IN THE BUBBLE JOHN THACKARA - witz cultural

at the same time.’’ 35 Speech interfaces are difficult to design effectively at
the best of times, he says; they flourish most where they employ normal
human conversational techniques—but that’s terribly hard to achieve.
Other designers agree that natural language is so heavily dependent on
shared understanding, a shared knowledge base, and shared cultural experiences
that there is no evidence that computers will ever be able to understand
human language in the way other humans do. ‘‘So much of our
communication is based on nuance, gesture, and inflection, that although
it might be a year or two before computers can recognize our words, it
might be decades—if ever—before computers correctly infer our meaning,’’
say Winograd and Flores. 36 Even then, the simplest of tasks can go wrong.
I once visited the research labs of Sharp, in Japan, where our host boldly
stated ‘‘Open!’’ (in English) to a sliding glass door. It stayed shut, however,
until he went behind a curtain and flipped a switch. ‘‘Sorry,’’ said
my rueful host, ‘‘it was set to Japanese.’’ The upshot is that while speech
may help blind and disabled people interact with computers, it’s unlikely
to become the dominant way we connect with the planet and its vital
signs.
Pssst!
Literacy 173
Natural-language interfaces may elude us, but those based on other
kinds of sound are more promising. The advent of ubiquitous computing
has accelerated interest in sound as a medium of interaction with the
environment—human-made or otherwise. So although computers may disappear,
they are unlikely to go quietly. Research into the ‘‘sonification of
hybrid objects’’ proceeds apace. This is the use of sound to display data,
monitor systems, and provide enhanced user interfaces for computers and
virtual-reality systems. 37 Behind the beeps and squawks that emanate from
the technical devices that fill our lives lies a growing body of research into
sonification. In one pan-European project called The Sounding Object,
researchers from diverse fields, such as experimental psychology, signal
processing, human-computer interaction, and acoustics, are developing a
phenomenology and a psychophysics of ‘‘sound events’’ (which I think
means noises) that are relevant for interaction with and among artifacts. 38
Researchers at Stanford University are even using synthesized human
vowel sounds to help clinicians interpret data as they investigate tissue in