The Eyes Tell a Lot

Public Relations Office        March 21, 2013

Takeshi Kawabata, Professor at the School of Science and Technology

Profile of Takeshi Kawabata

I’ve seen something nice in a shop window, and I want to buy it. But where’s a sales clerk? There’s one, but he seems busy. I’ll keep looking at him. Oh, he’s noticed me. “Can I help you?” he says, with a big smile.

When I’m waiting for someone in a public place, I don’t stare at the people around me. If my eyes happen to meet theirs, I look away naturally and casually, trying to convey the nuance of “Sorry. Never mind.”

Why do people subconsciously (or consciously) react to the eyes of others? It’s partly because the eyes convey a message regarding whether or not you’d like to continue your face-to-face communication. If you talk with someone while gazing steadily at him or her, it indicates that you’d like to continue the conversation. On the other hand, if you don’t maintain eye contact, it implies that you’re not interested in the conversation.

Needless to say, speech, or verbal information, plays an important role in making oneself understood in social life. However, body movements, hand gestures, facial expressions etc. actually play an even more important role in communication. The information provided through these nonverbal communication channels is called “paralanguage information.”

While speech conveys the basic meaning of what you’d like to say, nonverbal information adds more complex aspects, such as nuance and feelings, to the basic meaning. People learn these functions through many conversations. The gaze is one of the nonverbal communication channels. Actually, it is often said that “the eyes tell a lot.”

I conduct research on computers that can verbally communicate with human beings in a natural manner. I work on verbal communication technologies, such as speech recognition technology that enables a character in a computer software program, or a robot, to understand what’s being said by human beings, and speech synthesis technology that enables a robot and a computer to speak a human language. However, I believe that the key to my research resides in how we deal with nonverbal communication elements.

Let me explain one character that I’m developing now. She first looks around, making the observer feel that she’s thinking about something. When you say to her “Excuse me,” she quickly faces the front and begins to look at you, indicating that she’s ready to listen to you. In subsequent conversation with you, she nods at appropriate times, her gaze moving up and down, implying that she is following what you are saying. It might be good to arrange it so that, if she can’t hear you because of some noise, she tilts her head and says “Pardon?” The integration of nonverbal communication elements would make the conversation more active and rhythmical.

From this perspective, I think that it would be better to include paralanguage information in e-mail, a typical form of text communication. Maybe the use of pictographs in e-mail could indicate that you’re trying to express your “gaze.”

Profile of Takeshi Kawabata

Completed graduate program at the Department of Electrical and Communication Engineering, School of Engineering, Tohoku University. Doctor of Engineering. Worked at the Musashino Electrical Communication Laboratory of Nippon Telegraph and Telephone Public Corporation, ATR Interpreting Telephony Research Laboratories, Carnegie Mellon University (as a visiting researcher), NTT Basic Research Laboratories, and other institutes. In 2003 began work as Professor at the Department of Informatics of School of Science and Technology of Kwansei Gakuin University. In 2009, began work as Professor at the Department of Human System Interaction.
Prof. Kawabata conducts research on computers that can engage in natural conversations with human beings, exploring a new signal-processing paradigm integrating the complex network theory.