Confederacy of a Dunce - Accessible research

Wednesday, May 03, 2006

« Serious cycling ahead | Main | Capital adventures »

One real challenge for me as a wannabe scientist is communicating my work effectively. I've had plenty of practice writing for an advanced audience (peer-reviewed publications in academic journals) but practically none intended for the general public. Until now, that is. I was recently asked to write a short piece about the sign language research I've involved in, to be published in a magazine whose primary readership is British signers. Fortunately I received plenty of feedback from one of my colleagues, a native signer who is substantially more experienced in writing for the community. It'll be interesting to see how people respond to it (if they respond at all). But I thought I'd post the article here as well, just to give a general idea of the kind of work I'm doing with sign language. All flaws, overstatements and errors of any kind are strictly my own and should not be attributed to my co-authors in any way. If you'd like to see the "academic journal" paper reporting this work please drop me a line.

Words, signs and imagery: when the language makes the difference.

In spoken languages, the way most words sound or are written has nothing to do with the actual thing each word represents. For example, neither the sound of the English word "hammer", nor its written form, have any resemblance to a hammer. Nor do the sounds of words for hammer in other languages ("kakas" in Hungarian, "martello" in Italian, "kanazuchi" in Japanese, "mlot" in Polish to mention but a few). This fact has led to a view held by many researchers: that "language" can be defined, in part, by the fact that words do not resemble the things they refer to. However, this view is based almost entirely upon research on spoken languages. Words referring to real-world things and events almost never resemble those things and events, because the languages are limited to sounds that can be produced by voice.

On the other hand, signed languagescan express far more information about the real world. In BSL, for example, numerous signs look like the object or action they refer to. For example, the sign HAMMER is produced by moving a fist up and down, resembling the typical use of a hammer, and the sign SCISSORS is produced by moving the index and middle fingers outward and inward, as if they were a pair of scissors. This sort of similarity between things in the world, and the form of signs, has often led people who know little about signed languages to the wrong conclusion that sign languages are somehow inferior to spoken languages, we know that research on sign languages shows otherwise.

We conducted an experiment designed to test whether the visual information of BSL signs affects native signers' judgements of similarity between different kinds of objects and actions, compared to English words which do not resemble things and events in the world. We used three types of signs/words: tools (e.g. SCISSORS, BROOM), tool-actions (e.g. DRILLING, DRAWING), and body-actions (e.g. SLAPPING, PUNCHING). The difference between objects (tools) and actions (tool actions and body actions) seems to be extremely important in understanding how language works, especially as they have very different roles in sentences (objects are nouns, actions are usually verbs). In BSL, all three kinds of signs are visually linked to their meanings, but tools and tool-actions are especially similar. For both of them, BSL signs look like the act of using a tool, while for body-actions, the signs look like the body movements:

Stills from BSL video clips used in the experiment. Arrows reflect direction of movement.

In this experiment, we showed BSL signers groups of three video-clips and asked them to decide which two of the signs were the most similar in meaning. We also asked English speakers to do a similar task, but using spoken words instead of video-clips. This kind of task has been used quite often to examine words' meanings in spoken languages, and speakers of a language tend to agree highly in their judgements. Because we wanted to look at the effect of the visual properties of signs, we looked at signers' and speakers' judgements of signs and words referring to tool actions. BSL signers were twice as likely to judge tool actions as more similar to tools than to body actions, while instead English speakers were four times as likely to judge tool actions as more similar to body actions. For example, SAWING and SPANNER were chosen as similar by BSL users 77% of the time, while English speakers chose them only 5% of the time; in contrast, SAWING and SCRATCHING were chosen as similar by BSL users only 5% of the time, but English speakers chose them 50% of the time.

The visual properties of signs are quite important in signed languages. Because BSL signs for both tools and tool-actions are so similar to each other, resembling the act of using a tool, BSL signers were far more likely to think of the two as very similar in meaning. This was very different from English speakers who thought of objects and actions as extremely different in meaning, even though both of them were related to tools. These results show that language cannot be defined by a lack of resemblance to things in the world, and that research on signed languages is an important part of understanding language itself as a whole.

language

Wednesday, May 03, 2006 12:21:12 PM (GMT Standard Time, UTC+00:00)