Natural language processing tools for computer assisted language learning

Natural language processing tools for computer assisted language learning
Anne Vandeventer Faltin (Geneva)

1 Introduction

Computer assisted language learning (CALL) emerged in the early days of computers. Since the early 1960’s, CALL software was designed and implemented. In the drill-and-practice days of language teaching methodology, what could, better than a computer, repeatedly and relentlessly propose exercises to the learners? Since then, advances in second language acquisition (SLA) have helped the shift of interest from repetitive exercises to more communicative tasks.

The range of exercises offered in CALL software has always been constrained by the kind of feedback judged necessary for the learning process and its availability on a computer. Thus, too often, exercises tend to elicit answers easily classified as right or wrong. This leads to all sorts of multiple-choice-questions, fill-in-the-blanks, and similar exercise types which computers can very easily correct automatically. For slightly more complex exercises requiring the learners to write a few words, either no feedback is provided or pattern matching techniques are used, requiring exercise authors to enter many different correct and wrong answers. This very tedious and time consuming task yet only provides appropriate feedback when the learner types in one of the expected answers.

True communicative tasks are in need of more “intelligent” devices to provide appropriate feedback to the users. Natural language processing (NLP) tools seem to be the obvious answer. NLP tools can be advantageously used for the correction of free production exercises where they can diagnose, in a generic way, many kinds of mistakes. Besides exercise correction, NLP tools can also be used as additional help resources within CALL software, to enable the learners to get more of the content materials: written texts can be listened to with the help of a speech synthesizer, sentences can be analyzed with a sentence structure viewer, verb conjugations can be verified with a conjugation tool, to give only a few examples.

A very complete and interesting introduction to NLP in the CALL context can be found in (Nerbonne 2003), which moreover contains very extensive bibliography. Collections of papers discussing the use of NLP for CALL include (Schulze et al. 1998), (Holland et al. 1995), and (Swartz/Yazdani 1992), among others.

In this paper, we illustrate the usefulness of NLP for CALL based on the presentation of three NLP tools [1] designed for French and integrated within a CALL software, namely: (i) a sentence structure viewer displaying information on the words of a sentence and on its principal constituents (section 2); (ii) an error diagnosis system, including a spell checker, a grammar checker, and a coherence checker [2] (section 3); and (iii) a conjugation tool (section 4). Section 5 concludes this paper.

2 Sentence structure viewer

A sentence structure viewer is a resource tool provided to help language learners grasp the structure of a sentence. This can help them understand the sentence better, correct some of their mistakes, or appropriate grammatical subtleties of the language.

Our sentence structure viewer has two different outputs, a linear color display (section 2.2) as well as a hierarchical structure tree display (section 2.3). However, the core of the tool is exactly the same for both outputs: a syntactic parser, described in section 2.1, is used to provide a rich syntactic analysis of the sentences. The information resulting from the parsing process is screened and displayed differently by each distinct output.
2.1 Syntactic parsing

Syntactic analysis is provided by a pre-existing syntactic analyzer, the Fips parser. Fips has been fully described elsewhere (Wehrli 97, Laenzlinger/Wehrli 1991), but we nevertheless very briefly present the main characteristics of this parser as they are important both for the sentence structure viewer and for the error diagnosis system (section 3).

~ by jeanehistoria on December 29, 2009.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: