The ICoN Corpus of Academic Written Italian (L1 and L2)

The ICoN Corpus of Academic Written Italian (L1 and L2) Federica Cominetti, Mirko Tavosanis Università di Firenze, Università di Pisa Piazza di San Marco, 4, Firenze, Italy, via Santa Maria 36, Pisa, Italy fedcominetti@gmail.com, mirko.tavosanis@gmail.com Abstract This paper describes the ICoN corpus, a corpus of academic written Italian, some of the directions of research it could open, and some of the first outcomes of research conducted on it. The ICoN corpus includes 2,115,000 tokens written by students having Italian as L2 students (level B2 or higher) and 1,769,000 tokens written by students having Italian as L1; this makes it the largest corpus of its kind. The texts included in the corpus come from the online examinations taken by 787 different students for the ICoN Degree Program in Italian Language and Culture for foreign students and Italian citizens residing abroad. The texts were produced by students having 41 different L1s, and 18 different L1s are represented in the corpus by more than 20,000 tokens. The corpus is encoded in XML files; it can be freely queried online and it is available upon request for research purposes. The paper includes the discussion of preliminary research in the field of collocations, showing that, in the texts included in the corpus, while learners and natives do use multiword expressions in a similar way, learners can overuse relatively infrequent forms of multiword adverbials, or use some adverbials in a non-standard way. Keywords: Corpus linguistics, Learners, Italian, multiword expressions, collocations 1. Introduction Learner corpora are useful and widely used in the study of language and language learning. In this paper we describe the building of the largest (to date) learner corpus of academic written Italian as L2 and of a large comparable corpus of academic written Italian as L1, collectively called “The ICoN corpus”, already partially described in Tavosanis (2014) and (2016). The corpus can be queried online while its complete contents are available on request for research purposes.1 2. The Context of Learner Corpora Research and the Status of Learner Corpora for Written Italian Being a learner corpus, the ICoN corpus belongs to a class of corpora well understood in theory and much appreciated for its practical uses (for an overview, see Granger, Gilquin and Meunier 2015). Learner corpora today number in the hundreds: the list of learner corpora maintained by Magali Paquot at the Université Catholique de Louvain includes to date 166 entries, while the Learner Corpus Bibliography maintained by the Learner Corpus Association (at the address www.learnercorpusassociation.org/resources/lcb/) includes more than 1,500 entries. In the research tradition of the field, “in terms of medium and text type, the dominant focus was – and to a large extent still is – on writing, in particular essay writing” (Granger - Gilquin -Meunier 2015, 2). Well known examples of this include for example the ICLE (International Corpus of Learner English) corpus. This corpus (described in detail in Granger, Dagneaux, Meunier, Paquot, 2009) includes 2.5 millions of words coming from 3,640 unabridged essays, with an average length of …

Content truncated to 500 words