The Xi’an Multi-Language Learner Corpus

<h3>Introduction</h3>
<p>The Xi&rsquo;an Multi-Language Learner Corpus was developed by <a href="https://en.xisu.edu.cn/">Xi'an International Studies University (XISU)</a>. It is comprised of 526 argumentative essays in 15 languages by Chinese L1 university students studying second languages, along with student metadata and writing prompts. It was developed to support second language learner research and to provide a database for cross-linguistic comparison of second languages.</p>
<h3>Data</h3>
<p>The essays were produced by undergraduate students at XISU and <a href="https://web.ymu.edu.cn/gjxy/index.htm">Yunnan Minzu University (YMU)</a> in response to writing prompts prepared by the corpus development team. Data was collected in 2023 and 2024. Participating students were linguistic majors or studying one of the foreign languages available at XISU and YMU. Off-topic essays and incomplete texts were excluded</p>
<p>All texts were cleaned and formatted. No changes were made to the texts in relation to grammatical tense or turn of phrase accuracy.</p>
<p>Text and token counts by language are as follows:</p>
<table>
<thead>
<tr>
<th class="language">Language</th>
<th class="numeric">texts</th>
<th class="numeric">tokens</th>
</tr>
</thead>
<tbody>
<tr>
<td class="language">Arabic</td>
<td class="numeric">8</td>
<td class="numeric">1,762</td>
</tr>
<tr>
<td class="language">English</td>
<td class="numeric">107</td>
<td class="numeric">32,822</td>
</tr>
<tr>
<td class="language">Filipino</td>
<td class="numeric">10</td>
<td class="numeric">1,371</td>
</tr>
<tr>
<td class="language">French</td>
<td class="numeric">129</td>
<td class="numeric">39,944</td>
</tr>
<tr>
<td class="language">German</td>
<td class="numeric">78</td>
<td class="numeric">10,941</td>
</tr>
<tr>
<td class="language">Hindi</td>
<td class="numeric">16</td>
<td class="numeric">2,972</td>
</tr>
<tr>
<td class="language">Indonesian</td>
<td class="numeric">14</td>
<td class="numeric">2,630</td>
</tr>
<tr>
<td class="language">Korean</td>
<td class="numeric">24</td>
<td class="numeric">2,630</td>
</tr>
<tr>
<td class="language">Malay</td>
<td class="numeric">36</td>
<td class="numeric">5,208</td>
</tr>
<tr>
<td class="language">Persian</td>
<td class="numeric">12</td>
<td class="numeric">1,751</td>
</tr>
<tr>
<td class="language">Russian</td>
<td class="numeric">33</td>
<td class="numeric">8,018</td>
</tr>
<tr>
<td class="language">Swahili</td>
<td class="numeric">10</td>
<td class="numeric">1,840</td>
</tr>
<tr>
<td class="language">Thai</td>
<td class="numeric">12</td>
<td class="numeric">1,661</td>
</tr>
<tr>
<td class="language">Turkish</td>
<td class="numeric">22</td>
<td class="numeric">3,719</td>
</tr>
<tr>
<td class="language">Urdu</td>
<td class="numeric">15</td>
<td class="numeric">3,645</td>
</tr>
</tbody>
</table>
<p>&nbsp;</p>
<p><a href="https://lancsbox.lancs.ac.uk/">LancsBox X 4.0</a> was used for counting Swahili, Persian, French, Urdu, and Hindi tokens. <a href="https://www.laurenceanthony.net/software/antconc/">AntConc 4.2.4</a> was used for counting tokens in the other languages.</p>
<p>The essays and writing prompts are stored in UTF-8 encoded plain text files. Metadata is presented in .csv files.</p>