raza@cs.hw.ac.uk (raz(a)) (03/02/90)
Speech synthesis systems are already with us and speech recognition is improving all the time, within the next 100 years we can expect to see the arrival of machines whose main communication medium is speech (or iconic for the hearing impaired). Compare this with the advent of the pocket calculator, this was heralded as the first step towards an innumerate society. What is the likelyhood of these developments leading to a society that is illiterate ? is it a bad thing ? will this mean a reliance of our society on this kind of technology ? Raza Hussain janet : raza@uk.ac.hw.cs computer science, arpanet : raza@cs.hw.ac.uk heriot watt uni., Edinburgh uucp : ..ukc!cs.hw.ac.uk!raza
marie@ernie.Berkeley.EDU (Marie desJardins) (03/03/90)
In article <808@odin.cs.hw.ac.uk> raza@cs.hw.ac.uk (raz(a)) writes: >Speech synthesis systems are already with us and speech recognition is >improving all the time... Compare this with the advent of the >pocket calculator, this was heralded as the first step towards an >innumerate society. What is the likelyhood of these developments >leading to a society that is illiterate ? You raise an interesting question. But I don't believe widespread illiteracy will be a problem (at least not as a result of speech synthesis) for several reasons: - Consider, as you mention, the invention of the pocket calculator. Yes, in a way, we have come to depend on the calculator to do even relatively simple calculations. On the other hand, the reliability is significantly higher than when all calculations had to be performed by hand, ease and speed of performing calculations has been greatly increased, and most importantly, most people can still add and subtract (notwithstanding McDonald's workers who can't make change until the computer tells them how much to return), although they probably can't rattle off multiplication tables as readily as schoolchildren could in the early 1900s. (Of course, I never did learn my multiplication tables, so I don't think this is a great loss. :-) - Consider also the invention of the printing press. Prior to the widespread availability of printed texts, the oral tradition was much stronger than it is today. Educated people were expected to memorize literature, poetry, famous speeches, and so forth. That skill was simply supplanted by other skills when it became superfluous. (Although I must admit I am still amazed by the idea that somebody could memorize the entire Iliad!) - Finally, I really can't imagine speech synthesis ever completely replacing the written word. There are certain freedoms the written word provides that synthesized speech can't: you can quickly scan material, you can read at your own pace, pausing whenever you feel like it without requesting that the system stop, and so forth. Here's a question, then: can anybody else envision a communication method that could completely eradicate the written word? Even a system that uses graphics and speech synthesis liberally wouldn't (in my mind) have the ability to transmit information simply and conveniently that written text have. Here's another question (although I'm getting away from the results of AI technology, so I guess this is now officially in the wrong group): how many people think that widespread use of calculators has actually led to "innumeracy"? [To forestall responses of the McDonald's server sort as raised above, I would argue that those people probably wouldn't have had very good arithmetic skills anyway, and the real effect of the computer-cash-register in that case is to make it easier for people with poor arithmetic skills to work in McDonald's.] Marie desJardins marie@ernie.berkey.edu
dmocsny@uceng.UC.EDU (daniel mocsny) (03/04/90)
In article <34666@ucbvax.BERKELEY.EDU> marie@ernie.Berkeley.EDU.UUCP (Marie desJardins) writes: > - Consider, as you mention, the invention of the pocket > calculator. Yes, in a way, we have come to depend on > the calculator to do even relatively simple calculations. [...] > - Consider also the invention of the printing press. Prior > to the widespread availability of printed texts [...] Educated > people were expected to memorize literature, poetry, > famous speeches, and so forth. That skill was simply > supplanted by other skills when it became superfluous. Until we find some way to increase human intelligence, the capacity of a given person can think is essentially fixed. Since intelligence is a necessary ingredient in the individual's ability to generate wealth, the only way the individual can generate more wealth with this finite resource is to progressively engineer away the need to think. People can do this in four main ways: 1. Build machines that can replace some of the mechanical aspects of thinking, and relegate mental tasks to them. 2. Engineer society to support wealth creation with less thought. For example, consider transportation technology. The problem of negotiating a rough trail requires much more thought than the problem of traversing a smoothly paved road. The greater thought required to grade and pave the road pays off in the thought saved during its repeated use. The next step up from the road is the railroad, upon which large vehicles may be safely directed with least thought of all. Standards are another essential aspect of engineering society to reduce thought requirements. A little reflection shows this to be obvious. E.g., consider how much extra thought you must waste when you move repeatedly back and forth between several computer terminals with spuriously variable keyboard layouts. The computer industry, of course, is today imposing thought requirements on society that are vastly in excess of what is necessary. In other words, horizontal fragmentation in the computer industry is destroying wealth on the scale of a large natural or artificial disaster. 3. Multiply the consequences of thought by scaling up industrial processes. For example, in chemical engineering, you can design a 1000 ton/day process with the same labor (i.e., thought) requirement as a 100 ton/day process. It's a little bit harder than simply multiplying everything on the blueprint by a capacity of 10, but not much harder. If you build a 10 m^2 mixing tank, you must hire a worker to watch it. The same worker can as well watch a 100 m^2 tank. 4. Fragment thought requirements into narrow domains and permit each person to specialize in just one or few. People think much more effectively after long practice. Since no one person can yet master all the skills necessary to maintain a complex technological society, we have no choice but to divide labor. (This relates to point 3; each specialist must multiply her/his effectiveness by a large enough factor to make up for the thousands of other people who do not have her/his skills.) >Here's a question, then: can anybody else envision a communication >method that could completely eradicate the written word? Of course. You already satisfy *many* or your information/communication requirements with such a system. How does your motor cortex communicate with your musculo-skeletal system (muscles, motor neurons, proprioceptors, etc.) to allow you to walk across the room without falling down? Hint: it isn't by composing written reports and minutes and shipping them off to its "branch offices." Obviously, however, your question pertains to the difficult task of coupling autonomous intelligent entities into an organic, symbiotic whole, when they have no built-in organic links. I suppose we should look to the human nervous system for guidance here. No matter how efficient you may be at generating, distributing, and assimilating text, that can never be as good as *just* *knowing*. Dan Mocsny dmocsny@uceng.uc.edu
aipdc@castle.ed.ac.uk (Paul D. Crowley) (03/04/90)
Computer interfaces should see to use the fastest channel available for communication. For input, this means speech. For output, it still means writing. I think the twenty-six will be around for a long while yet. -- \/ o\ "I say we grease this rat-fuck son-of-a-bitch Paul D Crowley /\__/ right now. No offense." - Aliens. aipdc@uk.ac.ed.castle
lws@comm.WANG.COM (Lyle Seaman) (03/06/90)
raza@cs.hw.ac.uk (raz(a)) writes: >Speech synthesis systems are already with us and speech recognition is >improving all the time, within the next 100 years we can expect to see >the arrival of machines whose main communication medium is speech (or >iconic for the hearing impaired). Compare this with the advent of the >pocket calculator, this was heralded as the first step towards an >innumerate society. >What is the likelyhood of these developments leading to a society that >is illiterate ? is it a bad thing ? will this mean a reliance of our >society on this kind of technology ? If you say "mostly illiterate", might be. But look at it this way. ** Why do we consider literacy a good thing? ** Well, it allows one-to-many information transfer, asynchronously, which is crucial for communication, education, modern business, etc. But is literacy *intrinsically* good? I think not. If another technology comes along and replaces written language, then it will only be because it is *more useful* than written language. And yes, we may become reliant on this kind of technology, but we are already reliant on the telephone which has replaced the role of written language considerably. So what's wrong with reliance on a technology? A: Perhaps someday the technology may become unavailable. What would we do then? Well, what we would we do if telephones suddenly disappeared (unlikely)? I imagine we'd write a lot more letters. And our society _would_ be different. So I don't believe that society AS WE NOW KNOW IT will rely on such a technology, I think that the society produced as a result of the technology would unavoidably depend on it. On the other hand, it is doubtful that speech technology alone could replace written language. The bandwidth of the visual channel is so much greater than the bandwidth of the aural channel. Yes, spoken language does convey certain nuances that written language may fail to convey (sarcasm, for instance). But sarcasm is not usually an essential feature of physics texts. Being able to control the rate of assimilation is very important to understanding information-dense subjects, being able to back up and re-read a phrase or a paragraph, easily and quickly. So in summary: No, I don't think speech technology alone will cause wide-spread illiteracy. But if some technology did do so, I don't think we would suffer by it. Lyle. lws%comm.wang.com@uunet.uu.net
dmocsny@uceng.UC.EDU (daniel mocsny) (03/06/90)
In article <1990Mar5.173358.25523@comm.WANG.COM> lws@comm.WANG.COM (Lyle Seaman) writes: >Well, what we would we do if telephones suddenly >disappeared (unlikely)? I imagine we'd write a lot more letters. And >our society _would_ be different. So I don't believe that society AS It's not unlikely at all, at least locally and temporarily. The telephone network has many critical exchanges whose failure can paralyze communications for tens of thousands to hundreds of thousands of customers. Thanks to the general reliability of the phone system, these failures are rare. But they do occur, and the well-known result is bedlam and severe economic losses. If telephones suddenly disappeared, we would not simply "write a lot more letters." Rather, our entire financial and industrial systems would collapse within days if not hours, authorities would be unable to maintain order, and starvation would set into large cities within weeks. >WE NOW KNOW IT will rely on such a technology, I think that the society >produced as a result of the technology would unavoidably depend on it. Well yes, that is exactly true. As long as human population continues to increase, and social institutions become ever more complex, we need more technological inputs to keep the show running. Once we incorporate a technology, we can't simply pull the plug on it. We've already burned the bridge behind us. Our only choice is to press on, replacing each old technology with more effective new technologies. Or else allow 90% of the population to die. >On the other hand, it is doubtful that speech technology alone could >replace written language. The bandwidth of the visual channel is so >much greater than the bandwidth of the aural channel. Yes, spoken >language does convey certain nuances that written language may fail to >convey (sarcasm, for instance). But sarcasm is not usually an essential >feature of physics texts. Being able to control the rate of assimilation >is very important to understanding information-dense subjects, being able >to back up and re-read a phrase or a paragraph, easily and quickly. I find that I experience rapid fatigue listening to a technical discussion, whereas I can read the same level of material for hours if necessary. I suspect most of the problem stems from being unable to control the rate or content of the incoming data. Also, the spoken word is inherently serial data, whereas text + graphics can take on a parallel structure. For example, by laying out many text blocks on a desk or workstation screen, the information worker can establish a spatial framework for organizing and digesting complex information. Since humans are good at spatial reasoning, apparently the ability to associate glancing in a given location with obtaining a certain fact is a great adjunct to learning. Consider how transparent the menu structure of a familiar program can become. You can formulate "questions" for the machine simply by a mechanical movement or gesture. This may be easier than abstracting those questions into verbal queries. For a speech-based interface to be generally useful, it would have to contain at least some of the abilities of the human expert. The human expert is able to listen to the scrambled queries from the human client, deduce what the client is *really* asking, and then tell the client. But I don't know how the speech-based interface would derive an advantage over an equally smart graphics interface. Of course, in field and real-time applications, speech-based interfaces would be very useful, for the worker who doesn't have any free hands, must keep visual attention locked on real-world objects, or must receive alarm messages quickly. Dan Mocsny dmocsny@uceng.uc.edu
smoliar@vaxa.isi.edu (Stephen Smoliar) (03/06/90)
In article <1990Mar5.173358.25523@comm.WANG.COM> lws@comm.WANG.COM (Lyle Seaman) writes: >raza@cs.hw.ac.uk (raz(a)) writes: > >>Speech synthesis systems are already with us and speech recognition is >>improving all the time, within the next 100 years we can expect to see >>the arrival of machines whose main communication medium is speech (or >>iconic for the hearing impaired). Compare this with the advent of the >>pocket calculator, this was heralded as the first step towards an >>innumerate society. > >>What is the likelyhood of these developments leading to a society that >>is illiterate ? is it a bad thing ? will this mean a reliance of our >>society on this kind of technology ? > >If you say "mostly illiterate", might be. But look at it this way. > ** Why do we consider literacy a good thing? ** >Well, it allows one-to-many information transfer, asynchronously, which is >crucial for communication, education, modern business, etc. But is literacy >*intrinsically* good? I think not. If another technology comes along and >replaces written language, then it will only be because it is *more useful* >than written language. This is all very well and good as it stands, but I think it misses the cautionary point which was initially raised. The two advantages that written communication has over verbal communication is that it is NOT REAL TIME and it is basically RANDOM ACCESS. In other words, I can take as much time as I want with a piece of printed text and read it in any order which serves my attempts to extract information from it. Yes, I can record the spoken word; but I shall never have the flexibility of access I have with printed text. Now I am not yet sure of this, but I would be willing to try to mount an argument that written material is more conducive to thought than spoken material. When I am dealing with spoken text, I take it as in comes; and then it goes. My degree of retention is very much a function of the skill of the speaker, enhanced by my own powers of attention. If the material is written and I miss something, I can always go back and read it again. I think the danger we must confront is that we may become too enamored of communication which becomes more and more real-time. This removes us from situations in which we can ponder at our leisure, exercising those "little grey cells" to find the message in the text. We work with written material in learning physics because learning physics is hard work; we cannot simply absorb it from the right kind of real-time presentation. If communication through speech-processing systems discourages our ability to work with printed texts, then it may ultimately erode our ability to think for ourselves. ========================================================================= USPS: Stephen Smoliar USC Information Sciences Institute 4676 Admiralty Way Suite 1001 Marina del Rey, California 90292-6695 Internet: smoliar@vaxa.isi.edu "Only a schoolteacher innocent of how literature is made could have written such a line."--Gore Vidal
gilham@csl.sri.com (Fred Gilham) (03/07/90)
Lyle Seaman (I think) writes:
If you say "mostly illiterate", might be. But look at it this way.
** Why do we consider literacy a good thing? **
Well, it allows one-to-many information transfer, asynchronously,
which is crucial for communication, education, modern business, etc.
But is literacy *intrinsically* good? I think not. If another
technology comes along and replaces written language, then it will
only be because it is *more useful* than written language.
====================
The thing that concerns me about this statement is that it doesn't
address the question: More useful to whom? Taking television as an
example, the question has two answers: Television is useful to
advertisers and to governments (the former mostly in the USA, the
latter in many other countries). Television (in its current form) has
definite disadvantages as a communications medium (limited access,
passive reception, a strong tendency to trivialize everything it
touches), yet it is the most popular medium in the U.S. apart from
speech. (In some families it even supplants that....)
Television has broken many promises. It was supposed to educate; in
the U.S. it undermines education. It was supposed to create a sort of
global village that everyone can participate in; it has homogenized
our culture without increasing our ability to affect one another as
individuals. By some sort of process that nobody forsaw, its prime
purpose is now to attract viewers. Anything that does not perform
sufficiently well in meeting that demand falls by the wayside.
There is no guarantee that any change, technological or otherwise,
will be an ``advance.''
--
Fred Gilham gilham@csl.sri.com
"The culture of any period is a mixture of that which docilely caters
to passing whims and fancies and that which transcends these things --
and may also pass judgement on them." -- Stanislaw Lem