Understanding the Derivative of y with Respect to x

  • Thread starter Thread starter Karol
  • Start date Start date
  • Tags Tags
    Derivative
Karol
Messages
1,380
Reaction score
22

Homework Statement


Snap2.jpg

Isn't the derivative of y with respect to x Defined as ##~\frac{dy}{dx}##?
What and how do i have to prove?

Homework Equations


The chain rule:
$$\frac{dy}{dx}=\frac{dy}{du}\frac{du}{dx}$$

The Attempt at a Solution


$$\frac{dy}{dx}=\frac{dy/dt}{dx/dt}$$
 
Physics news on Phys.org
What a horrible question. dy and dx are not numbers, and dx/dy is not a ratio.

What you can probably do is studying d y(x(t))/dt and then abusing the notation sufficiently until you get dy/dx, showing that t does not matter.
 
  • Like
Likes fresh_42
What is y[x(t)]? isn't it just y(t)?
For example, the example given in the book:
$$x=t+\frac{1}{t},~~y=t-\frac{1}{t}~\rightarrow~x^2-y^2=4$$
isn't ##~y[x(t)]=t-\frac{1}{t}~##?
 
Karol said:
What is y[x(t)]? isn't it just y(t)?
For example, the example given in the book:
$$x=t+\frac{1}{t},~~y=t-\frac{1}{t}~\rightarrow~x^2-y^2=4$$
isn't ##~y[x(t)]=t-\frac{1}{t}~##?
No. What is ##y(x)## if ##x## was just a real number?
 
mfb said:
What a horrible question. dy and dx are not numbers, and dx/dy is not a ratio.

What you can probably do is studying d y(x(t))/dt and then abusing the notation sufficiently until you get dy/dx, showing that t does not matter.

In Nonstandard Analysis, ##dy## and ##dx## are actual infinitesimals and ##dy/dx## really is a ratio; the "standard part" of the ratio equals the ordinary derivative.

However, I have serious doubts that the OP's course is using full-fledged non-standard analysis, but is instead likely using a form of heuristic argument that has been around since the days of Leibnitz.
 
  • Like
Likes mfb
fresh_42 said:
No. What is ##y(x)## if ##x## was just a real number?
$$y=\sqrt{(x-2)(x+2)}$$
 
Karol said:
$$y=\sqrt{(x-2)(x+2)}$$
This is not how ##y## is defined.
 
fresh_42 said:
No. What is ##y(x)## if ##x## was just a real number?
This is not how ##y## is defined.
If x were a real number then ##~y=x-\frac{1}{x}##
But what is y[x(t)]?
$$y[x(t)]=y\left[ t+\frac{1}{t} \right]=\left(t+\frac{1}{t} \right)-\frac{1}{\left(t+\frac{1}{t} \right)}\neq t-\frac{1}{t}$$
I don't understand
 
Last edited:
You have ##y(sth.)= sth. \,- \frac{1}{sth.}##. Next this something is ##x(t) = sth. = t + \frac{1}{t}##. What do we get by substitution of that something?
 
  • #10
fresh_42 said:
What do we get by substitution of that something?
We get the value of y, but y isn't a function of x, each (x and y) are functions of t. we just join the points (x,y) produced by t.
 
  • #11
Karol said:
We get the value of y, but y isn't a function of x, each (x and y) are functions of t. we just join the points (x,y) produced by t.
You have ##y(t)=t-\frac{1}{t}##. This means, if ##t## is replaced by ##x##, then we get ##y(x)=x-\frac{1}{x}##. And if then the function ##x=x(t)= t + \frac{1}{t}## is parameterized by ##t##, we get
$$
(y \circ x)(t)=y(x(t))=x(t) -\frac{1}{x(t)} = (t+\frac{1}{t})- \frac{1}{t+\frac{1}{t}}
$$
Now you can calculate either ##\frac{d}{dt}(y\circ x)(t)## as function of ##t## or use the chain rule and calculate
$$
\frac{d}{dt}(y\circ x)(t)= \frac{d}{dx}y(x) \cdot \frac{d}{dt}x(t) = \frac{d}{dx} (x-\frac{1}{x}) \cdot \frac{d}{dt} (t+\frac{1}{t})
$$
and then substitute ##x=t+\frac{1}{t}## and hope both quotients are equal.
 
  • #12
It is very tiring to prove that both these equations are equal.
And even so, what did i prove? that ##~\frac{d}{dx}y(x) \cdot \frac{d}{dt}x(t) = \frac{d}{dt}y[x(t)]##
Is it what the question asked? in this example:
$$y=\sqrt{(x-2)(x+2)}~\rightarrow~y'(x)=...$$
And i have to prove it equals dy/dx
I leave this question since i don't want to tire you and me, i think i won't get further.
Thank you fresh_42
 
  • #13
Karol said:
$$y=\sqrt{(x-2)(x+2)}$$

No: there are two solutions
$$y(x) = \pm \sqrt{(x-2)(x+2)}$$
This really is true: if you plot the points ##(x(t),y(t))## for ##t> 0## you will find they lie on the curve ##y = -\sqrt{(x-2)(x+2)}## for ##0 < t \leq 1## and on the curve ##y = + \sqrt{(x-2)(x+2)}## for ##1 \leq t < \infty##. This is one branch of the hyperbola ##x^2 - y^2 = 4##. For ##t < 0## you get the opposite branch of the hyperbola, opening out to the left of ##x = -2##.
 
  • Like
Likes Karol
  • #14
Thank you Ray, you all are specialists, they drew the plot in the book and it's as you've said
 
Back
Top