Input for the Future: Speech Recognition Software is Gaining a "Voice"

Unfortunately, few districts have been able to provide adequate time for their teachers to evaluate the new technologies they expect them to integrate into their instruction. However, there is a recently improved technology that will hopefully inspire districts to find adequate time.

Voice recognition can change the way we interact with our computers and has the potential to deeply impact teaching and learning in general. In the last two years voice recognition has significantly improved and is rapidly gaining in popularity.

As educators, we need to proactively examine this technology rather than reactively assimilating it after business and consumers have already adopted it. We need to carefully assess this technology for its potential to change learning and find the best ways to integrate it into our goals and the school environment.

Description

Voice recognition (or speech recognition) is a computer application that lets people control a computer by speaking to it. Users can tell their computer to execute commands or they can use it to transcribe what they say. (This column will focus on using voice recognition for writing.)

Presently, there are two kinds of voice recognition software available: discrete and continuous. Discrete voice recognition has been around for at least ten years and requires the user to speak...one...word...at...a...time. Continuous voice recognition, which is newer, allows the user to speak at a more or less normal rate.

My Assessment Begins

I recently bought the latest voice recognition software to assess its progress. This software's capabilities have vastly improved over the last two versions I reviewed. I first saw quadriplegics at Johns Hopkins use voice recognition 10 years ago. For them, that software was incredibly powerful. But, for people who can type, even those who hunt and peck, it seemed slow.

Three years ago an improved version of the discrete software was released. Due...to...the...fact...that...each...word...had...to...be...spoken...discreetly...and that the voice training required a lot of time, the software was still somewhat limiting. The newest versions of voice recognition are empowering to users, take very little time to initially train to their voice and accept continuous speech.

Benefits

Several benefits of the new version are almost immediately evident. First, writing with pencil and paper and even keyboard and word processor has always seemed a chore, which slowed the flow of ideas and lacked the elaborate and defining qualities of conversation. The process involved far too much time.

Now, the limiting factor is not the physical act of writing, but how quickly one can generate and compose ideas. With the new voice recognition software the challenge moves from the mechanical writing process to generating, composing and editing ideas directly. Another benefit is that correcting and processing work is far less physically taxing since ones hands are doing much less work. Carpal tunnel syndrome may well be kept at bay.

These benefits will also apply to many students and therefore improve schooling (after all, reading and writing are the two core skills used most at school).

Imagine not having to worry as much about teaching keyboarding. Imagine how the teaching of reading and writing will change when students can say a word and see it typed up in front of them. Imagine how special needs students will be freed from their difficulties and enabled to regain confidence in themselves as writers. Imagine how quickly ideas will make it from the mind to the paper. Imagine how students' writing will change as their focus shifts from getting their ideas down on paper or word processor to processing their ideas, editing their work, and sharing it with others.

Even beginning readers/writers will be able to sit down at a computer and see the words they speak transformed into written words. Students previously slowed by gross and fine motor coordination difficulties will not have to struggle with handwriting or keyboarding, but, instead, can focus on the composition processes of "writing."

It's conceivable that even early elementary students will become processors of words and ideas. It might be that they could quickly jump from a stage of relative transcription to advanced revision and editing.

Issues and Difficulties

Along with the benefits there are potential issues. One issue may be teacher and parent acceptance. Writing in this manner will seem quite different to them. In most school settings I can imagine some teachers and a small but vocal parent group condemning the changes they foresee (similar to the calculator debate).

They imagine a world in which students' handwriting and keyboarding skills atrophy. They ask, "What will our kids do when they don't have a computer with voice recognition capabilities easily accessible? Won't this be a crutch that keeps students from really learning to write?"

While these teachers and parents may be too focused on the physical processes of writing, these questions raise legitimate concerns. They are exactly the types of questions teachers must consider before the technology overruns them. Still, we must remember that this does not have to be an either or situation. Aside from the input of the text, using voice recognition for writing is not very different than writing with a pencil or word processor. It is simply a way to get words onto paper.

However, if we are not careful, I can imagine districts spending a lot of money and time on this technology with little significant change. Instead of becoming processors of words and ideas, computers with voice recognition will be used to churn out endless stories and reports full of typed speech, which have had superficial, if any, editing or revision.

Without proper instruction, students' style of language may not move from an informal, conversational style to the more formal style of writing. Teachers will need to refocus their instruction more on composing and editing ideas and not their input; no mean feat.

Another potential issue is teachers' ability to understand and teach the new "skills" needed to write this way. Teachers must help students learn the program's commands and features; how to compose ideas in their head; and how to say those ideas aloud in a manner that the software can understand (discrete or continuous enunciation). In addition, students will need to learn how to attend to all three tasks at once.

A third, related issue, will be deciding at what age students will have the cognitive and academic skills needed to benefit from this technology. To use the software available today, students would need to have some ability to sight read to train the software and have sufficient reading skills to correct misrecognized words and pick corrections from the alternative word lists. As the software improves, these needs will decrease and the appropriate age will decline.

Finally, teachers interested in having their students use this technology will need to consider what adaptations they will need to make to their teaching and the classroom environment.

Will it be feasible in a group environment? Under what conditions?

Conclusion

The technology required for quality voice recognition is here today and viable. Teachers who want to use the newest software will need to have a relatively new computer (see the individual vendors' requirements at their websites listed below). I think anyone who looks carefully at this software will find it appropriate for a range of students.

The possible changes in writing will be due to the result of good teaching of "writing" with the technology supporting the input of words rather than the technology itself. As schools begin to use this technology we are likely to first see sites that merely use it to mechanize writing. As schools gain experience with this technology and teachers reflect carefully on how it changes writing, we will begin to see more instances where teachers teach students to focus on composing and editing ideas. Their students' writing will improve and mature at a faster rate and many students previously marginalized in our educational process because of their limited ability to "write" will flourish as writers.

Teachers will need to have the time to work with this technology themselves; time to work with students writing with this technology; time to reflect on and discuss the changes, benefits, and drawbacks; and time to modify their goals and methods. Districts that don't provide this time to their teachers are more likely to merely benefit from the mechanization whereas those districts that do provide the time necessary to learn and reflect on the technology will see better writers.

Web Resource to explore for more information

Macintosh Voice Recognition Software

Windows Voice Recognition Software

  • Dragon Systems (NaturallySpeaking, DragonDictate)
  • IBM (ViaVoice)
  • Learnout and Hauspie (Voice Express) &endash; www.lhs.com/voicexpress
  • Phillips (Free Speech)


 © 2000, Jordy Whitmer

MACUL Newsletter Version - http://macul.org/newsletter/2000/march2000/sigee.html