Datasets:

WINGNUS
/

ACL-OCL

	{
	"paper_id": "P85-1026",
	"header": {
	"generated_with": "S2ORC 1.0.0",
	"date_generated": "2023-01-19T09:39:37.414895Z"
	},
	"title": "REPAIRING REFERENCE IDENTIFICATION FAILURES BY RELAXATION",
	"authors": [
	{
	"first": "Bradley",
	"middle": [
	"A"
	],
	"last": "Goodman",
	"suffix": "",
	"affiliation": {
	"laboratory": "BBN Laboratories I0 Moulton Street Cambridge. Mass. 02238",
	"institution": "",
	"location": {}
	},
	"email": ""
	}
	],
	"year": "",
	"venue": null,
	"identifiers": {},
	"abstract": "The goal of thls work is the enrichment of human-machlne mteractIons in a natural language",
	"pdf_parse": {
	"paper_id": "P85-1026",
	"_pdf_hash": "",
	"abstract": [
	{
	"text": "The goal of thls work is the enrichment of human-machlne mteractIons in a natural language",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Abstract",
	"sec_num": null
	}
	],
	"body_text": [
	{
	"text": "We want to provide a framework less restrictive than earlier ones by allowing a speaker leeway tn forming an utterance about a task and in determining the conversational vehicle to deliver it, A speaker and listener cannot be assured to have the same beliefs, contexts, backgrounds or goals at each point in a conversation. As a result, dlfflcultles and mistakes arise when a listener interprets a speakers utterance. These mistakes can lead to various kinds of mlsunderstandlngs between speaker and hstener. including reference failures or failure to understand the speaker's mtentlon. We call these mtsunderstandlngs mlscommunmatlon Such m~stakes constitute a klnd of \"ill-formed\" input that can slow down end possibly break down communication. Our goal ~s to recognize and Isolate such mlscommunlcattons and circumvent them. Thls paper will hlghhght a particular class of mlscommunlcatlon -reference problems -by descrlbmg a case study, includlng techniques for avoldlng failures of reference",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "envlronment. 1",
	"sec_num": null
	},
	{
	"text": "Cohen, Perrault and Allen showed in thelr paper \"Beyond Question Answering\" [8~ that \",.. users of cluestlon-answerzng systems expect them to do more than just answer isolated questions --they expect systems to engage tn conversation.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "I Introduction",
	"sec_num": null
	},
	{
	"text": "In doing ~o, the system ts expected to allow users to be less than meticulously hteral ~n conveying their zntentlons, and tt is expected to make hnguxstlc and pragmatic use of the previous discourse.\" Following in thelr footsteps, we want to build robust natural language processing systems that can detect and recover from mlsc~mmunlcatton.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "I Introduction",
	"sec_num": null
	},
	{
	"text": "The development of such systems requires s study on how people communicate and how they recover from problems In communication. This paper summarizes the results of a dissertation [13] that tnvestlgates the kinds of mlscommunlcatlon that occur in human communication with a special emphasis on reference prooiems, i.e.. problems a listener has determining whom or what a speaker ts talking about. We have written computer programs and algorithms that demonstrate h~w one could handle such problems m",
	"cite_spans": [
	{
	"start": 180,
	"end": 184,
	"text": "[13]",
	"ref_id": "BIBREF14"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "I Introduction",
	"sec_num": null
	},
	{
	"text": "the context of a natural language understand2ng system. The study of mzscommunlcatlon is a necessary task wlthm such a context since any computer capable of communlcat~ng with humans tn natural language must be tolerant of the tmprecIse, lll-devlsed or complex utterances that people often use.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "IThis reseorcn was suDDorted in port by the Oefenee Advonce4 Reseorch Pro~ect Aqency under \u00a2ontr=ct Neee14--77--C-~378.",
	"sec_num": null
	},
	{
	"text": "Our current research [25, 26] views most dialogues as being cooperatlve and goal directed, l,e.. a speaker and hstener work together to achieve a common goal.",
	"cite_spans": [
	{
	"start": 21,
	"end": 25,
	"text": "[25,",
	"ref_id": null
	},
	{
	"start": 26,
	"end": 29,
	"text": "26]",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "IThis reseorcn was suDDorted in port by the Oefenee Advonce4 Reseorch Pro~ect Aqency under \u00a2ontr=ct Neee14--77--C-~378.",
	"sec_num": null
	},
	{
	"text": "The interpretation of an utterance involves Identifying the underlying plan or goal that the utterance reflects [5. I, 23] . Thls plan, however, is rarely, d ever, obvious at the surface sentence level. A central issue In the interpretation of utterances ts the transformation of sequences of imprecise, zlldevised or complex utterances into well-speclhed plans that might be carried out by dialogue participants. Within thls context, mlscommunlcatlon can occur.",
	"cite_spans": [
	{
	"start": 112,
	"end": 122,
	"text": "[5. I, 23]",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "IThis reseorcn was suDDorted in port by the Oefenee Advonce4 Reseorch Pro~ect Aqency under \u00a2ontr=ct Neee14--77--C-~378.",
	"sec_num": null
	},
	{
	"text": "We ere particularly concerned with cases of mxscommunlcatlon from the heater's viewpoint, such as when the hearer is mattentlve to. confused about, or misled about the zntentlons of the speaker.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "IThis reseorcn was suDDorted in port by the Oefenee Advonce4 Reseorch Pro~ect Aqency under \u00a2ontr=ct Neee14--77--C-~378.",
	"sec_num": null
	},
	{
	"text": "In ordinary exchanges speakers usually make assumptions regarding what thelr listeners know about a topic of discussion. They w111 leave out details thought to be superfluous [2. 19] . Since the speaker really does not know exactly what a listener knows about a topic, tt ts easy to make statements that can be misinterpreted or not understood by the listener because not enough details were presented. One principal source of trouble Is the description constructed by the speaker to refer to an actual object in the world.",
	"cite_spans": [
	{
	"start": 175,
	"end": 182,
	"text": "[2. 19]",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "IThis reseorcn was suDDorted in port by the Oefenee Advonce4 Reseorch Pro~ect Aqency under \u00a2ontr=ct Neee14--77--C-~378.",
	"sec_num": null
	},
	{
	"text": "The descmptlon can be tmpreclse, confused, ambiguous or over!v speclflc. It might be interpreted under the wrong context. This leads to dlfflculty for the hstener when figuring out what oblect ~s being described, that Is. ref.erence identification errors. Such descriptions are \"allformed\" input, the blame for ill-formedness may lie partly with the speaker and partly with the listener The speaker may have been sloppy or not taken the hearer into consideration, the listener may be either remiss or unwilling to admit he can't understand the speaker and to ask the speaker for clarification, or may slmply feel that he has understood when he zn fact has not.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "IThis reseorcn was suDDorted in port by the Oefenee Advonce4 Reseorch Pro~ect Aqency under \u00a2ontr=ct Neee14--77--C-~378.",
	"sec_num": null
	},
	{
	"text": "ts part of an on-going effort to develop a reference Identlfzcatmn and plan recognition mechanism that can exhibit more \"human-hke ' tolerance of such utterances.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Thls work",
	"sec_num": null
	},
	{
	"text": "Our goal zs to build a more robust system that can handle errorful utterances, and ~hat can be incorporated in exlstlng systems.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Thls work",
	"sec_num": null
	},
	{
	"text": "As a start, we have concentrated on reference tdentlflcatzon.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Thls work",
	"sec_num": null
	},
	{
	"text": "In conversation people use imperfect descriptions to communicate about objects; sometimes their partners succeed zn understanding and occasionally they fail. Any computer hoping to play the part of a listener must be capable of taking what the speaker says and either deleting, adapting or clarifying it.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Thls work",
	"sec_num": null
	},
	{
	"text": "We are developing a theory of the use of extensional descrlptlons that will help explam how people successfully use such imperfect descriptions. We call thls the theory of reference mlscommunlcation Section 2 of this paper highlights some aspects of normal communication and then provides a general discussion on the types of miscommunlcatlon that occur In conversation, concentrating primarily on reference problems and motivating many of them with Illustrative protocols.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Thls work",
	"sec_num": null
	},
	{
	"text": "Section 3 presents possible ways around some of the problems of miscommunxcation in reference. Motivated there is a partial Implementation of a reference mechanism that attempts to overcome many reference problems. We are following the task-omented paradigm of Grosz [14] since it Is easy to study (through videotapes).",
	"cite_spans": [
	{
	"start": 267,
	"end": 271,
	"text": "[14]",
	"ref_id": "BIBREF15"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Thls work",
	"sec_num": null
	},
	{
	"text": "It places the world In front of you (a primarily extensional world), and It limits the dlscusslon whlle still providing a rlch environment for complex descriptions.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Thls work",
	"sec_num": null
	},
	{
	"text": "The task chosen as the target for the system Is the assembly of a toy water pump. J must repair the descrlptlon, reahzmg that A doesnt really mean \"one hole but is referring to t,~e 'two' holes. J apparently does this since he doesnt complain about as description and correctly attaches the BASEVALVE to the TUBEBASE Figure lib) shows the configuration of the pump after the TUBEBASE is attached to the MAINTUBE \"n Lme I0, [n Lme 13. J interprets \"a red plastic piece\" to refer to the .VOZZLE When A adds the relative clause \"that has four gi=mos on it.\" J is forced to drop the NOZZLE as the referent and to se{ect the SLLDEV~LVE In Lmes i7 and 18, A'S description \"the other--the open part of the maln tube. the lower valve\" is ambiguous, and J selects the wrong slte, namely the TUBEBAEE, in which to insert the SLIDEVALVE.",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 317,
	"end": 329,
	"text": "Figure lib)",
	"ref_id": null
	}
	],
	"eq_spans": [],
	"section": "Thls work",
	"sec_num": null
	},
	{
	"text": "Since The Toy Water Pump",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Thls work",
	"sec_num": null
	},
	{
	"text": "People must and do manage to resolve lots of (potentaal) mascommumcataon In everyday conversataon. Much of it as resolved subconscaously wlth the hstener unaware that anything is wrong, Other mlscommumcatlon is resolved wath the listener actively deleting or replacang mformataon m the speakers utterance until It flts the current context. Sometimes thls resolutlon Is postponed until the questlonable part of the utterance is actually needed.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "C 2 Miscommunication",
	"sec_num": null
	},
	{
	"text": "Shll. when all these fail. the hstener can ask the speaker to clarlfy what was said. 2 There are many aspects of an utterance that the hstener can become confused about and that can lead to mascommunacatton. The hstener can become confused about what the speaker intends for the referents, the actaons, and the goals described by the utterance,",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "C 2 Miscommunication",
	"sec_num": null
	},
	{
	"text": "Confuslons often appear to result from confhct between the current state of the conversation. the overall goal of the speaker, or the manner In which the speaker presented the anformatlon. However, when the hstener steps back and is able to discover what k~nd of confuslon ~s occurring, then the confusion can qulte possibly be resolved.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "C 2 Miscommunication",
	"sec_num": null
	},
	{
	"text": "Thls sectaon attempts to motlvate a paradlgm for the kinds of conversation that we studled and traes to point out places m the paradlgm that leave room for mlscommumcatlon. ~'.1.1 Effects of the structure of task-oriented dialogues",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Causes of mlscommunication",
	"sec_num": "2.1"
	},
	{
	"text": "Task-oriented conversatlons have a speclfic goal to be achleved: the performance of a task (e.g.. [14] ). The partlclpants in the dlalogue can have the same skill level and they can slmply work together to accomplish the task; or one of them, the expert, could know more and could direct the other, the apprentlce. to perform the task.",
	"cite_spans": [
	{
	"start": 98,
	"end": 102,
	"text": "[14]",
	"ref_id": "BIBREF15"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Causes of mlscommunication",
	"sec_num": "2.1"
	},
	{
	"text": "We have concentrated prlmarlly on the latter case -due to the protocols that we examlned -but many of our observations can be generahzed to the former case, too.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Causes of mlscommunication",
	"sec_num": "2.1"
	},
	{
	"text": "We will refer to thls as the apprentlce-expert domaln.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Causes of mlscommunication",
	"sec_num": "2.1"
	},
	{
	"text": "The vlewpomts of the expert and apprentlce differ greatly In apprentlce-expert exchanges.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Causes of mlscommunication",
	"sec_num": "2.1"
	},
	{
	"text": "The expert, having an understandlng of the functlonahty of the elements in the task. has more of a feel for how the elements work together, how they go together, and how the indlvldual elements can be used.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Causes of mlscommunication",
	"sec_num": "2.1"
	},
	{
	"text": "The apprentlce normally has no such knowledge and must base hls declslons on perceptual features such as shape [15] . The structure of the task affects the structure of the dlalogue [14}. partlcularly through the center of attentlon of the expert and apprentlce.",
	"cite_spans": [
	{
	"start": 111,
	"end": 115,
	"text": "[15]",
	"ref_id": "BIBREF16"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Causes of mlscommunication",
	"sec_num": "2.1"
	},
	{
	"text": "Thls is the phenomenon called focus [14. 20. 24] . whlch, in taskorlented dlalogues Is a very real and operational thlng (e.g., focus is used In resolving anaphorac references). Shafts ~n focus correspond dlrectly to the task, ats subtasks, the oblects an a task and the subpleces of each object Focus and focus shifts are governed by many rules [14. :~0, 24] Confusaon may result when expected shafts do not take place.",
	"cite_spans": [
	{
	"start": 36,
	"end": 48,
	"text": "[14. 20. 24]",
	"ref_id": null
	},
	{
	"start": 346,
	"end": 359,
	"text": "[14. :~0, 24]",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Causes of mlscommunication",
	"sec_num": "2.1"
	},
	{
	"text": "For example. If the expert changes focus to an object but never discusses Its subpaeces ~such as an obvaous attachment surface) or never bothers to talk about the object reasonably soon after its antroductlon (Le., between the tame of ~ts mtroductlon and its use. without digressing in a wellstructured way In between (see [20] )), then the apprentlce may become confused, leavang hlm r~pe for mlscommunlcatlon. The reverse anfluence between focus and oblects can lead to trouble, too. A shzft In focus by the expert that does not have a manHestatlon In the apprentlce's world wall also perplex the apprentice Focus also influences how descr:ptlons are formed [15, 2] . The level of detail requlred in a description depends directly on the elements currently highlighted by the focus If the oblect to be descrabed Is samflar to other element~ m focus, the expert must be more speclhc m the formulation of the descraptlon or may conslder shlftmg focus away from the posslbly ambiguous objects to one where the amblgulty wont occur.",
	"cite_spans": [
	{
	"start": 323,
	"end": 327,
	"text": "[20]",
	"ref_id": "BIBREF22"
	},
	{
	"start": 660,
	"end": 664,
	"text": "[15,",
	"ref_id": "BIBREF16"
	},
	{
	"start": 665,
	"end": 667,
	"text": "2]",
	"ref_id": "BIBREF1"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Causes of mlscommunication",
	"sec_num": "2.1"
	},
	{
	"text": "In thls section we will make It clear that people do m:scommunlcate and yet they often manage to flx thlngs.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Consequences of miscommunicatlon",
	"sec_num": "2.2"
	},
	{
	"text": "We will look at speclfic forms of mlscommunlcatlon and descrlbe ways to detect them. We will hzghhght relatlonsh;ps between different mlscommunzcat;on problems but won't necessarzly demonstrate ways to resolve each of them.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Consequences of miscommunicatlon",
	"sec_num": "2.2"
	},
	{
	"text": ";n [17).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "2An analysis of clarification suodialogues can be found",
	"sec_num": null
	},
	{
	"text": "There are many ways hearers can get confused during a conversation. Figure 2 outlines some of them that were derived from analyzing the water pump protocols.",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 68,
	"end": 76,
	"text": "Figure 2",
	"ref_id": null
	}
	],
	"eq_spans": [],
	"section": "Instances of mtscommun/cation",
	"sec_num": "2.2.1"
	},
	{
	"text": "This section defines and illustrates many of them through numerous excerpts.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Instances of mtscommun/cation",
	"sec_num": "2.2.1"
	},
	{
	"text": "Each excerpt is marked in parentheses to show what modality of communication was used (see [9] for a description about the collection of these excerpts). Each bracketed portion of the excerpt explains what was occurring at that point in the dialogue. The confusions themselves, coupled with the description at the end of this section on how to recognize when one of them is occurring, provides motivation for the use of the algorithm outlined in Section 3 as a means for repairing communication problems. We will only discuss referent confusion tn this paper.",
	"cite_spans": [
	{
	"start": 91,
	"end": 94,
	"text": "[9]",
	"ref_id": "BIBREF9"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Instances of mtscommun/cation",
	"sec_num": "2.2.1"
	},
	{
	"text": "The other forms of confusion -Action. Goal, and Cogmtive Load -are described in [11. 13] . Another categorization of confusmns that lead to conversation failure can be found in [22] .",
	"cite_spans": [
	{
	"start": 80,
	"end": 88,
	"text": "[11. 13]",
	"ref_id": null
	},
	{
	"start": 177,
	"end": 181,
	"text": "[22]",
	"ref_id": "BIBREF25"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Instances of mtscommun/cation",
	"sec_num": "2.2.1"
	},
	{
	"text": "\u2022 Figure 2 :",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 2,
	"end": 10,
	"text": "Figure 2",
	"ref_id": null
	}
	],
	"eq_spans": [],
	"section": "Instances of mtscommun/cation",
	"sec_num": "2.2.1"
	},
	{
	"text": "A taxonomy of confusmns Referent ~onfuslon occurs when the listener is unable to correctly determine what the speaker is referring to with a particular descrlptmn.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Instances of mtscommun/cation",
	"sec_num": "2.2.1"
	},
	{
	"text": "[t occurs when the descriptions In the utterance are ambiguous or imprecise, when there IS confusion between the speaker and listener about what the current focus or context Is, or when the descriptions in the utterance are either incorrect or incompatible with the current or global context.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Instances of mtscommun/cation",
	"sec_num": "2.2.1"
	},
	{
	"text": "Ambiguous (and. thus, imprecise) descnptxons can cause confusion about the referent.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "Excerpt 2 below illustrates a case where the speaker's description is underspecxfled -it does not provide enough dated to prune the set of possible referents down to one. In Line 4 and 5, S describes the location to place a peg into a hole by giving spatial information.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "Since the location is given relative to another location by \"in the hole at the green end\", it defines a region where the peg might go instead of a specific location.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "In this particular case, there are three possible holes to choose from that are near the green end. The listener chooses one -the wrong one -and inserts the peg into it.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "Because this dialogue took place face to face, S is able to correct the ambiguity in Lines 6 and 7.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "A speaker's description can be imprecise in several possible ways. (1) It may contain features that do not readily apply in the domain. In fine 3, Excerpt 3, the feature \"funny\" has no relevance to the listener. It is not until A provides a fuller description in Lines 5 to 8 that E is able to select the proper piece.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "(2) It may use a vague head noun coupled with few or no feature values (and context alone does not necessarily suffice to distinguish the object). In Excerpt 4, Line 9, \"attachment\"",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "is vague because all objects in the domain are attachable parts. The expert's use of \"attachment\" was most likely to signal the action the apprentice can expect to take next.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "The use of the feature value \"clear'* provides little benefit either because three clear, unused parts exist.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "The size descriptor \"little\" prunes this set of possible referents down to two contenders.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "(3) Enough feature values are provided but at least one value is too vague leading to trouble.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "In Excerpt 5, Line 3, the use of the attribute value \"rounded\" to describe the shape does not sufficiently reduce the set of four possible referents (though, in this particular instance, A correctly identifies it) because the term is applicable to numerous parts In the dommn.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "A more precise shape descriptor such as \"bell-shaped\" or \"cylindrical\" would have been more beneficial to the listener, Excerpt 3 (Telephone) E: I. All right. ",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Erroneous Specificity",
	"sec_num": null
	},
	{
	"text": "Focus confusion can occur when the speaker sets up one focus and then proceeds with another one without letting the listener know of the switch (i.e., a focus shift occurs without any indication). An opposite phenomenon can also happen -the listener may feel that a focus shift has taken place when the speaker actually never intended one.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "These really are very similar -one Is viewed more strongly from the perspective of the speaker and the other from the listener.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "Excerpt 6 below lUustrates an mstance of the first type of focus confusion.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "In the excerpt, the speaker (S) shifts focus without notifying the listener (P) of the switch. As the excerpt begins, P ,s holding the TUBEBASE. S provides in Lines 1 to 16 mstructzons for P to attach the CAP and the SPOUT to outlets OUTLETI and OUTLET2, respectively, on the MAINTUSE.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "Upon P's successful completion of these attachments. S switches focus m Lines 17 to 20 to the TUSESASE assembly and requests P to screw tt on to the bottom of the M,e/NTUSE. White P completes the task. S realizes she left out a step in the assemblythe placement of the SLIDEVALVE into OUTLET2 of the M,eJNTUSE before the SPOUT ts placed over the same outlet.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "S attempts to correct her mistake by requesting P to remove \"the pies \"~ piece in ~nes 22 and 23.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "Since S never indicated a shift in focus from the TUSESASE back to the IPOUT, P mterprets \"the pies\" to refer to the TUSESASE. 3. on the side holes that are 3The whole ward here is \"pleetic.\" People in general tend to be good ot proceedinq before heorin 9 the whole utteronce or even the whole word. 30. it fits loosely in there.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "Excerpt 7 below demonstrates the latter type of focus confuszon that occurs when the speaker (S) sets up one focus -the M,4]NTUBE, which is the correct focus In this case -but then proceeds in such a manner that the listener (J) thinks a focus shift to another piece, the TUBESASE, has occurred.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "Thus, Line 15 refers to \"the lower side hole in the M,41NTUBE\" for S and \"the hole in the TUBEBASE\" for J. J has no way of realizing that he has focused incorrectly unless the description as he interprets it doesn't have a real world correlate (here something does satisfy the description so J doesn't sense any problem) or if, later in the exchange, a conflict arises due to the mistake (e.g,, a requested action can not be performed).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "In Line 31, J inserts a piece into the wrong hole because of the misunderstanding in Line 15. Line 31 hints that J may have become suspicious that an ambiguity existed but since the task was successfully completed (i.e., the red piece fit into the hole in the base), and since S did not provide any clarification, he assumed he was correct. Misfocus can also occur when the speaker inadvertently lefts to distinguish the proper focus because he did not notice a possible ambiguity; or when, through no fault of the speaker, the listener just fails to recognize a switch in focus indicated by the speaker. ~xcerpt 7 above is an example of the first type because S failed to notice that an amblguzty existed since he never explicitly brought the TUBEBASE either into or out of focus.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "He just assumed that J had the same perspective as hzm -a perspective in which uo ambiguity occurred.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Improper Focus",
	"sec_num": null
	},
	{
	"text": "Context differs from focus.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Wrong Context",
	"sec_num": null
	},
	{
	"text": "The context of a portion of a conversation is concerned with the po:nt of the discussion in that fragment and with the set of objects relevant to that discussion, though not attended to currently.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Wrong Context",
	"sec_num": null
	},
	{
	"text": "Focus pertains to the elements which are currently being attended to in the context. For example, two people can share the same context but have different focus assignments wt~hm it -we're both talking about the water pump but you're describing the MA/NTUB\u00a3 and I'm descrlbmg the AIRCH,4MB\u00a3,q.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Wrong Context",
	"sec_num": null
	},
	{
	"text": "Alternatively, we could JUst be uslng different contexts -I think you're talking about taking the pump apart but you're talking about replh^lng the pump with new parts -m both cases we m~v be sharing the same focus -the pump -but our conte~,s are totally off from one another. ~",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Wrong Context",
	"sec_num": null
	},
	{
	"text": "The kinds of misunderstandings that can occur because of context problems are similar to those for focus problems:",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Wrong Context",
	"sec_num": null
	},
	{
	"text": "(1) the speaker might set up or be xn one context for a discussion and then proceed in another one without effectively letting the listener know of the change, (2) the listener may feel a change in context has taken place when in fact the speaker never Intended one, or (:3) the Listener fails to recognize an indicated context switch by the speaker.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Wrong Context",
	"sec_num": null
	},
	{
	"text": "Context affects reference because it helps define the set of available oblects that are possible contenders for the referent of the speaker's descriptions.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Wrong Context",
	"sec_num": null
	},
	{
	"text": "If the contexts of the speaker and listener differ, then m\|sreference might result.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Wrong Context",
	"sec_num": null
	},
	{
	"text": "An analogy (see [I0] for \u2022 discusslon on analogies) is a useful way to help descrlbe an object by attemptlng to be more precise by using shared past expemence and knowledge -espec:ally shape and functional reformation.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Bad AnaloEy",
	"sec_num": null
	},
	{
	"text": "If that past experxence or knowledge doesn't contain the reformation the speaker assumes it does or isn't there, then trouble occurs. Thus. one more way referent confusion can occur Is by describing an oh}act using \u2022 poor analogy. An analogy used to describe an object might not be spec:fic 4Groez [14, lS] would dem~ril~ this as o difference in \"task DIane J ~ile Rai\u00a2ltlNnt [2e, 21] m~uld say that the \"c0mlmmjcativa gCNlie\" dJffare\u00a2l.",
	"cite_spans": [
	{
	"start": 298,
	"end": 306,
	"text": "[14, lS]",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Bad AnaloEy",
	"sec_num": null
	},
	{
	"text": "enough -confusing the listener because several pieces might conform to the analogy or, tn fact, none at all appear to fit because discovering a mapping between the analogous object and some piece in the environment Is too difficult.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Bad AnaloEy",
	"sec_num": null
	},
	{
	"text": "In Excerpt 8, J at first has trouble correctly satisfying A's functional analogy \"stopper\" in \"the bag blue stopper\", but finally selects what he considers to be the closest match to \"stopper\".",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Bad AnaloEy",
	"sec_num": null
	},
	{
	"text": "A: I. Okay. Now. In other cases tt might be too specificconfusing the listener because none of the available referents appear to fit it.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Excerpt 8 (Telephone)",
	"sec_num": null
	},
	{
	"text": "In Line 8 of Excerpt 6, \"nozzle-looking\" forms a poor shape analogy because the object being referred to actually Is an elbowshaped spout.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Excerpt 8 (Telephone)",
	"sec_num": null
	},
	{
	"text": "The \"nozzle-looklng\" part of the description convinced the listener that what he was looking for was something specific like a nozzle (which xs a small spout). Sometimes, when an oblect xs a clear representative of a specified analogy class, the apprent2ce may become confused, wondering why the expert bothered to form an analogy mstead of just directly describing the object as a member of the class. Hence, tt would not be surprising d the apprentice tgnoreu the best representatnve of the class for some less obvious exemplar. Thus, for example, It ts better to say \"nozzle\" instead of \"nozzle-looking.\" In Excerpt 9, the description \"hippopotamus face shape\" (a shape analogy) tn Lines 2 and 3, and \"champagne top\" (a shape analogy) in Line 9. ere too speclhc and the hstener ts unable to easily find something close enough to match either of them.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Excerpt 8 (Telephone)",
	"sec_num": null
	},
	{
	"text": "He can't discover a mapping between the oblect in the analogy and one in the real world. A description is incompatible when (1) one or more of the specified conditions, i.e., the feature values, do not satisfy any of the pieces; (2) when one or more specified constraints do not hold (e.g.. saying \"the loose one\" when all objects are tightly attached). or (3) if no one object satisfies al_~l of the features specified in the description.",
	"cite_spans": [
	{
	"start": 357,
	"end": 360,
	"text": "(3)",
	"ref_id": "BIBREF2"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Excerpt 8 (Telephone)",
	"sec_num": null
	},
	{
	"text": "In Lines 7 and 8 of Excerpt 9 above, M's use of \"the two peg holes\" leads to bewilderment for the listener because the described object has no holes in it. M actually meant \"two pegs\".",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Excerpt 8 (Telephone)",
	"sec_num": null
	},
	{
	"text": "Part of our research has been to examine how a listener discovers the need for a repair of an utterance or a description during communication.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Detecting miscommunicatlon",
	"sec_num": "2.2.2"
	},
	{
	"text": "The incompatibility of a referent or action is one signal of possible trouble.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Detecting miscommunicatlon",
	"sec_num": "2.2.2"
	},
	{
	"text": "The appearance of an obstacle that blocks one from achieving a goal is another indication of a problem.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Detecting miscommunicatlon",
	"sec_num": "2.2.2"
	},
	{
	"text": "Two kinds of incompat~btltty, action or referent. appear In the taxonomy of confusions.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Incompatibillty",
	"sec_num": null
	},
	{
	"text": "The strongest hint that there is a reference problem occurs when the listener finds no real world object to correspond to the speaker's description.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Incompatibillty",
	"sec_num": null
	},
	{
	"text": "This can occur when (1) one or more of the specified feature values xn the description are not satisfied by any of the pieces (e.g. saying \"the orange cap\" when none of the objects are orange~. {2) when one or more specified constraints do not hold (e.g., saying \"the red plug that fits loosely\" when all the red plugs attach tightly), or (3) If no one object satisfies all of the features specified m the description (I.e., ther'e-ts, for each feature, an object that exhibits the specified feature value, but no one object exhibits all of the values). An action problem xs likely ~f I l) the listener cannot perform the action specified by the speaker because of some obstacle; (2) the hstener performs the action but does not arrlve at its intended effect (I.e., a specified or default constramt lsnt satisfied); or (3) the current action affects a previous action tn an adverse way, yet the speaker has given no sign of any importance to this side-effect.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Incompatibillty",
	"sec_num": null
	},
	{
	"text": "This blockage can result m confusion for the hstener because he did not expect the speaker to give him tasks that could not be achieved.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal obstacle A goal obstacle occurs when a goal (or subgoa[) one is trying to achieve ts blocked",
	"sec_num": null
	},
	{
	"text": "Often. though, it points out for the hstener that some mlscommunication (such as mlsreference) has occurred.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal obstacle A goal obstacle occurs when a goal (or subgoa[) one is trying to achieve ts blocked",
	"sec_num": null
	},
	{
	"text": "Goal redundancy occurs when the requested goal (or subgoal) is already satisfied. In some sense, xt xs a special klnd of goal obstacle where the goal to be fulfilled is blocked because it is already satisfied. It is a simple goal obstacle because nothmg has to be done to get around it. However, it can lead to confusion on Given all the possibilities for confusion, when confusions do occur, they must be resolved If the task is to be performed.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "This section explores the problem of fixing reference failures.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "Reference Identification is a search process where a listener looks for something in the world that satisfies a speaker's uttered description. A computatlonal scheme for performing reference has evolved from work by other artificial intelligence researchers (e.g., see [14] ). That tradltlonal approach succeeds if a referent ~s found, or falls d no referent ts found {see Figure 3(a) ).",
	"cite_spans": [
	{
	"start": 269,
	"end": 273,
	"text": "[14]",
	"ref_id": "BIBREF15"
	}
	],
	"ref_spans": [
	{
	"start": 373,
	"end": 384,
	"text": "Figure 3(a)",
	"ref_id": "FIGREF2"
	}
	],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "However, a reference identlficatlon component must be more versatile than those constructed m the traditional manner. The excerpts provided m the prevlous section show that the traditional approach is wrong because people's real behavlor zs much more elaborate.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "In particular. hsteners often find the correct referent even when the speaker's descrlpt)on does not describe any object In the world.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "For example, a speaker could descrlbe a blue block as the \"turquoise block.\" Most listeners would go ahead and assume that the blue block was the one the speaker meant.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "A key feature to reference identlficatlon is \"negotlatlon.\" Negotlatlon in reference ldentlhcatlon comes in two forms. First. It can occur between the listener and the speaker.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "The listener can step back, expand greatly on the speaker's descrlptlon of a plausible referent, and ask for conhrmatlon that he has indeed found the correct referent.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "For example, a hstener could mltlate negotiation wlth 'Tin confused. Are you talking about the thlng that is klnd of flared at the top?",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "Couple inches long. It's kind of blue.\" Second. negotiation can be wlth oneself.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "Thls type of negotiation, called self-negotlatlon. Ls the one that we are most concerned wlth in thls research. The listener conslders aspects of the speaker's descrzptlon, the context of the commumcatlon, and the listener's own abdltles. He then apphes that dehberatlon to determine whether one referent candldate :s better than another or. if no candidate Is found, what are the most likely places for error or confuslon.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "Such negotlatlon can result in the listener testing whether or not a partlcular referent works.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "For example, linguistic descrlptlons can influence a listener's perception of the world.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "The listener must ask himself whether he can percelve one of the oblects in the world the way the speaker described it. in some cases, the listener's perceptlon may overrule the descrlptlon because the listener can't percelve ~t the way the speaker described it. Thus. the reference process becomes multi-step and resumable This computational model, which I call \"FWIM\" for \"Find What I Mean\", is more faithful to the data than the traditional model (see Figure 3 ",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 455,
	"end": 463,
	"text": "Figure 3",
	"ref_id": "FIGREF2"
	}
	],
	"eq_spans": [],
	"section": "Goal redundancy",
	"sec_num": null
	},
	{
	"text": "When a description fails to denote a referent In the real world properly, It Is possible to repair tt by a relaxatlon process that ignores or modifies parts of the descrlptlon.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The relaxation component",
	"sec_num": "3.2"
	},
	{
	"text": "Since a description can speclfy many features of an object, the order In which parts of It are relaxed Is crucial (i.e.. relaxing Ln different orders could yield matches to different objects)",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The relaxation component",
	"sec_num": "3.2"
	},
	{
	"text": "There are several kinds of relaxation possible One can ignore a constituent,",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The relaxation component",
	"sec_num": "3.2"
	},
	{
	"text": "replace It with something close, replace it with a related value, or change focus (i.e.. consider a different group of objects.).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The relaxation component",
	"sec_num": "3.2"
	},
	{
	"text": "This section descrlbes the overall relaxatlon component that draws on knowledge sources about descriptions and the real world as it tries to relax an errorful description to one for which a referent can be sdentlfied.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "The relaxation component",
	"sec_num": "3.2"
	},
	{
	"text": "Identifying the referent of a description requires finding an element in the world that corresponds to the speaker's description (where every feature specified in the description is present In the element in the world but not necessarily vice versa).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Find a referent using a reference mechamsm",
	"sec_num": "3.2.1"
	},
	{
	"text": "The initial task of our reference mechanism Is to determine whether or not a search of the (taxonomic) knowledge base that we use to model the world Is necessary. For example, the reference component should not bother searchingunless specifically requested to do so -for a referent for indefinite noun phrases (which usually describe new or hypothetical objects) or extremely vague descriptions (which do not clearly describe an oblect because they are composed of imprecise feature values).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Find a referent using a reference mechamsm",
	"sec_num": "3.2.1"
	},
	{
	"text": "A number of aspects of discourse pragmattcs can be used in that determination (eg., the use of a delctlc In a definite noun phrase, such as \"thls X\" or \"the last X\", hints that the object was either mentioned previously or that it probably was evoked by some previous reference, and that it is searchable) but we will not examine them here. A CONCEPT Is like a set. representing those elements described by it. A SUPERC link ('==>\") is used between concepts to show set Inclusion.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Find a referent using a reference mechamsm",
	"sec_num": "3.2.1"
	},
	{
	"text": "For example, consider Figure 3 Assuming that a search of the knowledge base Is considered necessary, then a reference search mechanism ts revoked.",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 22,
	"end": 30,
	"text": "Figure 3",
	"ref_id": "FIGREF2"
	}
	],
	"eq_spans": [],
	"section": "Find a referent using a reference mechamsm",
	"sec_num": "3.2.1"
	},
	{
	"text": "The search mechanism uses the KL-One Classther [16] With respect to reference, this means that all possible (descriptions of) referents of the descrlptlon will be subsumed by tt after It has been classLhed rote the knowledge base taxonomy.",
	"cite_spans": [
	{
	"start": 47,
	"end": 51,
	"text": "[16]",
	"ref_id": "BIBREF17"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Find a referent using a reference mechamsm",
	"sec_num": "3.2.1"
	},
	{
	"text": "If more than one candidate referent Is below (when a descrlptlon A Is subsumed by B. we say A ts \"below\" B) the classified description, then, unless a quantifier in the description specified more than one element, the speaker's description is ambiguous.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Find a referent using a reference mechamsm",
	"sec_num": "3.2.1"
	},
	{
	"text": "If exactly one descr~ptlon Is below it, then the intended referent is assumed to have been found.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Find a referent using a reference mechamsm",
	"sec_num": "3.2.1"
	},
	{
	"text": "Finally, if no referent is found below the classified descrxption, the relaxation component is invoked.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Find a referent using a reference mechamsm",
	"sec_num": "3.2.1"
	},
	{
	"text": "We will only consider the last case in the rest of the paper.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Find a referent using a reference mechamsm",
	"sec_num": "3.2.1"
	},
	{
	"text": "It is necessary to determine whether or not the lack of a referent for a description has to do with the description itself (i.e.. reference failure) or outside forces that are causing reference confusion. For example, the problem may be with the flow of the conversation and the speaker's and hsteners perspectives on it; it may be due to mcorrect attachment of a modifier; it may be due to the action requested; and so on.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Collect votes for or against relaxing the description",
	"sec_num": "3.2.2"
	},
	{
	"text": "Pragmatic rules are Invoked to decide whether or not the descrxptlon should be relaxed.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Collect votes for or against relaxing the description",
	"sec_num": "3.2.2"
	},
	{
	"text": "These rules will not be discussed here so we will assume that the problem lies in the speakers description.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Collect votes for or against relaxing the description",
	"sec_num": "3.2.2"
	},
	{
	"text": "If relaxation Is demanded, then the system must (1) find potential referent candidates, (2l determine which features in the speaker's description to relax and in what order, and use those ordered features to order the potential candidates with respect to the preferred ordering of features, and (3~ determine the proper relaxation techniques to use and apply them to the description.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Perform the relaxation of the description",
	"sec_num": "3.2.3"
	},
	{
	"text": "Before relaxation can take place, potential candidates for referents (which denote elements in the listener's visual scene) must first be found These candidates are discovered by performing a \"walk\" tn the knowledge base taxonomy in the general vlclmty of the speakers classified description. A KL-One partial marcher is used to determme how close the candidate descriptions found during the walk are to the speakers description, The partial metcher generates a numerical score to represent how well the descrlptlons match (after first generating scores at the feature level to help determme how the features are to be aligned end how well they match). This score is based on information about KL-One and does not take mto account any information about the task domain. The ordering of features and candidates for relaxation described below takes Into account the task domain. The set of best descriptions returned by the marcher (as determined by some cutoff score) are selected as referent candidates.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Find potential referent candidates",
	"sec_num": null
	},
	{
	"text": "At this peat the reference system inspects the speaker's description and the candidates, decides wtltch features to relax and in what order. 5 and generates a master ordering of features for relaxation.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "Once the feature order Is created, the reference system uses 50f course, om=a one ~rticular candidate is selected. then deciding which features to relax is relatively tr(vial -one simply c(mporee feature by feature between the candidate description (the target) and the speaker's description (the ~ttern) and notes any discrepancies.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "that ordering to determine the order in which to try relaxing the candidates.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "We draw pr;martly on sources of linguistic knowledge, pragmatic knowledge, discourse knowledge, domam knowledge, perceptual knowledge, hierarchical knowledge, and trial and error knowledge durmg this repair process. A detailed treatment of all of them can be found In [12, 27, 13] . These knowledge sources are consulted to determine the feature ordering for relaxation.",
	"cite_spans": [
	{
	"start": 268,
	"end": 272,
	"text": "[12,",
	"ref_id": "BIBREF12"
	},
	{
	"start": 273,
	"end": 276,
	"text": "27,",
	"ref_id": null
	},
	{
	"start": 277,
	"end": 280,
	"text": "13]",
	"ref_id": "BIBREF14"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "We represent information from each knowledge source as a set of relaxation rules. These rules are written in a PROLOG-Iike language. Figure 5 illustrates one such linguistic knowledge relaxation rule. This rule is motivated by the observation in the excerpts that speakers typ~cally add more important informatlon at the end of a descrlpt~on (where they are separated from the ma~n part of the descrlpt~on and thus provided more emphasis).",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 133,
	"end": 141,
	"text": "Figure 5",
	"ref_id": "FIGREF5"
	}
	],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "Since the syntactic constituents often at the end are relatlve clauses or predicate complements, we created this more specdic relaxatlon rule. However. a more general and more applicable rule is that information presented at the end of a descrlptlon is usually more promment. E.g.. Rel ox-Feot ure-Be f ore(v 1 .v2)",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "<-ObjectOeecr(d), Feat ureOeec r i ptor(v! ),",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "FectureOescr iptor(v2),",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "FecturelnOeecr i pt ion(vf .d).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "Feat urel nOesc r i pt i on(v2 .d). The partial ordermgs are then zntegrated to form a d~rected graph.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "For example. perceptual knowledge may say to relax color However. ~f the color value was asserted ~n a relative clause. linguistic knowledge would rank color lower. ~.e.. placmg ~t later ~n the list of things to relax.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "Smce different knowledge sources generally have different partial orderlngs of features, these differences can lead to a conflict over which features to relax.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "It Is the job of the best candidate algorithm to resolve the d~sagreements among knowledge sources. It's goal ts to order the referent candidates, Ci, so that relaxation ~s attempted on the best candzdates first Those candidates are the ones that conform best to a proposed feature ordering. To start, the algorithm exammes pairs of candidates and the feature order~ngs from each knowledge source.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "For each candidate C i. the algorithm scores the effect of relaxlng the speaker's orlglnal descrlpt~on to C i. using the feature ordering from one knowledge source.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "The score reflects the goal of mln~mlz:ng the number of features relaxed whale try3ng to relax the features that are \"earhest\" sn the feature ordermg.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "It repeats ~ts scoring of C i for each knowledge source, and sums up its scores to form Ci's total score. The Ci's are then ordered by that score. Figure 6 provides a graphic description of th~s process.",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 147,
	"end": 155,
	"text": "Figure 6",
	"ref_id": "FIGREF1"
	}
	],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "A set of objects ~n the real world are selected by the partial marcher as potent~a\| candidates for the referent.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "These candidates are shown across the top of the figure.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "The lines on the right side of each box correspond to the set of features that describe that object.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "The speaker's descrlpt~on ts represented in the center of the figure.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "The set of specified features and their assigned feature value (e.g., the pair Color-Maroon) are also shown there.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "A set of partial orderings are generated that suggest which features in the speaker's description should be relaxed first -one ordering for each knowledge source (shown as \"l~nguist~c,\" \"Perceptual.\" and \"H~erarchlcaI\" in the figure). These are put together to form a directed graph that represents the possible, reasonable ways to relax the features specified tn the speakers description.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "Finally. the referent candidates are reordered using the information expressed ~n the speaker's description and in the directed graph of features. Reordering referent candldates Once a set of ordered, potential candldates are selected, the relaxation mechanlsm begms step 3 of relaxatlon; it trles to find proper relaxation methods to relax the features that have lust been ordered ~success tn flndlng such methods \"justifies\" relaxing the descrlptlon).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "It stops at the first candidate which zs reasonable.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Order the features and candidates for relaxation",
	"sec_num": null
	},
	{
	"text": "Relaxation can take place wlth many aspects of a speaker's descrlptlon:",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Determine which relaxation methods to apply",
	"sec_num": null
	},
	{
	"text": "wlth complex relatlons specified In the descrlptlon, wlth indlvldual features of a referent specified by the descrlptlon, and with the focus of attention in the real world where one attempts to find a match.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Determine which relaxation methods to apply",
	"sec_num": null
	},
	{
	"text": "Complex relatlons speclfted in a speaker's descrlptlon include spatlal relations (e.g.. \"the outlet near the top of the tube\">, comparatives (e.g. \"the larger tube\") and superlatlves (e.g., \"the longest tube\").",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Determine which relaxation methods to apply",
	"sec_num": null
	},
	{
	"text": "These can be relaxed. The slmpler features of an object (such as slze or color) that are speclfied in the speaker's descrlptton are also open to relaxation.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Determine which relaxation methods to apply",
	"sec_num": null
	},
	{
	"text": "Often the objects in focus In the real world implicitly cause other objects to be In focus [14, 2{] ]. The subparts of an object ~n focus, for example, are reasonable candidates for the referent of a fazhng description and should be checked. At other times, the speaker might attribute features of a subpart of an object to the whole object (e.g., describing a plunger that Is composed of a red handle, a metal rod, a blue cap, and a green cup as \"the green plunger\").",
	"cite_spans": [
	{
	"start": 91,
	"end": 99,
	"text": "[14, 2{]",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Determine which relaxation methods to apply",
	"sec_num": null
	},
	{
	"text": "In these cases, the relaxation mechanism utilizes the part-whole relation in object descriptions to suggest a way to relax the speaker's description.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Determine which relaxation methods to apply",
	"sec_num": null
	},
	{
	"text": "Relaxation of a description has a few global strategies that can be followed for each part of the description: (I) drop the errorful feature value from the description altogether, (2) weaken or tighten the feature value but keep its new value close to the specified one, or (S) try some other feature value. Each of those procedures are specialists that attempt to first relax the feature value to one \"near\" the current one (e.g., one would prefer to first relax the color \"red\" to \"pink\" before relaxing it to \"blue\") and then. d that fails, to try relaxing it to any of the other possible values.",
	"cite_spans": [
	{
	"start": 111,
	"end": 114,
	"text": "(I)",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Determine which relaxation methods to apply",
	"sec_num": null
	},
	{
	"text": "If those fail. the feature would simply be ignored.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Determine which relaxation methods to apply",
	"sec_num": null
	},
	{
	"text": "This section describes how a referent identification system can handle a mlsreference using the scheme outlined in the previous section.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "For the purposes of thls example, assume that the water pump objects currently in focus include the CAP. the MAINTUBE. the AIRCHAMBER and the STAND (see Figure l{a) for a picture of these parts)",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 153,
	"end": 165,
	"text": "Figure l{a)",
	"ref_id": null
	}
	],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Assume also that the speaker tries to describe two of the objects. \". two devices that are clear piastlc One of them has two openings on the outside with threads on the end, and its about five inches long.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The other one tsa rounded piece with a turquoise base on it. Both are tubular.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The rounded piece fits loosely over...\". The reference system can find a unique referent for the first obJect but not for the second.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The relaxation algorithm will be shown below to reduce the set of referent candidates for the second description down to two.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "It. then. requires the system/listener to try out those candidates to determine if one. or both, fits loosely. The protocols exhibit a similar result when the listener uses \"fits loosely\" to get the correct referent (eg.. Excerpt 6 exemplifies where the \"fit\" can confirm that the proper referent was found). A representation of each of the water pump objects that are currently under consideration is presented in Figure 8 Each provides a physical description of the object -in terms of its dimensions, the basic 3-D shapes composing it, and its physical features -and a basic functional description of the obJect.",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 415,
	"end": 423,
	"text": "Figure 8",
	"ref_id": null
	}
	],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The first entry in each representation tn Figure 8 {that entry is shown in uppercase) defines the basic kind of entity being described {e.g.. \"TUBE\" means that the object being described is some kind of tube)",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 42,
	"end": 50,
	"text": "Figure 8",
	"ref_id": null
	}
	],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The words in mixed case refer to the names of features and the words in uppercase refer to possible fillers of those features from things in the water pump world.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The \"Subpart\" feature provides a place for an embedded description of an object that is a subpart of a parent object. Such subparts can be referred to on their own or as part of the parent object. The \"Orientation\" feature, used in the representations in Figure 8 . provides a rotation and translation of the object from some standard orientation to the oblects current orientation in 3-D space. The standard orientation provides a way to define relative positions such as \"top,\" \"bottom,\" or \"slde.\" At this point, the knowledge sources and their associated rules that were mentloned earlier apply. These rules attempt to order the feature values m the speaker's description for relaxation.",
	"cite_spans": [],
	"ref_spans": [
	{
	"start": 255,
	"end": 264,
	"text": "Figure 8",
	"ref_id": null
	}
	],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "First. we'll order the features m DescrZ using lingulstlc knowledge. Linguistic analysls of Deser2, \"... are clear plastlc ... a rounded pace wlth a turquoise base ... Both are tubular ... fits loosely over .... \" tells us that the features were specified using the following modifiers. Observations from the protocols (as described by the rules developed In [13] ) has shown that people tend to relax first features specified as adlectlves, then as preposltlonal phrases and finally as relative clauses or predicate complements.",
	"cite_spans": [
	{
	"start": 359,
	"end": 363,
	"text": "[13]",
	"ref_id": "BIBREF14"
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Thls suggests relaxation of Descr2 in the order:",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "]Shape} < \|Color.Subport\| < \|Tronsporency.COmpOSi t ;on.Analogical-ShaDe ,F; t \|",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The set of features on the left side of a \"<\" symbol is relaxed before the set on the rlght side The order that the features inside the braces. \")~\", are relaxed is left unspecified {i.e., any order of relaxation Is alrlght) Perceptual Information about the domain also provldes suggestlons.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Whenever a feature has feature values that are close, then one should be prepared to relax any of them to any of the others (we call thls the \"clustered feature value rule\")",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "[n thls example. smce the colors are all very close -BLUE. TURQUOISE, and VIOLET -then Color may be a reasonable thing to relax.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Hxerarchlcal Information about how closely related one feature value Is to another can also be used to determine what to relax. The Shape values are a good example. A CYLINDRICAL shape Is also a CONICAL shape, which Is also a 3-D ROUND shape.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Hence. It Is very reasonable to match ROUNDED to CYLINDRICAL. All of these suggestions can be put together to form the order: ~Sho~e.Co~or\| < ~Su~l)art~ < \|Trangporeflcy ,Compos i t i on.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Ana\| og i ca I--Shope. F i I: \|.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The referent candldates MAINTUBE. STAND, and .41R CHAMBER can be examined and possibly ordered for relaxation using the above feature ordering For this example, the relaxation of Descr2 to any of the candidates requires relaxing their SHAPE and COLOR features.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Since they each require reiaxmg the same features, the candidates can not be ordered w, th respect to each other (i.e., none of the possible feature orders is better for relaxing the candidates).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Hence. no one candidate stands out as the most likely referent.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "While no orderlng of the candidates was posslble. the order generated to relax the features In the speaker's description can be used to guide the relaxation, of each candldate.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The relaxation methods mentioned at the end of the last section come Into use here.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Generate-Simdar-Shape-Values can determine that HEMISPHERICAL and CYLINDRICAL shapes of the AIR CHAMBER are close to the 3D-ROUND shape.. This holds equally true for the cyhndrlcal shapes of the MAINTUBE and the STAND.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Generate-Similar-Color-Values next trms relaxing the Color TURQUOISE.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "It determmes the colors BLUE and GREEN as the best alternates.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Here only two clear winners exist -the AIR CHAMBER and the STAND -while the MAINTUBE is dropped as a candidate smce it Is reasonable to relax TURQUOISE to BLUE or to GREEN but not to VIOLET Subpart, Transparency, Analoglcal-Shape, and Composition provide no further help {though. the fact that the AIR CHAMBER has both CLEAR and OPAQUE subparts mght put it slightly lower than the ST,hVD whose subparts are all CLEAR. Thls difference. however, is not slgndicant.).",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Thls leaves trial and error attempts to try to complete the FIT action. The one (if any) that fits -and fits loosely -Is selected as the referent.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "The protocols showed that people often do just that -reducing their set of choices down as best they can and then taking each of the remalnmg chmces and trying out the requested action on them",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "An example on handling a misreference",
	"sec_num": "3.3"
	},
	{
	"text": "Our goal m thls work Is to budd robust natural language understanding systems, allowmg them to detect and avold mlscommunlcatlon.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "The goal is not to make a perfect listener but a more tolerant one that could avold many mistakes, though still wrong on occasion.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "In Section 2, we mtroduced a taxonomy of mlscommunlcatlon problems that occur tn expertapprentice dialogues. We showed that reference mistakes are one kind of obstacle to robust communication.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "To tackle reference problems, we descrlbed how to extend the succeed/fad paradigm followed by previous natural language researchers",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "We represented real world objects hlerarchlcallv in a knowledge base using a representation language, KL-One.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "that follows in the tradition of semantlc networks and frames.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "In such a representatlon framework, the reference identification task looks for a referent by comparing the representation of the speakers Input to elements in the knowledge base by using a matching procedure.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "Failure to find a referent in previous reference identlhcatlon systems resulted In the unsuccessful termination of the reference task We claim that people behave better than this and exphcltly illustrated such cases in an expert-apprentlce domain about toy water pumps.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "We developed a theory of relaxation for recovering from reference failures that provides a much better model for human performance. When people are asked to identify objects, they go about it m a certain way. flnd candidates, adjust as necessary, re-try, and, if necessary, glve up and ask for help. We claim that relaxation is an Integral part of this process and that the particular parameters of relaxation differ from task to task and person to person.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "Our work models the relaxation process and provldes a computatlonal model for experimenting w~th the different parameters.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "The theory incorporates the same language and physical knowledge that people use m performing reference identification to guide the relaxation process. Thls knowledge Is represented as a set of rules and as data m a hierarchical knowledge base. Rule-based relaxation provided a methodical way to use knowledge about language and the world to find a referent.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "The hlererchxcal representation made It posslble to tackle issues of Impreclslon and overspecification In a speakers description.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "It allows one to check the position of a description in the hierarchy and to use that position to fudge Imprecision and over-speclfication and to suggest possible repairs to the descriptlon.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "Interestingly. one would expect that \"closest\" match would suffice to solve the problem of finding a referent.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "We showed, however, that it doesn't usually provide you with the correct referent.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "Closest match isn't sufficient because there are many features associated wlth an object and, thus. determimng whlch of those features to keep and which to drop Is a difficult problem due to the combinatorlcs and the effects of context.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	},
	{
	"text": "The relaxation method described circumvents the problem by using the knowledge that people have about language and the physical world to prune down the search space.",
	"cite_spans": [],
	"ref_spans": [],
	"eq_spans": [],
	"section": "Conclusion",
	"sec_num": "4"
	}
	],
	"back_matter": [
	{
	"text": "(C~mpol, t ion PI, ~I~T The above analysls using the partial matcher provldes no clear winner smce the differences are so close causing the scores generated for the candldates to be almost exactly the same (i.e.. the only difference was In the score for Transparency).All candidates. hence, will be retained for now.",
	"cite_spans": [
	{
	"start": 19,
	"end": 23,
	"text": "~I~T",
	"ref_id": null
	}
	],
	"ref_spans": [],
	"eq_spans": [],
	"section": "annex",
	"sec_num": null
	}
	],
	"bib_entries": {
	"BIBREF0": {
	"ref_id": "b0",
	"title": "A Plan-Based Approach to Speech Act Recognztion",
	"authors": [
	{
	"first": "James",
	"middle": [
	"F"
	],
	"last": "Allen",
	"suffix": ""
	}
	],
	"year": 1979,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Allen. James F. A Plan-Based Approach to Speech Act Recognztion. Ph.D. Th.. University of Toronto. 1979.",
	"links": null
	},
	"BIBREF1": {
	"ref_id": "b1",
	"title": "Planning AIat~Tal Language Utterances to Satisf!/ Multiple Goals",
	"authors": [
	{
	"first": "Douglas",
	"middle": [
	"E"
	],
	"last": "Appelt",
	"suffix": ""
	}
	],
	"year": 1981,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Appelt, Douglas E. Planning AIat~Tal Language Utterances to Satisf!/ Multiple Goals. Ph.D Th., Stanford Unlverslty, 1981.",
	"links": null
	},
	"BIBREF2": {
	"ref_id": "b2",
	"title": "A 3tr~ctural Paradigm /or Representing Knowledge",
	"authors": [
	{
	"first": "Ronald",
	"middle": [
	"J"
	],
	"last": "Brachman",
	"suffix": ""
	}
	],
	"year": 1977,
	"venue": "Also",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Brachman, Ronald J. A 3tr~ctural Paradigm /or Representing Knowledge. Ph.D. Th., Harvard Umverslty, 1977 Also, Technlcal Report No. 3605. Bolt Beranek and Newman Inc.",
	"links": null
	},
	"BIBREF3": {
	"ref_id": "b3",
	"title": "Repair Theory A Generative Theory of Bugs m Procedural Sk111s",
	"authors": [
	{
	"first": "John",
	"middle": [],
	"last": "Brown",
	"suffix": ""
	},
	{
	"first": "Kurt",
	"middle": [],
	"last": "Seely",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Vanlehn",
	"suffix": ""
	}
	],
	"year": 1980,
	"venue": "",
	"volume": "4",
	"issue": "",
	"pages": "379--426",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Brown, John Seely and Kurt VanLehn. \"Repair Theory A Generative Theory of Bugs m Procedural Sk111s.\" Cognitive Science ~, 4 (1980), 379-426",
	"links": null
	},
	"BIBREF4": {
	"ref_id": "b4",
	"title": "On Kno~vlng What to Sa?l. Planning Speech Acts",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Cohen",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Philip R",
	"suffix": ""
	}
	],
	"year": 1978,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Cohen. Philip R On Kno~vlng What to Sa?l. Planning Speech Acts. Ph.D. Th., University of Toronto, 1978.",
	"links": null
	},
	"BIBREF6": {
	"ref_id": "b6",
	"title": "Beyond Question Answering",
	"authors": [
	{
	"first": "C",
	"middle": [],
	"last": "Perrault",
	"suffix": ""
	},
	{
	"first": "J",
	"middle": [],
	"last": "Allen",
	"suffix": ""
	}
	],
	"year": null,
	"venue": "KnowLedge Representation and Natural Language Processing. W. Lehnart and M. Ringle",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "C Perrault and J. Allen. Beyond Question Answering. In KnowLedge Representation and Natural Language Processing. W. Lehnart and M. Ringle, Ed..",
	"links": null
	},
	"BIBREF7": {
	"ref_id": "b7",
	"title": "The need for Referent Identlficatlon as a Planned Actlon",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Cohen",
	"suffix": ""
	},
	{
	"first": "R",
	"middle": [],
	"last": "Phlhp",
	"suffix": ""
	}
	],
	"year": 1981,
	"venue": "Proceedings of IJCAI-81",
	"volume": "",
	"issue": "",
	"pages": "31--35",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Cohen. Phlhp R. The need for Referent Identlficatlon as a Planned Actlon. Proceedings of IJCAI-81. Vancouver. B.C., Canada, August. 1981, pp. 31-35.",
	"links": null
	},
	"BIBREF8": {
	"ref_id": "b8",
	"title": "Dependencies of Discourse Structure on the Modahty of Communlcatlon. Telephone vs. Teletype. Proceedings of ACL",
	"authors": [
	{
	"first": "Phlhp",
	"middle": [
	"R"
	],
	"last": "Cohen",
	"suffix": ""
	},
	{
	"first": "Scott",
	"middle": [],
	"last": "Fertlg",
	"suffix": ""
	},
	{
	"first": "Kathy",
	"middle": [],
	"last": "Start",
	"suffix": ""
	}
	],
	"year": 1982,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "28--35",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Cohen, Phlhp R, Scott Fertlg and Kathy Start. Dependencies of Discourse Structure on the Modahty of Communlcatlon. Telephone vs. Teletype. Proceedings of ACL, Toronto. Ont., Canada, June, 1982, pp. 28-35.",
	"links": null
	},
	"BIBREF9": {
	"ref_id": "b9",
	"title": "The Pragmatlcs of Referring and the Modahty of Communlcatlon",
	"authors": [
	{
	"first": "Philip",
	"middle": [
	"R"
	],
	"last": "Cohen",
	"suffix": ""
	}
	],
	"year": 1984,
	"venue": "Computational Linguistics",
	"volume": "10",
	"issue": "",
	"pages": "97--146",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Cohen, Philip R. \"The Pragmatlcs of Referring and the Modahty of Communlcatlon.\" Computational Linguistics 10, 2 (April-June 1984). 97-146.",
	"links": null
	},
	"BIBREF10": {
	"ref_id": "b10",
	"title": "The Structure of Analogical Models In Science. Bolt Beranek and",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Gentner",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Dedre",
	"suffix": ""
	}
	],
	"year": null,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Gentner. Dedre. The Structure of Analogical Models In Science. Bolt Beranek and Newman Inc..",
	"links": null
	},
	"BIBREF11": {
	"ref_id": "b11",
	"title": "Mlscommunlcatlon an Task-Oriented Dialogues",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Goodman",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Bradley",
	"suffix": ""
	}
	],
	"year": 1982,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Goodman. Bradley A. Mlscommunlcatlon an Task- Oriented Dialogues. KRNL Group Working Paper, Bolt BeraneK and Newman Inc., April 1982.",
	"links": null
	},
	"BIBREF12": {
	"ref_id": "b12",
	"title": "Repairing Miscommunlcatlon: Relaxation m Reference",
	"authors": [
	{
	"first": "Bradley",
	"middle": [
	"A"
	],
	"last": "Goodman",
	"suffix": ""
	}
	],
	"year": null,
	"venue": "Proceedings of AAAI-83",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Goodman, Bradley A. Repairing Miscommunlcatlon: Relaxation m Reference. Proceedings of AAAI-83.",
	"links": null
	},
	"BIBREF14": {
	"ref_id": "b14",
	"title": "Communication and Miscomccr,~n~cation",
	"authors": [
	{
	"first": "Bradley",
	"middle": [
	"A"
	],
	"last": "Goodman",
	"suffix": ""
	}
	],
	"year": 1984,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Goodman, Bradley A. Communication and Miscomccr,~n~cation. Ph.D. Th., University of Illinols. Urbane, 1984.",
	"links": null
	},
	"BIBREF15": {
	"ref_id": "b15",
	"title": "University of Californla, Berkeley. 1977. Also, Technical Note 151",
	"authors": [
	{
	"first": "Barbara",
	"middle": [
	"J"
	],
	"last": "Gross",
	"suffix": ""
	}
	],
	"year": null,
	"venue": "Ph.D. Th",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Gross, Barbara J. The Representation and Use of Focus in Dialogue Under'standing. Ph.D. Th., University of Californla, Berkeley. 1977. Also, Technical Note 151. Stanford Research Instltute.",
	"links": null
	},
	"BIBREF16": {
	"ref_id": "b16",
	"title": "Focusing and descriptions in natural language dialogues",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Gross",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [
	"J"
	],
	"last": "Barbara",
	"suffix": ""
	}
	],
	"year": 1981,
	"venue": "Elements of Discourse Understanding, Joshi. Webber and Sags",
	"volume": "",
	"issue": "",
	"pages": "84--89",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Gross. Barbara J. Focusing and descriptions in natural language dialogues. In Elements of Discourse Understanding, Joshi. Webber and Sags, Ed.,Cambrldge University Press, 1981, pp. 84-ID5.",
	"links": null
	},
	"BIBREF17": {
	"ref_id": "b17",
	"title": "Bolt Beranek and Newman Inc. Also Consul Note # 5, USC/Information Sciences Institute",
	"authors": [
	{
	"first": "Thomas. A ](l-One",
	"middle": [],
	"last": "Lipkis",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Classifier",
	"suffix": ""
	}
	],
	"year": 1981,
	"venue": "Proceedings of the 1981 KL-One Workshop",
	"volume": "",
	"issue": "",
	"pages": "128--145",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Lipkis, Thomas. A ](L-ONE Classifier. Proceedings of the 1981 KL-One Workshop, June, 1982, pp. 128-145. Report No. 4842, Bolt Beranek and Newman Inc. Also Consul Note # 5, USC/Information Sciences Institute. October 1981.",
	"links": null
	},
	"BIBREF18": {
	"ref_id": "b18",
	"title": "A Plan Recogmtion Model for Clarlfication Subdialogues",
	"authors": [
	{
	"first": "Diane",
	"middle": [
	"J"
	],
	"last": "L4tman",
	"suffix": ""
	},
	{
	"first": "James",
	"middle": [
	"F"
	],
	"last": "Allen",
	"suffix": ""
	}
	],
	"year": 1984,
	"venue": "Proceedings of Coling84, Stanford Umverslty",
	"volume": "",
	"issue": "",
	"pages": "302--311",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "L4tman, Diane J. and James F. Allen. A Plan Recogmtion Model for Clarlfication Subdialogues. Proceedings of Coling84, Stanford Umverslty, Stanford, CA., July, 1984, pp. 302-311.",
	"links": null
	},
	"BIBREF19": {
	"ref_id": "b19",
	"title": "Realization. Proceedings of the",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Mark",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "William",
	"suffix": ""
	}
	],
	"year": 1981,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "78--89",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Mark. William. Realization. Proceedings of the 1981 ](L-One Workshop, June, 1982, pp. 78-89. Report No. 4842, Bolt Beranek and Newman Inc.",
	"links": null
	},
	"BIBREF20": {
	"ref_id": "b20",
	"title": "Recurslon In Text and Its Use in Language Generation",
	"authors": [
	{
	"first": "]",
	"middle": [],
	"last": "Mckeown",
	"suffix": ""
	},
	{
	"first": "R",
	"middle": [],
	"last": "(athleen",
	"suffix": ""
	}
	],
	"year": null,
	"venue": "Proceedings of AAAI-83",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "McKeown, ](athleen R. Recurslon In Text and Its Use in Language Generation. Proceedings of AAAI-83.",
	"links": null
	},
	"BIBREF22": {
	"ref_id": "b22",
	"title": "Conversational Coherency",
	"authors": [
	{
	"first": "Rachel",
	"middle": [],
	"last": "Relchman",
	"suffix": ""
	}
	],
	"year": 1978,
	"venue": "Cognitive Science",
	"volume": "2",
	"issue": "4",
	"pages": "283--327",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Relchman, Rachel. \"Conversational Coherency.\" Cognitive Science 2. 4 (1978). 283-327.",
	"links": null
	},
	"BIBREF23": {
	"ref_id": "b23",
	"title": "Plain Speaking: A Theory and Grammar of 3pontaneo~s Discourse",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Relchman",
	"suffix": ""
	}
	],
	"year": null,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Relchman. Rachel. Plain Speaking: A Theory and Grammar of 3pontaneo~s Discourse. Ph.D Th..",
	"links": null
	},
	"BIBREF24": {
	"ref_id": "b24",
	"title": "Also",
	"authors": [
	{
	"first": "Harvard",
	"middle": [],
	"last": "Umverslty",
	"suffix": ""
	}
	],
	"year": 1981,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Harvard Umverslty, 1981. Also, Technical Report No. 4861, Bolt Beranek and Newman Inc.",
	"links": null
	},
	"BIBREF25": {
	"ref_id": "b25",
	"title": "Recogmzmg antended meamng and speaker's plans",
	"authors": [
	{
	"first": "",
	"middle": [],
	"last": "Ringle",
	"suffix": ""
	},
	{
	"first": "Bertram",
	"middle": [],
	"last": "Martin",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [
	"C L"
	],
	"last": "Bruce ; ~3] Sidner",
	"suffix": ""
	},
	{
	"first": "D",
	"middle": [
	"J"
	],
	"last": "Israel",
	"suffix": ""
	}
	],
	"year": 1981,
	"venue": "Proceedings of the Internatlone, l Joint Conference In Artlfictal Intelhgence. The International Joint Conferences on Artlficai Intelligence",
	"volume": "",
	"issue": "",
	"pages": "203--208",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Ringle. Martin and Bertram Bruce. Conversation Failure. In Knowledge Representation and Hatlzral Language Processing, W. Lehnart and M. RIngle. Ed.,Lawrence Erlbaum Assocmtes, 1981. [:~3] Sidner. C L.. and Israel, D.J. Recogmzmg antended meamng and speaker's plans. Proceedings of the Internatlone, l Joint Conference In Artlfictal Intelhgence. The International Joint Conferences on Artlficai Intelligence. Vancouver. B.C. August. 1981, pp. 203-208.",
	"links": null
	},
	"BIBREF26": {
	"ref_id": "b26",
	"title": "To,yards cz Computational Theory of Definite AnaphoTa Comprehension in English Discourse",
	"authors": [
	{
	"first": "Candace",
	"middle": [],
	"last": "Sldner",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Lee",
	"suffix": ""
	}
	],
	"year": 1979,
	"venue": "Ph.D. Th",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Sldner, Candace Lee. To,yards cz Computational Theory of Definite AnaphoTa Comprehension in English Discourse. Ph.D. Th., Massachusetts Instltute of Technology, 1979. Also, Report No. TR-537, MIT AI Lab",
	"links": null
	},
	"BIBREF28": {
	"ref_id": "b28",
	"title": "Research an Knowledge Representatlon for Natural Language Understanding Report No. 4785, Bolt Beranek and",
	"authors": [
	{
	"first": "M",
	"middle": [],
	"last": "Bates",
	"suffix": ""
	},
	{
	"first": "R",
	"middle": [
	"J"
	],
	"last": "Bobrow",
	"suffix": ""
	},
	{
	"first": "R",
	"middle": [
	"J"
	],
	"last": "Brachman",
	"suffix": ""
	},
	{
	"first": "P",
	"middle": [
	"R"
	],
	"last": "Cohen",
	"suffix": ""
	},
	{
	"first": "D",
	"middle": [
	"J"
	],
	"last": "Israel",
	"suffix": ""
	},
	{
	"first": "J",
	"middle": [
	"B L"
	],
	"last": "Schmolze",
	"suffix": ""
	},
	{
	"first": "W",
	"middle": [
	"A"
	],
	"last": "Webber",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Woods",
	"suffix": ""
	}
	],
	"year": null,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "M. Bates, R. J. Bobrow, R. J. Brachman, P. R. Cohen, D. J. Israel, J. Schmolze. B. L. Webber, W. A. Woods. Research an Knowledge Representatlon for Natural Language Understanding Report No. 4785, Bolt Beranek and Newman Inc..",
	"links": null
	},
	"BIBREF36": {
	"ref_id": "b36",
	"title": "Research an Knowledge Representation for Natural Language Understanding -Annual Report. I",
	"authors": [
	{
	"first": "M",
	"middle": [],
	"last": "Vilem",
	"suffix": ""
	}
	],
	"year": 1982,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Vilem, M. Research an Knowledge Representation for Natural Language Understanding - Annual Report. I September 1982 -31 August 1983. Technical Report 5421. BBN Laboratories. Cambradge. MA, 1983.",
	"links": null
	},
	"BIBREF41": {
	"ref_id": "b41",
	"title": "Research m Knowledge Representation for Natural Language Understanding -Annual Report, I",
	"authors": [
	{
	"first": "M",
	"middle": [],
	"last": "Vilem",
	"suffix": ""
	}
	],
	"year": 1983,
	"venue": "BBN Laboratorles Inc",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Vilem, M. Research m Knowledge Representation for Natural Language Understanding - Annual Report, I September 1983 -31 August 1984. Technical Report 5894. BBN Laboratorles Inc., Cambrldge, MA, 1984.",
	"links": null
	},
	"BIBREF42": {
	"ref_id": "b42",
	"title": "A Forma~ App~'oach to /~.scourse Anapho1\"a",
	"authors": [
	{
	"first": "Bonnle",
	"middle": [],
	"last": "Webber",
	"suffix": ""
	},
	{
	"first": "",
	"middle": [],
	"last": "Lynn",
	"suffix": ""
	}
	],
	"year": null,
	"venue": "",
	"volume": "",
	"issue": "",
	"pages": "",
	"other_ids": {},
	"num": null,
	"urls": [],
	"raw_text": "Webber, Bonnle Lynn. A Forma~ App~'oach to /~.scourse Anapho1\"a. Ph.D. Th., Harvard University.",
	"links": null
	}
	},
	"ref_entries": {
	"FIGREF0": {
	"uris": null,
	"text": "Figure I:",
	"num": null,
	"type_str": "figure"
	},
	"FIGREF1": {
	"uris": null,
	"text": "Face-to-Face) S 1. And place 2. the blue cap that's left [P takes CAP]",
	"num": null,
	"type_str": "figure"
	},
	"FIGREF2": {
	"uris": null,
	"text": "Approaches to reference ]dentdlcatlon",
	"num": null,
	"type_str": "figure"
	},
	"FIGREF3": {
	"uris": null,
	"text": ". The SuperC from Concept B to Concept A is like stating BCA for two sets A and B An INDIVIDUAL CONCEPT ts used to guarantee that the subset speclhed by a concept Is unique The [ndlvldual Concept D shown m the figure Is dehned to be a unique member of the subset specified by Concept C ROLEs on concepts are like normal attributes and slot",
	"num": null,
	"type_str": "figure"
	},
	"FIGREF4": {
	"uris": null,
	"text": "quo I (syntoc t ic-f orm(v t .d), \"ADJ\"). ;'quo I (synt a\u00a2t ic-f orm(v2.d), \"REL-CLS\") A sample relaxation rule Each knowledge source produces ~ts own partial ordermg of features.",
	"num": null,
	"type_str": "figure"
	},
	"FIGREF5": {
	"uris": null,
	"text": "pm-c~al FI -~ \u00a2o1\u00a2-f~ tl oe fz P~ \u00a2 -) N\|eeet.\u00a2tnlceJ f3 -) F~I:I~ f2 F3 fZ oe f~ oe F, F4 -) Size f3 fa f4 Ct~ of/~rtu.s I~ ,*~,~r~;~ Figure 8:",
	"num": null,
	"type_str": "figure"
	},
	"FIGREF6": {
	"uris": null,
	"text": "provides a simplified and hnearlzed vlew of the actual KL-One representatlon of the speaker'",
	"num": null,
	"type_str": "figure"
	},
	"TABREF0": {
	"text": ",=scribes the two holes on the BAjEVALVE as \"the httle hoie\"",
	"html": null,
	"num": null,
	"content": "<table><tr><td colspan=\"8\">The water pump Is reasonably</td><td colspan=\"2\">complex, containing</td><td>four</td></tr><tr><td colspan=\"4\">subassemblies</td><td>that</td><td/><td>are</td><td colspan=\"2\">built</td><td>from</td><td>plastic</td><td>tubes,</td></tr><tr><td colspan=\"10\">nozzles, valves, plungers, and caps that can be screwed</td></tr><tr><td colspan=\"3\">or pushed</td><td colspan=\"3\">together.</td><td/><td colspan=\"2\">A large</td><td>corpus</td><td>of dialogues</td></tr><tr><td colspan=\"3\">concerning</td><td>thls</td><td colspan=\"2\">task</td><td colspan=\"2\">was</td><td colspan=\"2\">collected</td><td>by Cohen</td><td>(see</td></tr><tr><td colspan=\"10\">[7. 8. 9]). These dialogues contained</td><td>instructions</td><td>from</td></tr><tr><td>an</td><td colspan=\"2\">\"expert\"</td><td colspan=\"2\">to</td><td>an</td><td colspan=\"4\">\"apprentice\"</td><td>that</td><td>explain</td><td>the</td></tr><tr><td colspan=\"10\">assembly of the toy water pump,</td><td>Both participants</td></tr><tr><td>were</td><td colspan=\"3\">working</td><td>to</td><td colspan=\"3\">achieve</td><td>a</td><td>common</td><td>goal</td><td>-</td><td>the</td></tr><tr><td colspan=\"3\">successful</td><td colspan=\"6\">assembly of the pump</td><td>Thls domain Is rlch</td></tr><tr><td>m</td><td colspan=\"3\">perceptual</td><td colspan=\"4\">information,</td><td colspan=\"2\">allowing</td><td>for</td><td>complex</td></tr><tr><td colspan=\"3\">descriptions</td><td colspan=\"4\">of elements</td><td colspan=\"2\">in it.</td><td>The</td><td>data</td><td>provide</td></tr><tr><td colspan=\"10\">examples of imprecision, confusion, and ambiguity as</td></tr><tr><td colspan=\"10\">we!l as attempts to correct these problems</td></tr><tr><td/><td colspan=\"2\">The</td><td colspan=\"3\">following</td><td colspan=\"3\">exchange</td><td>exemplifies</td><td>one</td><td>such</td></tr><tr><td colspan=\"3\">situation.</td><td colspan=\"7\">Here A Is instructing J to assemble part of</td></tr><tr><td colspan=\"4\">the water pump.</td><td colspan=\"6\">Refer to Figure l(a) for a picture</td><td>of</td></tr><tr><td colspan=\"3\">the pump.</td><td colspan=\"7\">A and J are communicating</td><td>verbally</td><td>but</td></tr><tr><td colspan=\"2\">neither</td><td colspan=\"8\">can see the other. (The bracketed</td><td>text In the</td></tr><tr><td colspan=\"2\">excerpt</td><td colspan=\"7\">tells what was actually</td><td>occurring</td><td>while each</td></tr><tr><td colspan=\"2\">utterance</td><td colspan=\"5\">was spoken.)</td><td colspan=\"3\">Notlce the complexity</td><td>of the</td></tr><tr><td colspan=\"2\">speaker's</td><td colspan=\"4\">descriptions</td><td/><td>and</td><td>the</td><td>resultant</td><td>processing</td></tr><tr><td colspan=\"7\">required by the listener,</td><td colspan=\"3\">Thls dialogue illustrates</td><td>when</td></tr><tr><td colspan=\"2\">listeners</td><td colspan=\"2\">repair</td><td colspan=\"4\">the speakers</td><td colspan=\"2\">description</td><td>in order to</td></tr><tr><td colspan=\"4\">flnd a referent,</td><td colspan=\"6\">when they repair their mztlal reference</td></tr><tr><td colspan=\"10\">choice once they are given more information,</td><td>and when</td></tr><tr><td colspan=\"6\">they fall t ~. choose</td><td colspan=\"2\">a proper</td><td colspan=\"2\">referent</td><td>In Linp 7, A</td></tr><tr><td>:[</td><td/><td/><td/><td/><td/><td/><td/><td/></tr></table>",
	"type_str": "table"
	},
	"TABREF1": {
	"text": "the SL/DEVALVE flts, J doesn't L~nes 27 and 28, J indicates that A dld not glve him enough znformatlon to perform the requested . that has two little teeth sticking 3. out of the bottom of it.",
	"html": null,
	"num": null,
	"content": "<table><tr><td/><td/><td/><td/><td/><td colspan=\"2\">A. 5. Okay</td><td>On that take the</td></tr><tr><td/><td/><td/><td/><td/><td/><td>6. brlght shocking pink piece of plastic</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">[J takes BASEVALVE]</td></tr><tr><td/><td/><td/><td/><td/><td/><td>7. and stick the little hole over the</td></tr><tr><td/><td/><td/><td/><td/><td/><td>teeth.</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">[J starts to install the BASEVALVE. backs off, looks</td></tr><tr><td/><td/><td/><td/><td/><td/><td>at it again and then goes ahead and</td></tr><tr><td/><td/><td/><td/><td/><td/><td>installs it]</td></tr><tr><td/><td/><td/><td/><td/><td/><td>J. 8 Okay</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">A: 9 Now screw that blue cap onto</td></tr><tr><td/><td/><td/><td/><td/><td/><td>I0. the bottom of the maln tube.</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">[J screws TUBEBASE onto MAINTUBE]</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">J. 11. Okay</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">A. 12 Now. there's a--</td></tr><tr><td/><td/><td/><td/><td/><td/><td>13. a red plastic piece</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">[J starts for NOZZLE]</td></tr><tr><td/><td/><td/><td/><td/><td/><td>14 that has four gizmos on It.</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">[J switches to SLIDEVALVE]</td></tr><tr><td/><td/><td/><td/><td/><td>J.</td><td>15 Yes.</td></tr><tr><td/><td/><td/><td/><td/><td>A</td><td>16 Okay</td><td>Put the ungtzmoed end In the</td></tr><tr><td/><td/><td/><td/><td/><td/><td>uh</td></tr><tr><td/><td/><td/><td/><td/><td/><td>17 the other--the open</td></tr><tr><td/><td/><td/><td/><td/><td/><td>18 part</td><td>of the</td><td>maln</td><td>tube,</td><td>the</td><td>lower</td></tr><tr><td/><td/><td/><td/><td/><td/><td>valve</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">[3 puts SLIDEVALVE into hole in TUBEBASE, but A</td></tr><tr><td/><td/><td/><td/><td/><td/><td>meant OUTLET2 of MAINTUBE]</td></tr><tr><td/><td/><td/><td/><td/><td>I</td><td>19 All right</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">A 20 !t ;ust hts loosely</td><td>It .doesnt</td></tr><tr><td/><td/><td/><td/><td/><td/><td>'~I have to f'.t right. Okay. then take</td></tr><tr><td/><td/><td/><td/><td/><td/><td>.~2 the clear plastic elbow ]omt</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">[J takes SPOUT]</td></tr><tr><td/><td/><td/><td/><td/><td>J</td><td>23 All right</td></tr><tr><td/><td/><td/><td/><td/><td>A</td><td>$4 And put tt over the bottom opening,</td></tr><tr><td/><td/><td/><td/><td/><td/><td>too.</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">[J trees installing SPOUT on TI/BEBASE]</td></tr><tr><td/><td/><td/><td/><td/><td>l</td><td>-~ Okay</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"2\">a. 28. Okay</td><td>Now. take the--</td></tr><tr><td/><td/><td/><td/><td/><td/><td>27 Which end am I supposed</td><td>to put It</td></tr><tr><td/><td/><td/><td/><td/><td/><td>over')</td></tr><tr><td>detect</td><td>any trouble.</td><td colspan=\"3\">L~nes 20 and 21 keep I from</td><td/><td>28 Do you know \u00b0</td></tr><tr><td colspan=\"4\">thinking that something is wrong because</td><td>the part fits</td><td/></tr><tr><td colspan=\"3\">loosely, In action. In Lme 30. J further</td><td>compounds</td><td>the error in</td><td colspan=\"2\">A. -:'9 Put the--put the--the big end--</td></tr><tr><td/><td/><td/><td/><td/><td/><td>30 the blg end over it.</td></tr></table>",
	"type_str": "table"
	},
	"TABREF3": {
	"text": ": 8. now take the clear little 9. attachment 10. and put on the hole where you 11. just put the red cap on 12. make sure it points 13. upward",
	"html": null,
	"num": null,
	"content": "<table><tr><td>R: 7. ok</td><td/><td/><td/><td/><td/></tr><tr><td>R: 14. ok</td><td/><td/><td/><td/><td/></tr><tr><td>F, xeerpt 5 (Teletype)</td><td/><td/><td/><td/><td/></tr><tr><td>S; I. Ok,</td><td/><td/><td/><td/><td/></tr><tr><td>2. put the red nozzle on the outlet</td><td/><td/><td/><td/><td/></tr><tr><td>3. of the rounded clear chamber</td><td/><td/><td/><td/><td/></tr><tr><td>4. ok?</td><td/><td/><td/><td/><td/></tr><tr><td/><td colspan=\"5\">'s another funny little</td></tr><tr><td/><td colspan=\"2\">4. red thing, a</td><td/><td/><td/></tr><tr><td>[A is</td><td>confused,</td><td colspan=\"2\">examines</td><td>both</td><td>NOZZLE</td><td>and</td></tr><tr><td/><td colspan=\"2\">SX.,mr-VALVE ]</td><td/><td/><td/></tr><tr><td/><td colspan=\"5\">5. little teeny red thing that's</td></tr><tr><td/><td colspan=\"2\">6. some--should</td><td colspan=\"3\">be somewhere on</td></tr><tr><td/><td colspan=\"5\">7. the desk, that has um--there's</td></tr><tr><td/><td colspan=\"4\">8. like teeth on one end.</td><td/></tr><tr><td colspan=\"2\">[E takes SLIDEVALVE]</td><td/><td/><td/><td/></tr><tr><td/><td>A: 9. Okay.</td><td/><td/><td/><td/></tr><tr><td/><td colspan=\"4\">E: 10. It's a funny-loo--hollow,</td><td/></tr><tr><td/><td colspan=\"5\">11. hollow projection on one end</td></tr><tr><td/><td colspan=\"5\">12. and then teeth on the other.</td></tr><tr><td/><td colspan=\"4\">Excerpt 4 (Teletype)</td><td/></tr><tr><td/><td colspan=\"5\">A: I. take the red thing with the</td></tr><tr><td/><td colspan=\"2\">2. prongs on it</td><td/><td/><td/></tr><tr><td/><td colspan=\"5\">3. and fit it onto the other hole</td></tr><tr><td/><td colspan=\"2\">4. of the cylinder</td><td/><td/><td/></tr><tr><td/><td colspan=\"4\">5. so that the prongs are</td><td/></tr><tr><td/><td colspan=\"2\">6. sticking out</td><td/><td/><td/></tr><tr><td>2O7</td><td/><td/><td/><td/><td/></tr></table>",
	"type_str": "table"
	},
	"TABREF4": {
	"text": "4. on the cylinder, [P lays down TUBEBASE]5. the side hole that is farthest 6. from the green end. [P puts CAP on OUTLET! of MAINTUBE] P: 7. Okay.",
	"html": null,
	"num": null,
	"content": "<table><tr><td colspan=\"2\">S; 8. And take the nozzle-looking</td></tr><tr><td>9. piece,</td><td/></tr><tr><td>[P grabs NOZZLE]</td><td/></tr><tr><td>10. no</td><td/></tr><tr><td colspan=\"2\">11. I mean the clear plastic one,</td></tr><tr><td>[P takes SPOUT]</td><td/></tr><tr><td colspan=\"2\">12. and place it on the other hole</td></tr><tr><td>[P identifies O~</td><td>of MA1NTUBE]</td></tr><tr><td colspan=\"2\">13. that's left,</td></tr><tr><td colspan=\"2\">14. so that nozzle points away</td></tr><tr><td>15. from the</td><td/></tr><tr><td colspan=\"2\">[P installs SPOUT on OUTLET2 of MAINTUBE]</td></tr><tr><td>16. right.</td><td/></tr><tr><td>P: 17. Okay.</td><td/></tr><tr><td>S: 18. Now</td><td/></tr><tr><td>19. take the</td><td/></tr><tr><td colspan=\"2\">20. cap base thing</td></tr><tr><td>[P takes TUBEBASE]</td><td/></tr><tr><td colspan=\"2\">21. and screw it onto the bottom,</td></tr><tr><td colspan=\"2\">[P sorewsTUBEBASE on)L~3NTUBE]</td></tr><tr><td>22, ooops,</td><td/></tr><tr><td colspan=\"2\">[S realizes she has forgotten to have P put</td></tr><tr><td colspan=\"2\">SLIDL~ALVE into OUTLET2 of MAINTUBE]</td></tr><tr><td colspan=\"2\">23. un-undo the pies</td></tr><tr><td colspan=\"2\">[P starts to take TUBEBASE off MAINTUBE]</td></tr><tr><td>24. no</td><td/></tr><tr><td colspan=\"2\">25. the clear plastic thing that I</td></tr><tr><td colspan=\"2\">26. told you to put on</td></tr><tr><td>[P removes SPOUT]</td><td/></tr><tr><td>27. sorry.</td><td/></tr><tr><td colspan=\"2\">28. And place the little red thing</td></tr><tr><td>[P takes $LID~ALVZ]</td><td/></tr><tr><td colspan=\"2\">29. tn there first,</td></tr><tr><td colspan=\"2\">[P mserts SLXD~ALVZ into OUTLET~ of M[AINT~E]</td></tr></table>",
	"type_str": "table"
	},
	"TABREF5": {
	"text": "And put the small end into 29. that hole on the air tube--30. on the big tube. ; 31. On the very bottom? starts to put it into the bottom hole of TUBEBASE -though he indicates he is unsure of himself]",
	"html": null,
	"num": null,
	"content": "<table><tr><td/><td/><td>hcerpt</td><td>7 (Telephone)</td><td>S: 32. On the bottom, 33. Yes.</td></tr><tr><td/><td colspan=\"2\">S: 1. Um now.</td></tr><tr><td/><td/><td colspan=\"2\">2. Now we're getting a little</td></tr><tr><td/><td/><td colspan=\"2\">3. more difficult.</td></tr><tr><td/><td colspan=\"2\">J: 4. (laughs)</td></tr><tr><td/><td colspan=\"3\">S: 5. Pick out the large air tube</td></tr><tr><td colspan=\"3\">[l picks up SAND]</td></tr><tr><td/><td/><td colspan=\"2\">6. that has the plunger in it.</td></tr><tr><td colspan=\"4\">[J puts down STAND. takes PLUNGER/MAINTUB~.</td></tr><tr><td/><td/><td>assembly]</td></tr><tr><td/><td colspan=\"2\">J: 7. Okay.</td></tr><tr><td/><td colspan=\"3\">S: 8. And set it on ~ts base,</td></tr><tr><td colspan=\"4\">[J puts down idAINTUBE, standing vertically, on the</td></tr><tr><td/><td/><td>TABLE]</td></tr><tr><td/><td/><td colspan=\"2\">9. which is blue now,</td></tr><tr><td/><td/><td>10. rzght?</td></tr><tr><td colspan=\"4\">[J has shifted focus to the TUBEBASE]</td></tr><tr><td/><td colspan=\"2\">J: 11. Yeah.</td></tr><tr><td/><td colspan=\"3\">$, 12. Base is blue.</td></tr><tr><td/><td/><td>13. Okay.</td></tr><tr><td/><td/><td>14. Now</td></tr><tr><td/><td/><td colspan=\"2\">15. You've got a bottom hole still</td></tr><tr><td/><td/><td>16. to be filled,</td></tr><tr><td/><td/><td>17. correct?</td></tr><tr><td/><td colspan=\"2\">J: 18. Yeah.</td></tr><tr><td colspan=\"4\">[J answers this with MAINTUBE still sittint on the</td></tr><tr><td/><td/><td colspan=\"2\">TABLE; he shows no indication of what</td></tr><tr><td/><td/><td colspan=\"2\">hole he thinks i8 meant -the one on</td></tr><tr><td/><td/><td colspan=\"2\">the MAINTUBE. OUTLET2, or the one in</td></tr><tr><td/><td/><td colspan=\"2\">the TUBEBASE]</td></tr><tr><td/><td>S.</td><td>19. Okay.</td></tr><tr><td/><td/><td colspan=\"2\">20. You have one red piece</td></tr><tr><td>[J</td><td>picks</td><td colspan=\"2\">21. remamm8? up ldA/NTUBE assembly and looks at</td></tr><tr><td/><td/><td colspan=\"2\">TUBEBASE, rotatine the MAINTUBE so</td></tr><tr><td/><td/><td colspan=\"2\">that TUBP-BASE is pointed up, and</td></tr><tr><td/><td/><td colspan=\"2\">sees the hole in it; he then looks at</td></tr><tr><td/><td/><td colspan=\"2\">the SLIDEVALVE]</td></tr><tr><td/><td colspan=\"2\">J: 22. Yeah.</td></tr><tr><td/><td colspan=\"2\">3. 23. Okay.</td></tr><tr><td/><td/><td colspan=\"2\">24. Take that red piece.</td></tr><tr><td colspan=\"3\">[j takes SIJDEVALVE]</td></tr><tr><td/><td/><td colspan=\"2\">25. It's got four little feet on</td></tr><tr><td/><td/><td>26. it?</td></tr></table>",
	"type_str": "table"
	},
	"TABREF15": {
	"text": "",
	"html": null,
	"num": null,
	"content": "<table><tr><td>Dlaerl:</td><td/><td/><td/><td/><td/><td/><td/></tr><tr><td colspan=\"6\">I DCVIC~ ITrensperency C~ARI</td><td/><td/></tr><tr><td/><td colspan=\"6\">*Composition PL.ASTIC)</td><td/></tr><tr><td/><td colspan=\"5\">iSuSpert I(~PL~NilIi</td><td/><td/></tr><tr><td/><td colspan=\"4\">ISUDpiTI 10111))</td><td/><td/><td/></tr><tr><td/><td colspan=\"2\">ISulleirt</td><td colspan=\"6\">{T~ADS Ilii-Politlon</td><td>[NDIt)</td></tr><tr><td/><td colspan=\"5\">IDlm4,nilonl I~nlth</td><td colspan=\"2\">S.OII</td></tr><tr><td/><td colspan=\"4\">ilAiloli\u00a9il-~hlll</td><td colspan=\"3\">T$111Jlilll</td></tr><tr><td colspan=\"9\">(FIT-INTO IOuter (D[VIC/ (Trlnsperencv</td><td>C'~TARI</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"4\">;Compos,t ,on PLASTIC)</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"4\">Shape R~O )</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"4\">( Ail@lo\| I \u00a2li -.~hllpe 'r~.~ljt~R )</td></tr><tr><td/><td/><td/><td/><td/><td colspan=\"3\">(Subpirl</td><td>~SA. ~\u00a2 q('olor T'~T~QtrOl.~T);Ib)</td></tr><tr><td/><td/><td colspan=\"2\">\u00a2 Inner</td><td>)</td><td/><td/><td/></tr><tr><td/><td/><td colspan=\"3\">ir,tCond\|t~on</td><td colspan=\"2\">LOOSEI )</td><td/></tr><tr><td colspan=\"2\">Figure 7:</td><td colspan=\"4\">The speaker's</td><td colspan=\"3\">descriptions</td></tr><tr><td colspan=\"5\">The first step</td><td colspan=\"4\">in the reference</td><td>process</td><td>ts the</td></tr><tr><td>actual</td><td colspan=\"2\">search</td><td colspan=\"6\">for a referent</td><td>in the knowiedge</td><td>base</td></tr><tr><td colspan=\"3\">The reference</td><td colspan=\"4\">identification</td><td/><td>process</td><td>is incremental in</td></tr><tr><td>nature,</td><td colspan=\"5\">l.e,, the listener</td><td colspan=\"3\">c~n begin the search</td><td>process</td></tr><tr><td>before</td><td colspan=\"3\">he hears</td><td colspan=\"5\">the complete</td><td>description</td><td>This was</td></tr><tr><td colspan=\"2\">observed</td><td colspan=\"3\">throughout</td><td colspan=\"4\">the videotape</td><td>excerpts</td><td>and the</td></tr><tr><td colspan=\"2\">algorithm</td><td colspan=\"3\">presented</td><td colspan=\"2\">here</td><td colspan=\"2\">is actually</td><td>deslgned</td><td>to be</td></tr><tr><td colspan=\"3\">incremental.</td><td colspan=\"2\">The</td><td colspan=\"3\">KL-One</td><td>Classifier</td><td>compares</td><td>the</td></tr><tr><td colspan=\"9\">features specified in the speaker's</td><td>descriptions</td><td>(Descrl</td></tr><tr><td colspan=\"6\">and the\" \"Outer\" feature</td><td colspan=\"3\">of Descr2 in Figure T) with the</td></tr><tr><td>features</td><td/><td colspan=\"2\">speclhed</td><td colspan=\"4\">for each</td><td>element</td><td>in the</td><td>EL-One</td></tr><tr><td colspan=\"2\">taxonomy</td><td colspan=\"2\">that</td><td colspan=\"4\">corresponds</td><td>to one</td><td>of the</td><td>current</td></tr><tr><td colspan=\"4\">objects of interest</td><td colspan=\"5\">in the real world.</td><td>Notice that some</td></tr><tr><td colspan=\"2\">features</td><td>are</td><td colspan=\"3\">directly</td><td colspan=\"3\">comparable.</td><td>For</td><td>example,</td><td>the</td></tr><tr><td colspan=\"3\">\"Transparency\"</td><td/><td colspan=\"3\">feature</td><td/><td>of</td><td>Descrl</td><td>and</td><td>the</td></tr><tr><td colspan=\"3\">\"Transparency\"</td><td colspan=\"3\">feature</td><td colspan=\"3\">of MAINTUBE are both equal to</td></tr><tr><td colspan=\"2\">\"CLEAR.\"</td><td colspan=\"2\">Other</td><td colspan=\"3\">features</td><td/><td>require</td><td>further</td><td>processing</td></tr><tr><td colspan=\"9\">before they can be compared.</td><td>The OPENING value of</td></tr><tr><td colspan=\"2\">\"Subpart\"</td><td colspan=\"3\">in Deecrl</td><td colspan=\"4\">is thought</td><td>of primarily</td><td>as a 2-D</td></tr><tr><td colspan=\"3\">cross-sectlon</td><td colspan=\"2\">(such</td><td colspan=\"4\">as a \"hole\"). while two CYLINDER</td></tr><tr><td colspan=\"2\">subparts</td><td colspan=\"7\">of MA/NTUBE are viewed as 13-D) cylinders</td></tr><tr><td colspan=\"7\">that have the \"Function\"</td><td colspan=\"2\">of being outlets, i.e., OUTLET-</td></tr><tr><td colspan=\"5\">ATTACHMENT-POINTS.</td><td/><td colspan=\"2\">To</td><td>compare</td><td>OPENING</td><td>and</td></tr><tr><td colspan=\"6\">CYLINDER, the inference</td><td colspan=\"3\">must be made that both things</td></tr><tr><td colspan=\"3\">can describe</td><td colspan=\"6\">the same thing</td><td>(similar</td><td>inferences</td><td>are</td></tr><tr><td colspan=\"2\">developed</td><td colspan=\"3\">in (18]).</td><td colspan=\"4\">One way this inference</td><td>can occur</td></tr><tr><td colspan=\"4\">ts by recurslvely</td><td colspan=\"4\">examining</td><td>the subparts</td><td>of MA/NTUBE</td></tr><tr><td>wlth</td><td>the</td><td colspan=\"2\">partial</td><td/><td colspan=\"3\">matcher</td><td>until</td><td>the</td><td>cylinders</td><td>are</td></tr><tr><td colspan=\"8\">examined at the 2-D level.</td><td>At that level, an end of the</td></tr><tr><td colspan=\"2\">cylinder</td><td colspan=\"5\">will be defined</td><td colspan=\"2\">as an OPENING</td><td>With that</td></tr><tr><td colspan=\"3\">examination,</td><td colspan=\"6\">the MAINTUBE can be seen as described</td></tr><tr><td colspan=\"3\">by Deeer I.</td><td/><td/><td/><td/><td/></tr><tr><td colspan=\"5\">Descr2 presents</td><td colspan=\"4\">different problems.</td><td>Descr2 refers</td></tr><tr><td colspan=\"8\">to an obJect that is supposed</td><td>to have a subpart</td><td>that is</td></tr><tr><td colspan=\"9\">TURQUOISE. The Classifier determines</td><td>that Descr2 could</td></tr><tr><td colspan=\"9\">not describe either the CAP or 3TAND because</td><td>both are</td></tr><tr><td colspan=\"9\">BLUE. It also could not describe</td><td>the MAINTUBE 6 or .aIR</td></tr><tr><td colspan=\"4\">CHAMBER since</td><td colspan=\"2\">each</td><td colspan=\"2\">has</td><td>subparts</td><td>that</td><td>are</td><td>either</td></tr><tr><td colspan=\"9\">VIOLET or BLUE. The Classifier places Descr2 as best it</td></tr><tr><td colspan=\"6\">can in the taxonomy,</td><td colspan=\"3\">showing no connections</td><td>between</td></tr><tr><td colspan=\"9\">6Sin\u00a9e Deacr~ refers to ~AJNTUSE. MAINTUBE could be</td></tr><tr><td colspan=\"5\">dropped as \u2022 potential</td><td colspan=\"4\">referent</td><td>candidate for Descr2.</td><td>We</td></tr><tr><td>will.</td><td colspan=\"2\">hommver,</td><td/><td/><td/><td/><td/></tr></table>",
	"type_str": "table"
	}
	}
	}
	}