Does Science Ever Change Its Mind? 科學會改變想法嗎？

🔊This was good science. It was mathematical, it was tested against observation, and it was improved by generations of astronomers in Greek, Arabic, and Latin. In 1543 Nicolaus Copernicus published a book that put the sun at the center and set the earth moving. He knew what to expect. “I can easily conceive, most Holy Father,” he wrote to the pope, “that as soon as some people learn that in this book … I ascribe certain motions to the Earth, they will cry out at once that I and my theory should be rejected.” He had kept the manuscript back for decades.

這是好的科學。它是數學的；它拿觀察來檢驗；一代又一代用希臘文、阿拉伯文與拉丁文寫作的天文學家改進了它。1543 年，尼古拉・哥白尼出版了一本書，把太陽放在中心，讓地球動了起來。他知道會遇到什麼。他寫給教宗：「至聖的聖父，我很容易想像得到：一旦有些人得知，我在這本書裡……把某些運動歸給了地球，他們會立刻大喊，說我和我的理論都應該被摒棄。」這份手稿，他壓了幾十年。

🔊In 1962 the historian Thomas Kuhn asked how such changes come about, and gave an answer that surprised philosophers. Most science, he said, is not a search for new theories. It is work done inside a paradigm, by which he meant a shared achievement that tells a community what the world contains, which questions are worth asking, and what a good answer looks like. Ptolemy’s astronomy was a paradigm, and Newton’s mechanics was another. Within one, scientists do what Kuhn called normal science. They solve puzzles, measuring a constant more exactly or extending the theory to a new case. They do not test the paradigm. They assume it, and when an experiment disagrees with it they suspect the experiment.

1962 年，歷史學家湯瑪斯・孔恩問：這樣的改變是怎麼發生的？而他給出的答案讓哲學家很意外。他說，大部分的科學並不是在尋找新的理論。它是在一個「典範」之內所做的工作。他說的典範，是一項共同的成就；它告訴一個社群：世界裡有些什麼、哪些問題值得問，以及一個好的答案長什麼樣子。托勒密的天文學是一個典範；牛頓的力學是另一個。在典範之內，科學家做的是孔恩所謂的「常態科學」。他們解謎：把一個常數量得更精確，或者把理論推廣到一個新的情況。他們並不檢驗典範。他們預設它；而當一個實驗與它不一致的時候，他們懷疑的是實驗。

🔊Kuhn thought this conservatism necessary. Only people who trust a framework will push it far enough to find where it breaks. Results that will not fit, which he called anomalies, pile up slowly. For a long time they are set aside. When there are too many of them, or one of them touches something central, the field enters a crisis. Rival ideas appear, and the young begin to take them seriously.

孔恩認為這種保守是必要的。只有信任一個框架的人，才會把它推得夠遠，遠到發現它在哪裡斷裂。怎麼也兜不攏的結果，他稱為「異常」；它們慢慢地堆積起來。有很長一段時間，它們被擱在一邊。等到異常太多，或者其中有一個碰到了核心的東西，這個領域就進入「危機」。對立的想法出現了；而年輕人開始認真看待它們。

🔊A revolution, for Kuhn, is the replacement of one paradigm by another, and it does not happen the way a theorem is proved. The two sides do not share enough to settle the matter by a single experiment. Copernicus’s system was at first no more accurate than Ptolemy’s. It still used epicycles, since Copernicus kept the old belief that heavenly motion must be circular. It also contradicted the best physics of the day, which said that a moving earth would leave falling stones and clouds behind. What it offered was a kind of order. In Ptolemy’s system every planet’s epicycle was mysteriously tied to the sun’s yearly motion. Copernicus explained this at a stroke: the yearly circle is ours. The backward loop of Mars is what we see when the faster earth overtakes it.

對孔恩來說，一場革命就是一個典範被另一個取代；而它發生的方式，並不像證明一條定理。雙方共有的東西不夠多，沒辦法用單一的實驗把事情定下來。哥白尼的體系一開始並不比托勒密的更準。它仍然用到本輪；因為哥白尼保留了那個舊信念：天體的運動必定是圓的。它還牴觸了當時最好的物理學；那套物理學說，如果地球在動，落下的石頭與天上的雲就會被拋在後面。它提供的是一種秩序。在托勒密的體系裡，每一顆行星的本輪，都莫名其妙地跟太陽每年的運動綁在一起。哥白尼一舉解釋了這件事：那個一年一圈的圓，是我們自己的。火星倒退著打的那個圈，是比較快的地球追過它的時候，我們所看到的景象。

🔊It took more than a century, and the work of Kepler, Galileo, and Newton, before the new system was plainly better by every measure. During that time, Kuhn argued, choosing it was a judgment that the evidence did not force. He compared the change to the drawing that can be seen as a duck or as a rabbit. The lines are the same, and what one sees in them has shifted. Scientists on either side of a revolution, he wrote, work in different worlds. The physicist Max Planck had said something harsher from his own experience: “A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die, and a new generation grows up that is familiar with it.”

要等到一個多世紀之後，經過克卜勒、伽利略與牛頓的工作，新的體系才在每一個標準上都明顯比較好。孔恩主張，在那段期間，選擇它是一個判斷，而證據並沒有強迫人做出這個判斷。他把這個改變比作那幅既可以看成鴨子、也可以看成兔子的圖。線條是一樣的；人在線條裡看到的東西變了。他寫道，革命兩邊的科學家，是在不同的世界裡工作。物理學家馬克斯・普朗克根據自己的經驗，說過更嚴厲的話：「一個新的科學真理得勝，並不是靠著說服它的反對者、使他們看見光明；而是因為它的反對者終於死了，而熟悉它的新一代長大了。」

🔊China had its own astronomy, with careful records of eclipses, comets, and “guest stars” going back two thousand years, and its own rival pictures of the sky. In the 1600s Jesuit missionaries brought European methods to the court in Beijing and showed that they predicted eclipses more accurately. In 1644 the German Jesuit Adam Schall was put in charge of the imperial calendar. Twenty years later an official named Yang Guangxian brought charges against him and the Chinese astronomers who worked with him. Yang’s objection was not mainly about accuracy. He is remembered for the sentence: “It is better for China to have no good calendar than to have Westerners in China.” Schall was condemned, and five Chinese astronomers were executed. Yang took over the bureau, his predictions failed, and in 1669 a public contest of predictions restored the Jesuits’ methods.

中國有自己的天文學，有兩千年來對日月食、彗星與「客星」的仔細紀錄，也有自己幾種互相對立的天體圖像。十七世紀，耶穌會的傳教士把歐洲的方法帶到北京的宮廷，並且顯示它們預測日月食更準確。1644 年，德國耶穌會士湯若望受命掌管欽天監。二十年後，一位名叫楊光先的官員，控告他以及與他共事的中國天文官。楊光先的反對，主要並不在準確與否。後人記得他的這句話：「寧可使中夏無好曆法，不可使中夏有西洋人。」湯若望被定罪；五位中國的天文官被處決。楊光先接掌了欽天監；他的預測失敗了；1669 年，一場公開的預測比試，恢復了耶穌會士的方法。

🔊The episode shows both of Kuhn’s points and one that tells against him. A way of doing astronomy was bound up with institutions, loyalties, and a view of the world, and it was defended on those grounds. Yet the two sides could agree on a test, the length of a shadow at noon on a named day, and the result decided the matter for the emperor. Paradigms may be different worlds. They can still be made to meet at particular points.

這段插曲同時顯示了孔恩的兩個論點，以及一個不利於他的論點。一種做天文學的方式，與制度、效忠的對象，以及一種世界觀綁在一起；而人們是基於這些理由來捍衛它的。然而雙方還是能對一項檢驗達成共識：在指定的那一天，正午的日影有多長；而結果為皇帝把事情定了下來。典範也許是不同的世界。它們仍然可以被安排在某些特定的點上相遇。

🔊Kuhn’s book made many readers conclude that science is a succession of fashions and that no paradigm is truer than another. He denied this for the rest of his life. Later theories, he said, are better at solving puzzles than earlier ones: more accurate, wider in scope, simpler, and more fruitful of new results. These values are shared across revolutions, and they are why science does not go round in circles. What he would not say is that later theories are closer to the truth about what the world is really like, since he saw no way to compare a theory with the world except through another theory.

孔恩的書使許多讀者得出結論：科學只是一連串的流行；沒有哪個典範比另一個更真。他在餘生裡一直否認這一點。他說，後來的理論比先前的理論更善於解謎：更準確、範圍更廣、更簡單，也更能結出新的成果。這些價值是跨越革命而共有的；而它們就是科學不會原地打轉的原因。他不肯說的是：後來的理論更接近「世界真正是什麼樣子」的真相；因為他看不出有什麼辦法，能不經由另一個理論，就拿一個理論去跟世界比較。

🔊That last step is still argued over. Realists point out that theories which survive a revolution usually keep the successes of the old ones as special cases, as Einstein’s equations give Newton’s at low speeds, and that this looks like getting nearer to something. Their opponents reply with history: most past theories that were once successful are now thought false, so we should expect the same of ours. Both sides accept that science changes its mind. It does so rarely, slowly, and under pressure, and when it has changed it can do more than it could before.

最後這一步，至今仍在爭論。實在論者指出：在革命中存活下來的理論，通常把舊理論的成功保留為特例，就像愛因斯坦的方程式在低速下給出牛頓的方程式；而這看起來像是在接近某個東西。他們的對手用歷史來回應：過去曾經成功的理論，大多數現在被認為是假的；所以我們應該預期，我們的理論也會一樣。雙方都接受：科學會改變想法。它很少這樣做，做得很慢，而且是在壓力之下；而改變之後，它能做的事比以前多。