[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: SETI public: An Alien language and Dolphins.



Marvin Minsky has covered this subject on his Web site:

http://www.media.mit.edu/~minsky/

Especially see this article:

http://www.media.mit.edu/~minsky/papers/AlienIntelligence.html

And don't forget that SETI will need biologists and sociologists as well.

Larry


At 07:51 PM 12/22/1999 +0000, Richard Burke-Ward wrote:

>>('H' has already detected the alien signal !!)    'A' does not have a
>>microphone, so 'H' cannot hear the actual voice of 'A'   It would be
>>foolish of 'A' to start any transmission with technical details about his
>>subject, so what does he do?
>
>Albert, the signal would therefore be a question of on-off binary 
>transmission, yes? What SETI assumes will happen is that the sender would 
>use the characteristics of the message as part of the starting point. 
>Frequency of transmission, say, might help define units of length. You 
>can build up a basic technical vocabulary this way. Or you could start 
>with maths - one pulse, then two, the three; and then interpolate other 
>strings which code for 'plus', 'minus', etc.
>
>Getting started is the easy bit, though. The tough bit is, what next? I 
>have tried this bottom-up approach to creating a communication, and it is 
>profoundly limited. The reason is that any 'language' created this way is 
>essentially self-referential: each word or syntactical structure is 
>defined purely by its relationship to other elements of the message sent 
>so far. There are no external reference points, so the language is 
>limited to symbolic/abstract communication. You can teach concepts like 
>true, flase, infinity, negativity, you can provide labels such as 'word' 
>or 'sentence' - but you can't get from there to a definition of an apple 
>tree. A bottom-up language can only describe ideas, by relating them to 
>observed features of the signal itself (such as frequency, or on-off 
>patterns). Essentially anything that shows a pattern (say, the periodic 
>table) can be described by creating a similar pattern within the signal 
>or within the coding system you're using. The problem is that natural 
>language, and most forms of human menaing and endeavour show no such 
>regularity, and in any case are non-universal.
>
>Receiver 'B' might recognise labels for the elements or mathematical 
>notation or Planck's constant - but it would be impossible to discern a 
>pattern which was based on moral values, social hierarchies, even 
>biuological imperatives - because these are things which are specific to 
>each culture, let alone to each species.
>
>So to get further, you need a method of *showing* what a label/word is 
>associated with. That's why Drake's message from Arecibo was a picture, 
>as was Encounter 2000's message. Plus the Pioneer probes had etched 
>plaques showing images, Voyager carried a CD full of photos. We use our 
>eyes, so we 'point' or 'show' what a word a means by using pictures. But 
>ETI may have radically different vision - or none at all.
>
>Your idea of using two people is interesting, in this context, though. 
>You could safely assume you're talking to another sighted person (well, 
>*fairly* safely: one of the things we did in the "Talking With Aliens" 
>film was give Kent Cullers from SETI Institute a replica of the Pioneer 
>plaque - he's blind and he could extract almost no information from it at 
>all - and he's HIUMAN!). So let's assume this is two sighted humans we're 
>talking about. The challenge is to find a form of representation - 
>pictures, whatever - that can convey concrete meanings to give the word 
>for 'tree' or 'diode'.
>
>And the trouble is, you can't. Because even two separate human cultures 
>can have profoundly different representational styles. Doug Vakoch at the 
>SETI Inst has examples of artwork from many cultures, all images of 
>people. Most of them were unrecognisable even when you knew what they 
>were. Pictures are a matter of convention, tailored to our cultural 
>perceptions as much as to the particular structure of our eyes. And where 
>there is nothing you can be sure will be constant across the two 
>cultures, there is no base reference point, and communication fails.
>
>Best example: there was one code the Germans never cracked in WW2 - one 
>based on Navajo mythology and language. Effectively, it could only be 
>understood by another Navajo. The meaning was in the culture, not in the 
>words. But the point is, when you are trying to communicate between 
>species, or between worlds, you cannot escape your own culture and your 
>own biology.
>
>Truly rich communication may only be possible face to face - because once 
>you share a physical context, the referentiality problem is diminished or 
>gone.
>
>>Perhaps SETI should have some linguistic specialists in its ranks??......
>
>Too right they should. The linguist couldn't begin to provide solutions, 
>but she or he could at least begin to delineate some of the challenges of 
>a post-contact world. Actually, Doug Vakoch, who's a recent addition, 
>makes a pretty impressive job of covering this territory. He's a social 
>psychologist, and has published a lot on the issue.
>
>Sorry about the extended lecture!
>
>Season's best to you all.
>
>Richard
>