File size: 79,091 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 |
{
"paper_id": "2020",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T07:33:41.162686Z"
},
"title": "A pseudonymisation method for language documentation corpora: An experiment with spoken Komi",
"authors": [
{
"first": "Niko",
"middle": [],
"last": "Partanen",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Helsinki Helsinki",
"location": {
"country": "Finland"
}
},
"email": "niko.partanen@helsinki.fi"
},
{
"first": "Rogier",
"middle": [],
"last": "Blokland",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Uppsala University Uppsala",
"location": {
"country": "Sweden"
}
},
"email": "rogier.blokland@moderna.uu.se"
},
{
"first": "Michael",
"middle": [],
"last": "Rie\u00dfler",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Eastern Finland Joensuu",
"location": {
"country": "Finland"
}
},
"email": "michael.riessler@uef.fi"
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "This article introduces a novel and creative application of the Constraint Grammar formalism, by presenting an automated method for pseudonymising a Zyrian Komi spoken language corpus in an effective, reliable and scalable manner. The method is intended to be used to minimize various kinds of personal information found in the corpus in order to make spoken language data available while preventing the spread of sensitive personal data about the recorded informants or other persons mentioned in the texts. In our implementation, a Constraint Grammar based pseudonymisation tool is used as an automatically applied shallow layer that derives from the original corpus data a version which can be shared for open research use. Teesiq Seo artikli tutvustas vahts\u00f5t ja loovat piirdmiisi grammatiga (PG) formalism\u00f5 pruuk\u00b4mist. Taas om metod\u00b4, kon PG pruugitas tuusjaos, et s\u00fcr\u00e4komi k\u00f5n\u00f5ldu keele korpus\u00f5 lindistuisi saassiq teg\u00fcs\u00e4he, kimm\u00e4he ja kontrol\u00b4mis\u00f5v\u00f5imalus\u00f5ga va\u0155onimmiga k\u00e4kkiq. Seo metod\u00b4om tett, et korpus\u00f5n saassiq k\u00f5n\u00f5l\u00f5jid\u00f5 andmit nii pall\u00b4o v\u00e4hemb\u00e4s v\u00f5ttaq, ku v\u00f5i, ja et tul\u00f5mit saassiq kergehe k\u00e4silde kontrolliq. Mi plaani perr\u00e4 pruugitas taad ku automaats\u00f5t kihti, mi\u00e4 tege korpus\u00f5 s\u00e4\u00e4ndses, et taad v\u00f5i kergehe uu\u0155mis\u00f5 jaos jakaq.",
"pdf_parse": {
"paper_id": "2020",
"_pdf_hash": "",
"abstract": [
{
"text": "This article introduces a novel and creative application of the Constraint Grammar formalism, by presenting an automated method for pseudonymising a Zyrian Komi spoken language corpus in an effective, reliable and scalable manner. The method is intended to be used to minimize various kinds of personal information found in the corpus in order to make spoken language data available while preventing the spread of sensitive personal data about the recorded informants or other persons mentioned in the texts. In our implementation, a Constraint Grammar based pseudonymisation tool is used as an automatically applied shallow layer that derives from the original corpus data a version which can be shared for open research use. Teesiq Seo artikli tutvustas vahts\u00f5t ja loovat piirdmiisi grammatiga (PG) formalism\u00f5 pruuk\u00b4mist. Taas om metod\u00b4, kon PG pruugitas tuusjaos, et s\u00fcr\u00e4komi k\u00f5n\u00f5ldu keele korpus\u00f5 lindistuisi saassiq teg\u00fcs\u00e4he, kimm\u00e4he ja kontrol\u00b4mis\u00f5v\u00f5imalus\u00f5ga va\u0155onimmiga k\u00e4kkiq. Seo metod\u00b4om tett, et korpus\u00f5n saassiq k\u00f5n\u00f5l\u00f5jid\u00f5 andmit nii pall\u00b4o v\u00e4hemb\u00e4s v\u00f5ttaq, ku v\u00f5i, ja et tul\u00f5mit saassiq kergehe k\u00e4silde kontrolliq. Mi plaani perr\u00e4 pruugitas taad ku automaats\u00f5t kihti, mi\u00e4 tege korpus\u00f5 s\u00e4\u00e4ndses, et taad v\u00f5i kergehe uu\u0155mis\u00f5 jaos jakaq.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "The research presented in this paper is predominantly relevant for documentary linguistics, aiming at the creation of a \"lasting multipurpose record of a language\" (Himmelmann, 2006, 1) , while applying a computational linguistic approach to an endangered Uralic language. Specifically, we are developing an automated method for pseudonymising the textual representation of a spoken language corpus in order to make the corpus data publishable while 1) preventing the spread of sensitive personal data, 2) overcome manual work in the process to the extent possible, and 3) keeping the pseudonymised data as one -more openly distributed -version of the original -and less openly distributed -data, rather than destroying the latter by overwriting or cutting away parts of them.",
"cite_spans": [
{
"start": 164,
"end": 185,
"text": "(Himmelmann, 2006, 1)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "To our knowledge, this is a novel approach in documentary linguistics, which so far seems to rely mostly on manual methods for pseudonymising (or anonymising) corpus data or bypasses the problem, typically by generally applying very restrict access protocols to corpus data preventing them from being openly published.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Computational linguistic projects aiming at corpus building for endangered languages, on the other hand, are rarely faced the problem of personal data protection because their corpora typically originate from written texts, which either are openly available to begin with or for which access rights have been cleared before the work with corpus building starts. Different from written-language data, corpora resulting from fieldwork-based spoken-language documentation invariably contain large amounts and various kinds of personal data. The reason for this is that documentary linguists transcribe authentic speech samples meant to be re-used in multidisciplinary research beyond structural linguistics.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Typically, recordings are done with members of small communities, where most individuals know each other, and common topics include oral histories about places, persons or events inevitably including personal information about the informants itself or other individuals.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Although the recorded speaker's informed consent to further processing and re-using the speech sample needs to be at hand in any case, fieldworkbased recordings nearly always include information which should not be made entirely openly available. Simultaneously, it is in the natural interest of documentary linguistics to make as many materials as widely available as possible. The approach discussed in this paper consequently attempts to find a suitable middle way in presenting our Zyrian Komi corpus to a wider audience, in this case mainly researchers such as linguists and anthropologists, whilst ensuring that the privacy of individual speakers is respected and the risk of miss-using personal data can be excluded.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Best practice recommendations related to the problem of personal data protection in Open Science are currently evolving (cf. Seyfeddinipur et al., 2019) , although relevant issues have been under discussion for a while already in the context of in Documentary Linguistics. The conventions and technical solutions described on this study have been developed in a specific situation, where the goal has been to publish and archive the corpus in the Language Bank of Finland (Blokland et al., 2020) , and can hopefully contribute to solving at least some of the many challenges still remaining open.",
"cite_spans": [
{
"start": 125,
"end": 152,
"text": "Seyfeddinipur et al., 2019)",
"ref_id": "BIBREF21"
},
{
"start": 472,
"end": 495,
"text": "(Blokland et al., 2020)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Since GDPR-related\u00b9 research practices are still evolving in Finland and there are as yet no clear guidelines, it is currently problematic to make audiovisual research materials containing identifiable personal information available. As there are fewer limitations when the material is anonymised or pseudonymised, we have explored this as a solution: a version of a corpus that does not contain identifiable personal information can be openly shared much more easily. The current approach we are considering is to share the current corpus with academic users through the Korp interface (Ahlberg et al., 2013) under so-called ACA conditions. This ensures that the users are authenticated as members of the academic community, and their identity is known. At this level, however, they can only access the versions that have passed through the \u00b9The General Data Protection Regulation (EU) 2016/679 pseudonymisation system described in this paper. We refer to this method as pseudonymisation, since the actual identity information is not discarded permanently from existence. We just derive a version where the personal data is minimised, and provide that to one particular user group.",
"cite_spans": [
{
"start": 587,
"end": 609,
"text": "(Ahlberg et al., 2013)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "We think this approach can be a satisfactory middle way to make the data accessible to the scientific community without needlessly revealing the personal information of individual speakers. We aim, however, to make the complete corpus available for research use with a specific application procedure, as is common with language documentation corpora in other archives.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "These methods to share the corpora primarily serve academic users in Europe, but not the community itself, and to resolve this issue our colleagues from the community have also made a selection of the corpus available as a 'community edition' at a website videocorpora.ru\u00b2, which is maintained and curated by FU-Lab (the Finno-Ugric Laboratory for the Support of Electronic Representation of Regional Languages in Syktyvkar, Russia). This, however, though trilingual (Zyrian Komi, Russian, English), is designed mainly from the point of view of and for community members, containing edited video versions aimed at an uncomplicated user experience, and does not (and is not primarily supposed to) satisfy the needs of many academic users.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "In order to reliably pseudonymise the transcriptions in a version of the corpus aimed at research use we have developed a workflow that uses existing rule-based NLP for Zyrian Komi. The approach consists of performing an analysis on all running text, with various strategies to manipulate and filter out proper nouns, such as person names and toponyms. This allows us to keep some of the naturalness of the running text, while removing and changing easily identifiable content. Another benefit of our approach is that it lets us show which recordings contain which types of personal information.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Our method uses Finite-State Morphology (henceforth FST), specifically HFST (Lind\u00e9n et al., 2009) , and Constraint Grammar (henceforth CG) (Karlsson, 1990; Karlsson et al., 2011) , specifically the CG-3 version (Bick and Didriksen, 2015) , and is applied to the corpus using a uralicNLP Python package (H\u00e4m\u00e4l\u00e4inen, 2019) . The use of rule-based NLP methods is, in our opinion, highly desirable in this context as we can thus \u00b2http://videocorpora.ru carefully control the entire process, and be certain about the achieved result. Although we benefit from the existence of a highly advanced Komi morphological analyser (Rueter, 2000) , we believe that this approach could also be applicable to other language documentation projects. In principle the analyser used for such projects would only need a very small lexicon containing those items to be pseudonymised, which could be very easily connected to the project's internal metadata.",
"cite_spans": [
{
"start": 71,
"end": 97,
"text": "HFST (Lind\u00e9n et al., 2009)",
"ref_id": null
},
{
"start": 139,
"end": 155,
"text": "(Karlsson, 1990;",
"ref_id": "BIBREF12"
},
{
"start": 156,
"end": 178,
"text": "Karlsson et al., 2011)",
"ref_id": "BIBREF13"
},
{
"start": 211,
"end": 237,
"text": "(Bick and Didriksen, 2015)",
"ref_id": "BIBREF1"
},
{
"start": 302,
"end": 320,
"text": "(H\u00e4m\u00e4l\u00e4inen, 2019)",
"ref_id": "BIBREF11"
},
{
"start": 617,
"end": 631,
"text": "(Rueter, 2000)",
"ref_id": "BIBREF19"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Much has been written on the problems with regard to the role of the linguistic consultants in language documentation, especially with regard to ethics and the acknowledgement of their role (see e.g. Rice, 2006; O'Meara and Good, 2010; Chelliah and Willem, 2010, 139-159; Dobrin and Berson, 2011; Bowern, 2015, 171-175) . This may refer to making their identity known or not in publications, corpora and other sources, though acknowledging their role in the material collected, whether or not they wish to be overtly acknowledged by name, should in any case be done. As our aim is to share material as openly as possible we avoided collecting sensitive and personal information that could potentially be harmful for the individuals and communities when building our documentary corpus, and focused primarily on narratives that document local culture and history. Unfortunately, the current interpretations of EU legislation still leave some unclarity as to how questions of personal data in our recordings should be addressed.",
"cite_spans": [
{
"start": 200,
"end": 211,
"text": "Rice, 2006;",
"ref_id": null
},
{
"start": 212,
"end": 235,
"text": "O'Meara and Good, 2010;",
"ref_id": "BIBREF15"
},
{
"start": 236,
"end": 271,
"text": "Chelliah and Willem, 2010, 139-159;",
"ref_id": null
},
{
"start": 272,
"end": 296,
"text": "Dobrin and Berson, 2011;",
"ref_id": "BIBREF6"
},
{
"start": 297,
"end": 319,
"text": "Bowern, 2015, 171-175)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Problem description",
"sec_num": "2"
},
{
"text": "Our research has been carried out in close cooperation with Zyrian Komi communities and native organisations, and we have made significant effort to ensure that our work is both accepted by the community and that the relevant materials are also available to community members. However, as e.g. Dorian (2010, 181) points out, consultants may not always fully grasp what exactly linguists plan on doing with the material they collect. Thereby we have to consider our own responsibility, independent of the informed consent provided by our language informants. We have to ensure that the ways the material is released to the public are appropriate.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Problem description",
"sec_num": "2"
},
{
"text": "The problem may be summarised such that even though we see no issues at the moment in sharing the complete dataset with the community, which we have already done in various ways (whilst taking into account the community's needs), and will also share it with researchers, most likely through a permission request with a description of intended use, it is currently not possible to share it entirely openly with the general public, as this would permanently expose community members' personal information. However, this kind of very formal and restricted method of distribution certainly hinders the active research use of the corpus, which is also something we do not want to happen. Thereby providing a pseudonymised version for the research use seems like a good alternative and something worth investigating further. When the pseudonymised version is available after an academically affiliated login in Korp, it is simple to familiarise oneself with the corpus to decide whether the complete dataset is needed for planned research. This also minimises the unnecessary redistribution of the entire corpus, as individuals do not need to access the complete dataset to evaluate its usability. In the same vein, with this information we can also create derived datasets that can be used in different experiments, but contain only minimal amount of personal data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Problem description",
"sec_num": "2"
},
{
"text": "The semantic tags that associate proper nouns in the Komi morphological analyser are: Sem/Mal (for a male forename), Sem/Fem (female forename), Sem/Patr-Mal, Sem/Patr-Fem (patronym), Sem/Sur, Sem/Sur-Mal, Sem/Sur-Fem (surname) and Sem/Plc (toponym). In addition to proper nouns numerals need specific attention, especially in constructions that are dates or years. Potentially, all tokens in the corpus tagged for one of these semantic categories contain information that either directly reveals the identity of individuals or information that can be easily used for revealing the identity of individuals when combined in combination with other data. The relevant identities can concern the recorded speaker(s) themselves (for instance the own name, names of parents and other relatives, or the place or date of birth) or they concern other individuals to which the recorded speaker is referring to. At the same time, however, spoken recordings contain names of places that are so large and general that they can be mentioned without in fact conveying a great deal of personal information. For example, cities such as Syktyvkar and Moscow have populations large enough such that identifying a person is usually not possible. The same is true for large bodies of water such as the rivers Izhma or Pechora, or mountain ranges like Ural mountains -they span so many localities that they do not actually identify any of them. All these larger entities are treated through the method presented in Section 3.1. However, for very small settlements it is important to be more careful, as some locations only consist of individual houses, and speakers can thereby be identified more easily than would usually be the case. Example 1, which is taken from the introductory part of a personal interview, clearly illustrates how an individual utterance can contain different types of identifiable personal information, of which some are larger entities, i.e. the capital of the Komi Republic Syktyvkar, that can remain unmodified.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "(1) M\u0435\u043d\u04e7 \u0448\u0443\u04e7\u043d\u044b \u0410\u043b\u0435\u043a\u0441\u0430\u043d\u0434\u0440, \u043c\u0435 \u043e\u043b\u0430 \u0412\u0435\u0440\u0442\u0435\u043f\u044b\u043d, \u0432\u0435\u043b\u04e7\u0434\u0447\u0438 \u0421\u044b\u043a\u0442\u044b\u0432\u043a\u0430\u0440\u044b\u043d.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "men\u0259 1 . _ \u0283u-\u0259n\u0268 call-3 . _ a\u028eeksandr, Aleksander Sem/Mal me 1 _ ol-a live-1 . _ vertep-\u0268n, Vertep- Sem/Plc vel\u0259t\u0255\u02d0-i study-1 . _ s\u0268kt\u0268vkar-\u0268n Syktyvkar- Sem/Plc",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "'My name is Aleksander, I live in Vertep, I studied in Syktyvkar.'",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "We can therefore construct a list of major settlements that occur in our corpus, and allow those to pass unchanged through our pseudonymisation system. This is necessary, as the analyser would otherwise mark all places with the tag Sem/Plc. With regard to smaller localities, however, we have two options: 1) either mark them with an empty placeholder, or 2) replace the value with a specific \"standard village\". At the moment we have opted to use empty placeholders, although there are various options that should be considered. Example 1 illustrates how, using this logic, we can distinguish large localities of the type Syktyvkar from small ones such as Vertep.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "Since the full name of each person we have worked with is included in our metadata database, it has been easy to evaluate whether all names are present in the Komi FST. Similarly, our metadata includes all names of recording locations, places of residence and places of birth. In practice, however, there are more place names mentioned in the narratives than those present in the metadata database, as the information in the database has originally been collected from those same interviews that were recorded and transcribed, i.e. the metadata referring to e.g. place names is limited to actual locations where people were born or lived or where recordings where made; place names merely mentioned in speech (like 'I visited Bangkok') have not been specially listed anywhere in our material. The work presented here is in principle one path toward constructing a more structured database of locations present in the corpus, which could be of high relevance for various types of linguistic and nonlinguistic research using our data.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "Another data type that potentially contains information sensitive to personal identities consists of dates. In the case of small local populations this can be true even for incomplete dates, i.e. indicating only the month of a year or even the year alone. Example 2 illustrates how this kind of information could be present in the corpus.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "There is a tendency for such numbers to occur in formulaic expressions, especially as direct replies to questions about the age and such properties. In this kind of situation, when the numbers are pronounced in a very literary manner and are all in Komi, it is relatively easy to identify such segments. For instance, we can write a CG rule that targets sequences that contain the word for 'year' and a preceding sequence of numerals. Another way to target these segments would be to look into them as replies to questions where this content is asked. This would take advantage of the conversationality of the recordings.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "However, there are particular challenges in those instances where the numbers are non-standard or in Russian, such as \u043f\u044f\u0442\u043e\u0433\u043e 'fifth' in Example 2:",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "(2) \u041d\u043e \u043c\u0435 \u0440\u04e7\u0434\u0438\u0442\u0447\u0438 \u043f\u044f\u0442\u043e\u0433\u043e \u0434\u0435\u043a\u0430\u0431\u0440\u044f \u0441\u044e\u0440\u0441 \u04e7\u043a\u043c\u044b\u0441\u0441\u0451 \u043a\u0432\u0430\u0439\u0442\u044b\u043c\u044b\u043d \u0432\u0438\u0442\u0435\u0434 \u0432\u043e\u044b\u043d. The problem here is essentially that parts of the sentence are in Russian. Therefore, we cannot analyse it with the Komi analyser alone. The method has not yet been fully implemented for multilingual data, as processing such material contains numerous mixed forms and other problems. The process currently planned is to pass all unrecognised words through a Russian analyser, which, however, would also demand some consistency in tagging schemes used across these analysers. Cross-linguistic annotation schemes cannot always be straightforwardly matched (see discussion in Rueter and Partanen, 2019) , but for a number of tasks any improvement here is very beneficial.",
"cite_spans": [
{
"start": 641,
"end": 667,
"text": "Rueter and Partanen, 2019)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Method",
"sec_num": "3"
},
{
"text": "Since CG does not allow us to modify the transcribed words directly, our method has been implemented through CG rules that add additional tags and prefixes to the available FST readings. For example, basic semantic tags do not need to be edited at this point, with the exception of locations that we want to keep, as described above. The CG rule that adds an additional tag for locations sufficiently large to be kept intact is very simple: When this rule is applied, the later processing steps do not apply to locations marked with the Sem/LargePlc tag.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation logic",
"sec_num": "3.1"
},
{
"text": "This rule depends on the list LARGE-PLACES which holds information about all the larger towns and settlements that we want to retain in the pseudonymised corpus. The list is manually compiled and contains some tens of generic large locations in Russia and elsewhere, among them common holiday destinations. It could be possible, however, to also connect this list to common open databases such as Wikidata,\u00b3 in order to let the rule automatically apply to all settlements that have a population, for instance, over half a million. This, however, would move from the current direction where the rules are edited based on our own observations and thorough knowledge of the material, although the changes are implemented through the CG.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation logic",
"sec_num": "3.1"
},
{
"text": "\u00b3https://www.wikidata.org For removing birthday data we have experimented with a set of rules that are specific for that context. The rule Explicit years below briefly illustrates this logic, although the actual implementation is slightly more complicated with more word order variation included.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation logic",
"sec_num": "3.1"
},
{
"text": "Explicit years 1 ADD:find-dates-years (Date) 2 TARGET (Num) OR (Ord) 3 ((1* (\"\u0432\u043e\")) OR (1* (\"\u0433\u043e\u0434\"))) ; OR (-1*(\"\u0440\u04e7\u0434\u0438\u0442\u0447\u044b\u043d\u044b\"))) ;",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation logic",
"sec_num": "3.1"
},
{
"text": "Such contextual rules are useful as we can be relatively sure that this date is a date of birth, which is then tagged accordingly. There are, however, so many instances of dates that are without contiguous context that we have decided to use a rule that removes all years and dates. However, having the explicit information available about possible dates of birth in the corpus is very important and increases the accountability of the corpus creators.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation logic",
"sec_num": "3.1"
},
{
"text": "All in all the system is relatively simple, consisting of some tens of CG rules. The actual removal of the sensitive tokens is done by a script reading the tagset that is specifically inserted through our rules. The script removes the actual tokens while leaving in the resulting pseudonymised corpus a placeholder-token, including the belonging morphosyntactic tags coming from the FST analyser.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation logic",
"sec_num": "3.1"
},
{
"text": "The process is implemented as Python functions that are currently used as a part in the script pipeline that convert from the original corpus data in XML format used by ELAN into VRT format needed by Korp. There is, however, no reason why the same methods could not be adapted to other environments not working with ELAN or Korp.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation logic",
"sec_num": "3.1"
},
{
"text": "The whole pipeline has been published in Zenodo (Partanen, 2019) and GitHub\u2074.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Implementation logic",
"sec_num": "3.1"
},
{
"text": "The quality of the system was evaluated with one pass through the complete corpus, and another more qualitative examination of one individual recording \u2074https://github.com/langdoc/ langdoc-pseudonymization of five speakers. This gives a relatively good impression of the accuracy and also the usability of the method. If the resulting text is unusable with too many omitted sections it is clear that the method is not of particular use for researchers.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "4"
},
{
"text": "In the complete corpus currently 2% of all tokens get marked as being possible proper names of persons, places or dates. During evaluation we also found that it was necessary to adjust the system so that both Russian and Komi language versions of the names of settlements are included in the analyser, as the speakers may use both. Some toponyms for smaller places, for instance \u041c\u043e\u0445\u0447\u0430, were missing from the analyser, as were some less-used patronyms, for instance \u041f\u0430\u0440\u0444\u0451\u043d\u043e\u0432\u0438\u0447 and \u0410\u0440\u0441\u0435\u043d\u0442\u044c\u0435\u0432\u043d\u0430. Adding these is, obviously, a trivial task. However, looking also at the Russian analyser benefits the infrastructure at large, as these same names occur in various languages spoken in Russia.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "4"
},
{
"text": "Interestingly, our evaluation run revealed also situations where the system en passant removes ambiguously tagged content. One such example is the lemma \u0411\u0443\u0440\u0430, which could be analysed as a surname or as an adverb. However, to our knowledge it is only used as an adverb with the meaning 'well' and never as a surname. The problem is related to various names that are foreign in cultural context in Komi, but in theory could be foreign names. Also several common Russian words have a potential surname reading, which by our evaluation is not relevant in our corpus. These are, for example, \u0413\u043e\u0440\u044f\u0447\u0438\u0439, \u0414\u0435\u043d and \u0413\u043e\u0442\u043e\u0432. We have relaxed the the system with additional rules to ignore such cases, but with careful consideration only.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "4"
},
{
"text": "One culturally important feature of our method is that it can correctly detect native Komi names, such as the multi-word \u041f\u0430\u0441\u044c \u041a\u043e\u043b\u044f. Among surnames a category of individuals who are so well known that they do not need to be removed is still under consideration. With some names this is clear, for instance, all tokens \u0412\u0438\u0445\u043c\u0430\u043d 'Wichmann' occurring in the corpus refer to the Finnish researcher Yrj\u00f6 Wichmann. Similarly, names such as 'Jesus' or 'Lenin' are kept, as these names are not used as a given name or nick name in our cultural context. Other surnames, such as \u041b\u044b\u0442\u043a\u0438\u043d 'Lytkin', may either refer to the well known Komi researcher Vasily Lytkin or other persons with the same name. Therefore, this part of the system needs refinement.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "4"
},
{
"text": "One benefit of using an orthographic transcription system for a spoken-language corpus (instead of phonemic transcription, cf. Blokland et al., 2015; Gerstenberger et al., 2016) is that orthographically proper nouns are consistently written with initial uppercase. Their parsing and verification for our pseudonymisation system is therefore an easy process. Our evaluation run revealed a list of approximately 7000 tokens with initial uppercase letters, that were not being recognised by the analyser yet. It has been possible to go through this list manually while collecting the forms. As these were primarily items missing from the lexicon, their inclusion was simple. Note also that since our field recordings have been carried out in a limited number of communities, the same toponyms are repeated in different recordings. All work with including these missing names in the analyser's lexicon files rapidly improves the performance of the analyser overall. All in all, our examination resulted in approximately 250 new lemmas being added into the lexicon files of the Komi morphological analyser.",
"cite_spans": [
{
"start": 127,
"end": 149,
"text": "Blokland et al., 2015;",
"ref_id": "BIBREF3"
},
{
"start": 150,
"end": 177,
"text": "Gerstenberger et al., 2016)",
"ref_id": "BIBREF8"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "4"
},
{
"text": "In one particular category of toponyms, names derive from common nouns designating landscape features, such as \u0414\u0456 /di/ 'island' and \u0401\u043b\u044c /jo\u028e/ 'stream'. In such instances the pseudonymisation is done only when they are written with capital letter and are in singular. Although this rule overgeneralises a few relevant common nouns in sentence initial position, it seems to handle this problem sufficiently. Some concepts are, however, so generic that to our knowledge they do not refer uniquely to individual settlements specifically, i.e. \u042f\u0433 /jag/ 'forest\u00b4, \u041a\u0443\u0440\u044a\u044f /kurja/ 'bay\u00b4and \u041a\u0430\u0442\u044b\u0434 /kat\u0268d/ 'downstream\u00b4. The proper noun reading is left at place when the form is written in lower case, but does not have any other possible interpretations. This is against our transcription conventions, but it is beneficial that the system has some robustness for such instances where the spelling is by mistake deviating from our guidelines.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "4"
},
{
"text": "The evaluation of all tokens marked in the entire corpus as potentially containing personal information revealed that out of 8000 pseudonymised tokens only 4% were mistakenly removed. If we would had pseudonymised all items, which are semantically tagged as proper names or dates, without implementing further rules as described above, the ratio of mistakenly removed forms would had been almost 50% of the tagged ones. This is so primarily because very high frequency words such as \u0418\u0437 /iz/ 'stone; Ural mountain; negation verb form' and \u041a\u043e\u043c\u0438 /komi/ 'Komi (Republic); Komi (an ethnic group, a language)' would had been tagged for pseudonymisation and removed by the script as well. This illustrates well, that such a task as writing rules for pseudonymisation can only be done with careful understanding of the data and its cultural context. The forms that cannot be proper nouns under any circumstances have been added to a separate list. There is always the possibility to edit such instances also at the level of the analyser itself.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "4"
},
{
"text": "Later evaluation should be linked to explicit tests that demonstrate that the rules are working under the desired conditions. However, already now the pipeline proposed in this paper has proven itself to be highly effective for the pseudonymisation of a large spoken-language corpus resulting from fieldwork recordings of an endangered language. The specific merits of this system are that it is easy to extend, and through rule-based implementation its precision can be very reliably evaluated and adjusted.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion & Further work",
"sec_num": "5"
},
{
"text": "One possible, and already planned, utilization for our method is the selection of sentences that can be included into dictionaries as examples. There is a general need to display in different web interfaces spoken language sentences that illustrate how a word is used, and through our method we could automatise the task to select example sentences. This could be combined into modern dictionary interfaces such as those discussed by Rueter et al. (2017) and H\u00e4m\u00e4l\u00e4inen and Rueter (2018) .",
"cite_spans": [
{
"start": 434,
"end": 454,
"text": "Rueter et al. (2017)",
"ref_id": "BIBREF20"
},
{
"start": 459,
"end": 487,
"text": "H\u00e4m\u00e4l\u00e4inen and Rueter (2018)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion & Further work",
"sec_num": "5"
},
{
"text": "We have described a method that is effective, reliable, and meets a concrete need in corpus data processing. We have also presented a novel and creative application for Constraint Grammar. We want to stress that besides removing or editing the marked personal information this method could also be used to evaluate how much of this kind of information individual transcriptions contain, thereby providing rough metrics about their level of sensitivity.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion & Further work",
"sec_num": "5"
},
{
"text": "Since no anonymization or pseudonymisation method is perfectly reliable, the materials cannot be made entirely available without further manual verification. We believe, however, that the results we have achieved reach a level that does allow relatively open distribution with only basic user authentication, for example, within an academic research context. As the system described contains many changing elements: Komi FST, CG and the corpus itself, testing and refinement will necessarily continue.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion & Further work",
"sec_num": "5"
}
],
"back_matter": [
{
"text": "The authors of this paper collaborate within the project \"Language Documentation meets Language Technology: The Next Step in the Description of Komi\", funded by the Kone Foundation, Finland. Special thanks to the University of Helsinki for funding Niko Partanen's travel.Thanks to Jack Rueter for his valuable and continuous work on the computational description of Komi, and to Marina Fedina's team at FU-Lab in Syktyvkar for their ongoing efforts in building Komi language resources and language technology.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgments",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Korp and Karp-a bestiary of language resources: the research infrastructure of Spr\u00e5kbanken",
"authors": [
{
"first": "Malin",
"middle": [],
"last": "Ahlberg",
"suffix": ""
},
{
"first": "Lars",
"middle": [],
"last": "Borin",
"suffix": ""
},
{
"first": "Markus",
"middle": [],
"last": "Forsberg",
"suffix": ""
},
{
"first": "Martin",
"middle": [],
"last": "Hammarstedt",
"suffix": ""
},
{
"first": "Leif-J\u00f6ran",
"middle": [],
"last": "Olsson",
"suffix": ""
},
{
"first": "Olof",
"middle": [],
"last": "Olsson",
"suffix": ""
},
{
"first": "Johan",
"middle": [],
"last": "Roxendal",
"suffix": ""
},
{
"first": "Jonatan",
"middle": [],
"last": "Uppstr\u00f6m",
"suffix": ""
}
],
"year": 2013,
"venue": "Proceedings of the 19th Nordic Conference of Computational Linguistics (NODALIDA 2013)",
"volume": "",
"issue": "",
"pages": "429--433",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Malin Ahlberg, Lars Borin, Markus Forsberg, Martin Hammarstedt, Leif-J\u00f6ran Olsson, Olof Olsson, Johan Roxendal, and Jonatan Uppstr\u00f6m. 2013. Korp and Karp-a bestiary of language resources: the research infrastructure of Spr\u00e5kbanken. In Proceedings of the 19th Nordic Conference of Computational Linguistics (NODALIDA 2013), pages 429-433.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Cg-3-beyond classical constraint grammar",
"authors": [
{
"first": "Eckhard",
"middle": [],
"last": "Bick",
"suffix": ""
},
{
"first": "Tino",
"middle": [],
"last": "Didriksen",
"suffix": ""
}
],
"year": 2015,
"venue": "Proceedings of the 20th Nordic conference of computational linguistics",
"volume": "109",
"issue": "",
"pages": "31--39",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Eckhard Bick and Tino Didriksen. 2015. Cg-3-beyond classical constraint grammar. In Proceedings of the 20th Nordic conference of computational linguistics, NODALIDA 2015, May 11-13, 2015, Vilnius, Lithua- nia, 109, pages 31-39. Link\u00f6ping University Elec- tronic Press.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "Spoken Komi Corpus. The Language Bank of Finland version",
"authors": [
{
"first": "Rogier",
"middle": [],
"last": "Blokland",
"suffix": ""
},
{
"first": "Marina",
"middle": [],
"last": "Fedina",
"suffix": ""
},
{
"first": "Niko",
"middle": [],
"last": "Partanen",
"suffix": ""
},
{
"first": "Michael",
"middle": [],
"last": "Rie\u00dfler",
"suffix": ""
}
],
"year": 2020,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rogier Blokland, Marina Fedina, Niko Partanen, and Michael Rie\u00dfler. 2020. Spoken Komi Corpus. The Language Bank of Finland version.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "Language documentation meets language technology",
"authors": [
{
"first": "Rogier",
"middle": [],
"last": "Blokland",
"suffix": ""
},
{
"first": "Ciprian",
"middle": [],
"last": "Gerstenberger",
"suffix": ""
},
{
"first": "Marina",
"middle": [],
"last": "Fedina",
"suffix": ""
},
{
"first": "Niko",
"middle": [],
"last": "Partanen",
"suffix": ""
},
{
"first": "Michael",
"middle": [],
"last": "Rie\u00dfler",
"suffix": ""
},
{
"first": "Joshua",
"middle": [],
"last": "Wilbur",
"suffix": ""
}
],
"year": 2015,
"venue": "First International Workshop on Computational Linguistics for Uralic Languages",
"volume": "",
"issue": "",
"pages": "8--18",
"other_ids": {
"DOI": [
"10.7557/5.3457"
]
},
"num": null,
"urls": [],
"raw_text": "Rogier Blokland, Ciprian Gerstenberger, Marina Fedina, Niko Partanen, Michael Rie\u00dfler, and Joshua Wilbur. 2015. Language documentation meets language tech- nology. In First International Workshop on Computa- tional Linguistics for Uralic Languages, 16th January, 2015, Troms\u00f8, Norway, number 2015:2 in Septen- trio Conference Series, pages 8-18. The University Library of Troms\u00f8.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Linguistic fieldwork: A practical guide",
"authors": [
{
"first": "Claire",
"middle": [],
"last": "Bowern",
"suffix": ""
}
],
"year": 2015,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Claire Bowern. 2015. Linguistic fieldwork: A practical guide. Springer.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "Handbook of descriptive linguistic fieldwork",
"authors": [
{
"first": "L",
"middle": [],
"last": "Shobhana",
"suffix": ""
},
{
"first": "Jules",
"middle": [],
"last": "Chelliah",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Willem",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Shobhana L Chelliah and Jules Willem. 2010. Handbook of descriptive linguistic fieldwork. Springer Science & Business Media.",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "In The Cambridge handbook of endangered languages",
"authors": [
{
"first": "M",
"middle": [],
"last": "Lise",
"suffix": ""
},
{
"first": "Joshua",
"middle": [],
"last": "Dobrin",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Berson",
"suffix": ""
}
],
"year": 2011,
"venue": "",
"volume": "",
"issue": "",
"pages": "187--211",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lise M Dobrin and Joshua Berson. 2011. Speakers and language documentation. In The Cambridge hand- book of endangered languages, pages 187-211. Cam- bridge University Press.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Documentation and responsibility",
"authors": [
{
"first": "",
"middle": [],
"last": "Nancy C Dorian",
"suffix": ""
}
],
"year": 2010,
"venue": "Language & Communication",
"volume": "30",
"issue": "3",
"pages": "179--185",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Nancy C Dorian. 2010. Documentation and responsibil- ity. Language & Communication, 30(3):179-185.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Utilizing language technology in the documentation of endangered Uralic languages",
"authors": [
{
"first": "Ciprian",
"middle": [],
"last": "Gerstenberger",
"suffix": ""
},
{
"first": "Niko",
"middle": [],
"last": "Partanen",
"suffix": ""
},
{
"first": "Michael",
"middle": [],
"last": "Rie\u00dfler",
"suffix": ""
},
{
"first": "Joshua",
"middle": [],
"last": "Wilbur",
"suffix": ""
}
],
"year": 2016,
"venue": "",
"volume": "4",
"issue": "",
"pages": "29--47",
"other_ids": {
"DOI": [
"10.3384/nejlt.2000-1533.1643"
]
},
"num": null,
"urls": [],
"raw_text": "Ciprian Gerstenberger, Niko Partanen, Michael Rie\u00dfler, and Joshua Wilbur. 2016. Utilizing language technol- ogy in the documentation of endangered Uralic lan- guages. 4:29-47.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Advances in synchronized XML-MediaWiki dictionary development in the context of endangered Uralic languages",
"authors": [
{
"first": "Mika",
"middle": [],
"last": "H\u00e4m\u00e4l\u00e4inen",
"suffix": ""
},
{
"first": "Jack",
"middle": [],
"last": "Rueter",
"suffix": ""
}
],
"year": 2018,
"venue": "Proceedings of the eighteenth EURALEX international congress",
"volume": "",
"issue": "",
"pages": "967--978",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mika H\u00e4m\u00e4l\u00e4inen and Jack Rueter. 2018. Advances in synchronized XML-MediaWiki dictionary develop- ment in the context of endangered Uralic languages. In Proceedings of the eighteenth EURALEX interna- tional congress, pages 967-978.",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Language documentation",
"authors": [
{
"first": "Nikolaus",
"middle": [],
"last": "Himmelmann",
"suffix": ""
}
],
"year": 2006,
"venue": "Essentials of Language Documentation, number 178 in Trends in Linguistics. Studies and Monographs",
"volume": "",
"issue": "",
"pages": "1--30",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Nikolaus Himmelmann. 2006. Language documenta- tion. In Jost Gippert, Ulrike Mosel, and Nikolaus Himmelmann, editors, Essentials of Language Docu- mentation, number 178 in Trends in Linguistics. Stud- ies and Monographs, pages 1-30. Mouton de Gruyter.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "UralicNLP: An NLP library for Uralic languages",
"authors": [
{
"first": "Mika",
"middle": [],
"last": "H\u00e4m\u00e4l\u00e4inen",
"suffix": ""
}
],
"year": 2019,
"venue": "Journal of Open Source Software",
"volume": "4",
"issue": "37",
"pages": "",
"other_ids": {
"DOI": [
"10.21105/joss.01345"
]
},
"num": null,
"urls": [],
"raw_text": "Mika H\u00e4m\u00e4l\u00e4inen. 2019. UralicNLP: An NLP library for Uralic languages. Journal of Open Source Soft- ware, 4(37):1345.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Constraint grammar as a framework for parsing running text",
"authors": [
{
"first": "Fred",
"middle": [],
"last": "Karlsson",
"suffix": ""
}
],
"year": 1990,
"venue": "In COLNG",
"volume": "3",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Fred Karlsson. 1990. Constraint grammar as a frame- work for parsing running text. In COLNG 1990 Vol- ume 3: Papers presented to the 13th International Con- ference on Computational Linguistics.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Constraint Grammar: a languageindependent system for parsing unrestricted text",
"authors": [
{
"first": "Fred",
"middle": [],
"last": "Karlsson",
"suffix": ""
},
{
"first": "Atro",
"middle": [],
"last": "Voutilainen",
"suffix": ""
},
{
"first": "Juha",
"middle": [],
"last": "Heikkilae",
"suffix": ""
},
{
"first": "Arto",
"middle": [],
"last": "Anttila",
"suffix": ""
}
],
"year": 2011,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Fred Karlsson, Atro Voutilainen, Juha Heikkilae, and Arto Anttila. 2011. Constraint Grammar: a language- independent system for parsing unrestricted text, vol- ume 4. Walter de Gruyter.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Hfst tools for morphology-an efficient open-source package for construction of morphological analyzers",
"authors": [
{
"first": "Miikka",
"middle": [],
"last": "Krister Lind\u00e9n",
"suffix": ""
},
{
"first": "Tommi",
"middle": [],
"last": "Silfverberg",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Pirinen",
"suffix": ""
}
],
"year": 2009,
"venue": "International Workshop on Systems and Frameworks for Computational Morphology",
"volume": "",
"issue": "",
"pages": "28--47",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Krister Lind\u00e9n, Miikka Silfverberg, and Tommi Piri- nen. 2009. Hfst tools for morphology-an efficient open-source package for construction of morpholog- ical analyzers. In International Workshop on Sys- tems and Frameworks for Computational Morphology, pages 28-47. Springer.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Ethical issues in legacy language resources. Language & Communication",
"authors": [
{
"first": "O'meara",
"middle": [],
"last": "Carolyn",
"suffix": ""
},
{
"first": "Jeff",
"middle": [],
"last": "Good",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "30",
"issue": "",
"pages": "162--170",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Carolyn O'Meara and Jeff Good. 2010. Ethical issues in legacy language resources. Language & Communica- tion, 30(3):162-170.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "langdoc/langdocpseudonymization: Language documentation corpus pseudonymization method",
"authors": [
{
"first": "",
"middle": [],
"last": "Niko Partanen",
"suffix": ""
}
],
"year": 2019,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {
"DOI": [
"10.5281/zenodo.3595506."
]
},
"num": null,
"urls": [],
"raw_text": "Niko Partanen. 2019. langdoc/langdoc- pseudonymization: Language documentation corpus pseudonymization method.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Ethical issues in linguistic fieldwork: An overview",
"authors": [],
"year": 2006,
"venue": "Journal of Academic Ethics",
"volume": "4",
"issue": "1-4",
"pages": "123--155",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Keren Rice. 2006. Ethical issues in linguistic field- work: An overview. Journal of Academic Ethics, 4(1- 4):123-155.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Survey of Uralic Universal Dependencies development",
"authors": [
{
"first": "Jack",
"middle": [],
"last": "Rueter",
"suffix": ""
},
{
"first": "Niko",
"middle": [],
"last": "Partanen",
"suffix": ""
}
],
"year": 2019,
"venue": "Workshop on Universal Dependencies",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jack Rueter and Niko Partanen. 2019. Survey of Uralic Universal Dependencies development. In Workshop on Universal Dependencies, page 78. Association for Computational Linguistics.",
"links": null
},
"BIBREF19": {
"ref_id": "b19",
"title": "\u0425\u0435\u043b\u044c\u0441\u0438\u043d\u043a\u0438\u0441\u0430 \u0443\u043d\u0438\u0432\u0435\u0440\u0441\u0438\u0442\u0435\u0442\u044b\u043d \u043a\u044b\u0432 \u0442\u0443\u044f\u043b\u044b\u0441\u044c \u0418\u0436\u043a\u0430\u0440\u044b\u043d \u043f\u0435\u0440\u044b\u043c\u0441\u0430 \u0441\u0438\u043c\u043f\u043e\u0437\u0438\u0443\u043c \u0432\u044b\u043b\u044b\u043d \u043b\u044b\u0434\u0434\u044c\u04e7\u043c\u0442\u043e\u0440",
"authors": [
{
"first": "Jack",
"middle": [
"M"
],
"last": "Rueter",
"suffix": ""
}
],
"year": 2000,
"venue": "\u041f\u0435\u0440\u043c\u0438\u0441\u0442\u0438\u043a\u0430 6 (Proceedings of Permistika 6 conference)",
"volume": "",
"issue": "",
"pages": "154--158",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jack M. Rueter. 2000. \u0425\u0435\u043b\u044c\u0441\u0438\u043d\u043a\u0438\u0441\u0430 \u0443\u043d\u0438\u0432\u0435\u0440\u0441\u0438\u0442\u0435\u0442\u044b\u043d \u043a\u044b\u0432 \u0442\u0443\u044f\u043b\u044b\u0441\u044c \u0418\u0436\u043a\u0430\u0440\u044b\u043d \u043f\u0435\u0440\u044b\u043c\u0441\u0430 \u0441\u0438\u043c\u043f\u043e\u0437\u0438\u0443\u043c \u0432\u044b\u043b\u044b\u043d \u043b\u044b\u0434\u0434\u044c\u04e7\u043c\u0442\u043e\u0440. In \u041f\u0435\u0440\u043c\u0438\u0441\u0442\u0438\u043a\u0430 6 (Proceedings of Per- mistika 6 conference), pages 154-158.",
"links": null
},
"BIBREF20": {
"ref_id": "b20",
"title": "Synchronized Mediawiki based analyzer dictionary development",
"authors": [
{
"first": "Jack",
"middle": [],
"last": "Michael Rueter",
"suffix": ""
},
{
"first": "Mika",
"middle": [
"Kalevi"
],
"last": "H\u00e4m\u00e4l\u00e4inen",
"suffix": ""
}
],
"year": 2017,
"venue": "3rd International Workshop for Computational Linguistics of Uralic Languages Proceedings of the Workshop. Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Jack Michael Rueter, Mika Kalevi H\u00e4m\u00e4l\u00e4inen, et al. 2017. Synchronized Mediawiki based analyzer dic- tionary development. In 3rd International Workshop for Computational Linguistics of Uralic Languages Proceedings of the Workshop. Association for Com- putational Linguistics.",
"links": null
},
"BIBREF21": {
"ref_id": "b21",
"title": "Public access to research data in language documentation: Challenges and possible strategies. Language Documentation and Conservation",
"authors": [
{
"first": "Mandana",
"middle": [],
"last": "Seyfeddinipur",
"suffix": ""
},
{
"first": "Felix",
"middle": [],
"last": "Ameka",
"suffix": ""
},
{
"first": "Lissant",
"middle": [],
"last": "Bolton",
"suffix": ""
},
{
"first": "Jonathan",
"middle": [],
"last": "Blumtritt",
"suffix": ""
},
{
"first": "Brian",
"middle": [],
"last": "Carpenter",
"suffix": ""
},
{
"first": "Hilaria",
"middle": [],
"last": "Cruz",
"suffix": ""
},
{
"first": "Sebastian",
"middle": [],
"last": "Drude",
"suffix": ""
},
{
"first": "Patience",
"middle": [
"L"
],
"last": "Epps",
"suffix": ""
},
{
"first": "Vera",
"middle": [],
"last": "Ferreira",
"suffix": ""
},
{
"first": "Ana",
"middle": [
"Vilacy"
],
"last": "Galucio",
"suffix": ""
}
],
"year": 2019,
"venue": "",
"volume": "13",
"issue": "",
"pages": "545--563",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mandana Seyfeddinipur, Felix Ameka, Lissant Bolton, Jonathan Blumtritt, Brian Carpenter, Hilaria Cruz, Sebastian Drude, Patience L Epps, Vera Ferreira, Ana Vilacy Galucio, et al. 2019. Public access to research data in language documentation: Challenges and possible strategies. Language Documentation and Conservation, 13:545-563.",
"links": null
}
},
"ref_entries": {}
}
} |