File size: 85,470 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 |
{
"paper_id": "2021",
"header": {
"generated_with": "S2ORC 1.0.0",
"date_generated": "2023-01-19T01:13:26.569130Z"
},
"title": "Representation of Yine (Arawak) Morphology by Finite State Transducer Formalism",
"authors": [
{
"first": "Adriano",
"middle": [
"M"
],
"last": "Ingunza",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Universidad Cat\u00f3lica del Per\u00fa",
"location": {
"addrLine": "San Miguel",
"settlement": "Lima",
"country": "Peru"
}
},
"email": ""
},
{
"first": "John",
"middle": [
"E"
],
"last": "Miller",
"suffix": "",
"affiliation": {},
"email": ""
},
{
"first": "Arturo",
"middle": [],
"last": "Oncevay",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "University of Edinburgh",
"location": {
"country": "Scotland"
}
},
"email": ""
},
{
"first": "Roberto",
"middle": [],
"last": "Zariquiey",
"suffix": "",
"affiliation": {
"laboratory": "",
"institution": "Universidad Cat\u00f3lica del Per\u00fa",
"location": {
"addrLine": "San Miguel",
"settlement": "Lima",
"country": "Peru"
}
},
"email": ""
}
],
"year": "",
"venue": null,
"identifiers": {},
"abstract": "We represent the complexity of Yine (Arawak) morphology with a finite state transducer (FST) based morphological analyzer. Yine is a low-resource indigenous polysynthetic Peruvian language spoken by approximately 3,000 people and is classified as 'definitely endangered' by UNESCO. We review Yine morphology focusing on morphophonology, possessive constructions and verbal predicates. Then we develop FSTs to model these components proposing techniques to solve challenging problems such as complex patterns of incorporating open and closed category arguments. This is a work in progress and we still have more to do in the development and verification of our analyzer. Our analyzer will serve both as a tool to better document the Yine language and as a component of natural language processing (NLP) applications such as spell checking and correction.",
"pdf_parse": {
"paper_id": "2021",
"_pdf_hash": "",
"abstract": [
{
"text": "We represent the complexity of Yine (Arawak) morphology with a finite state transducer (FST) based morphological analyzer. Yine is a low-resource indigenous polysynthetic Peruvian language spoken by approximately 3,000 people and is classified as 'definitely endangered' by UNESCO. We review Yine morphology focusing on morphophonology, possessive constructions and verbal predicates. Then we develop FSTs to model these components proposing techniques to solve challenging problems such as complex patterns of incorporating open and closed category arguments. This is a work in progress and we still have more to do in the development and verification of our analyzer. Our analyzer will serve both as a tool to better document the Yine language and as a component of natural language processing (NLP) applications such as spell checking and correction.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Abstract",
"sec_num": null
}
],
"body_text": [
{
"text": "Yine is a low resource indigenous polysynthetic Peruvian language of the Arawak family spoken by approximately 3,000 people living near the Ucayali and Madre de Dios rivers, tributary rivers of the Amazon. Yine is considered \"definitely endangered\" according to the UNESCO Atlas of the World's Languages in danger (Moseley, 2010) .",
"cite_spans": [
{
"start": 314,
"end": 329,
"text": "(Moseley, 2010)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "As noted by Zariquiey et al. (2019) , although Yine has a typologically oriented descriptive grammar, documentation and further study of several grammatical aspects are still urgently needed since the Yine language is at risk of entering into an obsolescent and consequently disappearing status. Therefore, such work is vital to not only adequately document the Yine language, but also to support its continued vitality through computer assisted tools such as spell-checkers and machine translators.",
"cite_spans": [
{
"start": 12,
"end": 35,
"text": "Zariquiey et al. (2019)",
"ref_id": "BIBREF18"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Formal and computational representation of morphology is considered a \"solved problem\" based * Authors contributed equally on Beesley and Karttunen's work and seminal Finite State Morphology text (Beesley and Karttunen, 2003; Karttunen and Beesley, 2005) . This does not mean that representing a language is either easy or fast, especially for the case of polysynthetic languages such as Yine.",
"cite_spans": [
{
"start": 209,
"end": 225,
"text": "Karttunen, 2003;",
"ref_id": "BIBREF1"
},
{
"start": 226,
"end": 254,
"text": "Karttunen and Beesley, 2005)",
"ref_id": "BIBREF8"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Our goal is to construct a high coverage finite state transducer (FST) morphological analyzer both to document and preserve the Yine language, and to use it in NLP applications, such as spell checking and correction, that might promote language vitality. Our contributions at this point are: 1. a partial functioning morphological analyzer for nominal and verbal constructions including possessive constructions and verbal predicates, and 2. various project decisions and FST patterns employed so far in construction of the analyzer. Given the incomplete implementation, it is too early to report meaningful project results.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Representation of Yine morphology by a FST is a work in progress. This paper describes relevant morphological features of Yine, representation of these features by FST, particularly challenging representation problems, a preliminary evaluation, and our current and planned future states.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Introduction",
"sec_num": "1"
},
{
"text": "Beesley and Karttunen (2003) 's Finite State Morphology text is a highly valuable resource for representing morphology by an FST. There are also numerous morphological analyses with FST representations available. Most relevant to this task are analyses performed for other indigenous Peruvian languages: Shipibo-Konibo (Cardenas and Zeman, 2018) , Quechua (Rios, 2010) and pan-Ashaninka (Ortega et al., 2020; Castro Mamani, 2020) . In particular, the last work includes applications of the FST to spell-checking and segmentation.",
"cite_spans": [
{
"start": 12,
"end": 28,
"text": "Karttunen (2003)",
"ref_id": "BIBREF1"
},
{
"start": 319,
"end": 345,
"text": "(Cardenas and Zeman, 2018)",
"ref_id": "BIBREF3"
},
{
"start": 356,
"end": 368,
"text": "(Rios, 2010)",
"ref_id": "BIBREF15"
},
{
"start": 387,
"end": 408,
"text": "(Ortega et al., 2020;",
"ref_id": "BIBREF11"
},
{
"start": 409,
"end": 429,
"text": "Castro Mamani, 2020)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "While we do not apply our work to spell checking in this paper, that is one of our planned goals.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Previously we had attempted to develop a Hun-Spell 1 based spell corrector, but found it too limiting given the polysynthetic nature of the Yine language. This is consistent with Lind\u00e9n (2010, 2014) , who found that FST correctors were essential to achieve performance on par with English for morphologically complex, and typically low resource, languages.",
"cite_spans": [
{
"start": 179,
"end": 198,
"text": "Lind\u00e9n (2010, 2014)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Software, tutorials, and examples for constructing FST morphology are available from the Finite State Morphology book website. 2 We use the Foma library 3 by Hulden (2009) , compatible with FST Morphology, and available, along with some fine tutorials. Both applications offer a Python API, but neither is under active development. There is limited community support for Foma.",
"cite_spans": [
{
"start": 127,
"end": 128,
"text": "2",
"ref_id": null
},
{
"start": 158,
"end": 171,
"text": "Hulden (2009)",
"ref_id": "BIBREF6"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Related Work",
"sec_num": "2"
},
{
"text": "Yine (ISO 639-3: pib) may be considered a morphosyntactically complex language due to its highly polysynthetic profile (mainly related to verbal structures). As noted by Aikhenvald (2020) , Arawak languages are synthetic, predominantly head marking and suffixing, with a complex verbal morphology. Yine presents three open word classes: nouns, verbs, and adjectives (mostly by derivation); and four closed word classes: pronouns, adverbs, demonstratives and numerals. In this section, we will only discuss the pronominal system, and some features associated with the verbal and nominal morphology, since they are relevant to the current state of representation of Yine morphology by the FST formalism.",
"cite_spans": [
{
"start": 170,
"end": 187,
"text": "Aikhenvald (2020)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Linguistic Profile and Resources",
"sec_num": "3"
},
{
"text": "As in almost all polysynthetic languages, Yine may express in just one word meanings that would require a whole sentence in other languages. This is illustrated by a complex predicative construction in (1), and a full possessive construction, in (2). Our morphological analysis is based on Hanson (2010)'s grammatical description; glosses have been adapted to the UniMorph schema (Kirov et al., 2018) .",
"cite_spans": [
{
"start": 380,
"end": 400,
"text": "(Kirov et al., 2018)",
"ref_id": "BIBREF9"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Morphological profile",
"sec_num": "3.1"
},
{
"text": "(1) niklokgimatanaktatkalu \u00f8-nikloka-gima-ta ARGNO3SM-swallow-QUOT-LGSPEC1 -na-kta-tka-lu 1 http://hunspell.github.io/ 2 http://www.fsmbook.com 3 https://fomafst.github.io/ -LGSPEC2-INDF-PFV-ARGAC3SM '(The huge snake) swallowed him up somehow, reportedly.'",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphological profile",
"sec_num": "3.1"
},
{
"text": "(2) ragmunateymana r-gagmuna-te-yma-na PSS3P-tree-PSSD-COM/INS-PSS3P 'With their trees' Note that Yine's morphological complexity involves vowel deletion as seen in (1) and morphemes that may be accounted for as circumfixes, as is the case of possession marking in (2) where possessor indexation is achieved with two elements: prefix rand the suffix -na. Its implications for FST expression are very interesting and will be discussed in \u00a74 and \u00a75. In the remaining subsections we present some of the mentioned features. Specifically, we present morphophonological rules, possessive constructions, verbal morphology aspects and argument indexing systems in relation with verbal predicates.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphological profile",
"sec_num": "3.1"
},
{
"text": "Yine presents a rich set of morphophonological processes such as vowel deletion and rhotacism of liquid consonants. These processes are presented below.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphophonological overview",
"sec_num": "3.2"
},
{
"text": "Deletion between stem and suffix occurs when a specific group of suffixes trigger the deletion of the final vowel in the attached stem as shown in (3), where the frequentative suffix -je triggers the deletion of the stem's final vowel. However, this can only occur if vowel deletion does not generate a cluster of three consonants which is an overall restriction in the language as can be seen in (4), where the stem remains complete in its overt realization and avoids the sequence /mkj/. Prefixing of possessive morphemes triggers other morphophonological processes that will be explained in \u00a73.3. In (5) we see /l/ rhotacism, which occurs when an /l/ initial suffix mutates /l/ to /r/ when attached to a stem ending in i, e, u or n. Example (6) shows how the suffix behaves when attached to a different ending stem. Note that it also occurs an internal-boundary vowel deletion process triggered by the third person suffix.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphophonological overview",
"sec_num": "3.2"
},
{
"text": "(5) pnikanru p-nika-ni-lu ARGNO2S-eat-DED-ARGAC3SM 'You will eat it (masc)' (6) pniklu p-nika-lu ARGNO2S-eat-ARGAC3SM 'You eat it (masc)'",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphophonological overview",
"sec_num": "3.2"
},
{
"text": "It is important to notice that the set of morphophonological rules developed by Hanson (2010) is neither exhaustive nor conclusive. The author mentions that a complete description of the morphological patterns of the language is still needed and leaves many issues open for further study. Thus, our application of them is based not only on the explicit description of Hanson (2010) but also in the examples presented by the author which entails some systematizable rules for our work. For example, examples (7) and (8) and how how the same 1PL object morpheme wu triggers vowel deletion in (7) and does not in (8) where it would create an identical consonant cluster ww. So, although vowel deletion seems to be lexically specified as mentioned by the author, phonological constraints seem to be highly relevant. There are other morphophonological rules applied in word formation which need to be studied in depth. Rules applied to prefixation processes, are presented in the next section.",
"cite_spans": [
{
"start": 80,
"end": 93,
"text": "Hanson (2010)",
"ref_id": "BIBREF5"
},
{
"start": 368,
"end": 381,
"text": "Hanson (2010)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Morphophonological overview",
"sec_num": "3.2"
},
{
"text": "Possessive constructions in Yine are formed by a possessor prefix (and if needed a linked possessor suffix), a possessed nominal root and, when needed, a 'possession status' suffix. Both morphological elements (i.e. the possessor prefix and the possession status suffixes) are determined by the semantics of the root they attach in terms of alienability. According to Hanson (2010) and Aikhenvald (2020) , nominals are lexically specified for alienable versus inalienable possession.",
"cite_spans": [
{
"start": 368,
"end": 381,
"text": "Hanson (2010)",
"ref_id": "BIBREF5"
},
{
"start": 386,
"end": 403,
"text": "Aikhenvald (2020)",
"ref_id": "BIBREF0"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Possessive constructions",
"sec_num": "3.3"
},
{
"text": "Alienability is a category that makes a morphosyntactic distinction between possession that can be terminated (alienables) and possession that cannot (inalienable) (Payne, 2007) . Of course, this is a language specific categorization. For example, in Yine, concepts such as house or language, are inalienable but a concept like husband is alienable. Nevertheless, concepts like mother or hand tend to be classified as inalienable in those languages that reflect this distinction in their grammar. Additionally, in Yine inalienable nouns present an internal sub-classification distinguishing between kinship terms (like mother or son) and non-kinship terms (like hand or house).",
"cite_spans": [
{
"start": 164,
"end": 177,
"text": "(Payne, 2007)",
"ref_id": "BIBREF12"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Possessive constructions",
"sec_num": "3.3"
},
{
"text": "Depending on the noun root class and its initial consonant, Yine possessive constructions will use one of the three pronominal sets for possessor indexing.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Possessive constructions",
"sec_num": "3.3"
},
{
"text": "Class 1 prefixes attach indistinctly to alienable or inalienable roots but only to those beginning with /g/. This consonant is always replaced by the pronoun. Additionally, if the first consonant is followed by a /u/, it mutates to a /i/ (this is always true with the exception of the 2PL prefix).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Possessive constructions",
"sec_num": "3.3"
},
{
"text": "Class 2 prefixes attach also to alienable and inalienable roots with exception of non-kinship inalienable roots. Regarding morphophonology, this class does not attach to stems beginning with /g/ and does not replace the initial consonant of the stem. Classes 1 and 2 are almost identical, only differing in the 3rd person masculine/plural prefix: class 1 uses /r/ and class 2 uses a \u00f8 form.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Possessive constructions",
"sec_num": "3.3"
},
{
"text": "Class 3 prefixes are attached only with those inalienable stems that do not begin with /g/. In the examples below we present the application of each pronominal class. The class 1 prefix pronoun for 1st person singular and its morphophonological effects on an alienable root is shown in (9), Class 2 prefix pronoun for 2nd person singular attached to an inalienable root is shown in (10), and Class 3 prefix pronoun for 3rd person plural is shown in (11). Finally, Class forms for 3rd person plural are shown in 2 A last consequence of lexical specification of nominal stems is the usage of the so called 'possessed status suffixes'. These are affixed to alienable stems when possessor is expressed, as shown in (9) with -te, and to inalienable stems when possessor is not expressed as in 13where -chi is used.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Possessive constructions",
"sec_num": "3.3"
},
{
"text": "Hanson (2010) treats morphological elements corresponding exclusively to the verbal stem separately from verbal predicate elements. She makes this separation to better leverage the commonality between verbal, nominal and adjectival predicates also attested to in Yine. Verbal stem morphology is exclusive to verbal stems, whereas predicative morphology may be applied to any predicate type.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Verbal and verbal predicate morphology",
"sec_num": "3.4"
},
{
"text": "Verbal stem morphology includes noun incorporants, oblique markers, evidentials, adverbial incorporants, aspect and subordination information, stem closure morphology, applicative suffixes and voice and mood morphemes.Verbal stem complexity is shown in (12). Notice that the example is not a simple stem but a predicate. Bolded morphemes correspond to what Hanson (2010) considers stem morphology.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Verbal and verbal predicate morphology",
"sec_num": "3.4"
},
{
"text": "(12) rustakatsyeggimatanrona r-gistaka-tsa-yegi-gima ARGNO3P-cut-cord.of-PROX-QUOT -ta-na- -LGSPEC1-LGSPEC2 -lo-na -ARGAC3SF-ARGNO3P",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Verbal and verbal predicate morphology",
"sec_num": "3.4"
},
{
"text": "'They cut the rope near her, reportedly' Argument indexing and 'external aspect' specification do not correspond to the verbal stem but to the predicative morphology. Argument indexing is achieved by using prefixation for subjects and suffixation for objects. As for possessor indexing, 3PL forms are indexed by two morphological elements: prefix r and suffix -na. The pronominal forms are almost the same as the ones used for possessive constructions. The main distinction is that only classes 1 are 2 are used. Pronominal indexes are also classified in two classes and follow a regular pattern.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Verbal and verbal predicate morphology",
"sec_num": "3.4"
},
{
"text": "Linguistic resources used for this paper such as analysis and corpora, come from three principal sources: Hanson (2010) which is a comprehensive typological oriented grammar, a Yine-Spanish/Spanish-Yine dictionary by Wise (1986) , and a theoretical guide developed by Zapata et al. (2017) . Additionally, we used a Yine corpus by Bustamante et al. (2020) for evalution purposes (see \u00a76).",
"cite_spans": [
{
"start": 217,
"end": 228,
"text": "Wise (1986)",
"ref_id": null
},
{
"start": 268,
"end": 288,
"text": "Zapata et al. (2017)",
"ref_id": "BIBREF17"
},
{
"start": 330,
"end": 354,
"text": "Bustamante et al. (2020)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Available linguistic resources",
"sec_num": "3.5"
},
{
"text": "In the FST morphology formalism ( Figure 1A ), parallel language representations (tapes) are mapped one to the other, where by convention the upper tape corresponds to the morphological analysis and and lower tape corresponds to the word form. Each level accepts (generates) valid strings in their respective tape, and either level can be transduced to corresponding (possibly multiple) strings on the other level. FSTs can be stacked so that a lower or upper tape feeds into the corresponding tape of another FST. In summary: 1. words can be transduced to morphological analyses, 2. morphological analyses can be transduced to words, 3. only valid representations are accepted (generated) on either side, 4. a valid input representation may result in multiple output representations, and 5. transducers can be stacked to multiple levels.",
"cite_spans": [],
"ref_spans": [
{
"start": 34,
"end": 43,
"text": "Figure 1A",
"ref_id": "FIGREF4"
}
],
"eq_spans": [],
"section": "Finite State Morphology",
"sec_num": "4"
},
{
"text": "Scripting for FST (see Figure 1B) Python scripts to FST source files. All morphological analysis is coded in FST, consistent with efforts by (Cardenas and Zeman, 2018; Ortega et al., 2020; Castro Mamani, 2020) for other Amazonian languages.",
"cite_spans": [
{
"start": 141,
"end": 167,
"text": "(Cardenas and Zeman, 2018;",
"ref_id": "BIBREF3"
},
{
"start": 168,
"end": 188,
"text": "Ortega et al., 2020;",
"ref_id": "BIBREF11"
},
{
"start": 189,
"end": 209,
"text": "Castro Mamani, 2020)",
"ref_id": "BIBREF4"
}
],
"ref_spans": [
{
"start": 23,
"end": 33,
"text": "Figure 1B)",
"ref_id": "FIGREF4"
}
],
"eq_spans": [],
"section": "Finite State Morphology",
"sec_num": "4"
},
{
"text": "The FST language can be viewed as divided into regular expressions (defining finite state machines (FSMs)) typically used for string searching or pattern matching, advanced operators on FSMs or FSTs, and a meta-language for interacting with FSTs. Regular expressions largely suffice for the analysis tape; cross-product, rewrite rule, composition, and containment advanced operators are essential for operating on FSMs and FSTs; define, apply, file related, and virtual stack machine related meta-commands let us construct and interact with FSTs and the operating system.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Finite State Morphology",
"sec_num": "4"
},
{
"text": "FST components may also be grouped functionally as lexical, post-lexical and memory filters; lexicon; and alterations ( Figure 1C ). Filters which restrict lexical generation precede the lexicon; they serve to restrict the allowable combinations of constituent morphemes that might be generated by the lexicon. The lexicon, originates all constituent morphemes from both open and closed morpheme classes generating all possible valid (mostly) lexical sequences.",
"cite_spans": [],
"ref_spans": [
{
"start": 120,
"end": 129,
"text": "Figure 1C",
"ref_id": "FIGREF4"
}
],
"eq_spans": [],
"section": "Finite State Morphology",
"sec_num": "4"
},
{
"text": "Sometimes it is difficult to prospectively generate only valid analyses, and so filters may be used to prevent over-generation. Similarly, some problems of over-generation (e.g., duplication) are more readily solved after generation with post-lexical filters. Phonological and morphophonological processing often imposes constraints on surface form realization of the morphological analyses, e.g., final vowel elision or rhotacism. Such constraints are implemented as alterations of the lexical analysis. Long range or discontinuous morphological relations are not readily handled by FSTs, but with use of limited memory based filters, with diacritic flags, even these problems can be resolved.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Finite State Morphology",
"sec_num": "4"
},
{
"text": "We chose to divide and conquer the analyzer project based on (Hanson, 2010) 's Yine Grammar structure. We define common terms, closed class morphemes, and open class roots, followed by higher level constructs expressible as single words: adjective, noun, noun phrase, verb, nominalization, predicate, and clause.",
"cite_spans": [
{
"start": 61,
"end": 75,
"text": "(Hanson, 2010)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Finite State Morphology",
"sec_num": "4"
},
{
"text": "In the next section on morphological analysis we will see cogent examples combining language analysis from the previous section and finite state morphology described here.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Finite State Morphology",
"sec_num": "4"
},
{
"text": "We report several morphological analyses and snippets of corresponding FST code. FST is a multiuse term applying to simple definitions, regular expressions, filters, alterations, lexicon and the entire analyzer. All the terms beginning with /\u2022/ are symbolic terms defined in file common-u.foma; their corresponding implementation specific and Unimorph terms are substituted on evaluation. Listing 1 shows a snippet of label definitions. Listing 2: Noun root snippet",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Morphological analysis",
"sec_num": "5"
},
{
"text": "Open word vocabulary is processed using Python scripts to construct root constituents with coded lexical information. The snippet in listing 2 defines noun roots, kamruru and gagmuna with form, alienability, and possessor prefix class. Inalienable nouns are further marked with \u2022Kin when a kinship term. Alienable nouns are marked for their possessed suffix type. Possessor prefix class is largely determinable from alienability, kinship, and whether the initial sound segment is /g/, but it was more convenient, to index it directly. Note the use of define to define the FST of all noun roots and assign it to NRoot. The form {kamruru} is expanded to a string of characters and available on both the upper and lower tapes of this transducer. The regular expression [\u2022NRoot \u2022PossPfx3 \u2022Inalienable]:0 groups together the sequence of analysis terms on the upper tape as .NROOT.3.NALN and and maps them to \u00f8 on the lower tape via the : cross-product operation.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Open word categories",
"sec_num": "5.1"
},
{
"text": "Yine noun roots from the example just above are shown in (13) and (14). Inalienable nouns are preferentially possessed and are marked with the suffix -chi when unpossessed. Alienable nouns can readily occur without a possessor (unmarked) and are marked with their possessed suffix when possessed. 4 13 Listing 3: Noun possession regexes",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Noun root examples",
"sec_num": "5.2"
},
{
"text": "The noun shown in (15) is copied from (2) above. Word construction shows several phenomena taken into account by the FST: 1. possessor class 1 (stem with initial /g/), 2. comitative noun case, 3. elision alteration of initial /g/, 4. discontinuous dependency for possessor 3 rd person plural. The snippet in listing 4 shows 3 rd person singular and plural prefixes from possessor prefix class 1. For the singular case, t \u20223SgFPssr -is written to the upper tape, and t^g to the lower tape. The intermediate flag^g subsequently triggers an alteration due to the initial /g/. The plural case adds complexity with a diacritic flag being set to positive by @P.PSSR.3PL@ for both upper and lower tapes, in addition to writing r \u20223PlPssr -to the upper tape and r^g to the lower tape. The diacritic flag with feature PSSR remembers its setting and permits completion of the word with the PSS3P Listing 4: Possessor paradigm 1 (initial 'g') The snippet in listing 5 presents three mutually exclusive noun case alternatives of which comitative is matched in analysis; and so the comitative -yma \u2022Com is written to the upper tape and yma to the lower tape. None of the cases trigger vowel elision. The snippet in listing 6 decides whether or not to show the PSS3P suffix based on the PSSR diacritic flag setting. If the flag setting meets the 3PL requirement, then -na \u20223PlPssr is written to the upper tape and^Vu na is written to the lower (intermediate) tape. The intermediate flag^Vu subsequently triggers an alteration of final vowel elision except for /u/. If the PSSR diacritic flag is not set then nothing is written to either tape; in this way the FST can accept the discontinuous 3 rd person plural possessor.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Nominal example",
"sec_num": "5.3"
},
{
"text": "define Pron3PlNSfx [ %-:\"^Vu\" \"@R.PSSR.3PL@\" {na} \u2022 3PlPssr:0 | \"@D.PSSR@\" ];",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Nominal example",
"sec_num": "5.3"
},
{
"text": "Listing 6: Possessor 3rd person plural suffix The snippet in listing 7 generates the noun from optional possessor class 1 prefix, noun root, optional noun plural, optional noun case and diacritic flag determined 3rd person plural suffix. The alteration FSTs are composed with the lexical output to handle changes due to initial /g/, final vowel elision, or final vowel elision for vowels other than /u/. Listing 7: Noun generation 6 Diacritic flags are a powerful yet difficult to understand addition to FST. See (Hulden, 2011) for a brief introduction and (Beesley and Karttunen, 2003, pp 339-373) for an in depth explanation with examples.",
"cite_spans": [
{
"start": 431,
"end": 432,
"text": "6",
"ref_id": null
},
{
"start": 570,
"end": 598,
"text": "Karttunen, 2003, pp 339-373)",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Nominal example",
"sec_num": "5.3"
},
{
"text": "The word ragmunateymana shows application of both the initial /g/ and final vowel elision except for /u/ alterations. The snippet in listing 8 shows how an initial gi is rewritten as /u/ or /g/ is rewritten as /\u00f8/ after the^g intermediate flag in the lower tape; subsequently the flag itself is erased from the lower tape.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Nominal example",
"sec_num": "5.3"
},
{
"text": "In ragmunateymana the initial /g/ of the noun root is elided and the /r/ of the pronoun prefix added. The case for final vowel other than /u/ elision is more complex, in that the vowel is not elided if it would result in a three consonant cluster. Such is the case here and so the final /a/ of -yma need not elide before -na. Since the three consonant cluster includes nasal consonants, the final /a/ could be elided resulting in the alternative valid word form ragmunateymna (Hanson, 2010) . Listing 8: 'g' alteration",
"cite_spans": [
{
"start": 476,
"end": 490,
"text": "(Hanson, 2010)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Nominal example",
"sec_num": "5.3"
},
{
"text": "The verb predicate shown in (16) is not testified to by the Yine corpus, but rather is a tour de force act of word creation based on the grammar by Hanson, comparable to verb predicate phrase creation in non-polysynthetic languages. The analysis shown is based on the FST analysis and shows several important word generation features: 1. subject prefix class 1 (stem with initial /g/), 2. associate prefix gim-, 3. alteration due to initial /g/, 4. discontinuous dependency of form for 3 rd person plural, 5. multiple incorporants for verb stem, 6. open category noun incorporant, 7. marker for closure of incorporants, 8. multiple incorporants for verb predicate, 9. vowel elision. The subject pronoun prefix class 1 (with inital /g/) is similar to that of possessor prefix class 1 with nouns. Discontinuous behavior for \u2022Subj3Pl is also similar to that for \u20223PlPssr, noun possessor 3 rd person plural, with the obvious difference that the 3 rd person plural subject suffix marker -na is now very distant from the prefix! Adding the associative prefix gimto the verb root triggers 'g' alteration for roots with initial /g/ similar to subject class 1. The FST, see listing 9, writes gim \u2022Assocto the upper tape and gim^g to the lower tape. The intermediate flag^g subsequently triggers 'g' alteration if the stem has initial /g/ as is the case here for the verb gustaka. Listing 9: 'g' alteration with gim-A huge difference in relation nouns is that verbs and verb predicates can have several incorporated morphemes including open noun class morphemes. Individual closed form incorporants are similar in structure to NounCase (listing 5) and VerbAssoc (listing 9) above. With verb stems, multiple incorporants can appear, but each incorporant type only once, and according to Hanson (2010) , the order of incorporants is flexible. The snippet in listing 10 shows forming the union of individual incorporants, and the snippet in listing 11 shows how this union is repeated over 1 to 9 iterations. While not obvious from the union (because everything is via definitions), the lexical form and analysis for each incorporant are written to the upper tape and the lexical form and a unique filter flag are written to the lower tape. The filter flags will be used to enforce the no more than one of each incorporant type rule. 7",
"cite_spans": [
{
"start": 1776,
"end": 1789,
"text": "Hanson (2010)",
"ref_id": "BIBREF5"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Verb predicate mega example",
"sec_num": "5.4"
},
{
"text": "define VerbIncorporantsNoCoda [ %-:0 NounAlienPoss 0:\"^I.A\" ... | VerbAspect2 0:\"^I.H\" | VerbAspect3 0:\"^I.I\" ];",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Verb predicate mega example",
"sec_num": "5.4"
},
{
"text": "Listing 10: Verb stem incorporant union When verb stem incorporants are used, they must be followed by marking of incorporant list closure, or by a causative which also effects closure, [VerbClosure | VerbCausative] . While repetition for 1 to 9 iterations of the union of incorporants assures no more than 9 incorporants, it does not prevent repetition of some of the incorporants. This is where the filter flags, e.g., \"^I.H\", are used. Com-posing~[detectIncorporantDuplicates] with the lower tape from verb incorporants excludes all cases where the same filter flag is repeated, thus eliminating repeated incorporants from the FST. Listing 12: Incorporant test for duplicates Alienable possessed nouns or inalienable nouns (possessed root form) can serve as incorporants. This augments the expressiveness of the verb stem dramatically in that the number of verb stem combinations now gets multiplied by the number of alienable nouns and by the number of inalienable nouns. Gloss (16) incorporates the possessed alienable noun siji-ne, 'corn'.",
"cite_spans": [
{
"start": 186,
"end": 215,
"text": "[VerbClosure | VerbCausative]",
"ref_id": null
},
{
"start": 439,
"end": 479,
"text": "Com-posing~[detectIncorporantDuplicates]",
"ref_id": null
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Verb predicate mega example",
"sec_num": "5.4"
},
{
"text": "Elision processes are the same or similar for nouns and we don't repeat the FST code here. Note that with so many components in the word and multiple elision processes it is not obvious to the non-native speaker, how to derive the final word form with all applied elisions and other alterations.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Verb predicate mega example",
"sec_num": "5.4"
},
{
"text": "There may be multiple analyses for individual words of the language and similarly multiple word representations for the same analysis. This ambiguity can happen because: 1. elision of final vowels of morphemes so that forms are no longer distinct, 2. elision is optional so that inherently there are multiple forms, or 3. the same form is used across multiple morphemes. Language use is a constant process of negotiation between ambiguity of expression and efficiency of communication.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Ambiguity",
"sec_num": "5.5"
},
{
"text": "For unit testing of noun, verb, and verb predicate analyses, we constructed forms for several distinct analyses each of 20 nouns sampled over possessor class and 20 verbs sampled over subject and object classes. Diverse analyses varied possessor/subject/object person, number, and gender as well as noun or verb incorporants. While resulting derived forms were largely consistent with analyses, we discovered and corrected several cases of lexically specified vowel elision and rhotacism not covered in Hanson (2010)'s grammar.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "6"
},
{
"text": "For coverage on test data we sampled words matching on known root forms with 25 each of noun roots and verb roots sampled at random from a Yine corpus by Bustamante et al. (2020) . This resulted in many out of vocabulary words from longer root forms than those used for selection. Yet, there remained numerous other words unrecognized (not covered) by the analyzer even though sharing the expected root. So we performed a detail error analysis from a sub-sample of 63 unrecognized words to diagnose errors and make model improvements.",
"cite_spans": [
{
"start": 154,
"end": 178,
"text": "Bustamante et al. (2020)",
"ref_id": "BIBREF2"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "6"
},
{
"text": "The error analysis is reported in table 1. Some forms suffered from multiple errors and so error counts exceed the number of words sampled. For nouns major reasons for lack of coverage are: 1. morpheme not in FST vocabulary, 2. non-verbal predicate, 3. verbalizer changed category to verb, 4. noun root entry incorrect. For verbs major reasons for lack of coverage are: 1. morpheme not in FST vocabulary, 2. elision and rhotacism alterations, 3. nominalizer changed category to noun, 4. morpheme has more flexible order.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "6"
},
{
"text": "Corrections and improvements from easy to hard are: 1. Correct out of vocabulary, entry, and orthographic errors of roots on vocabulary spreadsheets. 2. Correct intermediate flags and alterations for elision and rhotacism. 3. Add missing suffixes and more flexible order for morpheme out of vocabulary and order errors. 4. Prioritize development of non-verbal predicate, nominalizer, and verbalizer functions to address non-verbal predicate and change of category errors. Cardenas and Zeman (2018) ",
"cite_spans": [
{
"start": 472,
"end": 497,
"text": "Cardenas and Zeman (2018)",
"ref_id": "BIBREF3"
}
],
"ref_spans": [],
"eq_spans": [],
"section": "Evaluation",
"sec_num": "6"
},
{
"text": "We have shown our initial steps in developing noun, verb, verb predicate and pronoun categories for a morphological model of the Yine language, illustrating analyses performed and FST patterns used to solve challenging problems. Testing for analyzer coverage with real world data revealed several deficiencies, some expected (nominalizers, verbalizers, non-verbal predicate) and some surprises (unexpected elision, rhotacism, and missing morpheme errors). We will continue to improve the analyzer by fixing problems and adding major word categories and functions, now with added emphasis on testing with external data. Goals for the analyzer include both language documentation and use as a component of natural language processing (NLP) applications such as spell checking and low resource machine translation.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Conclusion",
"sec_num": "7"
},
{
"text": "The annotations shown in (13, 14) use standard fourline glossing format customary in contemporary grammatical description. Output from the FST morphological analyzer is added as a fifth line of the gloss.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Intermediate flags are an essential technique for triggering alterations. See alteration rule examples in (Hulden, 2011).",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
},
{
"text": "Beesley and Karttunen (2003, pp 299-230) explains a lexical filter version of this. In our implementation, filter flags are written to the lower tape and post-lexical filters applied to eliminate duplicate incorporant types.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "",
"sec_num": null
}
],
"back_matter": [
{
"text": "We are grateful to the bilingual teachers from NOPOKI Nimia Acho and Remigio Zapata. Similarly, we acknowledge the research grant of the \"Consejo Nacional de Ciencia, Tecnolog\u00eda e Innovaci\u00f3n Tecnol\u00f3gica\" (CONCYTEC, Peru) under the contract 183-2018-FONDECYT-BM-IADT-MU.",
"cite_spans": [],
"ref_spans": [],
"eq_spans": [],
"section": "Acknowledgements",
"sec_num": null
}
],
"bib_entries": {
"BIBREF0": {
"ref_id": "b0",
"title": "Morhology in arawakan languages",
"authors": [
{
"first": "Alexandra",
"middle": [],
"last": "Aikhenvald",
"suffix": ""
}
],
"year": 2020,
"venue": "Oxford Research Encyclopedia of Linguistics",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Alexandra Aikhenvald. 2020. Morhology in arawakan languages. Oxford Research Encyclopedia of Lin- guistics.",
"links": null
},
"BIBREF1": {
"ref_id": "b1",
"title": "Finite State Morphology. CSLI Studies in Computational Linguistics Online. CSLI publications, Stanford University",
"authors": [
{
"first": "R",
"middle": [],
"last": "Kenneth",
"suffix": ""
},
{
"first": "Lauri",
"middle": [],
"last": "Beesley",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Karttunen",
"suffix": ""
}
],
"year": 2003,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Kenneth R. Beesley and Lauri Karttunen. 2003. Finite State Morphology. CSLI Studies in Computational Linguistics Online. CSLI publications, Stanford Uni- versity, Stanford, CA, USA.",
"links": null
},
"BIBREF2": {
"ref_id": "b2",
"title": "No data to crawl? monolingual corpus creation from PDF files of truly low-resource languages in Peru",
"authors": [
{
"first": "Gina",
"middle": [],
"last": "Bustamante",
"suffix": ""
},
{
"first": "Arturo",
"middle": [],
"last": "Oncevay",
"suffix": ""
},
{
"first": "Roberto",
"middle": [],
"last": "Zariquiey",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 12th Language Resources and Evaluation Conference",
"volume": "",
"issue": "",
"pages": "2914--2923",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Gina Bustamante, Arturo Oncevay, and Roberto Zariquiey. 2020. No data to crawl? monolingual corpus creation from PDF files of truly low-resource languages in Peru. In Proceedings of the 12th Lan- guage Resources and Evaluation Conference, pages 2914-2923, Marseille, France. European Language Resources Association.",
"links": null
},
"BIBREF3": {
"ref_id": "b3",
"title": "A morphological analyzer for Shipibo-konibo",
"authors": [
{
"first": "Ronald",
"middle": [],
"last": "Cardenas",
"suffix": ""
},
{
"first": "Daniel",
"middle": [],
"last": "Zeman",
"suffix": ""
}
],
"year": 2018,
"venue": "Proceedings of the Fifteenth Workshop on Computational Research in Phonetics, Phonology, and Morphology",
"volume": "",
"issue": "",
"pages": "131--139",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Ronald Cardenas and Daniel Zeman. 2018. A mor- phological analyzer for Shipibo-konibo. In Proceed- ings of the Fifteenth Workshop on Computational Re- search in Phonetics, Phonology, and Morphology, pages 131-139, Brussels, Belgium. Association for Computational Linguistics.",
"links": null
},
"BIBREF4": {
"ref_id": "b4",
"title": "Ashaninka-morph",
"authors": [
{
"first": "Richard Alexander Castro",
"middle": [],
"last": "Mamani",
"suffix": ""
}
],
"year": 2020,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Richard Alexander Castro Mamani. 2020. Ashaninka-morph. github at https://github.com/hinantin/AshMorph.",
"links": null
},
"BIBREF5": {
"ref_id": "b5",
"title": "A Grammar of Yine (Piro)",
"authors": [
{
"first": "Rebecca",
"middle": [],
"last": "Hanson",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Rebecca Hanson. 2010. A Grammar of Yine (Piro).",
"links": null
},
"BIBREF6": {
"ref_id": "b6",
"title": "Foma: a finite-state compiler and library",
"authors": [
{
"first": "",
"middle": [],
"last": "Mans Hulden",
"suffix": ""
}
],
"year": 2009,
"venue": "Proceedings of the 12th Conference of the European Chapter of the Association for Computational Linguistics",
"volume": "",
"issue": "",
"pages": "29--32",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mans Hulden. 2009. Foma: a finite-state compiler and library. In Proceedings of the 12th Conference of the European Chapter of the Association for Com- putational Linguistics, pages 29-32. Association for Computational Linguistics.",
"links": null
},
"BIBREF7": {
"ref_id": "b7",
"title": "Morphological analysis with fsts",
"authors": [
{
"first": "",
"middle": [],
"last": "Mans Hulden",
"suffix": ""
}
],
"year": 2011,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mans Hulden. 2011. Morphological anal- ysis with fsts. Document in github: https://fomafst.github.io/morphtut.html.",
"links": null
},
"BIBREF8": {
"ref_id": "b8",
"title": "Twenty-five years of finite-state morphology",
"authors": [
{
"first": "Lauri",
"middle": [],
"last": "Karttunen",
"suffix": ""
},
{
"first": "Kenneth",
"middle": [
"R"
],
"last": "Beesley",
"suffix": ""
}
],
"year": 2005,
"venue": "Inquiries into Words, Constraints and Contexts. CSLI publications",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Lauri Karttunen and Kenneth R. Beesley. 2005. Twenty-five years of finite-state morphology. In Antti Arppe, Lauri Carlson, Krister Lind\u00e9n, Jussi Pi- itulainen, Mickael Suominen, Martti Vainio, Hanna Westerlund, and Anssi Yli-Jyr\u00e4, editors, Inquiries into Words, Constraints and Contexts. CSLI publi- cations, Stanford University.",
"links": null
},
"BIBREF9": {
"ref_id": "b9",
"title": "Unimorph 2.0: Universal morphology",
"authors": [
{
"first": "Christo",
"middle": [],
"last": "Kirov",
"suffix": ""
},
{
"first": "Ryan",
"middle": [],
"last": "Cotterell",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Sylak-Glassman",
"suffix": ""
},
{
"first": "G\u00e9raldine",
"middle": [],
"last": "Walther",
"suffix": ""
},
{
"first": "Ekaterina",
"middle": [],
"last": "Vylomova",
"suffix": ""
},
{
"first": "Patrick",
"middle": [],
"last": "Xia",
"suffix": ""
},
{
"first": "Manaal",
"middle": [],
"last": "Faruqui",
"suffix": ""
},
{
"first": "Arya",
"middle": [
"D"
],
"last": "Mccarthy",
"suffix": ""
},
{
"first": "Sandra",
"middle": [],
"last": "K\u00fcbler",
"suffix": ""
},
{
"first": "David",
"middle": [],
"last": "Yarowsky",
"suffix": ""
},
{
"first": "Jason",
"middle": [],
"last": "Eisner",
"suffix": ""
}
],
"year": 2018,
"venue": "Proceedings of the Eleventh International Conference on Language Resources and Evaluation",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christo Kirov, Ryan Cotterell, John Sylak-Glassman, G\u00e9raldine Walther, Ekaterina Vylomova, Patrick Xia, Manaal Faruqui, Arya D. McCarthy, Sandra K\u00fcbler, David Yarowsky, Jason Eisner, , and Mans Hulden. 2018. Unimorph 2.0: Universal morphol- ogy. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018). European Language Resources Asso- ciation (ELRA).",
"links": null
},
"BIBREF10": {
"ref_id": "b10",
"title": "Atlas of the world's languages in danger",
"authors": [],
"year": 2010,
"venue": "UNESCO",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Christopher Moseley, editor. 2010. Atlas of the world's languages in danger, 3rd edition. UNESCO, Paris, France.",
"links": null
},
"BIBREF11": {
"ref_id": "b11",
"title": "Overcoming resistance: The normalization of an Amazonian tribal language",
"authors": [
{
"first": "John",
"middle": [],
"last": "Ortega",
"suffix": ""
},
{
"first": "Richard",
"middle": [
"Alexander"
],
"last": "Castro-Mamani",
"suffix": ""
},
{
"first": "Jaime Rafael Montoya",
"middle": [],
"last": "Samame",
"suffix": ""
}
],
"year": 2020,
"venue": "Proceedings of the 3rd Workshop on Technologies for MT of Low Resource Languages",
"volume": "",
"issue": "",
"pages": "1--13",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "John Ortega, Richard Alexander Castro-Mamani, and Jaime Rafael Montoya Samame. 2020. Overcom- ing resistance: The normalization of an Amazonian tribal language. In Proceedings of the 3rd Workshop on Technologies for MT of Low Resource Languages, pages 1-13, Suzhou, China. Association for Compu- tational Linguistics.",
"links": null
},
"BIBREF12": {
"ref_id": "b12",
"title": "Describing Morphosyntax: A guide for field linguists",
"authors": [
{
"first": "E",
"middle": [],
"last": "Thomas",
"suffix": ""
},
{
"first": "",
"middle": [],
"last": "Payne",
"suffix": ""
}
],
"year": 2007,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Thomas E Payne. 2007. Describing Morphosyntax: A guide for field linguists. Cambridge University Press.",
"links": null
},
"BIBREF13": {
"ref_id": "b13",
"title": "Creating and weighting hunspell dictionaries as finite-state automata",
"authors": [
{
"first": "Tommi",
"middle": [
"A"
],
"last": "Pirinen",
"suffix": ""
},
{
"first": "Krister",
"middle": [],
"last": "Lind\u00e9n",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Tommi A. Pirinen and Krister Lind\u00e9n. 2010. Creating and weighting hunspell dictionaries as finite-state au- tomata. Investigationes Linguisticae.",
"links": null
},
"BIBREF14": {
"ref_id": "b14",
"title": "Stateof-the-art in weighted finite-state spell-checking",
"authors": [
{
"first": "Tommi",
"middle": [
"A"
],
"last": "Pirinen",
"suffix": ""
},
{
"first": "Krister",
"middle": [],
"last": "Lind\u00e9n",
"suffix": ""
}
],
"year": 2014,
"venue": "Proceedings of the 15th International Conference on Computational Linguistics and Intelligent Text Processing, CICLing",
"volume": "",
"issue": "",
"pages": "519--532",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Tommi A. Pirinen and Krister Lind\u00e9n. 2014. State- of-the-art in weighted finite-state spell-checking. In Proceedings of the 15th International Conference on Computational Linguistics and Intelligent Text Pro- cessing, CICLing 2014, pages 519-532, Berlin, Hei- delberg. Springer-Verlag.",
"links": null
},
"BIBREF15": {
"ref_id": "b15",
"title": "Applying Finite-State Techniques to a Native American Language: Quechua",
"authors": [
{
"first": "Annette",
"middle": [],
"last": "Rios",
"suffix": ""
}
],
"year": 2010,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Annette Rios. 2010. Applying Finite-State Techniques to a Native American Language: Quechua. Ph.D. thesis, Universita\u1e97 Zurich.",
"links": null
},
"BIBREF16": {
"ref_id": "b16",
"title": "Diccionario Piro. Summer Institute of Linguistics",
"authors": [],
"year": 1986,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Mary Ruth Wise, editor. 1986. Diccionario Piro. Sum- mer Institute of Linguistics, Yarinacocha, Per\u00fa.",
"links": null
},
"BIBREF17": {
"ref_id": "b17",
"title": "Gu\u00eda te\u00f3rica del idioma yine",
"authors": [
{
"first": "Remigio",
"middle": [],
"last": "Zapata",
"suffix": ""
},
{
"first": "Nimia",
"middle": [],
"last": "Acho",
"suffix": ""
},
{
"first": "Gerardo",
"middle": [],
"last": "Zerdin",
"suffix": ""
}
],
"year": 2017,
"venue": "",
"volume": "",
"issue": "",
"pages": "",
"other_ids": {},
"num": null,
"urls": [],
"raw_text": "Remigio Zapata, Nimia Acho, and Gerardo Zerdin. 2017. Gu\u00eda te\u00f3rica del idioma yine. Universidad Cat\u00f3lica Sedes Sapientae.",
"links": null
},
"BIBREF18": {
"ref_id": "b18",
"title": "Obsolescencia ling\u00fc\u00edstica, descripci\u00f3n gramatical y documentaci\u00f3n de lenguas en el per\u00fa: hacia un estado de la cuesti\u00f3n",
"authors": [
{
"first": "Roberto",
"middle": [],
"last": "Zariquiey",
"suffix": ""
},
{
"first": "Harald",
"middle": [],
"last": "Hammarstr\u00f6m",
"suffix": ""
},
{
"first": "M\u00f3nica",
"middle": [],
"last": "Arakaki",
"suffix": ""
},
{
"first": "Arturo",
"middle": [],
"last": "Oncevay",
"suffix": ""
},
{
"first": "John",
"middle": [],
"last": "Miller",
"suffix": ""
}
],
"year": 2019,
"venue": "Lexis",
"volume": "43",
"issue": "2",
"pages": "271--337",
"other_ids": {
"DOI": [
"10.18800/lexis.201902.001"
]
},
"num": null,
"urls": [],
"raw_text": "Roberto Zariquiey, Harald Hammarstr\u00f6m, M\u00f3nica Arakaki, Arturo Oncevay, John Miller, Aracelli Gar- c\u00eda, and Adriano Ingunza. 2019. Obsolescencia ling\u00fc\u00edstica, descripci\u00f3n gramatical y documentaci\u00f3n de lenguas en el per\u00fa: hacia un estado de la cuesti\u00f3n. Lexis, 43(2):271-337.",
"links": null
}
},
"ref_entries": {
"FIGREF0": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "you sleep (usually)'"
},
"FIGREF1": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "They taught us (how) to hunt'."
},
"FIGREF4": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "Language views: A) Upper analysis and lower form, B) By level/domain, C) By function."
},
"FIGREF5": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "define \u2022 NRoot \".NROOT\"; define \u2022 VRoot \".VROOT\"; ... define \u2022 Quot \".QUOT\"; # quote define \u2022 Infer \".INFER\"; # inference"
},
"FIGREF8": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "define gAlteration [[g i -> u || \"^g\" _] .o. [g -> 0 || \"^g\" _ ] .o. [\"^g\" -> 0] ];"
},
"FIGREF9": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "-gustaka-siji-ne ARGNO3P-LGSPEC3-cut-corn-PSSD -gima-nanu-je-ta -QUOT-EXTNS-HAB-LGSPEC1 -ya-nu-pa-lu-na -APPL-DED-ALL-ARGAC3SM-ARGNO3P'It is said that they, and someone else (usually) cut their (masc) corn during a specific time lapse' r.ARGNO3P-gim.LGSPEC3-gustaka.VROOT.AMBI -siji.NROOT.2.ALN.ne-ne.PSSD-gima.QUOT -nanu.EXTNS-je.HAB-ta.LGSPEC1-ya.APPL -nu.DED-pa.ALL-lu.ARGAC3SM-na.ARGNO3P"
},
"FIGREF10": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "define VerbAssoc [{gim} [ \u2022 Assoc %-]:\"^g\"];"
},
"FIGREF11": {
"type_str": "figure",
"uris": null,
"num": null,
"text": "a snippet for the FST of all duplicate filter flags. Each line such as $[\"^I.A\" ? * \"^I.A\"] denotes the language containing that filter flag duplicated, and the union over all such flags denotes the union of languages with duplicate flags. Taking the complement of this results in all languages without duplicate flags, and composing this complement with the actual group of incorporants, excludes any cases where there are duplicate flags. This is a powerful operator! define detectIncorporantDuplicates [ $[\"^I.A\" ? * \"^I.A\"] | $[\"^I.B\" ? * \"^I.B\"] ... | $[\"^I.I\" ? * \"^I.I\"]];"
},
"TABREF0": {
"type_str": "table",
"text": "includes an optional Lexc language for lexicons and an expansive FST language. While Lexc is a good fit for ordered concatenative morphology and is accessible for entering inventories of open category roots, it is not a natural fit for the highly agglutinative polysynthetic Yine language with its relatively free order of suffixes. Instead open category root inventories are edited in spreadsheets and exported via Possibly multiple stacked transducers sawli.NROOT.2.ALN.te-te.PSSD",
"num": null,
"html": null,
"content": "<table><tr><td>A</td><td>B</td><td>C</td></tr><tr><td>Analysis</td><td/><td/></tr><tr><td/><td>High level Lexc language for lexicons</td><td>Lexical filters</td></tr><tr><td/><td>Regular expressions defining FSMs Advanced operators on FSMs and FSTs</td><td>Lexicon Post-lexical filters</td><td>Memory filters</td></tr><tr><td>sawlite</td><td>Meta-language for</td><td>Alterations</td></tr><tr><td>Form</td><td>interacting with FST</td><td/></tr></table>"
},
"TABREF2": {
"type_str": "table",
"text": ".NROOT.1.ALN.te-te.PSSD Listing 3 shows how noun possession is defined by FST. Inalienable unpossessed state is marked with -chi by selecting inalienable nouns, $[\u2022Inalienable], from noun roots, NRoot, writing the noun root and -chi \u2022Unposs on the upper tape, and noun root and^V chi on the lower tape. $[\u2022Inalienable] is a lexical filter which when composed, .o., with noun roots from the lexicon selects only inalienable noun roots. The intermediate flag^V subsequently triggers a final vowel elision, defined by VElision. 5 Alienable possessed state is defined similarly except that for possessed suffix -te there is no final vowel elision.",
"num": null,
"html": null,
"content": "<table><tr><td>define NounInalienUnposs $[ \u2022 Inalienable]</td></tr><tr><td>.o. [NRoot %-:\"^V\" {chi} \u2022 Unposs:0]</td></tr><tr><td>.o. VElision;</td></tr><tr><td>define NounTe $[ \u2022 Alienable \u2022 PossSfxte]</td></tr><tr><td>.o. [NRoot %-:0 {te} \u2022 Poss:0];</td></tr><tr><td>kamrurchi</td></tr><tr><td>kamruru-chi</td></tr><tr><td>work-UNPSSD</td></tr><tr><td>'(the unpossessed) work'</td></tr><tr><td>kamruru.NROOT.3.NALN-chi.UNPSSD</td></tr><tr><td>(14) gagmunate</td></tr><tr><td>gagmuna-te</td></tr><tr><td>tree-PSSD</td></tr><tr><td>'(a possessed) tree'</td></tr></table>"
},
"TABREF5": {
"type_str": "table",
"text": "obtained 78.9% average coverage over multiple domains on test data for a completed FST morphology of an Amazonian polysynthetic language. Our \u224815% coverage in",
"num": null,
"html": null,
"content": "<table><tr><td>Error</td><td colspan=\"2\">Nouns Verbs</td></tr><tr><td>Root out of vocabulary</td><td>9</td><td>6</td></tr><tr><td>Morpheme out of vocabulary</td><td>7</td><td>10</td></tr><tr><td>Morpheme out of order</td><td>1</td><td>3</td></tr><tr><td>Elision incorrect</td><td>0</td><td>9</td></tr><tr><td>Rhotacism incorrect</td><td>0</td><td>6</td></tr><tr><td>Orthographic mismatch</td><td>0</td><td>2</td></tr><tr><td>Change of category</td><td>5</td><td>7</td></tr><tr><td>Non-verbal predicate</td><td>7</td><td>0</td></tr><tr><td>Root entry incorrect</td><td>5</td><td>1</td></tr><tr><td>Error counts</td><td>34</td><td>44</td></tr><tr><td>Sample size</td><td>30</td><td>33</td></tr><tr><td>Total words sampled</td><td>574</td><td>1292</td></tr><tr><td>Percentage recognized</td><td colspan=\"2\">11.3% 16.1%</td></tr></table>"
},
"TABREF6": {
"type_str": "table",
"text": "Lack of Coverage Reasons a preliminary evaluation on multiple domain test data should be interpreted as a measure of the effort still to go on this project. Our goal remains a high coverage FST morphological analyzer.",
"num": null,
"html": null,
"content": "<table/>"
}
}
}
} |