公元前3世纪的中国:真理，谎言和互联网

原文载于：《大西洋》月刊2011年12月29日

                    原文作者：Rebecca J. Rosen（丽贝卡?J.罗森）

     丽贝卡?J.罗森是《大西洋》月刊助理编辑。她曾经做过《威尔逊通报》季刊的助理编辑，在那里她主持杂志“本质”栏目。

       上个礼拜，一条令人震惊的消息帖爆微博：为了一双新耐克运动鞋扭打，一个年轻人丧生。

     《巴尔的摩太阳报》的记者在调查此事时，却发现这是一个蓄意编造的故事。的确，运动鞋曾吸引了一群人观看，但是没有人丧生。这是传统媒体惯用的得意伎俩，旧时的侦探打个电话就能查明真相。

     Slashdot资讯科技网发了一条消息：“网络暴民发假信息时我们应该怎么办？”发帖人问道：“毕竟，即便有人发现信息是错误的，他的声音也难以压倒错误信息的鼓噪。甚至有人声称，谁要是质疑这一信息，他就是无耻。对付这种网络暴民，难道我们就没有办法？怎样才能坚持良好的新闻传统，以正压邪，保持新闻的准确性？”

       现在不良信息充斥网络。也正是因为有了互联网，不良信息才得以快速传播。但是很清楚，也正是互联网使查证事实变得容易，使真实信息传播更广。哥伦比亚大学研究生卢克斯? 格拉夫指出，网络求证事实手段正在兴起，现在线网络平台的“求证能力”是2004—2010年的两倍多。正如一位资讯科技网的网民所说，任何虚假资讯都会在评论中受到反驳，一个非同寻常的网络评论系统会作出裁定，资讯科技网和维基网等机构的“事实求证”系统或网民提供的正确信息就会在那里出现。当一个叫作“事实核实” （Politifact）的查证机构把它确定为当年谎言时，其内容就会在网络到处可见。

       更是人感兴趣的是，既然如此，那么谷歌网和拆谎网（Snopes）为什么还会发布错误信息？

       有一种说法，真的，是神话般的说法，人类的脑子里接收到一个信息，然后由这个信息推论产生出想法和意见。但是如果是按照常规的方式理解，就会落入事前已经想好的陷阱。比如确认偏误（一种吸纳支持自己已有想法的倾向）就是这种理解方式，用这种方式去理解，人类的逻辑推论能力就会变得很低下。梅谢尔和斯佩贝尔在说到为什么人的逻辑推论能力那么低下时说，不是因为信息不足，而是因为我们使用了错误标准去衡量。人脑推论的存在，不是用来为我们提供外部世界准确图画的，而是用来构建和获得话语的，用梅谢尔和斯佩贝尔的话所说就是“争论”。人的头脑在他人说话捕捉瑕疵，要比从自己说话中捕捉瑕疵的能力强得多。在各样测试中可以看出，这一现象在两个人或一组人对话时，要比一个人独白时更为明显。

       正因为我们喜欢用另一种方式思维，事实就不再是我们了解外部世界的核心，起码按照先验方式思维是这样的。它们只是一种反驳其他人的工具而已。

 这就是为什么互联网带来了事实求证黄金时代的原因：互联网是辩驳的最佳媒介，而事实则是辩驳的生命线。以拆谎网为例，它帮助我们杜绝过快转发邮件，但却很少有人浏览拆谎网，去了解那些信口开河的东西。

       梅谢尔和斯佩贝尔关于人脑是如何获取新闻和政治交流准确价值核心的理论是：我们把“价值核心”它当做可靠性的代名词，因为实实在在的错误是相反评论的最好靶标。不管你是否有充分事实，关键是我们如何来评价你，无论你是一个政治家，一家出版机构或者是评论员。但不幸的是，准确性从来都不是质量的代名词。质量是更难得到的东西。

       我曾经做过事实求证员，曾仔细审查过数十万字文本中的拼写错误和错用数字。几乎总是这样，当事实出现错误时，你可以纠正它，而不需要去对评论中的字做一一修改。你要做的是删掉那些没有事实根据的片段，输入正确的。但是有些片段，数字是正确的，而事实却是大错而特错的。“从无处看世界”式的新闻或纷纷扰扰似是而非的报道可能是很尖锐的，但是它们却无助于解释问题和意见，更不用说它不能激发情感和同情了。

       我们一直都相信，真理会重现，真相会大白。我们将会得到更好的信息，作出更正确的决定，选举更好的领导，生活过得更美好。这是一幅大有希望的画卷，但它和人们所做的（世界所作的）却不一致。当你把梅谢尔和斯佩贝尔的理论牢记在心时，你就会理解，是记述、主意和思想，而不是事实，才能为世界提了供动力。对资讯科技网的帖子来说，喜讯是互联网正在滋长着准确性，坏消息则是准确性距离我们还很遥远。

 Truth, Lies, and the Internet

 Dec 29 2011, 2:23 PM ET 23

 Rebecca J. Rosen - Rebecca J. Rosen is an associate editor at The Atlantic. She was previously an associate editor at The Wilson Quarterly, where she spearheaded the magazine's In Essence section

 Last week, a horrifying story floated around Twitter: A young man had been killed in a scuffle over a new pair of Nikes.

 But when two Baltimore Sun reporters looked into it, they found that the story was a fiction. The sneakers attracted crowds, but no one was killed. It was a triumph for the traditional media, the ancient breed of sleuths who call people on the telephone and find out the truth.

 A commenter on the tech site Slashdot picked up the thread: "What do we do when the Internet mob is wrong?" the poster asked:

 After all, if one of the crowd discovered the error, the signal would barely rise above the noise. There are people claiming that anyone questioning the facts is being disrespectful. Is there something we can do about the mobocracy? How can we support the best traditions of journalism while fixing the worst? How can we nurture accuracy?

 Sure, there is bad information all over the Internet, and because of the Internet, it can spread more rapidly. But it's also clear that the Internet is making fact-checking easier and more widespread than ever. Lucas Graves, a doctoral candidate at Columbia, notes that fact-checking is on the rise; mentions of "fact check" more than doubled in Nexis between 2004 and 2010. And, as one Slashdotter writes, any wrong Slashdot piece will be disproved in the comments, voted up in the site's unique commenting system. The good information is out there, whether it's provided by institutions like FactCheck.org or the denizens of sites such as Wikipedia or Slashdot. And when the fact-checking shop Politifact royally screws up its Lie of the Year, the rebuttals are everywhere.

 A more interesting question is why, in this age of Google and Snopes, does misinformation persist? As a few of the Slashdot commenters note, plenty of urban legends that have been eminently checkable on Snopes for years continue to circulate. I suspect this can at least be partially explained by an intriguing theory of how the mind works, advanced last spring by Hugo Mercier and Dan Sperber, two cognitive scientists.

 There is a belief -- a myth, really -- that the human mind takes in information, and then reasons through it to produce ideas and opinions. But humans are notoriously poor at reasoning as it is conventionally understood, predictably falling into known traps, such as the confirmation bias (the tendency to absorb information that supports what one already thinks). Mercier and Sperber argued that the explanation for why human reasoning is so poor isn't because it's deficient, but because we've measured it against the wrong standard. Human reason doesn't exist to provide us with a more accurate picture of the world; it exists to structure and promote discourse, or what Mercier and Sperber term "argument." The human mind is better at spotting the flaws in someone else's argument than its own, and in groups or pairs can do much better on a variety of tests than when flying solo.

 As much as we like to think otherwise, facts, at least according to this schema, aren't at the core of how we understand the world, but they sure are useful for rebutting the way other people do.

 This is the reason why the Internet has brought a Golden Age of Fact Checking: The Internet is a medium perfect for rebuttals, and facts are the lifeblood of rebuttals. Snopes, for example, helps us refute a too-quickly-forwarded email, but few people would just browse Snopes to learn random things.

 Mercier and Sperber's theory about how the mind works gets to the core of why we value accuracy in journalism and political communications: We use it as a proxy for credibility, because factual mistakes are the easiest targets for an argument to the contrary. Whether or not you have your facts straight is how we judge you -- as a politician, a publication, or a pundit.

 But this is unfortunate, because accuracy is not always a good proxy for quality. Quality, however, is much more difficult to assess.

 I have been a fact checker. I have scrutinized tens if not hundreds of thousands of words of text for misspellings and misplaced digits. Almost always, when a fact is wrong, you can correct it without so much as changing a word of the surrounding argument. What you're doing is inoculating the piece from the charge of not having the facts straight. But the piece can be just as wrongheaded once the numbers are correct. View-from-nowhere journalism or he-said-she-said reporting can be entirely accurate, but do little to help explain an issue or an idea, to say nothing of inspiring empathy or compassion.

 We continue to believe that the truth will out and the facts will save us. We will have better information and make better decisions, elect better leaders, have better government, be better off. That's a pretty hopeful picture but not one that's in line with how humans -- or the world -- work. When you take Mercier and Sperber's theory to heart you understand that narratives, ideas, and ideologies are what fuel the world, not facts. For the Slashdot poster, the good news is that the Internet is nurturing accuracy. The bad news is that accuracy only takes us so far.