Suppose I have a vectorvalued function of four variables, $\mathbf{F}:\mathbb{R}^4\rightarrow \mathbb{R}^3$, such that $\mathbf{F}(x,y,z,t)$.

Now suppose $x,y,z$ are function of $t$, such that $\mathbf{F}(x(t),y(t),z(t),t)$. Is $\mathbf{F}$ still a function of four variables, $\mathbb{R}^4\rightarrow \mathbb{R}^3$?

A similar question:

Suppose I instead have a vectorvalued function of three variables, $\mathbf{F}:\mathbb{R}^3\rightarrow \mathbb{R}^3$, such that $\mathbf{F}(x,y,z)$.

Now if $\mathbf{F}(x(t),y(t),z(t))$, is $\mathbf{F}$ still a function of three variables, $\mathbb{R}^3\rightarrow \mathbb{R}^3$?

  • $\begingroup$ $\mathbf{F}$ is defined as a function $\mathbb{R}^4\to\mathbb{R}^3$. The definition doesn't change. But you may think of $t\mapsto F(x(t),y(t),z(t),t)$ which is a function $\mathbb{R}\to\mathbb{R}^3$ and a function in one variable. $\endgroup$ – Mundron Schmidt Jun 30 '17 at 10:58
  • 1
    $\begingroup$ Related: Can we add variables and create new functions? <> There's a serious implicit issue of mathematical grammar here: the phrases "such that $\mathbf{F}(x(t), y(t), z(t))$" and "if $\mathbf{F}(x(t), y(t), z(t))$" contain no verb. That is, "$\mathbf{F}(x(t), y(t), z(t))$" is neither a condition (as required by "such that...") nor a statement (as required by "if..."). Based on your earlier question, I think this is not merely an incidental aspect of the question. $\endgroup$ – Andrew D. Hwang Jun 30 '17 at 11:08
  • $\begingroup$ @AndrewD.Hwang What is the correct mathematical grammar in this question? English is not my native language. $\endgroup$ – JDoeDoe Jun 30 '17 at 11:53
  • 1
    $\begingroup$ I'd be inclined to write, "Now suppose $x$, $y$, $z$ are functions of $t$, and define $g(t) = \mathbf{F}(x(t),y(t),z(t),t)$." The equality sign acts as a verb; "$g(t) = \mathbf{F}(x(t),y(t),z(t),t)$" is a logical condition. (By contrast, "such that $\mathbf{F}(x(t),y(t),z(t),t)$" is analogous to writing "such that shoes" or "such that green".) Doing so should help clarify that $\mathbf{F}$ is a function of four variables, but $g$ is a function of just one variable (as John Hughes explains). $\endgroup$ – Andrew D. Hwang Jun 30 '17 at 12:10
  • 1
    $\begingroup$ Similarly, writing "if $g(t) = \mathbf{F}(x(t),y(t),z(t),t)$, ..." clearly separates the two functions $g$ (a function of one variable) and $\mathbf{F}$ (a function of three variables) whose identification seems to be the source of confusion. :) $\endgroup$ – Andrew D. Hwang Jun 30 '17 at 12:11

$F$ is still a function of four variables. When you write the thing you wrote, you're really defining something new implicitly (in a way that physicists often do, and without renaming, which tends to confuse lots of other folks).

For instance, I'd write things this way:

Given \begin{align} F& :\Bbb R^4 \to \Bbb R^3 \\ x& :\Bbb R \to \Bbb R \\ y& :\Bbb R \to \Bbb R \\ z& :\Bbb R \to \Bbb R \\ \end{align} let \begin{align} G& :\Bbb R \to \Bbb R^3: t \mapsto F(x(t), y(t), z(t), t). \end{align} Now it's evident that $G$ is a function of one variable, arrived at by composing several functions.

As I said, a physicist will sometimes write $F(t) = F(x(t), y(t), z(t), t)$ (or something like that), using the name "F" for both the 4-argument function $F$ and the one-argument function $G$. A typical example of this is "We have a time-varying electric field $E$ on all of 3-space, and a particle whose trajectory is given by functions $x,y,z$ of time. Then the electric field at time $t$ [implicit meaning: the electric field, at the point where the particle happens to be at time $t$] is $E(t) = E(x(t), y(t), z(t), t)$. And the derivative of this is just $E'(t) = \ldots$." I always find things like this baffling

  • $\begingroup$ Thanks! First you write $F :\Bbb R^4 \to \Bbb R$, shouldn't it be $F :\Bbb R^4 \to \Bbb R^3$? And also, is it neccesary to separately specify the domain and codomain of $x,y,z$? I mean, aren't they "included" in $\Bbb R^4$? $\endgroup$ – JDoeDoe Jun 30 '17 at 15:59
  • $\begingroup$ Yes -- I failed to see "vectorvalued" in your original note, and though "real valued" (even though my example used the electric FIELD). I'll edit. $\endgroup$ – John Hughes Jun 30 '17 at 17:05

I'm going to focus on the 4-variable case. The 3-variable case is similar.

In general, a function of $n$ variables is a function that takes $n$ arguments. In particular, a function with domain $\Bbb R^n$ is a function of $n$ variables.

On its own, $F$ is a function of 4 variables because it is a map from $\Bbb R^4$ to $\Bbb R^3$.

What you are looking at is what happens when $F$ is composed with another function. Let $\varphi: \Bbb R \to \Bbb R^4$ such that $t \to (x(t), y(t), z(t), t)$. Then $\varphi$ is a function of one variable (in this case $t$) which outputs vectors in $\Bbb R^4$. Then, the function $F(x(t), y(t), z(t), t)$ is just the function composition $(F \circ \varphi)(t)$. This function is a map from $\Bbb R$ to $\Bbb R^3$:

$$ \Bbb R \underbrace{\mapsto}_{\varphi} \Bbb R^4 \underbrace{\mapsto}_{F} \Bbb R^3 $$

So $F$ (which corresponds to $F(x,y,z,t)$) is a function of $4$ variables, but $F \circ \varphi$ (which corresponds to $F(x(t),y(t),z(t),t)$ ) is a function of one variable.


The number of variables is the number of elements you need to change to change the value of your function. In your case, knowing that $x,y,z$ are functions of $t$, you are implicitly saying that $\bf F$ can be viewed as a function of $t$ only too.

In symbols: you have a function of four variables $$\mathbf{F}:\bf{R}^4\longrightarrow \mathbf{R}^4$$ when the variables are $x,y,z,t$. But you also have the functions $x,y,z:\mathbf{R}\longrightarrow \mathbf{R}$ defined by $t\mapsto x(t),y(t),z(t)$ respectively, so you have the function $$\mathbf{R}\longrightarrow \mathbf{R}^4$$ defined by $t\mapsto (x(t),y(t),z(t),t)$. The composite $$\mathbf{R}\longrightarrow \mathbf{R}^4\overset{\mathbf{F}}{\longrightarrow} \mathbf {R}^4$$ is defined by $$t\mapsto (x(t),y(t),z(t),t)\mapsto \mathbf{F}(x(t),y(t),z(t),t)$$ and realises $\mathbf{F}$ as a function of one single variable.

Edit: about derivatives

With the above notations and assuming that $\mathbf{F}$ (that I will now call $F$) has the minimum required regularity (for instance, it is differentiable partially with respect to all the variables) you can take the partial derivatives: $$\frac{\partial F}{\partial x}, \frac{\partial F}{\partial y}, \frac{ \partial F}{\partial z}, \frac{\partial F}{\partial t}$$ Moreover, if even your functions $t\mapsto x(t),y(t),z(t)$ are differentiable with respect to $t$, you can define the derivatives $ \frac{dx}{d t}, \frac{d y}{d t}, \frac{dz}{dt}$; formally speaking, you are doing the same also for the last variable $t$, considering the identity function $t\mapsto t$ whose derivative is $1$. When you want to consider the total derivative of $F$ with respect to $t$, you are just differentiating the composite function $t\mapsto (x(t),y(t),z(t),t)\mapsto F(x(t),y(t),z(t),t)$ and this, recalling the derivative rule for composite functions, amounts to consider: $$ \frac{d F}{d t} = \frac{\partial F}{\partial x} \frac{dF}{dt}+ \frac{\partial F}{\partial y} \frac{d y}{dt}+ \frac{\partial F}{\partial z} \frac{d z}{d t} + \frac{\partial F}{\partial t}\cdot 1$$

In particular, note that $\partial F/\partial t\neq dF/dt$.

  • $\begingroup$ Thanks! I guess the partial derivative of $\mathbf{F}(x(t),y(t),z(t),t)$ is nonsense? So I can only take the ordinary derivative wrt $t$, $\frac{d}{dt}\mathbf{F}(x(t),y(t),z(t),t)$? What is the proper prime notation for this, $\mathbf{F}(x'(t),y'(t),z'(t),t')$, $\mathbf{F}'(x(t),y(t),z(t),t)$ or something else? $\endgroup$ – JDoeDoe Jun 30 '17 at 16:10
  • $\begingroup$ Dear @JDoeDoe, partial derivatives are perfectly defined if there is enough regularity for $F$; please see my edit $\endgroup$ – Caligula Jul 3 '17 at 9:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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