32
$\begingroup$

There are many, many ways of writing the derivative of a function $y=f(x)$:

$$\frac{d}{dx}y, \frac{dy}{dx},\frac{d}{dx}f(x), \frac{df}{dx}, \dot y, D_x f,f',y',f'(x),f_x$$

and so on.

Students often feel uncomfortable switching back and forth between the various forms. And in fact, it is not necessary to use more than one form of notation in most beginning calculus courses.

However, knowing more than one form of notation may be helpful in future classes.

Should an instructor encourage their students to use different forms of derivative in their writing?

$\endgroup$
3
  • 8
    $\begingroup$ I would strongly suggest, not to mix up notations like $f'$ and $f'(x)$. The first one is for a function, the second one for the function's value. $\endgroup$
    – Anschewski
    Mar 20, 2014 at 21:23
  • 16
    $\begingroup$ @Anschewski, yes, you are right, but many mathematicians are very inconsistent with this, as are texts. For that matter, $df/dx$ is strange unless $x$ is somehow the "sacred" name for the input to $f$, and so on. Or "the function $x^2$", where the presumption is that the input is $x$. Coping with these language abuses is part of the issue. I've known people who tried to explain the distinctions to students, with not much success. $\endgroup$ Mar 20, 2014 at 22:56
  • 5
    $\begingroup$ @paulgarrett, I know this is not handled consistently, neither in school nor at university. Let's discuss this in a new question. $\endgroup$
    – Anschewski
    Mar 21, 2014 at 9:31

10 Answers 10

34
$\begingroup$

The different notations are to a considerable extent outgrowth of slightly different ways of looking at "the derivative." To me these different ways is the important thing to discuss, the notation is a by product. And, I think it is quite useful to spend some time on discussing these different ways and to mention the different notations in that context. (However I understand this is not always possible, see the end.)

For example:

  • the notation $f'$ and $\dot{y}$ stresses the (physical) idea of having some function that describes something (some process) and one does something to it to analyze/understand its properties.

  • the notation $D_x f$ emphasizes the fact that one can think of the situation of having an operator (a function!) $D_x$ that assigns to each (under suitable conditions) function $f$ another function $D_xf$.

  • the notation $\frac{df}{dx}$ emphasizes the definition as a limit of the quotients of differences.

  • the notation $df(x)$, in particular when used like $f(x+h) = f(x) + df(x)h + r(h)$ stresses that the derivative at a some point is actually a linear function that just happens to be identifiable with its slope.

There would also be different ways to split things up and to motivate things.

If one does this the different notations come more naturally and one then can and I think should vary one's own usage and exercises a bit. Still, I would not insist or even much encourage students using different notations themselves. There is some clear merit in having them understand the different notations; for them being able to use different ones (fuently) this is much less clear to me. There is also a small risk that they would use them, but would use them not really 'idiomatically,' as the notations arein some sense not completely interchangable.

For example even an example in OP seems slightly of to me (though some place else different conventions might be in place): as far as I know, the dot-notation is mainly used by physicist and alike, and does typically not denote just some derivative, but writing $\dot{y}$ "means" derivative by time, which "would never" be denoted $x$ but $t$. [For me, as a pure mathematician, the above is strange to write, but I really think that if $y=xt^2$, then to some it is completely obvious that $\dot{y}=2xt$ and not $\dot{y}=t^2$.]

This answer assumes of course that one has time to do this discussion, and implicitly also that the students or a considerable part there of will actually see the higher dimensional theory for example. If this is not the case, I think I would not bother much with different notations. Pick one (based on your own preference or possibly informed by the predominant convention in whatever other field the students might study, if there is one) and stick to it, and just mention once that there are different notations, so that students are not completely surprised when the open some other book. However, if some student prefers to use a different one, accept this, too (as long it is clear and coherent).

$\endgroup$
2
  • 2
    $\begingroup$ Might I add that Leibniz's notation is handier than Lagrange's notation in certain situations requiring the chain rule in single variable calculus. I tend to prefer it for integration by substitution, implicit differentation, differentiation of parametric equations and finding derivatives of compositions of three or more functions. $\endgroup$
    – J W
    Mar 29, 2014 at 13:19
  • 8
    $\begingroup$ Just a nitpick: In physics courses, $x$ and $y$ usually represent the $x$-coordinate and the $y$-coordinate of some object's location, and it's likely that both $y$ and $x$ are functions of time so in fact if $y = xt^2$, then you'd usually get $\dot{y} = \dot{x}t^2 + 2xt$. $\endgroup$ Aug 28, 2014 at 22:32
15
$\begingroup$

Yes, students will encounter several forms of derivative notation, inside and outside of classes. Therefore, a calculus class should give them fluency in reading these.

Whether they stick to one notation in their writing is not a big deal. More importantly: the lectures and questions in calculus classes should use more than one notation.

$\endgroup$
6
  • 2
    $\begingroup$ Why the down-vote? $\endgroup$ Mar 20, 2014 at 20:59
  • 1
    $\begingroup$ @JonEricson, I could add that "in my personal experience, I have seen many notations for the derivative", but that would add little to the documentation already in the question. I have added the word "therefore" to show that there is a reasoned argument here, rather than a bald statement of opinion. What more would be helpful or desirable? $\endgroup$
    – user173
    Mar 25, 2014 at 18:33
  • 2
    $\begingroup$ @Matt F. This is a place where detailed (and longer) answers really are more valuable than a couple of terse paragraphs. To me, answers to the question must deal with the tension between confusing students with new notation and exposing them to the same. As it stands, this answer asserts just one side without considering the other. How do you deal with students getting confused when you switch from one notation to another? Are there techniques to introduce new notation that minimize confusion? $\endgroup$ Mar 25, 2014 at 18:42
  • 2
    $\begingroup$ @JonEricson: I often find shorter answers more valuable, but thanks for explaining your view. $\endgroup$
    – user173
    Mar 25, 2014 at 18:50
  • 2
    $\begingroup$ For what it's worth, I think this is a very useful answer, because it's true - Leibniz and prime notation are both very widespread, and as educators we have to, at least in part, prepare students for the world they will actually encounter. I suppose @user173 could have given a couple standard references. $\endgroup$
    – kcrisman
    Dec 15, 2016 at 1:16
12
$\begingroup$

I think it is important to introduce the $d/dx$, prime and, to a much lesser extent, dot notations for the following reasons:

  • $\frac{d}{dx}$ makes it quite clear which variable is being differentiated. Therefore, it is quite clear that $\frac{d}{dx}(a+x^2)=2x$ whereas $\frac{d}{da}(a+x^2) = 1$. In both of these calculations I assume that the complementary variable is held fixed in the differentiation. What do these look like in the prime notation? Well, $$ (a+x^2)' = 2x $$ is what my students would probably write. Perhaps, on occassion $(a+x^2)'=1$ would be offered by the sarcastic outlier. To actually be clear, we have to include the parenthetical evaluation: $$ (a+x^2)'(x) = 2x \qquad \& \qquad (a+x^2)'(a) = 1. $$ Even so, some might be tempted to write $(a+x^2)'(a) = 2a$. Which is it? The prime notation has danger in this regard.
  • the prime notation is efficient. I think this may be one of the greatest aspects: $$ (f+g)' = f'+g' \qquad (cf)'=cf' \qquad (fg)'=f'g+fg' $$ and $$ \left(\frac{f}{g} \right)' = \frac{f'g-fg'}{g^2} \qquad (f \circ g)'=(f' \circ g)g' $$ It does require more writing to get these across in the $d/dx$ notation. Although, I do prefer to include the argument for the chain rule $(f \circ g)'(x)=f'(g(x))g'(x)$.
  • the $\dot{x}$ and $\ddot{x}$ are fun. I just like to say $x$-double-dot. It takes me back to physics and so it has a certain nostalgic value for me personally. Pragmatically, these dot formulas are very pretty for expressing kinetic energy formulas for Lagrangian mechanics. When you use $\dot{x}$ as a variable it is a lot nicer than $x'$ or $dx/dt$. In variational calculus, the $\dot{x}$ appears as an "independent" variable from $x$.
$\endgroup$
3
  • 1
    $\begingroup$ Excellent answer! I add to the discussion of the Leibniz notation the advantage that it makes it much easier to discuss the chain rule (I find that calculus students understand $\frac{\text dy}{\text dx} = \frac{\text dy}{\text du}\cdot\frac{\text du}{\text dx}$ much more easily than they do the corresponding 'primed statement'); and of the prime or dot notation that they make differential equations much more compact, at a time when students are (hopefully) ready for the abstraction of dealing with functions as objects to be subjected to analysis. $\endgroup$
    – LSpice
    Aug 29, 2014 at 23:13
  • $\begingroup$ @LSpice Thanks. I also agree the apparent cancellation of $du$ is very appealing to starting students. In fact, at the very beginning I often force them to write the given function explicitly in terms of the $u$ as to drive the idea home. However, by the semester's end I try to expose them to all the common notations. Furthermore, if an abuse is common, I try to make it, but at the same time apologize for the misdeed. $\endgroup$ Aug 31, 2014 at 18:04
  • $\begingroup$ $(a+x^2)^\prime$ is irredeemable because $(a+x^2)^\prime(1)$ is ambiguous. (Is it $1$ or $2$?) The correct disambiguation is $(x \mapsto a+x^2)^\prime$. $\endgroup$
    – Jordan
    Feb 11, 2020 at 15:22
8
$\begingroup$

Short answer: I don't think that is a good idea to encourage students to use different notation on their own. But I would encourage the lecturer and the TAs to do so and tell the students different notations.

More detailed answer:

For many students, the topic itsself is hard enough. I don't think you should distract them from the most important things for notation.

But you can tell them, that there are different ways to write down things and the most important point is the notation should be consistent. For example, I've seen it several times that at the point of the definition, there were several methods mentioned to write down the new object (here: the derivative) and then in the lecture one of these notations was used, but on the exercices another notation was used (this is - as I later learned - due to the fact that someone else was makeing exercises, but he hasn't talked about notation with the professor). Students will learn that different people use different notation and every notation has some advantages and disadvantages. Of course, it could be that students are confused at the first point ("What does that strange symbol in the exercises mean?" - You can explain it and you can write more than a symbol in the first exercise where it appears. Later on it should be clear).

You should encourage the students to make up their minds what a good notation is to their understanding. And the students should use that notation consequently as long as they are satisfied. (This changes from time to time. I for example was a big fan of the $\frac{\partial}{\partial x}f$ until I took a class about partial differential equations where I discovered that my old notation was a nightmare. Than I changed my whole notation).

I think, it is important that students should be flexible and open-minded to notation: It can be that in every lecture there is a different notation, in books and later in papers, people use different notation (in bad ones, the notation chances from chapter to chapter). But one should never overemphasize the discussion of notation since there are more important things to teach. And you should not jump arround with notation (otherwise the notation gets the focus and students will ask you about the meaning of the symbols rather than about the content).

Remark: Most of my arguments given is not specific to the notation of derivatives, but also in a general context. Maybe the most easy examples where some students have problems with notation is the example where an equation should not be solves for $x$, but for a different variable. Another example would be the use of another symbol than $f$ for a function.

$\endgroup$
4
$\begingroup$

You only need $\frac{d}{dx}$, $\frac{dy}{dx}$, and $f^\prime$.

$\frac{d}{dx}\Psi$, where $\Psi$ is an expression whose value depends on that of $x$, is the rate of change of the value of $\Psi$ with respect to $x$.

$$\frac{d}{dx}\Psi = \lim_{h \rightarrow 0} \frac{f\left(x+h\right) - f\left(x\right)}{h}\ \text{where}\ f = k \mapsto \left(\Psi,\ \text{given}\ x=k\right)$$

$\frac{dy}{dx}$ is the rate of change of a dependent variable ($y$) with respect to an independent variable ($x$). Even though $y$ is simply an expression whose value depends on $x$ (making $\frac{dy}{dx}$ and $\frac{d}{dx}y$ synonyms), $\frac{dy}{dx}$ is nice because its form reminds us of the intuition that comes from $\frac{\Delta y}{\Delta x}$.

$$\frac{dy}{dx}= \lim_{h \rightarrow 0} \frac{f\left(x+h\right) - f\left(x\right)}{h}\ \text{where}\ f = k \mapsto \left(y,\ \text{given}\ x=k\right)$$

$f^\prime$, where $f$ is a function with one argument, is the rate of change of change of $f$'s output with respect to $f$'s input.

$$f^\prime = k \mapsto \lim_{h \rightarrow 0} \frac{f\left(k+h\right) - f\left(k\right)}{h}$$

Some caveats:

$\frac{df}{dx}$ and $\frac{d}{dx} f$ are meaningless because $f$ is a function, not a variable or an expression. The value of $f$ does not change with $x$; the value of $f\left(x\right)$ changes with $x$. So use $\frac{d}{dx} f\left(x\right)$ instead. Note that $\frac{d}{dx} f\left(x\right)$ is a synonym of $f^\prime\left(x\right)$. Also note that $\frac{d}{dx} f\left(a\right)=0$ because $f\left(a\right)$ is not an expression that depends on $x$.

Likewise, $y^\prime$ is meaningless because $y$ is a variable, not a function. $y$ could depend on any number of variables, so there's no unambiguous way to write $y$ as a function. But you could theoretically write: $\left( k \mapsto \left(y,\ \text{given}\ x=k \right)\right)^\prime\left(x\right)$ which is a synonym of $\frac{dy}{dx}$.

$\endgroup$
3
$\begingroup$

I'm for using different notations (particularly $\frac{\mathrm{d} f}{\mathrm{d} x}$ and $f'(x)$, as they are useful for emphasizing different aspects; and occasionally $\mathrm{D} f$ if the operator aspect is central). But don't overdo it, saying the same thing in five different ways "just because we can" will only confuse the reader. That they are expressions of different ways of seeing derivatives, from much before they were really understood, is incidental. And those views might even harm understanding, better leave them out until the concept is crystal clear and solid.

$\endgroup$
1
$\begingroup$

I think that the only possible answer is: it depends. Is fluency with various notation a goal of the course? If yes, you have your answer, if no, you would better concentrate on other stuff. Dealing with notation is hard and can be confusing and time-consuming, so don't engage in it unless you have good reasons to.

As an example, for engineering students it might make sense to have your course using the notation they will find in other courses, or to emphasize a notation which is more suited to your topics and a notation they will find elsewhere, to help them relate what you teach and the context in which they will have use it.

$\endgroup$
0
$\begingroup$

A thought from Steven Krantz's How to Teach Mathematics, 3rd Ed., published by the American Mathematical Society (sec. 2.12):

If a textbook uses notation or other conventions that you do not like, then don't use that book. You really are obliged to follow the notation and definitions and other paradigms in the text you have chosen. Otherwise all but the gifted students will be lost. If you repeatedly criticize the text as the course proceeds, then you will be sending a confusing message to the students: Why did you choose this book if it is obviously so full of flaws? Isn't it your job to select a text that you can teach from?

Granted that a textbook is unlikely to vary between multiple different notations, then the recommendation here is likely "no".

$\endgroup$
2
  • 1
    $\begingroup$ I'm not so sure about that in calculus, though (and perhaps only in calculus). Standard textbooks in the USA often explicitly introduce both Leibniz and "prime" notation, and then use them both. Perhaps that is a function of publishers wanting to go to the widest audience and hence throwing in the kitchen sink. Anyway, Krantz is beating a straw man to some extent, though it's not an unimportant point - you have to really be able to commit IF you are going to go against a book's notation, and it should only be done under duress. $\endgroup$
    – kcrisman
    Dec 15, 2016 at 1:19
  • $\begingroup$ Krantz's argument is fallacious for two reasons: (1) few instructors have the freedom as individuals to choose their own text; (2) for a subject like freshman calc, there is a finite number of texts to choose from, and most of them are extremely similar, because the publishers think it's good business to cover the common denominator. $\endgroup$
    – user507
    Feb 11, 2020 at 22:19
0
$\begingroup$

I think some reasonable fluency is needed. But I would mostly stick with dy/dx and y'.

I would also avoid using the most obscure (least used) notation when you are in the middle of introducing a new derivation or something complicated (e.g. integration by parts). Try to stay with something familiar in those cases (probably dy/dx).

$\endgroup$
-1
$\begingroup$

If one ponders seriously the question "If $d/dx$ is an operator, on what does it operate?", one comes to the conclusion that it doesn't operate on functions in the modern sense of the word (i.e. objects that map elements from a subset of $\mathbb{R}$ to elements of $\mathbb{R}$). Instead, it operates on objects of a different kind, for which the historically correct name is functions of $x$.

(Warning: although we use these other objects all the time, it seems that they were never formalized in modern mainstream mathematics. So we are in the awkward situation of teaching something which we can't really make precise. But I'll just pretend that we all know what a functions of $x$ is.)

If we compare the prime and the $d/dx$ notations from this perspective, we see that they are not two different notations for the same thing, but operators on two different spaces. Eulers prime operates on (modern) functions, while Leibniz' $d/dx$ operates on functions of $x$.

This is the convention I currently use when I teach. So for example I would never write $(x^2+1)'$, since $x^2+1$ is not a function (in the modern sense), but a function of $x$. Instead I write $\frac{d}{dx}(x^2+1)$. (There are other reasons for avoiding the prime notation when differentiating functions of $x$, like the impossibility to do the chain rule with it.). Similarly, if $f$ is a function like $x\mapsto (x^2+1)$ I would never write $\frac{df}{dx}$ since $f$ is not a function of $x$. Instead I write $f'$. On the other hand, if $g$ is a function depending on a parameter, like $x\mapsto x^2+c$, then it does make sense to write $\frac{dg}{dc}$, but this is not the same as $g'$.

$\endgroup$
7
  • $\begingroup$ Care to explain the downvote? $\endgroup$ Sep 27, 2018 at 6:45
  • $\begingroup$ This seems like one possible point of view on the interpretation of d/dx, but the question isn't asking for an interpretation of d/dx. $\endgroup$
    – user507
    Feb 11, 2020 at 22:21
  • 1
    $\begingroup$ @BenCrowell My point was the one needs at least two different notations for two different purposes, but probably Jordans answer is easier to understand for most people here. Btw. you say it's one possible point of view. Do you know another one? $\endgroup$ Feb 11, 2020 at 23:34
  • $\begingroup$ Also, I explain in more detail which derivative notations are ok and which are not here. $\endgroup$ Feb 11, 2020 at 23:39
  • 1
    $\begingroup$ Regarding $dg/dc$ in the last sentence, notice that $dg/dc$ is (like $g'$) another function, the constant function with value $1$ (so different from the function $g'$, as you noted). $\endgroup$ Aug 20, 2020 at 22:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.