Machine Literature
Nov 2019





Computers have played a great role in promoting human literature since it was created. People don't have to spend too long borrowing books from libraries in other cities or going to other countries to find inspiration. It is obvious that computers have improved the efficiency of creation. More and more writers publish their works on the Internet, which makes their gifts available and give the unnoticed talent an opportunity to be popular in the new era. In addition, computers have made irreplaceable contributions to the publication, preservation and dissemination of literature. It can be said that computers have made modern literature prosperous which are used as tools to show people’s creativity and diversity.

Human creativity is infinite. However, Natsume Soseki said, “People like to cut up the vast world into pieces, draw their own fields and stick to them.” In early 1912, Albert Gleizes and Jean Metzinger, the judges of the French independent salon for cubism, rejected Marcel Duchamp's work named Nude Descending a Staircase No.2. The reason was that they were worried that the painting might provoke ridicule from Italian futurists, because there is movement in the painting which is what Italian Futurists interested about. They claimed in On Cubism that “Cubism opposes all systems and pursues complete liberation”, but their practice was to bind others and themselves in the form of their creation. The replacement of old and new styles cannot free art. All innovation are towards to establish rules and authority.

Too much inspiration is stifled by the secular rules, social morality and common sense. We live in a world where chairs are chairs and tables are tables. The Chair Project, the work of Philipp Schmitt, inspires me a lot. The computers learn a lot about the chairs made by human beings through algorithms and design new chairs which have no logic at all. Most of them do not even have a place to sit down, but they are so natural that each one is unique as a perfect individual. They inadvertently break the human view of chairs as the place where people sit down and challenge them. Is creativity a unique characteristic of human beings? Computers are no longer the tools created to work for human beings when they create these unique chairs, but creators.

The relationship between human beings and computers began to quietly reverse. The creation of computers is not as regular as that of human beings. All of them seem to be just different data. I call the trait emotionless and absolutely neutral. In 2001: Space Odyssey, there is a paragraph: “When the light was covering the cave, the moongazing man found his father dead. He did not understand that' this old thing' was his father and the relationship between father and son. However, he still felt vaguely uneasy when he saw the body. Later, this uneasiness will turn into sadness.” Emotion is an essential element in creation. The emotional factors and subjective views of the creators will be added, leading to their works moving in completely different directions and increasing the chances of creating world-famous works. Nothing but the meaning of the work is remained after those being taken away. The work that is emotionless and absolutely objective is rare because people have feelings. They are born and cannot be separated from the body. Therefore, an emotionless and absolutely neutral work would be very interesting.

The way of thinking of human beings is totally different from that of computers. The human mind is vast and unconstrained, while computers do with a lot of in-depth calculations. An ordinary person can learn cooking, golf, astronomy, Freud's psychoanalysis and anything he wants to learn, while machines usually spend a lot of time perfecting the algorithm of a task. For example, the artificial intelligence of playing Go researched by Google's Deepmind studio can defeat the top players in the world, but it knows nothing about chess. The algorithm engineers conceive logically what a computer neural network should look like, because none of them really learn anatomy to understand the neural networks of humans. Human neural networks are horizontal, while computer neural networks are vertical. This work is not to discuss which is better or worse, but to illustrate that the creativity of computers may break through some areas that human beings dare not overstep when creating, thus obtaining an innovative perspective.

自从计算机被创造出来以后,它对人类的文学起到了极大的推动作用。人们通过计算机,不用花费数十天去其他城市的图书馆借阅书籍,也不用远赴它国寻找灵感。显而易见,计算机提高了人类创作的效率。更多的作家把作品发表在网络上,网络使他们的才华得以面世,让被尘封的遗珠迎来新时代的喜爱。同时,计算机为文学的出版,保存与传播也做出了无可替代的贡献。可以说,计算机造就了现代文学的昌盛。人类把计算机当成工具,来展示自己的创造性,与多元性。






人类的创造力是无限的,但夏目漱石说过“人,喜欢把海阔天高的世界用小刀零切碎割,画出自己的领域,并在其中画地为牢。只在固守立足之地,任何时候也不越雷池一步”。就像1912年初,法国立体主义独立沙龙的评审委员格雷兹(Albert Gleizes)和梅景奇以担心“这张画因出现了意大利未来主义热衷的运动感,可能会惹意大利未来主义者的嘲笑”为由拒绝了杜尚的《下楼的裸女》。格雷兹和梅景奇等人在《论立体主义》这本书上宣称:“立体主义虽然被指责为一个体系,却是反对一切体系的,可立体主义要的是彻底的解放”,可他们的做法却是用自己创造的形式来束缚别人和自己。在一切人类创作新旧风格的更替,都不能让艺术重获“自由”,一切革新的步骤都是通向那个最终目标:建立规则和权威。




世俗的规则,社会的道德,人类的常识扼杀了太多新鲜的灵感,让我们活在一个椅子就是椅子,桌子就是桌子的世界。菲利浦(Philipp Schmitt)的作品《椅子项目(the Chair Project)》给我了我很大的启发,计算机通过算法深度学习了大量人类制造的椅子,并设计出新的椅子。这些计算机设计的椅子毫无逻辑可言,大部分连坐下的地方都没有,但却是那么浑然天成,每一个都是一个独特完美的个体。无意中,这个计算机的设计打破了人类数千年对椅子的定义——坐下的地方。同时,向人类提出这样的挑战“创造力是人类独有的特性吗”?当计算机创作出这些独一无二的椅子时,计算机已经不是一个被创造的来替人类工作的工具,而是一个创造者。



这个时候,人类和计算机的关系开始悄悄的颠倒。计算机的创作没有人类创作时那么多规则可言,所有的一切在它们眼中只是不同的数据而已,我将这种特质称为无情绪和绝对中立。在《2001:太空漫游》中有一段:“光掩入山洞的时候,望月者发现了父亲已经在夜里死了。他并不明白‘这个老东西’就是他的父亲,父子关系还不在他的理解范围之内。然而当他看到那具羸弱的尸体时,心里还是隐约感到一阵不安——后来,这种不安才会演化为哀伤”。情绪是人类创作中必不可少的一个元素,在创作的过程中,创作者会加入大量情绪的因素和主观的看法,这些东西会在创作中发生改变,导致作品走向截然不同的方向,也增加了创造闻名于世作品的机会。但是如果将创作者的情绪和主观看法拿走,剩下的就是作品本身的意义。一个无情绪和绝对客观的作品在人类艺术史中是凤毛麟角,因为人是有感情的,感情与生俱来,又随着身体和大脑回归黄土,它是不可剥离于身体之外的。正因为这样一个无情绪和绝对中立的作品会是很有趣的。





人类的思考方式和计算机的思考方式是截然不同的,人类的思维是广阔无垠并且天马行空,而计算机且是通过大量的深度计算。一个正常的人,可以学习做菜,学习打高尔夫,学习天文学,学习佛洛依德精神分析法,可以学习任何想要学习的事情。而机器通常会花费大量的时间完善一个任务的算法,就像Google公司的Deepmind工作室,他们研究的围棋人工智能可以战胜世界上最顶尖的围棋选手,但是对国际象棋一无所知。算法工程师搭建计算机的神经网络时,他们用以逻辑来构思计算机的神经网络应该是什么样,因为在他们中没人真正的学习过解剖学来了解人类的神经网络。人类的的神经网络是横向的,而计算机的神经网络是纵向的。这个作品的本质并不是讨论孰优孰劣,而是像说明计算机的创造力可能会突破当下人类创作时的一些不敢逾越的区域,从而得到新的视角。