File size: 126,507 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 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 | {
"paper_id": "2021",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T13:27:40.812054Z"
},
"title": "Semi-automatic discourse annotation in a low-resource language: Developing a connective lexicon for Nigerian Pidgin",
"authors": [
{
"first": "Marian",
"middle": [],
"last": "Marchal",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Science and Technology",
"location": {}
},
"email": "marchal@coli.uni-saarland.de"
},
{
"first": "Merel",
"middle": [],
"last": "Scholman",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Science and Technology",
"location": {}
},
"email": "m.c.j.scholman@coli.uni-saarland.de"
},
{
"first": "Vera",
"middle": [],
"last": "Demberg",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Language Science and Technology",
"location": {}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "Cross-linguistic research on discourse structure and coherence marking requires discourse-annotated corpora and connective lexicons in a large number of languages. However, the availability of such resources is limited, especially for languages for which linguistic resources are scarce in general, such as Nigerian Pidgin. In this study, we demonstrate how a semi-automatic approach can be used to source connectives and their relation senses and develop a discourse-annotated corpus in a low-resource language. Connectives and their relation senses were extracted from a parallel corpus combining automatic (PDTB end-to-end parser) and manual annotations. This resulted in Naija-Lex, a lexicon of discourse connectives in Nigerian Pidgin with English translations. The lexicon shows that the majority of Nigerian Pidgin connectives are borrowed from its English lexifier, but that there are also some connectives that are unique to Nigerian Pidgin.",
"pdf_parse": {
"paper_id": "2021",
"_pdf_hash": "",
"abstract": [
{
"text": "Cross-linguistic research on discourse structure and coherence marking requires discourse-annotated corpora and connective lexicons in a large number of languages. However, the availability of such resources is limited, especially for languages for which linguistic resources are scarce in general, such as Nigerian Pidgin. In this study, we demonstrate how a semi-automatic approach can be used to source connectives and their relation senses and develop a discourse-annotated corpus in a low-resource language. Connectives and their relation senses were extracted from a parallel corpus combining automatic (PDTB end-to-end parser) and manual annotations. This resulted in Naija-Lex, a lexicon of discourse connectives in Nigerian Pidgin with English translations. The lexicon shows that the majority of Nigerian Pidgin connectives are borrowed from its English lexifier, but that there are also some connectives that are unique to Nigerian Pidgin.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Resources such as discourse-annotated corpora (e.g. PDTB, Prasad et al., 2008) and connective lexicons (e.g. DimLEX, Stede and Umbach, 1998) have proven highly valuable in investigating discourse structure and coherence marking, but they are only available for a limited number of languages. One promising way to expand these resources is to project annotations from languages for which such resources are available to underresourced languages. The current paper sets out to investigate how automatic discourse parsing can be combined with manual annotation to construct a connective lexicon in a low-resource language. We apply this method to create Naija-Lex, a lexicon of Nigerian Pidgin, but the method can also be adapted to apply to other low-resource languages.",
"cite_spans": [
{
"start": 58,
"end": 78,
"text": "Prasad et al., 2008)",
"ref_id": "BIBREF18"
},
{
"start": 117,
"end": 140,
"text": "Stede and Umbach, 1998)",
"ref_id": "BIBREF26"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Pidgin and creole languages, such as Nigerian Pidgin (commonly referred to as 'Pidgin' or 'Naija'), tend to be less complex on a syntactic and morphological level than their lexifiers (Szmrecsanyi et al., 2009) . This raises interesting questions with respect to their discourse complexity, such as: how is discourse structure realized in contact languages? And how does this affect discourse marking? In addition, because pidgins are still in an evolutionary stage, they are a testing ground for examining how connectives emerge. However, the scarcity of linguistic resources, such as discourseannotated corpora, for these languages makes it difficult to answer these questions. Another contribution of this paper is the creation of a discourseannotated layer for a parallel corpus of Nigerian Pidgin, which can be used to study these questions.",
"cite_spans": [
{
"start": 184,
"end": 210,
"text": "(Szmrecsanyi et al., 2009)",
"ref_id": "BIBREF27"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We adopt a semi-automatic approach using automatic connective identification and annotation projection, combined with manual annotation, to initialize a list of discourse connective candidates in Nigerian Pidgin. This allows us to evaluate how existing tools can be adopted to investigate lowresource languages. More specifically, we assess the usefulness of the PDTB end-to-end parser (Lin et al., 2014) in identifying connectives in informal spoken English translated text, to source connectives in the target language.",
"cite_spans": [
{
"start": 386,
"end": 404,
"text": "(Lin et al., 2014)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "The main contributions of the present research are the following:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 We identify challenges in constructing a lexicon in an under-resourced language and present a methodology for sourcing discourse connectives and their associated relation senses in such a language.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 We create a lexicon of discourse connectives in Nigerian Pidgin with English translations (Naija-Lex), which enables further research on discourse structure and coherence marking in Pidgin.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "\u2022 We add a discourse-annotated layer to an existing Nigerian Pidgin-English parallel corpus containing explicit connectives and their relation senses.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Section 2 discusses earlier work on automatic discourse annotation and lexicon construction, after which Nigerian Pidgin is introduced in Section 3. Section 4 describes the method used to identify the connectives in Pidgin. We elaborate on the Naija-Lex lexicon that we created using this method in Section 5 and evaluate our method in Section 6.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Connective inventories have been developed for various languages, including German (Stede and Umbach, 1998) , French (Roze et al., 2012) , Chinese (Zhou and Xue, 2015) and English (Das et al., 2018) , among others (see also . Recently, these efforts have been extended with several multi-lingual connective databases (Bourgonje et al., 2017; Kurfal\u0131 et al., 2020) .",
"cite_spans": [
{
"start": 83,
"end": 107,
"text": "(Stede and Umbach, 1998)",
"ref_id": "BIBREF26"
},
{
"start": 117,
"end": 136,
"text": "(Roze et al., 2012)",
"ref_id": "BIBREF22"
},
{
"start": 147,
"end": 167,
"text": "(Zhou and Xue, 2015)",
"ref_id": "BIBREF34"
},
{
"start": 180,
"end": 198,
"text": "(Das et al., 2018)",
"ref_id": "BIBREF8"
},
{
"start": 317,
"end": 341,
"text": "(Bourgonje et al., 2017;",
"ref_id": "BIBREF1"
},
{
"start": 342,
"end": 363,
"text": "Kurfal\u0131 et al., 2020)",
"ref_id": "BIBREF11"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related work 2.1 Lexicon creation",
"sec_num": "2"
},
{
"text": "One challenging aspect about constructing a connective lexicon for a new language is defining the category of connectives. Traditionally, the PDTB (Prasad et al., 2008) has placed strong syntactic restrictions on which lexical items can function as connectives, considering only subordinating and coordinating conjunctions and adverbials. point out that for multilingual lexicons, such a strict definition is not feasible, as languages differ highly in the way coherence relations are expressed. Still, restrict the class of connectives syntactically, by posing that a connective cannot be inflected or modified, and that verbs should be excluded for this reason.",
"cite_spans": [
{
"start": 147,
"end": 168,
"text": "(Prasad et al., 2008)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related work 2.1 Lexicon creation",
"sec_num": "2"
},
{
"text": "Similar to certain other lexicons (e.g. LDM-PT, Mendes and Lejeune, 2016) , the lexicon we present here will maintain less strict requirements on the syntactic category and modifiability of the lexical item. In this way, we can allow for more variation that might be present in a relatively young language and obtain a more complete overview of discourse marking in Nigerian Pidgin.",
"cite_spans": [
{
"start": 48,
"end": 73,
"text": "Mendes and Lejeune, 2016)",
"ref_id": "BIBREF14"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related work 2.1 Lexicon creation",
"sec_num": "2"
},
{
"text": "In the last decade, shallow discourse parsing of explicit connectives has received considerable attention in the computational discourse community (e.g. Lin et al., 2014; Pitler and Nenkova, 2009; Wang and Lan, 2015; . One challenge for such automatic parsers is that a discourse connective often also has a non-connective function, which is difficult to distinguish. Still, automatic discourse parsers for English achieve an F1 of almost 95% in identifying explicit connectives (Lin et al., 2014; Wang and Lan, 2015) .",
"cite_spans": [
{
"start": 153,
"end": 170,
"text": "Lin et al., 2014;",
"ref_id": "BIBREF13"
},
{
"start": 171,
"end": 196,
"text": "Pitler and Nenkova, 2009;",
"ref_id": "BIBREF17"
},
{
"start": 197,
"end": 216,
"text": "Wang and Lan, 2015;",
"ref_id": "BIBREF29"
},
{
"start": 479,
"end": 497,
"text": "(Lin et al., 2014;",
"ref_id": "BIBREF13"
},
{
"start": 498,
"end": 517,
"text": "Wang and Lan, 2015)",
"ref_id": "BIBREF29"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Connective identification",
"sec_num": "2.2"
},
{
"text": "Discourse parsers are still very much restricted in terms of language and domain, because of a strong focus on the use of the PDTB for training. Parsers on languages other than English often reach much lower performance (e.g. Xue et al., 2016) , although the recent DisRPT connective detection task has yielded some high performing systems in other languages as well (Muller et al., 2019; Yu et al., 2019) . However, these parsers still require training data, which is not available for low-resource languages.",
"cite_spans": [
{
"start": 226,
"end": 243,
"text": "Xue et al., 2016)",
"ref_id": "BIBREF30"
},
{
"start": 367,
"end": 388,
"text": "(Muller et al., 2019;",
"ref_id": "BIBREF15"
},
{
"start": 389,
"end": 405,
"text": "Yu et al., 2019)",
"ref_id": "BIBREF31"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Connective identification",
"sec_num": "2.2"
},
{
"text": "Using parallel corpora might therefore be a promising method to identify connectives and then project their annotations to a low-resource language. One possible difficulty is that these corpora might be from different domains than what the parsers have been trained on, which leads to lower performance (Ramesh and Yu, 2010; Knaebel and Stede, 2020; Scholman et al., 2021) .",
"cite_spans": [
{
"start": 325,
"end": 349,
"text": "Knaebel and Stede, 2020;",
"ref_id": "BIBREF10"
},
{
"start": 350,
"end": 372,
"text": "Scholman et al., 2021)",
"ref_id": "BIBREF23"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Connective identification",
"sec_num": "2.2"
},
{
"text": "We will use the PDTB end-to-end parser (Lin et al., 2014) , as it has been shown to perform better on spontaneous spoken data than other parsers (Scholman et al., 2021). However, performance of the parser drops considerably when comparing spoken language with the newspaper text data it was trained on. We will therefore also assess its reliability on an out-of-domain translated text.",
"cite_spans": [
{
"start": 39,
"end": 57,
"text": "(Lin et al., 2014)",
"ref_id": "BIBREF13"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Connective identification",
"sec_num": "2.2"
},
{
"text": "Several previous studies have used research available on connectives in one language (mostly English) to identify connectives in another language. Zhou and Xue (2012) and Das et al. (2020) have used bilingual dictionaries to translate English connectives to Chinese and Bangla respectively as part of identifying connectives in these target languages to construct a lexicon. However, no dictionaries are available for Nigerian Pidgin.",
"cite_spans": [
{
"start": 147,
"end": 166,
"text": "Zhou and Xue (2012)",
"ref_id": "BIBREF33"
},
{
"start": 171,
"end": 188,
"text": "Das et al. (2020)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Connective projection",
"sec_num": "2.3"
},
{
"text": "Another line of research in using cross-lingual information to source connectives is to use parallel corpora. Versley (2010) projected discourse connectives across an English-German parallel corpus, but found that simple word alignment is not enough to reliably identify connectives in text and that performance increased when bootstrapping using an established connective list. combined a German connective lexicon with paral-lel text corpora to find Dutch connectives and subsequently manually checked and annotated them for relation sense. However, this approach does not automatically distinguish between connective and non-connective usages of connectives in the source languages, which always needs to be resolved manually. The current research attempts to resolve this by using a shallow discourse parser to identify connectives in the translation text automatically.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective projection",
"sec_num": "2.3"
},
{
"text": "A more similar approach to that adopted in the current study is that of Laali and Kosseim (2014) . They used the PDTB end-to-end parser to identify English discourse connectives in the Europarl corpus source text, after which they used word alignment to identify connectives in the parallel French texts. This mapping was further improved with additional filtering based on frequency and syntactic information. However, performance is highly dependent on the length and syntactic categories of discourse connectives, which vary between languages. In addition, automatic word alignment can only be applied successfully if enough parallel data is available, while resources for the language under consideration here are limited.",
"cite_spans": [
{
"start": 72,
"end": 96,
"text": "Laali and Kosseim (2014)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Connective projection",
"sec_num": "2.3"
},
{
"text": "One possible drawback of using translation text to source connectives is that it draws on the assumption that coherence relations in the source text and the translated text remain the same. However, this is not necessarily the case. Discourse relations often allow multiple interpretations, but a translated text might only contain the interpretation of the translator (i.e. in cases where the original connective is disambiguated). Secondly, there are several factors guiding when discourse relations are implicitated or explicitated in translation text, such as specific features of the target language and the relation sense (Becher, 2011; Zufferey and Cartoni, 2014) . The distribution of explicit relations in the translation text thus does not necessarily reflect the distribution of explicit relations in the source text. This would influence the likelihood of connectives found in the source language by only taking into account the connectives in the translation text. We will therefore also assess how well the explicit connectives in the translated text reflect the discourse relations in the source text.",
"cite_spans": [
{
"start": 628,
"end": 642,
"text": "(Becher, 2011;",
"ref_id": "BIBREF0"
},
{
"start": 643,
"end": 670,
"text": "Zufferey and Cartoni, 2014)",
"ref_id": "BIBREF35"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Connective projection",
"sec_num": "2.3"
},
{
"text": "Finally, connective lexicons contain information about the relation senses that can be signalled by a certain connective. The PDTB end-to-end parser provides relation sense labels for the connectives it identifies, which could be projected to the con-nectives in the parallel text. This was, however, not taken into account by Laali and Kosseim (2014) . The present research therefore evaluates and extends their methodology for lexicon creation with a parallel corpus.",
"cite_spans": [
{
"start": 327,
"end": 351,
"text": "Laali and Kosseim (2014)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Connective projection",
"sec_num": "2.3"
},
{
"text": "Nigerian Pidgin is a contact language spoken all over Nigeria. Although officially a pidgin, it is also often considered a creole (e.g., Courtin et al., 2018) , as it is a relatively stable language with a fullydeveloped vocabulary and grammar. In addition to approximately 5 million native speakers, there are estimated to be almost 100 million speakers of Nigerian Pidgin as a second language. It is often also referred to simply as 'Pidgin' by Nigerian Pidgin speakers themselves.",
"cite_spans": [
{
"start": 137,
"end": 158,
"text": "Courtin et al., 2018)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Nigerian Pidgin",
"sec_num": "3"
},
{
"text": "English is one of the main lexifiers of Nigerian Pidgin, with many words having a similar form and meaning as the English origin, as can be seen in Example (1).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Nigerian Pidgin",
"sec_num": "3"
},
{
"text": "(1) I say 'toh, me I sabi operate computer small small'. I said 'well, I know a bit about computers'.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Nigerian Pidgin",
"sec_num": "3"
},
{
"text": "Nigerian Pidgin is also influenced by other European languages (e.g. Portuguese, illustrated by sabi in the example above) as well as various indigenous languages it has been in contact with, such as Hausa, Igbo and Yoruba. Nigerian Pidgin is characterized by focus constructions (see e.g. me in (1)), which are often indicated with the focus particle na (Caron et al., 2019) . In addition, serial verb constructions are typical for Nigerian Pidgin. As a result, coordinating conjunctions are less frequent in Nigerian Pidgin than in English (Courtin et al., 2018) . An example of such a serial verb construction is provided in Example (2), which has been translated in English with the coordinating conjunction and. This suggests that discourse relations might be expressed differently in Nigerian Pidgin.",
"cite_spans": [
{
"start": 355,
"end": 375,
"text": "(Caron et al., 2019)",
"ref_id": "BIBREF3"
},
{
"start": 542,
"end": 564,
"text": "(Courtin et al., 2018)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Nigerian Pidgin",
"sec_num": "3"
},
{
"text": "(2) Con dey hustle, con hustle money go give am.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Nigerian Pidgin",
"sec_num": "3"
},
{
"text": "Then we hustled, got money and went to give it to her.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Nigerian Pidgin",
"sec_num": "3"
},
{
"text": "Linguistic research and resources on Nigerian Pidgin are limited, although this has shifted in recent years. As part of a larger project studying the syntactic and prosodic structure of Nigerian Pidgin, Caron et al. (2019) are in the process of creating a 500,000 word corpus of spoken data, part of which is already available and will be used in the current study. We also note related work from the NLP community in studying Pidgin to explore new approaches for low-resource natural language generation (Chang et al., 2020) and translation (Ogueji and Ahia, 2019) . However, no previous research has examined the discourse structure of Nigerian Pidgin.",
"cite_spans": [
{
"start": 203,
"end": 222,
"text": "Caron et al. (2019)",
"ref_id": "BIBREF3"
},
{
"start": 505,
"end": 525,
"text": "(Chang et al., 2020)",
"ref_id": "BIBREF4"
},
{
"start": 542,
"end": 565,
"text": "(Ogueji and Ahia, 2019)",
"ref_id": "BIBREF16"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Nigerian Pidgin",
"sec_num": "3"
},
{
"text": "The data we use in our analysis comes from the gold section of the Naija Treebank (UD NSC Corpus), 1 a parallel corpus of Nigerian Pidgin utterances with English translations. This dataset consists of 140,859 words collected in various locations, and represents 87 speakers. The sampling of speakers aimed at balancing age, sex, education, linguistic and geographic background. The genres recorded cover life stories, speeches, radio programs, free conversations, cooking recipes, comments on current state of affairs, etc. The translation of the Nigerian Pidgin sentences into English was done by a team of native speakers of Nigerian Pidgin, and aimed at remaining as faithful as possible to the structure and style of the original utterances (Caron et al., 2019) .",
"cite_spans": [
{
"start": 745,
"end": 765,
"text": "(Caron et al., 2019)",
"ref_id": "BIBREF3"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Methodology 4.1 Data",
"sec_num": "4"
},
{
"text": "We extracted the original Nigerian Pidgin utterances (referred to as Source Text -ST), together with their English translations (referred to as Translated Text -TT) and the Nigerian Pidgin POS tags from the NSC corpus. In total, the dataset used in this study contains 9,242 Nigerian Pidgin utterances, divided into three subsets: dev (n=991), train (n=7,279), and test (n=972). Figure 1 displays the workflow used in this study. As Step 1, we ran an automatic end-to-end PDTB classifier on the English text to extract the TT connectives, by disambiguating between connective and non-connective usages of connective tokens. This resulted in the extraction of 4,592 TT connective tokens (dev: n=454, train: n=3,691, test: n=446), consisting of 41 different TT connective types, in 3,389 utterances. In addition, the parser provides the relation sense (PDTB 2.0) of each English connective.",
"cite_spans": [],
"ref_spans": [
{
"start": 379,
"end": 387,
"text": "Figure 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "Methodology 4.1 Data",
"sec_num": "4"
},
{
"text": "As a next step, the first two authors both manually annotated the Nigerian Pidgin counterpart of all English connectives in the dev set in order to obtain an initial dictionary of English-Nigerian Pidgin connective mappings. Following the criteria from the PDTB (Prasad et al., 2008) , only Nigerian Pidgin words from the following syntactic categories were initially considered as connective candidates: coordinating conjunctions, subordinating conjunctions, adverbs, and adpositions. Of 446 TT connectives, 336 were mapped onto an equivalent in the ST. This mapping functioned as input for an initial (seed) connective dictionary, which contained 30 Pidgin connective candidates and 52 unique co-occurrence types (excluding TT connective\u223cimplicit ST pairing).",
"cite_spans": [
{
"start": 262,
"end": 283,
"text": "(Prasad et al., 2008)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Construction of the lexicon",
"sec_num": "4.2"
},
{
"text": "This initial dictionary was then used to predict the Nigerian Pidgin equivalent of the extracted English connectives in the train set in Step 3. For each utterance containing an English connective in the TT, the POS-tagged ST was searched for one of the Nigerian Pidgin equivalents that had been mapped onto this connective in the dev set. In order to predict a mapping for cases where a single Pidgin connective was mapped to multiple English connectives or vice versa we calculated normalized pointwise mutual information (NPMI) of co-occurrences in our dictionary. The connective match with the highest NPMI was selected as the target connective. For example, English until could be a translation of Nigerian Pigin till or sotay. If the ST of a TT containing until would contain both till and sotay, sotay would be selected as its association with until is stronger than that of till.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Construction of the lexicon",
"sec_num": "4.2"
},
{
"text": "In cases where the NPMI did not resolve to a one-to-one mapping (e.g. the TT text contained two instances of until, with only one equivalent in the ST), the Nigerian Pidgin connective with the closest relative position in the sentence was selected. When no Pidgin translation equivalent was found, the four words closest to the relative position in the TT were searched for another Pidgin connective that had not been mapped to this English connective yet to extend the initial dictionary with new candidate mappings.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Construction of the lexicon",
"sec_num": "4.2"
},
{
"text": "To examine whether the data set contained other Nigerian Pidgin connectives that did not occur in the dev subset, we manually annotated selected instances from the train set where (a) an English connective had been identified in the TT that did not appear in the dev set and (b) where no Nigerian Pidgin translation equivalent was found (Step 4). We also manually examined all instances where the Pidgin connective candidates 'wey' or 'sey' were extracted from the ST, since these are frequently used as complementizers rather than connectives. In this annotation round we expanded the connective dictionary by relaxing restrictions on the syntactic category, due to high co-occurrence of English connectives with certain auxiliary verbs and lexical phrases in Pidgin. We then further verified the dictionary by checking the new English-Nigerian Pidgin connective mappings that did not occur in the seed dictionary and consulted a native Pidgin-speaking linguist to verify (a) whether connectives that did not have an English origin were indeed a connective in Pidgin and (b) whether the English translations were valid.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Construction of the lexicon",
"sec_num": "4.2"
},
{
"text": "The updated dictionary was used to improve the automatically identified Nigerian Pidgin connectives in the train set (Step 5) and to construct the connective lexicon. The lexicon was complemented with automatic relation sense labels and the frequency data from the corpus.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Construction of the lexicon",
"sec_num": "4.2"
},
{
"text": "In total, 4,186 Nigerian Pidgin connective tokens were identified. They were retrieved based on 40 English connective types in the translation text. The Nigerian Pidgin connective tokens consisted of 77 types, which were categorized in 57 entries in the lexicon. Naija-Lex is made available online. 2 For each connective entry, the lexicon contains information on its frequency, alternative forms, syntactic category(/ies), English translation equivalents and non-connective usage. In addition, the various relation senses that the connective can signal are included, together with an example of the Nigerian Pidgin connective in every sense and the relation sense distribution.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Naija-Lex: A lexicon of Nigerian Pidgin connectives",
"sec_num": "5"
},
{
"text": "The majority of the connective entries (n=39) are derived from connectives in the English lexifier. Their phonology has often been adapted slightly to that of Nigerian Pidgin, leading to multiple variants of the same connective. For example, because occurs in Nigerian Pidgin as because, but also as cause and cos, and dough has been adapted from English though. Only 18 connective types in the lexicon did not originate from English connectives. Examples of these are abi (English: or), wey be sey (English: when) and sotay (English: until). Some of these Pidgin connectives have evolved from English non-connective words. Consider con (originating from English 'come') and make. Both have evolved from an English verb and can now be used as a main verb in Pidgin, in which case they do not function as connectives, as well as an auxiliary, in which case they are frequently translated as connectives expressing causal and temporal connectives in English (see Examples (3) and (4)).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective origin and syntactic category",
"sec_num": "5.1"
},
{
"text": "(3) I con realise sey omo na dis Pidgin na im make us connect like dat.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective origin and syntactic category",
"sec_num": "5.1"
},
{
"text": "Then I realized that, wow, this Pidgin brought us together.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective origin and syntactic category",
"sec_num": "5.1"
},
{
"text": "(4) De want hold am make e no fall. They want to hold them so they won't fall off.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective origin and syntactic category",
"sec_num": "5.1"
},
{
"text": "Con is used frequently in narration, connecting the events in one utterance with the following events in the next utterance. Like many other Nigerian Pidgin auxiliary verbs, con is positioned between the subject and the main verb of the clause. By contrast, make is only found in clause-initial position. In a non-connective usage, it often functions as a directive, its meaning equivalent to 'should' in English. This causal function has likely extended to its usage as a causal discourse connective. Note that make does not satisfy the criterion of not being modifiable (cf. : depending on the pronoun, it is inflected to meh or mey. We still consider it a connective, because it can also be used to signal a causal relation between two abstract, clausal arguments (similar to causal discourse verbs, see e.g. Danlos, 2006) .",
"cite_spans": [
{
"start": 812,
"end": 825,
"text": "Danlos, 2006)",
"ref_id": "BIBREF7"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Connective origin and syntactic category",
"sec_num": "5.1"
},
{
"text": "Other unique Pidgin connectives have evolved from Pidgin words that originally did not express a discourse function. The connective naim is a contraction of the auxiliary verb na, which is often used as a focus particle, and the 3SG pronoun im and has been lexicalized to a causal/temporal connective, equivalent to then or so. An example of its connective and non-connective use can be found in (5) and (6) respectively:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective origin and syntactic category",
"sec_num": "5.1"
},
{
"text": "(5) E get one lady, hm she just enter inside shop o. Naim she say she wan do facials. There was one lady, she just went into the shop. Then she said she wanted to do facials.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective origin and syntactic category",
"sec_num": "5.1"
},
{
"text": "(6) And na im dey give am. And he's the one giving it to her.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective origin and syntactic category",
"sec_num": "5.1"
},
{
"text": "The most frequent syntactic category of the connective variants was subordinating conjunctions (n=16), followed by adverbs (n=15) and adpositions (n=11). The lexicon also contains one particle as Pidgin connective (shey, English so), which is another syntactic class traditionally not considered to contain connectives in English.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective origin and syntactic category",
"sec_num": "5.1"
},
{
"text": "In terms of frequency of connectives in the Nigerian Pidgin data, the data shows that if was the most frequent (n=643), followed by so (n=606) and and (n=511). The most frequent connectives that are not English cognates are con and naim, with 394 and 77 occurrences respectively. classes (based on PDTB2 level 2 relation labels).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective and relation sense distribution",
"sec_num": "5.2"
},
{
"text": "Temporal and Cause were the most frequent explicit relations in the corpus. Table 1 also shows that Conjunction relations were frequently implicit in the ST and had been explicitated in the TT. Importantly, the distribution is similar in the manually annotated dev set, indicating that this distribution is not an artefact of the automatic annotation. When considering the relational distribution of each connective, we find that most connectives have one clear dominant meaning. This is illustrated in Figure 2 . For example, Nigerian Pidgin den is almost exclusively used in temporal relations. The connective as is predominantly used as a temporal connective, but similar to English, has also maintained its causal meaning. The non-cognate connectives con and naim, however, are used in various relations, mostly Temporal and Cause. Although it might seem surprising that additional connectives have developed for relations that can also be signalled by English cognate connectives (e.g. den for Temporal and so for Cause), we note that these relations are also the most frequent, which might explain why speakers are more likely to find new ways to express them. Alternatively, these new connectives might have evolved precisely because of the polyfunctional usage. The connectives borrowed from English clearly signal one dominant relation, whereas con and naim can be used to underspecify these discourse relations.",
"cite_spans": [],
"ref_spans": [
{
"start": 76,
"end": 83,
"text": "Table 1",
"ref_id": "TABREF0"
},
{
"start": 503,
"end": 511,
"text": "Figure 2",
"ref_id": "FIGREF1"
}
],
"eq_spans": [],
"section": "Connective and relation sense distribution",
"sec_num": "5.2"
},
{
"text": "6 Evaluation of the method",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective and relation sense distribution",
"sec_num": "5.2"
},
{
"text": "To evaluate accuracy of the first step in our pipeline we manually annotated the TT in the development set for presence of connectives, according to the criteria set by . Connectives were identified in the English TT by the PDTB end-to-end parser with relatively high reliability (precision = 85.2%, recall = 85.6%, F1 = 85.4%). This is lower than the parser's performance on the test set of the PDTB (i.e. 95.76%), but that was expected, because the parser was trained on text from the written newspaper domain.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Automatic connective identification in TT",
"sec_num": "6.1"
},
{
"text": "A closer examination of the performance per connective shows that precision is especially low on so, so that, also and later (<= 50%). In spoken language, so is often used as a discourse marker rather than a connective. 3 For so that, the majority of mistakes were caused by the parser identifying so with a demonstrative that as a connective. Low recall was reached for afterwards and yet, which occurred infrequently and were not detected by the parser. The inaccuracy of automatic connective identification is thus mainly due to a small number of connectives. Manual examination of these instances can improve the accuracy significantly.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Automatic connective identification in TT",
"sec_num": "6.1"
},
{
"text": "The PDTB end-to-end parser was also used to extract relation senses for the connectives in the TT. In order to evaluate the parser's reliability, the second author and an independent coder annotated the English connectives in the dev set using PDTB2 labels. Agreement was high (Cohen's \u03ba: 0.88, CI [0.85, 0.92]; AC 1 = 0.89, CI [0.86, 0.92], level 2 labels) (see Spooren and Degand, 2010, for IAA standards in discourse relation annotation). Disagreements were resolved by the first author.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Relation sense identification in TT",
"sec_num": "6.2"
},
{
"text": "For the connectives that the parser had identified correctly, the label provided by the parser was identical to the gold label in 78.8% of cases (Cohen's \u03ba: 0.74, CI [0.70, 0.79]; AC 1 = 0.77, CI [0.72, 0.81]).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Relation sense identification in TT",
"sec_num": "6.2"
},
{
"text": "Again, reliability was characterized by low performance on a small number of relation senses. More specifically, there was extremely low precision on Concession relations (0.04), as well as low recall on Contrast relations (0.23), because many Concession relations were identified as Contrast relations by the parser. Since the parser cannot distinguish between these relations reliably, these senses were merged in further analyses. Note that these distinctions are notoriously difficult to annotate, even by experts (Robaldo and Miltsakaki, 2014; Zufferey et al., 2012) . Similarly, recall was low on Synchrony relations (0.30), since many of them were considered Asynchronous in manual annotation. These senses were therefore also collapsed. Adopting this 8-way classification resulted in very high reliability: (Cohen's \u03ba: 0.90, CI [0.86, 0.94]; AC 1 = 0.91, CI [0.88, 0.95]).",
"cite_spans": [
{
"start": 518,
"end": 548,
"text": "(Robaldo and Miltsakaki, 2014;",
"ref_id": "BIBREF21"
},
{
"start": 549,
"end": 571,
"text": "Zufferey et al., 2012)",
"ref_id": "BIBREF36"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Relation sense identification in TT",
"sec_num": "6.2"
},
{
"text": "The connective in fact was classified by the parser as signalling a Conjunction relation instead of Restatement, leading to a low precision (0.20) on this relation sense as well. The parser classified the two instances of as if in the development set as Concession, whereas the manual annotation was Expansion. Note that the PDTB 2.0 manual also indicates that both of these interpretations are possible for these connectives.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Relation sense identification in TT",
"sec_num": "6.2"
},
{
"text": "To evaluate how reliable the automatic connective mapping was with the updated dictionary (i.e.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective mapping from TT to ST",
"sec_num": "6.3"
},
{
"text": "Step 5 in Figure 1 ) and how the lexicon extends to new data, the accuracy of the final dictionary was determined based on its performance on the test set. 87.9% of the English connectives in the test set were correctly matched with their Pidgin connective. Sixteen connectives in the test set (3.6% of the total TT connectives) were incorrectly identified as implicit in ST, because their Pidgin translation had not co-occurred with this TT connective in the initial set. When allowing for the automatic mapping to identify new TT-ST pairings, the overall accuracy of the mapping in the test set increased to 90.9% (precision: 92.0, recall: 97.9, F1: 94.9).",
"cite_spans": [],
"ref_spans": [
{
"start": 10,
"end": 18,
"text": "Figure 1",
"ref_id": "FIGREF0"
}
],
"eq_spans": [],
"section": "Connective mapping from TT to ST",
"sec_num": "6.3"
},
{
"text": "The remainder of the incorrect mappings were mostly due to the English connective being incorrectly identified by the parser as a connective (n=12) or to one or multiple possible Nigerian Pidgin translations of the English connective occurring in the utterance, while they were not a translation of the English connective. Again, performance on individual connectives was high, except for a few Nigerian Pidgin connectives: sey, wey, for, na and like all reached F1 scores lower than 70%. This is likely because they also frequently occur with a non-connective meaning.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective mapping from TT to ST",
"sec_num": "6.3"
},
{
"text": "Finally, we evaluate how many Nigerian Pidgin connectives are missed when annotating only a subset of the corpus manually. In the Pidgin test set, we identified six Pidgin connectives that did not occur in the lexicon so far. Two of them were corpus annotation variants (i.e. different spelling/POS-tag) of connectives that had already been extracted from the initial set. The remainder of the new Pidgin connectives were separate entries and their English TT equivalent had either been mapped onto a new mapping or had incorrectly been identified as implicit in the ST. Extending the lexicon using additional datasets can therefore be reliably done using manual annotation of only those English TT connectives that either have not occurred previously or that are mapped onto a ST connectives that so far has not co-occurred with the English connective. 4",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Connective mapping from TT to ST",
"sec_num": "6.3"
},
{
"text": "A final premise of this method is that the relation marked in the TT reflects the relation expressed by the connective in the ST (cf. Laali and Kosseim, 2014) . To evaluate this assumption, we manually annotated the relation sense of the Nigerian Pidgin connectives that had been identified in the dev set. Inter-annotator agreement on a subset of the data was sufficient (Cohen's \u03ba: 0.77, CI [0.59, 0.94]; AC 1 = 0.82, CI [0.68, 0.96] ). These manual annotations of the ST were then compared with the projected automatic annotations of the PDTB parser. Since the analysis above shows that distinctions within the class of Temporal relations and Comparison are not reliably annotated, these classes were combined. Agreement between the parser's label and the manual annotation on this 8-way classification was high (Cohen's \u03ba: 0.90, CI [0.86, 0.94] ; AC 1 = 0.91, CI [0.88, 0.94] ) and only slightly lower than the reliability of the parser on the translation text. Using automatic annotation of a translation text to extract the relation sense of connectives in a source text is therefore well feasible. 5",
"cite_spans": [
{
"start": 134,
"end": 158,
"text": "Laali and Kosseim, 2014)",
"ref_id": "BIBREF12"
},
{
"start": 420,
"end": 429,
"text": "CI [0.68,",
"ref_id": null
},
{
"start": 430,
"end": 435,
"text": "0.96]",
"ref_id": null
},
{
"start": 815,
"end": 832,
"text": "(Cohen's \u03ba: 0.90,",
"ref_id": null
},
{
"start": 833,
"end": 842,
"text": "CI [0.86,",
"ref_id": null
},
{
"start": 843,
"end": 848,
"text": "0.94]",
"ref_id": null
},
{
"start": 864,
"end": 873,
"text": "CI [0.88,",
"ref_id": null
},
{
"start": 874,
"end": 879,
"text": "0.94]",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Relation sense projection from ST to TT",
"sec_num": "6.4"
},
{
"text": "The present research shows how extending discourse research to contact languages provides more insight in how discourse structure is realized crosslinguistically. We find a similar relational distribution between English and Pidgin. Moreover, we identified a relatively large variety of discourse markers in Nigerian Pidgin. Taken together, this indicates that the discourse structure of Pidgin is not necessarily less complex than in English, even though pidgins are generally considered to be less syntactically complex.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "Furthermore, the analysis reveals that most of the discourse connectives in Nigerian Pidgin are derived from English discourse markers. However, discourse relations in Pidgin are not always marked by lexical items from a limited set of syntactic categories (cf. : Auxiliary verbs can also signal coherence relations in Nigerian Pidgin. We therefore recommend researchers working on identifying connectives in different languages to be flexible in what items might signal coherence relations. Cross-linguistic variation should be taken into account when defining restrictions on what can constitute a discourse connective. In order to do so, more research on coherence marking in different languages (especially non-Indo-European languages) is needed.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "Nigerian Pidgin is mainly a spoken language. This affects our analysis in several ways. Firstly, the nature of spoken language influences the distribution of connectives and the relation senses they mark. A smaller number of connective types are used in spoken compared to written text (Crible and Cuenca, 2017) . At the same time, connectives in spoken language are used in a wider variety of functions. This should therefore be taken into account when considering the frequency of connectives and their relation sense distribution in the lexicon.",
"cite_spans": [
{
"start": 286,
"end": 311,
"text": "(Crible and Cuenca, 2017)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "Secondly, the performance of the parser is in-fluenced by the change in domain. A comparison with the parser's performance on a different English discourse-annotated spoken corpus (DiscoSPICE, Rehbein et al., 2016) revealed that the parser makes comparable mistakes in identifying connectives in dialogues by native speakers of English (see also Scholman et al., 2021) . Inaccuracies of the parser in the current study are thus for the most part due to the fact that our data is spoken.",
"cite_spans": [
{
"start": 180,
"end": 214,
"text": "(DiscoSPICE, Rehbein et al., 2016)",
"ref_id": null
},
{
"start": 346,
"end": 368,
"text": "Scholman et al., 2021)",
"ref_id": "BIBREF23"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "Despite the fact that the process described in this paper contains a number of manual steps, we believe that this method can easily be extended to other low-resource languages. The annotations were for the most part done by the second author, who is semi-proficient in Pidgin. In addition, comparing instances of connective candidates that were automatically identified in the corpus allowed a non-native speaker with little knowledge of the language (the first author) to manually verify the data. Although the native speaker we consulted was a trained linguist, the additional required judgments can be also be obtained in such a way that no linguistic knowledge is required from the native speaker. Furthermore, the required amount of manual annotations does not increase considerably with larger datasets. Firstly, by annotating the dev set alone, only a few infrequent connectives had been missed (6 out of 77 connective variants) and annotating a small subset of the data helps to identify these cases. Secondly, larger resources lead to a stronger reliability of statistical approaches. A combination of manual and automatic approaches therefore seems a fruitful way to construct resources for other under-researched languages as well.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "Another challenge in identifying connectives in contact languages specifically, is the large amount of variation and the frequent use of code-switching. Especially for an English-lexified language like Nigerian Pidgin, it can be difficult to distinguish between cases in which the utterance is English or Nigerian Pidgin. In fact, some of the connectives that we initially identified as ST connectives were in fact English connectives that had been used during code-switching. Frequency filters, which have also been adopted in fully-automatic approaches to cross-lingual connective identification (Laali and Kosseim, 2014) , are therefore crucial. Lexical items that only occur a few times or are only used by a small number of speakers should be checked manually. To distinguish between code-switched and infrequent connective uses, the context should be taken into account. If the connective is embedded in an English-only sentence, it is likely not a connective in the target language.",
"cite_spans": [
{
"start": 598,
"end": 623,
"text": "(Laali and Kosseim, 2014)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "The similarities between Nigerian Pidgin and English could also have been employed as an advantage. For example, as one reviewer suggests, Levenshtein's distance can be used to find Nigerian Pidgin connective candidates that are similar to the English connective. However, there are a few drawbacks to such an approach. The Nigerian Pidgin version of an English connective sometimes diverged strongly from its English cognate (e.g. because -cos) and some of the Nigerian Pidgin connectives with an English origin were not cognates of the English connective, but of a synonym (e.g. the Pidgin meaning-equivalent of instead is rader). Levenshtein's distance might therefore not always be a reliable measure. More importantly, our goal was to establish the usefulness of existing resources to research discourse structure in low-resource languages in general, not English-lexified contact languages only. In addition, in searching for Nigerian Pidgin connectives, we did not want to create a bias towards English connective cognates, as one of our research goals was to examine how discourse connectives emerge and to establish which connectives would not originate from English.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Discussion",
"sec_num": "7"
},
{
"text": "This study presents Naija-Lex, a lexicon of connectives in Nigerian Pidgin, which was constructed by combining automatic discourse parsing with manual annotation. In doing so, we showed that automatic discourse parsing is a valuable resource in constructing a discourse-annotated database to extract connectives. However, construction of a seed set based on a subset of the data and manual annotation of selected connectives and relation senses is necessary to obtain a reliable database. In addition, we made a first step to building a Nigerian-Pidgin discourse-annotated corpus by annotating ST connectives that were explicit in the TT. We will extend this corpus with implicit relations in further work.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "8"
},
{
"text": "https://universaldependencies.org/ treebanks/pcm_nsc/index.html",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "https://osf.io/xns9z/?view_only= 710a1eca318f46b9b10584c3b980beec",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Note that the difference between discourse marker and connective so was reflected in the Pidgin POS-tags. Discourse marker uses of so were mostly annotated as adverbs, whereas connective uses of so were mostly subordinating conjunctions.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "This amounts to less than 5% of the data for our corpus.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "To reduce noise in the corpus and the lexicon, we manually checked the relation senses that accounted for less than 5% of the cases for a given Pidgin connective.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "We are highly grateful to Emeka Felix Onwuegbuzia for his valuable insights on Nigerian Pidgin and evaluating a subset of the connectives. This research was funded by the German Research Foundation (DFG) as part of SFB 1102 \"Information Density and Linguistic Encoding\".",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgements",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "When and why do translators add connectives?: A corpus-based study",
"authors": [
{
"first": "Viktor",
"middle": [],
"last": "Becher",
"suffix": ""
}
],
"year": 2011,
"venue": "Target. International Journal of Translation Studies",
"volume": "23",
"issue": "1",
"pages": "26--47",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Viktor Becher. 2011. When and why do translators add connectives?: A corpus-based study. Target. Inter- national Journal of Translation Studies, 23(1):26- 47.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Toward a bilingual lexical database on connectives: Exploiting a German/Italian parallel corpus",
"authors": [
{
"first": "Peter",
"middle": [],
"last": "Bourgonje",
"suffix": ""
},
{
"first": "Yulia",
"middle": [],
"last": "Grishina",
"suffix": ""
},
{
"first": "Manfred",
"middle": [],
"last": "Stede",
"suffix": ""
}
],
"year": 2017,
"venue": "Proceedings of the Fourth Italian Conference on Computational Linguistics (CLiC-it)",
"volume": "",
"issue": "",
"pages": "1--6",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Peter Bourgonje, Yulia Grishina, and Manfred Stede. 2017. Toward a bilingual lexical database on con- nectives: Exploiting a German/Italian parallel cor- pus. In Proceedings of the Fourth Italian Confer- ence on Computational Linguistics (CLiC-it), pages 1-6, Rome, Italy.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Constructing a lexicon of Dutch discourse connectives",
"authors": [
{
"first": "Peter",
"middle": [],
"last": "Bourgonje",
"suffix": ""
},
{
"first": "Jet",
"middle": [],
"last": "Hoek",
"suffix": ""
},
{
"first": "Jacqueline",
"middle": [],
"last": "Evers-Vermeul",
"suffix": ""
},
{
"first": "Gisela",
"middle": [],
"last": "Redeker",
"suffix": ""
}
],
"year": 2018,
"venue": "Computational Linguistics in the Netherlands Journal",
"volume": "8",
"issue": "",
"pages": "163--175",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Peter Bourgonje, Jet Hoek, Jacqueline Evers-Vermeul, Gisela Redeker, Ted Sanders, and Manfred Stede. 2018. Constructing a lexicon of Dutch discourse connectives. Computational Linguistics in the Netherlands Journal, 8:163-175.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "A surface-syntactic UD treebank for Naija",
"authors": [
{
"first": "Bernard",
"middle": [],
"last": "Caron",
"suffix": ""
},
{
"first": "Marine",
"middle": [],
"last": "Courtin",
"suffix": ""
},
{
"first": "Kim",
"middle": [],
"last": "Gerdes",
"suffix": ""
},
{
"first": "Sylvain",
"middle": [],
"last": "Kahane",
"suffix": ""
}
],
"year": 2019,
"venue": "TLT 2019, Treebanks and Linguistic Theories",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Bernard Caron, Marine Courtin, Kim Gerdes, and Syl- vain Kahane. 2019. A surface-syntactic UD tree- bank for Naija. In TLT 2019, Treebanks and Lin- guistic Theories, Syntaxfest.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Unsupervised Pidgin text generation by pivoting English data and self-training",
"authors": [
{
"first": "Ernie",
"middle": [],
"last": "Chang",
"suffix": ""
},
{
"first": "David",
"middle": [
"Ifeoluwa"
],
"last": "Adelani",
"suffix": ""
},
{
"first": "Xiaoyu",
"middle": [],
"last": "Shen",
"suffix": ""
},
{
"first": "Vera",
"middle": [],
"last": "Demberg",
"suffix": ""
}
],
"year": 2020,
"venue": "International Conference on Learning Representations",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ernie Chang, David Ifeoluwa Adelani, Xiaoyu Shen, and Vera Demberg. 2020. Unsupervised Pidgin text generation by pivoting English data and self-training. In International Conference on Learning Represen- tations (ICLR).",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Establishing a language by annotating a corpus",
"authors": [
{
"first": "Marine",
"middle": [],
"last": "Courtin",
"suffix": ""
},
{
"first": "Bernard",
"middle": [],
"last": "Caron",
"suffix": ""
},
{
"first": "Kim",
"middle": [],
"last": "Gerdes",
"suffix": ""
},
{
"first": "Sylvain",
"middle": [],
"last": "Kahane",
"suffix": ""
}
],
"year": 2018,
"venue": "Annotation in Digital Humanities",
"volume": "2155",
"issue": "",
"pages": "7--11",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Marine Courtin, Bernard Caron, Kim Gerdes, and Syl- vain Kahane. 2018. Establishing a language by annotating a corpus. In annDH 2018 Annotation in Digital Humanities, volume 2155, pages 7-11. CEUR.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Discourse markers in speech: Characteristics and challenges for corpus annotation",
"authors": [
{
"first": "Ludivine",
"middle": [],
"last": "Crible",
"suffix": ""
},
{
"first": "Maria-Josep",
"middle": [],
"last": "Cuenca",
"suffix": ""
}
],
"year": 2017,
"venue": "Dialogue and Discourse",
"volume": "8",
"issue": "2",
"pages": "149--166",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ludivine Crible and Maria-Josep Cuenca. 2017. Dis- course markers in speech: Characteristics and chal- lenges for corpus annotation. Dialogue and Dis- course, 8(2):149-166.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "discourse verbs\" and discourse periphrastic links",
"authors": [
{
"first": "Laurence",
"middle": [],
"last": "Danlos",
"suffix": ""
}
],
"year": 2006,
"venue": "Proceedings of KON-VENS 2006",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Laurence Danlos. 2006. \"discourse verbs\" and dis- course periphrastic links. In Proceedings of KON- VENS 2006 (Konferenz zur Verarbeitung nat\u00fcrlicher Sprache) Universit\u00e4t Konstanz, page 160.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Constructing a lexicon of English discourse connectives",
"authors": [
{
"first": "Debopam",
"middle": [],
"last": "Das",
"suffix": ""
},
{
"first": "Tatjana",
"middle": [],
"last": "Scheffler",
"suffix": ""
},
{
"first": "Peter",
"middle": [],
"last": "Bourgonje",
"suffix": ""
},
{
"first": "Manfred",
"middle": [],
"last": "Stede",
"suffix": ""
}
],
"year": 2018,
"venue": "Proceedings of the 19th Annual SIGdial Meeting on Discourse and Dialogue",
"volume": "",
"issue": "",
"pages": "360--365",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Debopam Das, Tatjana Scheffler, Peter Bourgonje, and Manfred Stede. 2018. Constructing a lexicon of En- glish discourse connectives. In Proceedings of the 19th Annual SIGdial Meeting on Discourse and Dia- logue, pages 360-365.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "DiMLex-Bangla: A lexicon of Bangla discourse connectives",
"authors": [
{
"first": "Debopam",
"middle": [],
"last": "Das",
"suffix": ""
},
{
"first": "Manfred",
"middle": [],
"last": "Stede",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 12th Language Resources and Evaluation Conference",
"volume": "",
"issue": "",
"pages": "1097--1102",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Debopam Das, Manfred Stede, Soumya Sankar Ghosh, and Lahari Chatterjee. 2020. DiMLex-Bangla: A lexicon of Bangla discourse connectives. In Pro- ceedings of the 12th Language Resources and Eval- uation Conference, pages 1097-1102.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Contextualized embeddings for connective disambiguation in shallow discourse parsing",
"authors": [
{
"first": "Ren\u00e9",
"middle": [],
"last": "Knaebel",
"suffix": ""
},
{
"first": "Manfred",
"middle": [],
"last": "Stede",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the First Workshop on Computational Approaches to Discourse",
"volume": "",
"issue": "",
"pages": "65--75",
"other_ids": {
"DOI": [
"10.18653/v1/2020.codi-1.7"
]
},
"num": null,
"urls": [],
"raw_text": "Ren\u00e9 Knaebel and Manfred Stede. 2020. Contextual- ized embeddings for connective disambiguation in shallow discourse parsing. In Proceedings of the First Workshop on Computational Approaches to Discourse (CoDi 2020), pages 65-75, Online. As- sociation for Computational Linguistics.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Ted-MDB Lexicons: TrEn-ConnLex, PtEnConnLex",
"authors": [
{
"first": "Murathan",
"middle": [],
"last": "Kurfal\u0131",
"suffix": ""
},
{
"first": "Sibel",
"middle": [],
"last": "Ozer",
"suffix": ""
},
{
"first": "Deniz",
"middle": [],
"last": "Zeyrek",
"suffix": ""
},
{
"first": "Am\u00e1lia",
"middle": [],
"last": "Mendes",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the First Workshop on Computational Approaches to Discourse",
"volume": "",
"issue": "",
"pages": "148--153",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Murathan Kurfal\u0131, Sibel Ozer, Deniz Zeyrek, and Am\u00e1lia Mendes. 2020. Ted-MDB Lexicons: TrEn- ConnLex, PtEnConnLex. In Proceedings of the First Workshop on Computational Approaches to Discourse, pages 148-153.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Inducing discourse connectives from parallel texts",
"authors": [
{
"first": "Majid",
"middle": [],
"last": "Laali",
"suffix": ""
},
{
"first": "Leila",
"middle": [],
"last": "Kosseim",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of COLING 2014, the 25th International Conference on Computational Linguistics: Technical Papers",
"volume": "",
"issue": "",
"pages": "610--619",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Majid Laali and Leila Kosseim. 2014. Inducing dis- course connectives from parallel texts. In Proceed- ings of COLING 2014, the 25th International Con- ference on Computational Linguistics: Technical Pa- pers, pages 610-619.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "A PDTB-styled end-to-end discourse parser",
"authors": [
{
"first": "Ziheng",
"middle": [],
"last": "Lin",
"suffix": ""
},
{
"first": "Min-Yen",
"middle": [],
"last": "Hwee Tou Ng",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Kan",
"suffix": ""
}
],
"year": 2014,
"venue": "Natural Language Engineering",
"volume": "20",
"issue": "2",
"pages": "151--184",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ziheng Lin, Hwee Tou Ng, and Min-Yen Kan. 2014. A PDTB-styled end-to-end discourse parser. Natural Language Engineering, 20(2):151-184.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "LDM-PT: A Portuguese lexicon of discourse markers",
"authors": [
{
"first": "Am\u00e1lia",
"middle": [],
"last": "Mendes",
"suffix": ""
},
{
"first": "Pierre",
"middle": [],
"last": "Lejeune",
"suffix": ""
}
],
"year": 2016,
"venue": "Conference Handbook of TextLink-Structuring Discourse in Multilingual Europe Second Action Conference",
"volume": "",
"issue": "",
"pages": "89--92",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Am\u00e1lia Mendes and Pierre Lejeune. 2016. LDM-PT: A Portuguese lexicon of discourse markers. In Confer- ence Handbook of TextLink-Structuring Discourse in Multilingual Europe Second Action Conference, pages 89-92. Debrecen University Press.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Tony: Contextual embeddings for accurate multilingual discourse segmentation of full documents",
"authors": [
{
"first": "Philippe",
"middle": [],
"last": "Muller",
"suffix": ""
},
{
"first": "Chlo\u00e9",
"middle": [],
"last": "Braud",
"suffix": ""
},
{
"first": "Mathieu",
"middle": [],
"last": "Morey",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of the Workshop on Discourse Relation Parsing and Treebanking",
"volume": "",
"issue": "",
"pages": "115--124",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Philippe Muller, Chlo\u00e9 Braud, and Mathieu Morey. 2019. Tony: Contextual embeddings for accurate multilingual discourse segmentation of full docu- ments. In Proceedings of the Workshop on Dis- course Relation Parsing and Treebanking 2019, pages 115-124. Association for Computational Lin- guistics.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Pidgi-nUNMT: Unsupervised neural machine translation from West African Pidgin to English",
"authors": [
{
"first": "Kelechi",
"middle": [],
"last": "Ogueji",
"suffix": ""
},
{
"first": "Orevaoghene",
"middle": [],
"last": "Ahia",
"suffix": ""
}
],
"year": 2019,
"venue": "NeurIPS 2019 Workshop on Machine Learning for the Developing World (ML4D)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kelechi Ogueji and Orevaoghene Ahia. 2019. Pidgi- nUNMT: Unsupervised neural machine translation from West African Pidgin to English. In NeurIPS 2019 Workshop on Machine Learning for the Devel- oping World (ML4D).",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Using syntax to disambiguate explicit discourse connectives in text",
"authors": [
{
"first": "Emily",
"middle": [],
"last": "Pitler",
"suffix": ""
},
{
"first": "Ani",
"middle": [],
"last": "Nenkova",
"suffix": ""
}
],
"year": 2009,
"venue": "Proceedings of the ACL-IJCNLP 2009 Conference Short Papers",
"volume": "",
"issue": "",
"pages": "13--16",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Emily Pitler and Ani Nenkova. 2009. Using syntax to disambiguate explicit discourse connectives in text. In Proceedings of the ACL-IJCNLP 2009 Confer- ence Short Papers, pages 13-16, Suntec, Singapore. Association for Computational Linguistics.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "The Penn Discourse TreeBank 2.0",
"authors": [
{
"first": "Rashmi",
"middle": [],
"last": "Prasad",
"suffix": ""
},
{
"first": "Nikhil",
"middle": [],
"last": "Dinesh",
"suffix": ""
},
{
"first": "Alan",
"middle": [],
"last": "Lee",
"suffix": ""
},
{
"first": "Eleni",
"middle": [],
"last": "Miltsakaki",
"suffix": ""
},
{
"first": "Livio",
"middle": [],
"last": "Robaldo",
"suffix": ""
},
{
"first": "Aravind",
"middle": [],
"last": "Joshi",
"suffix": ""
},
{
"first": "Bonnie",
"middle": [],
"last": "Webber",
"suffix": ""
}
],
"year": 2008,
"venue": "Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08)",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Milt- sakaki, Livio Robaldo, Aravind Joshi, and Bon- nie Webber. 2008. The Penn Discourse TreeBank 2.0. In Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC'08), Marrakech, Morocco. European Lan- guage Resources Association (ELRA).",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "Identifying discourse connectives in biomedical text",
"authors": [
{
"first": "Ramesh",
"middle": [],
"last": "Balaji Polepalli",
"suffix": ""
},
{
"first": "Hong",
"middle": [],
"last": "Yu",
"suffix": ""
}
],
"year": 2010,
"venue": "AMIA Annual Symposium Proceedings",
"volume": "2010",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Balaji Polepalli Ramesh and Hong Yu. 2010. Iden- tifying discourse connectives in biomedical text. In AMIA Annual Symposium Proceedings, volume 2010, page 657. American Medical Informatics As- sociation.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Annotating discourse relations in spoken language: A comparison of the PDTB and CCR frameworks",
"authors": [
{
"first": "Ines",
"middle": [],
"last": "Rehbein",
"suffix": ""
},
{
"first": "Merel",
"middle": [],
"last": "Scholman",
"suffix": ""
},
{
"first": "Vera",
"middle": [],
"last": "Demberg",
"suffix": ""
}
],
"year": 2016,
"venue": "Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC'16)",
"volume": "",
"issue": "",
"pages": "1039--1046",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ines Rehbein, Merel Scholman, and Vera Demberg. 2016. Annotating discourse relations in spoken language: A comparison of the PDTB and CCR frameworks. In Proceedings of the Tenth Inter- national Conference on Language Resources and Evaluation (LREC'16), pages 1039-1046, Portoro\u017e, Slovenia. European Language Resources Associa- tion (ELRA).",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Corpusdriven semantics of concession: Where do expectations come from?",
"authors": [
{
"first": "Livio",
"middle": [],
"last": "Robaldo",
"suffix": ""
},
{
"first": "Eleni",
"middle": [],
"last": "Miltsakaki",
"suffix": ""
}
],
"year": 2014,
"venue": "Dialogue & Discourse",
"volume": "5",
"issue": "1",
"pages": "1--36",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Livio Robaldo and Eleni Miltsakaki. 2014. Corpus- driven semantics of concession: Where do expecta- tions come from? Dialogue & Discourse, 5(1):1- 36.",
"links": null
},
"BIBREF22": {
"ref_id": "b22",
"title": "Lexconn: a french lexicon of discourse connectives. Discours. Revue de linguistique, psycholinguistique et informatique. A journal of linguistics, psycholinguistics and computational linguistics",
"authors": [
{
"first": "Charlotte",
"middle": [],
"last": "Roze",
"suffix": ""
},
{
"first": "Laurence",
"middle": [],
"last": "Danlos",
"suffix": ""
},
{
"first": "Philippe",
"middle": [],
"last": "Muller",
"suffix": ""
}
],
"year": 2012,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Charlotte Roze, Laurence Danlos, and Philippe Muller. 2012. Lexconn: a french lexicon of discourse connectives. Discours. Revue de linguistique, psy- cholinguistique et informatique. A journal of linguis- tics, psycholinguistics and computational linguistics, (10).",
"links": null
},
"BIBREF23": {
"ref_id": "b23",
"title": "Comparison of methods for explicit discourse connective identification across various domains",
"authors": [
{
"first": "C",
"middle": [
"J"
],
"last": "Merel",
"suffix": ""
},
{
"first": "Tianai",
"middle": [],
"last": "Scholman",
"suffix": ""
},
{
"first": "Frances",
"middle": [],
"last": "Dong",
"suffix": ""
},
{
"first": "Vera",
"middle": [],
"last": "Yung",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Demberg",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the First Workshop on Computational Approaches to Discourse",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Merel C.J. Scholman, Tianai Dong, Frances Yung, and Vera Demberg. 2021. Comparison of methods for explicit discourse connective identification across various domains. In Proceedings of the First Work- shop on Computational Approaches to Discourse (CoDi 2020).",
"links": null
},
"BIBREF24": {
"ref_id": "b24",
"title": "Coding coherence relations: Reliability and validity",
"authors": [
{
"first": "P",
"middle": [
"M S"
],
"last": "Wilbert",
"suffix": ""
},
{
"first": "Liesbeth",
"middle": [],
"last": "Spooren",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Degand",
"suffix": ""
}
],
"year": 2010,
"venue": "Corpus Linguistics and Linguistic Theory",
"volume": "6",
"issue": "2",
"pages": "241--266",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Wilbert P. M. S. Spooren and Liesbeth Degand. 2010. Coding coherence relations: Reliability and validity. Corpus Linguistics and Linguistic Theory, 6(2):241- 266.",
"links": null
},
"BIBREF25": {
"ref_id": "b25",
"title": "Connective-Lex: A web-based multilingual lexical resource for connectives. Discours. Revue de linguistique, psycholinguistique et informatique. A journal of linguistics, psycholinguistics and computational linguistics",
"authors": [
{
"first": "Manfred",
"middle": [],
"last": "Stede",
"suffix": ""
},
{
"first": "Tatjana",
"middle": [],
"last": "Scheffler",
"suffix": ""
},
{
"first": "Am\u00e1lia",
"middle": [],
"last": "Mendes",
"suffix": ""
}
],
"year": 2018,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Manfred Stede, Tatjana Scheffler, and Am\u00e1lia Mendes. 2018. Connective-Lex: A web-based multilingual lexical resource for connectives. Discours. Revue de linguistique, psycholinguistique et informatique. A journal of linguistics, psycholinguistics and com- putational linguistics, (24).",
"links": null
},
"BIBREF26": {
"ref_id": "b26",
"title": "DiMLex: A lexicon of discourse markers for text generation and understanding",
"authors": [
{
"first": "Manfred",
"middle": [],
"last": "Stede",
"suffix": ""
},
{
"first": "Carla",
"middle": [],
"last": "Umbach",
"suffix": ""
}
],
"year": 1998,
"venue": "36th Annual Meeting of the Association for Computational Linguistics and 17th International Conference on Computational Linguistics",
"volume": "2",
"issue": "",
"pages": "1238--1242",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Manfred Stede and Carla Umbach. 1998. DiMLex: A lexicon of discourse markers for text generation and understanding. In 36th Annual Meeting of the Asso- ciation for Computational Linguistics and 17th In- ternational Conference on Computational Linguis- tics, Volume 2, pages 1238-1242.",
"links": null
},
"BIBREF27": {
"ref_id": "b27",
"title": "Between simplification and complexification: Nonstandard varieties of English around the world",
"authors": [
{
"first": "Benedikt",
"middle": [],
"last": "Szmrecsanyi",
"suffix": ""
},
{
"first": "Bernd",
"middle": [],
"last": "Kortmann",
"suffix": ""
},
{
"first": "Geoffrey",
"middle": [],
"last": "Sampson",
"suffix": ""
},
{
"first": "David",
"middle": [],
"last": "Gil",
"suffix": ""
},
{
"first": "Peter",
"middle": [],
"last": "Trudgill",
"suffix": ""
}
],
"year": 2009,
"venue": "Language complexity as an evolving variable",
"volume": "",
"issue": "",
"pages": "64--79",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Benedikt Szmrecsanyi, Bernd Kortmann, Geoffrey Sampson, David Gil, and Peter Trudgill. 2009. Be- tween simplification and complexification: Non- standard varieties of English around the world. In Language complexity as an evolving variable, pages 64-79. Oxford University Press.",
"links": null
},
"BIBREF28": {
"ref_id": "b28",
"title": "Discovery of ambiguous and unambiguous discourse connectives via annotation projection",
"authors": [
{
"first": "Yannick",
"middle": [],
"last": "Versley",
"suffix": ""
}
],
"year": 2010,
"venue": "Proceedings of Workshop on Annotation and Exploitation of Parallel Corpora (AEPC)",
"volume": "",
"issue": "",
"pages": "83--82",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yannick Versley. 2010. Discovery of ambiguous and unambiguous discourse connectives via annotation projection. In Proceedings of Workshop on Annota- tion and Exploitation of Parallel Corpora (AEPC), pages 83-82.",
"links": null
},
"BIBREF29": {
"ref_id": "b29",
"title": "A refined endto-end discourse parser",
"authors": [
{
"first": "Jianxiang",
"middle": [],
"last": "Wang",
"suffix": ""
},
{
"first": "Man",
"middle": [],
"last": "Lan",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the Nineteenth Conference on Computational Natural Language Learning -Shared Task",
"volume": "",
"issue": "",
"pages": "17--24",
"other_ids": {
"DOI": [
"10.18653/v1/K15-2002"
]
},
"num": null,
"urls": [],
"raw_text": "Jianxiang Wang and Man Lan. 2015. A refined end- to-end discourse parser. In Proceedings of the Nine- teenth Conference on Computational Natural Lan- guage Learning -Shared Task, pages 17-24, Beijing, China. Association for Computational Linguistics.",
"links": null
},
"BIBREF30": {
"ref_id": "b30",
"title": "CoNLL 2016 shared task on multilingual shallow discourse parsing. Proceedings of the CoNLL-16 shared task",
"authors": [
{
"first": "Nianwen",
"middle": [],
"last": "Xue",
"suffix": ""
},
{
"first": "Sameer",
"middle": [],
"last": "Hwee Tou Ng",
"suffix": ""
},
{
"first": "Attapol",
"middle": [],
"last": "Pradhan",
"suffix": ""
},
{
"first": "Bonnie",
"middle": [],
"last": "Rutherford",
"suffix": ""
},
{
"first": "Chuan",
"middle": [],
"last": "Webber",
"suffix": ""
},
{
"first": "Hongmin",
"middle": [],
"last": "Wang",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Wang",
"suffix": ""
}
],
"year": 2016,
"venue": "",
"volume": "",
"issue": "",
"pages": "1--19",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Nianwen Xue, Hwee Tou Ng, Sameer Pradhan, At- tapol Rutherford, Bonnie Webber, Chuan Wang, and Hongmin Wang. 2016. CoNLL 2016 shared task on multilingual shallow discourse parsing. Proceed- ings of the CoNLL-16 shared task, pages 1-19.",
"links": null
},
"BIBREF31": {
"ref_id": "b31",
"title": "Gum-Drop at the DISRPT2019 shared task: A model stacking approach to discourse unit segmentation and connective detection",
"authors": [
{
"first": "Yue",
"middle": [],
"last": "Yu",
"suffix": ""
},
{
"first": "Yilun",
"middle": [],
"last": "Zhu",
"suffix": ""
},
{
"first": "Yang",
"middle": [],
"last": "Liu",
"suffix": ""
},
{
"first": "Yan",
"middle": [],
"last": "Liu",
"suffix": ""
},
{
"first": "Siyao",
"middle": [],
"last": "Peng",
"suffix": ""
},
{
"first": "Mackenzie",
"middle": [],
"last": "Gong",
"suffix": ""
},
{
"first": "Amir",
"middle": [],
"last": "Zeldes",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of Discourse Relation Parsing and Treebanking (DIS-RPT2019)",
"volume": "",
"issue": "",
"pages": "133--143",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yue Yu, Yilun Zhu, Yang Liu, Yan Liu, Siyao Peng, Mackenzie Gong, and Amir Zeldes. 2019. Gum- Drop at the DISRPT2019 shared task: A model stacking approach to discourse unit segmentation and connective detection. In Proceedings of Dis- course Relation Parsing and Treebanking (DIS- RPT2019), pages 133-143.",
"links": null
},
"BIBREF32": {
"ref_id": "b32",
"title": "The DIS-RPT 2019 shared task on elementary discourse unit segmentation and connective detection",
"authors": [
{
"first": "Amir",
"middle": [],
"last": "Zeldes",
"suffix": ""
},
{
"first": "Debopam",
"middle": [],
"last": "Das",
"suffix": ""
},
{
"first": "Erick",
"middle": [
"Galani"
],
"last": "Maziero",
"suffix": ""
},
{
"first": "Juliano",
"middle": [],
"last": "Antonio",
"suffix": ""
},
{
"first": "Mikel",
"middle": [],
"last": "Iruskieta",
"suffix": ""
}
],
"year": 2019,
"venue": "Proceedings of the Workshop on Discourse Relation Parsing and Treebanking",
"volume": "",
"issue": "",
"pages": "97--104",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Amir Zeldes, Debopam Das, Erick Galani Maziero, Ju- liano Antonio, and Mikel Iruskieta. 2019. The DIS- RPT 2019 shared task on elementary discourse unit segmentation and connective detection. In Proceed- ings of the Workshop on Discourse Relation Parsing and Treebanking 2019, pages 97-104.",
"links": null
},
"BIBREF33": {
"ref_id": "b33",
"title": "PDTB-style discourse annotation of Chinese text",
"authors": [
{
"first": "Yuping",
"middle": [],
"last": "Zhou",
"suffix": ""
},
{
"first": "Nianwen",
"middle": [],
"last": "Xue",
"suffix": ""
}
],
"year": 2012,
"venue": "Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics",
"volume": "1",
"issue": "",
"pages": "69--77",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yuping Zhou and Nianwen Xue. 2012. PDTB-style dis- course annotation of Chinese text. In Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 69-77.",
"links": null
},
"BIBREF34": {
"ref_id": "b34",
"title": "The Chinese Discourse TreeBank: A Chinese corpus annotated with discourse relations. Language Resources and Evaluation",
"authors": [
{
"first": "Yuping",
"middle": [],
"last": "Zhou",
"suffix": ""
},
{
"first": "Nianwen",
"middle": [],
"last": "Xue",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "49",
"issue": "",
"pages": "397--431",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Yuping Zhou and Nianwen Xue. 2015. The Chinese Discourse TreeBank: A Chinese corpus annotated with discourse relations. Language Resources and Evaluation, 49(2):397-431.",
"links": null
},
"BIBREF35": {
"ref_id": "b35",
"title": "A multifactorial analysis of explicitation in translation. Target",
"authors": [
{
"first": "Sandrine",
"middle": [],
"last": "Zufferey",
"suffix": ""
},
{
"first": "Bruno",
"middle": [],
"last": "Cartoni",
"suffix": ""
}
],
"year": 2014,
"venue": "International Journal of Translation Studies",
"volume": "26",
"issue": "3",
"pages": "361--384",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sandrine Zufferey and Bruno Cartoni. 2014. A multi- factorial analysis of explicitation in translation. Tar- get. International Journal of Translation Studies, 26(3):361-384.",
"links": null
},
"BIBREF36": {
"ref_id": "b36",
"title": "Empirical validations of multilingual annotation schemes for discourse relations",
"authors": [
{
"first": "Sandrine",
"middle": [],
"last": "Zufferey",
"suffix": ""
},
{
"first": "Liesbeth",
"middle": [],
"last": "Degand",
"suffix": ""
},
{
"first": "Andrei",
"middle": [],
"last": "Popescu-Belis",
"suffix": ""
},
{
"first": "Ted",
"middle": [
"J M"
],
"last": "Sanders",
"suffix": ""
}
],
"year": 2012,
"venue": "Proceedings of the 8th Joint ISO-ACL/SIGSEM Workshop on Interoperable Semantic Annotation (ISA-8)",
"volume": "",
"issue": "",
"pages": "77--84",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Sandrine Zufferey, Liesbeth Degand, Andrei Popescu- Belis, and Ted J. M. Sanders. 2012. Empirical val- idations of multilingual annotation schemes for dis- course relations. In Proceedings of the 8th Joint ISO- ACL/SIGSEM Workshop on Interoperable Semantic Annotation (ISA-8), pages 77-84, Jeju Island, Korea.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"text": "Schematic overview of the workflow for the construction of the connective lexicon (and discourseannotated corpus) for Nigerian Pidgin.",
"uris": null,
"num": null,
"type_str": "figure"
},
"FIGREF1": {
"text": "Relation sense distribution of the twelve most frequent Nigerian Pidgin connectives.",
"uris": null,
"num": null,
"type_str": "figure"
},
"TABREF0": {
"text": "",
"html": null,
"type_str": "table",
"content": "<table><tr><td>provides information regarding the rela-</td></tr><tr><td>tion sense distribution found in the corpus, based on</td></tr><tr><td>projected labels originally assigned to the English</td></tr><tr><td>connectives. Connectives in Nigerian Pidgin were</td></tr><tr><td>used to express a total of eight different relation</td></tr></table>",
"num": null
}
}
}
} |