星星之火,何时燎原?Minority Rules?

作者: Sophie_娘子 | 来源:发表于2018-07-12 21:44 被阅读223次

    范仲淹在《岳阳楼记》的结尾发出千古一叹,:“噫!微斯人,吾谁与归?”

    这句感慨,半是苦恼,半是庆幸:“如果不是你,还有谁与我志同道合呢?”幸亏有你,让我不至孤单;但只有你我,也还是太少了,那些庙堂之高的大事,还是需要更多的声音和力量啊……

    这大概是所有仁人志士的心中所叹。年幼的时候曾经怀揣着孩子般的天真幻想着以一己之力改变这个世界。渐渐大了,懂得积少成多、众人拾柴火焰高的道理,交志同道合的朋友,汇聚更多的声音。

    然而一个新运动、新思想的萌芽和推广,到底需要几个志同道合的朋友?需要发出多大的声音?星星之火,要蔓延多久,才能有燎原之势?

    这期《Science》宾夕法尼亚大学的Centola 教授即致力于这一研究。她的课题组在互联网上招募志愿者(所有活动都在互联网上进行,志愿者并不互相见面),并将志愿者任意分组,随机配对。被配对的志愿者会看到一张任意图片(如下图)。两人在可选择的范围内给图片一个名字,名字相同会得到金钱上的奖励,名字不同则会有同等的惩罚。一轮结束后则再次随意配对。如此轮转,在金钱利益的驱使下,志愿者逐渐达成一致,每张图片开始有固定的名字。

    游戏规则

    此时,课题组的成员也开始加入游戏,作为“顽固少数派”,他们会每次坚持给图片一个与共识不同的名字。可想而知,若是在一轮游戏中被随机与研究者配对,那么两方则会给出迥异的答案,也会得到相应的惩罚。

    Centola 提出的问题是,“顽固少数派”需要占据实验大众多少的百分比,才能改变大众的想法,从而使大家达到新的共识?

    答案是:24.3%.

    “顽固少数派”改变大众共识的最少人数比

    从图中可以看出,当“顽固少数派”(灰色)的人数低于24.3%时,多数派(黑色)并没有偏离已经确认的共识。而当少数派多于25%时,多数派逐渐达成新的共识。

    这是一个有趣的数据,基本是四分之一的百分比。很难想象一个单位、社区、学校、团体能够有四分之一的坚定不移的少数派。四分之一的女权铁杆支持者?四分之一的同性恋平权者?四分之一的宗教改革者?四分之一的革命者?这似乎是对“少数派”的略显苛刻的要求。

    与此同时,我也有些怀疑,此实验还是低估了“多数派”的顽固。在现实生活中,人们对于一些原则、法律、信仰、现状的坚持,其驱动力往往高于金钱和利益,是受很多不同社会因素影响的:成长环境、舆论环境、个人道德观和价值观等。相较于少数派,他们所承受的压力更小,也因此更加容易坚持自己的立场。而实验中的志愿者,毫无负担、思想包袱、价值原则,可以任意改变一个图片的名称。由此看来,四分之一的百分比也许还是低估了。

    另一件值得思考的事情是,并不是所有的少数派,在人数相等的情况下,都会发出同等大小的声音。如同性恋者应该远不及总人数的四分之一,然而我们却一直清晰地听到他们的呐喊;女人是人类的一半,可女权仍然任重而道远。同样,基督教和伊斯兰教相比、审美观的多元性和单一性相比、亚洲人与黑人相比,存在感相差非常之大。有些“少数派”可以未见其人、先闻其声,没有四分之一的人数、却有四分之一的架势。难怪连五毛党都上了《Science》,有些话重复的多了、分贝大了,别人信了,大概连自己也信了。

    路漫漫其修远兮,吾将上下而求索。思想进步的过程是漫长的(但并不是所有的“改变”都可称之为“进步”)。少数派从被打压、被质疑,到积少成多,后来又打压、质疑他人,来回轮转。或许社会学给我们的思考就是:怎样处多数、又怎样处少数。酌贪泉而觉爽,处涸辙以犹欢。

    噫!微斯人,吾谁与归?


    I don’t know how many people have read the children’s book Frindle. Anyhow, if you haven’t, I’m about to spoil it for you.

    Cover of book “Frindle” by Andrew Clements

    Nick is an obnoxious fifth grader who decides to question the origin of every word to stall time in Mrs. Granger’s classroom. After hearing Mrs. Granger’s explanation, Nick decides to create a new English word, “Frindle”, to replace one that already existed, “pen”.

    The idea was taken up by other equally obnoxious fifth graders, and Mrs. Granger, who had a deep and lasting respect for the English language, and for the word “pen”, is understandably annoyed. She starts to hand out detention to the children heard using the word “Frindle”, which seems to strangely excite the students even more and goad them into flaunting the word in her presence. Soon the whole school is put in detention, to the exasperation of parents and bus drivers, who threatened to go on strike. The whole town was suitably amused, both by the children and also by the almost theatrical rage on the part of Mrs. Granger. The word “Frindle” started showing up everywhere as an inside joke.

    In the epilogue of the book, Mrs. Granger sent the grown Nick a gift along with a letter. The gift was the newest edition of the dictionary, edited to include the word “frindle”, meaning, “pen”, and the letter explained that she had willingly played the villain to the word to aid its spread like wildfire.

    As children, the book had certainly been fun to read. It’s always exciting to make trouble, create something new, and get in the teacher’s hair while at it. But as adults, this remains an interesting phenomenon worth examining: what does it take to change the status quo? How many does it take? What is the critical tipping point? What is the price of change?

    A team led by Penn professor Damon Centola set out to address this problem. The experimental set up was easy enough. Volunteers were recruited to take part in online tests where they were randomly put into groups and then randomly paired. Each pair was shown a picture, in which case each participant gave the picture a name simultaneously. If the participants agreed, both were rewarded monetarily. If they disagreed, they were penalized. Although no global "consensus" was outwardly encouraged, players eventually reached a convention based on pair-wise interactions, stimulated by financial gains.

    Then researchers joined the game as "committed minorities", hoping to overthrow the convention by giving the picture a novel name. Critical mass was determined as the minimal percentage of "committed minorities" required to tip the old convention in favor of the new. After experimenting with various groups of different sizes, the critical mass was established as 24.3% of the population.

    Time series showing adoption of the alternative convention by noncommitted subjects

    As is shown in the graph above, when the committed minority (gray) stays below 25%, the established convention (black) does not change. When committed minorities surpass 25%, a new consensus is eventually reached.

    24.3% actually seems to be an unoffending and unsurprising number, in my opinion. In actuality, minority groups that fluctuate around this percentage certainly have their voices heard. However, just qualitatively thinking, some issues, like marriage equality, language/culture recognition, seem to have more weight even when their proponents account for far less than 24.3% of the population; while issues such as gender equality seem to make poor progress even when the percentage of its proponents reach 40% or more. Of course, the study assumes that minorities are diehard and resistant to change, thus "committed". However, in real life interactions, people are under a lot more influences and pressures, and may wish to navigate a middle ground and not draw harsh lines. How do these "not-entirely-committed" minorities affect the critical mass? In addition, there also exists extremists who are steadfast in their beliefs and very vocal about pronouncing them. Does such an individual maybe have the effect of multiple "committed minorities"? These two scenarios could easily be accounted for in the experimental set up.  "Committed minorities" could change their answers to agree with convention every once in a while, thus accounting for the "not-entirely-committed", which to be honest, seems to be closer to average human behavior; vocal minorities can be modeled by maybe tipping the financial reward in their favor, e.g. receiving a bigger reward when agreeing with extremists and a bigger punishment when disagreeing with them compared to the average public.

    Something that's not so easily modeled, however, is that people's beliefs and convictions in life are hardly ever driven by trivial monetary gains. Faith is perhaps defined by the ability to hold onto things when circumstances are tough. Whereas the mass majority will change their names for a picture to gain something in an experiment, it is hardly likely that they will so easily change their minds when confronted face to face. Human memory, history, self-pride, identity, dignity. All are mixed into how we see the world and see issues, and these may be the most resistant to change. As to internet interactions, the internet has allowed us to expand from only interacting with each other in person, in the here and now, to being in touch with literally the whole word, while hiding under the sweet cloak of anonymity. Online, one can change one's opinions at will, without the complications of pride and embarrassment one might have when dealing with friends and acquaintances face-to-face. One might expect the model then, to be more reflective of internet interactions as opposed to face-to-face ones. However, I find it just happens that it is on the internet where people speak their minds the loudest and become the most extreme, sometimes shedding even the pretence of decency. Anyone who’s had an argument with a stranger online know that it is almost futile to attempt changing their minds.

    So, how to get a new word you created to be included in the dictionary? In a case as inoffensive as this, the answer seems to be make lots of friends, and make sure that at least 1 in 4 of them is committed to using the new word. As to social, political, religious and ideological changes, I feel the model falls short.

    Reference

    Centola, Damon & Becker, Joshua & Brackbill, Devon & Baronchelli, Andrea. (2018). Experimental evidence for tipping points in social convention. Science. 360. 1116-1119. 10.1126/science.aas8827.

    相关文章

      网友评论

      • 史东:使用数学的方法,去研究社交概率的时候,是很有意思的。可惜的是,用的是完全陌生人模式。如果要想体现出复杂的人际模型所带来的影响,道路还很漫长。
        Sophie_娘子:@史东 嗯,是啊。这个模型更适合模拟网上的互动:一个人有多大可能被陌生网友改变?需要多少个陌生网友众口一词?模拟现实生活的话,局限性很大…

      本文标题:星星之火,何时燎原?Minority Rules?

      本文链接:https://www.haomeiwen.com/subject/bmcdpftx.html