导数 | 符号历史 / 理论框架物理拓展

注：机翻，未校。

---

A ‘Symbolic’ History of the Derivative

导数的 “符号化” 历史

克莱门西・蒙特尔

Clemency Montelle

新西兰坎特伯雷大学

University of Canterbury New Zealand

1 引言

How often do you have to tell your students to brush up on their notation? When they have dropped limit notation, forgotten critical modulus signs, mixed up their integrals, muddled up their derivatives, how do you convey to them the importance of recording it right?

你多久就得提醒学生们复习一下数学符号的使用？当他们遗漏极限符号、忘记关键的绝对值符号、混淆积分和导数时，你如何向他们传达正确书写符号的重要性呢？

What of your exasperation as they fail to appreciate the precision that mathematical notation affords them—notation which has been developed and refined over centuries, and notation that will continue to be improved for centuries more. Indeed, mathematics is the one subject in which they can really express exactly what they mean. How can we help them appreciate the symbolism they use?

当他们意识不到数学符号所带来的精确性时，你会有多恼怒啊！这些符号经过了几个世纪的发展和完善，而且在未来的几个世纪里还会继续改进。实际上，数学是一门能让他们精确表达自己想法的学科。我们怎样才能帮助他们理解所使用的符号呢？

This paper provides a light-hearted look at various notational conventions concerning the derivative. It will briefly cover the various proposals for its symbolic representation over history and the reasons behind the prevalence of the various manifestations over others. The examination of the history of the development of calculus notations is not only fascinating, but suggests a paradigm for the development of future notations. An examination of the reasons behind the failures and successes of past notation can equip one with a certain foresight regarding the future of newly introduced symbols, as particular areas of mathematics expand.

本文以轻松的视角探讨导数的各种符号表示约定。它将简要介绍历史上对导数符号表示的各种提议，以及为何某些表示形式比其他形式更流行的原因。研究微积分符号的发展历史不仅饶有趣味，还能为未来符号的发展提供一种范例。分析过去符号成败的原因，能让人在数学的特定领域不断拓展时，对新引入符号的未来发展具备一定的前瞻性。

This capsule is intended for undergraduate calculus classes and should ideally be offered directly before or after covering the derivative and its various applications, or indeed right in the middle when the students need a bit of a break from theory and practice! It may also be suitable for higher level secondary school mathematics.

这个教学单元适用于本科微积分课程，最好在讲授导数及其各种应用之前或之后进行，或者在学生需要从理论和实践中稍作休息时，安排在课程中间时段。它也适合高中高年级的数学教学。

This capsule is intended to take 30 minutes and contains visual material which may be presented to the students as well as directed questions to encourage them to think critically about notation.

这个教学单元预计时长为 30 分钟，包含可展示给学生的可视化材料，以及引导性问题，以鼓励他们批判性地思考数学符号。

2 导数

How many ways to represent the derivative have you come across in mathematics? Are they really equivalent? Who developed them and what can history reveal about the success of some notations over others? Among the many disagreements Newton and Leibniz are remembered for, they had a big one over notation. They both independently developed distinctive notation for the derivative when they published their results in calculus. In turn, allegiance in both British and European mathematical communities to strictly one or the other persisted for almost half a century until Leibniz’s notation finally prevailed. We will look at this scuffle and its consequences.

在数学中，你见过多少种表示导数的方式？它们真的等价吗？是谁发明了这些表示方法？从历史中能看出为什么某些符号比其他符号更成功吗？在牛顿和莱布尼茨众多为人熟知的争论中，符号表示法就是一个重大分歧点。他们在发表微积分成果时，各自独立地发明了独特的导数符号。结果，英国和欧洲的数学界在近半个世纪里，一直分别严格坚守其中一种符号表示法，直到莱布尼茨的符号最终占据上风。我们将探讨这场纷争及其影响。

人物	时间	导数符号表示
牛顿	1671 年	$\dot{x},\ddot{x},\dot{\vec{x}}$（点标记法）
莱布尼茨	1675 年	$d$ 相关符号，如 $\frac{dx}{dy}$（最初形式类似 $dxaddy$ 或 $dx:dy$）
拉格朗日	1797 年	$f^{\prime }(x),f^{\prime \prime }(x),f^{\prime \prime \prime }(x)$（撇标记法）
伯努利和阿尔博加斯特	1800 年	$D$ 相关符号，如 $DF(a)$ 表示 $F(a+x)$ 展开式中 $x$ 一次项系数

3 艾萨克・牛顿（1643 - 1727）

To understand Newton’s use of symbolism, we need to be familiar with the basics of his conception of calculus. In Newton’s mind, the fundamental notion of calculus was concerned with motion. Each variable in an equation could be ultimately interpreted as a distance reckoned with respect to time. Therefore, central to his account and thus his symbolic devising, was the notion of time.

为了理解牛顿对符号的使用，我们需要熟悉他的微积分基本概念。在牛顿看来，微积分的基本概念与运动相关。方程中的每个变量最终都可以解释为相对于时间的距离。因此，时间概念是他的理论以及符号设计的核心。

Throughout his mathematical researches, Newton developed several distinct ways to represent the derivative. The first of these was contained in De Analysi 1669 (but not published until 1711), in which among many other things, he developed a systematic method for differentiation. Notation-wise, Newton set a ‘little zero’, literally $\circ$, as a very small interval of time and $op$ and $oq$ as the corresponding small increments over which $x$ and $y$ change in this interval.

在他的数学研究过程中，牛顿提出了几种不同的导数表示方法。第一种方法包含在 1669 年的《分析学》中（但直到 1711 年才出版），在这本书里，他还开发了一种系统的微分方法。在符号表示上，牛顿用一个 “小零”，即 $\circ$，表示极短的时间间隔，$op$ 和 $oq$ 则表示在这个时间间隔内 $x$ 和 $y$ 的相应微小增量。

The second notational system appeared in about 1671, where $p$ and $q$ were reoriented symbolically as well as mathematically. Newton developed what he called Fluxional Calculus, in which he consciously avoided the use of infinitesimals as he considered these mathematically imprecise. He imagined that a curve was the result of the motion of a particle with respect to time. This curve, which was the path traced by the particle, could be then analyzed via changes in $x$ and in $y$. In order to consider the various properties of the curve, he defined fluents, or flowing quantities, as the variables $x$ and $y$ themselves and fluxions the velocity in the $x$ and $y$ direction respectively with respect to time. Thus, Newton’s fluxions were equivalent to our first derivative.

第二种符号系统大约出现在 1671 年，此时 $p$ 和 $q$ 在符号和数学意义上都有了重新定义。牛顿发展了他所谓的 “流数术”，他有意识地避免使用无穷小量，因为他认为这在数学上不够精确。他设想曲线是一个粒子随时间运动的轨迹。这条曲线，即粒子的运动路径，可以通过 $x$ 和 $y$ 的变化来分析。为了研究曲线的各种性质，他将流量（或流动量）定义为变量 $x$ 和 $y$ 本身，而流数则是 $x$ 和 $y$ 方向上相对于时间的速度。因此，牛顿的流数相当于我们现在所说的一阶导数。

Newton represented fluxions notationally as:

牛顿用以下符号表示流数：

$$\dot{x}\ddot{x}\dot{\vec{x}}$$

which are equivalent to the modern $\frac{dx}{dt}$, $\frac{d^{2}x}{d{t}^2}$ and $\frac{d^{3}x}{d{t}^3}$ respectively. These are occasionally referred to as ‘pricked letters’(for obvious reasons!), but more commonly today the system is called ‘dot - notation’.As you can see, by repeating the dots and introducing dashes Newton could present fluxions of fluxions, or fluents of fluents! The natural succession therefore was:

它们分别相当于现代的 $\frac{dx}{dt}$、$\frac{d^{2}x}{d{t}^2}$ 和 $\frac{d^{3}x}{d{t}^3}$。这些符号有时被称为 “带点字母”（原因显而易见！），但如今更常用的叫法是 “点标记法”。如你所见，通过重复加点和引入短横线，牛顿可以表示流数的流数，或流量的流量！因此，自然的序列是：

$$\Vert \Vert \Vert x\dot{x}\dot{x}\dot{x}\dot{x}\dot{\tilde{x}}\dot{\tilde{x}}$$

each term being the fluxion of the term preceding.

每一项都是前一项的流数。

Historians often connect particular features of Newton’s personality (in particular being private and guarded) and his continually evolving account of calculus, as responsible for the poor reception of his notation and indeed its ultimate demise. That he seemed to write for himself, they assume, had the consequence that he tended to adopt the notation he thought of at the time and put little thought into its practicality for wider audiences and broader mathematical function.

历史学家常常将牛顿的个性特点（尤其是他的内向和谨慎）以及他不断演变的微积分理论，与他的符号表示法不受欢迎乃至最终被弃用联系起来。他们认为，牛顿似乎是为自己而写作，这导致他倾向于采用当时想到的符号，而很少考虑这些符号对更广泛受众的实用性以及在更广泛数学领域的功能。

English mathematicians of the eighteenth century remained loyal to Newton’s notation, but ultimately the tradition lost favor. It is perhaps because of the inability of the notational system to allow for a richer development in the field of calculus, and to incorporate and represent for the equivalents in multi - variate calculus and calculus of variations. Newton’s system was also somewhat cumbersome, because of its orientation around the notion of time. Any expression containing a derivative not taken with respect to time was quite involved. For example, in order to express the derivative $\frac{dz}{dx}$, Newton had to write:

18 世纪的英国数学家们一直坚守牛顿的符号表示法，但最终这种传统不再受到青睐。这可能是因为这种符号系统无法促进微积分领域更丰富的发展，也难以在多元微积分和变分法中体现和表示相应的概念。牛顿的符号系统也有些繁琐，因为它围绕时间概念展开。任何包含非对时间求导的表达式都相当复杂。例如，为了表示导数 $\frac{dz}{dx}$，牛顿不得不写成：

$$\dot{z}:\dot{x}$$

which literally meant $\frac{dz}{dt}:\frac{dx}{dt}$

其字面意思是 $\frac{dz}{dt}:\frac{dx}{dt}$。

Newton’s dot - notation, however, didn’t disappear completely from the mathematical scene. It is used occasionally when expressing differential equations, particularly in engineering contexts, and can still be seen today in some textbooks.

然而，牛顿的点标记法并没有完全从数学领域消失。在表示微分方程时，偶尔还会用到它，尤其是在工程领域，如今在一些教科书中仍能看到。

4 戈特弗里德・威廉・莱布尼茨（1646 - 1716）

Leibniz, Newton’s European contemporary, in contrast took the greatest pains to develop his notation, believing that good notation was of primary importance, not only for representation of mathematical concepts, but for human understanding. In fact, he credited all his mathematical insights to his notation. The symbolic representation of a mathematical concept was, in his opinion, to be designed with the greatest attention—capturing a mathematical idea with carefully selected notation would help mathematicians make progress. The importance of notation seems to have been fostered in his early days as a student, as historians record episodes which reveal Leibniz’s teachers instilling in him a reverence for simplicity and clarity in discourse.

莱布尼茨与牛顿是同时代的欧洲人，与牛顿形成对比的是，莱布尼茨竭尽全力地发展他的符号表示法。他认为，好的符号不仅对于数学概念的表达至关重要，对于人们的理解也同样关键。事实上，他将自己所有的数学见解都归功于他的符号。在他看来，数学概念的符号表示应该精心设计，用精心挑选的符号来捕捉数学思想，有助于数学家取得进步。符号的重要性似乎在他学生时代就已扎根，历史学家记载，莱布尼茨的老师向他灌输了在表达中追求简洁和清晰的理念。

Leibniz’s account of the calculus was centered, in contrast to Newton, on the differential, a variable $dx$ which was an arbitrarily small finite increment upon the variable $x$. In a letter dated November 11th 1675, he introduced the symbol $dx$ and $dy$ as the differentials of $x$ and $y$ and something like $\frac{dx}{dy}$ for what we have come to call the derivative, but not in that exact form—rather expressed as:

与牛顿不同，莱布尼茨的微积分理论以微分概念为核心。微分是变量 $x$ 上一个任意小的有限增量，用变量 $dx$ 表示。在 1675 年 11 月 11 日的一封信中，他引入了符号 $dx$ 和 $dy$，分别表示 $x$ 和 $y$ 的微分，以及类似 $\frac{dx}{dy}$ 的形式来表示我们现在所说的导数，但最初并非完全是这种形式，而是写成：

$$dxaddy$$

or

或者

$$dx:dy$$

He wrote in this letter: …idem est $dx$ and $\frac{x}{d}$, id est differentia inter duas $x$ proximas

他在信中写道：……$dx$ 和 $\frac{x}{d}$ 是一样的，即两个相邻 $x$ 值之间的差

$...dx$ and $\frac{x}{d}$ are the same, that is, the difference between two proximate $x$’s

Higher derivatives were to be denoted as

高阶导数表示为：

$$ddx$$

being, in turn, an infinitesimally small increment upon the variable $dx$, and so on.

它是变量 $dx$ 上一个无穷小的增量，以此类推。

Almost immediately after this, he discovered the product and quotient rules. It was as if the symbolism suggested these concepts and manipulations.

几乎在这之后，他就发现了乘积法则和商法则。似乎正是这种符号表示法启发了这些概念和运算。

…videndum an $dxdy$ idem sit quod $d\overline{xy}$, et an $\frac{dx}{dy}$ idem quod $d\frac{x}{y}$

… 需考察 $dxdy$ 是否等同于 $d\overline{xy}$ ，以及 $\frac{dx}{dy}$ 是否等同于 $d\frac{x}{y}$ 。

…let us now examine whether $dxdy$ is the same thing as $d\overline{xy}$ and whether $\frac{dx}{dy}$ is the same things as $d(\frac{x}{y})$

… 现在让我们考察 $dxdy$ 是否与 $d\overline{xy}$ 相同，以及 $\frac{dx}{dy}$ 是否与 $d(\frac{x}{y})$ 相同 。

These rules were followed shortly by his articulation of the power rule and the rule for roots with respect to the derivative. Other rules fundamental to differentiation, such as the chain rule, seem trivial using his notation. His notation also allowed him in 1684 to solve a problem that he described as “one of the most difficult and most beautiful problems of applied mathematics, which without our differential calculus…no one could attack with any such ease.” This is the problem of finding a curve whose subtangent is a given constant $a$.

在这些法则之后，他很快又阐述了求导的幂法则和根式法则。其他一些基本的求导法则，比如链式法则，用他的符号表示起来似乎很简单。1684 年，他的符号表示法还帮助他解决了一个他称之为 “应用数学中最难且最美的问题之一，如果没有我们的微积分…… 没人能如此轻松地解决” 的问题。这个问题就是找到一条曲线，其次切线长度为给定常数 $a$。

Leibniz’s notation offered a simple, yet clear representation of the infinitesimal demands of the calculus, which no doubt facilitated its spread. His use of the ‘$d$’ which acted upon something emphasized the operator quality of the process. This proved important in later developments. His notation was immediately embraced by the continental mathematical community to their great advantage.

莱布尼茨的符号简洁而清晰地表示了微积分中无穷小的概念，这无疑促进了它的传播。他使用的 “$d$” 作用于某个量，强调了这个过程的算子性质。这在后来的发展中被证明非常重要。他的符号立即被欧洲大陆的数学界所接受，为他们带来了很大的便利。

There are many reasons as to why it ultimately prevailed. One of the most important was the fact that it could incorporate further details and features as mathematical developments and insights in calculus deepened. Furthermore, it recorded the right amount of information without obscuring anything, but neither was it too cumbersome.

它最终占据主导地位有诸多原因。其中最重要的一点是，随着微积分领域数学研究的深入发展和认知的加深，这种符号表示法能够融入更多的细节和特性。此外，它记录的信息量恰到好处，既不会遗漏关键信息，也不会过于繁杂。

One of the main distinctions in the approach to notation by Newton and by Leibniz is one of motivation. Newton searched for a means of expression for concepts he already conceived of. Leibniz, on the other hand, developed a platform in order to research new mathematics, with the result that, in many cases, the resulting symbolism proved to be suggestive of new mathematical insights and relations—just as he had intended.

牛顿和莱布尼茨在符号表示法上的主要区别之一在于动机。牛顿是为他已有的概念寻找表达方式；而莱布尼茨则是为研究新数学搭建一个平台，结果在很多情况下，他创造的符号正如他所期望的那样，能启发新的数学见解和关系。

In fact Leibniz’s ‘symbolic’ ambitions went beyond mathematics. He went so far as to develop a characteristica universalis or “universal characteristic”, founded on an alphabetic - like indexing of human thought. Through it, each element essential to human thought could be then represented and then a method of combining these symbols to indicate more complex notions. “It is obvious that if we could find characters or signs suited for expressing all our thoughts as clearly and as exactly as arithmetic expresses numbers or geometry expresses lines, we could do in all matters…For all investigations which depend on reasoning would be carried out by transposing these characters and by a species of calculus.” (Preface to the General Science, 1677. Revision of Rutherford’s translation in Jolley [5] p.234.)

事实上，莱布尼茨的 “符号化” 抱负不止于数学领域。他甚至还开发了一种 “通用字符表” 或 “通用特征符号”，它基于对人类思维类似字母索引的方式构建。通过这种符号系统，人类思维中的每一个基本元素都能被表示出来，并且可以通过组合这些符号来表达更复杂的概念。“显然，如果我们能找到像算术表示数字、几何表示线条那样清晰精确地表达我们所有思想的字符或符号，我们就能在所有事情上…… 因为所有基于推理的研究都可以通过变换这些字符并运用某种演算来完成。”（《通用科学序言》，1677 年。卢瑟福翻译版本修订，乔利 [5]，第 234 页）

5 约瑟夫 - 路易斯・拉格朗日（1736 - 1813）

At the end of the eighteenth century, Lagrange gave a definitive new approach to the calculus, being dissatisfied with earlier accounts. In his mathematical description, he purposefully avoided all references to fluxions and fluents, and to infinitesimals, (and indeed limits for that matter), because he believed they were unable to be defined precisely enough in a mathematical context.

18 世纪末，拉格朗日对早期的微积分理论感到不满，从而提出了一种全新的、具有决定性意义的微积分研究方法。在他的数学描述中，他有意避开了所有关于流数、流量以及无穷小（实际上还有极限）的概念，因为他认为在数学语境中，这些概念无法得到足够精确的定义。

Lagrange emphasized the centrality of the concept of the function and established a new basis for calculus with recourse to algebraic processes. Any function, Lagrange argued, can be represented as a power series. Given then some function $f(x)$, its related function $f(x+i)$ (with $i$ indeterminate) can be expanded:

拉格朗日强调函数概念的核心地位，并借助代数方法为微积分建立了新的基础。他认为，任何函数都可以表示为幂级数。假设有函数 $f(x)$，与之相关的函数 $f(x+i)$（其中 $i$ 为未定元）可以展开为：

$$f(x+i)=f(x)+p(x)i+q(x)i^2+r(x)i^3+\cdots$$

where $p(x),q(x),\cdots$ are independent of $i$. The first coefficient $p(x)$ of $i$ can be identified with the ratio $\frac{dy}{dx}$, and likewise for higher order ratios. He named $p(x)$ the “fonction dérivée” (from which we get ‘derivative’), because of the fact that it is ‘derived’ from the initial function $f(x)$. He denoted all higher order derivatives in turn:

其中 $p(x),q(x),\cdots$ 与 $i$ 无关。$i$ 的一次项系数 $p(x)$ 可以等同于 $\frac{dy}{dx}$，更高阶的系数也有类似的对应关系 。他将 $p(x)$ 命名为 “fonction dérivée”（由此我们得到了 “derivative”，即导数这个词），因为它是由初始函数 $f(x)$ “推导” 而来的。他依次表示所有高阶导数为：

$$f^{\prime },f^{\prime \prime },f^{\prime \prime \prime },\text{ etc.}$$

to emphasize the fact that the functions $p,q,r$ are all likewise “derived” from the initial function $f(x)$.

以此强调函数 $p,q,r$ 同样都是由初始函数 $f(x)$ “推导” 出来的。

The notation of Lagrange is used extensively today, as an important and useful adjunct of other notations current in modern calculus.

拉格朗日的符号表示法如今被广泛使用，它是现代微积分中其他常用符号的重要且实用的补充。

原函数与导函数的命名
 
9. Nous appellerons la fonction $fx$, fonction primitive, par rap - port aux fonctions $f^{\prime }x$, $f^{\prime \prime }x$, etc. qui en dérivent, et nous appelle - rons celles - ci, fonctions dérivées, par rapport à celle - là. Nous nommerons de plus la première fonction dérivée $f^{\prime }x$, fonction prime; la seconde fonction dérivée $f^{\prime \prime }x$, fonction seconde; la troi - sième fonction dérivée $f^{\prime \prime \prime }x$, fonction tierce, et ainsi de suite.
我们将函数 $fx$ 称为 原函数，相对于由它导出的函数 $f^{\prime }x$、$f^{\prime \prime }x$ 等而言。我们把后者称为 导函数。我们还将一阶导函数 $f^{\prime }x$ 称为 一阶导数；二阶导函数 $f^{\prime \prime }x$ 称为 二阶导数；三阶导函数 $f^{\prime \prime \prime }x$ 称为 三阶导数，依此类推 。
 
De la même manière, si $y$ est supposée une fonction de $x$, nous dénoterons ses fonctions dérivées par $y^{\prime }$, $y^{\prime \prime }$, $y^{\prime \prime \prime }$, etc., de sorte que $y$ étant une fonction primitive, $y^{\prime }$ sera sa fonction prime, $y^{\prime \prime }$ en sera la fonction seconde, $y^{\prime \prime \prime }$ la fonction tierce, et ainsi de suite.
同样地，如果 $y$ 被视为 $x$ 的函数，我们用 $y^{\prime }$、$y^{\prime \prime }$、$y^{\prime \prime \prime }$ 等表示它的导函数。这样，若 $y$ 是原函数，$y^{\prime }$ 将是它的 一阶导数，$y^{\prime \prime }$ 是 二阶导数，$y^{\prime \prime \prime }$ 是 三阶导数，依此类推。
 
De sorte que $x$ devenant $x+i$, $y$ deviendra
所以当 $x$ 变为 $x+i$ 时，$y$ 将变为

$$y+y^{\prime }i+\frac{y^{\prime \prime }{i}^2}{2}+\frac{y^{\prime \prime \prime }{i}^3}{2\cdot 3}+\text{etc.}$$
Ainsi, pourvu qu’on ait un moyen d’avoir la fonction prime d’une fonction primitive quelconque, on aura, par la simple répétition des mêmes opérations, toutes les fonctions dérivées, et par conséquent tous les termes de la série qui résulte du développement de la fonc - tion primitive.
因此，只要有一种方法能求出任意原函数的一阶导数，通过重复相同的运算，就能得到所有的导函数，进而得到原函数展开式中级数的所有项 。
 
Au reste, pour peu qu’on connaisse le calcul différentiel, on doit voir que les fonctions dérivées $y^{\prime }$, $y^{\prime \prime }$, $y^{\prime \prime \prime }$, etc. relatives à $x$, coïncident avec les expressions $\frac{dy}{dx}$, $\frac{d^{2}y}{d{x}^2}$, $\frac{d^{3}y}{d{x}^3}$, etc.

此外，只要对微分运算稍有了解，就会发现关于 $x$ 的导函数 $y^{\prime }$、$y^{\prime \prime }$、$y^{\prime \prime \prime }$ 等，分别与表达式 $\frac{dy}{dx}$、$\frac{d^{2}y}{d{x}^2}$、$\frac{d^{3}y}{d{x}^3}$ 等一致 。

Figure 5. Excerpt from Théorie des fonctions analytiques of Lagrange (1797) in which there is the first appearance of the prime notation for the derivative

图 5. 摘自拉格朗日 1797 年的《解析函数论》，其中首次出现了导数的撇标记法

6 路易・弗朗索瓦・安托万・阿尔博加斯特（1759 - 1803）

Indeed, many other notational systems were proposed for representing various concepts in calculus with varying reception from the mathematical community. We can’t give an exhaustive survey here, but will consider one final proposal which is significant because it conceives of the processes in the calculus as operations; the symbols become operators, which act on one function, to produce another. This notation is less common, but still has currency today in some mathematical contexts, particularly linear differential equations.

确实，人们提出了许多其他符号系统来表示微积分中的各种概念，它们在数学界的接受程度各不相同。我们无法在此进行详尽的综述，但将探讨最后一种具有重要意义的符号表示法，这种方法将微积分中的过程视为运算，符号成为作用于一个函数以产生另一个函数的算子。这种符号表示法不太常见，但在如今的一些数学领域，尤其是线性微分方程中，仍然在使用。

The first to champion this orientation was Arbogast in 1800, although such symbolism was used by Johann Bernoulli (1667 - 1748) earlier. Arbogast’s approach to calculus was similar to Lagrange. He used as his main symbol $D$, to indicate the process of differentiation. Similar to Lagrange, Arbogast supposes that $F(a+x)$ can be represented as the series:

最早倡导这种符号表示法的是 1800 年的阿尔博加斯特，不过约翰・伯努利（1667 - 1748）更早使用过类似的符号。阿尔博加斯特研究微积分的方法与拉格朗日类似。他用主要符号 $D$ 来表示求导过程。和拉格朗日一样，阿尔博加斯特认为 $F(a+x)$ 可以表示为级数：

$$a+bx+\frac{c}{1\cdot 2}{x}^2+\cdots$$

in which $a=Fa$. The symbol $D$ then is the operation on $Fa$ such that:

其中 $a=F(a)$。符号 $D$ 是对 $F(a)$ 的一种运算，使得：

$$DFa=b$$

and all higher derivatives:

并且所有高阶导数表示为：

$$DDFa=c$$

and so on. Arbogast also developed additional features which were not adopted much beyond their inception, but are historically interesting (see Figure 2).

以此类推。阿尔博加斯特还开发了一些其他特性，这些特性在创立后并未得到广泛采用，但从历史角度来看很有意思（见图 2）。

He states his motivations ([3] p. 209):

他阐述了自己的动机（[3]，第 209 页）：

To form the algorithm of derivations it became necessary to introduce new signs; I have given this subject particular attention, begin persuaded that the secret of the power of analysis consists in the happy choice and use of signs, simple and characteristic of the things which they are to represent. In this regard I have set myself the following rules:

为了构建求导算法，有必要引入新的符号；我特别关注这个问题，因为我深信分析学强大力量的奥秘在于恰当地选择和使用符号，这些符号要简单且能体现其所代表事物的特征。在这方面，我为自己制定了以下规则：

(1) To make notations as much as possible analogous to the received notations;

（1）使符号尽可能与已被接受的符号类似；

(2) Not to introduce notations which are not needed and which I can replace without confusion by those already in use;

（2）不引入不需要的符号，对于那些能用现有符号替代且不会引起混淆的，就不引入新符号；

(3) To select very simple ones, yet such that will exhibit all the varieties which the different operations require…

（3）选择非常简单的符号，但这些符号要能体现不同运算所需的各种变化……

As well as emphasizing the operator - like quality of the process, Arbogast’s notation found favor as it avoided reference to infinitesimals, which was appealing to many mathematicians, as succinctly expressed by a mathematical analyst from the early nineteenth century, Christian Kramp (1760 - 1826) ([3] pp. 210 - 211.):

阿尔博加斯特的符号表示法不仅强调了运算过程的算子性质，还因避免提及无穷小量而受到许多数学家的青睐，正如 19 世纪早期的数学分析家克里斯蒂安・克兰普（1760 - 1826）简洁表述的那样（[3]，第 210 - 211 页）：

Later researches have convinced me of the absolute inutility of this constant factor or divisor $dx$, as well as the notion of the infinitely small…[the adoption of the notation $D$] banishes all idea of the infinite and causes all this part of analysis to re - enter the domain of ordinary algebra.

后来的研究让我相信，这个常数因子或除数 $dx$ 以及无穷小的概念完全没有用处……（采用符号 $D$）消除了所有关于无穷的概念，使分析学的这一部分重新回归到普通代数的范畴。

7 结论

What makes a good notation? Is it how it responds to changes and developments? Its applicability to other fields? Its ability to be incorporated into new technology and discoveries? Should notation somehow reflect its function? Should we be conservative about our mathematical notation, or encourage innovation? Should we be revolutionary or retaining? Notation is pivotal for a mathematician and is often given very little consideration and taken for granted, not just by students, but by the very practitioners themselves!

什么样的符号才是好符号呢？是它对变化和发展的适应性吗？是它在其他领域的适用性吗？还是它融入新技术和新发现的能力？符号是否应该以某种方式反映其功能？我们在数学符号方面应该保守，还是鼓励创新呢？是应该进行彻底变革，还是保持传统？符号对于数学家至关重要，但它常常被忽视，不仅学生如此，就连专业的数学从业者也习以为常！

As we have seen, Newton and Leibniz feuded not only over the authorship of the calculus but also how notationally to express it. Correspondence reveals heavy criticism of the other’s approach: “Newton”, Newton himself once anonymously wrote, “does not confine himself to symbols”! Similar sentiments were returned from Leibniz and his supporters. Newton pursued research into the calculus through ideas of velocity and acceleration; Leibniz, through sums and differences. In turn, the English mathematical community subscribed to Newton’s methods and notation, the continent to those of Leibniz. As a result of these incompatible approaches, English and continental mathematicians effectively terminated all mathematical collaboration and exchange. In terms of notation, as the calculus was developed, Leibniz’s proposals turned out to be easier to work with and as a result, mathematical analysis on the continent progressed rapidly. To their loss, English mathematicians missed out on a century of stimulating and advancing mathematical activity, until they too adopted Leibniz’s notation by the eighteenth century. Indeed, the struggle to establish the expression for the derivative shows us the ways in which notational considerations can make a huge difference in mathematics.

正如我们所见，牛顿和莱布尼茨不仅在微积分的发明权上存在争议，在导数的符号表示上也有分歧。他们的通信中充满了对对方方法的严厉批评：牛顿曾匿名写道，“牛顿并不局限于符号”！莱布尼茨及其支持者也表达了类似的看法。牛顿从速度和加速度的概念出发研究微积分，而莱布尼茨则从求和与求差的角度入手。结果，英国数学界采用牛顿的方法和符号，欧洲大陆的数学界则采用莱布尼茨的。由于这些不相容的方法，英国和欧洲大陆的数学家实际上终止了所有的数学合作与交流。在符号表示方面，随着微积分的发展，事实证明莱布尼茨的符号使用起来更方便，因此欧洲大陆的数学分析迅速发展。英国数学家们因此错过了一个世纪充满活力且不断进步的数学活动，直到 18 世纪他们也采用了莱布尼茨的符号。的确，导数符号表示法的发展历程表明，符号的选择在数学领域会产生巨大的影响。

8 思考问题

1. Brainstorm some features you deem important to keep in mind when developing or evaluating a particular notational system. Here are some possible criteria to get you going: comprehensibility - simplicity - avoidance of ambiguity - uniqueness (i.e., make up symbols or borrow from other languages) - aesthetics - symbols which suggest their meaning - rhetorical potential - ability to represent multiple/additional features - minimisation of the strokes of a pen etc… Rank them from the most to the least important.

   集思广益，思考在开发或评估一种特定的符号系统时，你认为哪些特征是重要的。以下是一些可能的标准，供你参考：可理解性 - 简洁性 - 避免歧义 - 独特性（例如，创造新符号或借鉴其他语言的符号） - 美观性 - 符号能暗示其含义 - 具有启发思考的潜力 - 能够表示多种 / 额外的特征 - 书写简便性等等…… 按照重要程度从高到低对它们进行排序。

2. In what ways were the various proposals for the derivative a product of the concept their instigators were trying to emphasize?

   导数的各种符号表示提议在哪些方面体现了其提出者试图强调的概念？

3. Mathematical symbols not only represent a specific idea, they are a guide to operation by their careful placement. Select a notation that you use on a regular basis. What are its benefits? What are its limitations? What would you like to change about it?

   数学符号不仅代表特定的概念，其巧妙的布局还能指导运算。选择一种你经常使用的符号。它有哪些优点？存在哪些局限？你希望对它做出哪些改进？

4. We have looked at an episode from the History of Mathematics in which notation played an important role in facilitating mathematical insights. This is but one of many episodes which raises some very important questions for mathematicians regarding notation: To what extent is our ability to conceptualize and develop mathematics based upon the symbolism we employ? Consider this with respect to some specific examples, such as the integral sign, matrices, fractions, function notation, powers and indices.

   我们研究了数学史上符号在促进数学见解方面发挥重要作用的一个案例。这只是众多引发数学家对符号进行思考的案例之一：我们构建数学概念和推动数学发展的能力在多大程度上依赖于所使用的符号？结合一些具体例子，如积分符号、矩阵、分数、函数符号、幂和指数等，思考这个问题。

参考文献

[1] Jason Socrates Bardi, The Calculus Wars: Newton, Leibniz, and the Greatest Mathematical Clash of All Time, Thunder’s Mouth Press, New York NY, 2006, 86 - 88, 214 - 215.

[2] Carl B. Boyer, A History of Mathematics, 2nd Ed., John Wiley and Sons, 1991, 391 - 414.

[3] Florian Cajori, A History of Mathematical Notations, 2 Vols., Dover, 1993, reprinted from the work first published as two volumes by the Open Court Publishing Company, La Salle Illinois, in 1928 and 1929, 196 - 242.

[4] Victor Katz, A History of Mathematics An Introduction, 2nd ed., Addison Wesley Longman, 1998, 468 - 543.

[5] Nicholas Jolley, The Cambridge Companion to Leibniz, Cambridge University Press, 1995.

---

AN EXTENDED THEORETICAL FRAMEWORK FOR THE CONCEPT OF THE DERIVATIVE

导数概念的一个扩展理论框架

David Roundy,∗ Tevian Dray,† Corinne A. Manogue,‡ Joseph F. Wagner,§ and Eric Weber¶
August 27, 2014

Abstract
This paper extends the theoretical framework for exploring student understanding of the concept of the derivative, which was developed by Zandieh (2000). We expand upon the concept of a physical representation for the derivative by extending Zandieh’s map of the territory to provide higher resolution in regions that are of interest to those operating in a physical context. We also introduce the idea of “thick” derivatives, which are ratios of small but not infinitesimal changes, which are practically equivalent to the true derivative.

摘要

本文扩展了 Zandieh (2000) 开发的关于学生对导数概念理解的理论框架。我们通过扩展 Zandieh 的理论框架，增加了对导数的物理表示的理解，特别是在物理情境下感兴趣的区域提供了更高分辨率。我们还引入了“厚”导数的概念，即小但非无穷小的变化率，这些变化率实际上等同于真实的导数。

Key words: derivative, theoretical framework, physical, experiment

关键词：导数，理论框架，物理，实验

In this theoretical report we extend the theoretical framework for exploring student understanding of the concept of the derivative which was developed by Zandieh (2000). We expand upon the concept of a physical representation for the derivative. As with Zandieh’s original framework, this work is not meant to explain how or why students learn as they do, nor to propose a learning trajectory. Rather, this work extends Zandieh’s map of the territory, to provide higher resolution in regions that are of interest to those working with derivatives in a physical context.
在本理论报告中，我们扩展了 Zandieh (2000) 开发的关于学生对导数概念理解的理论框架。我们扩展了导数的物理表示的概念。与 Zandieh 的原始框架一样，本研究并不旨在解释学生为何以某种方式学习，也不提出学习路径。相反，本研究扩展了 Zandieh 的理论框架，特别是在物理情境下与导数相关的区域提供了更高分辨率。

In addition to focusing on the physical context, we discuss challenges that have arisen in applying Zandieh’s framework to an understanding of the derivative beyond the level of first-year calculus. This work is motivated by preliminary results of a project to study understanding of the derivative across STEM fields. In the process of interviewing physicists and engineers, we have identified shortcomings that arise when applying Zandieh’s framework beyond the level of first-year calculus, and in particular outside the field of mathematics.
除了关注物理情境外，我们还讨论了在将 Zandieh 的框架应用于超出微积分一年级水平的导数理解时所面临的挑战。这项工作是受到一个跨学科领域研究导数理解的项目的初步结果的启发。在采访物理学家和工程师的过程中，我们发现了在超出微积分一年级水平、特别是在数学领域之外应用 Zandieh 框架时出现的不足。

We have found that the concept image for the derivative of physicists and engineers contains substantial elements that are congruent with the three process-object layers identified by Zandieh, but lead to the introduction of new contexts and representations that could also be productive in the instruction of calculus.
我们发现，物理学家和工程师对导数的概念图像包含与 Zandieh 识别的三个过程-对象层次一致的重要元素，但这也导致了新情境和表示方法的引入，这些新元素在微积分教学中也可能是富有成效的。

Physicists and engineers live and work in a world full of uncertainty, and are accustomed to use the language of equality where there is actually approximation. This language reflects a somewhat “thicker” concept of the derivative than that held by mathematicians. Where a mathematician would speak of the slope of the secant line as an approximation for the derivative, a physicist or engineer might say that the slope of a line drawn between two carefully chosen measurements of a physical observable is the derivative (with some unspecified uncertainty).
物理学家和工程师生活和工作在一个充满不确定性的世界中，他们习惯于在实际是近似的地方使用等式语言。这种语言反映了一种比数学家所持有的“更厚”的导数概念。数学家会将割线的斜率视为导数的近似值，而物理学家或工程师可能会说，连接两个精心选择的物理可观测量测量值的直线的斜率就是导数（带有一些未指定的不确定性）。

As we will explain, this “thickness” derives from the impossibility of achieving exact results in physical or numerical contexts. Attempts to estimate a derivative over too small an interval, for example, could result in a highly erroneous estimate of a derivative due to numerical round-off error or limitations in experimental precision.
正如我们将解释的，这种“厚度”来自于在物理或数值情境中无法获得精确结果的现实。例如，尝试在过小的区间内估计导数可能会由于数值舍入误差或实验精度的限制而导致对导数的严重误估。

Theoretical background
理论背景

Concept Image

概念图像

In this work, we extend the theoretical framework of Zandieh (2000), which itself draws on the idea of concept image (Vinner, 1983). Vinner (1983) describes the concept image as the set of properties associated with a concept together with mental pictures of the concept. Thompson (2013) argues that the development of coherent meanings is at the heart of the mathematics that we want teachers to teach and what we want students to learn. He argued that meanings reside in the minds of the person producing them and the person interpreting them.
在本研究中，我们扩展了 Zandieh (2000) 的理论框架，该框架本身借鉴了概念图像（Vinner, 1983）的思想。Vinner (1983) 将概念图像描述为与概念相关的属性集合以及对概念的心理图像。Thompson (2013) 认为，发展连贯的意义是我们希望教师教授和学生学习的数学的核心。他认为，意义存在于产生意义的人和解释意义的人的心智中。

Zandieh’s framework for the concept of the derivative

Zandieh 的导数概念框架

Zandieh (2000) introduced a framework for the concept of the derivative, aimed at mapping student concept images at the level of first-year calculus. This framework maps out the correct concepts as understood by the mathematical community, and thus does not incorporate incorrect understandings. We reproduce in Fig. 1 below Zandieh’s outline of her framework. This table consists of columns corresponding to representations or contexts, and rows corresponding to process-object layers. The process-object framework is taken from Sfard (1991), who conceives of mathematics as proceeding through processes acting on objects, with those processes then becoming reified into objects.
Zandieh (2000) 引入了一个关于导数概念的框架，旨在绘制一年级微积分学生对导数的概念图像。该框架绘制了数学界所理解的正确概念，因此没有包含错误的理解。我们在下图 1 中重现了 Zandieh 的框架大纲。该表格包含对应于表示或情境的列，以及对应于过程-对象层次的行。过程-对象框架来自 Sfard (1991)，她认为数学是通过作用于对象的过程进行的，这些过程随后被物化为对象。

过程-对象层次
图形 语言 物理 符号 其他
斜率 变化率 速度 差商
比率 极限 函数

Figure 1: Zandieh’s outline of the framework for the concept of the derivative.
图 1：Zandieh 的导数概念框架大纲。

Representations

表示方法

Each of the representations in Zandieh’s table can be used to convey the concepts behind the three process-object layers. She also likens these columns to “contexts” in the sense that each of these provides a context within which we can think about the derivative. In the paragraphs below, we give a brief summary of each position in Fig. 1
Zandieh 表中的每一种表示方法都可以用来传达三个过程-对象层次背后的概念。她还将这些列比作“情境”，因为每一个都提供了一个我们可以思考导数的情境。在下面的段落中，我们对图 1 中的每个位置进行了简要总结。

Graphical. The graphical representation of the derivative is slope. At the ratio layer, this is the slope of a secant line between two points on the curve describing a function. When taking the limit, we arrive at the slope of the tangent line at a point. Finally, considering the derivative as a function requires us to recognize that the slope is different for different values of the independent variable.

图形 。导数的图形表示是斜率。在比率层，这是函数曲线上两点之间的割线斜率。当取极限时，我们得到某点处的切线斜率。最后，将导数视为一个函数需要我们认识到斜率随自变量的不同值而变化。

Verbal. The verbal representation for the derivative discussed by Zandieh is the “rate of change.” At the ratio layer, this is expressed as an “average rate of change.” When taking the limit, this becomes the “instantaneous rate of change.” Understanding this verbal description as a function requires us to visualize the instantaneous rate of change for the inputs over the domain of the function.

语言 。Zandieh 讨论的导数的语言表示是“变化率”。在比率层，这被表达为“平均变化率”。当取极限时，它变为“瞬时变化率”。将这种语言描述理解为一个函数需要我们想象函数定义域内输入的瞬时变化率。

Physical. The physical representation, or paradigmatic physical representation is velocity: average velocity, instantaneous velocity, and the velocity as a function of time. These physical concepts provide a language that we can use to understand the derivative: a large derivative means “faster” and a varying derivative means there is acceleration going on.

物理 。物理表示，或典范物理表示是速度：平均速度、瞬时速度以及速度随时间的函数。这些物理概念为我们提供了一种理解导数的语言：较大的导数意味着“更快”，变化的导数意味着存在加速度。

Symbolic. The symbolic representation of the derivative is the formal definition of the derivative in terms of the limit of a difference quotient. In this case, the distinction between the limit layer and the function layer can be subtle. They differ in the recognition that the variable describing the point at which the limit is taken can be treated as the argument of a function. Zandieh expresses this with a notational distinction between 在 $x_0$ and $x$.

符号 。导数的符号表示是用差商的极限形式定义导数。在这种情况下，极限层和函数层之间的区别可能很微妙。它们的区别在于认识到描述取极限点的变量可以被视为函数的自变量。Zandieh 通过在 $x_0$ 和 $x$ 之间的符号区别来表达这一点。

Other. Finally, we point out that Zandieh explicitly placed in her framework space for additional contexts. In particular, when discussing the physical context, she mentioned that there is a wide set of physical contexts for understanding the derivative. In this paper, we will discuss some of the subtleties we have encountered in investigating understanding of the derivative within the context of a mechanical system (AUTHOR).

其他 。最后，我们指出 Zandieh 在她的框架中明确为其他情境留出了空间。特别是，在讨论物理情境时，她提到理解导数的物理情境非常广泛。在本文中，我们将讨论我们在研究机械系统中导数理解时遇到的一些细微之处（作者）。

Extensions to Zandieh’s framework

对 Zandieh 框架的扩展

Likwambe and Christiansen (2008) extend Zandieh’s framework in three ways. Firstly, they recognize the importance in a concept image that we be able to make connections between different representations, and extend the use of the table to include arrows indicating that a student has made a connection between two representations or ideas. Secondly, they add a “non-layer” row, which indicates a recognition or use of that representation of the derivative without indication of an understanding of any of the three process-objects layers. Finally, Likwambe and Christiansen (2008) added a separate category for what they refer to as instrumental understanding, a term taken from Skemp (1978). Instrumental understanding (as opposed to relational understanding) refers to the knowledge of and ability to follow a procedure. Both Skemp (1978) and Lithner (2003) point out that instrumental understanding is commonly emphasized in both homework assignments and exams. Zandieh explicitly omits instrumental understanding from her framework, but Likwambe and Christiansen (2008) add an additional box for instrumental understanding, in order to include “the only learning exhibited by most of the interviewees.”

Likwambe 和 Christiansen (2008) 从三个方面扩展了 Zandieh 的框架。首先，他们认识到在概念图像中，我们能够建立不同表示之间的联系是很重要的，并扩展了表格的使用，增加了箭头，表明学生已经建立了两种表示或想法之间的联系。其次，他们增加了一个“非层次”行，这表明对导数的这种表示的认识或使用，但没有表明对三个过程-对象层次中的任何一个的理解。最后，Likwambe 和 Christiansen (2008) 增加了一个单独的类别，他们称之为工具性理解，这个术语来自 Skemp (1978)。工具性理解（与关系性理解相对）是指对程序的知识和执行能力。Skemp (1978) 和 Lithner (2003) 指出，工具性理解通常在家庭作业和考试中被强调。Zandieh 明确将工具性理解排除在她的框架之外，但 Likwambe 和 Christiansen (2008) 增加了一个额外的工具性理解框，以包括“大多数被访谈者所表现出的唯一学习”。

Extending Zandieh’s framework for the derivative

扩展 Zandieh 的导数框架

In our research on expert understanding of the derivative across disciplines, we have encountered several issues that led us to an extension of Zandieh’s framework for the derivative, with a particular focus on physical contexts. We propose a deeper understanding of the “physical” representation, and add an additional “numerical” representation, which fills out the Rule of Four: graphical, verbal, symbolic and numerical (Hughes-Hallett et al., 1998). In addition, we follow Likwambe and Christiansen (2008) in adding an instrumental understanding category that lives outside the three process-object layers. Figure 2 shows our framework for the concept of the derivative. This figure is modeled after Fig. 1, the framework of Zandieh, and is best understood in terms of the differences between these two frameworks. We have added one additional column labeled numerical (and removed the Other column to make space). We have added the instrumental understanding of Likwambe and Christiansen (2008) (which is to say, the rules of differentiation) as an entirely separate table, partially to reflect its weak connection to any other aspect of the concept of the derivative. Finally, we have added into each entry of the table (which Zandieh left blank) an iconic description of the concept meant by that entry. These entries are intended to aide in understanding the table by compactly describing the conception of the derivative indicated by that combination of row and column.
在我们对跨学科领域专家对导数理解的研究中，我们遇到了一些问题，这些问题促使我们扩展了 Zandieh 的导数框架，特别关注物理情境。我们提出了对“物理”表示的更深入理解，并增加了一个额外的“数值”表示，这完善了四则运算规则：图形、语言、符号和数值（Hughes-Hallett 等，1998）。此外，我们按照 Likwambe 和 Christiansen (2008) 的方法，增加了一个工具性理解类别，它位于三个过程-对象层次之外。图 2 展示了我们对导数概念的框架。这个图是根据 Zandieh 的框架图 1 模型设计的，最好通过这两个框架之间的差异来理解。我们增加了一个额外的列，标记为数值（并删除了其他列以腾出空间）。我们将 Likwambe 和 Christiansen (2008) 的工具性理解（即，微分规则）作为一个完全独立的表格添加进去，部分原因是反映其与导数概念的其他方面联系较弱。最后，我们在表的每个条目中（Zandieh 留空）添加了一个概念的标志性描述。这些条目旨在通过简洁地描述由行和列组合所指示的导数概念，帮助理解表格。

Changes in the framework

框架的变化

In this section, we discuss individually the extensions we have made to Zandieh’s framework.
在本节中，我们将分别讨论我们对 Zandieh 框架所做的扩展。

Physical. We begin by noting that the physical examples given by Zandieh (2000) each involve a time derivative: velocity, acceleration, and the time rate of change of temperature. We suggest that although these quantities do reside in a physical context, perhaps at least some uses of these phrases properly belong in the realm of verbal representation. We propose here a more “physical” (as opposed to verbal) concept of the physical representation of the derivative.
物理 。我们首先指出，Zandieh (2000) 给出的物理例子都涉及时间导数：速度、加速度以及温度随时间的变化率。我们认为，尽管这些量确实存在于物理情境中，但至少有一些对这些术语的使用可能更属于语言表示的范畴。在这里，我们提出了一个更“物理的”（而不是语言的）导数的物理表示概念。

We define the physical representation for the derivative to be a process to measure that derivative (see, for instance Roundy, Kustusch, & Manogue, 2014; Styer, 1999). Of course, the concept does not require us to actually perform a measurement, just to imagine one. However, we note that it is the process of measurement itself that is the physical representation. Actually obtaining a numerical measurement would (also) require the use of the numerical representation, and describing the measurement may involve a verbal or graphical representation (Roundy et al., 2014; Styer, 1999), but the measurement process itself is the physical representation of the concept of the derivative.

我们定义导数的物理表示为测量该导数的过程（见 Roundy, Kustusch, & Manogue, 2014；Styer, 1999）。当然，这个概念并不要求我们实际进行测量，只需想象一下即可。然而，我们指出，正是测量过程本身构成了导数的物理表示。实际获得一个数值测量结果（也需要）使用数值表示，而描述测量过程可能涉及语言或图形表示（Roundy 等，2014；Styer, 1999），但测量过程本身是导数概念的物理表示。

As an example, consider the derivative $\frac{dV}{dp}$ of the volume of a piston full of air with respect to the pressure on the piston, as controlled by a set of weights on the piston (illustrated in Fig. 2). At the ratio layer, one can say that you need to measure the volume twice, with two different pressures, and the derivative is the change in volume divided by the change in pressure.

例如，考虑一个充满空气的活塞相对于活塞上压力的导数 $\frac{dV}{dp}$，该压力由活塞上的砝码控制（如图 2 所示）。在比率层，可以说你需要测量两次体积，对应两个不同的压力，导数就是体积变化量除以压力变化量。

The limit layer imposes on this process the idea that the two pressures need to be quite similar in order for this ratio to “be” the derivative in the thick sense used by physicists and engineers. However, it is not desirable to choose too small a value for $\Delta p$, because this would result in an imprecise measurement, since the change in volume would be too small to be precisely measured, resulting in increased error in the value of the measured derivative. Finally, the function layer requires us to recognize that this ratio will depend on the pressure itself, and that to fully explore the derivative, we must perform repeated experiments—or more likely a single experiment in which we gradually add weight to the piston and repeatedly measure its volume. The physical representation of a derivative can often (but not always) be felt or perceived directly, which leads scientists to give derivatives names such as compressibility, velocity, thermal conductivity, etc. Qualitatively, the derivative $\frac{dV}{dp}$ describes the compressibility of the air: how easy it is to compress. We anticipate that as the piston is compressed at higher pressures, it will require more and more pressure to compress it further. Because the volume cannot be negative, we can conclude on physical grounds that the derivative must eventually approach zero as the pressure increases.

极限层要求这两个压力值必须非常接近，以便这个比率能够“成为”物理学家和工程师所使用的“厚”导数。然而，不建议选择过小的 $\Delta p$ 值，因为这会导致测量不精确，因为体积变化量太小而无法精确测量，从而导致测量导数的误差增加。最后，函数层要求我们认识到这个比率将依赖于压力本身，为了充分探索导数，我们需要进行多次实验——或者更有可能的是，进行一次实验，逐渐增加砝码重量并反复测量活塞的体积。导数的物理表示通常（但不总是）可以直接感知或触摸到，这促使科学家们给导数起了诸如压缩性、速度、热导率等名称。从定性上讲，导数 $\frac{dV}{dp}$ 描述了空气的压缩性：它有多容易被压缩。我们预计，随着活塞在更高压力下被压缩，它将需要越来越多的压力才能进一步压缩。由于体积不能为负，我们可以在物理基础上得出结论，随着压力的增加，导数最终会趋近于零。

Numerical. The numerical representation is the one member of the Rule of Four (Hughes-Hallett et al., 1998) that was not present in the framework of Zandieh (2000). We recognize a numerical representation of the derivative that is closely allied to but distinct from the physical representation. This representation parallels the formal symbolic concept of the derivative, but differs in ways that are of practical importance in the use of the derivative in the sciences and in numerical analysis.
数值 。数值表示是四则运算规则（Hughes-Hallett 等，1998）中唯一一个在 Zandieh (2000) 的框架中不存在的成员。我们认识到导数的数值表示与物理表示密切相关但又有所不同。这种表示与导数的形式符号概念平行，但在科学和数值分析中使用导数时具有实际重要性。

The numerical concept of the derivative begins with a ratio of change:
数值导数的概念从变化率开始：

$$\frac{y_2-y_1}{x_2-x_1}$$

where it is understood that the values in this equation are numerical values. When we take the limit numerically, we do not formally write ${\lim }_{\Delta x\to 0}$, and we do not apply a formal procedure. Rather we select a value of $\Delta x$ that is small, where small is understood in terms of the desired precision. As in the case of physical measurements, practically speaking it is possible to make the change $\Delta x$ too small, in this case due to truncation error in a computer or calculator. In this regard, when operating numerically we think of derivatives as having some “thickness,” in contrast to the formal definition which requires an infinitesimal limit. Finally, the derivative as function is understood as a sequence of numerical ratios of differences, just as a function can be understood numerically as an array of numbers or set of ordered pairs.
在这个方程中，这些值被理解为数值。当我们数值地取极限时，我们不会正式地写 ${\lim }_{\Delta x\to 0}$，也不会应用一个正式的程序。相反，我们会选择一个足够小的 $\Delta x$ 值，这里的“小”是根据所需的精度来理解的。正如在物理测量的情况下，实际上，$\Delta x$ 的变化可能会太小，这种情况下是由于计算机或计算器中的截断误差。在这方面，当我们数值地操作时，我们认为导数具有一定的“厚度”，这与需要无穷小极限的形式定义形成对比。最后，作为函数的导数被理解为一系列数值差的比率，就像函数可以被数值地理解为一组数字或有序对的集合。

Conclusions

结论

We have extended the framework of Zandieh (2000) in several ways: we have elaborated on the physical representation of the derivative; we have added a numerical representation of the derivative; and we have added space in the framework for the set of rules for finding symbolic derivatives. Each of these changes reflects an expansion of the table to incorporate additional answers to the prompt, “find the derivative.” By making use of the numerical representation of the derivative, one can answer the prompt numerically. Similarly, if the derivative is situated in a physical context, one can respond with a measurement process. Both of these responses require a conceptual understanding of the derivative in terms of ratio, limit and function, and involve a certain “thickness” in the derivative. In contrast, as pointed out by Zandieh, the instrumental-understanding approach to “find the derivative” using the rules for symbolic derivatives does not require a conceptual understanding of the derivative.
我们以多种方式扩展了 Zandieh (2000) 的框架：我们详细阐述了导数的物理表示；我们增加了导数的数值表示；并且我们在框架中为寻找符号导数的规则集留出了空间。这些变化都反映了表格的扩展，以纳入对“求导数”这一提示的更多回答。通过使用导数的数值表示，可以数值地回答这一提示。同样，如果导数处于物理情境中，可以用测量过程来回应。这两种回答都需要在比率、极限和函数方面对导数有概念性的理解，并且涉及导数的某种“厚度”。相比之下，正如 Zandieh 指出的，使用符号导数规则来“求导数”的工具性理解方法不需要对导数有概念性的理解。

References

参考文献

Hughes-Hallett, D., Flath, D., Gleason, A., Gordon, S., Lock, P., Lomen, D., … others (1998). Calculus: Single variable (2nd ed.). John Wiley & Sons Australia, Limited.
Hughes-Hallett, D., Flath, D., Gleason, A., Gordon, S., Lock, P., Lomen, D., … 等人 (1998). 微积分：单变量（第 2 版）。John Wiley & Sons Australia, Limited.

Likwambe, B., & Christiansen, I. M. (2008). A case study of the development of in-service teachers’ concept images of the derivative. Pythagoras(68), 22–31.
Likwambe, B., & Christiansen, I. M. (2008). 在职教师导数概念图像发展的案例研究。《毕达哥拉斯》(68)，22–31。

Lithner, J. (2003). Students’ mathematical reasoning in university textbook exercises. Educational studies in mathematics, 52(1), 29–55.
Lithner, J. (2003). 大学生在教科书练习中的数学推理。《数学教育研究》, 52(1), 29–55。

Roundy, D., Kustusch, M. B., & Manogue, C. (2014). Name the experiment! Interpreting thermodynamic derivatives as thought experiments. American Journal of Physics, 82(1), 39–46.
Roundy, D., Kustusch, M. B., & Manogue, C. (2014). 命名实验！将热力学导数解释为思想实验。《美国物理杂志》, 82(1), 39–46。

Sfard, A. (1991). On the dual nature of mathematical conceptions: Reflections on processes and objects as different sides of the same coin. Educational studies in mathematics, 22(1), 1–36.
Sfard, A. (1991). 数学概念的双重性质：对过程和对象作为同一事物的两面的反思。《数学教育研究》, 22(1), 1–36。

Skemp, R. R. (1978). Relational understanding and instrumental understanding. The Arithmetic Teacher, 9–15.
Skemp, R. R. (1978). 关系性理解与工具性理解。《算术教师》, 9–15。

Styer, D. F. (1999). A thermodynamic derivative means an experiment. American Journal of Physics, 67(12), 1094–1095.
Styer, D. F. (1999). 热力学导数意味着一个实验。《美国物理杂志》, 67(12), 1094–1095。

Thompson, P. W. (2013). In the absence of meaning. In K. Leatham (Ed.), Vital directions for research in mathematics education. New York, NY: Springer.
Thompson, P. W. (2013). 缺乏意义。见 K. Leatham (编)，《数学教育研究的重要方向》。纽约：Springer。

Vinner, S. (1983). Concept definition, concept image and the notion of function. International Journal of Mathematical Education in Science and Technology, 14, 293-305.
Vinner, S. (1983). 概念定义、概念图像与函数概念。《科学与技术数学教育国际杂志》, 14, 293-305。

Zandieh, M. (2000). A theoretical framework for analyzing student understanding of the concept of derivative. CBMS Issues in Mathematics Education, 8, 103–122.
Zandieh, M. (2000). 分析学生对导数概念理解的理论框架。《数学教育中的 CBMS 问题》, 8, 103–122。

---

via :

• A ‘Symbolic’ History of the Derivative
  https://www.researchgate.net/publication/297477507_A_‘symbolic’_history_of_the_derivative

• AN EXTENDED THEORETICAL FRAMEWORK FOR THE CONCEPT OF THE DERIVATIVE - framework.pdf
  https://bridge.math.oregonstate.edu/papers/framework.pdf