File size: 58,387 Bytes
6fa4bc9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 |
{
"paper_id": "O08-5001",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T08:02:43.168648Z"
},
"title": "Knowledge Representation and Sense Disambiguation for Interrogatives in E-HowNet",
"authors": [
{
"first": "Shu-Ling",
"middle": [],
"last": "Huang",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Academia Sinica",
"location": {}
},
"email": "slhuang@mail.nhcue.edu.tw"
},
{
"first": "Keh-Jiann",
"middle": [],
"last": "Chen",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Academia Sinica",
"location": {}
},
"email": "kchen@iis.sinica.edu.tw"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "In order to train machines to 'understand' natural language, we propose a meaning representation mechanism called E-HowNet to encode lexical senses. In this paper, we take interrogatives as examples to demonstrate the mechanisms of semantic representation and composition of interrogative constructions under the framework of E-HowNet. We classify the interrogative words into five classes according to their query types, and represent each type of interrogatives with fine-grained features and operators. The process of semantic composition and the difficulties of representation, such as word sense disambiguation, are addressed. Finally, machine understanding is tested by showing how machines derive the same deep semantic structure for synonymous sentences with different surface structures.",
"pdf_parse": {
"paper_id": "O08-5001",
"_pdf_hash": "",
"abstract": [
{
"text": "In order to train machines to 'understand' natural language, we propose a meaning representation mechanism called E-HowNet to encode lexical senses. In this paper, we take interrogatives as examples to demonstrate the mechanisms of semantic representation and composition of interrogative constructions under the framework of E-HowNet. We classify the interrogative words into five classes according to their query types, and represent each type of interrogatives with fine-grained features and operators. The process of semantic composition and the difficulties of representation, such as word sense disambiguation, are addressed. Finally, machine understanding is tested by showing how machines derive the same deep semantic structure for synonymous sentences with different surface structures.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Electronic dictionaries are designed for the purpose of providing users (or computers) convenient access to relevant knowledge of words to understand language. When we say that a sentence is 'understood', we mean that the concepts and the conceptual relations expressed by the sentence are unambiguously identified and we can make the correct inferences/responses. To have a computer understand a sentence, we must have a framework for representing lexical knowledge and performing semantic composition and disambiguation processes. other concepts. Compared to WordNet, HowNet's architecture provides richer information apart from hyponymy relations. It also enriches relational links between words via encoded feature relations. The advantages of HowNet are (a) its inherent properties are derived from encoded feature relations in addition to hypernym concepts, and (b) information regarding conceptual differences between different concepts and information regarding morph-semantic structure are encoded. Therefore, we adopt a similar mechanism to define word sense in E-HowNet, but represent concepts in a more accurate and flexible way by (a) defining new concepts by well-defined concepts, (b) providing a uniform representational framework for both function words and content words, and (c) embedding semantic composition and decomposition capabilities. More detailed discussions can be seen at [Chen et al. 2005] .",
"cite_spans": [
{
"start": 1402,
"end": 1420,
"text": "[Chen et al. 2005]",
"ref_id": "BIBREF1"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "In E-HowNet, we define each lexical sense by the composition of well-defined concepts and/or basic concepts, called sememes in HowNet. The sememes are linked to their sense equivalence WordNet synsets [Fellbaum 1998 ]. Take \u571f\u5730\u516c 'God of earth' as an example:",
"cite_spans": [
{
"start": 201,
"end": 215,
"text": "[Fellbaum 1998",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "(1)",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "\u571f\u5730\u516c Tu di gong 'God of earth' Def:{God|\u795e:telic={manage|\u7ba1\uf9e4:patient={land|\uf9d3\u5730},agent={~}}} Here, 'God' is the hypernym of the target word \u571f\u5730\u516c 'God of earth', 'manage' and 'land' are its related concepts. 'telic', 'patient', and 'agent' are relations which link these concepts. Obviously, to achieve mechanical understanding of natural language, the same or similar concepts must have the same or similar underlying semantic representation. However, natural language can be ambiguous. Different sentences might express the same meaning, and the same sentence can also express different meanings. The following sentences (2) and 3show the former phenomenon, and (4) and (5) show the latter:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "(2) \u6211\u80fd\u5426\u62cd\u7167\uff1f Wo neng fou pai zhao? Is it OK for me to take pictures?",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "(3) \u6211\u53ef\uf967\u53ef\u4ee5\u7167\u76f8\uff1f Wo ke bu ke yi zhao xiang? Can I take photos?",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "(4) \u571f\u5730\u516c\u6709\u653f\u7b56\u3002Tu di gong you zheng ce. The policy of public sharing of the land.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "(5) \u571f\u5730\u516c\u6709\u653f\u7b56\u3002Tu di gong you zheng ce. God of earth has his policy.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "Thus, transforming the surface structure of a sentence into a canonical semantic representation",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "Interrogatives in E-HowNet and simultaneously solving the problem of word sense ambiguity are major research issues. In summary, lexical semantic representation and composition (including disambiguation) are the most demanding techniques for understanding natural language by machines and the design of E-HowNet is aimed for these objectives.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "In this paper, we will take interrogatives as examples to demonstrate the mechanism of lexical semantic representation and composition in E-HowNet [Chen et al. 2004] . The goal is to achieve near canonical semantic representation for synonyms and sense equivalent sentences. Take sentences (2) and (3) as examples. Although their syntactic structure and surface strings are very different, by composing lexical sense representations, we hope the machine can 'understand' synonymy of sentences in different surface forms.",
"cite_spans": [
{
"start": 147,
"end": 165,
"text": "[Chen et al. 2004]",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "Analysis of interrogative constructions is of great interest to linguists, as well as to computer scientists, for example, those who are engaged in QA techniques. Interrogative constructions have played a central role in the development of modern syntactic theory. Ginzburg and A. Sag [2000] have pointed out that the interrogative has been at the heart of work in generative grammar, along with government and binding (GB) theory and head-driven phrase structure grammar (HPSG). Nonetheless, to date, most syntacticians take quite different approaches from semantic and pragmatic points of view on interrogatives. Taking questions in Mandarin Chinese as example, Shao [1996] has summed up the current study of interrogatives and listed the main research themes as follows: the type of question, interrogative particles, querying focus and its answer, degree of doubt, special interrogative sentence patterns, etc. Most of the above themes are purely grammatical analysis. To build a frame-based entity-relation knowledge representation model, we find interrogative construction a good and challenging example because it combines problems of syntax, semantics, and pragmatics.",
"cite_spans": [
{
"start": 265,
"end": 291,
"text": "Ginzburg and A. Sag [2000]",
"ref_id": "BIBREF5"
},
{
"start": 664,
"end": 675,
"text": "Shao [1996]",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "In the following section, we briefly describe the previous works for interrogatives. Then, we introduce our analysis of type classifications for interrogatives and their representation in E-HowNet. Next, we present the semantic composition process of interrogative sentences and the difficulties encountered. We conclude the paper by discussing our results and future work.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1."
},
{
"text": "Interrogatives in Chinese studies are traditionally attributed to the mood category of syntax. Ma [1935] wrote the first grammar book for Mandarin Chinese. He classified interrogatives into the mood category. Later, Li [1930] and Lv [1942] carried forward his viewpoint and influenced modern linguistic theory on interrogatives deeply. Most linguists consider there to be four grammatical types that explicitly mark an utterance as an interrogative [Lv 1942; Li and Thompson 1997; Tang 1983; Lu 1984; Shao 1996] . First, is a question which can be answered by 'yes' or 'no', called a factual question, true and false interrogative, or yes/no interrogative. Second, is a question which includes Wh-words such as 'who', 'what', and 'when', called a Wh-word interrogative, or an information seeking interrogative. Third, is a question which mentions two or more possible alternative answers, called a disjunctive interrogative or an either/or interrogative. Fourth, is a question which is composed of a statement followed by an A not A form, such as dui bu dui 'right or wrong ' ",
"cite_spans": [
{
"start": 98,
"end": 104,
"text": "[1935]",
"ref_id": null
},
{
"start": 216,
"end": 225,
"text": "Li [1930]",
"ref_id": null
},
{
"start": 233,
"end": 239,
"text": "[1942]",
"ref_id": null
},
{
"start": 449,
"end": 458,
"text": "[Lv 1942;",
"ref_id": null
},
{
"start": 459,
"end": 480,
"text": "Li and Thompson 1997;",
"ref_id": "BIBREF6"
},
{
"start": 481,
"end": 491,
"text": "Tang 1983;",
"ref_id": null
},
{
"start": 492,
"end": 500,
"text": "Lu 1984;",
"ref_id": null
},
{
"start": 501,
"end": 511,
"text": "Shao 1996]",
"ref_id": null
},
{
"start": 1074,
"end": 1075,
"text": "'",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Background",
"sec_num": "2."
},
{
"text": "For QA applications, we are more concerned about semantic discrimination of different interrogatives. Therefore, we take a sense-based approach to create a hierarchical classification which is guided by a layered semantic hierarchy of answer types, and eventually classify interrogative sentences into fine-grained classes, shown as (8):",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Our Classification of Interrogatives",
"sec_num": "3.1"
},
{
"text": "Interrogatives in E-HowNet 8interrogative -(A) true/false (yes/no) interrogative Who knows where I can find this game?",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Our Classification of Interrogatives",
"sec_num": "3.1"
},
{
"text": "Sentence (9) belongs to the true/false interrogative type and the entire statement is a querying focus. Dissimilarly, Sentence (10) indicates two querying foci through using different interrogatives, 'who' and 'where'. In other words, the true/false interrogative asks truth value of the positive predication of the sentence. Then, the Wh-interrogative is used to ask for information. By analyzing the querying focus, Wh-interrogatives can be further divided into four types: (B-a) asking factual information, such as time, location, quantity, and so forth; (B-b) asking relationship, such as kinship; (B-c) asking opinion or attitude, such as possibility, capacity, volition, etc.; and the last, (B-d) asking to choose an option. Sentence (10) refers to type (B-a). For the remaining types, we give an example of each as follows: Rice rinsing water is acidic or alkaline?",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Our Classification of Interrogatives",
"sec_num": "3.1"
},
{
"text": "Here, the fine-grain distinctions between interrogative type (B-a) and (B-b) and between interrogative type (A) and (B-c) are clarified below. For instance, sentence (11) refers to type (B-b), but why do we need to separate it from type (B-a) when they both use 'what' to make questions? In sentence (11), the question word \uf9fd\u9ebc she me 'what' asks for relationship but not the type of a frame element or the value of a semantic role as exemplified in (10). 2 The semantic representation of a complex relation is different from the representation of entities. Therefore, we differentiate between interrogative type (B-a) and (B-b). Chen et al. [2004] proposed a compositional mechanism to describe complex relations. For example, we express 'mother in law' as 14:",
"cite_spans": [
{
"start": 455,
"end": 456,
"text": "2",
"ref_id": null
},
{
"start": 629,
"end": 647,
"text": "Chen et al. [2004]",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Our Classification of Interrogatives",
"sec_num": "3.1"
},
{
"text": "(14) mother in law def:{human|\u4eba=mother(spouse({x:human|\u4eba}))}",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Our Classification of Interrogatives",
"sec_num": "3.1"
},
{
"text": "According to the representation model, when our querying focus is a complex relation, we put a question mark before the relation role, such as mother, spouse, parents, etc., to mark the query focus. Representational detail is shown in the next section.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Our Classification of Interrogatives",
"sec_num": "3.1"
},
{
"text": "Second, some may argue that there is no distinction between type (A) and (B-c). Example (12) is of type (B-c); we find it may have a yes/no answer, a typical characteristic of type (A).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Our Classification of Interrogatives",
"sec_num": "3.1"
},
{
"text": "Can he eat hot peppers?",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Ta ke bu ke yi chi la jiao?",
"sec_num": null
},
{
"text": "However, if we compare the meaning of sentences (12) and (15), we can still find a slight difference between a yes/no question and a question of asking opinion. Sentence (12) has the meaning of asking the hearer's permission, but (15) does not. [1996] has classified the A not A form into five classes according to A's part of speech, shown as follows: (a) A is a copula e.g. \u662f\uf967\u662f shi bu shi 'be or be not' (b) A is a modal word e.g. \u597d\uf967\u597d hao bu hao 'ok or not ok' (c) A is an auxiliary e.g. \u80af\uf967\u80af ken bu ken 'willing or not willing' (d) A is a verb e.g. \u61c2\uf967\u61c2 dong bu dong 'understand or not understand' (e) A is an adjective e.g. \u7f8e\uf967\u7f8e mei bu mei 'beautiful or not beautiful'. From the semantic perspective, we merge (a), (d), (e) and (b), (c) to re-divide these five categories into two categories, i.e. modal A not A interrogatives and other A not A interrogatives.",
"cite_spans": [
{
"start": 245,
"end": 251,
"text": "[1996]",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Ta ke bu ke yi chi la jiao?",
"sec_num": null
},
{
"text": "In E-HowNet, we made distinctions between content sense and relational sense thereby representing senses of content words and senses of function words in different ways. For instance, in (16), the content words 'bathe' and 'cold water' are represented differently from the function word 'with'. In this example, the function word 'with' plays the role of 'instrument' which links the relation between its argument 'cold water' and the matrix verb 'bathe'. ",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Knowledge Representation for Interrogatives",
"sec_num": "3.2"
},
{
"text": "{clean|\u4f7f\u6de8: patient={body|\u8eab\u9ad4}, instrument={water|\u6c34:temperature={cold|\uf92e}}}.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Result of semantic composition:",
"sec_num": null
},
{
"text": "In much the same way, interrogative words have more relational sense than content sense, so they are defined by semantic role to denote relational sense and use the operator '.Ques.' to mark the querying focus, i.e. the object or its discrimination features which speakers want to know.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Result of semantic composition:",
"sec_num": null
},
{
"text": "According to the classification of interrogatives above, we represent each type of interrogative as follows: 17 We use two operators, .Ques. and .Option., to denote querying focus or optional items. More detailed sense representations for interrogatives are shown in Table 1 . [Li et al. 1999] . To check the completeness of the above table and to find the distribution of query types, we randomly extracted 1% of the sentences with question marks from the Sinica corpus to see the coverage of the above table, and the results of the distribution are shown in Table 2 . There are 203 sentences, 9 of which do not contain any query word listed in Table 1 . Their query sense is expressed only by a question mark in the end of the sentence. However, we can insert 'ma' \u55ce or 'ne' \u5462 before the question mark to these sentences to make them true/false interrogatives, so statistically they are still counted as true/false interrogatives. We can conclude that most of the questions ask true/false value or factual information. Furthermore, we can classify each of them according to Table 1 . ",
"cite_spans": [
{
"start": 277,
"end": 293,
"text": "[Li et al. 1999]",
"ref_id": null
}
],
"ref_spans": [
{
"start": 267,
"end": 274,
"text": "Table 1",
"ref_id": "TABREF5"
},
{
"start": 560,
"end": 567,
"text": "Table 2",
"ref_id": "TABREF6"
},
{
"start": 646,
"end": 653,
"text": "Table 1",
"ref_id": "TABREF5"
},
{
"start": 1076,
"end": 1083,
"text": "Table 1",
"ref_id": "TABREF5"
}
],
"eq_spans": [],
"section": "Result of semantic composition:",
"sec_num": null
},
{
"text": "In addition to the above interrogatives, there are also some derived interrogative compounds, which are impossible to list comprehensively in our database. Thus, generating their sense representation automatically has to be accomplished. Most interrogative compounds are composed of an interrogative determinative and a noun. Since the number of interrogative determinatives is limited, have already been defined in Table 1 , and listed in (18), we design the rules in (19) to derive the sense representations of interrogative compounds. As the question determinatives \u5e7e ji and \u54ea na are ambiguous, in addition to the general rule (A),(B),(C),(D), we also provide two other rules (E) and (F) to disambiguate their word sense.",
"cite_spans": [],
"ref_spans": [
{
"start": 416,
"end": 423,
"text": "Table 1",
"ref_id": "TABREF5"
}
],
"eq_spans": [],
"section": "Knowledge Representation for Interrogative Compounds",
"sec_num": "3.3"
},
{
"text": "The previous discussion has been about semantic representation of lexical senses. To establish a formal system to handle the task of language understanding, we also need to address the issue of semantic composition. To understand Chinese sentences, after the word segmentation and parsing process, we get coarse-grained head-argument event structures of the sentences. Then, we try to project surface syntax onto the semantic structure for establishing truly integrated semantic relations. In Example (20), we identify the sentential head 'lose' after the parsing process, and, based on E-HowNet, the arguments of event 'lose' are 'possessor' and 'possession'; thus we know the 'data' here is the possession of 'lose'. Therefore, the result of composition is as follows: ",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Semantic Composition for Interrogatives",
"sec_num": "4."
},
{
"text": "To achieve the goal of automatic semantic composition, we have to solve the problem of word sense disambiguation. In Chinese, \uf9fd\u9ebc she me, \u600e\u9ebc ze me, \u600e\u6a23 ze yang, \u591a duo, \u54ea na, and \u5e7e ji are the most frequently used interrogatives, and they all have ambiguous senses. In addition to Rule (19), their sense disambiguation rules are discussed below:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Sense Disambiguation",
"sec_num": null
},
{
"text": "\uf9fd\u9ebc she me /\u4f55 he/ \u5565 sha \uf9fd\u9ebc she me 'what' plays the grammatical functions of adjective and pronoun, and there are two senses for each function. Accordingly, we generate four rules to disambiguate the word senses of \uf9fd\u9ebc she me, and the details are shown in Table 3 . \u4f55 he and \u5565 sha are its literary and slangy usage, hence, share the same disambiguation rules. In Mandarin Chinese, \uf9ba le, \u9084 hai and the negative words \u6c92 mei, \uf967 bu all mark the endpoint of event; therefore, we use them as constraints to disambiguate the senses of \u600e\u9ebc ze me. Additionally, when a modal word appears between \u600e\u9ebc ze me and the event, normally the event also has an endpoint, that is, \u600e\u9ebc ze me questions 'reason' here. Unlike",
"cite_spans": [],
"ref_spans": [
{
"start": 253,
"end": 260,
"text": "Table 3",
"ref_id": "TABREF10"
}
],
"eq_spans": [],
"section": "Sense Disambiguation",
"sec_num": null
},
{
"text": "Interrogatives in E-HowNet (36) \u660e\u5929\u600e\u6a23 ming tian ze yang 'how about tomorrow' Restore ellipsis:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Sense Disambiguation",
"sec_num": null
},
{
"text": "\u660e\u5929\u4e00\u8d77\u53bb\u600e\u6a23 ming tian yi qi qu ze yang 'how about go together tomorrow' def:{go|\u53bb:manner={together|\u5171\u540c},time={tomorrow|\u660e\u5929},willingness={.Ques.}} However, the method for recovering the omitted part of the surface sentence is out of the scope of this paper.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Sense Disambiguation",
"sec_num": null
},
{
"text": "\u591a duo \u591a duo 'how' also plays the role of an adverb. It's usually followed by an attribute value, such as \u751c tian 'sweet', \u8070\u660e cong ming 'smart', \u9060 yuan 'far', \u5927 da 'big'. It can be used to express feelings of exclamation or doubt. We can not simply distinguish these two senses by the context, but we need to rely on the tone. For this reason, we will deal only with the senses of doubt. Incidentally, it is always possible to turn a declarative statement into a question by using a slightly rising intonation pattern. For the same reason, we do not deal with such sentences and a few interrogative words, such as \u554a a, \u5427 ba, \u5462 ne, as well.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Sense Disambiguation",
"sec_num": null
},
{
"text": "\u591a duo 'how' with interrogative sense can be represent as below: The semantic roles of 'sweetness', 'smartness', 'distance', and ''size' are inferred from their respective values 'sweet|\u751c', 'smart|\u8070\u660e', 'far|\u9060', and 'big|\u5927' by checking the taxonomy of feature-value hierarchy in E-HowNet.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Sense Disambiguation",
"sec_num": null
},
{
"text": "(37) \u591a duo",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Sense Disambiguation",
"sec_num": null
},
{
"text": "For the disambiguation of \u54ea and \u5e7e, please see Section 3.3 (19) rules (E) and (F).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "\u54ea na \u5e7e ji",
"sec_num": null
},
{
"text": "To achieve near canonical semantic representation, we studied the semantic representation and composition of interrogatives. According to the semantic classification of interrogatives, we represent interrogatives in a hierarchy as follows: Although the surface structures of (2), (3) are different, we find that their result of composition is the same. It means that, by means of E-HowNet representation, a machine can judge the sense similarity of words, phrases, and sentences and achieve machine understanding. However, this is only an illustration by example. For future research, we will implement a parsing system incorporated with the E-HowNet model to perform semantic composition process practically. To achieve this goal, apart from sense disambiguation, we find that discordance between syntactic structure and semantic relations is another critical problem. Take The E-HowNet sense representation of (39) is: def:{hard|\u8f9b\u82e6:theme={travel|\uf983\ufa08:distance={far|\u9060}},degree={very|\u5f88},truth={.Ques.}}",
"cite_spans": [],
"ref_spans": [
{
"start": 870,
"end": 874,
"text": "Take",
"ref_id": null
}
],
"eq_spans": [],
"section": "Conclusion and Future Works",
"sec_num": "5."
},
{
"text": "Comparing the semantic representation with syntactic structure, we find rhetorical interrogative \uf967\u662f\u55ce bu shi ma 'Isn't it' is segmented into three words in syntax analysis, but in the semantic point of view, they are integrated into one word and represented as 'truth={.Ques.}'. There are still many types of discordance between syntactic structure and semantic relations that need to be studied. Furthermore, we have to find the mapping rules and match coarse-grained syntactic arguments to fine-grained semantic relations in the future. These results are applicable to both declarative sentences and interrogative sentences.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and Future Works",
"sec_num": "5."
},
{
"text": "In conclusion, this study sheds new light on designing better and accurate question-answering systems because E-HowNet representation of questions not only represents their senses, but also marks the focused information to be answered. In addition, the proposed representational scheme also provides a way to convert a given sentence into a near-canonical sense representation. Therefore, the design of Chinese QA system will be our future task.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and Future Works",
"sec_num": "5."
},
{
"text": "Tag interrogative is formed by adding a short A not A question form of certain verbs as a tag to a statement. In this paper, we regard it as a general A not A question type due to the same semantic performance.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "The disambiguation of \uf9fd\u9ebc, see section 4.1.1.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "In this paper, our focus is semantic representation, so we don't discuss the interrogative words '\u554a a';'\u5440 a' ; '\u56c9 luo'; '\u4e4e hu'; '\u5427 ba' or '\u5594 wo'. This is because the tone decides if they are interrogative words or not.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "The meaningful or dummy measure words are listed in Tai C.H. et al. \"A Semantic Composition Method for Deriving Sense Representations of Determinative-Measure Compounds in E-HowNet\" in the proceeding of ROCLING 2008, Taipei, Taiwan.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "This research was supported in part by the National Science Council under a Center Excellence Grant NSC 96-2752-E-001-001-PAE and Grant N S C96-2221-E-001-009.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgement",
"sec_num": null
},
{
"text": "this, when \u600e\u9ebc ze me appears between the modal word and event, the event does not have an endpoint, so \u600e\u9ebc ze me questions 'means' instead. Their differences can be found in example (33). In the case of entity+\u600e\u6a23 ze yang /\u5982\u4f55 ru he, it usually happens when the main verb of the sentence is omitted. In such cases, we must first recover the omitted part of the surface sentence based on the context then infer the complete sense. For example,",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "annex",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Extended-HowNet: multi-level concept definition and complex relation description",
"authors": [
{
"first": "K",
"middle": [
"J"
],
"last": "Chen",
"suffix": ""
},
{
"first": "S",
"middle": [
"L"
],
"last": "Huang",
"suffix": ""
},
{
"first": "Y",
"middle": [
"Y"
],
"last": "Shih",
"suffix": ""
},
{
"first": "Y",
"middle": [
"J"
],
"last": "Chen",
"suffix": ""
}
],
"year": 2004,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chen, K. J., S. L. Huang, Y. Y. Shih, and Y. J. Chen, \"Extended-HowNet: multi-level concept definition and complex relation description,\" 2004, Peking University.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Extended-HowNet-A Representational Framework for Concepts",
"authors": [
{
"first": "K",
"middle": [
"J"
],
"last": "Chen",
"suffix": ""
},
{
"first": "S",
"middle": [
"L"
],
"last": "Huang",
"suffix": ""
},
{
"first": "Y",
"middle": [
"Y"
],
"last": "Shih",
"suffix": ""
},
{
"first": "Y",
"middle": [
"J"
],
"last": "Chen",
"suffix": ""
}
],
"year": 2005,
"venue": "Proceedings of IJCNLP-05 Workshop on Lexical Semantic",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chen, K. J., S. L. Huang, Y.Y. Shih, and Y. J. Chen, \"Extended-HowNet-A Representational Framework for Concepts,\"In Proceedings of IJCNLP-05 Workshop on Lexical Semantic, 2005, Jeju Island, South Korea.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Modality and Modal Sense Representation in E-HowNet",
"authors": [
{
"first": "Y",
"middle": [
"S"
],
"last": "Chung",
"suffix": ""
},
{
"first": "S",
"middle": [
"L"
],
"last": "Huang",
"suffix": ""
},
{
"first": "K",
"middle": [
"J"
],
"last": "Chen",
"suffix": ""
}
],
"year": 2007,
"venue": "the Proceeding of the 21st Pacific Asia Conference on Language, Information and Computation (PACLIC)",
"volume": "",
"issue": "",
"pages": "136--145",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Chung, Y. S., S. L. Huang, and K. J. Chen, \"Modality and Modal Sense Representation in E-HowNet,\" In the Proceeding of the 21st Pacific Asia Conference on Language, Information and Computation (PACLIC), 2007, Seoul National University, pp. 136-145.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Interrogative Investigations: The Form, Meaning and Use of English Interrogatives",
"authors": [
{
"first": "J",
"middle": [],
"last": "Ginzburg",
"suffix": ""
},
{
"first": "I",
"middle": [
"A"
],
"last": "Sag",
"suffix": ""
}
],
"year": 2000,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ginzburg, J., and I. A. Sag, Interrogative Investigations: The Form, Meaning and Use of English Interrogatives, Stanford: CSLI Publications, 2000.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "\u77f3\u5b9a\u6829(Shi, Dingxu)\uff0c\u5171\u6027\u8207\u500b\u6027-\u6f22\u8a9e\u8a9e\u8a00\u5b78\u4e2d\u7684\u722d\u8b70\uff0c\uf963\u4eac\u8a9e\u8a00\u6587\u5316\u5927\u5b78\u51fa\u7248\u793e\uff0c\uf963 \u4eac\uff0c1999",
"authors": [
{
"first": "C",
"middle": [
"N"
],
"last": "Li",
"suffix": ""
},
{
"first": "S",
"middle": [
"A"
],
"last": "Thompson",
"suffix": ""
},
{
"first": "Mandarin",
"middle": [],
"last": "Chinese",
"suffix": ""
}
],
"year": 1997,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Li, C. N., and S. A. Thompson, Mandarin Chinese, The Crane Publishing Co., LTD., 1997. \u77f3\u5b9a\u6829(Shi, Dingxu)\uff0c\u5171\u6027\u8207\u500b\u6027-\u6f22\u8a9e\u8a9e\u8a00\u5b78\u4e2d\u7684\u722d\u8b70\uff0c\uf963\u4eac\u8a9e\u8a00\u6587\u5316\u5927\u5b78\u51fa\u7248\u793e\uff0c\uf963 \u4eac\uff0c1999.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Jianzhong)\uff0c\u99ac\u6c0f\u6587\u901a\uff0c\u5546\u52d9\u5370\u66f8\u9928\uff0c\u53f0\uf963\uff0c1935. Interrogatives in E-HowNet",
"authors": [
{
"first": "",
"middle": [],
"last": "\u99ac\u5efa\u5fe0(ma",
"suffix": ""
}
],
"year": null,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "\u99ac\u5efa\u5fe0(Ma, Jianzhong)\uff0c\u99ac\u6c0f\u6587\u901a\uff0c\u5546\u52d9\u5370\u66f8\u9928\uff0c\u53f0\uf963\uff0c1935. Interrogatives in E-HowNet",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Jianming)\uff0c \"\u95dc\u65bc\u73fe\u4ee3\u6f22\u8a9e\uf9e8\u7684\u7591\u554f\u8a9e\u6c23\u8a5e\"\uff0c\u6536\uf93f\u65bc\u4e2d\u570b\u8a9e\u6587\uff0c\u7b2c\u4e94\u671f\uff0c1984",
"authors": [
{
"first": "",
"middle": [],
"last": "\uf9d3\u5109\u660e(lu",
"suffix": ""
}
],
"year": null,
"venue": "",
"volume": "",
"issue": "",
"pages": "330--332",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "\uf9d3\u5109\u660e(Lu, Jianming)\uff0c \"\u95dc\u65bc\u73fe\u4ee3\u6f22\u8a9e\uf9e8\u7684\u7591\u554f\u8a9e\u6c23\u8a5e\"\uff0c\u6536\uf93f\u65bc\u4e2d\u570b\u8a9e\u6587\uff0c\u7b2c\u4e94\u671f\uff0c1984, pp. 330-332.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Chinhsi)\uff0c\u65b0\u8457\u570b\u8a9e\u6587\u6cd5(Chinese grammar on the national language)\uff0c\u5546\u52d9\u5370\u66f8 \u9928\uff0c\u53f0\uf963\uff0c1930",
"authors": [
{
"first": "",
"middle": [],
"last": "\uf989\u9326\u7199(li",
"suffix": ""
}
],
"year": null,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "\uf989\u9326\u7199(Li, Chinhsi)\uff0c\u65b0\u8457\u570b\u8a9e\u6587\u6cd5(Chinese grammar on the national language)\uff0c\u5546\u52d9\u5370\u66f8 \u9928\uff0c\u53f0\uf963\uff0c1930.",
"links": null
}
},
"ref_entries": {
"TABREF3": {
"text": "Does he like to eat hot peppers? Therefore, in Mandarin Chinese,\u53ef\uf967\u53ef\u4ee5 ke bu ke yi 'can or cannot' and \u559c\uf967\u559c\u6b61 xi bu xi huan 'like or do not like' both have the A not A form, but from the semantic point of view, they belong to different types. Shao",
"num": null,
"content": "<table><tr><td>Interrogatives in E-HowNet</td></tr><tr><td>(15) \u4ed6\u559c\uf967\u559c\u6b61\u5403\u8fa3\u6912\uff1f</td></tr><tr><td>Ta xi bu xi huan chi la jiao?</td></tr></table>",
"type_str": "table",
"html": null
},
"TABREF5": {
"text": "The interrogatives above are gathered from Li and Thompson's analysis and are integrated by checking over 1000 question titles manually in Baidu knows",
"num": null,
"content": "<table><tr><td>Question types</td><td>Words</td><td>Sense representation</td></tr><tr><td/><td>\u55ce ma 3</td><td/></tr><tr><td>true/false interrogatives</td><td/><td>def: truth={.Ques.}</td></tr><tr><td/><td>A not A</td><td/></tr><tr><td/><td>\u8ab0 shei 'who'</td><td>def: participant={animate: formal={.Ques.}}</td></tr><tr><td/><td>\u5e7e\u9ede\u9418 ji dian zhong ' what time'</td><td>def: time={.Ques.}</td></tr><tr><td/><td>\uf9fd\u9ebc 1 she me 'what'</td><td/></tr><tr><td/><td>\u4f55 1 he 'what'</td><td>def: participant={inanimate:formal={.Ques.}}</td></tr><tr><td>Wh-word interrogatives:</td><td>\u5565 1 sha 'what'</td><td/></tr><tr><td>asking factual information</td><td>\uf9fd\u9ebc 2 she me 'what'</td><td/></tr><tr><td/><td>\u4f55 2 he 'what'</td><td>def: formal={.Ques.}</td></tr><tr><td/><td>\u5565 2 sha 'what'</td><td/></tr><tr><td/><td>\u70ba\u4f55 wei he 'why'</td><td/></tr><tr><td/><td>\u4f55\u4ee5 he yi 'why'</td><td>def: reason={.Ques.}</td></tr><tr><td/><td>\u4f55\uf967 he bu 'why not'</td><td/></tr></table>",
"type_str": "table",
"html": null
},
"TABREF6": {
"text": "",
"num": null,
"content": "<table><tr><td>Question types</td><td>Words</td><td colspan=\"2\">Number Total(203)</td></tr><tr><td/><td>\u55ce ma</td><td>50</td><td/></tr><tr><td/><td>\u5462 ni</td><td>16</td><td/></tr><tr><td/><td>\u662f\u5426 shi fou 'whether'</td><td>10</td><td/></tr><tr><td>true/false interrogatives</td><td>\u6709\u6c92\u6709 you mei you 'have'</td><td>1</td><td>98</td></tr><tr><td/><td>\u662f\uf967\u662f shi bu shi 'is it'</td><td>6</td><td/></tr><tr><td/><td>\uf967\u662f\u55ce bu shi ma 'isn't it'</td><td>1</td><td/></tr><tr><td/><td>A not A</td><td>14</td><td/></tr><tr><td/><td>\u8ab0 shei 'who'</td><td>4</td><td/></tr><tr><td/><td>\u5e7e\u9ede\u9418 ji dian zhong 'what time'</td><td>2</td><td/></tr><tr><td/><td>\uf9fd\u9ebc 1 she me 'what'</td><td>14</td><td/></tr><tr><td/><td>\u4f55 1 he 'what'</td><td>1</td><td/></tr><tr><td/><td>\u5565 1 sha 'what'</td><td>0</td><td/></tr><tr><td>Wh-word interrogatives: asking factual information</td><td>\uf9fd\u9ebc 2 she me 'what'</td><td>13</td><td>94</td></tr><tr><td/><td>\u4f55 2 he 'what'</td><td>2</td><td/></tr><tr><td/><td>\u5565 2 sha 'what'</td><td>0</td><td/></tr><tr><td/><td>\u70ba\u4f55 wei he 'why'</td><td>2</td><td/></tr><tr><td/><td>\u4f55\u4ee5 he yi 'why'</td><td>0</td><td/></tr><tr><td/><td>\u4f55\uf967 he bu 'why not'</td><td>1</td><td/></tr></table>",
"type_str": "table",
"html": null
},
"TABREF7": {
"text": "The rules for deriving sense representations of interrogative compounds If the morphological structure of the interrogative compound is ID+head. \u4f55\u6642 he shi 'when' =def: time={.Ques.} e.g. \uf9fd\u9ebc\u5f62\uf9fa she me xing zhuang 'what shape' =def: shape={.Ques.} e.g. \u54ea\u8272 na se 'what color' =def: color={.Ques.}}",
"num": null,
"content": "<table><tr><td colspan=\"3\">Interrogatives in E-HowNet</td></tr><tr><td colspan=\"3\">(D) if head=MM(meaningful measure words)</td><td>then</td><td>def: {MM: {ID} }</td></tr><tr><td colspan=\"4\">e.g. \u591a\u5c11\u7897 duo shao wan 'how many bowls' =def:container={bowl|\u7897:</td></tr><tr><td>quantity={.Ques.}}</td><td/><td/></tr><tr><td>Other rules,</td><td/><td/></tr><tr><td colspan=\"4\">(E) while {\u7b2c di,\uf9b6\u62dc li bai,\u661f\u671f xing qi,\u897f\u5143 xi yuan,\u6c11\u570b ming quo}+\u5e7e ji</td></tr><tr><td colspan=\"4\">or \u5e7e ji + {\u9ede(\u9418)dain(zhong),\u6708 yue,\u865f hao,\u6b72 sui,\u9031\uf98e zhou nian,\uf98e nian}</td></tr><tr><td colspan=\"4\">(18) The sense representation of interrogative determinatives (ID)</td></tr><tr><td colspan=\"2\">\uf9fd\u9ebc she me /\u4f55 he / \u5565 sha 'what'</td><td colspan=\"2\">def:formal={.Ques.}</td></tr><tr><td>\u591a\u5c11 duo shao 'how many'</td><td/><td colspan=\"2\">def:quantity={.Ques.}</td></tr><tr><td>\u5e7e ji 'how many'or 'what'</td><td/><td colspan=\"2\">def:quantity={.Ques.}or</td></tr><tr><td>def:ordinal={.Ques.}</td><td/><td/></tr><tr><td>\u54ea na 'which'or 'where'</td><td/><td colspan=\"2\">def:quantifier={.Ques.}or</td></tr><tr><td>def:location={.Ques.}</td><td/><td/></tr><tr><td>(19) (A) if head !=semantic role</td><td/><td>then</td><td>def: {head:ID}</td></tr><tr><td/><td/><td/><td>}}.</td></tr><tr><td>Similarily,</td><td/><td/></tr><tr><td colspan=\"4\">e.g. \u5e7e\u4eba ji ren 'how many people' =def: {human|\u4eba:quantity={.Ques.}}</td></tr><tr><td>(B) if head=semantic role</td><td>then</td><td colspan=\"2\">def: role={.Ques.}</td></tr><tr><td colspan=\"3\">e.g. (C) if head=DM(dummy measure words) then</td><td>def: ID</td></tr></table>",
"type_str": "table",
"html": null
},
"TABREF10": {
"text": "",
"num": null,
"content": "<table><tr><td>Rules of disambiguation</td><td colspan=\"2\">Examples & E-HowNet representation</td></tr><tr><td>adjective 1:</td><td/><td/></tr><tr><td>Case 1: if \uf9fd\u9ebc/\u4f55/\u5565+semantic role</td><td colspan=\"2\">\uf9fd\u9ebc\u6642\u9593 she me shi jian 'what time' def: time={.Ques.}</td></tr><tr><td>then \uf9fd\u9ebc/\u4f55/\u5565=def: role={.Ques.}</td><td>\u5565\u50f9\u9322 sha jia qian 'what price'</td><td>def:cost={.Ques.}</td></tr><tr><td>(\uf9fd\u9ebc ask the value of the semantic role.)</td><td>\u4f55\u5730 he di 'what place'</td><td>def:location={.Ques.}</td></tr><tr><td>adjective 2:</td><td>\u4f55\u4eba he ren 'what person'</td><td/></tr><tr><td>Case 2: if \uf9fd\u9ebc/\u4f55/\u5565+entity (nominalized</td><td>def: {human|\u4eba: formal={.Ques.}}</td><td/></tr><tr><td>verbs are included ) then \uf9fd\u9ebc/\u4f55/\u5565=def: {entity:formal={.Ques.}} (\uf9fd\u9ebc ask the type/restriction of a frame element/ participant role.)</td><td colspan=\"2\">\u5565\u8b8a\u5316 sha bian hua 'what change' def: {change|\u8b8a\u5316:formal={.Ques.}} \uf9fd\u9ebc\uf967\u540c she me bu tong 'what difference'</td></tr><tr><td/><td colspan=\"2\">def: {difference|\uf967\u540c:formal={.Ques.}}</td></tr><tr><td>pronoun 1: verb+\uf9fd\u9ebc/\u4f55/\u5565</td><td>\u5403\uf9fd\u9ebc chi she me 'eat what'</td><td/></tr><tr><td>Case 3: if \uf9fd\u9ebc use as an interrogative pronoun then \uf9fd\u9ebc/\u4f55/\u5565 =def:{event:participant={.Ques.}}</td><td>def: {eat|\u5403: patient={.Ques.}} \uf96f\u5565 shuo sha 'talk what'</td><td/></tr><tr><td/><td>def: {speak|\uf96f:content={.Ques.}}</td><td/></tr><tr><td>pronoun 2: verb+\uf9fd\u9ebc/\u4f55/\u5565 or \uf9fd\u9ebc/\u5565+verb</td><td colspan=\"2\">\u62ff\uf9fd\u9ebc\u90fd\u53ef na she me dou ke 'It's OK to get anything'</td></tr><tr><td>Case 4: if \uf9fd\u9ebc use as an indefinite pronoun</td><td colspan=\"2\">def: {hold|\u62ff:patient={entity: quantity={all}}}</td></tr><tr><td>Then \uf9fd\u9ebc/\u4f55/\u5565={event:participant={x}} or{event:participant={entity: quantity={all}}}</td><td colspan=\"2\">\uf9fd\u9ebc\u4e5f\uf967\u6015 she me ye bu pa 'Be afraid of nothing' def: {.not.fear|\u5bb3\u6015:cause={entity: quantity={all}}}</td></tr></table>",
"type_str": "table",
"html": null
}
}
}
} |