JOVANA
Library Glossary Getting Started Three Levels Fields How it works Mission
Join the mission
Back to the library
數學 1684

求極大值與極小值的新方法

戈特弗里德·威廉·萊布尼茲

寥寥幾個符號——dx、dy——把「變化」本身變成了可以演算的代數。

Choose your version
In depth · the introduction

在萊布尼茲之前,求一條曲線的陡度,是一道只能一條曲線一條曲線去攻的難題。他把它變成了一套幾乎人人都能照著做的「配方」。

核心想法

微積分是「變化」的數學。它的一半,問的是一個聽起來很簡單的問題:一條曲線在某一個點上有多陡?對一段筆直的斜坡,答案很容易——上升量除以前進量。可真實的曲線是彎的,所以每一個點的陡度都不一樣。

萊布尼茲的妙招是「放大」。任何一條光滑曲線,只要看足夠小的一小段,它看上去就是直的。把那一小步橫向記作 dx,把隨之而來的那一點點上升記作 dy;那麼這個點的陡度,就只是 dy 除以 dx。而他真正的天才之處,是給出了一小串把這些微小片段組合起來的法則——於是你再也不必親手去「放大」,只要演算就行。

它是如何誕生的

萊布尼茲是律師、外交官,也是哲學家,數學是他很晚才自學、卻學得極出色的本事。1670 年代在巴黎,他不但理出了微積分的兩半,更同樣重要地,定下了我們至今仍在用的那套清爽記號——dx、dy,以及後來的積分號 ∫。1684 年 10 月,他把它發表了出來:在德國期刊《學者紀要》上,密密麻麻的六頁。

在海峽對岸,艾薩克·牛頓早約二十年就已得到一套等價的方法,卻大體只藏在自己手裡。當萊布尼茲率先發表,一場關於「功勞歸誰」的激烈爭吵隨之爆發。皇家學會的一個委員會——背後由牛頓暗中操控——判定萊布尼茲剽竊。如今歷史學家一致認為,兩人是各自獨立發現的。而今天的學生學到的,是萊布尼茲那套整潔的符號,不是牛頓的點。

它為何重要

一旦你能算出任何東西變化得有多快,你就能描述運動、熱、生長、電,甚至金錢。微積分成了物理與工程共同的語言:自然的規律,大多寫成關於「變化速率」的方程式,而這些方程式,都用著萊布尼茲的那個 d。幾乎沒有哪座橋、哪台引擎、哪艘太空船、哪個經濟模型,能離得開它。

一個可以想像的畫面

想像你夜裡開車上一座盤山路,車燈只照亮前方一公尺。在那被照亮的一公尺裡,路看起來就是一段筆直的斜坡,它的陡度,不過是「你爬升了多少」除以「你前進了多少」——這就是 dy 除以 dx。被照亮的那一段越短,它告訴你的、關於這個確切位置的陡度,就越精確。萊布尼茲的微積分,正是當那段亮路縮到無窮小時,關於這個比值的規則手冊。

可互動的曲線:在拋物線、三次曲線或正弦波中選一條,沿曲線滑動一個點並縮小間隔 Δx;一個小三角形(橫邊 dx、豎邊 dy)緊貼曲線,穿過兩點的割線一路轉動,直到與切線吻合,其斜率落到曲線真正的陡度上。

它的位置

笛卡兒(1637)讓代數與幾何聯姻,費馬也有一套求切線與極值的方法——但真正鑄成這件通用工具的,是萊布尼茲,以及獨立地,牛頓。從這裡,一條線徑直通向牛頓的《原理》,通向歐拉,通向本館中的每一個微分方程式:傅立葉的熱、馬克士威的場、薛丁格的波。現代科學中,每當有什麼在變化,描述它用的,都是萊布尼茲在這裡寫下的語言。

The original document
Original source text
G. W. Leibniz · Nova Methodus pro Maximis et Minimis · Acta Eruditorum (Oct. 1684): 467–473
Nova methodus pro maximis et minimis, itemque tangentibus, quae nec fractas nec irrationales quantitates moratur, et singulare pro illis calculi genus
(“A new method for maxima and minima, as well as tangents, which is impeded neither by fractional nor by irrational quantities, and a remarkable type of calculus for them.”)
Leibniz takes a curve referred to an axis, with ordinates v, w, y, z, and tangents VB, WC, YD, ZE meeting the axis at B, C, D, E. He then introduces his central device: a fixed but arbitrary increment, dx, against which every other small difference is measured.
Now some right line taken arbitrarily may be called dx, and the right line which shall be to dx, as v (or w, y, z, resp.) is to VB (or WC, YD, ZE, respect.) may be called dv (or dw, dy, dz, resp.), or the differentials.
Because dv stands to dx as the ordinate stands to its subtangent, the ratio dy/dx is, by definition, the slope of the tangent — and the little right triangle with legs dx and dy, its hypotenuse lying along the tangent, becomes the engine of the whole method.
From this single definition Leibniz states the rules of the calculus, in a form every student still learns: the differential of a constant is zero; d(ax) = a·dx; differentials carry through sums and differences, so d(z − y + w + x) = dz − dy + dw + dx; for a product, d(xv) = x·dv + v·dx; for a quotient, d(v/y) = (y·dv − v·dy)/y²; and for powers and roots, d(xⁿ) = n·xⁿ⁻¹·dx — and, he stresses, these hold whether the exponent is whole, fractional, or irrational.
He then reads meaning into the signs. Where the ordinates stop increasing and turn to decrease (or the reverse), the curve reaches a maximum or a minimum, and there the differential dy is nothing — zero. Where the curve passes from concave to convex, the point of inflection, it is the differences of the differences that change sign.
[ … ]
To show the method's reach, Leibniz closes by solving a problem posed by Florimond de Beaune — to find a curve whose subtangent is everywhere constant — and arrives at the logarithmic curve, a transcendental curve that ordinary algebra cannot capture. The same calculus, he notes, opens problems that had resisted every earlier approach.
Acta Eruditorum · Leipzig · October 1684