In [30], an ICALL system called Your Verbal Zone (YVZ) is described, which is used to support Turkish students’ English vocabulary learning. YVZ contains a morphological analyzer which is able to find the root of a morphologically complex word and return its base form together with affixes. It, of course, uses NLP in order to achieve that. In order to further support students, the system contains a built-in, bilingual dictionary, a number of examples related to word use and details of function and meaning of particular affixes.
In [13], the case for WordBricks is argued, an intelligent system supporting the process of writing grammatically correct sentences. The system features a grammar checker and allows students to produce free utterances. This also permits learners to experiment with language and test their own hypotheses on language structures. Feedback generation system is also implemented.
In [31], a program called C-DA, used for improving learners’ metacognitive reading strategies is described. The standalone program, also available in the Web browser, keeps track of learner answers to questions, and for each incorrect answer, it initiates the “mediation” mechanism which helps the learner solve the task. The mediation mechanism is designed to offer more implicit help at first, and if the error persists and the learner is unable to provide the correct answer, help gradually becomes more explicit.
An ICALL system called CASTLE is described in [32]. It provides an opportunity for language learners to practice their communicative skills, namely writing, through a number of predetermined role-play scenarios. Upon wrong input and depending on how the error is classified by the diagnosis module, CASTLE takes remedial action in the form of additional grammar exercises to stress certain formal language structures. In the same time, the system keeps an up-to-date student model which holds data concerning student proficiency for a particular topic, data about errors and proneness to commit them, and the estimates of the likely cause, so the system is able to adapt its behavior towards each learner.
A description of a Web Passive Voice Tutor is given in [33]. The system uses stereotype modeling to initialize the student model, applies NLP techniques for intelligent analysis of student solutions to tasks and tailors feedback and advice to each student separately. In addition, it uses the so called link annotation technique to guide the learner through the learning domain by suggesting the most appropriate activity to follow.
I-PETER, as described in [34], is an intelligent system for language learning that allows for three distinct ways of learning: identifying and mending holes in the existing knowledge of the subject matter; improving on the current level of knowledge of a selected part of the domain; selecting and practicing only selected concepts and/or subconcepts of the domain. In each case, the learner needs to give an answer to a particular question, and once it is evaluated by the system, the learner may request detailed theoretical explanation regarding the error. In addition, there is a possibility for the system to fully take charge of the learning process.