The Polish Learner Corpus

As learners acquire foreign languages, they tend to apply the structures of their mother tongue to a target language. This phenomenon, called language transfer, became the focus of second language acquisition more than half a century ago (Selinker 1969) and it is still crucial in language teaching and learning (Gass and Swlinker 1992). Knowing transfer errors, teachers are able to adjust classroom materials and be aware of features that may be easy or difficult for learners (Odlin 1989: 4). Moreover, similarities between languages can be helpful in language acquisition (1989: 27), especially as concerns genetically close languages such as Polish and Czech, in which there is a high level of mutual interference. Czech learners benefit from positive transfer while using common structures for both languages. They acquire the Polish language system much faster than non-Slavic learners. However, they also fall into the trap of using structures seemingly common for both languages but in fact caused by negative transfer. As a result, the production of Czech learners in the Polish language is characterised by a unique set of language errors typical for this group of learners. Our research (Kaczmarska & Zasina 2021) is the first attempt to explore this area using a learner corpus.

Polish in writings of czech-speaking learners on basis of learner corpus PoLKo Acquiring a foreign language, learners tend to apply structures from their mother tongue to a target language. Especially, it is the case of genetically close languages such as Polish and Czech where a mutual interference is high. Czech learners benefit from a positive transfer while using common structures for both languages. They acquire the Polish language system much faster than non-Slavic learners. However, they also get into a trap using structures seemingly common for both languages caused by a negative transfer. As a result, the production of Czech learners in the Polish language is characterised by a unique set of language errors, which is typical for this group of learners. Since previous studies did not pay enough attention to this type of errors, our research aims to fill the gap. Using data from a Polish learner corpus PoLKo, this pilot study analyses the most prominent errors in the area of valency, syntax, spelling, lexis, etc. Our observations deliver a suitable source for learners and their teachers that might be used for the preparation of tailor-made study materials. This approach undoubtedly has a chance to improve the teaching methods of Polish as a foreign language among Czech-speaking learners.

The present article addresses valency errors in writings of non-native speakers learning Polish as a foreign language. Valency is a key element in a foreign language acquisition, and yet there are no studies on valency errors based on empirical data for Polish as a foreign language. Therefore, this study presents the first attempt to examine valency errors based on data from the Polish Learner Corpus PoLKo. The pilot analysis deals with different proficiency levels (A1-C1) and nationalities (Slavic and non-Slavic) of learners. The corpus material has shown that valency errors are present across all language levels and for different nationalities. Nevertheless , valency errors are more common in prepositional phrases among learners with uninflected mother tongue. Further investigation into the subject is obviously needed, and it may be expected that it will bring new results and new conclusions, along with the development of PoLKo.

Please find the video presenetation here: https://www.youtube.com/watch?v=fJL7CbWXAHQ In recent years, learner corpora have become increasingly popular as a source for analysing L2 learners’ language (Gilquin, Granger, & Paquot, 2007: 322–323). However, in comparison with national corpora, which are available for many languages, we find relatively few learner corpora for other languages than English (Štindlová, 2013: 62–65). In particular, a learner corpus is still missing that could drive analyses of the language of non-native speakers of Polish (Zasina, 2019). Therefore, the goal of our poster is to present the first attempt to compile such a learner corpus and report on the ongoing project. The primary goal of the project is to collect learners’ writings in Polish as a foreign language at various levels of language proficiency. The collected material will be a basis for analysing the learners’ language, identifying the most common language errors, creating classroom materials, and improving modern teaching methods. In the first step, we are going to collect all available electronic texts, so as to gather a sizeable amount of starting material in the shortest possible time. In the second step of our project, we intend to focus more on hand-written texts and the rules for transcribing such texts in a computer-readable format, and on balancing the entire corpus in terms of first language and language level (CEFR). In the initial phase, the corpus will be prepared in the TeiTok environment (Janssen, 2016), where it can be easily modified during its creation. TeiTok will also be used for text transcription and managing all collected learners’ writings.