Uncovering How Humans Perceive and Verbalize Space

Kyoko Inoue (Associate Professor, Department of Foreign Languages and Liberal Arts)

The fields I study, cognitive anthropology and linguistic anthropology, have recently been gaining attention in connection with areas like natural language processing by computers and cognitive science for robotics. In fact, this type of research began out of the need for technological development. For example, when a person is told to "Walk along the wall," they can perform the action immediately. However, how many centimeters away from the wall can one be and still be considered "along" it? How many meters must one walk beside it for the action to qualify as "along"? When a robot is commanded to do something that humans take for granted, it cannot. As we delve deeper into human interaction in science and technology, we inevitably find ourselves needing to elucidate how humans perceive the external world and express it in words—in other words, to understand human information processing.

So, how do we perceive space, for example? We tend to think that we process it mostly in terms of "right" and "left," but some languages do not have this distinction. In Mayan languages, for instance, one might say something like, "This cup is on the mountain side." This method of using landmarks or cardinal directions, such as with "Yamanote" or the "East River," can be seen in English and Japanese as well. Furthermore, there is a way of framing things by likening objects and spaces to the human body, as in expressions like "a stone's throw away" (literally "the tip of the eye and nose" in Japanese) or "the spine of a book." It seems they can be broadly divided into these three types. By comparing various languages, I hope to formulate which methods are chosen in which situations in which languages, and ultimately, I aim to apply these findings to the recognition and information processing of computers and robots.