File size: 98,237 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
{
    "paper_id": "I13-1044",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T07:14:52.218764Z"
    },
    "title": "Weasels, Hedges and Peacocks: Discourse-level Uncertainty in Wikipedia Articles",
    "authors": [
        {
            "first": "Veronika",
            "middle": [],
            "last": "Vincze",
            "suffix": "",
            "affiliation": {
                "laboratory": "Research Group on Artificial Intelligence",
                "institution": "Hungarian Academy of Sciences",
                "location": {}
            },
            "email": "vinczev@inf.u-szeged.hu"
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "Uncertainty is an important linguistic phenomenon that is relevant in many areas of language processing. While earlier research mostly concentrated on the semantic aspects of uncertainty, here we focus on discourse-and pragmaticsrelated aspects of uncertainty. We present a classification of such linguistic phenomena and introduce a corpus of Wikipedia articles in which the presented types of discourse-level uncertainty-weasel, hedge and peacock-have been manually annotated. We also discuss some experimental results on discourse-level uncertainty detection.",
    "pdf_parse": {
        "paper_id": "I13-1044",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "Uncertainty is an important linguistic phenomenon that is relevant in many areas of language processing. While earlier research mostly concentrated on the semantic aspects of uncertainty, here we focus on discourse-and pragmaticsrelated aspects of uncertainty. We present a classification of such linguistic phenomena and introduce a corpus of Wikipedia articles in which the presented types of discourse-level uncertainty-weasel, hedge and peacock-have been manually annotated. We also discuss some experimental results on discourse-level uncertainty detection.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "In many areas of natural language processing, it is essential to distinguish between factual and nonfactual information. Thus, depending on the precise task, negated or uncertain propositions should be treated separately by e.g. information extraction systems or they should be neglected. For instance, in medicine, if it is uncertain whether the patient suffers from an illness, the doctor should undertake further examinations to determine the final diagnosis. In another case, only those pieces of news are relevant in news media that are true and come from a reliable source. Uncertain information or unreliable sources should not be part of the news. In order to be able to find uncertain propositions in a huge amount of texts, a reliable uncertainty detector is needed, which can be only developed if annotated resources are at hand.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Previous studies on uncertainty detection concentrated mostly on the semantic dimensions. Indeed, in many cases it is the lexical content (meaning) of the uncertainty marker (cue) that is responsible for uncertainty, i.e. it can be identified in texts with the help of semantic tools. However, there are other types of uncertainty which cannot be described by just concentrating on semantics. For instance, many may denote quite different approximations: in the sentence Many of the students did not read the book, many may signal about 60-70% of the students (or at least more than 50%), while in This airline loses many suitcases, many may be only 20% but this number is still high enough for passengers to call it many. Here, the context and world knowledge determine how the quantifier many should be interpreted.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Here, we will focus on pragmatics-and discourse-related aspects of uncertainty. We will examine the concepts of source, fuzziness and subjectivity and their connection with uncertainty. As a first contribution, we will present a languageindependent classification of such linguistic phenomena. As another contribution, we will introduce a corpus of Wikipedia articles in which linguistic cues of the presented types of discourselevel uncertainty have been manually annotated, hence empirical data on the frequency of such phenomena can also be provided. We will report the results of our experiments and we will also compare them with those of previous studies.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1"
            },
            {
                "text": "Different concepts and terms that are related to uncertainty phenomena are employed. Modality is usually associated with uncertainty (Palmer, 1986) , but the terms factuality (Saur\u00ed and Pustejovsky, 2012) , veridicality (de Marneffe et al., 2012) , evidentiality (Aikhenvald, 2004) and commitment (Diab et al., 2009) are also used. They all represent related but slightly different linguistic phenomena, which lie mostly in the category of semantic uncertainty. Propositions can be uncertain at the semantic level, that is, their truth value cannot be determined just given the speaker's mental state. Szarvas et al. (2012) offer a classi-fication of semantic uncertainty phenomena.",
                "cite_spans": [
                    {
                        "start": 133,
                        "end": 147,
                        "text": "(Palmer, 1986)",
                        "ref_id": "BIBREF17"
                    },
                    {
                        "start": 175,
                        "end": 204,
                        "text": "(Saur\u00ed and Pustejovsky, 2012)",
                        "ref_id": "BIBREF22"
                    },
                    {
                        "start": 220,
                        "end": 246,
                        "text": "(de Marneffe et al., 2012)",
                        "ref_id": "BIBREF6"
                    },
                    {
                        "start": 263,
                        "end": 281,
                        "text": "(Aikhenvald, 2004)",
                        "ref_id": null
                    },
                    {
                        "start": 297,
                        "end": 316,
                        "text": "(Diab et al., 2009)",
                        "ref_id": "BIBREF7"
                    },
                    {
                        "start": 602,
                        "end": 623,
                        "text": "Szarvas et al. (2012)",
                        "ref_id": "BIBREF23"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discourse-level Uncertainty",
                "sec_num": "2"
            },
            {
                "text": "Here, we use the term uncertainty similar to Szarvas et al. (2012) , who aimed at giving a unified framework for the above-mentioned phenomena: \"uncertain propositions are those [...] whose truth value or reliability cannot be determined due to lack of information\". They contrast semantic uncertainty with discourse-level uncertainty: if the scheme \"cue x but it is certain that not x\" is invalid (where x denotes a proposition, and cue denotes an uncertainty cue), that is, an uncertain proposition and its negated version cannot be coordinated, it is an instance of semantic uncertainty (e.g. ##It may be raining in New York but it is certain that it is not raining in New York).",
                "cite_spans": [
                    {
                        "start": 45,
                        "end": 66,
                        "text": "Szarvas et al. (2012)",
                        "ref_id": "BIBREF23"
                    },
                    {
                        "start": 178,
                        "end": 183,
                        "text": "[...]",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discourse-level Uncertainty",
                "sec_num": "2"
            },
            {
                "text": "Besides semantic uncertainty, uncertainty can be found at the level of discourse as well. Here, the missing or intentionally omitted information is not related to the propositional content of the utterance but to other factors. In contrast to semantic uncertainty (Szarvas et al., 2012) , the truth value of such propositions can be determined, but uncertainty arises if the proposition is analyzed in detail. For instance, the sentence Some people are running evokes questions like Who exactly are those people that are running? Here, the answer usually depends on the context, the speaker and the discourse and it cannot be determined out of context, thus henceforth such phenomena will be labeled discourse-level uncertainty.",
                "cite_spans": [
                    {
                        "start": 264,
                        "end": 286,
                        "text": "(Szarvas et al., 2012)",
                        "ref_id": "BIBREF23"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discourse-level Uncertainty",
                "sec_num": "2"
            },
            {
                "text": "We will carefully analyze discourse-level uncertainty phenomena below which are named after their most typical linguistic markers, i.e. cues. Although for the sake of simplicity we only provide English examples here, our categorization is based on pragmatic and cognitive considerations, and we will implicitly assume that our categories are language-independent. We will focus on Wikipedia articles, which -as indicated by previous studies (Ganter and Strube, 2009; Farkas et al., 2010 ) -seem to contain a certain amount of uncertainty phenomena like this. We will concentrate on three key aspects of discourse-level uncertainty, namely, sources, fuzziness and subjectivity.",
                "cite_spans": [
                    {
                        "start": 441,
                        "end": 466,
                        "text": "(Ganter and Strube, 2009;",
                        "ref_id": "BIBREF9"
                    },
                    {
                        "start": 467,
                        "end": 486,
                        "text": "Farkas et al., 2010",
                        "ref_id": "BIBREF8"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discourse-level Uncertainty",
                "sec_num": "2"
            },
            {
                "text": "The notion of source is important for deciding the reliability of information conveyed (Saur\u00ed and Pustejovsky, 2012; Wiebe et al., 2005; Nawaz et al., 2010) . It is not a matter of indifference to whom the information / opinion belongs to, espe-cially in news media: people are more likely to believe a statement if it is communicated by a reliable source as opposed to a piece of sourceless information. In the public mind, experts, scientists, ministers, etc. are viewed as credible sources (cf. Bell (1991) ) while unnamed or unidentifiable sources are considered less reliable. If some pieces of information are backed by a credible source, they are more likely to be treated as trustworthy, however, sourceless information is given less credence.",
                "cite_spans": [
                    {
                        "start": 87,
                        "end": 116,
                        "text": "(Saur\u00ed and Pustejovsky, 2012;",
                        "ref_id": "BIBREF22"
                    },
                    {
                        "start": 117,
                        "end": 136,
                        "text": "Wiebe et al., 2005;",
                        "ref_id": "BIBREF26"
                    },
                    {
                        "start": 137,
                        "end": 156,
                        "text": "Nawaz et al., 2010)",
                        "ref_id": "BIBREF15"
                    },
                    {
                        "start": 498,
                        "end": 509,
                        "text": "Bell (1991)",
                        "ref_id": "BIBREF2"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "Events with no obvious sources are called weasels in Wikipedia 1 (Ganter and Strube, 2009) : their source is missing or is specified only vaguely or too generally, hence, it cannot be exactly determined who the holder of the opinion is (undetermined source) as it is either not expressed or expressed by an indefinite noun phrase. Weasel sentences usually invoke questions like Who said that? and Who thinks that? The following sentence illustrates this:",
                "cite_spans": [
                    {
                        "start": 65,
                        "end": 90,
                        "text": "(Ganter and Strube, 2009)",
                        "ref_id": "BIBREF9"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "Some have claimed that Bush would have actually increased his lead if state wide recounts had taken place.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "The ultimate source of the proposition expressed in the embedded sentence is not known since it is denoted by the pronoun some. Thus, it is not known who provided the opinion and therefore it is uncertain whether this is an important (reliable) piece of information (e.g. the opinion of experts) or whether it should be ignored.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "Passive constructions which do not express the agent comprise a special type of weasels:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "It has been suggested [by whom?] that he should have involved Clinton much more heavily in his campaign.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "The sentence does not reveal who has suggested the involvement of Clinton in the campaign. Hence, the source of the information is unclear and the source is missing from the sentence.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "The basic idea behind weasel phenomena is the lack of a reference: it is not known who the source of the opinion is. This view is supported by the fact that a weasel candidate ceases to be uncertain if it is enhanced by citations:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "Most authors now prefer to place it within the genus Pezoporus, e.g. Leeton et al. (1998) .",
                "cite_spans": [
                    {
                        "start": 69,
                        "end": 89,
                        "text": "Leeton et al. (1998)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "The phrase most authors would indicate a weasel (it is not clear whose opinion is this) but the citation at the end of the sentence clearly identifies the source.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "In this paper, we extend the original notion of weasel and we argue that propositions that have an underspecified argument that would be relevant or is not common knowledge in the situation can be also viewed as weasels. Thus, a proposition is considered to be an instance of weasel if any of its relevant arguments is underspecified, i.e. it evokes questions like Who/what exactly? Which? Here, we give an example:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "While the Skyraider is not as iconic as some other aircraft, it has been featured in some Vietnam-era films such as The Green Berets (1968) and Flight of the Intruder (1991).",
                "cite_spans": [
                    {
                        "start": 126,
                        "end": 139,
                        "text": "Berets (1968)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "The sentence does not determine what kind of aircraft is considered iconic, so it is a vague or underspecified statement: we only know that there are \"iconic aircraft\", but no more details are specified. Again, the weasel type of uncertainty is expressed here by the adjectives some and other. Note that there is another occurrence of the word some in the sentence, but it does not denote any uncertainty in this case since the relevant Vietnamera films are then listed.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Weasels",
                "sec_num": "2.1"
            },
            {
                "text": "Another type of discourse-level uncertainty that will be discussed later on is called a hedge. Although a lot of studies used the term hedge, it may denote different linguistic phenomena for different authors. For instance, hedge means mostly speculation in the biomedical domain (see e.g. Medlock and Briscoe (2007) , Vincze et al. (2008) , and Farkas et al. (2010)). When contrasting epistemic modality and hedging, Rizomilioti (2006) categorizes approximators, passive voice and attribution to unnamed sources, among others, as instances of hedging and Hyland (1996) also cites them among common hedging devices.",
                "cite_spans": [
                    {
                        "start": 290,
                        "end": 316,
                        "text": "Medlock and Briscoe (2007)",
                        "ref_id": "BIBREF13"
                    },
                    {
                        "start": 319,
                        "end": 339,
                        "text": "Vincze et al. (2008)",
                        "ref_id": "BIBREF25"
                    },
                    {
                        "start": 556,
                        "end": 569,
                        "text": "Hyland (1996)",
                        "ref_id": "BIBREF10"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "Here, we understand hedge in the sense introduced by Lakoff (1973) . For him, hedges are \"words whose job is to make things fuzzier or less fuzzy\", that is, the exact meaning of some qualities or quantities is blurred by them. Intensifiers (very, much), deintensifiers (a bit, less) and circumscribers (approximately) also belong to this group. Their effect is to add uncertainty to some elements in the proposition: they shift the value of some quality / quantity and the truth value of the proposition can only be decided if it is known what the reference point in the discourse is as the following example shows: Specialized services will very often provide a much more reliable service based on trusted publications.",
                "cite_spans": [
                    {
                        "start": 53,
                        "end": 66,
                        "text": "Lakoff (1973)",
                        "ref_id": "BIBREF12"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "In this sentence, there are several hedge cues. First, there is often, which informs us that it is not always the case that specialized services provide much more reliable service. It is modified by the intensifier very, which indicates that it is almost always the case (but still not always). Next, their service is much more reliable than any other service (at least those relevant in the context), that is, it is very reliable.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "However, it should be noted that there is no absolute way to determine the truth value of this proposition without agreeing on what is meant by e.g. often: for now, let us say that often means at least seven out of ten times (but not ten times out of ten) and then very often may denote eight or nine times out of ten. It depends on the context, the speakers and the event described in the sentence to determine the reference point according to which the quantity or quality of events or entities can be evaluated. In the above example, the reference point may be 70%, and intensifiers denote that the quality or frequency of the event / entity is above the reference point, in this case, above 70%. Deintensifiers, however, assert that the quality or frequency is below the reference point.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "Circumscribers -as their name states -circumscribe the exact amount or quality of the event or entity, which can be above or below the reference point. To represent this visually, they denote a set around the reference point in which the exact amount or quality is situated (see Figure 1 below). Here are some linguistic examples:",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 279,
                        "end": 287,
                        "text": "Figure 1",
                        "ref_id": "FIGREF0"
                    }
                ],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "This may explain why it has a lower than average estimated albedo of\u02dc0.03.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "The duration of attacks averages 3-7 days.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "It is interesting to note that in such cases not only cue words but also cue characters are responsible for uncertainty: the tilde and hyphen in these specific cases. Moreover, there are cue words that function as circumscribers as well like approximately and another use of some:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "Amsterdam Zuidoost has approximately 86,000 inhabitants and consists of some 38,000 houses. Figure 1 shows the relationship of hedge types relative to the reference point. Thus, each type of hedge denotes a set in which the exact amount, quality or frequency of the relevant event or entity is situated but its exact place remains unclear.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 92,
                        "end": 100,
                        "text": "Figure 1",
                        "ref_id": "FIGREF0"
                    }
                ],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "Hedging is also one of the politeness strategies mentioned by Brown and Levinson (1987) : they may function as mitigators in order to minimize disagreement, and to acknowledge that the speaker is imposing a task on the hearer. In the request Could you please sort of correct this very short text for me? the phrase sort of is a hedge, and the \"very short\" text may in fact be rather long. Here, hedges have pragmatic functions and they do not refer to uncertainty.",
                "cite_spans": [
                    {
                        "start": 62,
                        "end": 87,
                        "text": "Brown and Levinson (1987)",
                        "ref_id": "BIBREF3"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Hedges",
                "sec_num": "2.2"
            },
            {
                "text": "Subjectivity by its very nature contains aspects of uncertainty. People's opinions may differ from each other concerning specific things or events: they do not necessarily agree on what is good, neutral or bad. Thus, we cannot unequivocally determine what is good or what is bad.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Peacocks",
                "sec_num": "2.3"
            },
            {
                "text": "Words that express unprovable qualifications or exaggerations are called peacock by Wikipedia editors. 2 Their meaning often inherently contain positive or negative subjective judgments, that is, they are polar expressions. Peacock terms include brilliant, excellent and best-known. Although their usage may be acceptable in other contexts, the objective style of Wikipedia editing requires that peacocks should be avoided.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Peacocks",
                "sec_num": "2.3"
            },
            {
                "text": "2 http://en.wikipedia.org/wiki/ Wikipedia:Manual_of_Style/Words_to_watch Although they are not called peacocks by Wikipedia editors, we classify other subjective elements as peacocks as well. For instance, editorial remarks that refer to the subjective opinion of the author of the article (like ironically and unfortunately) or contentious labels (controversial and legendary) may all express subjectivity in certain contexts, hence we treat them here as peacock terms. The uncertainty in their meaning again lies in the fact that it cannot be objectively judged what can be called excellent for instance -it can be only deduced from discourse or contextual information and it may differ from speaker to speaker.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Peacocks",
                "sec_num": "2.3"
            },
            {
                "text": "Here is a sentence with some peacock terms:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Peacocks",
                "sec_num": "2.3"
            },
            {
                "text": "Through the ardent efforts of Rozsnyai, the Philharmonia Hungarica quickly matured into one of Europe's most distinguished orchestras.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Peacocks",
                "sec_num": "2.3"
            },
            {
                "text": "The words ardent and most distinguished are clearly positive in polarity, and again it cannot be objectively decided what level of enthusiasm is called ardent or which orchestras belong to the most distinguished ones.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Peacocks",
                "sec_num": "2.3"
            },
            {
                "text": "All peacock terms are similar to hedges to some extent. They can be called scalar uncertainties since in both cases, a scale is involved in the interpretation of the uncertain term. In the case of peacock, there is a scale of polarity on which phrases can be judged as positive or negative whereas in the case of hedges, there is a scale on which there is a reference point, on the basis of which the uncertain part of the utterance is placed. Although they are similar, we suggest that peacocks and hedges be differentiated in our classification because peacocks are related to subjectivity while hedges are more neutral, hence they can be relevant for different NLP applications (e.g. in opinion mining, which seeks to collect subjective opinions on different topics, peacocks may prove more useful than hedges). Still, hedges shift the value of the quantity / quality mentioned in the text while peacocks denote a specific point on the scale, without modifying it, which again suggests that they should not be lumped in the same class.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Peacocks",
                "sec_num": "2.3"
            },
            {
                "text": "These days, uncertainty and modality detection is a widely studied area in natural language processing, which manifests itself in a number of corpora annotated for uncertainty in domains like biology (Medlock and Briscoe, 2007; Kim et al., 2008; Vincze et al., 2008; Nawaz et al., 2010) , medicine (Uzuner et al., 2009) , news media (Wilson, 2008; Saur\u00ed and Pustejovsky, 2009; Rubin, 2010) , and encyclopedia texts (Farkas et al., 2010) . Although some authors have called attention to the fact that the progressive nature of discourse and dimensions of time should be also taken into account (de Marneffe et al., 2012; Saur\u00ed and Pustejovsky, 2012) , as can be judged on the basis of available guidelines, most of these corpora make use of semantic uncertainty, with some exceptions that take into account pragmatic or discourse-level information as well (see below).",
                "cite_spans": [
                    {
                        "start": 200,
                        "end": 227,
                        "text": "(Medlock and Briscoe, 2007;",
                        "ref_id": "BIBREF13"
                    },
                    {
                        "start": 228,
                        "end": 245,
                        "text": "Kim et al., 2008;",
                        "ref_id": "BIBREF11"
                    },
                    {
                        "start": 246,
                        "end": 266,
                        "text": "Vincze et al., 2008;",
                        "ref_id": "BIBREF25"
                    },
                    {
                        "start": 267,
                        "end": 286,
                        "text": "Nawaz et al., 2010)",
                        "ref_id": "BIBREF15"
                    },
                    {
                        "start": 298,
                        "end": 319,
                        "text": "(Uzuner et al., 2009)",
                        "ref_id": "BIBREF24"
                    },
                    {
                        "start": 333,
                        "end": 347,
                        "text": "(Wilson, 2008;",
                        "ref_id": "BIBREF28"
                    },
                    {
                        "start": 348,
                        "end": 376,
                        "text": "Saur\u00ed and Pustejovsky, 2009;",
                        "ref_id": "BIBREF21"
                    },
                    {
                        "start": 377,
                        "end": 389,
                        "text": "Rubin, 2010)",
                        "ref_id": "BIBREF20"
                    },
                    {
                        "start": 415,
                        "end": 436,
                        "text": "(Farkas et al., 2010)",
                        "ref_id": "BIBREF8"
                    },
                    {
                        "start": 593,
                        "end": 619,
                        "text": "(de Marneffe et al., 2012;",
                        "ref_id": "BIBREF6"
                    },
                    {
                        "start": 620,
                        "end": 648,
                        "text": "Saur\u00ed and Pustejovsky, 2012)",
                        "ref_id": "BIBREF22"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "3"
            },
            {
                "text": "The concept of source has played a significant role in the literature. FactBank (Saur\u00ed and Pustejovsky, 2009) explicitly annotates the factuality of events according to their sources' perspective and Wiebe et al. (2005) also emphasize the role of sources annotated in the MPQA corpus for opinion mining. The notion of perspective -both in Nawaz et al. (2010) and in Morante and Daelemans (2011) -is similar to the one of sources applied in FactBank and MPQA. In Wikipedia, the lack of identifiable sources is explicitly discouraged by editors. They call such phenomena weasels (see also Ganter and Strube (2009) ) and weasel detection was one of the subtasks of the CoNLL-2010 shared task (Farkas et al., 2010) .",
                "cite_spans": [
                    {
                        "start": 80,
                        "end": 109,
                        "text": "(Saur\u00ed and Pustejovsky, 2009)",
                        "ref_id": "BIBREF21"
                    },
                    {
                        "start": 200,
                        "end": 219,
                        "text": "Wiebe et al. (2005)",
                        "ref_id": "BIBREF26"
                    },
                    {
                        "start": 339,
                        "end": 358,
                        "text": "Nawaz et al. (2010)",
                        "ref_id": "BIBREF15"
                    },
                    {
                        "start": 587,
                        "end": 611,
                        "text": "Ganter and Strube (2009)",
                        "ref_id": "BIBREF9"
                    },
                    {
                        "start": 689,
                        "end": 710,
                        "text": "(Farkas et al., 2010)",
                        "ref_id": "BIBREF8"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "3"
            },
            {
                "text": "The lack of source characteristics to weasels can be paired with a certain strategy that Hyland (1996) calls impersonal constructions. It is a type of writer-oriented hedges 3 in his system. It is interesting to note that in his system, the opposite of this strategy can also be found, which could be called anti-weasel: the writer emphasizes his responsibility by using first person pronouns. However, this latter strategy does not represent any form of uncertainty in our view.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "3"
            },
            {
                "text": "Fuzziness is another dimension of uncertainty. Lakoff (1973) gave an account of some lexical items -which he calls hedges -that \"make things fuzzier\", that is, words such as approximately, kind of, at least etc. Due to the presence of such words, the quality or quantity under investigation is shifted on a scale. If modified by the adverb very for instance, it moves towards one end of the scale on which this quality/quantity is determined. The phenomenon of hedging in scientific articles is analyzed and categorized according to the functions it can fulfill in Hyland (1996) .",
                "cite_spans": [
                    {
                        "start": 47,
                        "end": 60,
                        "text": "Lakoff (1973)",
                        "ref_id": "BIBREF12"
                    },
                    {
                        "start": 565,
                        "end": 578,
                        "text": "Hyland (1996)",
                        "ref_id": "BIBREF10"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "3"
            },
            {
                "text": "Subjectivity is also related to uncertainty. There is a great diversity among individual views and opinions: a feature of a product may be appreciated by some customers but it might be considered intolerable for others. Thus, what should be considered positive or negative seems subjective. Many approaches to subjectivity or sentiment analysis rely on lexicons and databases of subjective terms. For instance, the database SentiWord-Net (Baccianella et al., 2010) contains a subset of the synsets of the Princeton Wordnet with positivity, negativity and neutrality scores assigned to each concept, depending on the use of its sentiment orientation, thus it is a lexicon where subjective terms are listed and ranked. Wilson (2008) defines subjectivity clues as words and phrases that express private states, that is, individual opinions. She distinguishes lexical cues and syntactic cues that are responsible for subjectivity. She lists several modifiers among her syntactic clues of subjectivity like quite and really. However, in contrast with other subjective elements, we do not regard them as peacock cues since -as Wilson (2008) herself states -they \"work to intensify\", so in our system they are classified as hedge cues. On the other hand, some instances of biased language can also be classified as peacocks in our system (Recasens et al., 2013) .",
                "cite_spans": [
                    {
                        "start": 438,
                        "end": 464,
                        "text": "(Baccianella et al., 2010)",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 717,
                        "end": 730,
                        "text": "Wilson (2008)",
                        "ref_id": "BIBREF28"
                    },
                    {
                        "start": 1331,
                        "end": 1354,
                        "text": "(Recasens et al., 2013)",
                        "ref_id": "BIBREF18"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "3"
            },
            {
                "text": "Human communication and discourse is incremental in nature (Cristea and Webber, 1997) . Information may be added at a later point of the discourse that clarifies a previously missing piece of information. Applying this to discourse-level uncertainty, it may be the case that an apparent weasel phrase is elaborated on later in the discourse, or the exact value of an apparent hedge expression is later provided. In such cases, the phrases should not be marked as uncertain, which indicates the essential role of co-text -i.e. surrounding words in the text (Brown and Yule, 1983 ) -in detecting discourse-level uncertainty.",
                "cite_spans": [
                    {
                        "start": 59,
                        "end": 85,
                        "text": "(Cristea and Webber, 1997)",
                        "ref_id": "BIBREF5"
                    },
                    {
                        "start": 556,
                        "end": 577,
                        "text": "(Brown and Yule, 1983",
                        "ref_id": "BIBREF4"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Related Work",
                "sec_num": "3"
            },
            {
                "text": "In order to test the practical applicability of the new classification of discourse-level uncertainty phenomena, and to investigate the frequency of each uncertainty type, we also created an annotated corpus. We selected WikiWeasel, the Wikipedia subset of the CoNLL-2010 Shared Task corpora (Farkas et al., 2010) for annotation. By doing this, our results could be contrasted with those of the original annotation carried out specifically for the shared task. Moreover, as the corpus has recently been annotated for semantic uncertainty (Szarvas et al., 2012) , interesting comparisons can also be made between semantic and discourse-level uncertainty. The annotated corpus is available free of charge for research purposes at www.inf.u-szeged.hu/rgai/ uncertainty.",
                "cite_spans": [
                    {
                        "start": 292,
                        "end": 313,
                        "text": "(Farkas et al., 2010)",
                        "ref_id": "BIBREF8"
                    },
                    {
                        "start": 538,
                        "end": 560,
                        "text": "(Szarvas et al., 2012)",
                        "ref_id": "BIBREF23"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Annotated Corpus",
                "sec_num": "4"
            },
            {
                "text": "The dataset consists of 4,530 Wikipedia articles and 20,756 sentences. Texts were manually annotated by two linguists for linguistic cues denoting all types of discourse-level uncertainty, i.e. weasel, peacock and hedge. 200 articles were annotated by both linguists and the inter-annotator agreement rate for the categories weasel, peacock and hedge were 0.4837, 0.4512 and 0.4606, respectively (in terms of \u03ba-measure), which reflects that identifying discourse-level phenomena is not straightforward, however, it can be reasonably well solved considering the subjective nature of the task. During the annotation, special emphasis was laid on the discourse structure of the text. For instance, weasel cue candidates do not denote uncertainty when the sentence is enhanced with citations. Also, a weasel-like element may be elaborated on in the next sentence, thus it is not to be marked as weasel as in:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Statistical Data on the Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "Some ship names are references to other games created by Jordan Weisman. The \"Black Swan\" is a reference to a character from Crimson Skies, and also possibly to the ship Black Pearl from Pirates of the Caribbean.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Statistical Data on the Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "In order to attain the gold standard for the commonly annotated parts, the two annotators discussed problematic cases and reached a consensus for each case. The final version of the corpus contains these disambiguated cases.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Statistical Data on the Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "The dataset contains 10,794 discourse-level uncertainty cues 4 , which occur in 7,336 uncertain sentences. A sentence was considered to be uncertain if it contained at least one uncertainty cue. But, as the results show, many sentences include more than one uncertainty cue. Statistical data on the uncertainty cues found in the WikiWeasel corpus are listed in Table 1 , together with available data on semantic uncertainty types, taken from Szarvas et al. (2012) . As can be seen, most of the uncertainty cues found in the corpus belong to the discourselevel uncertainty class, the ratio of semantic to discourse-level uncertainty cues being 1:4. Among the types of discourse-level uncertainty, hedges are the most frequent, followed by weasels and peacocks. All this suggests that discourselevel uncertainty is very typical of Wikipedia articles, about 35% of the sentences being uncertain at the discourse level. As regards the specific classes, 3,807 (18.3%), 3,497 (16.8%) and 1,359 (6.5%) sentences contain at least one hedge, weasel or peacock cue, respectively.",
                "cite_spans": [
                    {
                        "start": 442,
                        "end": 463,
                        "text": "Szarvas et al. (2012)",
                        "ref_id": "BIBREF23"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 361,
                        "end": 368,
                        "text": "Table 1",
                        "ref_id": "TABREF1"
                    }
                ],
                "eq_spans": [],
                "section": "Statistical Data on the Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "On the number of different cues, Table 1 tells us that the set of linguistic cues expressing weasels are the most limited, with almost 100 cues. In contrast, peacock cues vary the most with 540 cues. This suggests that weasels have the most restricted vocabulary in contrast to peacocks, and hedges being in the middle. This also means that the average frequency of a weasel cue is much higher than that of a peacock cue: the average frequency of occurrence of weasel, hedge and peacock cues is 41.8, 18.24 and 3.54, respectively.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 33,
                        "end": 40,
                        "text": "Table 1",
                        "ref_id": "TABREF1"
                    }
                ],
                "eq_spans": [],
                "section": "Cue Distribution in the Corpus",
                "sec_num": "4.2"
            },
            {
                "text": "We did a more detailed analysis on the lexical distribution of the cues as well. The ten most frequent cues for each type are listed in Table 2 . These are responsible for about 86%, 45% and 42% of the occurrences of weasel, hedge and peacock cues, respectively. Thus, a limited vocabu- lary can account for over 85% of weasels. However, some terms can belong to more than one uncertainty type. For example, most occurs in all the three types (weasel: Most agree that this puts her at about 12 years of age, hedge: He spent most of his time working on questions of theology and peacock: Kathu is the district which covers the most touristical beach of Phuket), but some, many and several can all be instances of weasels and hedges. This is due to the linguistic variability of these items: e.g. some may refer to \"an indefinite quantity\" or \"something unspecified\".",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 136,
                        "end": 143,
                        "text": "Table 2",
                        "ref_id": "TABREF3"
                    }
                ],
                "eq_spans": [],
                "section": "Cue Distribution in the Corpus",
                "sec_num": "4.2"
            },
            {
                "text": "As can be seen, there are some overlapping cues among the types. This is especially so in the case of hedges and weasels: 25 cues can denote hedges or weasels as well, thus 25% of the weasel cues are ambiguous. These cues were also responsible for most of the differences between the two annotations, which indicates that their identification requires special attention both for human annotators and NLP tools: it is mostly the neighbouring words that can determine whether it is a weasel or hedge. For instance, if some occurs before a verb and constitutes a noun phrase on its own, then it is almost certainly a weasel cue (Some think that. . . ) but if it occurs before a noun denoting time, it is probably a hedge (some minutes ago).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Cue Distribution in the Corpus",
                "sec_num": "4.2"
            },
            {
                "text": "We carried out some baseline experiments on the corpus. We divided the corpus into training (80%) and test (20%) sets and applied a simple dictionary-based approach which classified each cue candidate as uncertain if it was tagged as uncertain in at least 50% of its occurrences in the training dataset. For ambiguous cues, the most frequent label was chosen (e.g. most was used as a peacock cue). Similar to the CoNLL-2010 shared task, we evaluated our results at the cue level as well as at the sentence level. Table 3 : Baseline results in terms of precision / recall / F-score. Table 3 shows that the peacock class is the most difficult to detect, which may be due to the fact that this class has the most diverse cues and thus applying a dictionary-based method leads to a lower recall. Still, the lower precision was due to the higher level of ambiguity concerning the most typical peacock cues (like most). As for hedges, a simple lexical approach can result in a good precision score, which suggests that hedge cues are less ambiguous than weasel or peacock cues. It is also seen that sentence-level results are significantly higher than cue-level results (ANOVA, p = 0.0026). Uncertain sentences typically contain more than one cue and in the former scenario, it is sufficient to recognize only one cue in the sentence to regard the sentence as uncertain and false negatives do not affect the performance significantly.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 513,
                        "end": 520,
                        "text": "Table 3",
                        "ref_id": null
                    },
                    {
                        "start": 582,
                        "end": 589,
                        "text": "Table 3",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Experiments",
                "sec_num": "5"
            },
            {
                "text": "If we compare the data with the CoNLL-2010 version of the corpus, it is seen that the new annotation scheme leads to many more cues (6,725 cue phrases in 4,718 uncertain sentences in the original version vs. 10,794 cues in 7,336 sentences in the version described here) and -although the datasets are not directly comparable -it gives a much better performance: the best system achieved an F-score of 60.2 on weasel detection at the sentence level and 36.5 at the cue level and no classes of cues were distinguished there (Farkas et al., 2010) . This difference may be attributed to several factors. First, not all hedge phenomena (used in the sense introduced here) were systematically annotated in the CoNLL-2010 corpus.",
                "cite_spans": [
                    {
                        "start": 522,
                        "end": 543,
                        "text": "(Farkas et al., 2010)",
                        "ref_id": "BIBREF8"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments",
                "sec_num": "5"
            },
            {
                "text": "Second, complex syntactic structures that contained several types of uncertainty were annotated as one complex cue (e.g. the phrase it has been widely suggested, which contains epistemic uncertainty (suggested), weasel (passive sentence with no agent) and hedge (widely) as well). Third, the CoNLL-2010 version did not distinguish subtypes of cues, i.e. semantic uncertainty and weasels were annotated in the same way. It was probably because of this lack of distinction that participants of the shared task got considerably lower results for Wikipedia articles than for biological papers, which contained fewer weasel cues (Farkas et al., 2010) . However, the new annotation makes it possible to select those types of uncertainty that are relevant for a given application, see Section 6.",
                "cite_spans": [
                    {
                        "start": 624,
                        "end": 645,
                        "text": "(Farkas et al., 2010)",
                        "ref_id": "BIBREF8"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments",
                "sec_num": "5"
            },
            {
                "text": "6 Discourse-level Uncertainty and NLP Detecting weasels is of utmost importance in every information extraction application where it should be known who the author/source is. Thus, information extraction applied for the news media may certainly profit from finding weasels, i.e. missing or undeterminable sources. Pieces of information without an identifiable (and reliable) source require special treatment: they will be excluded from the news or they will be communicated to the public in a special form, using phrases such as according to unnamed sources etc.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments",
                "sec_num": "5"
            },
            {
                "text": "In sentiment analysis and opinion mining, the identification of subjective terms is essential. These terms are often ambiguous hence a subjectivity word sense disambiguation is needed (Wiebe, 2012) . In our corpus, peacock terms and intensifiers -a subtype of hedges -are manually annotated, thus it can be used in the development and evaluation of tools that seek to disambiguate elements of a subjectivity lexicon in running texts.",
                "cite_spans": [
                    {
                        "start": 184,
                        "end": 197,
                        "text": "(Wiebe, 2012)",
                        "ref_id": "BIBREF27"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments",
                "sec_num": "5"
            },
            {
                "text": "Information retrieval may also be enhanced by detecting discourse-level uncertainty. In order to find relevant documents for queries that contain numbers, more specifically, to improve recall in such cases, it is important to handle numeric hedges. For instance, if someone looks for websites describing games appropriate for ten year old children, he also may be interested in games that are for children over eight. Thus, the search engine should be prepared for recognizing that the number specified in the query (\"ten\") is part of other numeric sets (e.g. \"over eight\") and in this way, more relevant hits can be retrieved.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments",
                "sec_num": "5"
            },
            {
                "text": "The linguistic processing of patents especially requires that hedges should be recognized. There is a tendency to generalize over the scope of the patent (i.e. hedges are used) in order to prevent further abuse (Osenga, 2006) . Thus, the scope of the patents can be expanded or other use cases can later be included in the patent. Hence, any NLP system that aims at patent processing must target hedge detection as well. Document classification may also profit from detecting discourse-level uncertainty since different genres of texts involve different types of uncertainty.For instance, papers in the humanities contain significantly more hedges than papers in sciences (Rizomilioti, 2006) . Thus, the frequency of hedges may be indicative of the domain of the text as well, which again may be exploited in document classification.",
                "cite_spans": [
                    {
                        "start": 211,
                        "end": 225,
                        "text": "(Osenga, 2006)",
                        "ref_id": "BIBREF16"
                    },
                    {
                        "start": 672,
                        "end": 691,
                        "text": "(Rizomilioti, 2006)",
                        "ref_id": "BIBREF19"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Experiments",
                "sec_num": "5"
            },
            {
                "text": "In this paper, we presented a classification of discourse-level uncertainty phenomena, and focused on the concepts of source, fuzziness and subjectivity. We also introduced a corpus of Wikipedia articles in which linguistic cues for each type of discourse-level uncertaintyweasel, peacock and hedge -were manually annotated. We carried out some baseline experiments on discourse-level uncertainty detection, which may prove useful in information extraction and retrieval, sentiment analysis and opinion mining.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusions",
                "sec_num": "7"
            },
            {
                "text": "In the future, we intend to develop a machinelearning based uncertainty detector. We would also like to investigate the distribution of weasels, hedges and peacocks in other types of texts (e.g. news media or scientific papers) and in other languages as our three categories are languageindependent. Moreover, to learn how domaindependent the model is, we plan to do some domain adaptation experiments as well.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusions",
                "sec_num": "7"
            },
            {
                "text": "http://en.wikipedia.org/wiki/ Wikipedia:Manual_of_Style/Words_to_watch",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "However, in our classification, it should be called a weasel.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "We should mention that our corpus contained 680 passive constructions, which were annotated as weasels. As we focus now on lexical cues of discourse-level uncertainty, and they belong to syntactic cues, the investigation of such cases will be subject to further studies.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            }
        ],
        "back_matter": [
            {
                "text": "This work was supported in part by the European Union and the European Social Fund through the project FuturICT.hu (grant no.: T\u00c1MOP-4.2.2.C-11/1/KONV-2012-0013).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Acknowledgments",
                "sec_num": null
            }
        ],
        "bib_entries": {
            "BIBREF1": {
                "ref_id": "b1",
                "title": "SentiWordNet 3.0: An Enhanced Lexical Resource for Sentiment Analysis and Opinion Mining",
                "authors": [
                    {
                        "first": "Stefano",
                        "middle": [],
                        "last": "Baccianella",
                        "suffix": ""
                    },
                    {
                        "first": "Andrea",
                        "middle": [],
                        "last": "Esuli",
                        "suffix": ""
                    },
                    {
                        "first": "Fabrizio",
                        "middle": [],
                        "last": "Sebastiani",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of LREC'10",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Stefano Baccianella, Andrea Esuli, and Fabrizio Sebas- tiani. 2010. SentiWordNet 3.0: An Enhanced Lexical Resource for Sentiment Analysis and Opin- ion Mining. In Proceedings of LREC'10, Valletta, Malta, May. ELRA.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "The language of the News Media",
                "authors": [
                    {
                        "first": "Allan",
                        "middle": [],
                        "last": "Bell",
                        "suffix": ""
                    }
                ],
                "year": 1991,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Allan Bell. 1991. The language of the News Media. Blackwell, Oxford.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "Politeness: Some universals in language usage",
                "authors": [
                    {
                        "first": "Penelope",
                        "middle": [],
                        "last": "Brown",
                        "suffix": ""
                    },
                    {
                        "first": "Stephen",
                        "middle": [
                            "C"
                        ],
                        "last": "Levinson",
                        "suffix": ""
                    }
                ],
                "year": 1987,
                "venue": "CUP",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Penelope Brown and Stephen C. Levinson. 1987. Politeness: Some universals in language usage. CUP, Cambridge, UK.",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "Discourse Analysis. CUP",
                "authors": [
                    {
                        "first": "Gillian",
                        "middle": [],
                        "last": "Brown",
                        "suffix": ""
                    },
                    {
                        "first": "George",
                        "middle": [],
                        "last": "Yule",
                        "suffix": ""
                    }
                ],
                "year": 1983,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Gillian Brown and George Yule. 1983. Discourse Analysis. CUP, Cambridge, UK.",
                "links": null
            },
            "BIBREF5": {
                "ref_id": "b5",
                "title": "Expectations in incremental discourse processing",
                "authors": [
                    {
                        "first": "Dan",
                        "middle": [],
                        "last": "Cristea",
                        "suffix": ""
                    },
                    {
                        "first": "Bonnie",
                        "middle": [],
                        "last": "Webber",
                        "suffix": ""
                    }
                ],
                "year": 1997,
                "venue": "Proceedings of ACL97/EACL97",
                "volume": "",
                "issue": "",
                "pages": "88--95",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Dan Cristea and Bonnie Webber. 1997. Expecta- tions in incremental discourse processing. In Pro- ceedings of ACL97/EACL97, pages 88-95. Morgan Kaufmann.",
                "links": null
            },
            "BIBREF6": {
                "ref_id": "b6",
                "title": "Did it happen? The pragmatic complexity of veridicality assessment",
                "authors": [
                    {
                        "first": "Marie-Catherine",
                        "middle": [],
                        "last": "De Marneffe",
                        "suffix": ""
                    },
                    {
                        "first": "Christopher",
                        "middle": [
                            "D"
                        ],
                        "last": "Manning",
                        "suffix": ""
                    },
                    {
                        "first": "Christopher",
                        "middle": [],
                        "last": "Potts",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "Computational Linguistics",
                "volume": "38",
                "issue": "",
                "pages": "301--333",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Marie-Catherine de Marneffe, Christopher D. Man- ning, and Christopher Potts. 2012. Did it happen? The pragmatic complexity of veridicality assess- ment. Computational Linguistics, 38:301-333.",
                "links": null
            },
            "BIBREF7": {
                "ref_id": "b7",
                "title": "Committed belief annotation and tagging",
                "authors": [
                    {
                        "first": "Mona",
                        "middle": [
                            "T"
                        ],
                        "last": "Diab",
                        "suffix": ""
                    },
                    {
                        "first": "Lori",
                        "middle": [
                            "S"
                        ],
                        "last": "Levin",
                        "suffix": ""
                    },
                    {
                        "first": "Teruko",
                        "middle": [],
                        "last": "Mitamura",
                        "suffix": ""
                    },
                    {
                        "first": "Owen",
                        "middle": [],
                        "last": "Rambow",
                        "suffix": ""
                    },
                    {
                        "first": "Vinodkumar",
                        "middle": [],
                        "last": "Prabhakaran",
                        "suffix": ""
                    },
                    {
                        "first": "Weiwei",
                        "middle": [],
                        "last": "Guo",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "Proceedings of LAW 2009",
                "volume": "",
                "issue": "",
                "pages": "68--73",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Mona T. Diab, Lori S. Levin, Teruko Mitamura, Owen Rambow, Vinodkumar Prabhakaran, and Weiwei Guo. 2009. Committed belief annotation and tag- ging. In Proceedings of LAW 2009, pages 68-73. The Association for Computer Linguistics.",
                "links": null
            },
            "BIBREF8": {
                "ref_id": "b8",
                "title": "The CoNLL-2010 Shared Task: Learning to Detect Hedges and their Scope in Natural Language Text",
                "authors": [
                    {
                        "first": "Rich\u00e1rd",
                        "middle": [],
                        "last": "Farkas",
                        "suffix": ""
                    },
                    {
                        "first": "Veronika",
                        "middle": [],
                        "last": "Vincze",
                        "suffix": ""
                    },
                    {
                        "first": "Gy\u00f6rgy",
                        "middle": [],
                        "last": "M\u00f3ra",
                        "suffix": ""
                    },
                    {
                        "first": "J\u00e1nos",
                        "middle": [],
                        "last": "Csirik",
                        "suffix": ""
                    },
                    {
                        "first": "Gy\u00f6rgy",
                        "middle": [],
                        "last": "Szarvas",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of the CoNLL-2010 Shared Task",
                "volume": "",
                "issue": "",
                "pages": "1--12",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Rich\u00e1rd Farkas, Veronika Vincze, Gy\u00f6rgy M\u00f3ra, J\u00e1nos Csirik, and Gy\u00f6rgy Szarvas. 2010. The CoNLL- 2010 Shared Task: Learning to Detect Hedges and their Scope in Natural Language Text. In Proceed- ings of the CoNLL-2010 Shared Task, pages 1-12, Uppsala, Sweden, July. ACL.",
                "links": null
            },
            "BIBREF9": {
                "ref_id": "b9",
                "title": "Finding Hedges by Chasing Weasels: Hedge Detection Using Wikipedia Tags and Shallow Linguistic Features",
                "authors": [
                    {
                        "first": "Viola",
                        "middle": [],
                        "last": "Ganter",
                        "suffix": ""
                    },
                    {
                        "first": "Michael",
                        "middle": [],
                        "last": "Strube",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "Proceedings of ACL-IJCNLP 2009",
                "volume": "",
                "issue": "",
                "pages": "173--176",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Viola Ganter and Michael Strube. 2009. Find- ing Hedges by Chasing Weasels: Hedge Detection Using Wikipedia Tags and Shallow Linguistic Fea- tures. In Proceedings of ACL-IJCNLP 2009, pages 173-176, Suntec, Singapore, August. ACL.",
                "links": null
            },
            "BIBREF10": {
                "ref_id": "b10",
                "title": "Writing without conviction? Hedging in scientific research articles",
                "authors": [
                    {
                        "first": "Ken",
                        "middle": [],
                        "last": "Hyland",
                        "suffix": ""
                    }
                ],
                "year": 1996,
                "venue": "Applied Linguistics",
                "volume": "17",
                "issue": "4",
                "pages": "433--454",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ken Hyland. 1996. Writing without conviction? Hedging in scientific research articles. Applied Lin- guistics, 17(4):433-454.",
                "links": null
            },
            "BIBREF11": {
                "ref_id": "b11",
                "title": "Corpus annotation for mining biomedical events from literature",
                "authors": [
                    {
                        "first": "Jin-Dong",
                        "middle": [],
                        "last": "Kim",
                        "suffix": ""
                    },
                    {
                        "first": "Tomoko",
                        "middle": [],
                        "last": "Ohta",
                        "suffix": ""
                    },
                    {
                        "first": "Jun'ichi",
                        "middle": [],
                        "last": "Tsujii",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "BMC Bioinformatics",
                "volume": "9",
                "issue": "10",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Jin-Dong Kim, Tomoko Ohta, and Jun'ichi Tsujii. 2008. Corpus annotation for mining biomedi- cal events from literature. BMC Bioinformatics, 9(Suppl 10).",
                "links": null
            },
            "BIBREF12": {
                "ref_id": "b12",
                "title": "Hedges: A study in meaning criteria and the logic of fuzzy concepts",
                "authors": [
                    {
                        "first": "George",
                        "middle": [],
                        "last": "Lakoff",
                        "suffix": ""
                    }
                ],
                "year": 1973,
                "venue": "Journal of Philosophical Logic",
                "volume": "2",
                "issue": "4",
                "pages": "458--508",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "George Lakoff. 1973. Hedges: A study in meaning criteria and the logic of fuzzy concepts. Journal of Philosophical Logic, 2(4):458-508.",
                "links": null
            },
            "BIBREF13": {
                "ref_id": "b13",
                "title": "Weakly Supervised Learning for Hedge Classification in Scientific Literature",
                "authors": [
                    {
                        "first": "Ben",
                        "middle": [],
                        "last": "Medlock",
                        "suffix": ""
                    },
                    {
                        "first": "Ted",
                        "middle": [],
                        "last": "Briscoe",
                        "suffix": ""
                    }
                ],
                "year": 2007,
                "venue": "Proceedings of the ACL",
                "volume": "",
                "issue": "",
                "pages": "992--999",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ben Medlock and Ted Briscoe. 2007. Weakly Super- vised Learning for Hedge Classification in Scientific Literature. In Proceedings of the ACL, pages 992- 999, Prague, Czech Republic, June.",
                "links": null
            },
            "BIBREF14": {
                "ref_id": "b14",
                "title": "Annotating Modality and Negation for a Machine Reading Evaluation",
                "authors": [
                    {
                        "first": "Roser",
                        "middle": [],
                        "last": "Morante",
                        "suffix": ""
                    },
                    {
                        "first": "Walter",
                        "middle": [],
                        "last": "Daelemans",
                        "suffix": ""
                    }
                ],
                "year": 2011,
                "venue": "Proceedings of CLEF",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Roser Morante and Walter Daelemans. 2011. Annotat- ing Modality and Negation for a Machine Reading Evaluation. In Proceedings of CLEF 2011.",
                "links": null
            },
            "BIBREF15": {
                "ref_id": "b15",
                "title": "Evaluating a meta-knowledge annotation scheme for bio-events",
                "authors": [
                    {
                        "first": "Raheel",
                        "middle": [],
                        "last": "Nawaz",
                        "suffix": ""
                    },
                    {
                        "first": "Paul",
                        "middle": [],
                        "last": "Thompson",
                        "suffix": ""
                    },
                    {
                        "first": "Sophia",
                        "middle": [],
                        "last": "Ananiadou",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "Proceedings of NESP 2010",
                "volume": "",
                "issue": "",
                "pages": "69--77",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Raheel Nawaz, Paul Thompson, and Sophia Anani- adou. 2010. Evaluating a meta-knowledge annota- tion scheme for bio-events. In Proceedings of NESP 2010, pages 69-77, Uppsala, Sweden, July. Univer- sity of Antwerp.",
                "links": null
            },
            "BIBREF16": {
                "ref_id": "b16",
                "title": "Linguistics and Patent Claim Construction",
                "authors": [
                    {
                        "first": "Kristen",
                        "middle": [],
                        "last": "Osenga",
                        "suffix": ""
                    }
                ],
                "year": 2006,
                "venue": "Rutgers Law Journal",
                "volume": "38",
                "issue": "61",
                "pages": "61--108",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Kristen Osenga. 2006. Linguistics and Patent Claim Construction. Rutgers Law Journal, 38(61):61-108.",
                "links": null
            },
            "BIBREF17": {
                "ref_id": "b17",
                "title": "Mood and Modality. CUP",
                "authors": [
                    {
                        "first": "Frank Robert",
                        "middle": [],
                        "last": "Palmer",
                        "suffix": ""
                    }
                ],
                "year": 1986,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Frank Robert Palmer. 1986. Mood and Modality. CUP, Cambridge.",
                "links": null
            },
            "BIBREF18": {
                "ref_id": "b18",
                "title": "Linguistic models for analyzing and detecting biased language",
                "authors": [
                    {
                        "first": "Marta",
                        "middle": [],
                        "last": "Recasens",
                        "suffix": ""
                    },
                    {
                        "first": "Cristian",
                        "middle": [],
                        "last": "Danescu-Niculescu-Mizil",
                        "suffix": ""
                    },
                    {
                        "first": "Dan",
                        "middle": [],
                        "last": "Jurafsky",
                        "suffix": ""
                    }
                ],
                "year": 2013,
                "venue": "Proceedings of ACL-2013",
                "volume": "",
                "issue": "",
                "pages": "1650--1659",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Marta Recasens, Cristian Danescu-Niculescu-Mizil, and Dan Jurafsky. 2013. Linguistic models for ana- lyzing and detecting biased language. In Proceed- ings of ACL-2013, pages 1650-1659, Sofia, Bul- garia, August. ACL.",
                "links": null
            },
            "BIBREF19": {
                "ref_id": "b19",
                "title": "Exploring epistemic modality in academic discourse using corpora",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Vassiliki Rizomilioti",
                        "suffix": ""
                    }
                ],
                "year": 2006,
                "venue": "Information Technology in Languages for Specific Purposes",
                "volume": "",
                "issue": "",
                "pages": "53--71",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Vassiliki Rizomilioti. 2006. Exploring epistemic modality in academic discourse using corpora. In Information Technology in Languages for Specific Purposes, pages 53-71. Springer US.",
                "links": null
            },
            "BIBREF20": {
                "ref_id": "b20",
                "title": "Epistemic modality: From uncertainty to certainty in the context of information seeking as interactions with texts. Information Processing & Management",
                "authors": [
                    {
                        "first": "L",
                        "middle": [],
                        "last": "Victoria",
                        "suffix": ""
                    },
                    {
                        "first": "",
                        "middle": [],
                        "last": "Rubin",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "",
                "volume": "46",
                "issue": "",
                "pages": "533--540",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Victoria L. Rubin. 2010. Epistemic modality: From uncertainty to certainty in the context of information seeking as interactions with texts. Information Pro- cessing & Management, 46(5):533-540.",
                "links": null
            },
            "BIBREF21": {
                "ref_id": "b21",
                "title": "FactBank: a corpus annotated with event factuality. Language Resources and Evaluation",
                "authors": [
                    {
                        "first": "Roser",
                        "middle": [],
                        "last": "Saur\u00ed",
                        "suffix": ""
                    },
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "",
                "volume": "43",
                "issue": "",
                "pages": "227--268",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Roser Saur\u00ed and James Pustejovsky. 2009. FactBank: a corpus annotated with event factuality. Language Resources and Evaluation, 43:227-268.",
                "links": null
            },
            "BIBREF22": {
                "ref_id": "b22",
                "title": "Are you sure that this happened? Assessing the factuality degree of events in text",
                "authors": [
                    {
                        "first": "Roser",
                        "middle": [],
                        "last": "Saur\u00ed",
                        "suffix": ""
                    },
                    {
                        "first": "James",
                        "middle": [],
                        "last": "Pustejovsky",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "Computational Linguistics",
                "volume": "38",
                "issue": "",
                "pages": "261--299",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Roser Saur\u00ed and James Pustejovsky. 2012. Are you sure that this happened? Assessing the factuality degree of events in text. Computational Linguistics, 38:261-299.",
                "links": null
            },
            "BIBREF23": {
                "ref_id": "b23",
                "title": "Crossgenre and cross-domain detection of semantic uncertainty",
                "authors": [
                    {
                        "first": "Gy\u00f6rgy",
                        "middle": [],
                        "last": "Szarvas",
                        "suffix": ""
                    },
                    {
                        "first": "Veronika",
                        "middle": [],
                        "last": "Vincze",
                        "suffix": ""
                    },
                    {
                        "first": "Rich\u00e1rd",
                        "middle": [],
                        "last": "Farkas",
                        "suffix": ""
                    },
                    {
                        "first": "Gy\u00f6rgy",
                        "middle": [],
                        "last": "M\u00f3ra",
                        "suffix": ""
                    },
                    {
                        "first": "Iryna",
                        "middle": [],
                        "last": "Gurevych",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "Computational Linguistics",
                "volume": "38",
                "issue": "",
                "pages": "335--367",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Gy\u00f6rgy Szarvas, Veronika Vincze, Rich\u00e1rd Farkas, Gy\u00f6rgy M\u00f3ra, and Iryna Gurevych. 2012. Cross- genre and cross-domain detection of semantic uncer- tainty. Computational Linguistics, 38:335-367.",
                "links": null
            },
            "BIBREF24": {
                "ref_id": "b24",
                "title": "Machine Learning and Rule-based Approaches to Assertion Classification",
                "authors": [
                    {
                        "first": "Ozlem",
                        "middle": [],
                        "last": "Uzuner",
                        "suffix": ""
                    },
                    {
                        "first": "Xiaoran",
                        "middle": [],
                        "last": "Zhang",
                        "suffix": ""
                    },
                    {
                        "first": "Tawanda",
                        "middle": [],
                        "last": "Sibanda",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "Journal of the American Medical Informatics Association",
                "volume": "16",
                "issue": "1",
                "pages": "109--115",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Ozlem Uzuner, Xiaoran Zhang, and Tawanda Sibanda. 2009. Machine Learning and Rule-based Approaches to Assertion Classification. Journal of the American Medical Informatics Association, 16(1):109-115, January.",
                "links": null
            },
            "BIBREF25": {
                "ref_id": "b25",
                "title": "The Bio-Scope Corpus: Biomedical Texts Annotated for Uncertainty, Negation and their Scopes",
                "authors": [
                    {
                        "first": "Veronika",
                        "middle": [],
                        "last": "Vincze",
                        "suffix": ""
                    },
                    {
                        "first": "Gy\u00f6rgy",
                        "middle": [],
                        "last": "Szarvas",
                        "suffix": ""
                    },
                    {
                        "first": "Rich\u00e1rd",
                        "middle": [],
                        "last": "Farkas",
                        "suffix": ""
                    },
                    {
                        "first": "Gy\u00f6rgy",
                        "middle": [],
                        "last": "M\u00f3ra",
                        "suffix": ""
                    },
                    {
                        "first": "J\u00e1nos",
                        "middle": [],
                        "last": "Csirik",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "BMC Bioinformatics",
                "volume": "9",
                "issue": "11",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Veronika Vincze, Gy\u00f6rgy Szarvas, Rich\u00e1rd Farkas, Gy\u00f6rgy M\u00f3ra, and J\u00e1nos Csirik. 2008. The Bio- Scope Corpus: Biomedical Texts Annotated for Uncertainty, Negation and their Scopes. BMC Bioinformatics, 9(Suppl 11):S9.",
                "links": null
            },
            "BIBREF26": {
                "ref_id": "b26",
                "title": "Annotating expressions of opinions and emotions in language. Language Resources and Evaluation",
                "authors": [
                    {
                        "first": "Janyce",
                        "middle": [],
                        "last": "Wiebe",
                        "suffix": ""
                    },
                    {
                        "first": "Theresa",
                        "middle": [],
                        "last": "Wilson",
                        "suffix": ""
                    },
                    {
                        "first": "Claire",
                        "middle": [],
                        "last": "Cardie",
                        "suffix": ""
                    }
                ],
                "year": 2005,
                "venue": "",
                "volume": "39",
                "issue": "",
                "pages": "164--210",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Janyce Wiebe, Theresa Wilson, and Claire Cardie. 2005. Annotating expressions of opinions and emo- tions in language. Language Resources and Evalu- ation, 39(2-3):164-210.",
                "links": null
            },
            "BIBREF27": {
                "ref_id": "b27",
                "title": "Subjectivity word sense disambiguation",
                "authors": [
                    {
                        "first": "Janyce",
                        "middle": [],
                        "last": "Wiebe",
                        "suffix": ""
                    }
                ],
                "year": 2012,
                "venue": "Proceedings of WASSA 2012",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Janyce Wiebe. 2012. Subjectivity word sense disam- biguation. In Proceedings of WASSA 2012, page 2, Jeju, Korea, July. ACL.",
                "links": null
            },
            "BIBREF28": {
                "ref_id": "b28",
                "title": "Fine-grained Subjectivity and Sentiment Analysis: Recognizing the Intensity, Polarity, and Attitudes of Private States",
                "authors": [
                    {
                        "first": "Wilson",
                        "middle": [],
                        "last": "Theresa Ann",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Theresa Ann Wilson. 2008. Fine-grained Subjectivity and Sentiment Analysis: Recognizing the Intensity, Polarity, and Attitudes of Private States. Ph.D. the- sis, University of Pittsburgh, Pittsburgh.",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF0": {
                "type_str": "figure",
                "num": null,
                "uris": null,
                "text": "Types of hedges."
            },
            "TABREF1": {
                "html": null,
                "type_str": "table",
                "text": "Uncertainty cues in WikiWeasel.",
                "num": null,
                "content": "<table/>"
            },
            "TABREF3": {
                "html": null,
                "type_str": "table",
                "text": "The most frequent discourse-level uncertainty cues in the WikiWeasel corpus.",
                "num": null,
                "content": "<table/>"
            }
        }
    }
}