按： Kindle 阅读器 的成功不但让亚马逊的老板 Jeff Bezos 感到吃惊 ，其实也会让所有关注这一事实的人感到吃惊。Bezos 预言纸质书籍将走向消亡。让人联想起 Eco 2003年11月1日在埃及亚历山大图书馆的演讲《书的未来》（译者：康慨）——

三种记忆

我们有三种记忆。第一种是生理上的，此种记忆由血肉形成，并归我们大脑支配。第二种是矿物的，在此意义上，人类已知有两种矿物形式的记忆：数千年前，有以陶板和石碑为载体的记忆，在埃及尤为著名，人们在其上刻下文字。第二种形式则是今日电脑的电子记忆，它以硅为基础。我们还知道另一种记忆，植物形式的记忆，首先是纸莎草纸，在埃及也同样著名，而后便是以纸制成的书。当然，史上最早的犊皮抄本也源自动物的身体，第一张纸也是由兽皮而非木材制成，但我们尽可以不去管它。我这样讲是为了简化植物形式的记忆与书籍的关系。

此地（译注：指亚历山大图书馆）过去始终致力于书籍的存护，将来也是如此，所以，它现在是，将来也会是植物记忆的圣殿。数百年来，图书馆一直是保存我们集体智慧的最重要的方式。它们始终都是一种全人类的大脑，让我们得以从中寻回遗忘，发现未知。请允许我做如下比喻：图书馆是一种最可能被人类效仿的神的智慧，有了它，就可在同一时刻看到并理解整个宇宙。人可以将得自一座大图书馆的信息存入心中，这使他有可能去习得上帝智慧的某些方面。换句话说，我们之所以发明图书馆，是因为我们自知没有神的力量，但我们会竭力效仿。

今天去建造，或翻新重建某座世界上最伟大的图书馆，听来像是一种挑战，或是挑衅。报纸或学报上，经常有某些作家面对这全新的电脑和互联网的时代，大谈即将来临的“书的死亡”。不管怎样，就算书会像方尖碑或陶制书写板那样消失，那也不会是什么废弃图书馆的好理由。正相反，图书馆会因保存对过去的发现而幸存下来，就像博物馆一样，同样的，我们之所以把罗塞塔石碑保存在博物馆里，是因为我们已没有在矿石表面刻下文件的习惯了。

不过，我对图书馆的盛赞大概有点过于乐观了。我属于那种始终相信印刷版图书仍有其未来的人，顺便说一句，所有担心它们消失的恐惧，不过是其他诸种恐惧，或是对某种东西将要终结的无尽恐惧中的一种，比如世界末日。

在多次采访中，我都被追问过这类问题：“新的电子媒体会让书籍消亡吗？网络会让文学消亡吗？新的超文本文明会消灭作家著述的观念吗？”如果你思维清楚，你就能看出来，这些问题各不相同，还要考虑到采访者提问时的思维方式。某人可能会想，如果你的回答是“不会的。别担心，万事无忧”，那么采访者就会疑虑全消。其实不然。如果你告诉这些人说，图书、文学、写作，都不会消失，他们又会是一副失望的表情。那么，怎样才能弄到独家新闻呢？发表某位诺贝尔奖得主去世的新闻不过一条消息而已，说他还活着，活得还挺好，这样的新闻不会有任何人感兴趣 ——也许除了那人自己以外。

法老不喜欢书写

今天我要做的，就是试着去解开纠缠在这些不同问题上的思维乱麻。阐明我们对这些不同问题的看法，可以帮助我们更好地去理解我们通常所说的图书、文本、文学、诠释等等。因而你会看到，一个愚蠢的问题是如何引出许多聪明的答案，或许还有，幼稚的采访如何产生了文化上的功能。

我们就从一个埃及故事开始，尽管这是希腊人讲的。据柏拉图在《斐德罗篇》所述，赫尔墨斯，或叫透特（译注：透特系埃及神话中的月神，掌魔法、智慧和写作，“赫尔墨斯”是希腊人对透特一词的希腊化改写），传说中书写的发明者，向法老塞穆斯（Thamus）展示了他的发明，法老赞扬了这种可让人类记住善忘之事的前所未闻的技术。但是，塞穆斯并未兴高采烈。“我的多才多艺的透特啊，” 他说，“记忆之力乃天赐神赋，须经不断训练才可保持长青。可人类一旦得到你的发明，将无需再磨炼其记忆之力。彼等记事，将不再因内在之努力，而仅仅借助外部工具的力量。”

我们能够理解塞穆斯的偏见。像任何新的技术发明一样，书写或许已经弱化了人类的力量，尽管其表面上是被取代和加强了。书写成了危险之术，因为它减弱了精神的力量，它给了人类一个僵化的灵魂，曲解的心智，一种矿物的记忆。

牧师仇视印刷术

当然，柏拉图写的是反语。柏拉图记下了他反对书写的理由。但是他也假托这些话是苏格拉底告诉他的，苏格拉底向来是只说不写（正因为他从未出书，所以他在学术争论中被人毁掉了）。现在，因为两个原因，无人再认同塞穆斯的偏见了。首先，我们知道书籍并不能使别人想我们之所想，正相反，书籍是一种激发更广阔思维的工具。只有在发明了书写之后，才有可能写出像普鲁斯特的《追忆似水年华》那样的出于自发记忆（spontaneous memory）的杰作。其次，如果人们从前为了记事而训练其记忆力的话，那么，在书写发明之后，他们仍然可以为了记住书中所说而训练记忆。书籍挑战并改进了记忆力，而不是使记忆麻痹。不管怎么说，法老都是一种永恒恐惧的范例：即一种对新技术成就将杀死我们认定的珍贵而有益之事的恐惧。

我故意用了“杀死”这个动词，是因为大概1400年之后，雨果在其《巴黎圣母院》中讲到牧师克洛德·孚罗洛悲伤地看着他的教堂尖塔时（也用了这个词）。《巴黎圣母院》发生在印刷术发明后的15世纪。在此之前，手稿保存在少数文化精英们的手里，他们教给大众的东西，不外乎那些圣经故事，基督和圣徒的生平，道德戒律。即便是民族的史迹，或是最基本的地理和自然科学观念（如未知民族的性格以及药草或石头的功效），也是得自教堂的画像。中世纪的教堂就像一档永不可撼动的电视节目，仿佛可以告诉人们需知的一切，既针对其日常生活，也为他们的永生服务。

可是这会儿，孚罗洛把一本印刷书放在他的桌上，低声说“ceci tuera cela”：这会杀了它，或者，用别的字眼来说，便是：这书会杀死教堂，字母会杀死画像。书将把人们从其最重要的价值观上移走，助长多余的信息，放任对经文的解读和愚蠢的好奇心。

超文本

在（20世纪）60年代，马歇尔·麦克卢汉写出了《古腾堡星系》（The Gutenberg Galaxy）一书，他在书中宣称，印刷术发明所提供的线性思维方式，通过电视图像或其他电子设备，即将被更加整合的认知和理解方式所代替。如果没有麦克卢汉，想必他的不少读者会先手指电视屏幕，然后再指着印刷的书，说“这会杀了它”。如果麦克卢汉还在我们中间，今天他也许会头一个写出“古腾堡反击”这样的东西来。当然了，电脑也是一种人们用以制作和编辑图像的工具，也的确通过图形来发布指令；但是同样，电脑也确已变成一种首要的字母工具。字词和线在其屏幕上游动，而且为了学会使用电脑，你必须得先学会读和写。

第一种“古腾堡星系”和第二种有什么不同吗？多的很。首先，只有（20世纪）80年代那种考古学意义上的文字处理器，才提供了一种线性写就的通讯，今天，只要它们呈现出一种超文本的结构，电脑就已不再是线性的了。说来奇怪，电脑是作为一种图灵机诞生的，每次只能执行一步，而事实上，在这种机器的内部，语言仍以这种方式工作，通过二进制的逻辑，0－1，0－1。然而，这种机器的输出已不再是线性的了：它就像一种符号焰火的爆炸。其模型完全不是直线的，它就像一个真正的星系，每个人都能在不同的星与星之间画出意想不到的连接，从而在任何新的视点上，画出全新的天像图。

而正是在这一点上，我们必须开始分而述之了，因为超文本结构通常意味着两种现象。首先，是一种文本式的超文本的存在。在传统书籍中，人必须以一种线性方式从左向右阅读（依据不同的文化，还可从右向左，或由上至下）。显然也可以跳着读——一下子翻到第300页——也可翻回来核对或重读第10页上的某些东西———但是这意味着体力劳动。与之相反，一种超文本格式的文本是一种多维度的网络，或者好比一座迷宫，其中每个点或节点都有与其他任何节点连接起来的可能。其次，是一种系统的超文本的存在。万维网（WWW）是所有超文本的光辉之母，一座全世界的图书馆，你能够，或者将会在短时间内，找出你需要的所有的书。网络是全部现有超文本的综合系统。

即将消亡的书

文本和系统的这种不同极其重要，我们过一会再说它。现在，先让我把常被问到的问题中最幼稚的那个搞掂，那里面的这种不同还不那么清晰。不过，通过回答这头一个问题，可以让我们弄清楚下一步的要点。这个幼稚的问题是：“超文本的磁盘，互联网，或是多媒体系统会让书籍消亡吗？”这个问题已经把我们带到了“这会杀了它”故事的最后一章。但是，甚至这个问题也令人迷惑，简洁一点的话，它可以有两种表达方式：（a）作为物理形式的书会消失吗？以及（b）作为一种虚拟形式的书会消失吗？

我先来回答第一个问题吧。哪怕在印刷术发明之后，书也从未成为获取信息的唯一手段。还有油画，大众图像印刷品，口授，等等。简而言之，书被证实是最适合传递信息的手段。书分两种：供阅读的书和供查阅的书。对“供阅读的书”而言，我把正常阅读它们的方式称作“侦探小说的方式”。你从第1页开始读起，作者一上来就告诉你罪案已经发生，随着案情的进展，你就逐段逐段地读下去，直到读完为止，最后你发现凶手是那个管家。此书的终结也就意味着你阅读体验的结束。请注意了，哪怕你在读——比如说哲学论文时，也会出现相同的过程。作者想让你打开书，从第1页开始读起，随着他提出的一系列的问题往下读，好看看他怎样得出最后必然的结论。当然，学者可以在重读这样一本书时，跳着来读，好把第一章的和最后一章的某两个陈述之间可能的联系隔离起来。他们也可以决定去隔离——比如说，“思想”这个词在一本指定著作中的每一次出现，这样，他们就会跳过好几百页，好把他们的注意力完全集中在那种意图之上。不过，外行恐怕会觉得这种阅读方式很不自然。

接下来是那种供查阅的书，例如手册和百科全书。百科全书的构想便是为了查阅，而完全不必从头读到尾。一个每晚睡前都要读《大英百科全书》，从头到尾读的人，恐怕是个引人发笑的怪人。通常，一个人翻开百科全书中的某一卷，是为了要了解或是想起拿破仑死在何时，要么就是硫酸的化学方程式。学者们使用百科全书的方式更为老练。举例来说，如果我想查查拿破仑有没有遇见康德的可能性，我会翻开我的百科全书的K卷和N卷：我找到拿破仑生于1769年，死于1821年，康德生于1724年，死于1804年，那时拿破仑已经当了皇帝。因此这二位没有会面的可能。为了确认这一点，我可能会查查康德或拿破仑的传记，但是对一生阅人无数拿破仑来说，一本短短的传记可能会将他与康德的会面忽略掉，而在康德的传记中，与拿破仑的会面便不会不提。简单地说，我必须在我图书馆的许多书架上快速翻阅许多本书；我必须记下笔记，好随后与我记下的所有数据进行比对。所有这一切都将让我付出艰苦的体力劳动。

是的，如果有超文本入替，我就可以浏览整个网络百科全书。我可以把一个已记录的事件，从一开始就与一系列散落于文本中的相似的事件联系起来；我可以从头到尾地比对；我可以去找一份所有以A打头的词汇清单；我可以去找所有拿破仑与康德的名字连在一起的事件；我可以比对他们生卒的数据——总之，我可以在几秒或几分钟之内做好我的工作。

超文本无疑会让百科全书和手册消亡。昨日，拥有一张装有整部百科全书的 CD－ROM已成为可能；今天，则可以上网查阅，其优势是，它允许把参考资料和非线性的补充信息混合起来使用。全部光盘，加上电脑，也只需一套印刷版百科全书五分之一的空间。一套印刷版的百科全书无法像一片CD－ROM那样方便运输，而且印刷版的百科全书的内容也不那么容易更新。现在，那些汗牛充栋，占据着我家和公立图书馆书架的百科全书，有望在不远的将来被消除出去，而且不会有什么抱怨他们消失的理由。让我们记住这一点吧，对许多人来说，一部卷帙浩瀚的百科全书是不可能实现的梦想，不是因为，或者不仅仅是因为买这些书的开销，还有装它们的书架占去的墙壁空间。拿我自己来说吧，当我以中世纪史学家的身份开始学者生涯时，我一度想在自家摆一套221卷的米涅（Migne）的《拉丁神父全集》（Patrologia Latina）。此书非常昂贵，但我还能付得起。我付不起的是一套可以装下221卷巨著的新公寓，要不然，我就只好把其他500部通常规格的大书清除出去。

不会消亡的书

那么，超文本的磁盘或万维网会取代供阅读的书吗？我们不得不再次就这个问题是关于书的物理形式还是虚拟形式做出决断。这一回，还是让我们先考虑物理问题吧。

好消息：书仍将是不可缺少的，这不仅仅是为了文学，也是为了一个供我们仔细阅读的环境，不仅仅是为了接受信息，也是为了要沉思并作出反应。读电脑屏幕跟读书是不一样的。想想学会一种新电脑程序的过程吧。通常，程序能把所有你需要的说明显示在屏幕上，但在大多数情况下，想了解此程序的用户还是会把说明打印出来，拿它们当书来读，要么就干脆买一本印刷版的说明书。不难想像，一种直观的程序可以把图书印刷和装订的过程讲得非常透彻，但是为了了解一种电脑程序如何编写，如何使用，我们还是需要一本印刷版的手册。

在电脑前呆上12个小时，我的眼睛就会像两个网球，我觉得非得找一把扶手椅，舒舒服服地坐下来，看看报纸，或者读一首好诗。所以，我认为电脑正在传播一种新的读写形式，但它无法满足它们激发起来的所有知识需求。请回忆一下，希伯来和早期阿拉伯人的文明都以一部书为基础，而这与他们都是游牧文明的事实无关。古代的埃及人可以将其记录刻在方尖石碑上，摩西和穆罕默德却没有。如果你想越过红海，或者从阿拉伯半岛到西班牙，比起方尖石碑来，卷轴书可能是记载《圣经》或《可兰经》的更具实用价值的工具。这也正是这两种文明的基础是厚文字而轻图像的书籍之原因所在。但是，书籍较之于电脑，还有另一个优势。即使是用只能保存70来年的现代酸纸印刷的书，也比磁介质更耐久。此外，它们不必受制于电力短缺和停电，也更不怕撞击。

到目前为止，书还是最经济，最灵活，最方便的信息传输方式，而且花费非常低。电脑通讯跑在你前面，书却会与你一同上路，而且步伐一致。如果你落难荒岛，没法给电脑接上电源，那么书仍然是最有价值的工具。就算你的电脑有太阳能电池，可你想躺在吊床上用它，也没那么容易。书仍然是落难时或日常生活中最好的伴侣。书是那种一旦发明，便无需再作改进的工具，因为它已臻完善，就像锤子、刀子、勺子或剪子一样。

按需印刷和电子书

有两种新发明即将投入企业化开发。其一是按需印刷：读者在查阅众多图书馆或出版社的目录之后，可以选出他所需的书，而后，操作员一按机器的电钮，便可将书按照读者喜欢的字体印出，并单独装订成书。这必将改变整个出版市场，或许会让书店走向灭亡，但它消灭不了书，也消灭不了图书馆，图书馆乃是寻获图书，扫描及重印的唯一之地。简单地说，每本书都将按购买者的愿望量身订做，就像过去的手抄本一样。

第二种发明是电子书（e－book），在其中插入微磁盘（micro－cassette），或将它连至互联网，便可当面印刷成书。即便在这种情况下，我们得到的仍会是书，尽管它不同于现有的书，好比今日之书不同于旧日的羊皮纸手抄本，亦如莎士比亚剧作的1623年首部对开本与最新的企鹅版之不同。然而，到目前为止，电子书尚未如其发明者所希望的那样，取得商业上的成功。我不断听说有些在电脑前长大，已不习惯翻阅图书的电脑迷，用电子书读了经典文学名著，但是我认为，这种现象仍然极为有限。普遍而言，人们似乎更喜欢传统的阅读方式，读印在纸上的诗或小说。电子书或许能证明在信息查询方面有用，就像字典或特殊的文件。它们会对上学时必须背着十来本书的学生们有帮助，但它们不会取代其他形式的书，例如我们喜欢在床头阅读的那种书。

事实上，尽管新技术设备层出不穷，但旧东西并未因此全然消亡。汽车跑得比自行车快，但并没有让自行车销声匿迹，新的技术进步也没让自行车焕然一新。新技术必然导致旧物废弃的想法往往过于单纯。照相术发明后，虽然画家们感到没有必要再像匠人那样复制现实了，但这并不意味着达盖尔（Daguerre，译注：银版照相术的发明人）的发明仅仅催生了抽象画法。在那种没有照相范例便存在不下去的现代绘画中，仍然有一整套传统：想想看，例如超写实主义（hyper－realism）。此刻，画家的眼睛通过摄影的眼睛看到现实。这意味着在文化史上，从来没有一物简单地杀死另一物这样的事例。当然，新发明总是让旧的发生深刻的变化。

诠释的界限

为了明判物理形式的书会消失这样一个矛盾的题目，可以说，有时这种恐惧不仅涉及到书，也涉及大多数印刷品。唉，如果有那么一天某人希望电脑，特别是文字处理软件，将有助于节约树木，那真是痴心妄想。相反，电脑促进了印刷品的生产。电脑为印刷文本创造出了新的制造和传播方式。为了校读或修改某份文本，如果它不是简短信函的话，那么人就需要将其打印出来，然后校读，而后在电脑上进行修改，并再把它打印出来。我认为，一个人要想写一篇几百页的文本并将其改定，不经多次打印是不可能的。

现在出现了新的超文本诗学，按照它的观点，无论是一本供阅读的书，还是一首诗，都可转换成超文本。就此而言，我们正在转向第二个问题，因为问题已不再是，或不仅是一个物理问题了，而是一个关系到创造性活动真正天性的问题，关系到阅读的过程，为了解开问题的谜团，我们首先要判断的是，通过一个超文本的链接，我们意欲何为。

请注意，如果这个问题涉及的是关于读者的无限，或无法确定的诠释的可能性，那这个问题就没什么可讨论的了。当然，它或许与乔伊斯作品的诗性有关，例如，他将其《芬尼根守灵夜》（Finnegans Wake）一书视为一种给患有理想的失眠者的理想中的读者读的文本。这个问题关系到诠释的界限，解构阅读的界限，以及过度诠释的界限，对此我已有专文论述。（译注：见艾柯《诠释与过度诠释》，三联书店1997年4月版）不：现在考虑的是无限，或至少是无法确定的大量诠释的事例，其原因不仅在于读者的主动，也在于文本本身的物理可动性，它们之所以被制造出来，就是为了被重写。为了理解这种类型的文本如何工作，我们应判明我们正在讨论的文本宇宙究竟是有限的和暂时的，有限的但却是无尽的，无尽的但却是有限的，还是无限的和无尽的。

系统和文本

首先，我们应将系统与文本区别开来。一种系统，例如一种语言系统，是一种由给定的自然语言所展现的全部可能性。一种语法规则的有限集合，容许讲话者制造出数量无限的语句，每个语言学的项目可依据其他语言学或语义学的项目得到诠释——词通过词义，事件通过事例，动物或花朵通过图像得到诠释，等等。

举个例子，翻开一部百科大辞典，它或会将狗定义为哺乳动物，那么，就翻到哺乳动物的条目，如果哺乳动物被定义为动物，就得到动物条目中去找，以此类推。同时，狗的属性可通过不同种类的狗的图示加以说明，如果它说，某种狗生活在拉普兰（译注：位于欧洲最北部，包括挪威北部、瑞典和芬兰以及苏联西北部的科拉半岛。大部分地处极圈之内），那你就得翻到拉普兰的条目，才能知道它的位置，大概就是这个意思。系统是有限的，一部百科全书在物理属性上也是有限的，但是你可以通过一种螺旋型运动环航其中，在这个意义上，它实质上是无限的。当然，这样说来，所有你能想得到的书就都被包括在其中了，只要有一部好字典和好的语法书。如果你能很好地使用英语辞典，那么你也能写出《哈姆雷特》，而且，前人同样为之的可能性微乎其微。如果把同样的文本系统交给莎翁和一个学童，那么他们写出《罗密欧与朱丽叶》的机率是一样的。

语法、辞典和百科全书是系统：你能用它们创造出所有你想要的文本。但是一个文本不能自成语言学或百科全书式的系统。一个特定的文本减低了系统构建封闭宇宙的无限或无尽的可能性。如果我说出一个句子，例如，“今天早上我吃过早餐了……”辞典就允许我开列许多可能的条目，并假定它们都在系统之内。但是，如果我明确地创造出我的文本，并说：“今天早上我吃了面包和黄油早餐”，那么我便把奶酪、鱼子酱、熏牛肉和苹果排除在外了。一个文本阉割了系统的无限可能性。《天方夜谭》能以许许多多种方式来加以诠释，但这故事发生在中东，不是在意大利，可以说，它讲的是阿里巴巴或莎赫扎德的故事，而无关一位决心猎捕白鲸的船长，或是某位游历了地狱、炼狱以及天堂的托斯卡纳诗人。

超文本创造了无限

再说说童话，例如《小红帽》，其文本始于一组给定的人物及情境——一个小女孩，一位母亲，一位外婆，一只狼，一片树林——并经过一系列限定的情节到达结局。当然，也可以把童话当做寓言来读，并赋予其情节和人物的行为以不同的道德含义，但是，《小红帽》是无法转换成《灰姑娘》的。《芬尼根守灵夜》的确是开放的，可以有很多种诠释，但它有一点是确定的，即它绝对不可能给出费马大定理（Fermat’s last theorem）的证明，或是伍迪·艾伦的全传。这一点看似微不足道，但许多解构学家的最根本错误，便是相信文本无所不能。其错谬显而易见。

现在，假设一下，一个暂时的和有限的文本，被许多词与词之间的链接超文本化地组织起来了。在辞典和百科全书中，“狼”这个字被潜在地相连至每一个部分构成其可能定义或描述的其他词（“狼”可与动物相连，可与哺乳动物、凶残、腿、毛皮、眼睛、森林，与那些有狼生活的国家的名字等等连在一起）。而在《小红帽》中，“狼”只能与这个字出现，或是使人明确感到其出现的那些段落连在一起。这一系列的可能的连接是暂时的和有限的，那么超文本的策略怎样才能被用来“打开”一个暂时的和有限的文本呢？

第一种可能性是使这一文本在物理性上受限，就这一层意义而言，一个故事可以通过不同作者的连续写作而加以丰富，在一种双重的意义上，比如说两维或三维的。但是我指的是特定的，如《小红帽》，第一位作者设定了开篇的场景，（女孩进了森林，）别的作者可以一个接一个地将故事发展下去，例如，让女孩没有遇到狼，而是阿里巴巴，让他们俩进入一座魔法城堡，然后邂逅一条有魔法的鳄鱼，凡此种种，这样，故事便能持续多年。但是在每种叙述都不相关的层面上，这个文本也可能是无限的，例如，当女孩进了森林，多个作者便可做出多种不同的选择。一个作者可以写女孩遇到了匹诺曹，另一位则可把她变成天鹅，或是进入金字塔，让她发现图坦卡蒙之子的宝藏。

这就是今天的可能，你可以在互联网上找到这种文学游戏的一些有趣的例子。

即兴演出的例子

对此，人们会提出作者身份和艺术作品的完整概念，作为一种有机整体能否幸存的问题。我想简单地告诉我的听众，这种事过去就有过，它既没有扰乱作者身份，也没有危及有机整体。第一个例子是意大利的职业艺术喜剧（译注：Commedia dell’arte，一种即兴戏剧形式，始自16世纪之意大利，流行至18世纪，今日仍有演出。系一种以流动剧团搭建露天舞台，并杂以戏法儿和杂技等娱乐观众的幽默喜剧形式，有角色固定的保留剧目，以及粗略的剧情大纲，即艾柯后文所称的Canovaccio），以Canovaccio，亦即故事大纲为基础，每次演出则有赖于演员的情绪和即兴发挥，因而每次演出都各不相同，所以，我们无法通过人称“有两个主子的滑稽仆人”（Arlecchino servo di due padroni）的一个单一作者，来判定任何一部单一的作品，而只能记录一个连续的演出系列，其大多数无疑已经失传，而它们全都互不相同。

另一个例子是即兴爵士乐演出（jazz jam session）。可以相信，《北新街蓝调》（译注：Basin Street Blues，爵士乐的经典之作）只有过一次通过后来的录音保存下来的独享尊荣的演出，但我们知道这不是真的。有多少次演出，就有多少种《北新街蓝调》，一旦两位或更多的表演者再度相会，并且试验他们对原始主旋律个性化和经过创新的版本，那么将来还会有许多我们尚不知道的《北新街蓝调》。我想说的是，我们已经习惯于集体流行艺术领域内作者身份缺失的观念，每位参与者都以其对爵士音乐的体验，在其中加入了某种东西，使之成为没有穷尽的故事。

这种实现自由创作的途径很受欢迎，并成为社会的文化组织构成的一个部分。

从有限中创造无限

然而，无尽和无限文本的创造活动，有别于现有的文本，后者也许可以被无限的方式加以诠释，但它在物理形式上仍然是有限的。与我们同时代的文化中，按照不同的标准，我们对贝多芬第五交响曲的新演出和某场《北新街》主旋律的新的即兴演奏，均加以接受和评价。在这个意义上，我看不出创作集体的迷人游戏，互联网上无限的故事，如何能把我们与作者的文学和艺术割裂开来。当然了，我们正在向一个更加自由的社会前进，自由的创造力将和对现有文本的诠释共存。我喜欢这样。但是不能说我们已经以新代旧了。我们新旧都要。

看电视时跳过广告是另一种与传统意义上的看电影无关的行为。一种超文本设备，允许我们创造出与对以往文本的诠释能力无关的新文本。我一直在极力寻找一种即是无限又是有限的文本环境的例证，却始终无法如愿。事实上，如果你有数量无限的元素来创作，那么为何还要把自己限制在一种有限宇宙的创造之内呢？这是个神学问题，一种宇宙的运动，人，或神，可在其中完成每种可能之事，但却为它自身立下了规条或界限，并创造出一个非常小而简单的宇宙。可是，我来考虑一下另一种可能性吧，即一开始便保证了一种基于有限元素的数量无限的可能性，就像一种符号系统，然而实际上却只有一种自由和创造力的幻像。

超文本可以提供一种自由发挥的幻像，即使是一种封闭的文本：一篇侦探小说可以被结构为这样一种方式，让读者能够选择自己的解决方案，决定其结尾是否让凶手指向管家、或主教、侦探、叙述者，作者或是读者。他们可以据此建立自己的个人故事。这样一种想法并不新鲜。在电脑发明之前，诗人和叙述者便梦想着一种完全开放的文本，让读者能够以不同的方式进行无限的再创作。这就是马拉美（译注：Stéphane Mallarmé，1842－1898，法国象征主义诗人）所赞美的“书”（Le Livre）的理念。雷蒙·格诺（译注：Raymond Queneau，1903－1976，法国诗人，小说家和出版家，后现代主义先驱）也发明了一种组合算法，借助于可能性，可以从句子的有限集中创作出成百万首诗歌，在（20世纪）60年代早期，马克斯·萨波塔（Max Saporta）写作并出版了一部小说，变换其页码的顺序，便可组成不同的故事。南尼·贝莱斯蒂尼（Nanni Balestrini）也给一台电脑输入了一组互不相关的诗歌，让机器以不同的组合方式创作出不同的诗。许多当代的作曲家也用机器来生成乐谱，这样便得到了不同的音乐表现。

所有这些可以物理形式移动的文本，造就了一种在读者方面绝对自由的印象，但这只是一种印象，一种自由的幻像。可让人以数量有限的元素去制造无限文本的方法已经存在了几千年，这就是字母表。使用字母数量有限的字母表，可制造出几十亿种文本，这正是从荷马到今天的人类所为。相反，一种刺激文本（stimulus－text）给我们带来的不是字母或词汇，而是已经建立的词汇排列，或是成页的记录，并不能让我们自由创造我们想要的东西。我们仅仅能够自由地，以适度多种的方式移动已有的文本块。一座考尔德活动雕塑（Calder mobile，译注：亚历山大·考尔德，1898－1976，美国现代主义雕塑家，以用铁丝制作能“动”的雕塑而闻名）的迷人之处，不在于它制造出了有无限可能的运动，而是因为我们欣赏艺术家强加于它的那种铁的规则，因为这种活动雕塑只以考尔德设想的方式活动。

在自由文本最后的边界上，也会有一种以封闭文本开始的文本。比如说，《小红帽》或《天方夜谭》，以及作为读者的我能以个人喜好加以修改的文本，如此详细阐述的一种第二文本，已不同于原始文本，它的作者是我自己，即使确定我的作者身份是为了对抗明确作者的观念。互联网对这些实验是开放的，其大多数美丽而有益。没有什么东西可以禁止某人写出一个小红帽把狼吃掉的故事。但是这与它的现实功用无关，它涉及的是书的深刻魅力。

阅读经验是唯一的

一本书给我们提供了一个文本，它在对多种解读开放时，告诉了我们某种无法改变的东西。假设你正在读托尔斯泰的《战争与和平》：你极其希望娜塔莎不要接受可怜的无赖安德烈的求婚；你极其希望那个了不起的王子一般的安德烈不要死去，让他和娜塔莎能白头偕老。如果你有一部超文本和交互式CD－ROM的《战争与和平》，便可依你所愿重写自己的故事；你可以创造无数的《战争与和平》，让彼埃尔· 别素霍夫杀死拿破仑，或是按照你的意愿，让拿破仑最后击败库图佐夫将军。多么自由，多么激动人心啊！每个布瓦尔（Bouvard）或佩居谢（Pécuchet）都能成为福楼拜。（译注：《布瓦尔和佩居谢》是福楼拜的一部未完成小说，主人公是两个费劲力气研究学问，最终却一事无成的誊写员，卡尔维诺称之为两位“堂·吉诃德式的作家”。）

唉，一本已经写出的书，其命运已经被压抑的作者的决定所确定，我们已无能为力。我们被迫接受命运，终于明白了我们无法改变宿命。一本超文本的和交互式的小说允许我们去实践自由和创造，我希望这种富于创意的活动能在未来的学校中实现。但是那种确已写出的小说，如《战争与和平》，所面对的不是我们想像力的无限可能，而是掌握着生与死的严格律法。

同样地，在《悲惨世界》中，维克多·雨果给我们提供了滑铁卢之战的美丽描述。雨果的滑铁卢与司汤达的截然不同，司汤达在《巴马修道院》中，透过他的英雄的眼睛看这场战役，而这位英雄的角度是在事件内部，这就不能理解其复杂性。相反，雨果从上帝的视点来描写这场战役，跟从每一处细节，让他透视整个场景的叙述居于统治地位。雨果不仅知道发生了的事，而且知道可能会发生的，以及实际不会发生的事。他知道，如果拿破仑得知在圣约翰山顶那边，有一道深沟，那么米约将军的铁甲奇兵就不会崩溃在英军脚下，然而他当时的情报却是模糊的或是缺失的。雨果知道，如果给冯·布罗将军当向导的牧羊人指了一条不同的路，那么普鲁士军队就无法及时赶到并击败法军。（译注：见《悲惨世界》第二部第一卷：“滑铁卢”）

的确，在角色扮演游戏中，某人可以重写滑铁卢，这样，格鲁希的一彪人马便可赶来救援拿破仑。但是雨果的滑铁卢的悲剧之美，正是让读者感到事情并不以他们的意愿为转移。悲剧文学作品的魅力，是让我们感到书中的英雄有逃脱其命运的可能，但却未能遂愿，原因在于他们的脆弱，他们的骄傲，或是他们的盲目。此外，雨果告诉我们：“这样一种晕眩，这样一种错误，这样一种毁灭，这样一种让整个历史为之震惊的失败，难道是某种无因之果吗？不……对即将到来的新时代而言，伟人的消失是必然的。某个无人可以反对的人，掌管着这一事件……上帝从这里经过，上帝经过了（Dieu a passé）。”

这就是每一部伟大的书所告诉我们的，上帝从这里经过，他的经过，既是为了信者，也是为了疑者。我们不可重写的书是存在的，因为其功能是教给我们必然性，只有在它们得到足够敬意的情况下，才会给我们以智慧。为了到达一个更高的知识境界和道德自由，它们可约束的课程不可或缺。

我希望并祝愿亚历山大图书馆继续收藏这种书，以使新读者得到阅读它们时那种不可替代的经验。万岁！这座植物记忆的殿堂。

（小标题为译者所加）

Vegetal and mineral memory: The future of books

The city of Alexandria played host on 1 November to the renowned Italian novelist and scholar Umberto Eco, who gave a lecture in English, on varieties of literary and geographic memory, at the newly opened Bibliotheca Alexandrina. Al-Ahram Weekly publishes the complete text of the lecture.

WE HAVE THREE TYPES OF MEMORY . The first one is organic, which is the memory made of flesh and blood and the one administrated by our brain. The second is mineral, and in this sense mankind has known two kinds of mineral memory: millennia ago, this was the memory represented by clay tablets and obelisks, pretty well known in this country, on which people carved their texts. However, this second type is also the electronic memory of today’s computers, based upon silicon. We have also known another kind of memory, the vegetal one, the one represented by the first papyruses, again well known in this country, and then on books, made of paper. Let me disregard the fact that at a certain moment the vellum of the first codices were of an organic origin, and the fact that the first paper was made with rugs and not with wood. Let me speak for the sake of simplicity of vegetal memory in order to designate books.

This place has been in the past and will be in the future devoted to the conservation of books; thus, it is and will be a temple of vegetal memory. Libraries, over the centuries, have been the most important way of keeping our collective wisdom. They were and still are a sort of universal brain where we can retrieve what we have forgotten and what we still do not know. If you will allow me to use such a metaphor, a library is the best possible imitation, by human beings, of a divine mind, where the whole universe is viewed and understood at the same time. A person able to store in his or her mind the information provided by a great library would emulate in some way the mind of God. In other words, we have invented libraries because we know that we do not have divine powers, but we try to do our best to imitate them.

To build, or better to rebuild, today one of the greatest libraries of the world might sound like a challenge, or a provocation. It happens frequently that in newspaper articles or academic papers some authors, facing the new computer and internet era, speak of the possible “death of books”. However, if books are to disappear, as did the obelisks or the clay tablets of ancient civilisations, this would not be a good reason to abolish libraries. On the contrary, they should survive as museums conserving the finds of the past, in the same way as we conserve the Rosetta Stone in a museum because we are no longer accustomed to carving our documents on mineral surfaces.
Yet, my praise for libraries will be a little more optimistic. I belong to the people who still believe that printed books have a future and that all fears à propos of their disappearance are only the last example of other fears, or of milleniaristic terrors about the end of something, the world included.

In the course of many interviews I have been obliged to answer questions of this sort: “Will the new electronic media make books obsolete? Will the Web make literature obsolete? Will the new hypertextual civilisation eliminate the very idea of authorship?” As you can see, if you have a well-balanced normal mind, these are different questions and, considering the apprehensive mode in which they are asked, one might think that the interviewer would feel reassured when your answer is, “No, keep cool, everything is OK”. Mistake. If you tell such people that books, literature, authorship will not disappear, they look desperate. Where, then, is the scoop? To publish the news that a given Nobel Prize winner has died is a piece of news; to say that he is alive and well does not interest anybody — except him, I presume.

WHAT I WANT TO DO TODAY is to try to unravel a skein of intertwined apprehensions about different problems. To clarify our ideas about these different problems can also help us to understand better what we usually mean by book, text, literature, interpretation, and so on. Thus you will see how from a silly question many wise answers can be produced, and such is probably the cultural function of naive interviews.

Let us start with an Egyptian story, even though one told by a Greek. According to Plato in Phaedrus when Hermes, or Theut, the alleged inventor of writing, presented his invention to the Pharaoh Thamus, the Pharaoh praised such an unheard of technique supposed to allow human beings to remember what they would otherwise forget. But Thamus was not completely happy. “My skillful Theut,” he said, “memory is a great gift that ought to be kept alive by continuous training. With your invention people will no longer be obliged to train their memory. They will remember things not because of an internal effort, but by mere virtue of an external device.”

We can understand the preoccupation of Thamus. Writing, like any other new technological invention, would have made torpid the human power which it pretended to substitute and reinforce. Writing was dangerous because it decreased the powers of mind by offering human beings a petrified soul, a caricature of mind, a mineral memory.
Plato’s text is ironical, naturally. Plato was writing down his argument against writing. But he was also pretending that his discourse was told by Socrates, who did not write (since he did not publish, he perished in the course of the academic fight.) Nowadays, nobody shares Thamus’s preoccupations for two very simple reasons. First of all, we know that books are not ways of making somebody else think in our place; on the contrary, they are machines that provoke further thoughts. Only after the invention of writing was it possible to write such a masterpiece of spontaneous memory as Proust’s A la Recherche du Temps Perdu . Secondly, if once upon a time people needed to train their memories in order to remember things, after the invention of writing they had also to train their memories in order to remember books. Books challenge and improve memory; they do not narcotise it. However, the Pharaoh was instantiating an eternal fear: the fear that a new technological achievement could kill something that we consider precious and fruitful.

I used the verb to kill on purpose because more or less 14 centuries later Victor Hugo, in his Notre Dame de Paris , narrated the story of a priest, Claude Frollo, looking in sadness at the towers of his cathedral. The story of Notre Dame de Paris takes places in the XVth century after the invention of printing. Before that, manuscripts were reserved to a restricted elite of literate persons, and the only thing to teach the masses about the stories of the Bible, the life of Christ and of the Saints, the moral principles, even the deeds of national history or the most elementary notions of geography and natural sciences (the nature of unknown peoples and the virtues of herbs or stones), was provided by the images of a cathedral. A mediaeval cathedral was a sort of permanent and unchangeable TV programme that was supposed to tell people everything indispensable for their everyday life, as well as for their eternal salvation.

Now, however, Frollo has on his table a printed book, and he whispers “ceci tuera cela”: this will kill that, or, in other words, the book will kill the cathedral, the alphabet will kill images. The book will distract people from their most important values, encouraging unnecessary information, free interpretation of the Scriptures, insane curiosity.
During the sixties, Marshall McLuhan wrote his book The Gutenberg Galaxy , where he announced that the linear way of thinking supported by the invention of printing was on the verge of being substituted by a more global way of perceiving and understanding through TV images or other kinds of electronic devices. If not McLuhan, then certainly many of his readers pointed their finger first at a TV screen and then to a printed book, saying “this will kill that”. Were McLuhan still among us, today he would have been the first to write something like “Gutenberg strikes back”. Certainly, a computer is an instrument by means of which one can produce and edit images, certainly instructions are provided by means of icons; but it is equally certainly that the computer has become first of all an alphabetic instrument. On its screen there run words and lines, and in order to use a computer you must be able to write and to read.
Are there differences between the first Gutenberg Galaxy and the second one? Many. First of all, only the archaeological word processors of the early eighties provided a sort of linear written communication. Today, computers are no longer linear in so far as they display a hypertextual structure. Curiously enough, the computer was born as a Turing machine, able to make a single step at a time, and in fact, in the depths of the machine, language still works in this way, by a binary logic, of zero-one, zero-one. However, the machine’s output is no longer linear: it is an explosion of semiotic fireworks. Its model is not so much a straight line as a real galaxy where everybody can draw unexpected connections between different stars to form new celestial images at any new navigation point.

YET IT IS EXACTLY AT THIS POINT that our unravelling activity must start because by hypertextual structure we usually mean two very different phenomena. First, there is the textual hypertext. In a traditional book one must read from left to right (or right to left, or up to down, according to different cultures) in a linear way. One can obviously skip through the pages, one — once arrived at page 300 — can go back to check or re- read something at page 10 — but this implies physical labour. In contrast to this, a hypertextual text is a multidimensional network or a maze in which every point or node can be potentially connected with any other node. Second, there is the systemic hypertext. The WWW is the Great Mother of All Hypertexts, a world-wide library where you can, or you will in short time, pick up all the books you wish. The Web is the general system of all existing hypertexts.

Such a difference between text and system is enormously important, and we shall come back to it. For the moment, let me liquidate the most naive among the frequently asked questions, in which this difference is not yet so clear. But it will be in answering this first question that we will be able to clarify our further point. The naive question is: “Will hypertextual diskettes, the internet, or multimedia systems make books obsolete?” With this question we have arrived at the final chapter in our this-will-kill-that story. But even this question is a confused one, since it can be formulated in two different ways: (a) will books disappear as physical objects, and (b) will books disappear as virtual objects?

Let me first answer the first question. Even after the invention of printing, books were never the only instrument for acquiring information. There were also paintings, popular printed images, oral teaching, and so on. Simply, books have proved to be the most suitable instrument for transmitting information. There are two sorts of book: those to be read and those to be consulted. As far as books-to-be-read are concerned, the normal way of reading them is the one that I would call the “detective story way”. You start from page one, where the author tells you that a crime has been committed, you follow every path of the detection process until the end, and finally you discover that the guilty one was the butler. End of the book and end of your reading experience. Notice that the same thing happens even if you read, let us say, a philosophical treatise. The author wants you to open the book at its first page, to follow the series of questions he proposes, and to see how he reaches certain final conclusions. Certainly, scholars can re-read such a book by jumping from one page to another, trying to isolate a possible link between a statement in the first chapter and one in the last. They can also decide to isolate, let us say, every occurrence of the word “idea” in a given work, thus skipping hundreds of pages in order to focus their attention only on passages dealing with that notion. However, these are ways of reading that the layman would consider as unnatural.

Then they are books to be consulted, like handbooks and encyclopaedias. Encyclopaedias are conceived in order to be consulted and never read from the first to the last page. A person reading the Encyclopaedia Britannica every night before sleeping, from the first to the last page, would be a comic character. Usually, one picks up a given volume of an encyclopaedia in order to know or to remember when Napoleon died, or what is the chemical formula for sulphuric acid. Scholars use encyclopaedias in a more sophisticated way. For instance, if I want to know whether it was possible or not that Napoleon met Kant, I have to pick up the volume K and the volume N of my encyclopaedia: I discover that Napoleon was born in 1769 and died in 1821, Kant was born in 1724 and died in 1804, when Napoleon was already emperor. It is therefore not impossible that the two met. In order to confirm this I would probably need to consult a biography of Kant, or of Napoleon, but in a short biography of Napoleon, who met so many persons in his life, a possible meeting with Kant can be disregarded, while in a biography of Kant a meeting with Napoleon would be recorded. In brief, I must leaf through many books on many shelves of my library; I must take notes in order to compare later all the data I have collected. All this will cost me painful physical labour.

Yet, with hypertext instead I can navigate through the whole net-cyclopaedia. I can connect an event registered at the beginning with a series of similar events disseminated throughout the text; I can compare the beginning with the end; I can ask for a list of all words beginning by A; I can ask for all the cases in which the name of Napoleon is linked with the one of Kant; I can compare the dates of their births and deaths — in short, I can do my job in a few seconds or a few minutes.

Hypertexts will certainly render encyclopaedias and handbooks obsolete. Yesterday, it was possible to have a whole encyclopaedia on a CD-ROM; today, it is possible to have it on line with the advantage that this permits cross references and the non-linear retrieval of information. All the compact disks, plus the computer, will occupy one fifth of the space occupied by a printed encyclopaedia. A printed encyclopaedia cannot be easily transported as a CD-ROM can, and a printed encyclopaedia cannot be easily updated. The shelves today occupied at my home as well as in public libraries by metres and metres of encyclopaedias could be eliminated in the near future, and there will be no reason to complain at their disappearance. Let us remember that for a lot of people a multivolume encyclopaedia is an impossible dream, not, or not only, because of the cost of the volumes, but because of the cost of the wall where the volumes are shelved. Personally, having started my scholarly activity as a medievalist I would like to have at home the 221 volumes of Migne’s Patrologia Latina . This is very expensive, but I could afford it. What I cannot afford is a new apartment in which to store 221 huge books without being obliged to eliminate at least 500 other normal tomes.

Yet, can a hypertextual disk or the WWW replace books to be read? Once again we have to decide whether the question concerns books as physical or as virtual objects. Once again let us consider the physical problem first.

Good news: books will remain indispensable, not only for literature but for any circumstances in which one needs to read carefully, not only in order to receive information but also to speculate and to reflect about it. To read a computer screen is not the same as to read a book. Think about the process of learning a new computer programme. Usually, the programme is able to display on the screen all the instructions you need. But usually users who want to learn the programme either print the instructions and read them as if they were in book form, or they buy a printed manual. It is possible to conceive of a visual programme that explains very well how to print and bind a book, but in order to get instructions on how to write, or how to use, a computer programme, we need a printed handbook.

After having spent 12 hours at a computer console, my eyes are like two tennis balls, and I feel the need of sitting down comfortably in an armchair and reading a newspaper, or maybe a good poem. Therefore, I think that computers are diffusing a new form of literacy, but they are incapable of satisfying all the intellectual needs they are stimulating. Please remember that both the Hebrew and the early Arab civilisations were based upon a book and this is not independent of the fact that they were both nomadic civilisations. The Ancient Egyptians could carve their records on stone obelisks: Moses and Muhammad could not. If you want to cross the Red Sea, or to go from the Arabian peninsula to Spain, a scroll is a more practical instrument for recording and transporting the Bible or the Koran than is an obelisk. This is why these two civilisations based upon a book privileged writing over images. But books also have another advantage in respect to computers. Even if printed on modern acid paper, which lasts only 70 years or so, they are more durable than magnetic supports. Moreover, they do not suffer from power shortages and black-outs, and they are more resistant to shocks.

Up to now, books still represent the most economical, flexible, wash-and-wear way to transport information at a very low cost. Computer communication travels ahead of you; books travel with you and at your speed. If you are shipwrecked on a desert island, where you don’t have the option of plugging in a computer, a book is still a valuable instrument. Even if your computer has solar batteries, you cannot easily read it while lying in a hammock. Books are still the best companions for a shipwreck, or for the day after the night before. Books belong to those kinds of instruments that, once invented, have not been further improved because they are already alright, such as the hammer, the knife, spoon or scissors.

TWO NEW INVENTIONS , however, are on the verge of being industrially exploited. One is printing on demand: after scanning the catalogues of many libraries or publishing houses a reader can select the book he needs, and the operator will push a button, and the machine will print and bind a single copy using the font the reader likes. This will certainly change the whole publishing market. It will probably eliminate bookstores, but it will not eliminate books, and it will not eliminate libraries, the only places where books can be found in order to scan and reprint them. Simply put: every book will be tailored according to the desires of the buyer, as happened with old manuscripts.

The second invention is the e-book where by inserting a micro- cassette in the book’s spine or by connecting it to the internet one can have a book printed out in front of us. Even in this case, however, we shall still have a book, though as different from our current ones as ours are different from old manuscripts on parchment, and as the first Shakespeare folio of 1623 is different from the last Penguin edition. Yet, up to now e-books have not proved to be commercially successful as their inventors hoped. I have been told that some hackers, grown up on computers and unused to browsing books, have finally read great literary masterpieces on e-books, but I think that the phenomenon remains very limited. In general, people seem to prefer the traditional way of reading a poem or a novel on printed paper. E-books will probably prove to be useful for consulting information, as happens with dictionaries or special documents. They will probably help students obliged to bring with them ten or more books when they go to school, but they will not substitute for other kinds of books that we love to read in bed before sleep, for example.

Indeed, there are a lot of new technological devices that have not made previous ones obsolete. Cars run faster than bicycles, but they have not rendered bicycles obsolete, and no new technological improvements can make a bicycle better than it was before. The idea that a new technology abolishes a previous one is frequently too simplistic. Though after the invention of photography painters did not feel obliged to serve any longer as craftsmen reproducing reality, this did not mean that Daguerre’s invention only encouraged abstract painting. There is a whole tradition in modern painting that could not have existed without photographic models: think, for instance, of hyper-realism. Here, reality is seen by the painter’s eye through the photographic eye. This means that in the history of culture it has never been the case that something has simply killed something else. Rather, a new invention has always profoundly changed an older one.
To conclude on this theme of the inconsistency of the idea of the physical disappearance of books, let us say that sometimes this fear does not only concern books but also printed material in general. Alas, if by chance one hoped that computers, and especially word processors, would contribute to saving trees, then that was wishful thinking. Instead, computers encourage the production of printed material. The computer creates new modes of production and diffusion of printed documents. In order to re- read a text, and to correct it properly, if it is not simply a short letter, one needs to print it, then to re-read it, then to correct it at the computer and to reprint it again. I do not think that one would be able to write a text of hundreds of pages and to correct it properly without reprinting it many times.
Today there are new hypertextual poetics according to which even a book-to-read, even a poem, can be transformed to hypertext. At this point we are shifting to question two, since the problem is no longer, or not only, a physical one, but rather one that concerns the very nature of creative activity, of the reading process, and in order to unravel this skein of questions we have first of all to decide what we mean by a hypertextual link.

Notice that if the question concerned the possibility of infinite, or indefinite, interpretations on the part of the reader, it would have very little to do with the problem under discussion. Rather, that would have to do with the poetics of a Joyce, for example, who thought of his book Finnegans Wake as a text that could be read by an ideal reader affected by an ideal insomnia. This question concerns the limits of interpretation, of deconstructive reading and of over-interpretation, to which I have devoted other writings. No: what are presently under consideration are cases in which the infinity, or at least the indefinite abundance of interpretations, are due not only to the initiative of the reader, but also to the physical mobility of the text itself, which is produced just in order to be re-written. In order to understand how texts of this genre can work we should decide whether the textual universe we are discussing is limited and finite, limited but virtually infinite, infinite but limited, or unlimited and infinite.
First of all, we should make a distinction between systems and texts. A system, for instance a linguistic system, is the whole of the possibilities displayed by a given natural language. A finite set of grammatical rules allows the speaker to produce an infinite number of sentences, and every linguistic item can be interpreted in terms of other linguistic or other semiotic items — a word by a definition, an event by an example, an animal or a flower by an image, and so on and so forth.

Take an encyclopaedic dictionary, for example. This might define a dog as a mammal, and then you have to go to the entry mammal, and if there mammals are defined as animals you must look for the entry animal, and so on. At the same time, the properties of dogs can be exemplified by images of dogs of different kinds; if it is said that a certain kind of dog lives in Lapland you must then go to the entry on Lapland to know where it is, and so on. The system is finite, an encyclopaedia being physically limited, but virtually unlimited in the sense you can circumnavigate it in a spiral-like movement, ad infinitum . In this sense, certainly all conceivable books are comprised by and within a good dictionary and a good grammar. If you are able to use an English dictionary well you could write Hamlet , and it is by mere chance that somebody did it before you. Give the same textual system to Shakespeare and to a schoolboy, and they have the same odds of producing Romeo and Juliet .

Grammars, dictionaries and encyclopaedias are systems: by using them you can produce all the texts you like. But a text itself is not a linguistic or an encyclopaedic system. A given text reduces the infinite or indefinite possibilities of a system to make up a closed universe. If I utter the sentence, “This morning I had for breakfast…”, for example, the dictionary allows me to list many possible items, provided they are all organic. But if I definitely produce my text and utter, “This morning I had for breakfast bread and butter”, then I have excluded cheese, caviar, pastrami and apples. A text castrates the infinite possibilities of a system. The Arabian Nights can be interpreted in many, many ways, but the story takes place in the Middle East and not in Italy, and it tells, let us say, of the deeds of Ali Baba or of Scheherazade and does not concern a captain determined to capture a white whale or a Tuscan poet visiting Hell, Purgatory and Paradise.

Take a fairy tale, like Little Red Riding Hood . The text starts from a given set of characters and situations — a little girl, a mother, a grandmother, a wolf, a wood — and through a series of finite steps arrives at a solution. Certainly, you can read the fairy tale as an allegory and attribute different moral meanings to the events and to the actions of the characters, but you cannot transform Little Red Riding Hood into Cinderella . Finnegans Wake is certainly open to many interpretations, but it is certain that it will never provide you with a demonstration of Fermat’s last theorem, or with the complete bibliography of Woody Allen. This seems trivial, but the radical mistake of many deconstructionists was to believe that you can do anything you want with a text. This is blatantly false.

Now suppose that a finite and limited text is organised hypertextually by many links connecting given words with other words. In a dictionary or an encyclopaedia the word wolf is potentially connected to every other word that makes up part of its possible definition or description (wolf is connected to animal, to mammal to ferocious, to legs, to fur, to eyes, to woods, to the names of the countries in which wolves exist, etc.). In Little Red Riding Hood , the wolf can be connected only with the textual sections in which it shows up or in which it is explicitly evoked. The series of possible links is finite and limited. How can hypertextual strategies be used to “open” up a finite and limited text?

The first possibility is to make the text physically unlimited, in the sense that a story can be enriched by the successive contributions of different authors and in a double sense, let us say either two-dimensionally or three-dimensionally. By this I mean that given, for instance, Little Red Riding Hood , the first author proposes a starting situation (the girl enters the wood) and different contributors can then develop the story one after the other, for example, by having the girl meet not the wolf but Ali Baba, by having both enter an enchanted castle, having a confrontation with a magic crocodile, and so on, so that the story can continue for years. But the text can also be infinite in the sense that at every narrative disjunction, for instance, when the girl enters the wood, many authors can make many different choices. For one author, the girl may meet Pinocchio, for another she may be transformed into a swan, or enter the Pyramids and discover the treasury of the son of Tutankhamen.
This is today possible, and you can find on the Net some interesting examples of such literary games.

AT THIS POINT one can raise a question about the survival of the very notion of authorship and of the work of art, as an organic whole. And I want simply to inform my audience that this has already happened in the past without disturbing either authorship or organic wholes. The first example is that of the Italian Commedia dell’arte , in which upon a canovaccio , that is, a summary of the basic story, every performance, depending on the mood and fantasy of the actors, was different from every other so that we cannot identify any single work by a single author called Arlecchino servo di due padroni and can only record an uninterrupted series of performances, most of them definitely lost and all certainly different one from another.

Another example would be a jazz jam session. We may believe that there was once a privileged performance of Basin Street Blues while only a later recorded session has survived, but we know that this is untrue. There were as many Basin Street Blues as there were performances of it, and there will be in future a lot of them that we do not know as yet, as soon as two or more performers meet again and try out their personal and inventive version of the original theme. What I want to say is that we are already accustomed to the idea of the absence of authorship in popular collective art in which every participant adds something, with experiences of jazz-like unending stories.

Such ways of implementing free creativity are welcome and make up part of the cultural tissue of society.

Yet, there is a difference between implementing the activity of producing infinite and unlimited texts and the existence of already produced texts, which can perhaps be interpreted in infinite ways but are physically limited. In our same contemporary culture we accept and evaluate, according to different standards, both a new performance of Beethoven’s Fifth and a new Jam Session on the Basin Street theme. In this sense, I do not see how the fascinating game of producing collective, infinite stories through the Net can deprive us of authorial literature and art in general. Rather, we are marching towards a more liberated society in which free creativity will coexist with the interpretation of already written texts. I like this. But we cannot say that we have substituted an old thing with a new one. We have both.
TV zapping is another kind of activity that has nothing to do with watching a movie in the traditional sense. A hypertextual device, it allows us to invent new texts that have nothing to do with our ability to interpret pre-existing texts. I have tried desperately to find an instance of unlimited and finite textual situations, but I have been unable to do so. In fact, if you have an infinite number of elements to play with why limit yourself to the production of a finite universe? It’s a theological matter, a sort of cosmic sport, in which one, or The One, could implement every possible performance but prescribes itself a rule, that is, limits, and generates a very small and simple universe. Let me, however, consider another possibility that at first glance promises an infinite number of possibilities with a finite number of elements, like a semiotic system, but in reality only offers an illusion of freedom and creativity.

A hypertext can give the illusion of opening up even a closed text: a detective story can be structured in such a way that its readers can select their own solution, deciding at the end if the guilty one should be the butler, the bishop, the detective, the narrator, the author or the reader. They can thus build up their own personal story. Such an idea is not a new one. Before the invention of computers, poets and narrators dreamt of a totally open text that readers could infinitely re-compose in different ways. Such was the idea of Le Livre , as extolled by Mallarmé. Raymond Queneau also invented a combinatorial algorithm by virtue of which it was possible to compose, from a finite set of lines, millions of poems. In the early sixties, Max Saporta wrote and published a novel whose pages could be displaced to compose different stories, and Nanni Balestrini gave a computer a disconnected list of verses that the machine combined in different ways to compose different poems. Many contemporary musicians have produced musical scores by manipulating which one can compose different musical performances.

All these physically moveable texts give an impression of absolute freedom on the part of the reader, but this is only an impression, an illusion of freedom. The machinery that allows one to produce an infinite text with a finite number of elements has existed for millennia, and this is the alphabet. Using an alphabet with a limited number of letters one can produce billions of texts, and this is exactly what has been done from Homer to the present days. In contrast, a stimulus-text that provides us not with letters, or words, but with pre-established sequences of words, or of pages, does not set us free to invent anything we want. We are only free to move pre-established textual chunks in a reasonably high number of ways. A Calder mobile is fascinating not because it produces an infinite number of possible movements but because we admire in it the iron-like rule imposed by the artist because the mobile moves only in the ways Calder wanted it to move.

At the last borderline of free textuality there can be a text that starts as a closed one, let us say, Little Red Riding Hood or The Arabian Nights , and that I, the reader, can modify according to my inclinations, thus elaborating a second text, which is no longer the same as the original one, whose author is myself, even though the affirmation of my authorship is a weapon against the concept of definite authorship. The Net is open to such experiments, and most of them can be beautiful and rewarding. Nothing forbids one writing a story where Little Red Riding Hood devours the wolf. Nothing forbids us from putting together different stories in a sort of narrative patchwork. But this has nothing to do with the real function and with the profound charms of books.

A BOOK OFFERS US A TEXT which, while being open to multiple interpretations, tells us something that cannot be modified. Suppose you are reading Tolstoy’s War and Peace : you desperately wish that Natasha will not accept the courtship of that miserable scoundrel Anatolij; you desperately wish that the marvellous person who is Prince Andrej will not die, and that he and Natasha will live together forever. If you had War and Peace on a hypertextual and interactive CD-ROM, you could rewrite your own story according to your desires; you could invent innumerable “War and Peaces”, where Pierre Besuchov succeeds in killing Napoleon, or, according to your penchants, Napoleon definitely defeats General Kutusov. What freedom, what excitement! Every Bouvard or Pécuchet could become a Flaubert!

Alas, with an already written book, whose fate is determined by repressive, authorial decision, we cannot do this. We are obliged to accept fate and to realise that we are unable to change destiny. A hypertextual and interactive novel allows us to practice freedom and creativity, and I hope that such inventive activity will be implemented in the schools of the future. But the already and definitely written novel War and Peace does not confront us with the unlimited possibilities of our imagination, but with the severe laws governing life and death.

Similarly, in Les Misérables Victor Hugo provides us with a beautiful description of the battle of Waterloo. Hugo’s Waterloo is the opposite of Stendhal’s. Stendhal, in La Charteuse de Parme , sees the battle through the eyes of his hero, who looks from inside the event and does not understand its complexity. On the contrary, Hugo describes the battle from the point of view of God, and follows it in every detail, dominating with his narrative perspective the whole scene. Hugo not only knows what happened but also what could have happened and did not in fact happen. He knows that if Napoleon had known that beyond the top of mount Saint Jean there was a cliff the cuirassiers of General Milhaud would not have collapsed at the feet of the English army, but his information in the event was vague or missing. Hugo knows that if the shepherd who had guided General von Bulow had suggested a different itinerary, then the Prussian army would have not arrived on time to cause the French defeat.
Indeed, in a role-play game one could rewrite Waterloo such that Grouchy arrived with his men to rescue Napoleon. But the tragic beauty of Hugo’s Waterloo is that the readers feel that things happen independently of their wishes. The charm of tragic literature is that we feel that its heroes could have escaped their fate but they do not succeed because of their weakness, their pride, or their blindness. Besides, Hugo tells us, “Such a vertigo, such an error, such a ruin, such a fall that astonished the whole of history, is it something without a cause? No… the disappearance of that great man was necessary for the coming of the new century. Someone, to whom none can object, took care of the event… God passed over there, Dieu a passé .”

That is what every great book tells us, that God passed there, and He passed for the believer as well as for the sceptic. There are books that we cannot re-write because their function is to teach us about necessity, and only if they are respected such as they are can they provide us with such wisdom. Their repressive lesson is indispensable for reaching a higher state of intellectual and moral freedom.

I hope and I wish that the Bibliotheca Alexandrina will continue to store this kind of books, in order to provide new readers with the irreplaceable experience of reading them. Long life to this temple of vegetal memory.