SEÑAL, The Computational Assistant for the Analysis and Assessment of Spanish Second Language Writing

Abstract

Essay writing is one of the staples of college language classes and one of the most important components of language learning and assessment (ACTFL 2012). From beginning to advanced courses, Spanish language classes in colleges nationwide use an array of essay tasks to assess Spanish as a second language (L2) writing proficiency. In this Technology Showcase presentation, we introduce SEÑAL, a modular NLP application intended to analyze and evaluate different components of essays written by students of Spanish as a second language (L2). Currently, the program has a web-based prototype which can be accessed in (señal.io)[https://señal.io]. SEÑAL can provide a lexical-syntactic analysis of the complexity of an L2 essay, information about the frequency of various grammatical forms (such as nouns, adjectives, adverbs, and various verb tenses), a machine learning L2 proficiency classifier, a machine translation identification system, and a sentiment analyzer of Spanish texts. Future modular additions include a topic detection and readership engagement analyzers. In this Technology Showcase, we will describe the components of SEÑAL and we will discuss the relevance of the program to facilitate objective measurements of language proficiency evaluation and grading. In particular, we argue for the integration of SEÑAL as a curricular grading tool in Spanish language programs, and as a Spanish Writing Assistant for Students. SEÑAL can also be employed as a SLA Research tool for assessing language development. SEÑAL will include future support for other modern languages, such as Portuguese and English.