17
This PDF includes a chapter from the following book: Linguistics for the Age of AI © 2021 Marjorie McShane and Sergei Nirenburg License Terms: Made available under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International Public License https://creativecommons.org/licenses/by-nc-nd/4.0/ OA Funding Provided By: The open access edition of this book was made possible by generous funding from Arcadia—a charitable fund of Lisbet Rausing and Peter Baldwin. The title-level DOI for this work is: doi:10.7551/mitpress/13618.001.0001 Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Linguistics for the Age of AI

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Linguistics for the Age of AI

This PDF includes a chapter from the following book:

Linguistics for the Age of AI

© 2021 Marjorie McShane and Sergei Nirenburg

License Terms:

Made available under a Creative CommonsAttribution-NonCommercial-NoDerivatives 4.0 International Public Licensehttps://creativecommons.org/licenses/by-nc-nd/4.0/

OA Funding Provided By:

The open access edition of this book was made possible by generous fundingfrom Arcadia—a charitable fund of Lisbet Rausing and Peter Baldwin.

The title-level DOI for this work is:

doi:10.7551/mitpress/13618.001.0001

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 2: Linguistics for the Age of AI

Page numbers in italic indicate a figure and page numbers in bold indicate a table on the corresponding page.

Index

Abduction, 22Abductive reasoning, 21–22Abelson, R. P., 6ACE. See Automatic Content Extraction (ACE) corpusAchalasia disease modellearning components of, 327–328ontological knowledge for, 323patient-authoring interface for, 320–321, 322

Acquisitionof lexicon, 103–104, 382of ontology, 98

Actionabilitydefinition of, 4, 62, 90judgement of, 13–15principle of least effort and, 14–15

ACT-R, 37Adjectives. See also Modificationnew-word learning of, 136unknown, 195–196

Advances in Cognitive Systems community, 396n1Adverbs. See also Modificationin constructions, 166semantic analysis of, 141

AECs. See Anaphoric event coreferences (AECs)Agent applications. See also Maryland Virtual

Patient (MVP) systembias-detection advisor, 331–343LEIA-robots, 343–347, 346

Agent architecture, 9–13Agent First principle, 23Agrammatic aphasia, 24AGREE-TO-AN-INTERVENTION-OR-NOT evaluation

function (MVP), 316–317Allen, J., 139, 353Al-Sayyid Bedouin Sign Language, 24Altmann, G. T. M., 17Ambiguitybenign, 15, 95, 209, 210, 220

lexical, 2–3, 290, 362–365morphological, 2pragmatic, 3of proper names, 208referential, 3residual, 247–254, 261–262, 290–291scope, 3semantic dependency, 3syntactic, 3

Analogy, reasoning by, 61, 252–253Analytic approach, 9Anaphoric event coreferences (AECs), 234–240,

393n45adjuncts in sponsor clause, 238–239coreference between events, 237–238coreference between objects in verb phrases, 238–239

ellipses-resolved meaning representation for, 234–235, 235

modals and other scopers, 239–240verbal/EVENT head of sponsor, 235–237

Anchors, memorychallenges of, 209–210definition of, 203example of, 204–205during Situational Reasoning, 297

Annotation. See Corpus annotationAntecedents. See SponsorsAnticipation, 12, 394n11Antonymy, 43Aphasia, 24Application-specific lexicons, 104–105Apresjan, Ju. D., 386n15Arguments, coreferencing of, 72Argument-taking wordsconstraints on, 211as constructions, 166–167lexical acquisition of, 74

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 3: Linguistics for the Age of AI

416 Index

Artificial neurons, 5Aspect, semantic analysis of, 162, 165Aspectual verbselided and underspecified events involving, 186–189, 192, 241–242

scopers and, 240Attempto, 97Automatic Content Extraction (ACE) corpus, 52Automatic disambiguation, 25, 74–76, 103, 370Automatic programming, 388n10Automatic syn-mapping process, 131–134

Babkin, P., 215, 365, 366Bad events/states. See Negative sentiment termsBaker, M., 49Bakhshandeh, O., 394n18Bar Hillel, Y., 5–6, 56Barrett’s metaplasia, 307Base-rate neglect, 337Basic Coreference Resolutionacross syntactic categories, 206–207anaphoric event coreferences, 234–240, 235, 393n45with aspectuals + OBJECTS, 241–242chain of coreference, 201, 208challenges of, 205–210, 391n4coreferential events expressed by verbs, 242–244decision-making after, 86definite descriptions, 204, 227–233ellipsis, 210–211example of, 203–212further exploration of, 244–245general principles of, 82implicatures, 206in knowledge-lean paradigm, 44–46personal pronouns, 206, 212–217pronominal broad referring expressions, 217–227, 392n28, 393n32

semantic relationships between referentially related entities, 207

terminology related to, 202–203words excluded from, 203

Basic Semantic Analysisalgorithms for, 141–142constructions, 165–175decision-making after, 85–86definite description processing, 227–228ellipsis, 186–193, 188fragmentary utterances, 193further exploration of, 199–200general principles of, 82, 141–143indirect speech acts, 175–176issues remaining after, 198–199metaphors, 180–185, 199metonymies, 185–186modification, 143–159nominal compounds, 176–180, 177, 200nonselection of optional direct objects, 193

proposition-level semantic enhancements, 160–165unknown words, 193–198

Batiukova, O., 102Baxter robot, 344BDI (belief-desire-intention) approach, 9, 11Beale, S., 237, 358, 362, 366Belief modality, 161Benign ambiguity, 15, 95, 209, 210, 220Berners-Lee, T., 27Besold, T. R., 40Bias-detection advisorclinician bias detection, 335–339memory support for bias avoidance, 333–334, 334patient bias detection, 339–343, 341, 342vision of, 331–333, 332

Bickerton, D., 23Big data, 4–5, 57, 63, 302, 324, 347Binding sets, 126–128, 127BLEU, 396n3Boas system, 328Boeing’s Computer Processable Language, 96Bos, J., 392n28Bounded rationality, theory of, 331Bowdle, B., 180Brants, T., 53Brick, T., 41Bridging references, 207, 229–231Broad referring expressions, resolution ofin machine translation, 392n28negative sentiment terms, 221–223, 226simple example of, 217–218in syntactically simple contexts, 219–221, 226, 393n32

using constructions, 218–219, 226using meaning of predicate nominals, 223–224, 227

using selectional constraints, 224–227Brooks, R., 38Brown, R. D., 20Byron, D., 392n28

Candidate sponsors, 202, 219, 226Carbonell, J. G., 21Cardinality of sets, 152–158Carlson, L., 20Carnap, R., 5Cartwright, N., 39, 89Case role labeling, 28Case studies, knowledge extraction from, 326–327CELT. See Controlled English to Logic Translation

(CELT)Centering Theory, 21Cerisara, C., 46Chain of coreference, 201, 208Changeconditions of, 190–191, 241–242events of, 270, 284

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 4: Linguistics for the Age of AI

Index 417

CHANGE-EVENTs, 198, 274–276missing values in, 279

Chinchor, N., 244Chomsky, N., 16Church, K., 7–8Cimiano, P., 100–101Cinková, S., 394n20Clark, H. H., 49Clinical bridges, 302Clinician biasesdetection of, 335–339memory support for avoidance of, 333–334, 334

Clinician training. See Maryland Virtual Patient (MVP) system

COCA corpus, exploration of, 140, 199–200, 245, 284Cognition modeling, 311–317decision-making evaluation functions, 314–317goals, 313–314learning through language interaction, 312–313, 312

Cognitive architecturecurrent views on, 36–38OntoAgent, 10–13, 11, 30, 115, 286–287, 287psycholinguistic evidence for, 17

Cognitive bias. See also Bias-detection advisorclinician biases, 335–339definition of, 331memory support for avoidance of, 333–334, 334patient biases, 339–343, 341, 342sources of, 331–332

Cognitive linguistics, 22–23, 386n19Cognitive load, 31, 90, 124, 223, 227, 256, 280Cognitive Science Laboratory, Princeton University, 42Cohen, K. B., 394n17Cohen, P., 100Cohen, P. R., 37Collocations, 108, 166, 391n18Combinatorial complexitycauses of, 134–135, 135sense bunching and, 135–138

Commandsindirect speech acts, 175–176semantic analysis of, 164–165, 164

Common Logic Controlled English, 97Communicative acts. See Dialog actsCommunity-level task formulation, 352Comparative filter, 371Comparatives, 270–279, 394n18classes of, 271–277, 272machine learning approach to, 394n18overview of, 270–271reasoning applied in, 277–279ungrounded and underspecified comparisons, 270–279

Component-level evaluation experimentsdifficult referring expressions, 365–366lexical disambiguation and establishment of semantic dependency structure, 362–365

multiword expressions, 358–362nominal compounding, 355–358verb phrase (VP) ellipsis, 366–369

Compositionality, 29, 386n27Compounds, nominal. See Nominal compounds

(NNs)Computational formal semantics, 18–20, 353Computational linguistics, 15, 89, 384Computer processable controlled languages, 96–97Computer Processable English, 96Concept grounding, 345Conceptsconcept instances versus, 98instances of, 98, 251mapping of, 105–106modifiers explained using combinations of, 149–150, 159

naming conventions for, 393n43prototypical relationships between, 265–266scripts and, 99types of, 70–71words versus, 98

Conditional filter, 371Confidence levels, 60, 65, 67Confirmation bias, 342Conflicts, property value, 228Constituency parses, 118, 119Constraintsautomatic disambiguation supported by, 74–76connected versus unconnected, 265–266

Construction grammar, 16–17, 165Constructionsconstituents in, 166–167definition of, 91, 165–166exclusion criteria for, 174lexical versus non-lexical, 167–168, 391n17null-semmed components of, 169–172object fronting, 167overlapping of, 174polysemy in, 174resolving personal pronouns with, 213–215, 217

resolving pronominal broad RefExes with, 218–219, 226

utterance-level, 172–173Context, 3–4Controlled English to Logic Translation (CELT),

96–97Controlled languages, 96–97Conventional metaphors, 180–185. See also

MetaphorsConversation acts. See Dialog actsCopular metaphors, 184–185Coreferenceof arguments, 72chain of, 208window of, 202

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 5: Linguistics for the Age of AI

418 Index

Coreference resolution. See Basic Coreference Resolution; Textual coreference

CoreNLP Natural Language Processing Toolkit, 117–118, 139, 356, 359, 388n3

coreference resolver, 213, 217, 227, 244, 294, 392n18

Cornell Natural Language for Visual Reasoning corpora, 158, 396n12

Corpus annotationof discourse structure, 21in evaluations, 351, 353–354, 366, 382event coreference links in, 244further exploration of, 139, 244“golden” text meaning representations, 93–94machine learning and, 15, 21, 44–46, 394n20manual, 19, 27, 52–55in the Prague Dependency Treebank, 55resources devoted to, 28semantic annotation in, 29times of events in, 162

Creative idiom use. See Idiomatic creativityCrosslinguistic differences, 100, 392n22Cryptography, 5Cybernetics, 5Cycorp, 37Cyc, 25–26, 98, 296

DARPA, Explainable AI program, 38–39Dcoref, 392n18Decision-makingafter basic semantic analysis, 82, 85–86after extended semantic analysis, 82, 87after pre-semantic analysis, 81, 84–85after pre-semantic integration, 81–82, 85after situational reasoning, 82, 87–88decision points in, 83–84, 83, 84in Maryland Virtual Patient (MVP) system, 313–317

Decision points, 83–84, 83, 84Default facet, 71, 146, 241Definite descriptionsawaiting Situational Reasoning, 233during Basic Coreference Resolution, 228–233during Basic Semantic Analysis, 227–228during Pre-Semantic Analysis, 227sponsors for, 229, 231–233

Delimit-scale routine, 145–147Demand-side approach, 302Dependency parses, 118, 119Dependency-syntax theory, 54–55Depletion effects, 333, 340Descriptive-pragmatic analyses, 20–22Descriptive rationality, 40DeVault, D., 41Dialog actsambiguity in, 290–291categories of, 47–48definition of, 47

detection of, 46–48, 387nn40–42indirect, 39, 81, 175–176, 199, 297–298residual ambiguity, 287, 290–291

Dialogue Act Markup in Several Layers (DAMSL) tag set, 47–48

Dictionariesmachine-readable, 42–43, 98, 103

Diogenes the Cynic, 39Direct objectsconstraints on, 75coreference between, 238–239ellipsis of, 392nonselection of optional, 193of sequential coordinate clauses, 215

Discourse. See PragmaticsDiscourse-structure annotation, 21Discourse-theoretic approaches. See PragmaticsDisease modelsdevelopment of, 304–305GERD example, 306–311, 307, 310, 311visualization of, 320–324

Disfluencies, stripping of, 124Distributional semantics, 28–29, 290, 386n27, 394n3Doctor-patient dialog, bias detection in, 342–343, 342Double agents, 303Dynamically computed valuesfor relative text components, 150–152of scalar attributes, 145–148

Dynamic programming, principle of, 61Dynamic sense bunching, 135–138Dynamic Syntax, 17

Effort modality, 161Eliminativism, 385n7Ellipsisaspectuals + OBJECTS, 189–193, 241–242definition of, 210–211gapping, 91, 192, 193, 211head noun, 192, 193lexically idiosyncratic, 189–190, 193in natural language processing, 211, 392n17verb phrase, 186–189, 188, 192, 366–369, 393n44

Empirical natural language processing (NLP), 27–29, 50–52, 290, 386n27, 394n3

Empiricism, 7–8End-system evaluations, 350English, J., 345English Gigaword corpus, 214, 365Enumerative lexicons, 102–103Episodic memory, 60, 68, 77, 287Epistemic modality, 113–114, 161, 163, 380Epiteuctic modality, 161Evaluationchallenges of, 349–350conclusions from, 381–382of difficult referring expressions, 365–366end-system, 349–350

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 6: Linguistics for the Age of AI

Index 419

holistic, 369–382of lexical disambiguation, 362–365of multiword expressions, 358–362of nominal compounding, 355–358task-oriented, 351–354of verb phrase ellipsis, 366–369

Evaluation functions, 314–317Evaluative attitudes, effect of, 340Evaluative modality, 161, 271, 377, 380Event coreference. See Textual coreferenceEvent ellipsislexically idiosyncratic, 189–190, 193verb phrase, 186–189, 188, 192, 366–369, 393n44

Event identity, 244Eventsanaphoric event coreferences, 234–240, 235, 245, 393n45

case roles for, 70of change, 270, 279, 284coreferential, 237–238, 242–244definition of, 70properties of, 70reference resolution of, 204

Event scripts, bridging via, 230–231Evolution of language, 23–25Exaggerations, 298–299Experiments, evaluationdifficult referring expressions, 365–366lexical disambiguation, 362–365multiword expressions, 358–362nominal compounding, 355–358verb phrase ellipsis, 366–369

Explainable AI program (DARPA), 38–39Explanatory AI (artificial intelligence), 13–15,

38–40, 301, 385n7Expletives, 170–171Exposure effect, 338–340Extended Semantic Analysisdecision-making after, 87fragments, 279–283, 394n20general principles of, 82incongruities, 254–264indirect modification, 262–264residual ambiguities, 247–254underspecification, 264–279

Extralinguistic information, 1

Facetsdefault, 146definition of, 71relaxable-to, 146sem, 146–147

False intuitions, 335–336Fast-lane knowledge elicitation strategy, 329, 329Feature matching, 214–215Fernández, R., 394n20Field-wide competitions, 44, 387n44

Fillmore, C., 385n11Filters, sentence extraction. See Sentence extraction

filtersFind-anchor-time routine, 388n2Finlayson, M. A., 19Fishing algorithm, 287–290Fixed expressions. See Multiword expressions (MWEs)Fleshing out algorithm, 287–290Focus, 20, 31Fodor, J., 104Folk psychology, 14, 39–40Forbus, K. D., 388n16Formal semantics, 17–20Formulaic language, 391n18. See also Multiword

expressions (MWEs)Fractured syntax, 287–290Fragments, 193, 211, 279–283FrameNet, 25–26, 28, 115, 385n11, 388n8Frame semantics, 26Framing sway, 342Functionalism, 385n7Fuzzy matching, 121–122

Gapping, 91, 192, 193, 211Garden-path sentences, 143, 390n2Gastroesophageal reflux disease (GERD)model, 306–311, 307, 310, 311

Generalized phrase structure grammar, 6, 385n4General-purpose lexicons, 104–105Generative grammar, 6, 16, 23, 139Generative lexicons, 102–103Genesis system, 19–20Gentner, D., 180GERD. See Gastroesophageal reflux diseaseGigaword corpus, 214, 365GLAIR, 37Goals, 61, 313–314“Golden” text meaning representations, 93–94Gonzalo, J., 43Google Translate, 56Graesser, G., 180Grounding, 48–49, 292, 387n41, 391n1

Hahn, U., 389n20Hajič, J. 55Halo effect, 341–342, 341Halo-property nests, 341–342, 341Handcrafted knowledge basesCyc, 25–26FrameNet, 25–26Semantic Web, 27VerbNet, 25–26

Harris, D. W., 82Hayes, P. J., 69Head-driven phrase structure grammar, 6, 385n4Head matching, 206, 391n8Head noun ellipsis, 192, 193

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 7: Linguistics for the Age of AI

420 Index

Hearst, M., 177Heuer, R. J., Jr., 332Heuristic evidence, incorporation ofempirical natural language processing (NLP), 27–29handcrafted knowledge bases, 25–27

Hidden meaningshyperbole, 298–299indirect speech acts, 175–176, 199, 297–298sarcasm, 298

Hierarchical task network (HTN) formalism, 344Hirschman, L., 244Hirst, G., 386n15Hobbs, J., 21–22, 349Holistic evaluationsconclusions from, 381experiment with filters, 374–381experiment without filters, 372–375limitations of, 369–371sentence extraction filters for, 371–372

Homographous prepositions, 129Horizontal incrementality, 77–82, 80Hovy, E., 7, 177, 244Hutchins, J., 56Hybrid evaluations, 354Hyperbole, 298–299Hypernyms, 42, 231–233Hyponyms, 99, 103, 104, 231–233

Ide, N., 103, 388n17Idiomatic creativity, 264detection of, 258–260, 258further exploration of, 284, 394n12semantic analysis of, 261–262sources of, 257–258, 394n14

Idioms, 284as constructions, 166creative use of, 257–262, 258, 264, 284, 394n12, 394n14

multiword expressions, 167, 174null-semmed constituents of, 169–172

Illocutionary acts. See Dialog actsIllusion of validity, 336–337Imperatives, 28, 164–165, 164Imperfect syn-maps, optimization of, 122, 126–128,

127Implicatures, 41, 80, 206, 376, 381Implied events, 173Imprecision, semantic, 60Incongruitiesdefinition of, 254idiomatic creativity, 257–262, 258, 264, 284, 394n12, 394n14

metonymies, 185–186, 254–255, 264, 394n10preposition swapping, 256–257, 264, 284

Incrementalitychallenges of, 40–41computational model of, 41–42

horizontal, 79, 80during natural language understanding, 77–82in psycholinguistics, 17vertical, 79, 80, 84

Incremental parser, 385n12Indirect modification, 158–159, 262–264Indirect objects, coreference between, 238–239Indirect speech acts, 39, 81, 175–176, 199, 297–298Inference, 22, 285Information theory, 5Inkpen, D., 386n15Instance coreference, 251between events, 237–238between objects in verb phrases, 238–239

Instances, concepts versus, 98Instance-type relationships, 207Interannotator agreement, 47, 53, 55, 351, 353, 366,

391n21Interlingual Annotation of Multilingual Text

Corpora project, 53Interrogatives. See QuestionsIntrasentential punctuation mark filter, 371Intuition, false, 335–336IS-A property, 70, 71, 99, 252Iteration values, of aspect, 162

Jackendoff, R., 23–24, 286Jarrell, B., 320Jelinek, F., 50Jeong, M., 47Johnson, K., 180, 393n44Johnson, M., 180–181, 184Jones, R. M., 396n2Journal of Pragmatics, The, 22Jumping to conclusions, 335Jurafsky, D., 56

Kahneman, D., 335, 338, 341Kamide, Y., 17KANT/KANTOO MT project, 96KDiff3, 366King, G. W., 7Klein, G., 335Knowledge-based approaches, 3–8, 33–34Knowledge-based evaluations, 353–354Knowledge bases, 61, 68–77, 115. See also Lexicons;

Ontologyautomatic extraction of, 42corpus annotations versus, 54episodic memory, 68, 77expansion of, 62handcrafted, 25–27in Situational Reasoning, 293–294specialized concepts in, 69

Knowledge bottleneck, 7, 33–34, 42, 384Knowledge elicitation methods (MVP), 328–331,

329, 330

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 8: Linguistics for the Age of AI

Index 421

Knowledge-lean paradigmcoreference resolution in, 21, 28–29, 44–46, 53–54definition of, 3–8

Knowledge Representation and Reasoning (KR&R) communities, 386n16

Knowledge representation language (KRL), 10, 94–97

Köhn, A., 40KR&R. See Knowledge Representation and

Reasoning (KR&R) communitiesKrál, P., 46KRL. See Knowledge representation language

(KRL)Kruijff, G. J. M., 41

Laërtius, Diogenes, 39Laird, J. E., 38Lakoff, G., 180, 182, 184Langley, P., 36Language-based learning, 33Language-centric reasoning, 61Language complexity, microtheory of, 371–372,

396n11Language-endowed intelligent agents (LEIAs).

See also Agent applications; Natural language understanding (NLU)

architecture of, 9–13future directions in, 383–384phenomenological stance for, 31–32, 386n28

Language evolution, 23–25Language Files (Mihalicek and Wilson), 139Language Hoax, The (McWhorter), 115Language independence, agent, 61Language Understanding Service, 287Largely language independent lexicons, 105–107,

389n24Lascarides, A., 394n20Leafgren, J., 386n17Learning, language-based. See also Lifelong

learningby being told, 336by LEIA-robots, 345–347, 346in the Maryland Virtual Patient (MVP) system, 324–328

prerequisite for, 33by reading, 299–300

Least effort, principle of, 14–15Lee, G. G., 47Lee, H., 213Legal sequences of actions, learning of, 345, 346LEIA-robots, 343–347, 346LEIAs. See Language-endowed intelligent agents

(LEIAs)Lemma annotator, 118Lenat, D., 26Levesque, H. J., 37Levi, J. N., 177

Levin, B., 26Lexical ambiguity, 2–3, 290, 362–365Lexical constructions, 167–168, 391n17Lexical disambiguation experiment, 362–365Lexical-functional grammar, 6, 385n4Lexical idiosyncrasy, 356–357Lexical lacunae, 363, 372, 382Lexically idiosyncratic event ellipses, 189–190,

193Lexical semantics, 17–18, 42–43Lexiconsacquisition of, 103–104, 382addition of construction senses to, 131–134application-specific, 104–105automatic disambiguation in, 74–76definition of, 68enumerative, 102–103features of, 73–76general-purpose, 104–105generative, 102–103incompleteness of, 121–122, 369–370issues of, 102–111as key to successful natural language understanding, 60, 142

largely language independent, 105–107, 389n24reuse across languages, 109–111

Lexico-syntactic constructionsresolving personal pronouns with, 213–215, 217resolving pronominal broad RefExes with, 218–219

Lifelong learningdefinition of, 2need for, 62, 101, 384new-word learning, 124–126, 299–300research implications of, 62

Light verb filter, 372Light verbs, 107–109Lin, J., 349Lindes, P., 38Linear grammar, 23Linguistic meaning, 286Linguistic scholarship, insight from, 15–25cognitive linguistics, 22–23language evolution, 23–25pragmatics, 20–22psycholinguistic, 17semantics, 18–20theoretical syntax, 16–17

Listing, 61Literal attributes, 70, 144–145Liu, B., 221Local properties, 99–100Locutionary acts. See Dialog actsLombrozo, T., 22Loops, 72Lower esophageal sphincter (LES), 306Lu, J., 45, 243–244

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 9: Linguistics for the Age of AI

422 Index

Machine-readable dictionaries, 42–43, 98, 103Machine translationbroad referring expressions in, 392n28historical overview of, 4–7, 56, 328–329paraphrase in, 353, 396n3

Mandatory syntactic constituents, 211, 392n16Manning, C. D., 22, 95Manual lexical acquisition, 382Mapping, syntactic. See Syntactic mapping

(syn-mapping)Marcus, G., 53Markables, selection of, 44, 351–352Martin, J. H., 56Maryland Virtual Patient (MVP) systemcognitive modeling for, 311–317disease model for GERD, 306–311, 307, 310, 311example system run, 317–320knowledge elicitation strategies for, 328–331, 329, 330

learning components of, 324–328omniscient agent in, 32ontological knowledge for, 321–324, 323ontological scripts for, 72–73, 305paraphrase and ontological metalanguage in, 113–115

patient-authoring interface for, 320–321, 322physiology modeling for, 304–305requirements for, 302–303traces of system functioning in, 324, 324vision and architecture of, 301–304, 303visualization of disease models in, 320–324

Matrix verbs, 163, 165McCarthy, J., 30McCrae, J., 100–101McCulloch, W. S., 5McShane, M., 105, 215, 237, 320, 358, 362, 365, 366,

369, 389n24, 392n22McWhorter, J. 100, 115Meaning, definition of, 50Meaning representations (MRs), 296. See also Text

meaning representations (TMRs)Measurement of progress. See EvaluationMechanical Translation and Computational

Linguistics journal, 4Mechanical Translation journal, 4Memoryanchors, 203–205, 209–210, 297episodic, 60, 68, 77, 287in LEIA-robots, 345, 347of referring expressions, 209–210, 297semantic, 287, 287

Mental actions, 10, 62, 346Merging, ontological, 100Meronymy, 43, 111bridging via, 230–231ontological paraphrase and, 113–115

MeSH, 389n20

Metacognition, 39–40Metalanguage, ontological, 60, 111–115Metaphors, 393n2conventional, 180–185copular, 184–185further exploration of, 199importance of, 180inventories and classifications of, 199in nominal compounding experiment, 357–358novel, 180–181past work on, 180–182, 391n21

Metonymic Mapping Repository, 255Metonymies, 185–186, 254–255, 264, 394n10Microtheoriescombinatorial complexity, 134–138, 135concept of, 2, 13, 16, 61, 88–93development of, 16incomplete coverage of, 370–371

Mihalicek, V., 139Mikulová, M., 55Miller, G., 25, 26, 42Mindreading, 14, 39–40, 69, 285, 298Minimal Recursion Semantics, 394n20Minsky, M., 30Mitkov, R., 44Modalityproperties of, 160–161semantic analysis of, 160–161, 161, 165types of, 161

Modals, coreference resolution and, 239–240ModelingBDI (belief-desire-intention) approach, 9explanatory power of, 13–15levels of, 13nature of, 62ontological, 10phenomenological stance for, 31–32, 386n28properties of, 388n16role of, 89–91simpler-first, 91situation, 10tenets of, 1

Modificationcombinations of concepts in, 149–150, 159dynamically computed values for relative text components, 150–152, 159

dynamically computed values for scalar attributes, 145–148, 159

indirect, 158–159, 262–264nonrestrictive, 393n40of null-semmed constituents, 170–172quantification and sets, 152–158, 159recorded property values, 143–145, 159restrictive, 229

Moldovan, D., 52Monti, J., 390n15, 391n18Morphological ambiguity, 2

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 10: Linguistics for the Age of AI

Index 423

MUC-7 Coreference Task, 21, 44, 53–54, 244MUC-7 coreference corpus, 53Multichannel grounding, 292Multiple negation filter, 371Multistep translation method, 97Multiword expressions (MWEs), 391n18addition of new construction senses for, 131–134, 382

alternative semantic representations of, 106–107challenges of, 174as constructions, 167–168evaluation experiment for, 358–362in lexicon, 74, 128light verbs in, 108–109optimization of imperfect syn-maps for, 28parser inconsistencies with, 121prevalence of, 390n15sense bunching and, 137

MVP. See Maryland Virtual Patient (MVP) system

Named entity recognition annotator, 118Narrow-domain systems, 41, 47, 62, 253Natural language (NL)controlled, 96–97knowledge representation language and, 94–98paraphrase in, 111–115, 389nn27–28

Natural language processing (NLP), 15–25, 41–42agent architecture of, 9–13, 11, 36–38ambiguity in, 2–4case roles for, 388n8context in, 3–4coreference resolution in, 21, 28–29, 44–46, 53–54corpus annotation, 15, 21, 28–29, 45, 52–55, 385n2, 386n18

definition of, 4ellipses in, 211, 392n17empirical, 27–29, 50–52field-wide competitions, 44, 387n44goals of, 7heuristic evidence incorporated into, 25–29historical overview of, 4–8incrementality in, 40–42knowledge-based approaches to, 3–8, 33–34knowledge bottleneck in, 7, 33–34, 42, 384knowledge-lean paradigm, 3–8, 21, 28–29, 44–46, 53–54

knowledge representation and reasoning, 386n16modeling in, 13–15natural language understanding compared to, 4, 12, 33–36

paths of development in, 7–8purview of, 8timeframe for projects in, 8

Natural language understanding (NLU). See also Agent applications; Microtheories; Text meaning representations (TMRs)

cognitive architecture of, 9–13, 11, 36–38

decision-making in, 63, 84–88decision points in, 83–84deep, 100–101example-based introduction to, 64–68future directions in, 383–384incrementality in, 77–82, 80, 81interaction with overall agent cognition, 61–62knowledge and reasoning needed for, 60–61methodological principles for, 60–62modules, 13natural language processing compared to, 4, 12, 33–36

nature of, 60ontological metalanguage in, 60, 111–115stages of, 79–84, 81 (see also individual stages)strategic preferences for, 62–63

Navarretta, C., 21Near synonyms, 103“Need more features” bias, 335Negative sentiment terms, 221–223, 226Neo-Whorfianism, 100, 115Neural networks, 15Newell, A., 388n14Newmeyer, F. J., 396n11Ng, V., 45, 243–244Nicaraguan Sign Language, 24Nirenburg, S., 95, 102, 345, 358, 362NLP. See Natural language processing (NLP)NLU. See Natural language understanding (NLU)No-main-proposition filter, 371–372Nominal compounds (NNs)basic semantic analysis of, 88, 128, 176–180, 177

challenges of, 178–179evaluation experiment for, 355–358lexicon-oriented approaches to, 179–180relation-selection approach to, 177–179relations in, 51–52search engine constraints and, 200underspecified analysis of, 180, 264–269, 279

Noncanonical syntax, 122, 124, 211Non-compositionality, 358, 387n2Non-lexical constructions, 167–168Nonliteral language, 20, 27, 145, 390n5Nonrestrictive modifiers, 393n40Nonselection of optional direct objects, 193Non-sentential utterances, 394n20Normative rationality, 40Nouns and noun phrasesdefinite descriptions, 203, 227–233dynamic sense bunching of, 136head noun ellipsis, 192, 193new-word learning of, 125–126, 194–195proper names, 68, 208, 229, 361coreference resolution of, 206

Novel metaphors, 180–181. See also MetaphorsNP-Defs. See Definite descriptions

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 11: Linguistics for the Age of AI

424 Index

Null-semmingdefinition of, 169modification of null-stemmed constituents, 170–172

pleonastics and light verbs, 107–109purpose of, 133typical uses of, 169–170

Object fronting, 167Objectswith aspectual verbs, 241–242definition of, 70direct, 75, 193, 215, 238–239, 392dynamic sense bunching of, 136indirect, 238–239ontological definition of, 71ontology-search strategies for, 248–252properties of, 70in verb phrases, coreference between, 238–239

Obligative modality, 161, 378, 380Observable causes, 39Occam’s razor, 182Ogden, C. K., 96O’Hara, T., 388n8Olsson, F., 45Onomasticon, 68, 77OntoAgent cognitive architecture, 30components of, 286–287, 287high-level sketch of, 10–13, 11knowledge bases for, 115

OntoElicit, 328–331, 329, 330Ontological Construction Repository, 265–267Ontological instances, 251Ontologically decomposable properties, objects

linked by, 249–251Ontological metalanguage, 60, 111–115Ontological paraphrase, 111–115, 389nn27–28Ontological paths, objects linked by, 252, 268–269Ontological Semantics, 30, 77, 89, 103, 152, 160–161,

161Ontology, 69–73availability of, 100benefits of, 71concepts in, 69–70content and acquisition of, 69–73crosslinguistic differences in, 100definition of, 68, 69example of, 66–67external resources for, 100facets in, 71issues of, 97–101for Maryland Virtual Patient (MVP) system, 321–324, 323

merging, 100objects and events in, 71object-to-object relations in, 248–254role of, 60

scripts in, 71–73, 305, 388n10, 394n5upper/lower divisions in, 100

Ontology merging, 100Onyshkevych, B., 102Open-domain vocabulary, 74Optional direct objects, nonselection of, 193Optionally transitive verbs, 193Ordered bag of concepts methodology, 211, 288–289Ordering, in scripts, 72Overlapping constructions, 174

Packed representation, 41Paraphrasechallenges of, 353detection of, 386n25machine translation and, 353, 396n3in natural language understanding, 111–115, 389nn27–28

in nominal compounds, 267–268PAROLE project, 389n23Paroubek, P., 350Parse trees, 139Parsingconstituency, 118, 119dependency, 118, 119error handling in, 129–130inconsistency in, 121parse trees, 139syntactic, 28–29

Part of speech (PoS) annotator, 118Partial event identity, 244Paths, objects linked by, 268–269Patient-authoring process (MVP), 320–321, 322Patient biases, detection of, 339–343, 341, 342PDT. See Prague Dependency Treebank (PDT)PENG Light Processable ENGlish, 96Penn Treebank, 52, 53Peptic stricture, 307Perception Interpreter Services, 287, 292Perfect syn-maps, requiring, 122Performance errors, 34, 256–257, 284, 394n11Perlocutionary acts. See Dialog actsPermeation, sense, 102–103Permissive modality, 161Perrault, C. R., 37Personal pronouns. See Pronouns, resolution ofPhenomenological stance, 31–32, 386n28Phrasemes. See Multiword expressions (MWEs)Phraseological units. See Multiword expressions

(MWEs)Physical actions, LEIA-robots, 344–347, 346Physiology modeling, MVP system, 304–305Piantadosi, S. T., 14Pinker, S., 55Plato, 39Pleonastic pronouns, 107–109, 203Plesionyms, 103

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 12: Linguistics for the Age of AI

Index 425

Poesio, M., 45Polylexical expressions, 391n18Polysemyin constructions, 174in fragments, 281–283in nominal compounding experiment, 357

Potential modality, 161Practical Effability, Principle of, 109Pragmatic ambiguity, 3Pragmaticsabductive reasoning, 21–22coreference resolution, 21, 28–29, 44–46corpus annotation, 15, 21, 28–29, 45, 52–55, 386n18descriptive-pragmatic analyses, 20dialog act detection, 46–48, 387nn40–42grounding, 48–49, 387n41reference resolution, 20–21relevance to agent development, 20–22textual inference, 22

Prague Dependency Treebank (PDT), 21, 54–55Preconditions of good practice, 330, 330Predicate nominals, resolving pronominal broad

referring expressions with, 223–224, 227Prepositionsin constructions, 166dynamic sense bunching of, 135–137homographous, 129prepositional phrase attachments, 128, 140preposition swapping, 256–257, 264, 284

Preprocessors, 28, 138Pre-Semantic Analysisconstituency parse, 118, 119decision-making after, 84–85definite description processing, 227dependency parse, 118, 119general principles of, 81outsourcing of, 60tool set for, 117–118

Pre-Semantic Integrationaddition of new construction senses in, 131–134combinatorial complexity in, 134–138, 135decision-making after, 85further exploration of, 139–140general principles of, 81–82new-word learning in, 124–126optimization of imperfect syn-maps in, 126–128, 127

parsing error handling in, 129–130reambiguation of syntactic decisions in, 128–129recovery from production errors in, 124syntactic mapping, 118–124, 120, 123

Preston, L. B., 396n11Priming effect, 12, 342Primitive properties, objects linked by, 248–249Princeton University, Cognitive Science Labora-

tory, 42Principle of least effort, 14–15

Principle of Practical Effability, 109Procedural semantics, 390n6Processing errors, 356Production errors, recovery from, 124, 290Pronominal broad referring expressions, resolution

of, 217–227in machine translation, 392n28negative sentiment terms, 221–223, 226simple example of, 217–218in syntactically simple contexts, 219–221, 226, 393n32

using constructions, 218–219, 226using meaning of predicate nominals, 223–224, 227using selectional constraints, 224–226, 227

Pronouns, resolution ofchallenges of, 206using externally developed engine, 213, 217using lexico-syntactic constructions, 213–215, 217

vetting of hypothesized pronominal coreferences, 215–217

PropBank, 28, 52–53, 388n8Proper namesambiguity of, 208in multiword expression experiment, 361repository of, 68sponsors for, 208, 229

Propertiesdefinition of, 143–144example of, 64–65facets of, 146–147local, 99–100in Maryland Virtual Patient (MVP) system, 306–311, 307, 311, 314, 322, 325, 334

objects linked by, 248–249ontologically decomposable, 249–251optimal inventory of, 70–71recorded values, 143–145semantically decomposable, 70of sets, 154–155stable, 293–294types of, 70, 144–145value mismatches, 228, 233

Property value conflicts, 228Proposition-level semantic enhancementsaspect, 162, 165commands, 164–165, 164matrix verbs, 163, 165modality, 160–161, 161, 165questions, 163–164, 165, 395n22

Prosodic features, 389n6Protégé environment, 26Prototypical concept relationships, 265–266Psycholinguistics, 17, 21, 40–42Pulman, S., 96Purver, M., 385n12Pustejovsky, J., 102

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 13: Linguistics for the Age of AI

426 Index

Quantification, 152–159Query expansion, WordNet and, 43Questions, 113–114, 395n22indirect speech acts, 175–176semantic analysis of, 163–165

Ramsay, A., 104Raskin, V., 102Rationality, normative versus descriptive, 40Reading, learning by, 299–300Reambiguation of syntactic decisions, 128–129Reasoning strategies. See also Situational Reasoningreasoning by analogy, 252–253role in overall agent cognition, 61

Recasens, Marta, 45Recorded property values, 143–145, 159Recoveryfrom parsing errors, 129–130from production errors, 124

Reference resolution, 202. See also Basic Corefer-ence Resolution

coreference resolution compared to, 202–203definition of, 201reference-resolution meaning procedures, 228–229, 233

terminology related to, 202–203Referential ambiguity, 3Referring expressions (RefExes)definition of, 202ellipsis in, 208evaluation experiment, 365–366implicatures in interpretation of, 206in machine translation, 392n28pronominal broad RefExes, 217–227, 392n28, 393n32

situational reference, 292–297sponsors, 202, 207–209, 293–297storing in memory, 209–210, 297universally known entities, 208

Reflexive pronouns, coreference resolution of, 213Related objects, ontology-search strategies forobjects clustered using vague property, 251–252objects filling case role of same event, 249objects linked by ontologically decomposable property, 249–252

objects linked by primitive property, 248–249RELATED-TO property, 251–252RELATION property, 70, 135, 144–145, 394n7Relation-selection approach, 177–179Relationships, semantic, 207Relative spatial expression filter, 371Relative text components, values for, 150–152, 159Relaxable-to facet, 71, 146, 241REQUEST-ACTION concept, 106, 131, 133, 164–165,

164, 289–290REQUEST-INFO concept. See QuestionsRequest-information dialog act, 47

Residual ambiguity, resolution methods for, 247–254

domain-based preferences, 253, 290ontology-search strategies for related objects, 248–252

reasoning by analogy, 252–253speech acts, 261–262, 290–291

Residual hidden meaningshyperbole, 298–299indirect speech acts, 297–298sarcasm, 298

Resnik, P., 349Restrictive modifiers, 229Riau Indonesian, 24Robotics, LEIAs in, 343–347, 346Roncone, A., 344Rosario, B., 177Rule-in/rule-out criteria, 351

Sample bias, 337–338Sampson, G., 53, 104Santa Barbara Corpus of Spoken American English,

124Sarcasm, 298Sayings. See Proverbial expressionsSCALAR-ATTRIBUTEs, 70, 159, 198dynamically computed values for, 145–149RANGE values for, 144–145

Scenic-route knowledge elicitation strategy, 329Schaefer, E. F., 49Schank, R., 6Scheutz, M., 38, 41, 347Schlangen, D., 394n20Schmid, H., 23Scope ambiguity, 3Scopers, 239–240Script-based bridging, 230–231Scripts, 71–73, 305, 388n10, 394n5SDRT, 394n20Search strategies. See Related objects, ontology-

search strategies forSEE-MD-OR-DO-NOTHING evaluation function (MVP),

314–316Selectional constraints, 224–227Semantically decomposable properties, 70Semantically null components. See Null-semmingSemantic analysis. See Basic Semantic Analysis;

Extended Semantic AnalysisSemantic constraints, 74–75Semantic dependency, 3, 362–365Semantic enhancements, proposition-levelaspect, 162, 165commands, 164–165, 164matrix verbs, 163, 165modality, 160–161, 161, 165questions, 163–164, 165, 395n22

Semantic memory, 287, 287

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 14: Linguistics for the Age of AI

Index 427

Semantic representations, 106–107Semantic role labeling, 28–29Semanticsabduction-centered, 21–22distributional, 28–29, 290, 386n27, 394n3formal, 18–20, 30frame, 26imprecision in, 60lexical, 17–18, 42–43Ontological Semantics, 30, 77, 89, 103, 152, 160–161, 161

procedural, 390n6relevance to agent development, 17–18sense bunching, 390n13sentence, 142truth-conditional, 6upper-case, 64, 387n2

Semantic structures. See Sem-strucsSemantic value, 82–83Semantic Web, 27Sem facet, 146–147, 241Sem-strucs (semantic structures), 65–66, 105–106,

109–111Sense bunching, 135–138Sense permeation, 102–103Sentence extraction filterscomparative, 371conditional, 371experiment using, 374–381intrasentential punctuation mark, 371light verb, 372multiple negation, 371no-main-proposition, 371–372relative spatial expression, 371set-based reasoning, 371

Sentence semantics, 142Sentence trimming, 221Sequential feature-matching, 214Set-based reasoning filter, 371Set expressions. See Multiword expressions (MWEs)Setsexamples of, 153–154expansion of, 155, 390n10notation for, 152–153properties of, 154–155set-member relationships, 207as sponsors, 231–232

7% rule, 147–148Shannon, C. E., 5Shirky, C., 27Sidner, C. L., 21Sign language, language evolution of, 24Simon, H., 331SIMPLE project, 105Simpler-first modeling, 91Simplifications, 63Siskind, J., 384

Situational Reasoningdecision-making after, 87–88definite description processing, 233fractured syntax in, 287–290general principles of, 82learning by reading, 299–300need for, 285–286OntoAgent cognitive architecture, 10–13, 11, 30, 115, 286–287

ordered bag of concepts methodology, 211, 288–289residual hidden meanings, 297–299residual lexical ambiguity, 290residual speech act ambiguity, 290–291situational reference, 292–297underspecified known expressions, 291underspecified unknown word analysis, 291–292

Situational reference, 292–297Sloppy coreference. See Type coreferenceSmall sample bias, 337–338, 340SNOMED, 389n20SOAR, 11, 37Social roles, guiding sponsor preferences with, 294Sortal incongruity, 255Source-code generation, 388n10Sowa, J. F., 96Span of text, 202, 206, 217, 219, 392n26Specify-approximation routine, 147–149Speech acts. See Dialog actsSpenader, J., 392n28Sponsor-head-identification algorithm, 236–237Sponsorsadjuncts in, 238–239challenges with, 207–209for definite descriptions (NP-Defs), 229, 231–233definition of, 202no-sponsor-needed instances, 208, 229for referring expressions, 202, 207–209, 293–297verbal/EVENT head of, 235–237

Spoonerisms, 394n11Ssplit annotator, 117Stable properties, 293–294Stanford CoreNLP. See CoreNLP Natural Language

Processing ToolkitStanford Natural Language Inference corpus,

386n18Steen, G., 181Stolcke, A., 47, 48Story understanding, 19–20Stoyanov, V., 44Straightforward constraint matching, 67Strict coreference. See Instance coreferenceStripping of disfluencies, 124Structuralism, 5Stuckardt, R., 45Subevents, 244Subsumption, 43, 113–115Supply-side approach, 302

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 15: Linguistics for the Age of AI

428 Index

Swapping of prepositions, 256–257, 264, 284Syn-mapping. See Syntactic mapping (syn-mapping)Synonyms, 69, 103, 267Synsets, 43Syn-strucs, dynamic sense bunching of, 135–136Syntactic ambiguity, 3Syntactic categories, coreference across, 206Syntactic ellipsis, 211Syntactic mapping (syn-mapping)automatic, 131–134basic strategy for, 118–124, 120, 123fractured syntax in, 287–290optimization of imperfect, 122, 126–128, 127

Syntactic parsing, 28–29, 77, 128–129, 139Syntactic simplicity, 220, 393n32Syntactic structure (syn-struc), 64–68Systemsdefinition of, 91–92system-vetting experiments, 369

Tarski, A., 5Task-oriented evaluations, 351–354Task-oriented methodology, 62–63Tavernise, S., 203ter Stal, W. G., 178Text meaning representations (TMRs). See also

Natural language understanding (NLU)aspect values of, 162–163confidence levels for, 65, 67episodic memory, 77example of, 64–68“golden,” 93–94role in overall agent cognition, 60–61sets in, 152–158storing in memory, 297TMR repository, 252–253understanding the formalism, 64–65

Textual coreference, 201. See also Basic Coreference Resolution

Textual inference, 22, 285Theoretical linguistics, 23construction grammar, 16–17, 165Dynamic Syntax, 17formal approaches to, 5–6generative grammar, 6, 16relevance to agent development, 16–17

Theories, 89Thesauri, 17, 103, 386n14Third-person pronouns, resolution of, 213–215Tiered-grammar hypothesis, 23–25Time of speech, 64, 388n2TMRs. See Text meaning representations (TMRs)Tokenize annotator, 117Topic, concept of, 20, 31TRAINS/TRIPS, 37, 124Transient relaxations, 306Transitive verbs, 28, 124–125, 193

Tratz, S., 177Traum, D., 47, 49, 353Trigger detection, 244Trimming, 221Troponymy, 43Truth-conditional semantics, 6Turn-taking, 387n41Type coreference, 237–239Type-versus-instance coreference decision

algorithm, 238

Uckelman, S. L., 40UMLS, 389n20Unconnected constraints, 265–266Underspecification, 82, 264–279, 390n13. See also

Basic Coreference Resolutioncoreference resolution and, 241–242definition of, 201events of change, 270, 279, 284known expressions, 291nominal compounds, 180, 264–269, 279ungrounded and underspecified comparisons, 270–279, 394n18

unknown word analysis, 291–292Under-the-hood panes (MVP), 324Undesirable things, referring expressions indicating,

221–223, 226Unger, C., 100–101Ungerer, F., 23Ungrounded and underspecified comparisons,

270–279classes of comparatives, 271–277, 272machine learning approach to, 394n18overview of, 270–271reasoning applied in, 277–279value sets for, 271

Unified Modeling Language (UML), 390n1Universal Grammar, 6Universally known entities, 208, 229, 392n12Unknown words, treatment ofduring Basic Semantic Analysis, 178, 194–198during Pre-Semantic Integration, 85, 124–126during Situational Reasoning, 291–292

Unobservable causes, 5, 39, 385n7Upper-case semantics, 64, 387n2Utterance-level constructions, 172–173Utterances, fragmentary, 193

Vague comparisonsin comparative construction, 273–274inward-looking, 275–276point of comparison located elsewhere in text, 276–277

Vague properties, objects clustered by, 251–252Validity, illusion of, 336–337Value facet, 71van der Vet, P. E., 178

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 16: Linguistics for the Age of AI

Index 429

Verbal actions, 62Verbal/EVENT head of sponsors, 235–237Verbmobil project, 40–41VerbNet, 25–26Verb phrase (VP) ellipsisaspect + OBJECT resolution, 241–242constructions, 187–189, 188,192coreference between objects in, 238–239definition of, 186–187, 192evaluation experiment, 366–369past work on, 393n44

Verbsaspectual, 186–189, 192, 240, 241–242in constructions, 166–167in copular metaphors, 184–185coreferential events expressed by, 242–244dynamic sense bunching of, 135–136light, 107–109matrix, 163, 165new-word learning of, 124–125optionally transitive, 193phrasal, 128–129transitive, 28, 124–125, 193unknown, 196

Véronis, J., 103Versley, Y., 45Vertical incrementality, 77–82, 80, 84Vetting sponsors for referring expressions, 293–296ViPER (Verb Phrase Ellipsis Resolver), 367–369Virtual ontological concepts, sem-strucs as,

105–106Visualization of disease models, 320–324under-the-hood panes, 324ontological knowledge, 321–324, 323patient-authoring interface, 320–321, 322

Visual meaning representations (VMRs), 296–297Volitive modality, 161VP ellipsis. See Verb phrase (VP) ellipsis

Wall Street Journal, 270, 356, 359Weaver, W., 5, 56Webber, B. L., 21Wiebe, J., 388n8Wiener, N., 5Wilkes-Gibbs, D., 49Wilks, Y., 6, 18–19, 34, 94–95, 388n17, 389n22Wilson, C., 139Window of coreference, 202Winograd, T., 6Winograd Schema Challenge, 213, 295–296Winston, P. H., 19Winther, R. G., 89Wittenberg, E., 23–24Wittgenstein, L., 94Woods, W. A., 6WordNet, 19, 26, 42–43, 95, 103, 115, 221Wordnets, 17–18, 98

Word sense disambiguation (WSD), 50–51Worknik, 115

XMRs, 296–297. See also Text meaning representa-tions (TMRs); Visual meaning representations (VMRs)

Yuret, D., 26

Zaenen, A., 22, 29, 33

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021

Page 17: Linguistics for the Age of AI

Downloaded from http://direct.mit.edu/books/book/chapter-pdf/1891686/9780262363136_c001200.pdf by guest on 28 May 2021