1. Natural language
- Posted by "Wallace B. Riley" <wryly at MINDSPRING.COM> Mar 28, 1998
- 881 views
Hello all - On this business of coding in 'natural language', has anybody thought of Esperanto? I am not an expert in that language or any language other than English. But I had the experience of participating in a performance of a choral work with the text in Esperanto a few years ago, and was surprised to note the number of words, the syntactical structure, and other characteristics that were similar to English and a few other languages to which I have been exposed. Again, I know a few words in German, and a few words in Spanish, and a *very* few words in French, but I can't carry on a conversation in any of them. I've tried. I even have problems in Australian English and 'English' English (as opposed to American English). As an amateur singer I have been exposed to Hebrew, Russian, Chinese(Mandarin) and maybe one or two others. I don't 'know' any of those languages but I could recognize some of the structural similarities in Esperanto. Somebody start a movement to program in Esperanto. You might trigger a landslide. Wally Riley wryly at mindspring.com
2. Re: Natural language
- Posted by Jacques Guy <j.guy at TRL.TELSTRA.COM.AU> Mar 29, 1998
- 923 views
Wallace B. Riley wrote: > On this business of coding in 'natural language', has anybody thought of > Esperanto? Yes, Esperanto has been used as the intermediate language in efforts at machine translation. Others have used Aymara (a South American language) because its grammar is entirely regular (far more regular in fact, than Esperanto) > I even have problems in Australian English and 'English' > English (as opposed to American English). That is perfectly normal. When I migrated to Australia I spoke perfect English, and understood spoken "English" English without any problem. But I couldn't understand a word of Australian English (the broader variety, I mean). A fellow teacher, who was Canadian, told me the same. It was so bad for him that he hadn't realized that the announcements at Sydney airport were made in English! > Somebody start a movement to program in Esperanto. You might trigger a > landslide. Ahem! A landslide it will be: you'll be buried in it. No, Esperanto is no good for programming. No human language is good for programming. A programming language is not at all a "language" in the sense that English, Hebrew, Esperanto, Chinese, Indonesian, are languages. I know: I taught French, Chinese, Latin, and ...programming in Simula 67!
3. Re: Natural language
- Posted by Michael Sabal <mjs at OSA.ATT.NE.JP> Mar 29, 1998
- 872 views
------ =_NextPart_000_01BD5B42.F41AAD00 Content-Transfer-Encoding: quoted-printable ---------- From: Wallace B. Riley[SMTP:wryly at MINDSPRING.COM] Sent: Sunday, 29 March, 1998 7:33 AM To: Multiple recipients of list EUPHORIA Subject: Natural language Hello all - On this business of coding in 'natural language', has anybody thought of Esperanto? I am not an expert in that language or any language other than English. = But I had the experience of participating in a performance of a choral work = with the text in Esperanto a few years ago, and was surprised to note the = number of words, the syntactical structure, and other characteristics that were similar to English and a few other languages to which I have been = exposed. Again, I know a few words in German, and a few words in Spanish, and a *very* few words in French, but I can't carry on a conversation in any = of them. I've tried. I even have problems in Australian English and = 'English' English (as opposed to American English). As an amateur singer I have = been exposed to Hebrew, Russian, Chinese(Mandarin) and maybe one or two = others. I don't 'know' any of those languages but I could recognize some of the structural similarities in Esperanto. Somebody start a movement to program in Esperanto. You might trigger a landslide. Wally Riley wryly at mindspring.com ------------- While my native language is English (and hence that would be the first = phase of the interface), I've lived in Japan for 2years teaching = English as a foreign language (and with some of my students, I think the = computer could learn faster with or without neural nets:)). I've = studied Spanish, French, Russian, and Japanese; and have sung in German, = Latin, and a few other languages I've never heard of. If you have the = syntax of Esperanto (SVO or SOV, etc.) and some kind of vocabulary list, = once my skeleton translator is finished for English, changing it to = another human language would be *very* easy. As far as other recent comments go, one of my intentions is to build = a web crawler to allow the program to build a knowledge base on its own. = A web crawler would also be required to answer such questions as = "What's tomorrow's weather?" and "How much does it cost to fly to St. = Petersburg? (Florida or Russia?)" However, I don't have much ability in = that kind of programming yet. Also, I posit that a computer does have a language. The machine = language (accessible through Euphoria) is the computer's native = language. It's not a human language, just like cat's and dog's don't = have human language. But it is a form of communication, which is the = base that's needed. I've read the research that's being done with = neural networks and teaching language to computers like it's taught to = infants. It's a valid idea and will probably work, but it wouldn't be = practical to implement on a large scale for a long time. Until then, I = believe I have a method that could work sooner. My "intermediate" = language, would act more like a native language for the computer. Once = built, the only thing required is a series of translation programs. = But....I'm too lazy and impatient, so I'm piecing together both at the = same time. The drawback: the scope of the intermediate language will be = slow in coming. It'll be a while before it will respond "Dammit Jim, = I'm a computer, not a painter!" :) BTW: If there are any Euphoria programmers in upstate New York, let me = know. I'll be going home sometime this next year. Mike Sabal mjs at osa.att.ne.jp http://home.att.ne.jp/gold/mjs/ ------ =_NextPart_000_01BD5B42.F41AAD00
4. Re: Natural language
- Posted by "Christopher K. Lester" <cklester at FLASH.NET> Mar 29, 1998
- 828 views
- Last edited Mar 30, 1998
------=_NextPart_000_006B_01BD5B59.261B9240 charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Regarding natural language searches, or a web crawler with the ability = to do its own searches, see www.askjeeves.com. They allow natural = language requests. Secondly, a human has language because it intends to communicate. = Computers have no such intention. I doubt they ever will. It's probably all semantics, but I think the better efforts would be = toward expert systems. Let's focus on having a bot that will search the = internet for information on certain items, then index those items. THAT = would be useful. Maybe askjeeves.com is doing this already...? I saw a television program on the Discovery channel the other day (or = somesuch channel) that featured a computer from Intel that could do 3 = billion calculations per second. It filled a small room. Extremely = powerful. Then they commented that it takes the computer a few moments = to emulate the function of ONE BRAIN CELL. Put together millions = (billions?) of those computers, and you'll have yourself a computer that = can emulate the brain. A computer with intelligence? Won't happen in our lifetime. Mike Sabal said some things, some of which follows: =20 As far as other recent comments go, one of my intentions is to = build a web crawler to allow the program to build a knowledge base on = its own. A web crawler would also be required to answer such questions = as "What's tomorrow's weather?" and "How much does it cost to fly to St. = Petersburg? (Florida or Russia?)" However, I don't have much ability in = that kind of programming yet. Also, I posit that a computer does have a language. The machine = language (accessible through Euphoria) is the computer's native = language. It's not a human language, just like cat's and dog's don't = have human language. But it is a form of communication, which is the = base that's needed. I've read the research that's being done with = neural networks and teaching language to computers like it's taught to = infants. It's a valid idea and will probably work, but it wouldn't be = practical to implement on a large scale for a long time. Until then, I = believe I have a method that could work sooner. My "intermediate" = language, would act more like a native language for the computer. Once = built, the only thing required is a series of translation programs. = But....I'm too lazy and impatient, so I'm piecing together both at the = same time. The drawback: the scope of the intermediate language will be = slow in coming. It'll be a while before it will respond "Dammit Jim, = I'm a computer, not a painter!" :) =20 Mike Sabal mjs at osa.att.ne.jp http://home.att.ne.jp/gold/mjs/ =20 =20 ------=_NextPart_000_006B_01BD5B59.261B9240 charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <!DOCTYPE HTML PUBLIC "-//W3C//DTD W3 HTML//EN"> <HTML> <HEAD> <META content=3Dtext/html;charset=3Diso-8859-1 = http-equiv=3DContent-Type> <META content=3D'"MSHTML 4.71.1712.3"' name=3DGENERATOR> </HEAD> <BODY bgColor=3D#c8e0d8> <DIV><FONT color=3D#000000 size=3D2>Regarding natural language searches, = or a web=20 crawler with the ability to do its own searches, see <A=20 href=3D"http://www.askjeeves.com">www.askjeeves.com.</A> They allow = natural=20 language requests.</FONT></DIV> <DIV><FONT color=3D#000000 size=3D2></FONT> </DIV> <DIV><FONT size=3D2>Secondly, a human has language because it intends to = communicate. Computers have no such intention. I doubt they ever=20 will.</FONT></DIV> <DIV><FONT size=3D2></FONT> </DIV> <DIV><FONT size=3D2>It's probably all semantics, but I think the better = efforts=20 would be toward expert systems. Let's focus on having a bot that will = search the=20 internet for information on certain items, then index those items. THAT = would be=20 useful. Maybe askjeeves.com is doing this already...?</FONT></DIV> <DIV><FONT size=3D2></FONT> </DIV> <DIV><FONT size=3D2>I saw a television program on the Discovery channel = the other=20 day (or somesuch channel) that featured a computer from Intel that could = do 3=20 billion calculations per second. It filled a small room. Extremely = powerful.=20 Then they commented that it takes the computer a few moments to emulate = the=20 function of ONE BRAIN CELL. Put together millions (billions?) of those=20 computers, and you'll have yourself a computer that can emulate the=20 brain.</FONT></DIV> <DIV><FONT size=3D2></FONT> </DIV> <DIV><FONT size=3D2>A computer with intelligence? Won't happen in our=20 lifetime.</FONT></DIV> <BLOCKQUOTE=20 style=3D"BORDER-LEFT: #000000 solid 2px; MARGIN-LEFT: 5px; PADDING-LEFT: = 5px">Mike=20 Sabal said some things, some of which follows:</BLOCKQUOTE> <BLOCKQUOTE=20 style=3D"BORDER-LEFT: #000000 solid 2px; MARGIN-LEFT: 5px; PADDING-LEFT: = 5px"><BR> =20 As far as other recent comments go, one of my intentions is to = build a=20 web crawler to allow the program to build a knowledge base on its = own. =20 A web crawler would also be required to answer such questions as=20 "What's tomorrow's weather?" and "How much does it = cost to=20 fly to St. Petersburg? (Florida or Russia?)" However, I = don't=20 have much ability in that kind of programming yet.<BR> Also, I = posit=20 that a computer does have a language. The machine language = (accessible=20 through Euphoria) is the computer's native language. It's not = a human=20 language, just like cat's and dog's don't have human language. = But it=20 is a form of communication, which is the base that's needed. = I've read=20 the research that's being done with neural networks and teaching = language to=20 computers like it's taught to infants. It's a valid idea and = will=20 probably work, but it wouldn't be practical to implement on a large = scale=20 for a long time. Until then, I believe I have a method that = could work=20 sooner. My "intermediate" language, would act more = like a=20 native language for the computer. Once built, the only thing = required=20 is a series of translation programs. But....I'm too lazy and=20 impatient, so I'm piecing together both at the same time. The=20 drawback: the scope of the intermediate language will be slow in=20 coming. It'll be a while before it will respond "Dammit = Jim, I'm=20 a computer, not a painter!" :)<BR><BR>Mike Sabal<BR><A=20 href=3D"mailto:mjs at osa.att.ne.jp">mjs at osa.att.ne.jp</A><BR><A=20 = ------=_NextPart_000_006B_01BD5B59.261B9240--
5. Re: Natural language
- Posted by Jacques Guy <j.guy at TRL.TELSTRA.COM.AU> Mar 30, 1998
- 828 views
Christopher K. Lester wrote: > Secondly, a human has language because it intends to communicate. > Computers have no such intention. Bull's eye! That is the fundamental difference. > Let's focus on having a bot that will search > the internet for information on certain items, then index those items. > THAT would be useful. That is far, far from a trivial task. I don't mean indexing and all that, but classifying the data so that it is useful. For instance, AltaVista is well nigh useless: it returns far too much garbage. The main problem, in my view, is the lack of proper clustering algorithms. As far as I know, they all assume that a point can belong only to one cluster (i.e., here, a Web document to one category). I was toying with a model that would allow a point to belong to several clusters when I got side-tracked.
6. Re: Natural language
- Posted by Irv Mullins <mountains at MINDSPRING.COM> Mar 30, 1998
- 839 views
At 02:19 PM 3/30/98 -0800, Jacques Guy wrote: >Christopher K. Lester wrote: > >> Secondly, a human has language because it intends to communicate. Except for politicians, of course, who have language for the exact opposite reason.* >> Computers have no such intention. Just think, people used to worry about _computers_ taking over the world! How silly of us. *note: some would argue that politicians are not human, since they are, after all, descended from lawyers.... Irv
7. Re: Natural language
- Posted by Kasey <kaeyb at GEOCITIES.COM> Mar 30, 1998
- 861 views
Christopher K. Lester wrote: > <snip> > I saw a television program on the Discovery channel the other day (or > somesuch channel) that featured a computer from Intel that could do 3 > billion calculations per second. It filled a small room. Extremely > powerful. Then they commented that it takes the computer a few moments > to emulate the function of ONE BRAIN CELL. Put together millions > (billions?) of those computers, and you'll have yourself a computer > that can emulate the brain. <snip> > Mike Sabal > mjs at osa.att.ne.jp > http://home.att.ne.jp/gold/mjs/ > > It was probably emulatating the entire funtion including chemical and electrical process, not just the function. my computer can easily run a multi-"cell" neural network and it's only a amd586-133 (slightly slower than a p90). Besides who said you have to mimick the human brain to get inteligence? KAsey
8. Re: Natural language
- Posted by Daniel Berstein <danielberstein at USA.NET> Mar 30, 1998
- 855 views
My thought about "Natural language programming": What use would it have a natural language propramming interface? Have you ever tried to explain a complex algorithm to someone without falling into control-structure definition, or data types? I just can't in plain Spanish (or english or whatever). We see it everyday on this listserver... each time someone needs help, we reply with of bunch of CODE... not natural language description/solution. In the case someone actually does a natural-language compiler, who defines the language? Doesn't language have social, ever evolving nature? Is it admisable to change your code after a new edition of Webster's dictionary? The world has spent more than half a century developing better interfaces with digital machines. The best we've done: high-level programming languages. What does it mean? We replace meaningless nemonics with plain english words (if..then, for..next,..). We've also developed a lot of methodologies (OO, structured, etc..) for helping *us* do the translation. Maybe in this field is the key for a more human-natural approach with the machine. I do see a usefulness in natural languages on computers... end user interfaces (voice recognition, OCR, etc..), not programming.