Can Computers Think? The History and Status of the Debate - Map 1 of 7
问题9
Issue Area: Can computers be creative?
问题域: 计算机会不会有创造力?
The link to the part of the map this discussion is about: http://www.macrovu.com/CCTWeb/CCT1/CCTMap1BeCreative.html
1. Common start point by Alan Turing (omitted)
97. (Disputing 1) Computers can never be creative. Computers only do what they are programmed to do; they have no originality or creative powers.
Note: similar debates play out in the "Free Will" arguments.
(反驳1) 计算机永远不可能有创造力。计算机只能做它们被要求做的;他们不具有原创性和创造力。
98. (Supporting 97) Anticipated by Alan Turing, 1950 Machine can never take us by surprise. Machines are entirely predictable in their behavior. Because they never do anything new, they can never surprise us.
"That's not surprising at all ..."
(支持97) 由阿兰·图灵1950年预见,机器永远不会使我们吃惊。机器的行为是完全可预测的。因为他们从不做新的事情,他们不会让我们吃惊。
99. (Disputing 98) Alan Turing 1950 Computers are not entirely predictable.The belief that computers are entirely predictable arises from the false assumption (widespread in philosophy and in mathematics) that humans can know everything that follows deductively from a set of premises. But humans learn new things in part through the working out of deductive consequences. Similarly, humans don't know everything a computer will do given some initial state of the computer; we learn new things in part by watching them performing their calculations. ( I like this one, and I guess this intrinsically has some links to such mathematical rules as goedel theorem, although it's on human's side -- translator )
(反驳98)阿兰·图灵1950 计算机不是完全可预测的。认为计算机完全可预测的想法来自于认为人类能够天然地知道一切有一组假设演绎出的结果的假想(这在哲学和数学中非常普遍)。但是人类有时能从这个演绎过程本身学到新知。类似地,人类不能从计算机的初始状态了解到所有这个计算机的后续运行状况;我们通过看他们执行这些计算也学到了新的东西。
100. (Disputing 98) Alan Turing, 1950 Machines frequently take us by surprise. Computers users and even experts are often surprised by the things that computers do.
"Wow! What a surprise!"
(反驳98)阿兰·图灵,1950 计算机时常能让我们吃惊。计算机使用者,甚至专家经常对计算机的所作所为感到吃惊。
101. (Disputing 100) Anticipated by Alan Turing, 1950 Surprise is a result of human creativity. Even if we are surprised by what a machine does, that reaction does not mean that the machine has done anything original or creative. It just means that the human made a creative prediction about what the computer would do, and was then surprised when the computer acted differently. (Or can we say, even if human is surprised at the outcome different than she expected, it is human that makes use of computer to do creative things?Though this argument sounds a bit too human centred. Or probably it's just a matter that creativity almost always means multiple answers, and human would be surprised regardless as long as the answer is not expected. -- translator)
(反驳100)阿兰·图灵预见,1950 惊奇是人类创造力的结果。即使我们对机器的行为感到吃惊,这个反应不代表机器做了任何有原创性的事情。那只是人对计算机可能做的事情做了一个具有创造力的预测,但对计算机的不同表现感到吃惊。
102. (Disputing 101) Alan Turing, 1950 The argument from human creativity applies to any case of surprise.You could always say that being surprised came from you, the interpreter, rather than from anything original on the other person's or machine's part. For example, if a human surprises you with a joke, then you could argue that the surprise was a result of your interpretation of the joke rather than anything creative on the joke teller's part. (This argument is neither so convincing nor justifiable on its own, as surprise is at least even from this point of view an interaction / mutual process that takes place on both sides; though it is obviously true that without right interpretation the the surprise can never happen -- translator)
(反驳101)阿兰·图灵,1950 人类创造力观点适用于任何惊讶的情形。你总能说惊奇的感觉来自于你本人——解释这个现象的人,而非任何起因于他人或机器的原创性事物。例如,如果一个人说了个笑话让你感到吃惊,然后你可以说这个惊奇的感觉是因为你对这个笑话的翻译(你懂这个笑话的含义——译者),而不是说笑话人的因素。
103. (Supporting 98) Countess of Lovelace. 1842 The analytical engine can never do anything original.The analytical engine (see sidebar, "The Analytical Engine") could never discover any new facts. It is limited to drawing out consequences of facts that it has been provided with. In contemporary terms, a computer can only do what it has been programmed to do.
"The analytical engine has no pretensions to orginate anything ... It can follow analysis; but it has no power of anticipating any analytical relations or truths."
(支持98)Lovelace伯爵夫人(Ada King,英国名诗人拜伦的女儿,也是他的唯一合法子嗣;某种意义上人类历史上首位程序员,ADA程序设计语言的命名来源。她与Babbage共事设计Analytical Engine“分析性机械”,一种早期机械计算机原型,虽然由于当时的工程技术条件从未设计制成,但Ada曾基于这个原型整理和撰写过一些笔记性文档,而这些文档包含了一些计算问题在这个机械上的算法实现以及对计算机科学早期发展的一些前瞻性意见——译者)分析性机械永远不能做任何原创性的事情。分析机器不能发现新的事实。它对基于提供给它的事实而推导出结果的能力也甚为有限。一个计算机只能做它被编码要做的事情。
104. (Disputing 103) Alan Turing, 1950 The analytical engine may have been able to think for itself. Ada Lovelace was justified in denying that the analytical engine could be creative, because she had no evidence that it was creative. But because the analytical engine was in fact a universal digital computer, it may have had far greater capabilities than she realized. With added speed and storage capacity the analytical engine may have been able to think for itself.
(反驳103)阿兰·图灵,1950 分析性机械也可能为他自己思考。很自然Ada Lovelace是因为她没有证据表明分析性机械能有创造力所以她否认了它能具有创造力。但是因为分析性机械事实上也是一种通用数字计算机,它可能具有比她所意识到的更大的能力。当增加速度和存储容量,分析性机械可能能自行思考。
The Analytical Engine -- Charles Babbage
Invented by Charles Babbabge circa 1860, the analytical engine was a mechanical computer composed of gears, cranks, and wheels, which could be programmed by punch cards. In principle, Babbage's analytical engine could carry out any of the calculations a modern electronic computer can, but due to construction and design costs the analytical engine was never built during Babbage's lifetime (several have been constructed since).
分析性机械——Charles Babbage
Charles Babbage大约在1860年发明,分析性机械是一种机械计算机,由齿轮,曲柄和转盘组成,能够通过打孔卡编程。理论上说,Babbage的分析性机械能够实现所有现代电子计算机能做的事情,但由于建造和设计成本,这个机器在Babbage在世时从没建成(此后有些被建造了)。
105. (Disputing 97) Computers have already been creative. Computer models that exhibit creativity or at least some component of creativity have already been developed.
(反驳97)计算机已经具有创造力了。能够展示创造力或至少一些创造力成分的计算机已经被开发出来了。
106. (Disputing 105) Douglas Hofstadter, 1995 The ELIZA effect. The ELIZA effect is a tendency to read more into computer performance than is warranted by their underlying code. For exmaple, the computerized psychotherapy program ELIZA (see "ELIZA", Map 2, Box 34) gives apparently sympathtic responses to human concerns, but in fact is only utilizing a set of canned responses.
Note: The ELIZA effect was recognized and described by ELIZA's creator, Joseph Weizenbaum, though he didn't give it that title.
(反驳106)Dauglas Hofstadter,1995 ELIZA效应。ELIZA效应通常泛指一种人将计算机的所作所为依照人的理解看得过于神奇的倾向。例如,计算机化心理治疗程序ELIZA(详见“ELIZA”,第二图,32框)能够根据人的顾虑给出同理思考的回应,但这事实上只是在利用一组既定的响应集合。
107. (Supporting 105) H. Gelernter, 1963 The geometry program. The geometry program is a system that works backward from geometric theorems, searching for their proofs by means-end analysis. This planning breaks down the problems using a hierarchy of goals and subgoals. To avoid impossible searches the program uses heuristics to select the most promising search paths.
(支持105)H. Gelernter,1963 几何学程序。几何学程序是一种从几何定理出发,通过方法-目的分析的方法反向寻找它们的证明的系统。这个方案基于目的和子目的层次结构将问题分解。为了(尽早避免)减少不必要的搜索,这个程序用启发式方法选择最有希望的搜索路径。
108. (Supporting 105) Philip Johnson-Laird, 1988a The jazz generator. The jazz generator produces chord sequences and uses them to improvise chords, bass-line melodies, and rhythms. (so far i haven't heard any music pleasant to listen to composed by computer, in this case it's jazz not the sort of music i like, so reasonable enough I don't have the right to evaluate it -- translator)
(支持105)Philip Johnson-Laird,1988a 爵士乐生成器。爵士乐生成器生成一组和弦序列,并将它们用来即兴演奏和弦,低音旋律和节奏。
109. (Supporting 105) Magaret Masterman, 1971 Haiku program. A program has been written that develops haiku (a style of Japanese poetry) through interaction with humans. The model provides poets with synonym lists to aid in word choice and also constrains line length to ensure that the haiku is properly formed. The haiku program can run without human interaction by making arbitrary choices from its synonym lists. (And what subject would the program choose/specify for the poem? -- translator)
What's on the screen of the computer: 'All white in the buds, I flash snow peaks in the spring, Bang the sun has fogged.'
(支持105)Magaret Masterman,1971 Haiku(俳句)程序。已有一个程序被做能通过和人类交流能产生俳句(一种日语诗歌)。这个模型能为诗人提供同类词表帮助用词选择并控制行长以确保写出符合规则俳句。这个程序也能在无人干预的情况下通过对同类次表的任意选择自信运行。
110. (Supporting 105) Jim Meehan, 1975 Implemented Model: TALE-SPIN. This program writes stories with characters that have goals and subgoals dependent on their motivations. Its characters cooperate in each other's plans and can form competitive relationships when necessary to achieve their goals. The program can also represent a wide range of communications between characters.
A demo, On it's screen it says: '... George Ant was very thirsty. Goerge wanted to get some water. George walked from his patch of ground across the meadow through the valley to a river bank.'
(支持105)Jim Meehan,1975 实现模型:TALE-SPIN(会说故事的程序)。这个程序能写出包含具有依赖于他们的动机的目的和子目的系统的人物的故事。这些人物能互相考虑对方的计划展开协作,也可能为达到各自目的而竞争。这个程序也能展示这些人物之间的各种各样交流。
111. (Supporting 105) Harold Cohen, B. Cohen, and P. Nii, 1984 Implemented Model: AARON. AARON produces visual art by selecting a random starting point on a canvas and then drawing lines from that point using a complex set of if-then rules.
(支持105)Harold Cohen, B. Cohen, 和P. Nii,1984 实现模型:AARON。AARON能够制造视觉艺术(也就是画画)——它选择画布上的一个随机的起始点,然后从那个点开始通过运用一组复杂的if-else条件判断来画线。
112. (Supporting 105) Margaret Boden, 1990 Implemented Model: Connectionist systems exhibit creativity. Connectionist networks (is that in some way similar to a artificial neural network --translator) can learn to recognize patterns without being specifically programmed to do so. Note: Also, see Map 4.
(支持105)Margaret Boden,1990 实现模型:联络系统能够展示出创造力。联络网能够学着记住各种模式,而不必被以特定方式编程以做到这点。(亦见图4)
'The computer recognizes that letter A without having been programmed to do so '
113. (Supporting 105) Sheldon Klein, 1975 Implemented Model: Book generator. This automatic novel writer generates 2,100-word mysteries. It develops a rudimentary plot based on the conflicting motivations of its characters and fits the model of a mystery story by revealing the murderer at the end.
(支持105)Sheldon Klein,1975 实现模型:书籍生成器。这个自动小说作家能生成2100个词的悬疑故事。它能基于其人物的互为矛盾的动机发展一个简单的故事情节,并通过在故事最终展露凶手而实现一个悬疑故事的要求。
On the screen, it says: '... Lady Buxley was near James. James caressed Lady Buxley with passion. James was Lady Buxley's lover ...'
114. (Disputing 105) Margaret Boden, 1977 The book generator is inadequate. The book-writing program's fiction is inadequate for the following reasons, (1) The stories are shapeless and rambling. (2) The specific motivational patterns are relatively crude and unstructured. (3) The identification of the murderer comes as a statement rather than as a discovery.
(反驳105)Margaret Boden,1977 书籍生成器不足以说明问题。书籍生成器产生的故事是不恰当的,有一下理由:(1)故事不成形且松散;(2)这些动机模式相对非常粗糙且缺乏条理。(3)对凶手的最终确定是通过陈述的方式,而不是通过一个逐步发现的过程。