ePark
Overview
The ePark corpus is a comprehensive and interactive resource for the preservation, learning, and revitalization of Indigenous languages of Taiwan. Developed by the Indigenous Languages Research and Development Foundation (ILRDF), this digital platform caters to a wide audience, including preschoolers, students, adults, and language teachers, with resources and tools designed to support various learning levels and linguistic goals. The corpus is available in all recognized 42 official dialects across the 16 different Formosan languages. The corpus is invaluable for documenting and preserving linguistic diversity, and it includes text, audio recordings, and translations, making it a comprehensive resource for research, education, and language revitalization.
Corpus Statistics
Amis Coastal
Amis Hengchun
Amis Malan
Amis Southern
Amis Xiuguluan
Bunun Junqun
Bunun Kaqun
Bunun Luanqun
Bunun Tanqun
Bunun Zhuoqun
Kavalan
Rukai Dawu
Rukai Dona
Rukai Eastern
Rukai Maolin
Rukai Wanshan
Rukai Wutai
Paiwan Central
Paiwan Eastern
Paiwan Northern
Paiwan Southern
Puyuma Jianhe
Puyuma Nanwang
Puyuma Xiqun
Puyuma Zhiben
Thao
Saaroa
Sakizaya
Yami
Atayal FourSeasons
Atayal Sekolik
Atayal Wanda
Atayal Wenshui
Atayal YilanZeaol
Atayal Zeaol
Seediq DeluValley
Seediq Duda
Seediq Tegudaya
Truku
Tsou
Kanakanavu
Saisiyat
Word count
58,810
36,726
35,979
36,318
37,022
52,965
29,637
30,240
29,700
28,829
55,429
29,959
28,596
30,129
27,665
24,434
54,258
34,340
35,996
59,648
37,996
35,216
65,186
37,017
39,555
55,072
42,787
53,657
63,684
38,123
59,188
30,977
36,910
40,168
39,189
39,565
41,166
57,792
56,323
56,555
49,011
50,629
Total audio
13.4h
10.2h
8.4h
9.2h
9.9h
16.2h
8.0h
7.9h
8.6h
8.1h
13.1h
9.5h
8.9h
9.4h
8.8h
8.4h
15.9h
9.4h
10.0h
13.9h
10.5h
9.4h
19.2h
9.9h
10.8h
11.5h
16.0h
14.0h
13.9h
9.2h
14.0h
8.8h
9.5h
10.3h
9.9h
11.7h
10.1h
12.4h
13.7h
15.2h
16.1h
12.5h
Transcribed
13.4h
10.2h
8.4h
9.2h
9.9h
16.2h
8.0h
7.9h
8.6h
8.1h
13.1h
9.5h
8.9h
9.4h
8.8h
8.4h
15.9h
9.4h
10.0h
13.9h
10.5h
9.4h
19.2h
9.9h
10.8h
11.5h
16.0h
14.0h
13.9h
9.2h
14.0h
8.8h
9.5h
10.3h
9.9h
11.7h
10.1h
12.4h
13.7h
15.2h
16.1h
12.5h
Untranscribed
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Translated words
English
9,727
9,425
9,306
10,323
9,283
8,075
8,106
8,611
8,121
8,537
8,866
7,576
7,940
8,183
7,272
6,537
7,545
10,154
9,303
10,006
9,830
9,218
9,445
9,887
10,139
9,302
7,646
9,160
9,657
9,456
10,220
9,145
10,605
9,497
9,716
9,356
10,044
8,937
9,401
9,137
7,706
8,897
Mandarin
58,794
36,710
35,979
35,968
36,665
52,957
29,621
30,224
29,684
28,791
55,400
29,935
28,262
30,061
27,485
24,367
54,228
34,324
35,795
59,405
37,915
35,182
65,096
36,993
39,539
55,056
42,771
53,441
63,684
38,073
58,840
30,913
36,886
40,152
39,173
39,428
41,085
57,776
56,307
56,518
48,989
50,581
Japanese
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Dutch
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Morphologically segmented
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Glossed words
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Access Details
Visit the ePark corpus online platform at https://web.klokah.tw/
The repo containing the ePark corpus in FormosanBank as well as the code to reconstruct the corpus can be found here.
Acknowledgments
The ePark corpus was developed through collaboration between ILRDF, educators, linguists, and Indigenous communities. Without the tremendous effort and collaborations of these entities, it wouldn't have been possible to have such a valuable resource in FormosanBank.
Corpus Notes
The corpus appears to use the standard orthography ("Ortho113" in FormosanBank nomenclature), with some exceptions:
Ortho113 specifies that only the Nanshi dialect uses
u, while the other dialects useo. However, there are a fair number ofos andus throughout the dialect, irrespective of dialect. Thestandardtier normalizes these to match the standard orthography. For the IPA transcriptions of theoriginaltier, we recognize eitherooruas referring to the same phoneme. .
Copyright
The content of the ePark corpus is released under the Creative Commons CC BY-NC-SA 4.0 license. Further information is available at https://web.klokah.tw/creativeCommons/.
Citation
In accordance with our Terms of Use, if you use this corpus or any product derived from this corpus in any publication, you must cite both FormosanBank and:
Indigenous Languages Research and Development Foundation. (2020). 族語E樂園. https://web.klokah.tw/
Last updated