Developing Language Learning Tools

Dennie Hoopingarner, Michigan State University
Paper delivered at the language workshop for the International Education Programs Service, February 22, 2007

1. The Web: medium of choice for computer-assisted language learning tools

It's not hard to see how the convergence of technology makes the Internet a natural destination for language learning. The web as a platform for language learning tools has several advantages. One advantage involves the display of multilingual text in a standardized manner. Before the web emerged as a platform, there was a plethora of methods for displaying languages that do not use the Roman alphabet. For example, there are at least five different ways of encoding simplified Chinese characters, and three other encodings for traditional characters. There are nine ways to encode Russian, four encodings for Turkish, and so on. Text that is formatted for a specific encoding only displays correctly on the computer screen when the display mechanism has the ability to render the text correctly. If a user did not have the proper software support for a given language, and did not adjust his or her computer's setting for the appropriate encoding for that language, the text would appear as illegible gibberish.

A solution was achieved in the Unicode standard. The purpose of creating the Unicode standard, as its name suggests, was to support the writing system of nearly every language. Instead of separate encodings for specific languages, the one encoding would contain definitions for not only the ASCII character set, consisting of the English alphabet, numbers, punctuation marks and some special symbols. The encoding would also encompass the extended roman alphabet, such as accented vowels for European languages, the Hebrew and Arabic alphabets, the Cyrillic alphabet, Korean and Japanese phonetic symbols and and even the fifteen thousand or so most-commonly used Chinese characters.

Only in the last few years have operating systems been able to support nearly every language out of the box. Beginning with Windows 2000 and Apple OS X, the text manipulation functions of the operating system are based on the Unicode standard. This gives software applications the ability to support Unicode natively, and developers do not have to require users to purchase and install special layers of software to enable computing in a specific language.

At almost the same time that operating systems were becoming world-savvy, the Internet connected computers into a global network. The implications of the Internet for teaching and learning are vast, and have been the topic of entire conferences. For language learning tools, it means a paradigm shift away from programs running on a standalone computer, and toward a network-based, collaborative, real-time, buzzword-infested way of delivering and accessing language instruction. Students demand 24/7 access to materials and resources, and institutions are seeing the value of providing it. Michael Wesch at Kansas State University says that the ultimate goal of interaction on the web is to reproduce the real world in a virtual environment, moving beyond virtual reality toward a reality that is alternate, but just as real. That's the direction in which we're moving. It might be fair to say that for a given aspect of your teaching, if you aren't at least moving toward doing it on the web, you're in danger of becoming irrelevant.

2. Review of the state of the art in web-based language learning tools

Although CALL has a relatively short history, already there are some common functions that seem to emerge over and again. Language teachers realized early on that the computer is really good at some things, perhaps even better than a human teacher. For example, the computer can provide students with audio and video input that can be exactly the same every time, or be varied every time by content or speed. Unlike the computer, human teachers can become impatient if they have to repeat the same thing over and over. I can personally attest to this fact. The computer is also good at prompting learners for responses, and providing consistent feedback.

In the bad old days, teachers had to write their own programs to create the input and interaction. This burdened language teachers with the job of becoming computer programmers. Relief from this roadblock came in the form of templates. These are programs that contain the software algorithms to collect and evaluate learner responses, and provide feedback. The language teachers no longer had to write programs, they could concentrate on preparing the content to enter into the system. The template fit the content into the appropriate slots in the program, and the end user received an interactive language practice exercise.

The web supports these traditional CALL functions, input and guided practice activities. Several powerful tools exist to help teachers create these programs. Hot Potatoes is a free tool, and Quia is a paid subscription service.

Example of Hot Potatoes: http://hotpot.uvic.ca/wintutor6/index.htm

Example of Quia: http://www.quia.com/mc/66200.html http://www.quia.com/cc/558086.html

These tools are easy to use. A language teacher can put a lot of content into the templates in a very short time, and create materials quite easily. However, you might have noticed that unlike some of the rich media programs that are out there, these are uni-media, namely, text-based. Second Language Acquisition research tells us that input is necessary for language acquisition, so it would be seem to be necessary to include media like sound and video into the exercises. Recent versions of Hot Potatoes and Quia finally included the ability to link to media files from the exercise page, but it requires teachers to edit the HTML source code of the web page. That puts us back to the bad old days of making teachers be computer programmers, not teachers and content experts.

A new kind of web application may offer a solution to the dilemma, as well as open the door to more re-usability, flexibility, and creative use of media resources. The approach is an example of Web 2.0 applications. The reasoning behind this approach goes something like this. Data often exists in isolation in separate places. You might have a video clip somewhere, a chunk of text somewhere else, a picture somewhere, and a text-based interactive exercise on another server. If each individual resource is identifiable by a standard formatting schema, then you should be able to combine them together in new ways. The combination of data from separate sources into a new package is called a "mashup."

Example mashups: http://mashups.clear.msu.edu/view.php?ID=MzI%3D

You can re-use existing data resources, and create your own resources using services such as YouTube. CLEAR offers a free tool called ViewPoint. ViewPoint will host your video or audio. Just upload it to the server. You can also record live audio, and video if your computer has a camera.

Resources have to be tagged with the HTML embed code. YouTube and ViewPoint provide the HTML code, which you can copy and paste into the Mashup. Hot Potatoes and Quia do not. However, CLEAR has a similar tool called SMILE, which provides the same functionality, plus the fact that it's embeddable.

3. The trap of instructional design

Language teachers have invested countless hours creating content and putting that content into templates and customized programs. The idea is that the teacher is the content expert. The content expert uses his/her expertise to create artifacts in the form of exercises, which the learners will use in order to build up their knowledge. There are practical and theoretical problems with that approach, however. You can see evidence of the practical problem in the inequality between the amount of time that teachers spend on creating the artifacts, and the amount of time that students spend using them. I have seen students finish in 10 minutes materials that took the teacher literally days to complete. Something is wrong with this picture.

On the theoretical side, the idea that the instructor can encapsulate knowledge into learning objects that then can be transplanted into the learners' brains is known as instructional design. The approach is well-documented, there are several models and rubriks for creating learning objects. Unfortunately, it is very un cost-effective. Don Bitzer, the father of the PLATO project that started in the 1960s, one of the largest computer-based instruction systems ever created, once estimated that is cost $300,000 to create each hour of instruction.

Since instructional design entered the scene, changes in learning and linguistic theories have led to new approaches to language pedagogy. It is probably safe to say that very few language teachers, especially those trained in language pedagogy, are strict adherents to instructional design. The essential problem with using instructional design with language is that language is so complex that it's impossible for the computer to anticipate every possible student response. Therefore, the computer is relegated to using objective type questions, like multiple choice, to assess students. And since assessing isn't interacting, and it isn't teaching, the ability of the computer to act as a language teacher is limited at best.

4. New vistas for language learning tools

That isn't to say that those people working in instructional design were wasting their time. On the contrary, I believe that their insights into the nature of how knowledge is organized in the brain can absolutely be applied to language learning. All that we need to save instructional design is a little modification to the approach.

The current dominant paradigm of learning theory in the field of education is constructivism. Essentially, it means that for learning to take place, learners must be active participants in the creation of their understanding. They can't be passive absorbers of information. Memorizing a dialog or vocabulary list, in other words, will not lead to language learning.

When we add technology to this approach to learning, it turns out that having learners create their own courseware is a remarkable way to facilitate learning. In fact, there are studies which indicate that creating courseware leads to better learning than using the same courseware.

The implication that I draw from this is that teachers should not be using the tools. Students should be using the tools. Students should be putting content into the templates, not teachers. Learning is an active process, and part of the process is organizing the information, and fitting it into frameworks of knowledge. The principles of instructional design offer a good model for doing just that. This is similar to the understanding of language acquisition and linguistic competence, and it holds great promise for second language teaching.