction introduced in the previous section is inspired by work in minimal model generation (Blackburn and Bos, 2008; Konrad, 2004) . Type satisfaction in the compositional process mirrors the theorem proving component, while construction of the specific model helps us distinguish what is inherent in the different manner of motion events. This latter aspect is the \"positive handle\", (Blackburn and Bos, 2008) which demonstrates the informativeness of a distinction in our simulation.",
                "cite_spans": [
                    {
                        "start": 123,
                        "end": 148,
                        "text": "(Blackburn and Bos, 2008;",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 149,
                        "end": 162,
                        "text": "Konrad, 2004)",
                        "ref_id": "BIBREF25"
                    },
                    {
                        "start": 417,
                        "end": 442,
                        "text": "(Blackburn and Bos, 2008)",
                        "ref_id": "BIBREF3"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Simulations as Minimal Models",
                "sec_num": "4"
            },
            {
                "text": "Simulation software must be able to map a predicate to a known behavior, its arguments to objects in the scene, and then prompt those objects to execute the behavior. A simple input sentence needs to be tagged and parsed and transformed into predicate/argument representation, and from there into a dynamic event structure, as in (Pustejovsky and Moszkowicz, 2011) . The event structure is interpreted as the transformation executed over the object or objects in each frame, and then rendered. We currently use only proper names to refer to objects in the scene, to simplify model generation, hence Ball1 and Floor. This facilitates easy object identification in this prototype development stage.",
                "cite_spans": [
                    {
                        "start": 330,
                        "end": 364,
                        "text": "(Pustejovsky and Moszkowicz, 2011)",
                        "ref_id": "BIBREF36"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Simulations as Minimal Models",
                "sec_num": "4"
            },
            {
                "text": "Given a tagged and dependency parsed sentence, we can the transform the parse into a predicate formula, using the root of the parse as the predicate, the subject as a singleton first argument, and all objects as an optional stack of subsequent arguments. The DITL expression forms the basis of the coded behavior. The first two initialization steps are coded into the behavior's start function while the the third, Kleene iterated step, is encoded in the behavior's update function.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Simulations as Minimal Models",
                "sec_num": "4"
            },
            {
                "text": "We use the freely-available game engine, Unity, (Goldstone, 2009) to handle all underlying graphics processing, and limited our object library to simple primitive shapes of spheroids, rectangular prisms, and planes. For every instance of an object, the game engine maintains a data structure for the object's virtual representation. Table 4 shows the data structure for Entity, the superclass of all movable objects. The position and scale of the object are represented as 3-vectors of floating point numbers. The rotation is represented as the Euler angles of the object's current rotation, also a 3-vector. This 3-vector is computed as a quaternion for rendering purposes. The transform matrix composes the position, scale, and quaternion rotation into the complete transformation applied to the object at any given frame. The geometry is a mesh. The points, edges, faces, and texture attributes that comprise the mesh are all immutable at the moment so the mesh type is considered atomic for our purposes. The collider contains the coordinates of the center of the object, minimum and maximum extents of the object's boundaries, and radius of the boundaries (for spherical objects).",
                "cite_spans": [
                    {
                        "start": 48,
                        "end": 65,
                        "text": "(Goldstone, 2009)",
                        "ref_id": "BIBREF19"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 333,
                        "end": 340,
                        "text": "Table 4",
                        "ref_id": "TABREF3"
                    }
                ],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "Behaviors can only be executed over Entity instances, so we also provide each one with a currentBehavior property, referencing the code to be executed over the object every frame that said behavior is being run. This code performs a transformation over the object at every step, generating a new state in a dynamic model of the event denoted by the a given predicate. Thus, the event 6 is decomposed into frame-by-frame transformations representing the \u03bd-transition from Section 3.2.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "We generate example simulations of behaviors in a sample environment, shown in Figure 4 , that consists of a sealed four-walled room that contains a number of primitive objects. The behaviors currently coded into our software map directly from DITL to the simulation. The various parts of the DITL formula that describes a given behavior are coded into the behavior's start or update functions in Unity. Below is one such C# code snippet: the per-frame transformation for roll. This \"translates\" the DITL expression (y := w; y",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 79,
                        "end": 87,
                        "text": "Figure 4",
                        "ref_id": "FIGREF7"
                    }
                ],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "= w; d(b loc ,y) < d(b loc ,w)) + , (z := v; z = v; z-b rot < v-b rot )",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": ",IN(y,Floor) + while explicitly calculating the value of the precise differences in location and rotation between each frame or time step. The variables moveSpeed, rotSpeed and radius are given explicit value. deltaTime refers to the time elapsed between frames.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "Translating a DITL formula into executable code makes evident the differences in minimal verb pairs, such as the ball (or box) rolled and the ball (or box) slid. When an object rolls, one area on the object must remain in contact with the supporting surface, and that area must be adjacent to the area contacting the surface in the previous time step. When an object slides, the same area on the object must contact the supporting surface. Compare the per-frame transformation for slide below to the given transformation for roll. This maps the DITL expression (y := w; y = w; d(b loc ,y) < d(b loc ,w),IN(y,Floor)) + . Here, the object's location changes along a path leading away from the start location, but does not rotate as in roll.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "DITL expressions and their coded equivalents can also be composed into new, more specific motions. The cross formula from Section 4 can be composed with that for roll to describe a \"roll across\" motion.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "In a model, a path verb such as cross does not necessarily need an explicit manner of motion specified. In a simulation, the manner needs to be given a value, requiring the composition of the path verb (e.g., cross) with one of a certain subsets of manner verbs specifying how the object moves relative to the supporting surface. Below are DITL expressions and code implementations for two cross predicates, the first a cross motion while sliding, the second a cross motion while rolling. At each frame, the distance between the object's current position and its previously computed destination is computed again, and the update moves the object away from its current position (d(b,y) < d(b,w)) toward the destination (d(b,z) > d(z,w)).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "Since no other manner of motion is specified, the object does not turn or rotate as it moves, but simply \"slides.\" Here the update is the same as above, but with the introduction of the rolling motion. In both code snippets, the non-changing value of transform.position.y implicitly maps the IN RCC condition in the DITL formulas, and keeps the moving object attached to the floor. If there exists a behavior corresponding to the predicate (by name) on an entity bearing the name of the predicate's first (subject) argument, the transformation encoded in that behavior is performed over the entity until an end condition specific to the behavior is met. The resulting animated motion depicts the manner of motion denoted by the predicate. Given a predicate of arity greater than 1, the simulator tries to prompt a behavior on the first argument that can be run using parameters of the subsequent arguments.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "A cross behavior, for example, divides the supporting surface into regions and attempts to move the crossing object from one region to the the opposite region. In figure 5, the bounds of Floor completely surround the bounds of Ball2 (IN(Ball2,Floor) in RCC8). This configuration makes it possible for the simulation to compute a motion moving the Ball2 object from one side of the Floor to the other.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "The left side of figure 5 shows a ball rolling and a box sliding, a depiction of two predicates: Box1 slid and Ball1 rolled. The right side depicts Ball2 crossed Floor (from the rear center to the front center). The starting state of each scene is overlaid semi-transparently while the in-progress state is fully opaque.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Generating Simulations",
                "sec_num": "5"
            },
            {
                "text": "In this paper, we describe a model for mapping natural language motion expressions into a 3D simulation environment. Our strategy has been to use minimal simulation generation as a conceptual debugging tool, in order to tease out the semantic differences between linguistic expressions; specifically those between verbs that are members of conventionally homogeneous classes, according to linguistic analysis.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussion and Conclusion",
                "sec_num": "6"
            },
            {
                "text": "It should be pointed out that our goal is different from WordsEye (Coyne and Sproat, 2001 ). While we are interested in using simulation generation to differentiate semantic distinctions in both lexical classes and compositional constructions, the goal behind WordsEye is to provide an enhanced interface to allow non-specialists create 3D scenes without being familiar with special software models for everyday objects and relations. There are obvious synergies between these two goals that can be pursued.",
                "cite_spans": [
                    {
                        "start": 66,
                        "end": 89,
                        "text": "(Coyne and Sproat, 2001",
                        "ref_id": "BIBREF5"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussion and Conclusion",
                "sec_num": "6"
            },
            {
                "text": "The simulations we create provide an interpretation of the given motion predicate over the given entity, but not the only interpretation. Just as Coyne et al. (2010) does for static objects in the WordsEye system, we must apply some implicit constraints to our motion predicates to allow them to be visually simulated. For instance, in the roll and slide examples given in Figure 5 , both objects are moving in the same direction-parallel to the back wall of the room object. Had the objects been moving perpendicular to the back wall or in any other direction, as long as they remained in con-tact with the floor at all times, the simulated motion would still be considered a \"roll\" (if rotating around an axis parallel to the floor), or a \"slide\" (if not), regardless of what the precise direction of motion is. Minimal pairs in a model have to be compared and contrasted in a discriminative way, and thus in modeling a slide predicate versus a roll predicate, knowing that the distinction is one of rotation parallel to the surface is enough to distinguish the two predicates in a model. In a simulation, the discriminative process requires that the two contrasting behaviors look different, and as such, the simulation software must be able to completely render a scene for each frame from behavior start to behavior finish, and so every variable for every object being rendered must have an assigned value, including the position of the object from frame to frame. If these values are left unspecified, the software either fails to compile or throws an exception. Thus, we are forced to arbitrarily choose a direction of motion (as well as direction of rotation, speed of rotation, speed of motion, etc.). As long as all nonchanging variables are kept consistent between a minimal pair of behaviors, we can evaluate the quantitative and qualitative differences between the values that do change. As simulations require values to be assigned to variables that can be left unspecified in an ordinary modeling process, simulations expose presuppositions about the semantics of motion verbs and of compositions that would not be necessary in a model alone.",
                "cite_spans": [
                    {
                        "start": 146,
                        "end": 165,
                        "text": "Coyne et al. (2010)",
                        "ref_id": "BIBREF6"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 373,
                        "end": 381,
                        "text": "Figure 5",
                        "ref_id": "FIGREF12"
                    }
                ],
                "eq_spans": [],
                "section": "Discussion and Conclusion",
                "sec_num": "6"
            },
            {
                "text": "In order to evaluate the appropriateness of a given simulation, we are currently experimenting with a strategy often used in classification and annotation tasks, namely pairwise similarity judgments (Rumshisky et al., 2012; . This involves presenting a user with a simple discrimination task that has a reduced cognitive load, comparing the similarity of the example to the target instances. In the present context, a subject is shown a specific simulation resulting from the translation from textual input, through DITL, to the visualization. A set of activity or event descriptions is given, and the subject is then asked to select which best describes the simulation shown; e.g., \"Is this a sliding?\", \"Is this a rolling?\". The results of this experiment are presently being evaluated.",
                "cite_spans": [
                    {
                        "start": 199,
                        "end": 223,
                        "text": "(Rumshisky et al., 2012;",
                        "ref_id": "BIBREF44"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussion and Conclusion",
                "sec_num": "6"
            },
            {
                "text": "The system is currently in the prototype stage and needs to be expanded in three main areas: ob-ject library, parsing pipeline, and predicate handling. Our object and behavior libraries are currently limited to geometric primitives and the motions that can be applied over them. While roll, slide, and cross behaviors can be scripted for spheres and cubes and shapes derived from them, a predicate like walk cannot be supported on the current infrastructure. Thus, we intend to expand the object library to include more complex inanimate objects (tables, chairs, or other household objects) as well as animate objects. Having an object library containing forms capable of executing greater numbers of predicates will allow us to implement those predicates.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussion and Conclusion",
                "sec_num": "6"
            },
            {
                "text": "The parsing pipeline described in Section 4 is only partially implemented, with the only completed parts being the latter stages, relating a formulas to a scripted behavior and its arguments. We intend to expand the parsing pipeline to include all the steps described in this paper: taking input as a simple natural language sentence, tagging and parsing it to extract the constituent parts of a predicate/argument representation, and using that output to prompt a behavior in software as a dynamic event structure. More robust parsing will afford us the opportunity to expand the diversity of predicates that the software can handle as well (Mc-Donald and Pustejovsky, 2014) . While currently limited to unary and binary predicates, we need to extend the capability to ternary predicates and predicates of greater arity, including the use of adjunct phrases and indirect objects. We are in the process of developing an implementation that uses Boxer (Curran et al., 2007 ) so that we can create first-order models from the dynamic expressions used here.",
                "cite_spans": [
                    {
                        "start": 642,
                        "end": 675,
                        "text": "(Mc-Donald and Pustejovsky, 2014)",
                        "ref_id": null
                    },
                    {
                        "start": 951,
                        "end": 971,
                        "text": "(Curran et al., 2007",
                        "ref_id": "BIBREF7"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussion and Conclusion",
                "sec_num": "6"
            },
            {
                "text": "This is consistent with the approach developed in(Fernando, 2009;Fernando, 2013). This approach to a dynamic interpretation of change in language semantics is also inspired bySteedman (2002).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "Cf.Groenendijk and Stokhof (1990) for dynamic updating, and Naumann (2001) for a related analysis.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "Many natural language predicates require reference to at least three states. These include the semelfactives mentioned above, as well as blink and iterative uses of knock and clap(Vendler, 1967;Dowty, 1979;Rothstein, 2008).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "Obviously, true synonyms in the lexicon would not be distinguishable in a model.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "These events are linguistic events, and not the same as \"events\" as used in software development or with event handlers.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            }
        ],
        "back_matter": [
            {
                "text": "We would like to thank David McDonald for comments and discussion. We would also like to thank the reviewers for several substantial suggestions for improvements and clarifications to the paper. All remaining errors are of course the responsibilities of the authors. This work was supported in part by the Department of the Navy, Office of Naval Research under grant N00014-13-1-0228. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the Office of Naval Research.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Acknowledgements",
                "sec_num": null
            }
        ],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "Towards a geometry of common sense: a semantics and a complete axiomatisation of merotopology",
                "authors": [
                    {
                        "first": "Nicholas",
                        "middle": [],
                        "last": "Asher",
                        "suffix": ""
                    },
                    {
                        "first": "Laure",
                        "middle": [],
                        "last": "Vieu",
                        "suffix": ""
                    }
                ],
                "year": 1995,
                "venue": "Proceedings of IJCAI95",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Nicholas Asher and Laure Vieu. 1995. Towards a ge- ometry of common sense: a semantics and a com- plete axiomatisation of merotopology. In Proceed- ings of IJCAI95, Montreal, Canada.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Spatial and linguistic aspects of visual imagery in sentence comprehension",
                "authors": [
                    {
                        "first": "K",
                        "middle": [],
                        "last": "Benjamin",
                        "suffix": ""
                    },
                    {
                        "first": "Shane",
                        "middle": [],
                        "last": "Bergen",
                        "suffix": ""
                    },
                    {
                        "first": "Teenie",
                        "middle": [],
                        "last": "Lindsay",
                        "suffix": ""
                    },
                    {
                        "first": "Srini",
                        "middle": [],
                        "last": "Matlock",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Narayanan",
                        "suffix": ""
                    }
                ],
                "year": 2007,
                "venue": "Cognitive Science",
                "volume": "31",
                "issue": "5",
                "pages": "733--764",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Benjamin K. Bergen, Shane Lindsay, Teenie Matlock, and Srini Narayanan. 2007. Spatial and linguistic aspects of visual imagery in sentence comprehen- sion. Cognitive Science, 31(5):733-764.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "Louder than words: The new science of how the mind makes meaning",
                "authors": [
                    {
                        "first": "K",
                        "middle": [],
                        "last": "Benjamin",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Bergen",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Benjamin K Bergen. 2012. Louder than words: The new science of how the mind makes meaning. Basic Books.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "Computational semantics",
                "authors": [
                    {
                        "first": "Patrick",
                        "middle": [],
                        "last": "Blackburn",
                        "suffix": ""
                    },
                    {
                        "first": "Johan",
                        "middle": [],
                        "last": "Bos",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "THEORIA. An International Journal for Theory, History and Foundations of Science",
                "volume": "18",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Patrick Blackburn and Johan Bos. 2008. Computa- tional semantics. THEORIA. An International Jour- nal for Theory, History and Foundations of Science, 18(1).",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "Folk psychology and the explanation of human behavior. The future of folk psychology: Intentionality and cognitive science",
                "authors": [
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Paul",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Churchland",
                        "suffix": ""
                    }
                ],
                "year": 1991,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "51--69",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Paul M Churchland. 1991. Folk psychology and the explanation of human behavior. The future of folk psychology: Intentionality and cognitive science, pages 51-69.",
                "links": null
            },
            "BIBREF5": {
                "ref_id": "b5",
                "title": "Wordseye: an automatic text-to-scene conversion system",
                "authors": [
                    {
                        "first": "Bob",
                        "middle": [],
                        "last": "Coyne",
                        "suffix": ""
                    },
                    {
                        "first": "Richard",
                        "middle": [],
                        "last": "Sproat",
                        "suffix": ""
                    }
                ],
                "year": 2001,
                "venue": "Proceedings of the 28th annual conference on Computer graphics and interactive techniques",
                "volume": "",
                "issue": "",
                "pages": "487--496",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Bob Coyne and Richard Sproat. 2001. Wordseye: an automatic text-to-scene conversion system. In Pro- ceedings of the 28th annual conference on Computer graphics and interactive techniques, pages 487-496. ACM.",
                "links": null
            },
            "BIBREF6": {
                "ref_id": "b6",
                "title": "Frame semantics in text-toscene generation",
                "authors": [
                    {
                        "first": "Bob",
                        "middle": [],
                        "last": "Coyne",
                        "suffix": ""
                    },
                    {
                        "first": "Owen",
                        "middle": [],
                        "last": "Rambow",
                        "suffix": ""
                    },
                    {
                        "first": "Julia",
                        "middle": [],
                        "last": "Hirschberg",
                        "suffix": ""
                    },
                    {
                        "first": "Richard",
                        "middle": [],
                        "last": "Sproat",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Knowledge-Based and Intelligent Information and Engineering Systems",
                "volume": "",
                "issue": "",
                "pages": "375--384",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Bob Coyne, Owen Rambow, Julia Hirschberg, and Richard Sproat. 2010. Frame semantics in text-to- scene generation. In Knowledge-Based and Intel- ligent Information and Engineering Systems, pages 375-384. Springer.",
                "links": null
            },
            "BIBREF7": {
                "ref_id": "b7",
                "title": "Linguistically motivated large-scale nlp with c&c and boxer",
                "authors": [
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Curran",
                        "suffix": ""
                    },
                    {
                        "first": "Stephen",
                        "middle": [],
                        "last": "Clark",
                        "suffix": ""
                    },
                    {
                        "first": "Johan",
                        "middle": [],
                        "last": "Bos",
                        "suffix": ""
                    }
                ],
                "year": 2007,
                "venue": "Proceedings of the 45th Annual Meeting of the Association for Computational Linguistics Companion Volume Proceedings of the Demo and Poster Sessions",
                "volume": "",
                "issue": "",
                "pages": "33--36",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "James Curran, Stephen Clark, and Johan Bos. 2007. Linguistically motivated large-scale nlp with c&c and boxer. In Proceedings of the 45th Annual Meet- ing of the Association for Computational Linguis- tics Companion Volume Proceedings of the Demo and Poster Sessions, pages 33-36, Prague, Czech Republic, June. Association for Computational Lin- guistics.",
                "links": null
            },
            "BIBREF8": {
                "ref_id": "b8",
                "title": "Word meaning and Montague grammar: The semantics of verbs and times in generative semantics and in Montague's PTQ",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "David R Dowty",
                        "suffix": ""
                    }
                ],
                "year": 1979,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "David R Dowty. 1979. Word meaning and Mon- tague grammar: The semantics of verbs and times in generative semantics and in Montague's PTQ, vol- ume 7. Springer.",
                "links": null
            },
            "BIBREF9": {
                "ref_id": "b9",
                "title": "Representing simple trajectories as oriented curves",
                "authors": [
                    {
                        "first": "C",
                        "middle": [],
                        "last": "Eschenbach",
                        "suffix": ""
                    },
                    {
                        "first": "C",
                        "middle": [],
                        "last": "Habel",
                        "suffix": ""
                    },
                    {
                        "first": "L",
                        "middle": [],
                        "last": "Kulik",
                        "suffix": ""
                    }
                ],
                "year": 1999,
                "venue": "FLAIRS-99, Proceedings of the 12th International Florida AI Research Society Conference",
                "volume": "",
                "issue": "",
                "pages": "431--436",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "C. Eschenbach, C. Habel, L. Kulik, et al. 1999. Rep- resenting simple trajectories as oriented curves. In FLAIRS-99, Proceedings of the 12th International Florida AI Research Society Conference, pages 431- 436.",
                "links": null
            },
            "BIBREF10": {
                "ref_id": "b10",
                "title": "From molecule to metaphor: A neural theory of language",
                "authors": [
                    {
                        "first": "Jerome",
                        "middle": [],
                        "last": "Feldman",
                        "suffix": ""
                    }
                ],
                "year": 2006,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Jerome Feldman. 2006. From molecule to metaphor: A neural theory of language. MIT press.",
                "links": null
            },
            "BIBREF11": {
                "ref_id": "b11",
                "title": "Situations in ltl as strings. Information and Computation",
                "authors": [
                    {
                        "first": "Tim",
                        "middle": [],
                        "last": "Fernando",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "",
                "volume": "207",
                "issue": "",
                "pages": "980--999",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Tim Fernando. 2009. Situations in ltl as strings. Infor- mation and Computation, 207(10):980-999.",
                "links": null
            },
            "BIBREF12": {
                "ref_id": "b12",
                "title": "Segmenting temporal intervals for tense and aspect",
                "authors": [
                    {
                        "first": "Tim",
                        "middle": [],
                        "last": "Fernando",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "The 13th Meeting on the Mathematics of Language",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Tim Fernando. 2013. Segmenting temporal intervals for tense and aspect. In The 13th Meeting on the Mathematics of Language, page 30.",
                "links": null
            },
            "BIBREF13": {
                "ref_id": "b13",
                "title": "Using orientation information for qualitative spatial reasoning",
                "authors": [
                    {
                        "first": "Christian",
                        "middle": [],
                        "last": "Freksa",
                        "suffix": ""
                    }
                ],
                "year": 1992,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Christian Freksa. 1992. Using orientation information for qualitative spatial reasoning. Springer.",
                "links": null
            },
            "BIBREF14": {
                "ref_id": "b14",
                "title": "Qualitative Spatial Change",
                "authors": [
                    {
                        "first": "Antony",
                        "middle": [],
                        "last": "Galton",
                        "suffix": ""
                    }
                ],
                "year": 2000,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Antony Galton. 2000. Qualitative Spatial Change. Oxford University Press, Oxford.",
                "links": null
            },
            "BIBREF15": {
                "ref_id": "b15",
                "title": "Fields and objects in space, time, and space-time",
                "authors": [
                    {
                        "first": "Antony",
                        "middle": [],
                        "last": "Galton",
                        "suffix": ""
                    }
                ],
                "year": 2004,
                "venue": "Spatial Cognition and Computation",
                "volume": "4",
                "issue": "1",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Antony Galton. 2004. Fields and objects in space, time, and space-time. Spatial Cognition and Com- putation, 4(1).",
                "links": null
            },
            "BIBREF16": {
                "ref_id": "b16",
                "title": "Interpreting definites using model generation",
                "authors": [
                    {
                        "first": "Claire",
                        "middle": [],
                        "last": "Gardent",
                        "suffix": ""
                    },
                    {
                        "first": "Karsten",
                        "middle": [],
                        "last": "Konrad",
                        "suffix": ""
                    }
                ],
                "year": 2000,
                "venue": "Journal of Language and Computation",
                "volume": "1",
                "issue": "2",
                "pages": "193--209",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Claire Gardent and Karsten Konrad. 2000. Interpreting definites using model generation. Journal of Lan- guage and Computation, 1(2):193-209.",
                "links": null
            },
            "BIBREF17": {
                "ref_id": "b17",
                "title": "Interpretation psychologized*. Mind & Language",
                "authors": [
                    {
                        "first": "Alvin I",
                        "middle": [],
                        "last": "Goldman",
                        "suffix": ""
                    }
                ],
                "year": 1989,
                "venue": "",
                "volume": "4",
                "issue": "",
                "pages": "161--185",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Alvin I Goldman. 1989. Interpretation psycholo- gized*. Mind & Language, 4(3):161-185.",
                "links": null
            },
            "BIBREF18": {
                "ref_id": "b18",
                "title": "Simulating minds: The philosophy, psychology, and neuroscience of mindreading",
                "authors": [
                    {
                        "first": "Alvin I",
                        "middle": [],
                        "last": "Goldman",
                        "suffix": ""
                    }
                ],
                "year": 2006,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Alvin I Goldman. 2006. Simulating minds: The phi- losophy, psychology, and neuroscience of mindread- ing. Oxford University Press.",
                "links": null
            },
            "BIBREF19": {
                "ref_id": "b19",
                "title": "Unity Game Development Essentials",
                "authors": [
                    {
                        "first": "Will",
                        "middle": [],
                        "last": "Goldstone",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Will Goldstone. 2009. Unity Game Development Es- sentials. Packt Publishing Ltd.",
                "links": null
            },
            "BIBREF20": {
                "ref_id": "b20",
                "title": "Dynamic predicate logic",
                "authors": [
                    {
                        "first": "Jeroen",
                        "middle": [],
                        "last": "Groenendijk",
                        "suffix": ""
                    },
                    {
                        "first": "Martin",
                        "middle": [],
                        "last": "Stokhof",
                        "suffix": ""
                    }
                ],
                "year": 1990,
                "venue": "Linguistics and Philosophy",
                "volume": "14",
                "issue": "",
                "pages": "39--100",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Jeroen Groenendijk and Martin Stokhof. 1990. Dy- namic predicate logic. Linguistics and Philosophy, 14:39-100.",
                "links": null
            },
            "BIBREF21": {
                "ref_id": "b21",
                "title": "Dynamic Logic",
                "authors": [
                    {
                        "first": "David",
                        "middle": [],
                        "last": "Harel",
                        "suffix": ""
                    },
                    {
                        "first": "Dexter",
                        "middle": [],
                        "last": "Kozen",
                        "suffix": ""
                    },
                    {
                        "first": "Jerzy",
                        "middle": [],
                        "last": "Tiuyn",
                        "suffix": ""
                    }
                ],
                "year": 2000,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "David Harel, Dexter Kozen, and Jerzy Tiuyn. 2000. Dynamic Logic. The MIT Press, 1st edition.",
                "links": null
            },
            "BIBREF22": {
                "ref_id": "b22",
                "title": "Dynamic logic",
                "authors": [
                    {
                        "first": "David",
                        "middle": [],
                        "last": "Harel",
                        "suffix": ""
                    }
                ],
                "year": 1984,
                "venue": "Handbook of Philosophical Logic",
                "volume": "II",
                "issue": "",
                "pages": "497--604",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "David Harel. 1984. Dynamic logic. In M. Gabbay and F. Gunthner, editors, Handbook of Philosophi- cal Logic, Volume II: Extensions of Classical Logic, page 497?604. Reidel.",
                "links": null
            },
            "BIBREF23": {
                "ref_id": "b23",
                "title": "Semantics and Cognition",
                "authors": [
                    {
                        "first": "Ray",
                        "middle": [],
                        "last": "Jackendoff",
                        "suffix": ""
                    }
                ],
                "year": 1983,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ray Jackendoff. 1983. Semantics and Cognition. MIT Press.",
                "links": null
            },
            "BIBREF24": {
                "ref_id": "b24",
                "title": "The proper treatment of measuring out, telicity, and perhaps even quantification in english",
                "authors": [
                    {
                        "first": "Ray",
                        "middle": [],
                        "last": "Jackendoff",
                        "suffix": ""
                    }
                ],
                "year": 1996,
                "venue": "Natural Language & Linguistic Theory",
                "volume": "14",
                "issue": "2",
                "pages": "305--354",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ray Jackendoff. 1996. The proper treatment of mea- suring out, telicity, and perhaps even quantification in english. Natural Language & Linguistic Theory, 14(2):305-354.",
                "links": null
            },
            "BIBREF25": {
                "ref_id": "b25",
                "title": "Model generation for natural language interpretation and analysis",
                "authors": [
                    {
                        "first": "Karsten",
                        "middle": [
                            "Konrad"
                        ],
                        "last": "",
                        "suffix": ""
                    }
                ],
                "year": 2004,
                "venue": "",
                "volume": "2953",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Karsten Konrad. 2004. Model generation for natural language interpretation and analysis, volume 2953. Springer.",
                "links": null
            },
            "BIBREF26": {
                "ref_id": "b26",
                "title": "Thematic relations as links between nominal reference and temporal constitution. Lexical matters",
                "authors": [
                    {
                        "first": "Manfred",
                        "middle": [],
                        "last": "Krifka",
                        "suffix": ""
                    }
                ],
                "year": 1992,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Manfred Krifka. 1992. Thematic relations as links be- tween nominal reference and temporal constitution. Lexical matters, 2953.",
                "links": null
            },
            "BIBREF27": {
                "ref_id": "b27",
                "title": "The spatial semantic hierarchy",
                "authors": [
                    {
                        "first": "Benjamin",
                        "middle": [],
                        "last": "Kuipers",
                        "suffix": ""
                    }
                ],
                "year": 2000,
                "venue": "Artificial Intelligence",
                "volume": "119",
                "issue": "1",
                "pages": "191--233",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Benjamin Kuipers. 2000. The spatial semantic hierar- chy. Artificial Intelligence, 119(1):191-233.",
                "links": null
            },
            "BIBREF28": {
                "ref_id": "b28",
                "title": "English verb class and alternations: a preliminary investigation",
                "authors": [
                    {
                        "first": "Beth",
                        "middle": [],
                        "last": "Levin",
                        "suffix": ""
                    }
                ],
                "year": 1993,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Beth Levin. 1993. English verb class and alternations: a preliminary investigation. University of Chicago Press.",
                "links": null
            },
            "BIBREF29": {
                "ref_id": "b29",
                "title": "Space in Language and Cognition: Explorations in Cognitive Diversity. Language, culture, and cognition",
                "authors": [
                    {
                        "first": "S",
                        "middle": [
                            "C"
                        ],
                        "last": "Levinson",
                        "suffix": ""
                    }
                ],
                "year": 2003,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "S.C. Levinson. 2003. Space in Language and Cog- nition: Explorations in Cognitive Diversity. Lan- guage, culture, and cognition. Cambridge University Press.",
                "links": null
            },
            "BIBREF30": {
                "ref_id": "b30",
                "title": "Interpreting Motion: Grounded Representations for Spatial Language",
                "authors": [
                    {
                        "first": "Inderjeet",
                        "middle": [],
                        "last": "Mani",
                        "suffix": ""
                    },
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Inderjeet Mani and James Pustejovsky. 2012. Inter- preting Motion: Grounded Representations for Spa- tial Language. Oxford University Press.",
                "links": null
            },
            "BIBREF31": {
                "ref_id": "b31",
                "title": "On the representation of inferences and their lexicalization",
                "authors": [
                    {
                        "first": "David",
                        "middle": [],
                        "last": "Mcdonald",
                        "suffix": ""
                    },
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Advances in Cognitive Systems",
                "volume": "3",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "David McDonald and James Pustejovsky. 2014. On the representation of inferences and their lexicaliza- tion. In Advances in Cognitive Systems, volume 3.",
                "links": null
            },
            "BIBREF32": {
                "ref_id": "b32",
                "title": "Contextual correlates of semantic similarity. Language and Cognitive Processes",
                "authors": [
                    {
                        "first": "G",
                        "middle": [],
                        "last": "Miller",
                        "suffix": ""
                    },
                    {
                        "first": "W",
                        "middle": [],
                        "last": "Charles",
                        "suffix": ""
                    }
                ],
                "year": 1991,
                "venue": "",
                "volume": "6",
                "issue": "",
                "pages": "1--28",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "G. Miller and W. Charles. 1991. Contextual corre- lates of semantic similarity. Language and Cogni- tive Processes, 6(1):1-28.",
                "links": null
            },
            "BIBREF33": {
                "ref_id": "b33",
                "title": "Language and perception",
                "authors": [
                    {
                        "first": "A",
                        "middle": [],
                        "last": "George",
                        "suffix": ""
                    },
                    {
                        "first": "Philip N Johnson-Laird",
                        "middle": [],
                        "last": "Miller",
                        "suffix": ""
                    }
                ],
                "year": 1976,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "George A Miller and Philip N Johnson-Laird. 1976. Language and perception. Belknap Press.",
                "links": null
            },
            "BIBREF34": {
                "ref_id": "b34",
                "title": "Reasoning about actions in narrative understanding",
                "authors": [
                    {
                        "first": "Srinivas",
                        "middle": [],
                        "last": "Narayanan",
                        "suffix": ""
                    }
                ],
                "year": 1999,
                "venue": "IJCAI",
                "volume": "99",
                "issue": "",
                "pages": "350--357",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Srinivas Narayanan. 1999. Reasoning about actions in narrative understanding. IJCAI, 99:350-357.",
                "links": null
            },
            "BIBREF35": {
                "ref_id": "b35",
                "title": "Aspects of changes: a dynamic event semantics",
                "authors": [
                    {
                        "first": "Ralf",
                        "middle": [],
                        "last": "Naumann",
                        "suffix": ""
                    }
                ],
                "year": 2001,
                "venue": "Journal of semantics",
                "volume": "18",
                "issue": "",
                "pages": "27--81",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ralf Naumann. 2001. Aspects of changes: a dynamic event semantics. Journal of semantics, 18:27-81.",
                "links": null
            },
            "BIBREF36": {
                "ref_id": "b36",
                "title": "The qualitative spatial dynamics of motion",
                "authors": [
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    },
                    {
                        "first": "Jessica",
                        "middle": [],
                        "last": "Moszkowicz",
                        "suffix": ""
                    }
                ],
                "year": 2011,
                "venue": "The Journal of Spatial Cognition and Computation",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "James Pustejovsky and Jessica Moszkowicz. 2011. The qualitative spatial dynamics of motion. The Journal of Spatial Cognition and Computation.",
                "links": null
            },
            "BIBREF37": {
                "ref_id": "b37",
                "title": "Deep semantic annotation with shallow methods",
                "authors": [
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    },
                    {
                        "first": "Anna",
                        "middle": [],
                        "last": "Rumshisky",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "James Pustejovsky and Anna Rumshisky. 2014. Deep semantic annotation with shallow methods. LREC Tutorial, May.",
                "links": null
            },
            "BIBREF38": {
                "ref_id": "b38",
                "title": "Annotation of compositional operations with glml",
                "authors": [
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    },
                    {
                        "first": "Anna",
                        "middle": [],
                        "last": "Rumshisky",
                        "suffix": ""
                    },
                    {
                        "first": "Olga",
                        "middle": [],
                        "last": "Batiukova",
                        "suffix": ""
                    },
                    {
                        "first": "Jessica",
                        "middle": [],
                        "last": "Moszkowicz",
                        "suffix": ""
                    }
                ],
                "year": 2014,
                "venue": "Computing Meaning",
                "volume": "",
                "issue": "",
                "pages": "217--234",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "James Pustejovsky, Anna Rumshisky, Olga Batiukova, and Jessica Moszkowicz. 2014. Annotation of com- positional operations with glml. In Harry Bunt, ed- itor, Computing Meaning, pages 217-234. Springer Netherlands.",
                "links": null
            },
            "BIBREF39": {
                "ref_id": "b39",
                "title": "The Generative Lexicon. Bradford Book",
                "authors": [
                    {
                        "first": "J",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    }
                ],
                "year": 1995,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "J. Pustejovsky. 1995. The Generative Lexicon. Brad- ford Book. Mit Press.",
                "links": null
            },
            "BIBREF40": {
                "ref_id": "b40",
                "title": "Dynamic event structure and habitat theory",
                "authors": [
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Proceedings of the 6th International Conference on Generative Approaches to the Lexicon (GL2013)",
                "volume": "",
                "issue": "",
                "pages": "1--10",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "James Pustejovsky. 2013a. Dynamic event structure and habitat theory. In Proceedings of the 6th Inter- national Conference on Generative Approaches to the Lexicon (GL2013), pages 1-10. ACL.",
                "links": null
            },
            "BIBREF41": {
                "ref_id": "b41",
                "title": "Where things happen: On the semantics of event localization",
                "authors": [
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Proceedings of ISA-9: International Workshop on Semantic Annotation",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "James Pustejovsky. 2013b. Where things happen: On the semantics of event localization. In Proceedings of ISA-9: International Workshop on Semantic An- notation.",
                "links": null
            },
            "BIBREF42": {
                "ref_id": "b42",
                "title": "A spatial logic based on regions and connections",
                "authors": [
                    {
                        "first": "David",
                        "middle": [],
                        "last": "Randell",
                        "suffix": ""
                    },
                    {
                        "first": "Zhan",
                        "middle": [],
                        "last": "Cui",
                        "suffix": ""
                    },
                    {
                        "first": "Anthony",
                        "middle": [],
                        "last": "Cohn",
                        "suffix": ""
                    }
                ],
                "year": 1992,
                "venue": "Proceedings of the 3rd Internation Conference on Knowledge Representation and REasoning",
                "volume": "",
                "issue": "",
                "pages": "165--176",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "David Randell, Zhan Cui, and Anthony Cohn. 1992. A spatial logic based on regions and connections. In Morgan Kaufmann, editor, Proceedings of the 3rd Internation Conference on Knowledge Representa- tion and REasoning, pages 165-176, San Mateo.",
                "links": null
            },
            "BIBREF43": {
                "ref_id": "b43",
                "title": "Two puzzles for a theory of lexical aspect: Semelfactives and degree achievements. Event structures in linguistic form and interpretation",
                "authors": [
                    {
                        "first": "Susan",
                        "middle": [],
                        "last": "Rothstein",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "",
                "volume": "5",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Susan Rothstein. 2008. Two puzzles for a theory of lexical aspect: Semelfactives and degree achieve- ments. Event structures in linguistic form and in- terpretation, 5:175.",
                "links": null
            },
            "BIBREF44": {
                "ref_id": "b44",
                "title": "Word sense inventories by non-experts",
                "authors": [
                    {
                        "first": "Anna",
                        "middle": [],
                        "last": "Rumshisky",
                        "suffix": ""
                    },
                    {
                        "first": "Nick",
                        "middle": [],
                        "last": "Botchan",
                        "suffix": ""
                    },
                    {
                        "first": "Sophie",
                        "middle": [],
                        "last": "Kushkuley",
                        "suffix": ""
                    },
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "LREC",
                "volume": "",
                "issue": "",
                "pages": "4055--4059",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Anna Rumshisky, Nick Botchan, Sophie Kushkuley, and James Pustejovsky. 2012. Word sense inven- tories by non-experts. In LREC, pages 4055-4059.",
                "links": null
            },
            "BIBREF45": {
                "ref_id": "b45",
                "title": "Grounding the lexical semantics of verbs in visual perception using force dynamics and event logic",
                "authors": [
                    {
                        "first": "Jeffrey",
                        "middle": [
                            "Mark"
                        ],
                        "last": "Siskind",
                        "suffix": ""
                    }
                ],
                "year": 2011,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {
                    "arXiv": [
                        "arXiv:1106.0256"
                    ]
                },
                "num": null,
                "urls": [],
                "raw_text": "Jeffrey Mark Siskind. 2011. Grounding the lexi- cal semantics of verbs in visual perception using force dynamics and event logic. arXiv preprint arXiv:1106.0256.",
                "links": null
            },
            "BIBREF46": {
                "ref_id": "b46",
                "title": "Plans, affordances, and combinatory grammar",
                "authors": [
                    {
                        "first": "Mark",
                        "middle": [],
                        "last": "Steedman",
                        "suffix": ""
                    }
                ],
                "year": 2002,
                "venue": "Linguistics and Philosophy",
                "volume": "25",
                "issue": "5-6",
                "pages": "723--753",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Mark Steedman. 2002. Plans, affordances, and combi- natory grammar. Linguistics and Philosophy, 25(5- 6):723-753.",
                "links": null
            },
            "BIBREF47": {
                "ref_id": "b47",
                "title": "How language structures space",
                "authors": [
                    {
                        "first": "Leonard",
                        "middle": [],
                        "last": "Talmy",
                        "suffix": ""
                    }
                ],
                "year": 1983,
                "venue": "Spatial Orientation: Theory, Research, and Application",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Leonard Talmy. 1983. How language structures space. In Herbert Pick and Linda Acredolo, editors, Spa- tial Orientation: Theory, Research, and Application. Plenum Press.",
                "links": null
            },
            "BIBREF48": {
                "ref_id": "b48",
                "title": "Lexicalization patterns: semantic structure in lexical forms",
                "authors": [
                    {
                        "first": "Leonard",
                        "middle": [],
                        "last": "Talmy",
                        "suffix": ""
                    }
                ],
                "year": 1985,
                "venue": "Language typology and semantic description",
                "volume": "3",
                "issue": "",
                "pages": "36--149",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Leonard Talmy. 1985. Lexicalization patterns: seman- tic structure in lexical forms. In T. Shopen, editor, Language typology and semantic description Vol- ume 3:, pages 36-149. Cambridge University Press.",
                "links": null
            },
            "BIBREF49": {
                "ref_id": "b49",
                "title": "Towards a cognitive semantics",
                "authors": [
                    {
                        "first": "Leonard",
                        "middle": [],
                        "last": "Talmy",
                        "suffix": ""
                    }
                ],
                "year": 2000,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Leonard Talmy. 2000. Towards a cognitive semantics. MIT Press.",
                "links": null
            },
            "BIBREF50": {
                "ref_id": "b50",
                "title": "Modal logic, transition systems and processes",
                "authors": [
                    {
                        "first": "Johan",
                        "middle": [],
                        "last": "Van Benthem",
                        "suffix": ""
                    },
                    {
                        "first": "Jan",
                        "middle": [],
                        "last": "Van Eijck",
                        "suffix": ""
                    },
                    {
                        "first": "Vera",
                        "middle": [],
                        "last": "Stebletsova",
                        "suffix": ""
                    }
                ],
                "year": 1994,
                "venue": "Journal of Logic and Computation",
                "volume": "4",
                "issue": "5",
                "pages": "811--855",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Johan van Benthem, Jan van Eijck, and Vera Ste- bletsova. 1994. Modal logic, transition systems and processes. Journal of Logic and Computation, 4(5):811-855.",
                "links": null
            },
            "BIBREF51": {
                "ref_id": "b51",
                "title": "Logic and the flow of information",
                "authors": [],
                "year": 1991,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Johannes Franciscus Abraham Karel van Benthem. 1991. Logic and the flow of information.",
                "links": null
            },
            "BIBREF52": {
                "ref_id": "b52",
                "title": "Linguistics in philosophy",
                "authors": [
                    {
                        "first": "Z",
                        "middle": [],
                        "last": "Vendler",
                        "suffix": ""
                    }
                ],
                "year": 1967,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Z. Vendler. 1967. Linguistics in philosophy. Cornell University Press Ithaca.",
                "links": null
            },
            "BIBREF53": {
                "ref_id": "b53",
                "title": "Prepositional aspect and the algebra of paths",
                "authors": [
                    {
                        "first": "J",
                        "middle": [],
                        "last": "Zwarts",
                        "suffix": ""
                    }
                ],
                "year": 2005,
                "venue": "Linguistics and Philosophy",
                "volume": "28",
                "issue": "6",
                "pages": "739--779",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "J. Zwarts. 2005. Prepositional aspect and the algebra of paths. Linguistics and Philosophy, 28(6):739- 779.",
                "links": null
            },
            "BIBREF54": {
                "ref_id": "b54",
                "title": "Event shape: Paths in the semantics of verbs",
                "authors": [],
                "year": 2006,
                "venue": "J. Zwarts",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "J. Zwarts. 2006. Event shape: Paths in the semantics of verbs.",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF0": {
                "uris": null,
                "text": "For Figure (F) relative to Ground (G): a. EC(F,G), throughout motion: b. DC(F,G), throughout motion: c. EC(F,G) followed by DC(F,G), throughout motion: d. Sub-part(F',F), EC(F',G) followed by DC(F',G), throughout motion: e. Containment of F in a Vehicle (V).",
                "type_str": "figure",
                "num": null
            },
            "FIGREF1": {
                "uris": null,
                "text": "With the surface as the ground argument, these verbs are defined in terms of two transitions.4",
                "type_str": "figure",
                "num": null
            },
            "FIGREF2": {
                "uris": null,
                "text": "Figure 3: Roll Motion",
                "type_str": "figure",
                "num": null
            },
            "FIGREF4": {
                "uris": null,
                "text": "1. pred := cross 1. pred := roll 2. x := Ball1 2. x := Ball1 3. y.push(Floor) cross(Ball1,[Floor]) roll(Ball1)",
                "type_str": "figure",
                "num": null
            },
            "FIGREF5": {
                "uris": null,
                "text": "The resulting predicates are represented inTable 3as expressions in Dynamic Interval Temporal Logic (DITL)(Pustejovsky and Moszkowicz, 2011), which are equivalent to the LTS expressions used above.cross(Ball1,Floor) loc(Ball1) := y, target(Ball1) := z; b := y; (y := w; y = w; d(b,y) < d(b,w), d(b,z) > d(z,w), IN(y,Floor)) + roll(Ball1) loc(Ball1) := y, rot(Ball1) := z; b loc := y, brot := z; (y := w; y = w; d(b loc ,y) < d(b loc ,w), IN(y,Floor)) + , (z := v; z = v; z-brot < v-brot) +",
                "type_str": "figure",
                "num": null
            },
            "FIGREF7": {
                "uris": null,
                "text": "Sample environment in top-down and perspective views.",
                "type_str": "figure",
                "num": null
            },
            "FIGREF8": {
                "uris": null,
                "text": "11) transform.rotation = new Vector3( 0.0,0.0,transform.rotation.z+ (rotSpeed * deltaTime)); transform.position = new Vector3( transform.position.x-radius * deltaTime,transform.position.y, transform.position.z);",
                "type_str": "figure",
                "num": null
            },
            "FIGREF9": {
                "uris": null,
                "text": "transform.position = new Vector3( transform.position.x-radius * deltaTime, transform.position.y, transform.position.z);",
                "type_str": "figure",
                "num": null
            },
            "FIGREF10": {
                "uris": null,
                "text": "13) loc(Ball1) := y, target(Ball1) := z; b := y; (y := w; y = w; d(b,y) < d(b,w), d(b,z) > d(z,w), IN(y,Floor)) + offset = transform.positiondestination; offset = Vector3.Normalize(offset); transform.position = new Vector3( transform.position.x-offset.x * radius * deltaTime, transform.position.y, transform.position.zoffset.z * radius * deltaTime);",
                "type_str": "figure",
                "num": null
            },
            "FIGREF11": {
                "uris": null,
                "text": "loc(Ball1) := y, target(Ball1) := z; b := y; (y := w; y = w;d(b loc ,y) < d(b loc ,w), d(b loc ,z) > d(z,w), (u := v; u = v; u-b rot < v-b rot ), IN(y,Floor)) + offset =transform.positiondestination; offset = Vector3.Normalize(offset); transform.rotation = new Vector3( 0.0,arccos(offset.z) * (360/PI * 2), transform.rotation.z+ (rotSpeed * deltaTime)); transform.position = new Vector3( transform.position.x-offset.x * radius * deltaTime, transform.position.y, transform.position.z-offset.z * radius * deltaTime);",
                "type_str": "figure",
                "num": null
            },
            "FIGREF12": {
                "uris": null,
                "text": "Roll and slide motions in progress (top), and cross motion in progress (bottom).",
                "type_str": "figure",
                "num": null
            },
            "TABREF0": {
                "text": "Dependency parses for Ball1 crossed Floor (top) and Ball1 rolled (bottom).",
                "html": null,
                "type_str": "table",
                "num": null,
                "content": "<table/>"
            },
            "TABREF1": {
                "text": "Transformation to predicate formula for Ball1 crossed Floor and Ball1 rolled.",
                "html": null,
                "type_str": "table",
                "num": null,
                "content": "<table/>"
            },
            "TABREF2": {
                "text": "DITL expressions for Ball1 crossed Floor and Ball1 rolled.",
                "html": null,
                "type_str": "table",
                "num": null,
                "content": "<table/>"
            },
            "TABREF3": {
                "text": "",
                "html": null,
                "type_str": "table",
                "num": null,
                "content": "<table/>"
            }
        }
    }
}