An interesting new study has kicked off in Belfast, Ireland, says this news item. Under the broad aegis of the Queen’s College project called SEMAINE, SAL, the Sensitive Artificial Listener, will be designed to sense the unspoken (ie non-verbal) signals that characterise what most humans use whilst communicating. I haven’t found out just what the longer acronym means yet, but it possibly derives from ‘SEnsory MAchine Interaction Network on Emotion’, since its earlier counterpart HUMAINE, led by the same Belfast researcher, apparently means ‘HUman MAchine Interaction Network on Emotion’.
The news appears in another magazine also, curiously similar, veritably identical in fact. What does that say about human communication, eh?However, I didn’t write this only to snipe at someone’s press release, regurgitated in different magazines. I can say this with a straight face, since I have already done my sniping in the previous paragraph.
What I am concerned about are the assumptions inherent in the statement of direction for SAL, as expressed by its leader. To whit: “A basic feature of human communication is that it is coloured by emotion. When we talk to another person, the words are carried on an undercurrent of signs that show them what attracts us, what bores us and so on. The fact that computers do not currently do this is one of the main reasons why communicating with them is so unlike interacting with a human. It is also one of the reasons we can find them so frustrating.”
Take a look at the picture embedded in both the magazine items. I mean, look at the young lady – not the one smiling at the wall, that’s one of the researchers, but the rather frustrated looking GRETA. Seems a bit of a two-dimensional creature, even though she apparently lacks the other attributes of her paparazzi-attracting counterparts, or even Ananova.
Why do I find the research statement questionable? It seems very reasonable, and obvious, and I must admit, that I would have nothing but admiration to express, were this me a couple of years back. But we live and learn, apparently, notwithstanding nasty comments from the wings, stage left.
One of the things that led me to think some more about communication is this statement from an extraordinary young lady, Amanda Baggs. She is autistic, and although we haven’t met, I would guess fairly severely, in a clinical sort of way. She can’t speak, and needs a special computer interface to make words, which the computer can then speak. She lives on her own, but in an assisted community building, designed for the elderly, because she needs some help to cope with mundane human needs.
Still, the video in the link above was made by her, working on her own. It is her own communication, as much as anything humans do can be their own, and perhaps more than this blogpost is my own. She makes a very powerful statement, but one that is strikingly familiar to many people from the Indian subcontinent. Not that I feel in any way that the familiarity is exclusive to us subcontinentals, but it strikes a chord quite resoundingly.
As she expresses, her normal communication is with the world around her, by sound, touch, colour, and many other senses. She finds it hard to limit her sense of communication to the merely verbal. Even the “undercurrent of signs”, quoted above, are but a fraction of her needs, and are indeed, a different subset. Many of those signs are not useful to her, either for ‘listening’ or ‘speaking’.
What use is a mere verbal inflection, for instance, to someone who has incredibly sensitive hearing, and perfect pitch, to boot, yet cannot speak?
She takes it further: what is the use of merely communicating with humans, when she has a whole world to communicate with, to enjoy communicating with, and to need communicating with?
Can a computer be used to deliver such a sense of communication, in an interactive manner?
I don’t see why not, as long as someone is asking the question. Is anyone waiting for the answer?
Why should we want such holistic communication? Part of the reason must surely be that many – perhaps most – young children are born with this facility, but are quickly taught to limit themselves to the kinds of communications that the parents and teachers can comprehend. Don’t fidget, don’t shout, don’t dance, don’t stick your fingers in your food/that dirt (etc, etc), da da, da da, da da, and all that jazz.
Is there a better way? Probably, and I won’t be surprised if that way can be done ‘better’, in the sense that it be done for all, not just the privileged few of the ancient ‘guru-shishya’ paradigm, by designing the right computers to be our teaching aids. “I know we’ve come a long way,” sang one my latter-day idols, but he ended that with an ominous, “where do the children play?”