File size: 86,131 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
{
    "paper_id": "O12-3003",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T08:03:13.837516Z"
    },
    "title": "Frequency, Collocation, and Statistical Modeling of Lexical Items: A Case Study of Temporal Expressions in Two Conversational Corpora 1",
    "authors": [
        {
            "first": "Sheng-Fu",
            "middle": [],
            "last": "Wang",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "National Taiwan University",
                "location": {
                    "addrLine": "3F, Le-Xue Building, No. 1, Sec. 4, Roosevelt Rd",
                    "postCode": "106",
                    "settlement": "Taipei Taiwan"
                }
            },
            "email": ""
        },
        {
            "first": "Jing-Chen",
            "middle": [],
            "last": "Yang",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "National Taiwan University",
                "location": {
                    "addrLine": "3F, Le-Xue Building, No. 1, Sec. 4, Roosevelt Rd",
                    "postCode": "106",
                    "settlement": "Taipei Taiwan"
                }
            },
            "email": ""
        },
        {
            "first": "Yu-Yun",
            "middle": [],
            "last": "Chang",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "National Taiwan University",
                "location": {
                    "addrLine": "3F, Le-Xue Building, No. 1, Sec. 4, Roosevelt Rd",
                    "postCode": "106",
                    "settlement": "Taipei Taiwan"
                }
            },
            "email": ""
        },
        {
            "first": "Yu-Wen",
            "middle": [],
            "last": "Liu",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "National Taiwan Normal University",
                "location": {
                    "addrLine": "No. 162, He-ping East Road, Section 1, Taipei",
                    "postCode": "106",
                    "country": "Taiwan"
                }
            },
            "email": ""
        },
        {
            "first": "Shu-Kai",
            "middle": [],
            "last": "Hsieh",
            "suffix": "",
            "affiliation": {
                "laboratory": "",
                "institution": "National Taiwan University",
                "location": {
                    "addrLine": "3F, Le-Xue Building, No. 1, Sec. 4, Roosevelt Rd",
                    "postCode": "106",
                    "settlement": "Taipei Taiwan"
                }
            },
            "email": "shukaihsieh@ntu.edu.tw"
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "This study examines how different dimensions of corpus frequency data may affect the outcome of statistical modeling of lexical items. Our analysis mainly focuses on a recently constructed elderly speaker corpus that is used to reveal patterns of aging people's language use. A conversational corpus contributed by speakers in their 20s serves as complementary material. The target words examined are temporal expressions, which might reveal how the speech produced by the elderly is organized. We conduct divisive hierarchical clustering analyses based on two different dimensions of corporal data, namely raw frequency distribution and collocation-based vectors. When different dimensions of data were used as the input, results showed that the target terms were clustered in different ways. Analyses based on frequency distributions and collocational patterns are distinct from each other. Specifically, statistically-based collocational analysis generally produces more distinct clustering results that differentiate temporal terms more delicately than do the ones based on raw frequency.",
    "pdf_parse": {
        "paper_id": "O12-3003",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "This study examines how different dimensions of corpus frequency data may affect the outcome of statistical modeling of lexical items. Our analysis mainly focuses on a recently constructed elderly speaker corpus that is used to reveal patterns of aging people's language use. A conversational corpus contributed by speakers in their 20s serves as complementary material. The target words examined are temporal expressions, which might reveal how the speech produced by the elderly is organized. We conduct divisive hierarchical clustering analyses based on two different dimensions of corporal data, namely raw frequency distribution and collocation-based vectors. When different dimensions of data were used as the input, results showed that the target terms were clustered in different ways. Analyses based on frequency distributions and collocational patterns are distinct from each other. Specifically, statistically-based collocational analysis generally produces more distinct clustering results that differentiate temporal terms more delicately than do the ones based on raw frequency.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "The study of gerontology is gaining wide attention, as the aging population is becoming a major issue in society. Research has noted that aging causes not only physiological changes for elderly people, but also affects their language production (Burk & Shafto, 2004) , cognitive load (Wilson, 2008) , context processing speed (Rush, Barch, & Braver 2006) , language performance patterns compared to younger individuals (Veliz, Riffo, & Arancibia, 2010) , etc.",
                "cite_spans": [
                    {
                        "start": 245,
                        "end": 266,
                        "text": "(Burk & Shafto, 2004)",
                        "ref_id": null
                    },
                    {
                        "start": 284,
                        "end": 298,
                        "text": "(Wilson, 2008)",
                        "ref_id": "BIBREF24"
                    },
                    {
                        "start": 326,
                        "end": 354,
                        "text": "(Rush, Barch, & Braver 2006)",
                        "ref_id": "BIBREF21"
                    },
                    {
                        "start": 419,
                        "end": 452,
                        "text": "(Veliz, Riffo, & Arancibia, 2010)",
                        "ref_id": "BIBREF22"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1."
            },
            {
                "text": "To study gerontology from a linguistic point of view, Green (1993) proposed that the phenomenon of gerontology could be studied through discourse analysis. Therefore, we use conversations the elderly and from younger speakers as our speech corpora and take these corpora as an input to exemplify the procedures and usage of lexical modeling.",
                "cite_spans": [
                    {
                        "start": 54,
                        "end": 66,
                        "text": "Green (1993)",
                        "ref_id": "BIBREF8"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1."
            },
            {
                "text": "As for the examination of temporal terms, we postulated that people's social roles, especially the elderly's, might be embedded in conversations where speakers share personal experiences or judgment of the past (Kuo, 2008) and the present. Thus, we presume that temporal expressions might serve as the anchoring points in a conversation-based corpus and might reveal a certain aspect of the speech behavior patterns of the elderly.",
                "cite_spans": [
                    {
                        "start": 211,
                        "end": 222,
                        "text": "(Kuo, 2008)",
                        "ref_id": "BIBREF14"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1."
            },
            {
                "text": "Statistical modeling can serve to describe a given set of data, be it diachronic subsets, register, or lexical units. Statistical models often take the so-called \"bottom-up\" approach that suits most corpus linguists' empirical state of mind. Moreover, nice and neat visualization is often a feat in such modeling techniques, to an extent that some of the models are called \"graph models\" (Widdows & Dorow, 2002) . When the proper behavior of lexical units and the structure of the lexicon are applied, statistical modeling may help us develop NLP-oriented lexicographic modules in the form of dictionaries, thesauri, and ontologies (Mitrofanova, Mukhin, Panicheva, & Savitsky, 2007) .",
                "cite_spans": [
                    {
                        "start": 388,
                        "end": 411,
                        "text": "(Widdows & Dorow, 2002)",
                        "ref_id": "BIBREF23"
                    },
                    {
                        "start": 632,
                        "end": 682,
                        "text": "(Mitrofanova, Mukhin, Panicheva, & Savitsky, 2007)",
                        "ref_id": "BIBREF17"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1."
            },
            {
                "text": "A glimpse at relevant studies would reveal that the most prominent kind of data input is related to the distributional patterns of the lexical items in corpora. The distributional data could be in the form of word frequencies and variability of frequencies (Gries & Hilpert, to appear) or the distribution of n-grams as a whole (Gries, Newman, & Shaoul, 2011) . The distributional pattern or dependency with syntactic patterns is also a prominent source of data input (Cimiano, Hotho, & Staab, 2004a , 2004b Lin, 1998; Pereira & Tishby 1992) . Target lexical items' dependency and co-occurrence with particular word types also has been taken as the basis of lexical modeling in some studies (Redington, Chater, & Finch, 1993) . Moreover, statistically-based collocational patterns are used for modeling similarities among lexical units of interest (Chen, 2009; Gries & Stefanowitsch, to appear) .",
                "cite_spans": [
                    {
                        "start": 257,
                        "end": 285,
                        "text": "(Gries & Hilpert, to appear)",
                        "ref_id": null
                    },
                    {
                        "start": 328,
                        "end": 359,
                        "text": "(Gries, Newman, & Shaoul, 2011)",
                        "ref_id": "BIBREF12"
                    },
                    {
                        "start": 468,
                        "end": 499,
                        "text": "(Cimiano, Hotho, & Staab, 2004a",
                        "ref_id": "BIBREF3"
                    },
                    {
                        "start": 500,
                        "end": 507,
                        "text": ", 2004b",
                        "ref_id": "BIBREF4"
                    },
                    {
                        "start": 508,
                        "end": 518,
                        "text": "Lin, 1998;",
                        "ref_id": "BIBREF16"
                    },
                    {
                        "start": 519,
                        "end": 541,
                        "text": "Pereira & Tishby 1992)",
                        "ref_id": "BIBREF18"
                    },
                    {
                        "start": 691,
                        "end": 725,
                        "text": "(Redington, Chater, & Finch, 1993)",
                        "ref_id": "BIBREF20"
                    },
                    {
                        "start": 848,
                        "end": 860,
                        "text": "(Chen, 2009;",
                        "ref_id": "BIBREF2"
                    },
                    {
                        "start": 861,
                        "end": 894,
                        "text": "Gries & Stefanowitsch, to appear)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1."
            },
            {
                "text": "The above-mentioned different methods, or rather, different data inputs, are considered to fall somewhere between raw distributional data and relational data, or between lexical items and syntactic patterns. In our study, we aim to compare the two endpoints of this methodological continuum, namely the \"frequency distributional data\" input and \"collocation data,\" in order to see how these different types of inputs may result in different data in lexical modeling. The former is the analysis based on the raw distributional data of lexical items, and the latter is based on more fine-grained examination of the relationship between lexical items, such as a particular item's collocational pattern with other lexical units in a corpus. By comparing the two kinds of analyses, we hope to reveal the extent to which these two methods demonstrate different patterns, thereby making contributions to the evaluation and selection of research methods in modeling lexical items.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Case Study of Temporal Expressions in Two Conversational Corpora",
                "sec_num": "39"
            },
            {
                "text": "This paper is organized as follows: Section 2 introduces the corpora used in this study, including data collection, guidelines for transcription, and annotation standards. Section 3 reports basic corporal information and preliminary analysis of six selected temporal expressions from the corpus. Section 4 demonstrates the methods and results of statistical modeling of temporal expressions, as well as a meta-analysis on different models. Section 5 and Section 6 summarize our findings and suggestions for further research respectively.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Case Study of Temporal Expressions in Two Conversational Corpora",
                "sec_num": "39"
            },
            {
                "text": "Speech data were collected from four pairs of elderly people. Each pair consisted of one male and one female speaker. All subjects are native speakers of Mandarin and Taiwanese Southern Min. One pair is from Changhua while the others are from Taipei. The mean age of the subjects is 65.75 years old (SD = 6.16). Each pair of speakers was asked to do a face-to-face conversation in Mandarin for 30 to 40 minutes. The designated conversational topic was the speakers' life experience in the past and the present. During the recording, other participants, such as the subject's relatives or the observer, might also be involved in the talk. All files were recorded by a digital recorder in the WAV format. The total length of the speech samples is 145 minutes.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Elderly Speaker Corpus",
                "sec_num": "2.1"
            },
            {
                "text": "Speech samples collected from the elderly people's conversations were then transcribed into Chinese characters, following Du Bois et al.'s transcription standards for discourse analysis (Du Bois, Schuetze-Coburn, Cumming, & Paolino, 1993) . Prosodic features and vocal qualities of the intonation units (IUs) were excluded from the transcription since they are not the main interest in this study. A short guideline of transcription standards is provided below.",
                "cite_spans": [
                    {
                        "start": 190,
                        "end": 238,
                        "text": "Bois, Schuetze-Coburn, Cumming, & Paolino, 1993)",
                        "ref_id": "BIBREF7"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Elderly Speaker Corpus",
                "sec_num": "2.1"
            },
            {
                "text": "Conversation samples were manually processed into several IUs. Each IU was labeled with a number on the left, as shown in Example (1). Sometimes, the subjects switched between Mandarin and other languages when speaking. Such cases of code-switching were enclosed in square brackets and labeled with L2 as well as the code for that non-Mandarin language. Example (3) demonstrates a case where a speaker switched between Mandarin and Taiwanese Southern Min (TSM). It should be noted that, since this study targeted elderly people's Mandarin speech performance, code-switching to languages other than Mandarin was excluded from our analysis.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Elderly Speaker Corpus",
                "sec_num": "2.1"
            },
            {
                "text": "(3)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Elderly Speaker Corpus",
                "sec_num": "2.1"
            },
            {
                "text": "268 SF: [L2 TSM \u55ae\uf9d7\uf902 TSM L2] single wheeler",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Elderly Speaker Corpus",
                "sec_num": "2.1"
            },
            {
                "text": "Laughter was also identified in the transcription. Each syllable of laughter was labeled with one token of the symbol @ (see Example 4a). Longer laughter was indicated by a single symbol @ with the duration in the parentheses (see Example 4b). Two @ symbols were placed at each end of an IU to show that the subject spoke while laughing (see Example 4c). The occurrence and duration of a pause in discourse were transcribed. Pauses were represented by dots: two dots for short pauses of less than 0.3 seconds, three dots for medium pauses between 0.3 and 0.6 seconds, and three dots with the duration specified in parentheses for pauses longer than 0.7 seconds. Example (5) below is the instance for pauses. The recorded utterances were not always audible or clear enough for the transcribers to identify what was being said. Each syllable of uncertain hearing was labeled with a capital X, as shown in Example (5) above. Last but not least, truncated words or IUs were represented by double hyphens --, as shown in Examples (1) and (5).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Elderly Speaker Corpus",
                "sec_num": "2.1"
            },
            {
                "text": "The transcription was automatically segmented and POS (part of speech) tagged through the CKIP Chinese Word Segmentation System provided by the Chinese Knowledge Information Processing (CKIP) group at Academia Sinica (2004). The segmentation and POS standards were based on the Sinica Corpus guidelines (1998). The annotated language samples then were checked manually. The procedure is described below.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Elderly Speaker Corpus",
                "sec_num": "2.1"
            },
            {
                "text": "First, every segmentation result derived from CKIP was examined and corrected if wrong, as in the following examples. Example (7a) is the original IU before segmentation and tagging. Through CKIP, we attain the result in Example (7b), which is falsely processed. Example 7cshows the proper segmentation after manual correction. Second, POS tags were viewed as correct if the main word classes were correct, while the details of their sub-classes were not of primary concern. For instance, in Example (8), the main word class of each POS tag (in this case, N, DE, V, or D) was examined, but not the sub-class tagging, as we give less consideration for the subcategories they belong to. Third, particles' tags were manually corrected to I for IU-initial particles 2 , and T for IU-final particles. If an IU contained nothing but particles, then the particles were tagged as I.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Elderly Speaker Corpus",
                "sec_num": "2.1"
            },
            {
                "text": "Finally, POS tags were removed for truncations (e.g. \u9019--), uncertain hearing (i.e. X), and code-switching. Given that truncations are not generally viewed as lexical items, they were not suitable for analysis at the lexical level.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The Elderly Speaker Corpus",
                "sec_num": "2.1"
            },
            {
                "text": "This corpus contains speech data collected by Master's students of the Graduate Institute of Linguistics at the National Taiwan University. The transcription follows Du Bois et al.'s standards for discourse analysis (1993) . The data was collected by graduate students in their early 20s, and most of the recruited speakers were similar in age to the data collectors. In other words, this corpus contains mostly \"youth speech,\" which is suitable as a complementary corpus to the elderly speakers' corpus for our analysis.",
                "cite_spans": [
                    {
                        "start": 216,
                        "end": 222,
                        "text": "(1993)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The NTU Conversational Corpus",
                "sec_num": "2.2"
            },
            {
                "text": "Although constant effort has been made in data collection of the NTU conversational corpus for more than ten years, little effort has been devoted to data organization and preprocessing. For this study, we selected a subset of face-to-face conversations between speakers (mostly students) less than 30 years old. The topics of these conversations were mostly everyday life experiences. The chosen subset, containing around 66,000 words, was tokenized and annotated the same way as the elderly speakers' corpus for further analysis.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The NTU Conversational Corpus",
                "sec_num": "2.2"
            },
            {
                "text": "The elderly speaker corpus contains 4,982 IUs of Mandarin utterances and 22,090 word tokens produced by all speakers. Elderly people's production in Mandarin contains 3,739 IUs, and there are 18,076 word tokens in total. The subset of the NTU conversational corpus used in this study contains 15,863 IUs and 65,742 word tokens.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Corpus Information & Preliminary Analysis",
                "sec_num": "3."
            },
            {
                "text": "The corpus processing tool used here is R (R. D. C. Team, 2010) , which allows us to perform tasks, including preprocessing, word frequency, KWIC (KeyWord In Context) extraction, and statistical modeling.",
                "cite_spans": [
                    {
                        "start": 52,
                        "end": 63,
                        "text": "Team, 2010)",
                        "ref_id": "BIBREF19"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Case Study of Temporal Expressions in Two Conversational Corpora",
                "sec_num": null
            },
            {
                "text": "We assume that time-related words may provide some vital clues to the elderly's and the youth's speech patterns, so the following analyses will focus on the subjects' use of temporal expressions. We are interested in how elderly people use \u73fe\u5728 (now) and \u4ee5\u524d (before), as well as other temporal expressions (tagged as Nd), to frame the present-and the past-related concepts, and how their use differs from the younger generations' employment of temporal expressions. Thus, the six most frequent temporal expressions from each corpus were selected for the analysis. Table 1 lists the frequency of the six target temporal expressions. As shown in the rankings, \"now\" is the most frequent temporal expression in both corpora. ",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 562,
                        "end": 569,
                        "text": "Table 1",
                        "ref_id": "TABREF1"
                    }
                ],
                "eq_spans": [],
                "section": "A Case Study of Temporal Expressions in Two Conversational Corpora",
                "sec_num": null
            },
            {
                "text": "In this section, we will present quantitative analyses with the help of hierarchical clustering, a data-driven approach, to see how the temporal terms of interest are grouped together with the frequency data extracted from our corpus.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Statistical Modeling of Temporal Expressions",
                "sec_num": "4."
            },
            {
                "text": "The clustering method employed here is divisive hierarchical clustering. This differs from agglomerative hierarchical clustering in that a group of entities is first divided into large groups before smaller groups are classified. Such a method is useful for finding a few clusters that are large in size (Rush, Barch, & Braver, 2006) . We would like to find out whether the terms for \"the present\" and \"the past\" can really be grouped into clusters that are different in temporality. Thus, divisive hierarchical clustering serves our need. In our current study, the clustering was conducted with the help of the dist() function in R (2010), which takes a table of correlations between the vectors containing different temporal expressions' frequency data or collocational data in the corpus and transforms it into a \"table of distances\" based on the square distance between these vectors.",
                "cite_spans": [
                    {
                        "start": 304,
                        "end": 333,
                        "text": "(Rush, Barch, & Braver, 2006)",
                        "ref_id": "BIBREF21"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Statistical Modeling of Temporal Expressions",
                "sec_num": "4."
            },
            {
                "text": "We executed a series of hierarchical clustering with different data input. The first analysis was run with the frequencies of the temporal terms across different files/texts in our corpus. Such an input was expected to capture the co-occurrence pattern of these temporal terms affected by individual speaker's style or idiolect, as well as by differences in the topic of conversation. The output is presented in Figure 1 , where \u73fe\u5728 (now) is separate from \u4ee5\u524d (before) under a major cluster on the left. Also, \u6700\u8fd1 (recently) stands independently from the other expressions, suggesting that temporal terms within a particular time domain are more likely to occur in the same text, which is really a conversational event in our corpora.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 412,
                        "end": 420,
                        "text": "Figure 1",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Modeling Results from the Elderly Speaker Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "Next, four clustering analyses were made based on the frequency data across subsets of different sizes. The sizes chosen for producing subsets were 10, 50, 200, and 500 words. Smaller subsets may reflect linguistic patterns in a few clauses, and larger subsets may reflect patterns in a larger unit, such as major or minor topics in the flow of conversation. As we can see in Figure 2 , \u73fe\u5728 (now) and \u4ee5\u524d (before) are classified in the same small cluster.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 376,
                        "end": 384,
                        "text": "Figure 2",
                        "ref_id": "FIGREF4"
                    }
                ],
                "eq_spans": [],
                "section": "Modeling Results from the Elderly Speaker Corpus",
                "sec_num": "4.1"
            },
            {
                "text": "Frequency We can also conduct clustering analysis according to how these terms collocate with other words in the corpus, on the premise that collocational patterns should reveal some characteristics of lexical items. Thus, two more analyses based on this assumption were given. The first analysis was done using each word type's collocational pattern (span = 3) with the six temporal terms as input. The second analysis was achieved through the dependency patterns of sentential particles (i.e. lah, hoNh, ah, oh, le, haNh, hioh, mah, as described by Li, 1999) , taking the temporal terms as input. There are two reasons for the inclusion of particle collocation. First, with regard to methodology, running more than one collocational test allows one to see whether collocational analyses with different approaches generate similar results. Second, sentential particles' dependency patterns might help us understand how the \"referent\" of each temporal expression is conceived and presented in discourse. The outcome is illustrated in Figure 3 . Again, \u73fe\u5728 (now) and \u4ee5\u524d (before) are clustered closely, showing that their collocational patterns might be similar.",
                "cite_spans": [
                    {
                        "start": 551,
                        "end": 560,
                        "text": "Li, 1999)",
                        "ref_id": "BIBREF15"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 1034,
                        "end": 1042,
                        "text": "Figure 3",
                        "ref_id": "FIGREF5"
                    }
                ],
                "eq_spans": [],
                "section": "Figure 1. Clustering based on frequencies in texts in the elderly speaker corpus. \"Height\" in the y-axis represents the furthest distance between the entities under a particular node in terms of the distance, based on the correlation of data vectors of these expressions. Thus, it is sensitive to how far apart the entities in question are.",
                "sec_num": null
            },
            {
                "text": ", Collocation",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Figure 1. Clustering based on frequencies in texts in the elderly speaker corpus. \"Height\" in the y-axis represents the furthest distance between the entities under a particular node in terms of the distance, based on the correlation of data vectors of these expressions. Thus, it is sensitive to how far apart the entities in question are.",
                "sec_num": null
            },
            {
                "text": "There is a potential problem in using raw frequencies in studying collocates. Collocates with high frequencies might simply be high frequency words, rather than being \"exclusively close\" to the terms of interest. Thus, we bring forth collexeme analysis (Gries, Hampe, & Sch\u00f6nefeld, 2005; Gries, 2007) , a statistical method developed for finding \"true collocates,\" that is, collocates with strong collocational strength (coll.strength hereafter). The coll.strength of each word type and particle was calculated and used as input for clustering analysis. The output is shown in Figure 4 . To sum up, \u73fe\u5728 (now) and \u4ee5\u524d (before) seem to intertwine, concerning their occurrences in different subsets of the corpus. This may suggest that, when elderly speakers talk about the past, the present follows as a contrast in time regarding the same subject matter, and vice versa. Only the by-text analysis shows a difference between the terms for the present and those for the past, suggesting that there are some conversations featuring more present than the past, and vice versa. Collocational strength analysis is another approach revealing a difference between \u73fe\u5728 (now) and \u4ee5\u524d (before), showing that, although the two terms usually are used closely, they still attract different words with different strengths. It should be noted that association patterns with particles do not distinguish between the present and the past. A possible explanation for this is that such a difference in pragmatic and discourse meaning is too fine-grained to be shown with information based on quantitative data. In other words, it shows that a quantitative method with corpus data has its limitation, especially when the annotation only functions at the basic POS level. Such findings of the temporal terms may in turn suggest that modeling lexical items is not a simple matter of looking for any types of analyzable data input. In addition to surface frequencies, taking collocational patterns into account, especially those based on statistical analyses, seems to be a requirement to capture the nuances among lexical items.",
                "cite_spans": [
                    {
                        "start": 253,
                        "end": 287,
                        "text": "(Gries, Hampe, & Sch\u00f6nefeld, 2005;",
                        "ref_id": "BIBREF10"
                    },
                    {
                        "start": 288,
                        "end": 300,
                        "text": "Gries, 2007)",
                        "ref_id": "BIBREF9"
                    }
                ],
                "ref_spans": [
                    {
                        "start": 577,
                        "end": 585,
                        "text": "Figure 4",
                        "ref_id": "FIGREF6"
                    }
                ],
                "eq_spans": [],
                "section": "Figure 1. Clustering based on frequencies in texts in the elderly speaker corpus. \"Height\" in the y-axis represents the furthest distance between the entities under a particular node in terms of the distance, based on the correlation of data vectors of these expressions. Thus, it is sensitive to how far apart the entities in question are.",
                "sec_num": null
            },
            {
                "text": "We performed similar analyses for the selected portion of the NTU conversational corpus, with a few modifications. First, the analysis based on the frequencies of temporal expressions in corpus subsets was only conducted with subset sizes of 10, 50, and 200 words, since the individual texts (conversations) in the NTU corpus are mostly not as long as those in our elderly speakers' corpus. For certain conversations, a 500-word window can cover almost all of the words in the text, making the analysis too similar to the by-text analysis we conducted. Second, the analyses of temporal expressions' raw collocation and collocational strength with particles were not conducted for the NTU corpus because the manual annotation on collocation between particles and temporal terms has not yet been completed. Figure 5 shows the results of clustering analysis based on raw frequencies of the temporal expressions in the NTU conversational corpus. Temporal scope seems to be the factor determining the clustering patterns. Expressions that have a less definite time frame, such as \u5f8c\uf92d (later) and \u4ee5\u524d (before), are grouped together, whereas expressions denoting a specific temporal scope, such as \u4eca\u5929 (today) and \u6628\u5929 (yesterday), are grouped under the same node. Figure 6 shows clustering results based on the target expressions' frequencies in the subsets of the data. The results are not clear-cut, lacking a consistent pattern across the outputs from different inputs. One of the more consistent patterns may be that \u73fe\u5728 (now) and \u5f8c\uf92d (later) are clustered together in two of the graphs above and are rather closely clustered in one of the two. This might imply that, within a context with a size smaller than the whole text, these two temporal expressions are commonly used together to form conversations.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 805,
                        "end": 813,
                        "text": "Figure 5",
                        "ref_id": "FIGREF7"
                    },
                    {
                        "start": 1253,
                        "end": 1261,
                        "text": "Figure 6",
                        "ref_id": "FIGREF8"
                    }
                ],
                "eq_spans": [],
                "section": "Modeling Results from the NTU (Youth) Conversational Corpus",
                "sec_num": "4.2"
            },
            {
                "text": "As for the result of the target temporal expressions' collocates in raw frequencies, clustering patterns differentiate expressions for the past, the present, and those with different time scopes. In Figure 7 , the expression \u4eca\uf98e (this year), which denotes the present within a bigger scope, is clustered away from other expressions. The expressions \u73fe\u5728 (now) and \u4eca \u5929 (today), both denoting the present within a smaller scope, are clustered together. Two expressions for the past, \u4ee5\u524d (before) and \u6628\u5929 (yesterday), are grouped under the same node. Finally, \u5f8c\uf92d (later), an expression for denoting time sequences, is clustered alone. This indicates that different types of temporal expressions collocate with other words in the corpus in different manners. Yet, when the input data are collocational strength patterns, which should reveal these expressions' true collocates, the results of the clustering change. \u73fe\u5728 (now) and \u4eca\u5929 (today), two expressions about the present, stand out in two distinct clusters, while \u5f8c\uf92d (later) also stands out to a certain extent. This reveals that expressions about the present are strongly and uniquely collocated with a certain group of words in the youth corpus, compared to other temporal terms. Such a result of modeling may imply that the youth use some particular patterns in structuring events or topics about the present.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 199,
                        "end": 207,
                        "text": "Figure 7",
                        "ref_id": "FIGREF9"
                    }
                ],
                "eq_spans": [],
                "section": "Modeling Results from the NTU (Youth) Conversational Corpus",
                "sec_num": "4.2"
            },
            {
                "text": "How do we evaluate all of these different results? The answer may not be surprising: We can do it with clustering analysis. The \"clustering\" package for R offers a function \"cutree\" for a simple quantification of different clustering, where each \"tree\" is quantified in terms of which cluster an item is clustered to. We collect the data for all of the trees shown above and execute clustering as meta-analysis. The outcome for the elderly speaker corpus is shown in Figure 8 , and the one for the NTU conversational corpus is shown in Figure 9 . An interesting pattern shows up in Figure 8 . There are two major clusters. The left one is based on frequency patterns of temporal terms, and the right one basically contains analyses of how these terms collocate or associate with other words or particles. Despite the curious occurrence of the \"by-500-words\" analysis in the right major cluster, the result of this meta-analysis seems to be able to characterize the major differences in terms of data input. More specifically, in the left major cluster, the \"by-text\" analysis is the first one singled out. This conforms to our impression that temporal terms are clustered differently, with \u73fe\u5728 (now) and \u6700\u8fd1 (recently) placed relatively far away from other past-related expressions. Moreover, in the right major cluster, the analyses with coll.strength are the first ones differentiated from the others. Again, this reflects that statistically-based analyses produce The evaluation of the clustering results of the NTU conversational corpus, shown in Figure 9 , clearly is different from the one of the elderly speakers' corpus. The only similar pattern is that the by-text analysis still stands out as a particular kind of clustering. Other clustering analyses are lumped together with no clearly emerging pattern. There are several possible reasons for such a result. First, one may argue that the relationship between clustering methods and their results does not hold a consistent pattern. On the other hand, it could mean that there are some differences in the use of the temporal items between these two corpora, hence, between speakers different in age. At this stage, the two corpora are not equal in size, so any claims about how elderly speakers' linguistic patterns actually differ from those of younger speakers' would appear to be unsound. To sum up, the evaluation method we propose here is a technique that can possibly reveal differences among methods of modeling or even differences between various corpora, as shown in our preliminary research results. For further development, similar investigations on better-controlled and comparable corpora must be conducted.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 467,
                        "end": 475,
                        "text": "Figure 8",
                        "ref_id": "FIGREF10"
                    },
                    {
                        "start": 536,
                        "end": 544,
                        "text": "Figure 9",
                        "ref_id": "FIGREF12"
                    },
                    {
                        "start": 582,
                        "end": 590,
                        "text": "Figure 8",
                        "ref_id": "FIGREF10"
                    },
                    {
                        "start": 1549,
                        "end": 1557,
                        "text": "Figure 9",
                        "ref_id": "FIGREF12"
                    }
                ],
                "eq_spans": [],
                "section": "Evaluation on Modeling Results",
                "sec_num": "4.3"
            },
            {
                "text": "Statistical modeling based on different types of data input displays different patterns, with modeling derived from frequencies and collocational patterns forming two major clusters (as revealed in the meta-analysis, which visualizes the difference between models based on quantitative data). In the \"frequency\" cluster, analysis based on distributional patterns is differentiated from the ones based on arbitrarily divided subsets. In the \"collocation\" cluster, statistically-oriented (i.e. collocational strength) analyses are distinguished from those based on surface collocational frequencies. For our present study, these findings are not overwhelmingly surprising because it is not hard to imagine the impact of the difference in texts and subsets on research, as well as the impact of surface frequencies and statistically-calculated relational patterns. Yet, when it comes to evaluating more types of modeling methods or inputs, meta-analysis of this kind provides a valuable means of choosing adequate methods. For instance, when researchers try to model different aspects of the lexical structure, the hierarchical modeling proposed here may help avoiding utilizing methods that are in fact very similar.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusion",
                "sec_num": "5."
            },
            {
                "text": "According to our analysis on temporal terms, the core expressions of the present and the past have very similar distributional patterns, showing that elderly speakers in the corpus tend to compare the present with the past in the same textual domains. The difference between these terms is disclosed only in models based on by-text frequency and statistical collocational analysis. The former shows that different speakers or conversation events may have their own preferred usage of temporal expressions. The latter indicates that these terms are still different in terms of their collocations, yet the difference can only be revealed through statistical tests on \"true collocates\" proposed by Gries (2007) . These findings can be seen as a pilot result of the linguistic patterns of aging people and young people in comparison.",
                "cite_spans": [
                    {
                        "start": 695,
                        "end": 707,
                        "text": "Gries (2007)",
                        "ref_id": "BIBREF9"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusion",
                "sec_num": "5."
            },
            {
                "text": "The primary purpose of this study was to attempt to highlight certain methodologies applicable to an elderly speaker corpus through several statistical approaches, rather than recklessly leaping to a conclusion that some universal elderly speech patterns are found in our corpus. The inclusion of the analysis on a younger speaker corpus helps us become more cautious with the claims we can make about our statistical approaches. Yet, to further explore the issue and confirm the validity of potential general linguistic patterns discovered in the current research, we must carefully conduct qualitative analyses of each temporal expression and interpret the results with the evidence from the quantitative methods we adopted previously. For example, the expansion of the size of the elderly speaker corpus does may alter the outcome of our statistical modeling.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Future work",
                "sec_num": "6."
            },
            {
                "text": "Also, the inequality in terms of the size of our two corpora and the different ways of data collection make the direct comparison between the elderly speaker corpus and the NTU (youth) conversational corpus difficult. In the future, it might be advisable to collect speech materials from a small number of younger speakers by asking them to narrate personal experiences and stories just like what we asked the elderly speakers to do. By doing so, we can directly compare this small corpus contributed by younger speakers with our elderly corpus, to see whether we can prompt even similar high-frequency temporal expressions for comparing two corpora of two generations of speakers.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "A Case Study of Temporal Expressions in Two Conversational Corpora",
                "sec_num": "53"
            },
            {
                "text": "Acknowledgement: Thanks Wang Chun-Chieh, Liu Chun-Jui, Anna Lofstrand, and Hsu Chan-Chia for their involvement in the construction of the elderly speakers' corpus and the early development of this paper.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "According to the standards provided by Sinica Corpus, I represents \"interjections\" which usually occur in the IU-initial position.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            }
        ],
        "back_matter": [],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "Analyzing linguistic data. A practical introduction to statistics using R",
                "authors": [
                    {
                        "first": "R",
                        "middle": [
                            "H"
                        ],
                        "last": "Baayen",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Baayen, R. H. (2008). Analyzing linguistic data. A practical introduction to statistics using R. Cambridge: Cambridge University Press.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Aging and language production. Current directions in psychological science",
                "authors": [
                    {
                        "first": "D",
                        "middle": [
                            "M"
                        ],
                        "last": "Bruke",
                        "suffix": ""
                    },
                    {
                        "first": "M",
                        "middle": [
                            "A"
                        ],
                        "last": "Shafto",
                        "suffix": ""
                    }
                ],
                "year": 2004,
                "venue": "",
                "volume": "13",
                "issue": "",
                "pages": "21--24",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Bruke, D. M. & Shafto, M. A. (2004). Aging and language production. Current directions in psychological science, 13, 21-24.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "Corpus, lexicon, and construction: A quantitative corpus approach to Mandarin possessive construction",
                "authors": [
                    {
                        "first": "C.-H",
                        "middle": [],
                        "last": "Chen",
                        "suffix": ""
                    }
                ],
                "year": 2009,
                "venue": "International Journal of Computational Linguigstics and Chinese Language Processing",
                "volume": "14",
                "issue": "",
                "pages": "305--340",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Chen, C.-H. (2009). Corpus, lexicon, and construction: A quantitative corpus approach to Mandarin possessive construction. International Journal of Computational Linguigstics and Chinese Language Processing, 14, 305-340.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "Comparing conceptual, divisive and agglomerative clustering for learning taxonomies from text",
                "authors": [
                    {
                        "first": "P",
                        "middle": [],
                        "last": "Cimiano",
                        "suffix": ""
                    },
                    {
                        "first": "A",
                        "middle": [],
                        "last": "Hotho",
                        "suffix": ""
                    },
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Staab",
                        "suffix": ""
                    }
                ],
                "year": 2004,
                "venue": "Proceedings of the European Conference of Artificial Intelligence",
                "volume": "",
                "issue": "",
                "pages": "435--439",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Cimiano, P., Hotho, A., & Staab, S. (2004a). Comparing conceptual, divisive and agglomerative clustering for learning taxonomies from text. in Proceedings of the European Conference of Artificial Intelligence, 435-439.",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "Clustering concept hierarchies from text",
                "authors": [
                    {
                        "first": "P",
                        "middle": [],
                        "last": "Cimiano",
                        "suffix": ""
                    },
                    {
                        "first": "A",
                        "middle": [],
                        "last": "Hotho",
                        "suffix": ""
                    },
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Staab",
                        "suffix": ""
                    }
                ],
                "year": 2004,
                "venue": "Proceedings of LREC 2004",
                "volume": "",
                "issue": "",
                "pages": "1--4",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Cimiano, P., Hotho, A., & Staab, S. (2004b). Clustering concept hierarchies from text. in Proceedings of LREC 2004, 1-4.",
                "links": null
            },
            "BIBREF5": {
                "ref_id": "b5",
                "title": "CKIP Chinese word segmentation system",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Ckip",
                        "suffix": ""
                    }
                ],
                "year": 2004,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "CKIP. (2004). CKIP Chinese word segmentation system. Retrieved June 2, 2011 from http://ckipsvr.iis.sinica.edu.tw/",
                "links": null
            },
            "BIBREF6": {
                "ref_id": "b6",
                "title": "Introduction to Sinica Corpus: A tagged balance corpus for Mandarin",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Ckip",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "CKIP. (1998). Introduction to Sinica Corpus: A tagged balance corpus for Mandarin Chinese. Taipei, Taiwan (R.O.C.): Academia Sinica.",
                "links": null
            },
            "BIBREF7": {
                "ref_id": "b7",
                "title": "Outline of discourse transcription",
                "authors": [
                    {
                        "first": "Du",
                        "middle": [],
                        "last": "Bois",
                        "suffix": ""
                    },
                    {
                        "first": "J",
                        "middle": [
                            "W"
                        ],
                        "last": "Schuetze-Coburn",
                        "suffix": ""
                    },
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Cumming",
                        "suffix": ""
                    },
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Paolino",
                        "suffix": ""
                    },
                    {
                        "first": "D",
                        "middle": [],
                        "last": "",
                        "suffix": ""
                    }
                ],
                "year": 1993,
                "venue": "Talking data: Transcription and coding in discourse research",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Du Bois J. W., Schuetze-Coburn, S., Cumming, S. & Paolino, D. (1993). Outline of discourse transcription. In Edwards J. A. & Lampert M. D. (Eds.), Talking data: Transcription and coding in discourse research. Hillsdale: New Jersey, Lawrence Erlbau.",
                "links": null
            },
            "BIBREF8": {
                "ref_id": "b8",
                "title": "Gerontolgy and the social construction of old age",
                "authors": [
                    {
                        "first": "B",
                        "middle": [
                            "S"
                        ],
                        "last": "Green",
                        "suffix": ""
                    }
                ],
                "year": 1993,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Green, B. S. (1993). Gerontolgy and the social construction of old age. New York: Aldine De Gruyter.",
                "links": null
            },
            "BIBREF9": {
                "ref_id": "b9",
                "title": "Collostructional analysis: Computing the degree of association between words and words/constructions. Retrived May",
                "authors": [
                    {
                        "first": "S",
                        "middle": [
                            "T"
                        ],
                        "last": "Gries",
                        "suffix": ""
                    }
                ],
                "year": 2007,
                "venue": "",
                "volume": "30",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Gries, S. T. (2007). Collostructional analysis: Computing the degree of association between words and words/constructions. Retrived May 30, 2011 from: http://www.linguistics.ucsb.edu/faculty/stgries/teaching/groningen/coll.analysis.r",
                "links": null
            },
            "BIBREF10": {
                "ref_id": "b10",
                "title": "Converging evidence: Bringing together experimental and corpus data on the association of verbs and constructions",
                "authors": [
                    {
                        "first": "S",
                        "middle": [
                            "T"
                        ],
                        "last": "Gries",
                        "suffix": ""
                    },
                    {
                        "first": "B",
                        "middle": [],
                        "last": "Hampe",
                        "suffix": ""
                    },
                    {
                        "first": "D",
                        "middle": [],
                        "last": "Sch\u00f6nefeld",
                        "suffix": ""
                    }
                ],
                "year": 2005,
                "venue": "Cognitive Linguistics",
                "volume": "16",
                "issue": "",
                "pages": "635--676",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Gries, S. T., Hampe, B., & Sch\u00f6nefeld, D., (2005). Converging evidence: Bringing together experimental and corpus data on the association of verbs and constructions. Cognitive Linguistics, 16, 635-676.",
                "links": null
            },
            "BIBREF11": {
                "ref_id": "b11",
                "title": "Variability-based neighbor clustering: A bottom-up approach to periodization in historical linguistics",
                "authors": [
                    {
                        "first": "S",
                        "middle": [
                            "T"
                        ],
                        "last": "Gries",
                        "suffix": ""
                    },
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Hilpert",
                        "suffix": ""
                    }
                ],
                "year": null,
                "venue": "Handbook on the history of English: Rethinking approaches to the history of English",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Gries, S. T. & Hilpert, M. (to appear). Variability-based neighbor clustering: A bottom-up approach to periodization in historical linguistics. In Nevalainen, T. & Traugott, E. C., (Eds.), Handbook on the history of English: Rethinking approaches to the history of English. Oxford: Oxford University Press.",
                "links": null
            },
            "BIBREF12": {
                "ref_id": "b12",
                "title": "N-grams and the clustering of registers. Empirical language research",
                "authors": [
                    {
                        "first": "S",
                        "middle": [
                            "T"
                        ],
                        "last": "Gries",
                        "suffix": ""
                    },
                    {
                        "first": "J",
                        "middle": [],
                        "last": "Newman",
                        "suffix": ""
                    },
                    {
                        "first": "C",
                        "middle": [],
                        "last": "Shaoul",
                        "suffix": ""
                    }
                ],
                "year": 2011,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Gries, S. T., Newman, J. & Shaoul, C. (2011). N-grams and the clustering of registers. Empirical language research. Retrived May 28, 2011 from: http://ejournals.org.uk/ELR/article/2011/1",
                "links": null
            },
            "BIBREF13": {
                "ref_id": "b13",
                "title": "Cluster analysis and the identification of collexeme classes",
                "authors": [
                    {
                        "first": "S",
                        "middle": [
                            "T"
                        ],
                        "last": "Gries",
                        "suffix": ""
                    },
                    {
                        "first": "A",
                        "middle": [],
                        "last": "Stefanowitsch",
                        "suffix": ""
                    }
                ],
                "year": null,
                "venue": "Empirical and Experimental Methods in Cognitive/Functional Research",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Gries, S. T. & Stefanowitsch, A. (to appear). Cluster analysis and the identification of collexeme classes. In S. Rice and J. Newman (Eds.), Empirical and Experimental Methods in Cognitive/Functional Research. Stanford, CA: CSLI.",
                "links": null
            },
            "BIBREF14": {
                "ref_id": "b14",
                "title": "Discourse and aging: A sociolinguistic analysis of elderly speech in Taiwan (NSC96-2411-H007-024)",
                "authors": [
                    {
                        "first": "S.-H",
                        "middle": [],
                        "last": "Kuo",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Kuo, S.-H. (2008). Discourse and aging: A sociolinguistic analysis of elderly speech in Taiwan (NSC96-2411-H007-024). Taipei, Taiwan (R.O.C.): National Science Council.",
                "links": null
            },
            "BIBREF15": {
                "ref_id": "b15",
                "title": "Utterance-final particles in Taiwanese: A discourse-pragmatic analysis",
                "authors": [
                    {
                        "first": "C",
                        "middle": [
                            "I"
                        ],
                        "last": "Li",
                        "suffix": ""
                    }
                ],
                "year": 1999,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Li, C. I. (1999). Utterance-final particles in Taiwanese: A discourse-pragmatic analysis. Taipei, Taiwan (R.O.C.): Crane Publishing Co.",
                "links": null
            },
            "BIBREF16": {
                "ref_id": "b16",
                "title": "Automatic retrieval and clustering of similar words",
                "authors": [
                    {
                        "first": "D",
                        "middle": [],
                        "last": "Lin",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "Proceedings of the 17 th International Conference on Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "768--774",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Lin, D. (1998). Automatic retrieval and clustering of similar words. in Proceedings of the 17 th International Conference on Computational Linguistics, 768-774.",
                "links": null
            },
            "BIBREF17": {
                "ref_id": "b17",
                "title": "Automatic word clustering in Russian texts",
                "authors": [
                    {
                        "first": "O",
                        "middle": [],
                        "last": "Mitrofanova",
                        "suffix": ""
                    },
                    {
                        "first": "A",
                        "middle": [],
                        "last": "Mukhin",
                        "suffix": ""
                    },
                    {
                        "first": "P",
                        "middle": [],
                        "last": "Panicheva",
                        "suffix": ""
                    },
                    {
                        "first": "V",
                        "middle": [],
                        "last": "Savitsky",
                        "suffix": ""
                    }
                ],
                "year": 2007,
                "venue": "Proceedings of the 10th International Conference on Text",
                "volume": "",
                "issue": "",
                "pages": "85--91",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Mitrofanova, O., Mukhin, A., Panicheva, P., & Savitsky, V. (2007). Automatic word clustering in Russian texts. Proceedings of the 10th International Conference on Text, Speech and Dialogue, 85-91.",
                "links": null
            },
            "BIBREF18": {
                "ref_id": "b18",
                "title": "Distributional similarity, phase transitions and hierarchical clustering",
                "authors": [
                    {
                        "first": "F",
                        "middle": [],
                        "last": "Pereira",
                        "suffix": ""
                    },
                    {
                        "first": "N",
                        "middle": [],
                        "last": "Tishby",
                        "suffix": ""
                    }
                ],
                "year": 1992,
                "venue": "AAAI Fall Symposium",
                "volume": "",
                "issue": "",
                "pages": "108--112",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Pereira, F. & Tishby, N. (1992). Distributional similarity, phase transitions and hierarchical clustering. Probabilistic approaches to natural language, papers from 1992 AAAI Fall Symposium, 108-112.",
                "links": null
            },
            "BIBREF19": {
                "ref_id": "b19",
                "title": "R: A language and environment for statistical computing",
                "authors": [
                    {
                        "first": "R",
                        "middle": [
                            "D C"
                        ],
                        "last": "Team",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "R. D. C. Team. (2010). R: A language and environment for statistical computing. Retrieved June 1, 2010 from: http://www.R-project.org",
                "links": null
            },
            "BIBREF20": {
                "ref_id": "b20",
                "title": "Distributional information and the acquisition of linguistic categories: A statistical approach",
                "authors": [
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Redington",
                        "suffix": ""
                    },
                    {
                        "first": "N",
                        "middle": [],
                        "last": "Chater",
                        "suffix": ""
                    },
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Finch",
                        "suffix": ""
                    }
                ],
                "year": 1993,
                "venue": "Proceedings of the Fifteenth Annual Conference of the Cognitive Science Society",
                "volume": "",
                "issue": "",
                "pages": "848--853",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Redington, M., Chater, N., & Finch, S. (1993). Distributional information and the acquisition of linguistic categories: A statistical approach. in Proceedings of the Fifteenth Annual Conference of the Cognitive Science Society, 848-853.",
                "links": null
            },
            "BIBREF21": {
                "ref_id": "b21",
                "title": "Accounting for cognitive aging: Context processing, inhibition or processing speed? Aging",
                "authors": [
                    {
                        "first": "B",
                        "middle": [
                            "K"
                        ],
                        "last": "Rush",
                        "suffix": ""
                    },
                    {
                        "first": "D",
                        "middle": [
                            "M"
                        ],
                        "last": "Barch",
                        "suffix": ""
                    },
                    {
                        "first": "T",
                        "middle": [
                            "S"
                        ],
                        "last": "Braver",
                        "suffix": ""
                    }
                ],
                "year": 2006,
                "venue": "Neuropsychology and Cognition",
                "volume": "13",
                "issue": "",
                "pages": "588--610",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Rush, B. K., Barch, D. M., & Braver, T. S. (2006). Accounting for cognitive aging: Context processing, inhibition or processing speed? Aging, Neuropsychology and Cognition, 13, 588-610.",
                "links": null
            },
            "BIBREF22": {
                "ref_id": "b22",
                "title": "Cognitive aging and language processing: Relevant issues. Revista de Ling\u00fc\u00edstica Te\u00f3rica y Aplicada",
                "authors": [
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Veliz",
                        "suffix": ""
                    },
                    {
                        "first": "B",
                        "middle": [],
                        "last": "Riffo",
                        "suffix": ""
                    },
                    {
                        "first": "B",
                        "middle": [],
                        "last": "Arancibia",
                        "suffix": ""
                    }
                ],
                "year": 2010,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "75--103",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Veliz, M., Riffo, B., & Arancibia, B. (2010). Cognitive aging and language processing: Relevant issues. Revista de Ling\u00fc\u00edstica Te\u00f3rica y Aplicada, 75-103.",
                "links": null
            },
            "BIBREF23": {
                "ref_id": "b23",
                "title": "A graph model for unsupervised lexical acquisition",
                "authors": [
                    {
                        "first": "D",
                        "middle": [],
                        "last": "Widdows",
                        "suffix": ""
                    },
                    {
                        "first": "B",
                        "middle": [],
                        "last": "Dorow",
                        "suffix": ""
                    }
                ],
                "year": 2002,
                "venue": "Proceedings of the 19th International Conference on Computational Linguistics",
                "volume": "",
                "issue": "",
                "pages": "1093--1099",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Widdows, D. & Dorow, B. (2002). A graph model for unsupervised lexical acquisition. in Proceedings of the 19th International Conference on Computational Linguistics, 1093-1099.",
                "links": null
            },
            "BIBREF24": {
                "ref_id": "b24",
                "title": "The effects of cognitive load on gait in older adults (Doctoral dissertation)",
                "authors": [
                    {
                        "first": "K",
                        "middle": [
                            "R"
                        ],
                        "last": "Wilson",
                        "suffix": ""
                    }
                ],
                "year": 2008,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Wilson, K. R. (2008). The effects of cognitive load on gait in older adults (Doctoral dissertation). Florida State University, Tallahassee, FL.",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF0": {
                "text": "SM: @(3.3) c. 828 O: @\u6c92 \u90a3\u9ebc \u56b4\u91cd la@ not that serious P. Frequency, Collocation, and Statistical Modeling of Lexical Items: 41 A Case Study of Temporal Expressions in Two Conversational Corpora",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF1": {
                "text": "...eh ..\u90fd \u662f..\u7236\u6bcd...(0.9)\u505a X P. all is parents do X Particles were transcribed in phonetic transcription to avoid disagreement on the employment of homophonic Mandarin characters, as Example (6) shows. Phonetic transcriptions for the particles included la,hoNh, a, o, le, haNh, hioh, and ma.   (6)    26 SM: hoNh.. a \u6211\u5011 \u4e8c\u5341 \u5e7e \u6b72 \u7d50\u5a5a P. P. we twenty more age get.married",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF2": {
                "text": "My father is his mother's brother.\" b. *\u6211 \u7238\u7238 \u662f \u4ed6\u5abd \u5abd\u7684 \u54e5\u54e5 I father is he.mom mom.DE brother c. \u6211 \u7238\u7238 \u662f \u4ed6 \u5abd\u5abd \u7684 \u54e5\u54e5 I father is he mom DE brother",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF4": {
                "text": ", and Statistical Modeling of Lexical Items: 45 A Case Study of Temporal Expressions in Two Conversational Corpora Clustering based on frequencies across subsets in the elderly speaker corpus. Upper left, with subsets of a size of 10 words. Upper right, of 50 words. Lower left, of 200 words. Lower right, of 500 words.",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF5": {
                "text": "Clustering based on association/collocation frequencies. Left, with all word types in the elderly speaker corpus. Right, with particles.",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF6": {
                "text": "Clustering based on coll.strength patterns. Left, with all word types in the elderly speaker corpus. Right, with particles. Frequency, Collocation, and Statistical Modeling of Lexical Items: 47 A Case Study of Temporal Expressions in Two Conversational Corpora",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF7": {
                "text": "Clustering based on frequencies in texts in the NTU conversational corpus.",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF8": {
                "text": "Clustering based on frequencies across subsets in the NTU conversational corpus. Left, with subsets of 10 words. Middle, of 50 words. Right, of 200 words. Frequency, Collocation, and Statistical Modeling of Lexical Items: 49 A Case Study of Temporal Expressions in Two Conversational Corpora",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF9": {
                "text": "Left, clustering based on collocational frequencies with all the word types in the NTU conversational corpus. Right, clustering based on collocational strength patterns.",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF10": {
                "text": "Clustering of various results from the elderly speaker corpus.",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF11": {
                "text": "Frequency, Collocation, and Statistical Modeling of Lexical Items:   51    A Case Study of Temporal Expressions in Two Conversational Corpora different patterns from the ones based on simple frequency values. What can be inferred from the patterns inFigure 8is that different types of data input certainly influence the outcome of clustering analysis.",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "FIGREF12": {
                "text": "Clustering of various results from the NTU conversational corpus.",
                "type_str": "figure",
                "num": null,
                "uris": null
            },
            "TABREF1": {
                "text": "",
                "content": "<table><tr><td/><td>Elderly Speaker Corpus</td><td/><td/><td>NTU (Youth) Corpus</td><td/></tr><tr><td colspan=\"2\">Rank. Term</td><td colspan=\"2\">Freq. Rank.</td><td>Term</td><td>Freq.</td></tr><tr><td>1</td><td>\u73fe\u5728(now)</td><td>169</td><td>1</td><td>\u73fe\u5728(now)</td><td>137</td></tr><tr><td>2</td><td>\u4ee5\u524d(before)</td><td>169</td><td>2</td><td>\u5f8c\uf92d(later)</td><td>76</td></tr><tr><td>3</td><td>\u5c0f\u6642\u5019(in one's childhood)</td><td>12</td><td>3</td><td>\u4eca\u5929(today)</td><td>52</td></tr><tr><td>4</td><td>\u6c11\u570b(R.O.C. year)</td><td>11</td><td>4</td><td>\u4ee5\u524d(before)</td><td>52</td></tr><tr><td>5</td><td>\u7576\u521d(back then)</td><td>9</td><td>5</td><td>\u6628\u5929(yesterday)</td><td>34</td></tr><tr><td>6</td><td>\u6700\u8fd1(recently)</td><td>6</td><td>6</td><td>\u4eca\uf98e(this year)</td><td>31</td></tr></table>",
                "type_str": "table",
                "html": null,
                "num": null
            }
        }
    }
}