File size: 106,571 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 |
{
"paper_id": "2020",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T02:10:11.814725Z"
},
"title": "Style Analysis of Argumentative Texts by Mining Rhetorical Devices",
"authors": [
{
"first": "Khalid",
"middle": [],
"last": "Al-Khatib",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Bauhaus-Universit\u00e4t Weimar",
"location": {
"settlement": "Weimar",
"country": "Germany"
}
},
"email": ""
},
{
"first": "Viorel",
"middle": [],
"last": "Morari",
"suffix": "",
"affiliation": {},
"email": "viorel.morari@averbis.com"
},
{
"first": "Benno",
"middle": [],
"last": "Stein",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Bauhaus-Universit\u00e4t Weimar",
"location": {
"settlement": "Weimar",
"country": "Germany"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "Using the appropriate style is key for writing a high-quality text. Reliable computational style analysis is hence essential for the automation of nearly all kinds of text synthesis tasks. Research on style analysis focuses on recognition problems such as authorship identification; the respective technology (e.g., n-gram distribution divergence quantification) showed to be effective for discrimination, but inappropriate for text synthesis since the \"essence of a style\" remains implicit. This paper contributes right here: it studies the automatic analysis of style at the knowledge-level based on rhetorical devices. To this end, we developed and evaluated a grammar-based approach for identifying 26 syntax-based devices. Then, we employed that approach to distinguish various patterns of style in selected sets of argumentative articles and presidential debates. The patterns reveal several insights into the style used there, while being adequate for integration in text synthesis systems.",
"pdf_parse": {
"paper_id": "2020",
"_pdf_hash": "",
"abstract": [
{
"text": "Using the appropriate style is key for writing a high-quality text. Reliable computational style analysis is hence essential for the automation of nearly all kinds of text synthesis tasks. Research on style analysis focuses on recognition problems such as authorship identification; the respective technology (e.g., n-gram distribution divergence quantification) showed to be effective for discrimination, but inappropriate for text synthesis since the \"essence of a style\" remains implicit. This paper contributes right here: it studies the automatic analysis of style at the knowledge-level based on rhetorical devices. To this end, we developed and evaluated a grammar-based approach for identifying 26 syntax-based devices. Then, we employed that approach to distinguish various patterns of style in selected sets of argumentative articles and presidential debates. The patterns reveal several insights into the style used there, while being adequate for integration in text synthesis systems.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "The decision for an adequate writing style plays a crucial role for an author who wants to achieve a particular goal, such as persuading the readers (Burton, 2007) . \"Style\" is an elusive concept which covers a wide range of techniques an author can follow, including justifying a conclusion by anecdotal evidence, using regular repetition of the same phrase, or raising questions and then answering them. In the literature on the subject, these techniques are called rhetorical devices (Johnson, 2016) .",
"cite_spans": [
{
"start": 149,
"end": 163,
"text": "(Burton, 2007)",
"ref_id": "BIBREF2"
},
{
"start": 487,
"end": 502,
"text": "(Johnson, 2016)",
"ref_id": "BIBREF17"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The automatic analysis of style has been addressed mostly by developing a set of style features (aka style indicators) such as the percentage of function words (Ganjigunte Ashok et al., 2013; Bergsma et al., 2012) . Those features have proven to be effective in various analysis tasks, such as genre classification and author recognition. However, they are not appropriate for typical text synthesis and writing assistance tasks, since they cannot reveal the \"essence of a style\" in an explicit and describable manner.",
"cite_spans": [
{
"start": 160,
"end": 191,
"text": "(Ganjigunte Ashok et al., 2013;",
"ref_id": "BIBREF9"
},
{
"start": 192,
"end": 213,
"text": "Bergsma et al., 2012)",
"ref_id": "BIBREF1"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "By contrast, analyzing the writing style based on rhetorical devices provides a mechanism to describe where, what, and how specific techniques are used. This kind of analysis is not only important for exploring content in social science (Niculae and Danescu-Niculescu-Mizil, 2014) , but it can also serve text synthesis systems by improving the quality of automatically generated texts (Hu et al., 2017) . Moreover, it can form the backbone of style suggestion tools. For example, when writing a text for which the desired specification (e.g., the genre) is given, adequate style techniques can be suggested to improve the text quality. In such a manner, new writers can learn to improve their texts and approach the quality of masterpieces written by top writers. Figure 1 illustrates the described connections.",
"cite_spans": [
{
"start": 237,
"end": 280,
"text": "(Niculae and Danescu-Niculescu-Mizil, 2014)",
"ref_id": "BIBREF25"
},
{
"start": 386,
"end": 403,
"text": "(Hu et al., 2017)",
"ref_id": "BIBREF15"
}
],
"ref_spans": [
{
"start": 765,
"end": 773,
"text": "Figure 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Rhetoric has been the subject of investigation amongst scholars since the time of ancient Greece. Meanwhile, a considerable number of rhetorical devices were developed and discussed in the literature. The most well-known collected lists of devices contain more than 500 devices (Lawrence et al., 2017) . Though various of them, such as irony and sarcasm, is hard to be computationally identified (Java, 2015) , there is still a sufficiently large portion of popular and-for our purpose-highly useful devices whose identification can be tackled with the current state of the art. Basically, rhetorical devices can be categorized according to different principles, where an important one is a linguistic level (lexical, syntactic, semantic, and pragmatic). For the time being, we will deal with syntax-based devices. Against the above background, this paper addresses three research questions. (1) How to identify syntax-based rhetorical devices in a text? (2) What are the most common patterns of using these devices?",
"cite_spans": [
{
"start": 278,
"end": 301,
"text": "(Lawrence et al., 2017)",
"ref_id": "BIBREF19"
},
{
"start": 396,
"end": 408,
"text": "(Java, 2015)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "(3) To which degree differ these patterns across different monological and dialogical argumentative texts? Within and across the texts' genres, topics, and authors? And across different opponent debaters?",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "To answer these questions, we develop a grammar-based approach for the identification of 26 rhetorical devices. The grammars are built on top of the outputs of a probabilistic context-free grammar parser, PCFG. For evaluation purpose, we create a corpus of 1718 texts which are labelled for rhetorical devices. The evaluation results show that our approach is able to identify the devices with an average of 0.70 in terms of F 1 . Based on the developed approach, we quantify and discuss the usage of devices in monological texts within and across different genres, topics, and authors using a subset of the New York Times annotated corpus (Sandhaus, 2008) . We also analyze the devices usage patterns in dialogical texts using a set of presidential debates from the American presidency project (Woolley and Gerhard, 2017) .",
"cite_spans": [
{
"start": 640,
"end": 656,
"text": "(Sandhaus, 2008)",
"ref_id": "BIBREF28"
},
{
"start": 795,
"end": 822,
"text": "(Woolley and Gerhard, 2017)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We consider the gained qualitative and quantitative insights about the usage of rhetorical devices as step forward to a new generation of semi-automated argumentative text generation and writing tools. All developed resources in this paper are made publicly available at www.webis.de",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Recently, Investigating rhetorical devices for style analysis has been considered in computational linguistics. Various devices at the semantic and pragmatic levels have been addressed singly such as irony (e.g., (C. Wallace et al., 2014) ), sarcasm (e.g., (Ghosh et al., 2015) ), evidence (e.g., (Rinott et al., 2015) ), and means of persuasion (e.g., (Duthie et al., 2016) ). In a notable work, Strommer (2011) work on identifying 'epanaphora'. They try to distinguish between accidental and intentional use of this device.",
"cite_spans": [
{
"start": 213,
"end": 238,
"text": "(C. Wallace et al., 2014)",
"ref_id": "BIBREF3"
},
{
"start": 257,
"end": 277,
"text": "(Ghosh et al., 2015)",
"ref_id": "BIBREF11"
},
{
"start": 297,
"end": 318,
"text": "(Rinott et al., 2015)",
"ref_id": "BIBREF27"
},
{
"start": 353,
"end": 374,
"text": "(Duthie et al., 2016)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Other studies target identifying a mix of syntax, and semantic devices. Gawryjo\u0142ek et al. (2009) addressed four rhetorical devices: 'anaphora', 'isocolon', 'epizeuxis', and 'oxymorons'. These devices were utilized to recognize the author of a set of documents. Java (2015) identified the four devices mentioned above in addition to nine new devices belonging to parallelism, repetition, and trope. The primary purpose of that work is to use the presence of a rhetorical device as a feature in machine learning models for authorship attribution. Since the authors consider syntax-based devices, we already considered five of their devices in our study. Regarding argumentaion, Lawrence et al. (2017) analyzed eight devices, six belong to the syntax and lexical levels, and two to the trope (i.e., semantic or pragmatic). Mainly, a pilot study was conducted to study the relation between argumentation structure and the identified devices.",
"cite_spans": [
{
"start": 72,
"end": 96,
"text": "Gawryjo\u0142ek et al. (2009)",
"ref_id": "BIBREF10"
},
{
"start": 676,
"end": 698,
"text": "Lawrence et al. (2017)",
"ref_id": "BIBREF19"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Few resources for rhetorical devices are publicly free. Up to our knowledge, the code of the previous studies is not available anywhere on the web. Hence, researchers often have to write a new piece of code every time they need to analyze style based on rhetorical devices. This paper resolves this problem considerably by providing a tool for identifying 26 different rhetorical devices. Our developed resources, including the code, will be made freely available.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "PCFG outputs have been employed for different tasks including response generation in dialogue (Yuan et al., 2015) , multiword expression identification (Green et al., 2011) , and the task at hand: identifying rhetorical devices (Gawryjo\u0142ek et al., 2009; Java, 2015) . However, we develop a set of original heuristic rules that map the devices' definitions to PCFG grammars. As far as we know, many devices from the 26 we identified have not been considered in any other study.",
"cite_spans": [
{
"start": 94,
"end": 113,
"text": "(Yuan et al., 2015)",
"ref_id": "BIBREF33"
},
{
"start": 152,
"end": 172,
"text": "(Green et al., 2011)",
"ref_id": "BIBREF13"
},
{
"start": 228,
"end": 253,
"text": "(Gawryjo\u0142ek et al., 2009;",
"ref_id": "BIBREF10"
},
{
"start": 254,
"end": 265,
"text": "Java, 2015)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Writing style analysis has been studied widely. The authorship recognition has been tackled in a large number of papers (e.g., (Sundararajan and Woodard, 2018) ). Besides, quality assessment research has involved applying several style analysis features (e.g., (Ganjigunte Ashok et al., 2013) ). In comparison to our analysis, we conducted a controlled analysis using 'matching' technique and we covered various aspects of monological and dialogical texts such as genre, topic, author, and debate opponent.",
"cite_spans": [
{
"start": 127,
"end": 159,
"text": "(Sundararajan and Woodard, 2018)",
"ref_id": "BIBREF30"
},
{
"start": 261,
"end": 292,
"text": "(Ganjigunte Ashok et al., 2013)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Rhetorical devices are the techniques of using the language to produce an effect on the target audience or readers (McKay and McKay, 2010) . For example, repeating particular phrases can produce effects such as emphasizing a certain argument, or evoking a specific emotion (Corbett, 1990) .",
"cite_spans": [
{
"start": 115,
"end": 138,
"text": "(McKay and McKay, 2010)",
"ref_id": "BIBREF23"
},
{
"start": 273,
"end": 288,
"text": "(Corbett, 1990)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Identification of Rhetorical Devices",
"sec_num": "3"
},
{
"text": "This paper targets syntax-based rhetorical devices. Particularly, we aim to identify 26 devices belonging to two main categories: (1) figurative syntax, which is referred as schemes in literature, and (2) ordinary syntax, which concerns the rules of well-formed structuring texts. The effect of the first attributed to using an artful deviation from the ordinary arrangement of words, while the effect of the second is coming from using a specific arrangement of words among other arrangements.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Identification of Rhetorical Devices",
"sec_num": "3"
},
{
"text": "In the next subsections, we detail the figurative and ordinary syntax devices and describe our approach for identifying them.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Identification of Rhetorical Devices",
"sec_num": "3"
},
{
"text": "Figurative devices center on arranging words artfully (Burton, 2007) . They are divided into four types: balance, inversion, omission and repetition.",
"cite_spans": [
{
"start": 54,
"end": 68,
"text": "(Burton, 2007)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Figurative Syntax Devices",
"sec_num": "3.1"
},
{
"text": "\u2022 The balance devices involve arranging the rhythm of thoughts. Hence, they can produce a sense of equivalence among the proposed ideas, or emphasize ideas' differences. For example, we can notice the contrast between ideas in the famous quote of Neil Armstrong: \"That's one small step for man, one giant leap for mankind\".",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Figurative Syntax Devices",
"sec_num": "3.1"
},
{
"text": "\u2022 The inversion devices concern changing the order of words, either to stress some ideas or to avoid the monotonous flow of a sentence. For example, \"Everybody's got troubles\" could be reordered to \"Troubles, everybody's got.\".",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Figurative Syntax Devices",
"sec_num": "3.1"
},
{
"text": "\u2022 The omission devices deal with removing words that readers can reveal intuitively. They are often used to imply unfinished thoughts or to keep a fast rhythm, such as: \"He came, he saw, he conquered.\".",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Figurative Syntax Devices",
"sec_num": "3.1"
},
{
"text": "\u2022 The repetition is the most frequent, and arguably, the most powerful. According to Aristotle, repetition is the key to a persuasive speech (Fahnestock, 2003) . Typically, repetition devices aim at influencing the emotional state of the reader by emphasizing or implicating a specific idea (Burton, 2007; Corbett, 1990 ). An example which illustrates the emotional impact of repetitions is the famous line from King Lear written by Shakespeare: \"Never, never, never, never, never.\" (M\u00fcller, 2006) . Table 1 shows an overview of the identified figurative devices in our work. The overview covers a definition, a formalization, and an example for each device belongs to balance, omission, or repetition 1 . Our formalization is grounded on the devices' definitions which are taken from a set of reliable sources such as 'Silva Rhetoricae': a comprehensive source for rhetoric on the web (Burton, 2007) .",
"cite_spans": [
{
"start": 141,
"end": 159,
"text": "(Fahnestock, 2003)",
"ref_id": "BIBREF8"
},
{
"start": 291,
"end": 305,
"text": "(Burton, 2007;",
"ref_id": "BIBREF2"
},
{
"start": 306,
"end": 319,
"text": "Corbett, 1990",
"ref_id": "BIBREF4"
},
{
"start": 483,
"end": 497,
"text": "(M\u00fcller, 2006)",
"ref_id": "BIBREF24"
},
{
"start": 886,
"end": 900,
"text": "(Burton, 2007)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [
{
"start": 500,
"end": 507,
"text": "Table 1",
"ref_id": null
}
],
"eq_spans": [],
"section": "Figurative Syntax Devices",
"sec_num": "3.1"
},
{
"text": "The formalization elements are: 'Cl' for clause, 'Phr' for phrase, 'W' for word, 'N' for noun, 'Vb' for verb, 'CC' for conjunction, 'COMMA' for comma, . . . for arbitrary intervening material, [. . . ] for word boundaries, {. . . } for phrase or clause boundaries, a = b for identity , and a = b for nonidentity. The elements of formalization are adopted from .",
"cite_spans": [
{
"start": 193,
"end": 201,
"text": "[. . . ]",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Figurative Syntax Devices",
"sec_num": "3.1"
},
{
"text": "Notice that we essentially concentrate on identifying the devices at the sentence-level, or across consecutive sentences. Besides, some rhetorical devices, according to their definitions, might overlap with (C6) Unless-cond.: Restricted version of if-conditional (its intrinsic meaning is narrowed down to \"Q in the case other than P\"). <... unless ...> You can't go on vacation unless you save some money. other devices in some special cases. This overlap is rare and partial. Nevertheless, we consider minimizing the possible overlaps among devices as much as possible in our formalization.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Figurative Syntax Devices",
"sec_num": "3.1"
},
{
"text": "From the ordinary syntax devices, we select conditionals, comparatives and superlatives, and passive voice. This selection is based on the impact of these devices on the readers (Martinet, 1960) .",
"cite_spans": [
{
"start": 178,
"end": 194,
"text": "(Martinet, 1960)",
"ref_id": "BIBREF22"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Ordinary Syntax Devices",
"sec_num": "3.2"
},
{
"text": "\u2022 The conditional devices entail the causality aspect of the language, and causality, in turn, could imply explaining an event. But it can also be used to argue about positive/negative consequences of a specific action such as \"If we were elected him, we would not have achievements\".",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Ordinary Syntax Devices",
"sec_num": "3.2"
},
{
"text": "\u2022 The comparatives and superlatives devices might be used to emphasize the superiority of an entity or idea, e.g., \"I will be the greatest jobs president that God ever created\".",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Ordinary Syntax Devices",
"sec_num": "3.2"
},
{
"text": "\u2022 The passive voice might be used to hide the subject of a negative action, or to stress the importance of an event, e.g., \"many mistakes were made, but the future will be great\". Table 1 provides an overview of the conditionals, comparatives and superlatives, and passive voice devices. The overview is analog to the one of the figurative category. The formalization is based on definitions from the same set of resources used in the figurative category. The elements of formalization are taken from The Penn Treebank POS Tag Set (Marcus et al., 1993) . 60, tricolon (60) and tetracolon (60). \u2020 we applied the 26 classifiers on the 'other' instances. If any of them labels an instance with its device, we consider the instance as wrongly classified. Table 3 : The precision, recall, and F 1 -score for identifying the rhetorical devices by category.",
"cite_spans": [
{
"start": 531,
"end": 552,
"text": "(Marcus et al., 1993)",
"ref_id": "BIBREF21"
}
],
"ref_spans": [
{
"start": 180,
"end": 187,
"text": "Table 1",
"ref_id": null
},
{
"start": 751,
"end": 758,
"text": "Table 3",
"ref_id": null
}
],
"eq_spans": [],
"section": "Ordinary Syntax Devices",
"sec_num": "3.2"
},
{
"text": "Here, we discuss the evaluation experiments of our approach for identifying the syntax-based rhetorical devices. First, we describe the newly created evaluation dataset. Then, we talk about the experimental settings and report on the obtained results. Finally, we address the limitations of our approach and perform an error analysis for its output.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Device",
"sec_num": null
},
{
"text": "Evaluation Dataset: Creating a dataset for rhetorical devices using manual annotation, even with crowd-sourcing, is extremely expensive and time consuming (Java, 2015) ; The reason behind this is the big number of devices, the potential overlaps between them, and the possibility for some devices to be spread across phrases, sentences, or even paragraphs. Thereby, we decided to follow a bunch of related research studies (e.g., (Java, 2015) ) and build the evaluation dataset as follows: We first identify a set of trustworthy sources on the web, which address the rhetorical devices, and have credibility as being developed by experts in rhetoric. Most of the selected sources are either mentioned or already used in some research studies, which speaks for their trustworthiness. From those sources, we use meta-data information (e.g., \"Example of Pysma:\") to collect a set of instances for our rhetorical devices. We found that targeting about 60 examples for each device is reasonable considering the size of content in the selected sources. We verified all the examples and ensured that there are no duplicated ones. Additionally, we accounted for the possible overlaps between the devices and minimized them adequately, i.e., all the examples for a device belong solely to this device. Unfortunately, two devices turned out to be considered only from few sources, and hence, we got less than 60 examples for them. We also collected 60 examples where none of the devices covered by our work is used. Overall, we collected 1718 examples: 1658 example distributed among the 26 devices and around 60 examples that belong to 'other'. The distribution is shown in Table 2 . This dataset, despite its relatively small size, is significantly larger than those that have been used for rhetorical devices in related work (Java, 2015) .",
"cite_spans": [
{
"start": 155,
"end": 167,
"text": "(Java, 2015)",
"ref_id": "BIBREF16"
},
{
"start": 430,
"end": 442,
"text": "(Java, 2015)",
"ref_id": "BIBREF16"
},
{
"start": 1818,
"end": 1830,
"text": "(Java, 2015)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [
{
"start": 1665,
"end": 1672,
"text": "Table 2",
"ref_id": "TABREF2"
}
],
"eq_spans": [],
"section": "Device",
"sec_num": null
},
{
"text": "Experimental Settings: The implementation of our approach was carried out using Apache Ruta TM (Rule-based Text Annotation) (Kluegl et al., 2016) . This tool provides a flexible language for identifying patterns in text spans intuitively. Thus, it facilitates identifying sophisticated patterns with a few lines of code. The implementation is performed on top of the outputs of Stanford Parser (Manning et al., 2014) , the version of 3.8.0. We evaluated our approach using the one-vs.-rest classification. That means we performed one classification experiment for each device; The instances of this device in the evaluation dataset is considered as the positive class, and the instances of the remaining devices as well as the 'other' as the negative class. The classifiers' effectiveness is reported in terms of precision, recall, and F 1 -score.",
"cite_spans": [
{
"start": 124,
"end": 145,
"text": "(Kluegl et al., 2016)",
"ref_id": "BIBREF18"
},
{
"start": 394,
"end": 416,
"text": "(Manning et al., 2014)",
"ref_id": "BIBREF20"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Device",
"sec_num": null
},
{
"text": "Classification Results: Table 2 shows the results of our experiments. Overall, we manage to identify the 26 devices with an average of 0.70 F 1 -score., which indicates a high effectiveness of our approach.",
"cite_spans": [],
"ref_spans": [
{
"start": 24,
"end": 31,
"text": "Table 2",
"ref_id": "TABREF2"
}
],
"eq_spans": [],
"section": "Device",
"sec_num": null
},
{
"text": "As for the \"figurative\" devices, the approach got high scores for the balance devices, including F 1score of 1.00 for 'pysma'. The 'isocolon' is the most challenging with F 1 -score of 0.68. As for the omission devices, the F 1 -scores range from 0.39 for 'asyndeton' and 0.69 for the 'hypozeugma'. These results are a bit lower than the other types. Most of the repetition devices have F 1 -score of about 0.73, except 'mesarchia' with 0.59, and 'mesodiplosis' with 0.39. Besides, \"ordinary\" devices got scores between 0.56 and 1.00. Interestingly, despite their simple syntax, comparatives and superlatives devices got the lowest scores. Table 3 shows the results of our approach regarding the six rhetorical categories that group the 26 devices. The F 1 scores range from 0.54 to 0.87. The best result is obtained for passive voice (0.87) and conditionals (0.82). Omission and repetition are the hardest to identify with 0.54 and 0.67 F 1 . Figure 2 shows an excerpt from a news editorial along with several rhetorical devices that our approach manages to identify.",
"cite_spans": [],
"ref_spans": [
{
"start": 640,
"end": 647,
"text": "Table 3",
"ref_id": null
},
{
"start": 944,
"end": 952,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Device",
"sec_num": null
},
{
"text": "Error Analysis: Despite the high effectiveness of our approach, it is subject to fail in some cases.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Device",
"sec_num": null
},
{
"text": "Concerning the \"figurative\" category, identifying the balance devices seems to be precise except for 'isocolon'. The identification of this device is based on the outputs of the syntax parser (i.e., POS tags) which are sometimes inaccurate, especially for long sentences. This has a negative impact on the precision score; for instance, \"It looks like the Libertarian candidate is racking up the percentage points in recent polls. As far as I can see the Libertarian candidate has over . . . .\". Here, the 'Libertarian candidate' makes the classifier of 'isocolon' treats it wrongly as a valid instance. For the omission devices, our approach manages to get 0.93 recall score for 'asyndeton' device, but only 0.25 for precision. We found that the abundance of commas, which we use as an indicator of the lack of conjunctions is insufficient to distinguish 'asyndeton' from other devices, especially 'enumeration'. For example, \"Old McDonald However, other experts say the current reliance on the Sabin oral live-virus vaccine has worked so well that great care should be taken before changing policies.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Device",
"sec_num": null
},
{
"text": "Passive Voice The injectable '' killed-virus '' vaccine was largely replaced by an oral vaccine made from live viruses, which is still being given to millions of American children.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Passive Voice",
"sec_num": null
},
{
"text": "The development of the new form of the Salk vaccine opens the way for it to be used in combination with other childhood vaccinations. Some health o cials, noting that it has been used in Europe and tested in the developing world, believe that it can be an e ective way to reduce immunizations and associated costs.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Passive Voice",
"sec_num": null
},
{
"text": "Dr. Frederick C. Robbins of Case Western Reserve University, chairman of the panel, said early attempts at the combined vaccine were abandoned in this country because of potency problems. Later successes with this approach in Europe, using an enhanced polio vaccine, have rekindled the idea of a combination approach, including the possibility of using both types of polio vaccines to merge their bene ts, he added.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Passive Voice",
"sec_num": null
},
{
"text": "Hypozeugma Asyndeton",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Passive Voice",
"sec_num": null
},
{
"text": "Study Under Way The Institute of Medicine, an adjunct of the National Academy of Sciences, is studying polio policy and is expected to submit recommendations to the Federal health authorities by April. At a recent public meeting in Washington, the committee heard suggestions for bringing back the inactivated-virus vaccine by combining it with the diphtheria, tetanus and pertussis shots. Enumeration Hypozeugma Hypozeugma Enumeration Passive Voice Asyndeton Figure 2 : An excerpt from a NYT news editorial. The rhetorical devices in each sentence are identified using our approach.",
"cite_spans": [],
"ref_spans": [
{
"start": 460,
"end": 468,
"text": "Figure 2",
"ref_id": null
}
],
"eq_spans": [],
"section": "Asyndeton Passive Voice",
"sec_num": null
},
{
"text": "had a pig, a dog, a cow and a horse.\" is identified as 'asyndeton', while it is actually 'enumeration'. As regards repetition, two devices there got low scores: the 'mesarchia' and 'mesodiplosis'. These devices have the least number of instances in our evaluation dataset. We also observed that our heuristic rules for defining the beginning and middle of sentences are the reason for some errors. For the \"ordinary\" category, the approach has promising results. However, the scores for the 'comparatives and superlatives' are moderate. Observing the errors there, we found that the main reason is again the inaccurate POS tags. For example, in the sentence 'the airport is further than the train station.', 'further' is tagged as comparative adverb instead of comparative adjective.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Asyndeton Passive Voice",
"sec_num": null
},
{
"text": "The 'other' class got a low F 1 -score. In addition to its restrictive way of evaluation that we followed, this score indicates that some devices' classifiers tend to have a lot of false positives.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Asyndeton Passive Voice",
"sec_num": null
},
{
"text": "To have a better idea regarding the effectiveness of our approach, we performed a manual inspection of the classifiers' outputs on a set of ten newspaper articles. We found that some devices such as 'isocolon' and 'asyndeton' indeed have many false positives. Besides, we found that the classifiers make more mistakes with very long sentences.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Asyndeton Passive Voice",
"sec_num": null
},
{
"text": "We rely on our identification approach to analyze the usage patterns of rhetorical devices in argumentative newspaper articles and presidential debates. First, we describe the acquisition and sampling of the analysis datasets. Then, we discuss the distribution of rhetorical devices there along with different article and debate aspects. The computed distributions illustrate various patterns of rhetorical devices and lead to several interesting insights.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Analysis of Rhetorical Devices",
"sec_num": "4"
},
{
"text": "Analysis Datasets: To conduct insightful analysis, we constructed two datasets for newspaper articles and presidential debates.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Analysis of Rhetorical Devices",
"sec_num": "4"
},
{
"text": "(1) Newspaper dataset: to construct this dataset, we used the NYT annotated corpus (Sandhaus, 2008) . The corpus comprises more than 1.8 million high-quality articles written by professional writers. It comes with many types of meta-data labeled by NYT staff, including the type of material (e.g., editorial), the author name, and the topic (e.g., sport). From this corpus, we sampled three subsets, each of which represents one of the three properties of genre, topic, and author. To conduct a controlled analysis, the sampling should account for the confounding variables. For example, studying the style in articles with a specific 'topic' can be influenced by their genres and authors. Hence, we first tried to resolve this issue with the stratification method (Tripepi et al., 2010) , which turned out to be not successful; despite the large size of the corpus, we found no information about the authors of about 40% of articles. Also, the distribution of articles in texts belong to the three properties are very skewed. The corpus includes much more reviews than editorials, for example. Many articles are written for 'politics' and few for 'sport', and some authors wrote tens of articles while others wrote only one. Therefore, we tried the matching technique (de Graaf et al., 2011), where we managed successfully to sample the three subsets. To preserve the balance between the subsets, we consider three instances for each propriety, i.e., the 'topic' subset includes 114 articles belong to science, education, and art. The 'genre' subset includes 89 articles belong to biography, editorial, and review. Finally, the 'authors' subset includes 159 articles written by Martin, Lewis, and Hevesi.",
"cite_spans": [
{
"start": 83,
"end": 99,
"text": "(Sandhaus, 2008)",
"ref_id": "BIBREF28"
},
{
"start": 765,
"end": 787,
"text": "(Tripepi et al., 2010)",
"ref_id": "BIBREF31"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Analysis of Rhetorical Devices",
"sec_num": "4"
},
{
"text": "(2) Debate dataset: we acquired this dataset based on the presidential debates from the American presidency project (Woolley and Gerhard, 2017). In particular, we extracted the entire set of debates that involve Donald Trump or/and Hillary Clinton. We think that these two characters are different in many aspects such as ideology, background, experience, opinions on different topics, etc. This difference could be reflected in their styles leading to interesting patterns. We created three subsets of the dataset: 'Trump vs. Clinton,' 'Trump vs. Not-Clinton', and 'Clinton vs. Not-Trump'. In this way, we can analyze the style of the two characters, and also address the question of whether they change their styles according to the debate opponent. In total, Clinton has 226 turns in her debates with Trump, and 1216 in her debates with the other candidates. Trump, on the other side, has 342 turns in his debates with Clinton, and 778 in his debates with the rest of candidates.",
"cite_spans": [
{
"start": 516,
"end": 590,
"text": "'Trump vs. Clinton,' 'Trump vs. Not-Clinton', and 'Clinton vs. Not-Trump'.",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Analysis of Rhetorical Devices",
"sec_num": "4"
},
{
"text": "Analysis Method: Basically, we applied our Identification approach (see Section 3) to the analysis datasets. In particular, a classifier for each device is applied to the articles or debates turns, resulting in the frequency of the device there. However, since our identification approach is not perfect, it is crucial to account for its errors. Hence, we followed the method used in (Al-Khatib et al., 2017) : for the frequency n of a rhetorical device rd in an instance i in a dataset. We computed a confidence interval for n, where the lowerbound = n * precision(rd), and the upperbound = n/recall(rd). Ultimately, The mean of the upper and lower bounds is the new frequency, which is normalized by the number of sentences in the articles/debate-turns belong to i. Accordingly, we computed the distributions of rhetorical devices in the analysis datasets and their subsets. The chi-squared test with 0.01 significant level is used to check whether the difference in the usage of rhetorical devices in the datasets and across their instances is significant, and the Cramer's V test is used to measure the effect-size of the distributions' difference.",
"cite_spans": [
{
"start": 384,
"end": 408,
"text": "(Al-Khatib et al., 2017)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Analysis of Rhetorical Devices",
"sec_num": "4"
},
{
"text": "Analysis Results: Figure 3 shows the distribution of rhetorical devices among the three authors (a), the three genres (b), the three topics (c), and the debate subsets (d). As expected, the style in newspaper articles (monologue) are significantly different than in debates (dialogue). Some analysis results for each of the datasets are as follows.",
"cite_spans": [],
"ref_spans": [
{
"start": 18,
"end": 26,
"text": "Figure 3",
"ref_id": "FIGREF2"
}
],
"eq_spans": [],
"section": "Analysis of Rhetorical Devices",
"sec_num": "4"
},
{
"text": "(1) Newspaper dataset: In addition to the significant difference among the three properties under studied, the results show a significant difference among the three authors. For example, Lewis and Hevesi use more repetition than Martin. Also, Lewis barely considers conditionals, in contrast to the other two authors. The results also show a significant difference between 'biography' and 'editorial' as well as 'editorial' and 'review', but not between 'review' and 'biography'. The reason might be that the articles in these two genres are written mainly to describe an entity. Interestingly, there is no significant difference between the three topics. Overall, our analysis suggests that the \"style\" identified by syntax-based rhetorical devices is primarily influenced by the 'author', and 'genre', while 'topic' has the least impact.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Analysis of Rhetorical Devices",
"sec_num": "4"
},
{
"text": "(2) Debate dataset: Interestingly, the results show that Clinton is more fond of 'comparatives' and 'passive voice' than Trump, which actually contradicts a widespread assumption (Gingell, 2016; Raskin, 2016) . However, our findings are mainly related to the debate genre. The style could be different in speeches, for example. We also found that Clinton uses 'asyndeton' more often than Trump. Since this device is very effective for making the turns easier to grasp, our finding this time is in line with (Raskin, 2016) , where they find that Clinton's language is 13% clearer and more direct than Trump's. The results indicate a significant difference between Clinton and Trump styles. More interestingly, while Clinton's style is significantly different when she debates with Trump than when she debates with the rest, Trump's style has no significant difference between his debates with Clinton and his debates with the rest. Apparently, unlike Clinton, Trump does not change his style depending on the opponent.",
"cite_spans": [
{
"start": 179,
"end": 194,
"text": "(Gingell, 2016;",
"ref_id": "BIBREF12"
},
{
"start": 195,
"end": 208,
"text": "Raskin, 2016)",
"ref_id": "BIBREF26"
},
{
"start": 507,
"end": 521,
"text": "(Raskin, 2016)",
"ref_id": "BIBREF26"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Analysis of Rhetorical Devices",
"sec_num": "4"
},
{
"text": "Writing style analysis has become a mature discipline, but it is mostly tackled from the recognition perspective. I.e., it can give strong classification results that, because of their intrinsic nature, cannot be transferred to constrained text generation or computational writing assistance. We address this shortcoming by proposing an approach for the explicit encoding and identification of rhetorical devices. In carefully designed experiments, we study the usage of these devices in different argumentative articles and presidential debates. The distributions show different patterns of style among three text's properties and provide new insights regarding style usage within the studied topics. The achieved F 1 classification performance (0.70) can be considered as very good for concrete multi-class classification setting; it shows that the applied approach has the potential to find its way into real-world argumentative text synthesis tools. We plan in the future to improve our grammars to minimize mistakes and increase the number of devices considering the inversion type of figurative syntax devices.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion and Future Work",
"sec_num": "5"
},
{
"text": "The inversion is left to future work due to its complexity.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Patterns of Argumentation Strategies across Topics",
"authors": [
{
"first": "Khalid",
"middle": [],
"last": "Al-Khatib",
"suffix": ""
},
{
"first": "Henning",
"middle": [],
"last": "Wachsmuth",
"suffix": ""
},
{
"first": "Matthias",
"middle": [],
"last": "Hagen",
"suffix": ""
},
{
"first": "Benno",
"middle": [],
"last": "Stein",
"suffix": ""
}
],
"year": 2017,
"venue": "Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing",
"volume": "17",
"issue": "",
"pages": "1362--1368",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Khalid Al-Khatib, Henning Wachsmuth, Matthias Hagen, and Benno Stein. 2017. Patterns of Argumentation Strategies across Topics. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing (EMNLP 17), pages 1362-1368. Association for Computational Linguistics.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Stylometric analysis of scientific articles",
"authors": [
{
"first": "Shane",
"middle": [],
"last": "Bergsma",
"suffix": ""
},
{
"first": "Matt",
"middle": [],
"last": "Post",
"suffix": ""
},
{
"first": "David",
"middle": [],
"last": "Yarowsky",
"suffix": ""
}
],
"year": 2012,
"venue": "Proceedings of the 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT 12)",
"volume": "",
"issue": "",
"pages": "327--337",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Shane Bergsma, Matt Post, and David Yarowsky. 2012. Stylometric analysis of scientific articles. In Proceed- ings of the 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT 12), pages 327-337. Association for Computational Linguistics.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "The forest of rhetoric (silva rhetoricae). Accessed on 16",
"authors": [
{
"first": "G",
"middle": [],
"last": "Burton",
"suffix": ""
}
],
"year": 2007,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "G. Burton. 2007. The forest of rhetoric (silva rhetoricae). Accessed on 16.08.2017.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Humans Require Context to Infer Ironic Intent (so Computers Probably do, too)",
"authors": [
{
"first": "Byron",
"middle": [
"C"
],
"last": "Wallace",
"suffix": ""
},
{
"first": "Laura",
"middle": [],
"last": "Do Kook Choe",
"suffix": ""
},
{
"first": "Eugene",
"middle": [],
"last": "Kertz",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Charniak",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of the 2014 Annual Meeting on Association for Computational Linguistics",
"volume": "1",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Byron C. Wallace, Do Kook Choe, Laura Kertz, and Eugene Charniak. 2014. Humans Require Context to Infer Ironic Intent (so Computers Probably do, too). In Proceedings of the 2014 Annual Meeting on Association for Computational Linguistics (ACL 14) -Volume 1.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Classical rhetoric for the modern student",
"authors": [
{
"first": "E",
"middle": [
"P J"
],
"last": "Corbett",
"suffix": ""
}
],
"year": 1990,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "E. P. J. Corbett. 1990. Classical rhetoric for the modern student. USA: Oxford University Press, 3 edition.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Matching, an Appealing Method to Avoid Confounding?",
"authors": [
{
"first": "M",
"middle": [],
"last": "A De Graaf",
"suffix": ""
},
{
"first": "K J",
"middle": [],
"last": "Jager",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Zoccali",
"suffix": ""
},
{
"first": "F W",
"middle": [],
"last": "Dekker",
"suffix": ""
}
],
"year": 2011,
"venue": "Nephron Clin Pract",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "M A de Graaf, K J Jager, C Zoccali, and F W Dekker. 2011. Matching, an Appealing Method to Avoid Confound- ing? Nephron Clin Pract.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Conditionals: A Comprehensive Empirical Analysis. Beitrage Zur Alexander-Von-Humboldt-Forschung",
"authors": [
{
"first": "R",
"middle": [],
"last": "Declerck",
"suffix": ""
},
{
"first": "S",
"middle": [],
"last": "Reed",
"suffix": ""
}
],
"year": 2001,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "R. Declerck and S. Reed. 2001. Conditionals: A Comprehensive Empirical Analysis. Beitrage Zur Alexander- Von-Humboldt-Forschung. Mouton de Gruyter.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Mining Ethos in Political Debate",
"authors": [
{
"first": "Rory",
"middle": [],
"last": "Duthie",
"suffix": ""
},
{
"first": "Katarzyna",
"middle": [],
"last": "Budzynska",
"suffix": ""
},
{
"first": "Chris",
"middle": [],
"last": "Reed",
"suffix": ""
}
],
"year": 2016,
"venue": "6th International Conference on Computational Models of Argument (COMMA 16)",
"volume": "",
"issue": "",
"pages": "299--310",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rory Duthie, Katarzyna Budzynska, and Chris Reed. 2016. Mining Ethos in Political Debate. In 6th International Conference on Computational Models of Argument (COMMA 16), pages 299-310.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Verbal and Visual Parallelism",
"authors": [
{
"first": "Jeanne",
"middle": [],
"last": "Fahnestock",
"suffix": ""
}
],
"year": 2003,
"venue": "Written Communication",
"volume": "20",
"issue": "2",
"pages": "123--152",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jeanne Fahnestock. 2003. Verbal and Visual Parallelism. Written Communication, 20(2):123-152.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Success with Style: Using Writing Style to Predict the Success of Novels",
"authors": [
{
"first": "Song",
"middle": [],
"last": "Vikas Ganjigunte Ashok",
"suffix": ""
},
{
"first": "Yejin",
"middle": [],
"last": "Feng",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Choi",
"suffix": ""
}
],
"year": 2013,
"venue": "Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "1753--1764",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Vikas Ganjigunte Ashok, Song Feng, and Yejin Choi. 2013. Success with Style: Using Writing Style to Predict the Success of Novels. In Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing, pages 1753-1764. Association for Computational Linguistics.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "An annotation tool for automatically detecting rhetorical figures",
"authors": [
{
"first": "J",
"middle": [],
"last": "Jakub",
"suffix": ""
},
{
"first": "Randy",
"middle": [
"A"
],
"last": "Gawryjo\u0142ek",
"suffix": ""
},
{
"first": "Chrysanne",
"middle": [],
"last": "Harris",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Dimarco",
"suffix": ""
}
],
"year": 2009,
"venue": "Proceedings, CMNAIX (Computational Models of Natural Argument)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jakub J. Gawryjo\u0142ek, Randy A. Harris, and Chrysanne DiMarco. 2009. An annotation tool for automatically detecting rhetorical figures. In Proceedings, CMNAIX (Computational Models of Natural Argument).",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Sarcastic or Not: Word Embeddings to Predict the Literal or Sarcastic Meaning of Words",
"authors": [
{
"first": "Debanjan",
"middle": [],
"last": "Ghosh",
"suffix": ""
},
{
"first": "Weiwei",
"middle": [],
"last": "Guo",
"suffix": ""
},
{
"first": "Smaranda",
"middle": [],
"last": "Muresan",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing",
"volume": "15",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Debanjan Ghosh, Weiwei Guo, and Smaranda Muresan. 2015. Sarcastic or Not: Word Embeddings to Predict the Literal or Sarcastic Meaning of Words. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP 15).",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Why superlatives are the absolute worst (unless you're Donald Trump",
"authors": [
{
"first": "James",
"middle": [],
"last": "Gingell",
"suffix": ""
}
],
"year": 2016,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "James Gingell. 2016. Why superlatives are the absolute worst (unless you're Donald Trump). https://www. theguardian.com/media/mind-your-language/2016/apr/15/. visited on 24.10.17.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Multiword Expression Identification with Tree Substitution Grammars: A Parsing Tour De Force with French",
"authors": [
{
"first": "Spence",
"middle": [],
"last": "Green",
"suffix": ""
},
{
"first": "Marie-Catherine",
"middle": [],
"last": "De Marneffe",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Bauer",
"suffix": ""
},
{
"first": "Christopher",
"middle": [
"D"
],
"last": "Manning",
"suffix": ""
}
],
"year": 2011,
"venue": "Proceedings of the Conference on Empirical Methods in Natural Language Processing",
"volume": "",
"issue": "",
"pages": "725--735",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Spence Green, Marie-Catherine de Marneffe, John Bauer, and Christopher D. Manning. 2011. Multiword Expres- sion Identification with Tree Substitution Grammars: A Parsing Tour De Force with French. In Proceedings of the Conference on Empirical Methods in Natural Language Processing, EMNLP '11, pages 725-735. Associa- tion for Computational Linguistics.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Constructing a Rhetorical Figuration Ontology",
"authors": [
{
"first": "R",
"middle": [],
"last": "Harris",
"suffix": ""
},
{
"first": "C",
"middle": [],
"last": "Dimarco",
"suffix": ""
}
],
"year": 2009,
"venue": "Symposium on Persuasive Technology and Digital Behaviour Intervention",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "R. Harris and C. DiMarco. 2009. Constructing a Rhetorical Figuration Ontology. In In Symposium on Persuasive Technology and Digital Behaviour Intervention.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Toward Controlled Generation of Text",
"authors": [
{
"first": "Zhiting",
"middle": [],
"last": "Hu",
"suffix": ""
},
{
"first": "Zichao",
"middle": [],
"last": "Yang",
"suffix": ""
},
{
"first": "Xiaodan",
"middle": [],
"last": "Liang",
"suffix": ""
},
{
"first": "Ruslan",
"middle": [],
"last": "Salakhutdinov",
"suffix": ""
},
{
"first": "Eric",
"middle": [
"P"
],
"last": "Xing",
"suffix": ""
}
],
"year": 2017,
"venue": "Proceedings of the 34th International Conference on Machine Learning",
"volume": "70",
"issue": "",
"pages": "1587--1596",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Zhiting Hu, Zichao Yang, Xiaodan Liang, Ruslan Salakhutdinov, and Eric P. Xing. 2017. Toward Controlled Generation of Text. In Proceedings of the 34th International Conference on Machine Learning, volume 70 of Proceedings of Machine Learning Research, pages 1587-1596. PMLR.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Characterization of Prose by Rhetorical Structure for Machine Learning Classification",
"authors": [
{
"first": "James",
"middle": [],
"last": "Java",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "James Java. 2015. Characterization of Prose by Rhetorical Structure for Machine Learning Classification. Ph.D. thesis.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "The Alphabet of Rhetoric",
"authors": [
{
"first": "R",
"middle": [],
"last": "Johnson",
"suffix": ""
}
],
"year": 2016,
"venue": "BiblioLife",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "R. Johnson. 2016. The Alphabet of Rhetoric. BiblioLife.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Uima ruta: Rapid development of rule-based information extraction applications",
"authors": [
{
"first": "Peter",
"middle": [],
"last": "Kluegl",
"suffix": ""
},
{
"first": "Martin",
"middle": [],
"last": "Toepfer",
"suffix": ""
},
{
"first": "Philip-Daniel",
"middle": [],
"last": "Beck",
"suffix": ""
},
{
"first": "Georg",
"middle": [],
"last": "Fette",
"suffix": ""
},
{
"first": "Frank",
"middle": [],
"last": "Puppe",
"suffix": ""
}
],
"year": 2016,
"venue": "Natural Language Engineering",
"volume": "22",
"issue": "",
"pages": "1--40",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Peter Kluegl, Martin Toepfer, Philip-Daniel Beck, Georg Fette, and Frank Puppe. 2016. Uima ruta: Rapid development of rule-based information extraction applications. Natural Language Engineering, 22:1-40.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Harnessing rhetorical figures for argument mining",
"authors": [
{
"first": "Johna",
"middle": [],
"last": "Lawrence",
"suffix": ""
},
{
"first": "Jackya",
"middle": [],
"last": "Visser",
"suffix": ""
},
{
"first": "Chris",
"middle": [],
"last": "",
"suffix": ""
}
],
"year": 2017,
"venue": "Argument & Computation",
"volume": "8",
"issue": "",
"pages": "289--310",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Johna Lawrence, Jackya Visser, and Chris Reed. 2017. Harnessing rhetorical figures for argument mining. Argu- ment & Computation, 8:289-310.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "The Stanford CoreNLP Natural Language Processing Toolkit",
"authors": [
{
"first": "Christopher",
"middle": [
"D"
],
"last": "Manning",
"suffix": ""
},
{
"first": "Mihai",
"middle": [],
"last": "Surdeanu",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Bauer",
"suffix": ""
},
{
"first": "Jenny",
"middle": [],
"last": "Finkel",
"suffix": ""
},
{
"first": "Steven",
"middle": [
"J"
],
"last": "Bethard",
"suffix": ""
},
{
"first": "David",
"middle": [],
"last": "Mcclosky",
"suffix": ""
}
],
"year": 2014,
"venue": "Association for Computational Linguistics (ACL) System Demonstrations",
"volume": "",
"issue": "",
"pages": "55--60",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christopher D. Manning, Mihai Surdeanu, John Bauer, Jenny Finkel, Steven J. Bethard, and David McClosky. 2014. The Stanford CoreNLP Natural Language Processing Toolkit. In Association for Computational Linguis- tics (ACL) System Demonstrations, pages 55-60.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Building a Large Annotated Corpus of English: The Penn Treebank",
"authors": [
{
"first": "Mitchell",
"middle": [
"P"
],
"last": "Marcus",
"suffix": ""
},
{
"first": "Mary",
"middle": [
"Ann"
],
"last": "Marcinkiewicz",
"suffix": ""
},
{
"first": "Beatrice",
"middle": [],
"last": "Santorini",
"suffix": ""
}
],
"year": 1993,
"venue": "Comput. Linguist",
"volume": "19",
"issue": "2",
"pages": "313--330",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mitchell P. Marcus, Mary Ann Marcinkiewicz, and Beatrice Santorini. 1993. Building a Large Annotated Corpus of English: The Penn Treebank. Comput. Linguist., 19(2):313-330.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Elements of General Linguistics",
"authors": [
{
"first": "Andr\u00e9",
"middle": [],
"last": "Martinet",
"suffix": ""
}
],
"year": 1960,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Andr\u00e9 Martinet. 1960. Elements of General Linguistics. Faber and Faber Ltd., London.",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "Classical Rhetoric 101. Accessed on 14",
"authors": [
{
"first": "Brett",
"middle": [],
"last": "Mckay",
"suffix": ""
},
{
"first": "Kate",
"middle": [],
"last": "Mckay",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Brett McKay and Kate McKay. 2010. Classical Rhetoric 101. Accessed on 14.08.2017.",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "Style",
"authors": [
{
"first": "Wolfgang",
"middle": [
"G"
],
"last": "M\u00fcller",
"suffix": ""
}
],
"year": 2006,
"venue": "Encyclopedia of Rhetoric",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Wolfgang G. M\u00fcller. 2006. Style. In Thomas O. Sloane, editor, Encyclopedia of Rhetoric. Oxford University Press, February.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"title": "Brighter than gold: Figurative language in user generated comparisons",
"authors": [
{
"first": "Vlad",
"middle": [],
"last": "Niculae",
"suffix": ""
},
{
"first": "Cristian",
"middle": [],
"last": "Danescu-Niculescu-Mizil",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of EMNLP",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Vlad Niculae and Cristian Danescu-Niculescu-Mizil. 2014. Brighter than gold: Figurative language in user generated comparisons. In Proceedings of EMNLP, October.",
"links": null
},
"BIBREF26": {
"ref_id": "b26",
"title": "Hillary clinton's acceptance speech as seen by the algorithms. the huffington post",
"authors": [
{
"first": "Robin",
"middle": [],
"last": "Raskin",
"suffix": ""
}
],
"year": 2016,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Robin Raskin. 2016. Hillary clinton's acceptance speech as seen by the algorithms. the huffington post. https: //www.huffingtonpost.com/robin-raskin. visited on 18.11.17.",
"links": null
},
"BIBREF27": {
"ref_id": "b27",
"title": "Show Me Your Evidence -An Automatic Method for Context Dependent Evidence Detection",
"authors": [
{
"first": "Ruty",
"middle": [],
"last": "Rinott",
"suffix": ""
},
{
"first": "Lena",
"middle": [],
"last": "Dankin",
"suffix": ""
},
{
"first": "Carlos",
"middle": [
"Alzate"
],
"last": "Perez",
"suffix": ""
},
{
"first": "M",
"middle": [
"Mitesh"
],
"last": "Khapra",
"suffix": ""
},
{
"first": "Ehud",
"middle": [],
"last": "Aharoni",
"suffix": ""
},
{
"first": "Noam",
"middle": [],
"last": "Slonim",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing",
"volume": "15",
"issue": "",
"pages": "440--450",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ruty Rinott, Lena Dankin, Carlos Alzate Perez, M. Mitesh Khapra, Ehud Aharoni, and Noam Slonim. 2015. Show Me Your Evidence -An Automatic Method for Context Dependent Evidence Detection. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing (EMNLP 15), pages 440-450. Association for Computational Linguistics.",
"links": null
},
"BIBREF28": {
"ref_id": "b28",
"title": "The new york times annotated corpus ldc2008t19. dvd. Philadelphia: Linguistic Data Consortium",
"authors": [
{
"first": "Evan",
"middle": [],
"last": "Sandhaus",
"suffix": ""
}
],
"year": 2008,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Evan Sandhaus. 2008. The new york times annotated corpus ldc2008t19. dvd. Philadelphia: Linguistic Data Consortium.",
"links": null
},
"BIBREF29": {
"ref_id": "b29",
"title": "Using rhetorical figures and shallow attributes as a metric of intent in text",
"authors": [
{
"first": "Claus",
"middle": [
"W"
],
"last": "Strommer",
"suffix": ""
}
],
"year": 2011,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Claus W. Strommer. 2011. Using rhetorical figures and shallow attributes as a metric of intent in text. Ph.D. thesis, University of Waterloo, Waterloo, Ontario, Canada.",
"links": null
},
"BIBREF30": {
"ref_id": "b30",
"title": "What represents \"style\" in authorship attribution?",
"authors": [
{
"first": "Kalaivani",
"middle": [],
"last": "Sundararajan",
"suffix": ""
},
{
"first": "Damon",
"middle": [
"L"
],
"last": "Woodard",
"suffix": ""
}
],
"year": 2018,
"venue": "Proceedings of the 27th International Conference on Computational Linguistics",
"volume": "",
"issue": "",
"pages": "2814--2822",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kalaivani Sundararajan and Damon L. Woodard. 2018. What represents \"style\" in authorship attribution? In Proceedings of the 27th International Conference on Computational Linguistics, COLING 2018, Santa Fe, New Mexico, USA, August 20-26,2018, pages 2814-2822.",
"links": null
},
"BIBREF31": {
"ref_id": "b31",
"title": "Stratification for confoundingpart 1: The mantel-haenszel formula",
"authors": [
{
"first": "Giovanni",
"middle": [],
"last": "Tripepi",
"suffix": ""
},
{
"first": "J",
"middle": [],
"last": "Kitty",
"suffix": ""
},
{
"first": "Friedo",
"middle": [
"W"
],
"last": "Jager",
"suffix": ""
},
{
"first": "Carmine",
"middle": [],
"last": "Dekker",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Zoccali",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Giovanni Tripepi, Kitty J Jager, Friedo W. Dekker, and Carmine Zoccali. 2010. Stratification for confounding - part 1: The mantel-haenszel formula.",
"links": null
},
"BIBREF33": {
"ref_id": "b33",
"title": "Response Generation in Dialogue Using a Tailored PCFG Parser",
"authors": [
{
"first": "Caixia",
"middle": [],
"last": "Yuan",
"suffix": ""
},
{
"first": "Xiaojie",
"middle": [],
"last": "Wang",
"suffix": ""
},
{
"first": "Qianhui",
"middle": [],
"last": "He",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the 15th European Workshop on Natural Language Generation (ENLG)",
"volume": "",
"issue": "",
"pages": "81--85",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Caixia Yuan, Xiaojie Wang, and Qianhui He. 2015. Response Generation in Dialogue Using a Tailored PCFG Parser. In Proceedings of the 15th European Workshop on Natural Language Generation (ENLG), pages 81-85. Association for Computational Linguistics.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"uris": null,
"num": null,
"text": "Envisioned tool for style checking and suggestion.",
"type_str": "figure"
},
"FIGREF1": {
"uris": null,
"num": null,
"text": "C7) Whether-cond.: Expresses alternative (disjunctive) conditions. <... whether ... or ...> Whether you are overweight or not, it is always better to watch your diet. (CS) Comp./Super. Adj. and Adv.: Used to compare differences between the two objects/states they modify. <... [JJR/JJS/RBR/RBS] ...> (comp. adjective): My house is larger. (PV) Passive Voice: Occurs when the object of an action is changed into the subject of a sentence. <... [to be] ... [VBN] ...> The problem is solved. Table 1: An overview of the (B) Balance, (O) Omission, and (R) Repetition figurative devices, and the (C) Conditionals, (CS) Comparatives and Superlatives, and (PV) Passive Voice ordinary devices.",
"type_str": "figure"
},
"FIGREF2": {
"uris": null,
"num": null,
"text": "The distribution of rhetorical devices among the authors (a), the genres (b), and the topics (c) in the newspaper articles dataset, and the distribution of rhetorical devices in the debate dataset (d).",
"type_str": "figure"
},
"TABREF0": {
"html": null,
"type_str": "table",
"text": "Lists a series of details, words or phrases. < ...W [CC | COMMA] W ...> Diligence, talent and passion will drive anybody to success. (B2) Isocolon: Similarly structured elements with the same length. < ...<Phr>a <Phr>a ...<Phr>a ...> Fill the armies, rule the air, and pour out the munitions. (B3) Pysma: Asking multiple questions successively. < ...< Cl? > < Cl? > ...>? Ex: Who are you? Why are you doing here (O1) Asyndeton: Omission of conjunctions between clauses. <Cla> COMMA <Cl b > COMMA <Clc> ... ) I came, I saw, I conquered. Placing last, in a construction containing several elements of equal value, the word(s) on which all of them depend. < ...[W]a , [W] b , [W]c ...Vb > Friends, Romans, countrymen, lend me your ears. . . Placing the verb that holds together the entire sentence either at the very beginning or the very ending of that sentence. < Vb ...> or, < ...Vb > Neither a borrower nor a lender be. Repetition at the end of a line, the word(s) that occurred at the beginning of the same line. < [W]a ...[W]a > Believe not all you can hear, tell not all you believe. Repetition of the same word(s) at the beginning and middle of successive sentences. < [W]a ...[W] b ...> < [W]a ...[W] b ...> I was looking for a paper. I was anxious for a paper. Repetition of the same word(s) at the end of successive sentences. < ...[W]a > < ...[W]a > O apple! wretched apple! Miserable apple! (R4) Mesodiplosis: Repetition of the same word(s) in the middle of successive sentences. < ...[W]a ...> < ...[W]a ...> There's no time like the future! There's no time like the past! (R5) Anadiplosis: Repetition of the last word(s) from the previous sentence at the beginning of the next. < ...[W]a > < [W]a ...> We ordered a pizza pie. A pizza pie that changed our lives. (R6) Diacope: Repetition of a word with one or more in between. < ...[W]a ...[W]a ...> The horror! Oh, the horror! (R7) Epizeuxis: Repetition of words with no others between. Several conjunctions in close succession (mainly between clauses). <Cla> CC <Cl b > CC <Clc> ... He ran and jumped and laughed for joy. If I had worked harder, I would have passed the exam. (C5) If-Counterf.: Statements that examine how a hypothetical change in a past experience could have affected the outcome of that experience. If [VBD+VBN], then [past modals] If I were you, I wouldn't come.",
"num": null,
"content": "<table><tr><td>Device definition | Formalization| Example</td></tr><tr><td>(B1) Enumeration: (O2) Hypozeugma: (R3) Epiphoza: < [W]a [W]a > Awake, awake and stand up O Jerusalem. (R8) Polysyndeton: Device definition | Formalization| Example (C1) If-cond. Zero: Conditionals which express general truths. If [VB/VBP/VBZ], then [VB/VBP/VBZ] If you heat ice, it melts. (C2) If-cond. One: Express situations which are very likely to happen in the future. If [VB/VBP/VBZ/VBG], then [MD+VB] If it rains, you will get wet. (C3) If-cond. Two: Expresses consequences that will not likely happen in the future. If [VBD], then [MD+VB] If it rained, you would get wet. (C4) If-cond. Three: Used to explain that present circumstances would be different if something different had happened in the past. If [VBD+VBN] , then [MD+VBN]</td></tr></table>"
},
"TABREF2": {
"html": null,
"type_str": "table",
"text": "The precision, recall, and F 1 -score for identifying the 26 rhetorical devices.",
"num": null,
"content": "<table><tr><td>Category</td><td colspan=\"4\">Instances Prec. Recall F1</td></tr><tr><td>(B) Balance</td><td>300</td><td>0.67</td><td colspan=\"2\">0.88 0.76</td></tr><tr><td>(O) Omission</td><td>180</td><td>0.4</td><td colspan=\"2\">0.81 0.54</td></tr><tr><td>(R) Repetition</td><td>420</td><td>0.6</td><td colspan=\"2\">0.77 0.67</td></tr><tr><td>(C) Conditionals</td><td>420</td><td>0.85</td><td>0.8</td><td>0.82</td></tr><tr><td>(CS) Comp.&Super.</td><td>278</td><td>0.59</td><td colspan=\"2\">0.62 0.60</td></tr><tr><td>(PV) Passive voice</td><td>60</td><td>0.78</td><td colspan=\"2\">0.98 0.87</td></tr></table>"
},
"TABREF3": {
"html": null,
"type_str": "table",
"text": "Title: HEALTH; Salk's Injectable Polio Vaccine May Be Revived Rhetorical Device Comp. Adv. Health o cials are considering a major change in the strategy of polio vaccination, using a new, more potent version of the injectable Salk vaccine that helped eradicate polio in the United States almost 30 years ago.",
"num": null,
"content": "<table/>"
}
}
}
} |