File size: 19,212 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
{
    "paper_id": "O00-1009",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T07:59:08.331120Z"
    },
    "title": "",
    "authors": [],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "Computer learner corpora have been widely used by SLA/EFL specialists since mid 1990s to gain better insights into authentic learner language. The work presented in this paper examines the inter-language of Taiwanese learners of English from a part-of-speech sequence perspective. Two pre-tagged corpora (one learner corpus and one native corpus) are involved in this work. The experimental results indicate that there are more than one third of eligible POS trigrams that are never practiced by the Taiwanese learners in their writing and the learners have stronger preference than native speakers in using pronouns, especially right after punctuations, verbs and conjunctions.",
    "pdf_parse": {
        "paper_id": "O00-1009",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "Computer learner corpora have been widely used by SLA/EFL specialists since mid 1990s to gain better insights into authentic learner language. The work presented in this paper examines the inter-language of Taiwanese learners of English from a part-of-speech sequence perspective. Two pre-tagged corpora (one learner corpus and one native corpus) are involved in this work. The experimental results indicate that there are more than one third of eligible POS trigrams that are never practiced by the Taiwanese learners in their writing and the learners have stronger preference than native speakers in using pronouns, especially right after punctuations, verbs and conjunctions.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "Rebecca H. Shih * , John Y. Chiang + and F. Tien +",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Part-of-speech Sequences and Distribution in a Learner Corpus of English",
                "sec_num": null
            },
            {
                "text": "With the recognition of its theoretical and practical potential, computer learner corpora (CLC) have been subsequently built up around the world since early 1990s. [1] CLC research aims to gain a better insight into learners' inter-language from the authentic data. The research often involves comparisons between inter-language that learners possess and native language on various linguistic features. For instance, the frequency distributions of most commonly-used words in a native and seven eastern European learner corpora are compared on various parts-of-speech categories [2] ; the use of complement clauses in terms of their frequencies in four learner corpora as contrasted with their native counterparts [3] is studied; the use of adverbial connectors by Swedish learners in comparison with the natives' is examined [4] . ",
                "cite_spans": [
                    {
                        "start": 164,
                        "end": 167,
                        "text": "[1]",
                        "ref_id": "BIBREF0"
                    },
                    {
                        "start": 579,
                        "end": 582,
                        "text": "[2]",
                        "ref_id": "BIBREF1"
                    },
                    {
                        "start": 714,
                        "end": 717,
                        "text": "[3]",
                        "ref_id": "BIBREF2"
                    },
                    {
                        "start": 826,
                        "end": 829,
                        "text": "[4]",
                        "ref_id": "BIBREF4"
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Introduction",
                "sec_num": "1."
            },
            {
                "text": "Perplexity, in speech recognition community, is often referred to as the number of equi-probable choices at each step of word prediction in a language model such as a bigram/trigram model under the assumption that a word depends merely on the previous one/two words. In this work, given a corpus L, the perplexity of the corpus, S(L), can be viewed as a measure of diversity for the next POS in a language model, and it is defined as:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Corpus Perplexity in Bigram and Trigram models",
                "sec_num": "3.1"
            },
            {
                "text": ") ( 2 ) ( L H L S = \u2211 = c c k H N L H ) ( 1 ) ( \u2211 \u2212 = k c c k P c k P k H ) | ( log ) | ( ) ( 2",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Corpus Perplexity in Bigram and Trigram models",
                "sec_num": "3.1"
            },
            {
                "text": "where H(L) is the entropy of the corpus L, N is the size of part-of-speech set, and P(k|c) is the probability that k will be the next POS when the current POS is c.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Corpus Perplexity in Bigram and Trigram models",
                "sec_num": "3.1"
            },
            {
                "text": ". In this experiment, the perplexities of BNC and TLCE corpora are calculated using both bigram and trigram models, and the results are shown in As can be seen in Table 1 , the perplexities of BNC corpus in the two language models are both greater than those of TLCE, especially in the trigram model where the degree of POS diversity in the learner corpus is only 2/3 of BNC's. The above phenomena can be explained by the limiting sentence structure varieties the learners possess.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 163,
                        "end": 170,
                        "text": "Table 1",
                        "ref_id": "TABREF1"
                    }
                ],
                "eq_spans": [],
                "section": "Corpus Perplexity in Bigram and Trigram models",
                "sec_num": "3.1"
            },
            {
                "text": "In order to further understand the limit of structure variety in learners' writing, the numbers of POS trigrams, i.e. sequences of three POSs, used in the two corpora are compared and shown in Table 2 : the number of POS trigrams in the corpora Under the same assumption, Figure 1 depicts the divergence of learners' use of trigrams from BNC, the optimum indicated by the square curve, on the scale of top-ranking trigrams in use. The diamond curve denotes the number of the learners' trigrams that overlap with BNC at the same rank. As illustrated, the learners' curve moves away from the optimum when the scope of the rank enlarges, especially after the rank of 1000. Figure 1 : The divergence of the use of POS trigrams",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 193,
                        "end": 200,
                        "text": "Table 2",
                        "ref_id": "TABREF2"
                    },
                    {
                        "start": 272,
                        "end": 280,
                        "text": "Figure 1",
                        "ref_id": null
                    },
                    {
                        "start": 670,
                        "end": 678,
                        "text": "Figure 1",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "Structure Variety",
                "sec_num": "3.2"
            },
            {
                "text": "As the learners have preference in using certain POS trigrams it is then desirable to understand the learners' preference in using POSs themselves as well. Figure 2 shows the POS distribution in each corpus, and only those taking up at east 5% of the corpus are indicated. Two significant phenomena are observed from the figure. ",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 156,
                        "end": 164,
                        "text": "Figure 2",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": "POS Distribution",
                "sec_num": "3.3"
            },
            {
                "text": "The results of the preliminary experiments above show that there are more than one third of BNC trigrams that the learners never practice in their writing, whereas there are 4.5% of TLCE trigrams which do not appear in the BNC's. It is intended to believe that this small proportion of TLCE trigrams is contributed from the learner's writing errors. However, increasing the size of the native speaker corpus to observe any changes in the distribution of the trigrams will clarify the findings. It is also worth looking into those BNC trigrams that the learners do not know or are not aware of, and then isolating those with high frequency for the pedagogical purpose.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussions and future work",
                "sec_num": "4."
            },
            {
                "text": "The experimental results also suggest that the learners use pronouns excessively in their writing and that they have stronger preference than native speakers in using pronouns right after punctuations, verbs and conjunctions but less preference after prepositions and nouns. Pronouns often appear in the informal register, and as the corpus is composed of college students' compositions as well as their weekly journals, the informality of the journals may contribute partly to their excessive use of pronouns. So, it is desirable in the next stage of the work to divide the learner corpus in terms of its different registers and compare their POS distributions with the native speaker corpus.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Discussions and future work",
                "sec_num": "4."
            }
        ],
        "back_matter": [
            {
                "text": "The authors would like to thank the National Science Council, Taiwan, for supporting this project, and Prof. Ching-Yuan Tsai for his insightful comment.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Acknowledgements",
                "sec_num": null
            }
        ],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "The International Corpus of Learner English, in English Language Corpora: Design, Analysis and Exploitation",
                "authors": [
                    {
                        "first": "S",
                        "middle": [],
                        "last": "Granger",
                        "suffix": ""
                    }
                ],
                "year": 1993,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "57--69",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Granger, S., The International Corpus of Learner English, in English Language Corpora: Design, Analysis and Exploitation, J. Aarts, P.d. Haan, and N. Oostdijk, Editors. 1993, Rodopi: Amsterdam. p. 57-69.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Overstatement in advanced learners' writing: stylistic aspects of adjective intensification",
                "authors": [
                    {
                        "first": "G",
                        "middle": [],
                        "last": "Lorenz",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "53--66",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Lorenz, G., Overstatement in advanced learners' writing: stylistic aspects of adjective intensification, in Learner English on Computer, S. Granger, Editor. 1998, Addison Wesley Longman Limited. p. 53-66.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "Comparing native and learner perspectives on English grammar: a study of complement clauses",
                "authors": [
                    {
                        "first": "D",
                        "middle": [],
                        "last": "Biber",
                        "suffix": ""
                    },
                    {
                        "first": "R",
                        "middle": [],
                        "last": "Reppen",
                        "suffix": ""
                    }
                ],
                "year": null,
                "venue": "Learner English on Computer",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Biber, D. and R. Reppen, Comparing native and learner perspectives on English grammar: a study of complement clauses, in Learner English on Computer, S.",
                "links": null
            },
            "BIBREF4": {
                "ref_id": "b4",
                "title": "The use of adverbial connectors in advanced Swedish learners' written English",
                "authors": [
                    {
                        "first": "M",
                        "middle": [],
                        "last": "Tapper",
                        "suffix": ""
                    }
                ],
                "year": 1998,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Tapper, M., The use of adverbial connectors in advanced Swedish learners' written English, in Learner English on Computer, S. Granger, Editor. 1998,",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF0": {
                "type_str": "figure",
                "text": "Firstly, although N(Noun) and VB(Verb) are the first two leading POSs in both corpora, there exists a distinct discrepancy of the percentage difference between the two. The difference in distribution percentage between N and VB in BNC reaches 9%, whereas merely 1% difference in TLCE. Secondly, PRON(pronoun), the 3rd highest distribution in the learner corpus but the 7 th in BNC, apparently is overused the learners. Distribution of Preceding POSs in PRON bigrams As the previous figure indicates the excessive use of PRON in the learner corpus, the phenomenon is further analyzed by examining the likelihood of each POS preceding PRON in the bigrams. Figure 3 shows the distribution of preceding POSs of PRON in each corpus. As seen, PUNC(punctuation), VB and CONJ(conjuction) are the three most likely POSs in TLCE to be followed by PRON, and the learners also have stronger preference in using these bigrams than the native speakers. By contrast, the bigrams, PREP(preposition)+PRON and N+PRON, are used more frequently by the native speakers than the learners. Distribution of Preceding POSs in PRON bigrams",
                "num": null,
                "uris": null
            },
            "TABREF0": {
                "content": "<table><tr><td>It is based on two corpora: Taiwanese Learner corpus of English (TLCE) and British</td></tr><tr><td>National Corpus (BNC). Both corpora are tagged by TOSCA tagger, using the</td></tr><tr><td>TOSCA-ICLE tagset. The details of the corpora and the tagger will be stated</td></tr><tr><td>subsequently in Section 2, which is followed by a series of experiments in Section 3.</td></tr><tr><td>Conclusions are drawn in Section 4 with future work.</td></tr><tr><td>2. Methodology</td></tr><tr><td>2.1 Corpora: TLCE and BNC</td></tr><tr><td>As stated in the introduction, CLC-research often compares non-native data with</td></tr><tr><td>native data in order to reveal the overuse and/or underuse phenomena in a learner</td></tr><tr><td>corpus. In this work, the Taiwanese Learner Corpus of English (TLCE) is under</td></tr><tr><td>investigation and the British National Corpus (BNC) is used for comparison. TLCE of</td></tr><tr><td>455,000 words is Appendix A)</td></tr></table>",
                "num": null,
                "html": null,
                "type_str": "table",
                "text": ", what particular areas of language behavior that are shared by learners with different backgrounds, and to what extent these phenomena appear in learner English. The aim of the work in this paper is to discover distinctive inter-language features of Taiwanese learners of English in terms of part-of-speech sequences and distribution. For instance, word forms such as takes, took, taken, and taking have the same lemma take. This function facilitates the collocation analysis under the same lemma. TOSCA operates with a lexicon, which currently contains about 160,000 lemma-tag pairs, covering about 90,000 lemmas. The TOSCA-ICLE tagset contains 270 different tags within 18 major word classes.For simplicity, only the major word classes are considered in the current study (see"
            },
            "TABREF1": {
                "content": "<table><tr><td>:</td></tr></table>",
                "num": null,
                "html": null,
                "type_str": "table",
                "text": ""
            },
            "TABREF2": {
                "content": "<table><tr><td>BNC</td><td>TLCE</td><td>overlap</td></tr><tr><td>2531</td><td>1649</td><td>1574</td></tr></table>",
                "num": null,
                "html": null,
                "type_str": "table",
                "text": "As seen in the table, there are 2531 trigram patterns in BNC, 1649 in TLCE, and 1574 in both. If those appearing in BNC can be viewed"
            }
        }
    }
}