Text to Speech, on the iPad

In June 2010, after 40 years of teaching classical Indian music, Nirmala Godhwani lost her ability to speak. Four months earlier she’d been diagnosed with amyotrophic lateral sclerosis, or ALS, and while Godhwani’s mind remained sharp, her motor skills deteriorated quickly. She grew frustrated trying out the various technologies for converting typed words into speech. Most have interfaces that rely heavily on graphics, including stick figures, to help users of all ages, literacy levels, and mental capacities recognize words. The graphics slowed Nirmala down, however, and made it difficult for her to form complex sentences. Before long, her hope that she could continue having substantive conversations “had been beaten out of her by this brutal disease,” says Ajay Godhwani, Nirmala’s nephew.

Then a 35-year-old senior director at a technology consulting company, Godhwani decided, along with Nirmala’s two sons, to create something better. The result, an iPad app called Verbally, is a lesson in economy. Godhwani consulted with speech therapists and computational linguists and was surprised to learn that only “about 200 words in the English language make up about 80 percent of daily conversations,” he says. So rather than stick figures, Verbally users see text buttons on the top half of the screen and a keyboard on the bottom. One tab shows about 50 of the most common words in English; another, a list of common phrases. Users can choose one of the text buttons or start typing. The app employs predictive text technology to recommend complete words and phrases based on the first few letters typed. The idea is to squeeze more information on each screen and reduce the number of steps it takes to form a sentence.