1

In An Introduction to Abstract Mathematics (Bond, Keane) they give this

$$\forall \, x \in \mathbb{R},\; \exists \, y \in \mathbb{R} \: \ni \: y^3 = x$$

(They say $\ni$ is "such that.") with the descriptive, Every real number has a cube root. Then they say this is a consequence of the Intermediate Value Theorem of Calculus. Can I get some detail on why this is?

Update

I guess I'm initially after the basic question of what is it about the set notation of the open sentence makes this a question of continuous-ness? That's what I need to understand first. I see "for all $x$ there exists a $y$ such that $y^3 = x$ and wonder why we're suddenly talking about continuous-ness and the IVT. How did we know to do that? It's like a primitive seeing water and making the leap that it's made of the stuff we breath and the stuff burning in the sun. Huh? all us other primitives say. And Every real number has a cube root doesn't help. Again, how is that IVT?

147pm
  • 920
  • The Intermediate Value Theorem applied against the function $f$ includes the premise that $f$ is a continuous function throughout the pertinent interval. $f(x) = x^{(1/3)}$ is known to be a continuous function for all finite real numbers $(x)$. – user2661923 Jun 17 '21 at 04:53
  • My main problem was the leap from set theory, set notation to functions. For beginners they're usually treated separately in texts, or one, then the other exclusively. For example, I presume the notion of a function being continuous in this example is an assumption taken from the fact that $x$ is universally quantified and $y$ being existentially quantified? IOW, what about the set notation tells me I'm dealing with a continuous function $f(x) = x^{1/3}$? – 147pm Jun 17 '21 at 15:56
  • see my answer... – user2661923 Jun 18 '21 at 02:34
  • @147pm At the point you're proving this statement, it's not appropriate to talk about a function $f(x)=x^{1/3}$ (it begs the question as the statement is trivial once we know such a function exists with domain $\mathbb{R}$, regardless of continuity). The function whose continuity matters, and whose existence we're not currently trying to prove, is $g(x) = x^3$. – Brian Moehring Jun 18 '21 at 03:11
  • @BrianMoehring Nice catch. Makes my answer of only arguable value. – user2661923 Jun 18 '21 at 03:13
  • I guess I'm still after the basic question of what is it about the set notation of the original open sentence makes this a question of continuous-ness? That's what I need to understand first. I see "for all $x$ there exists a $y$ such that $y^{3} = x$ and wonder why we're suddenly talking about continuousness and the IVT. How did we know to do that? It's like a primitive seeing water and making the leap that it's made of the stuff we breath and the stuff burning in the sun. Huh? all us other primitives say. And Every real number has a cube root doesn't help. Again, how is that IVT? – 147pm Jun 18 '21 at 03:19
  • IVT has the rough form $$\forall x \in (a,b), \exists y \in \mathbb{R}, f(y)=x$$ Compare that to your problem. – Brian Moehring Jun 18 '21 at 03:28
  • @Brian Moehring: Aha! Now we're getting somewhere! Can you direct me to a good intro text that explores this set theory-to-analysis connection? I would have never known, guessed this. – 147pm Jun 18 '21 at 03:30
  • @Brian Moehring: Or maybe I should put this aside and just get through Tom Apostol first... – 147pm Jun 18 '21 at 03:36

2 Answers2

3

I will relabel your notation. For all $a\in\mathbb{R}$, there exists $b\in\mathbb{R}$ such that $b^{3}=a$.

Suppose first that $a>0$. Define $f(x)=x^{3}-a$. Then $f$ is continuous on $\mathbb{R}$. Then $f(a+1)=(a+1)^{3}-a>0$ and $f(-(a+1))=(-a-1)^{3}-a<0$. By the intermediate value theorem, there exists $b\in\mathbb{R}$ such that $f(b)=0$, or equivalently, $b^{3}=a$.

The case for $a<0$ is similar.

The case when $a=0$ is trivial.

  • As I said in the above comment, I'm not clear on why, how the set notation becomes the guaranteed continuous function $f(x) = x^{(1/3)}$ how can we assume that by just looking at the set notation? Also, your step from defining $f(x) = x^3 - a$ to "Then $f$ is continuous." has me confused. Why? Can you direct me to more proof examples like yours? – 147pm Jun 17 '21 at 16:26
  • Just posted an answer to resolve the question of 147pm. – user2661923 Jun 18 '21 at 02:35
0

First, see the comment of Brian Moehring, following the original question. This indicates that the demonstration below, that $f(x) = x^{(1/3)}$ is a continuous function, is not pertinent.

So much analysis there, that I can't bring myself to delete it. However, I have added an Addendum to address the issue of the continuity of $(x^3).$


Responding to a comment/question from the OP:

Given $f:\Bbb{R}\to \Bbb{R}$ such that $f(x) = x^{(1/3)}.$

Demonstration that for all $x_0 \in \Bbb{R}, ~f(x)~$ is continuous at $x = x_0$:

Personally, I found the answers given here, here, and here to be somewhat sophisticated. It seems reasonable to (instead) provide a (primitive) $\epsilon, \delta$ proof.

Definition of continuity:

  • $f(x)$ is defined at $(x = x_0)$.
  • $\forall \epsilon > 0, ~\exists \delta > 0~$ such that $0 < |x - x_0| < \delta \implies |f(x) - f(x_0)| < \epsilon.$

Clearly, the first constraint above is satisfied, for all $x_0 \in \Bbb{R}.$

What the second constraint is saying, in effect, is that for each $(x_0)$, there exists a fixed finite limit $L$, such that $\lim_{x\to x_0} f(x) = L$ and that in all cases, $L = f(x_0)$.


Intermediate Results

Lemma 1:
$a < b \iff a^3 < b^3$.

Proof:
$g(x) = x^3 \implies g'(x) = 3x^2.$
Therefore, $g(x)$ is a strictly increasing function, throughout $\Bbb{R}.$

Lemma 2:

  1. $~~~\displaystyle 0 < \delta < x_0, ~~~0 < \epsilon < x_0^{(1/3)}, ~~~ \delta < \left[3x_0^{(2/3)}\epsilon - 3x_0^{(1/3)}\epsilon^2\right] ~~~\implies $
    $~~~\displaystyle \left[x_0^{(1/3)} - \epsilon\right] < (x_0 - \delta)^{(1/3)}.$

  2. $~~~\displaystyle 0 < \delta < x_0, ~~~0 < \epsilon, ~~~ \delta < \left[3x_0^{(2/3)}\epsilon\right] ~~~\implies $
    $~~~\displaystyle \left[x_0^{(1/3)} + \epsilon\right] > (x_0 + \delta)^{(1/3)}.$

Proof:
Using Lemma 1, I can show (for example) that $R < S$ by showing that $R^3 < S^3.$

$\displaystyle \left[x_0^{(1/3)} - \epsilon\right]^3 = x_0 - 3x_0^{(2/3)}\epsilon + 3x_0^{(1/3)}\epsilon^2 - \epsilon^3.$
By assumption, this is less than $(x_0 - \delta).$

Similarly,
$\displaystyle \left[x_0^{(1/3)} + \epsilon\right]^3 = x_0 + 3x_0^{(2/3)}\epsilon + 3x_0^{(1/3)}\epsilon^2 + \epsilon^3.$
By assumption, this is greater than $(x_0 + \delta).$


Proof of Second Constraint

From the premise, you have that
$$- \delta < (x - x_0) < \delta \iff (x_0 - \delta) < x < (x_0 + \delta).\tag{1}$$

From the conclusion you want to establish that
$$ - \epsilon < x^{(1/3)} - x_0^{(1/3)} < \epsilon \iff \left[x_0^{(1/3)} - \epsilon\right] < x^{(1/3)} < \left[x_0^{(1/3)} + \epsilon\right].\tag2$$

$\underline{\text{Case 1:}~x_0 > 0.}$

Use the following specifications: $$ \text{If} ~~\left[x_0^{(1/3)} > \epsilon\right] ~~\text{then}~~ \delta = \min\left\{ ~\frac{x_0}{2}, ~\left[x_0^{(2/3)}\epsilon - x_0^{(1/3)}\epsilon^2\right] ~\right\}.$$

$$ \text{Otherwise}~~ \delta = \min\left\{ ~\frac{x_0}{2}, ~\left[x_0^{(2/3)}\epsilon\right] ~\right\}.$$

It is assumed that $\epsilon > 0$ and that $\delta$ is as specified above.
From this, you know that $x_0 - \delta > 0.$

Using the specification, Lemma 1 and inequality 1 above, you have that
$\displaystyle 0 < (x_0 - \delta)^{(1/3)} < x^{(1/3)} < (x_0 + \delta)^{(1/3)}.$

Using the specification and Lemma 2, you know immediately that
$\displaystyle \left[x_0 + \delta\right]^{(1/3)} < \left[x_0^{(1/3)} + \epsilon\right].$

If $~~x_0^{(1/3)} - \epsilon \leq 0,~~$ then this immediately establishes that
$\displaystyle x_0^{(1/3)} - \epsilon < \left[x_0 - \delta\right]^{(1/3)}.$

Otherwise, the specification, with Lemma 2, establishes that
$\displaystyle x_0^{(1/3)} - \epsilon < \left[x_0 - \delta\right]^{(1/3)}.$

Therefore, the specification for $\delta$ establishes inequality (2) above.


$\underline{\text{Case 2:}~x_0 < 0.}$

Define $~y_0 = -x_0 \implies $

  • $y_0 > 0.$
  • $x_0^{(1/3)} = - y_0^{(1/3)}.$

For all $~x < 0,~$ define $~y = -x \implies $

  • $y > 0.$
  • $x^{(1/3)} = - y^{(1/3)}.$

Assume that $\epsilon > 0$.
Further assume that Case 1 is invoked, so that $\delta$ is specified so that whenever

$$- \delta < (y - y_0) < \delta \iff (y_0 - \delta) < y < (y_0 + \delta)\tag{3}$$

then, the following is satisfied:

$$ - \epsilon < y^{(1/3)} - y_0^{(1/3)} < \epsilon \iff \left[y_0^{(1/3)} - \epsilon\right] < y^{(1/3)} < \left[y_0^{(1/3)} + \epsilon\right].\tag4$$

Then, inequality 3 above is satisfied if and only if
$|y - y_0| < \delta \iff |x - x_0| < \delta.$

Further, inequality 4 above is satisfied is and only
$\left|y^{(1/3)} - y_0^{(1/3)}\right| < \epsilon \iff \left|x^{(1/3)} - x_0^{(1/3)}\right| < \epsilon.$

Therefore, continuity is established for all $x_0 < 0.$


$\underline{\text{Case 3:}~x_0 = 0.}$

Set $\displaystyle \delta = \frac{1}{2}\epsilon^3 \implies \delta^{(1/3)} < \epsilon,$ by Lemma 1.

Then, using Lemma 1 repeatedly,

$-\delta < x < \delta \implies$

$- \epsilon < -\delta^{(1/3)} < x^{(1/3)} < \delta^{(1/3)} < \epsilon.$

Therefore, continuity is established for $(x_0 = 0).$




Addendum
Let $h(x) = x^3.$
Prove that $h(x)$ is continuous everywhere.

First of all, I strongly suspect that this has been proven somewhere on MathSE. However, I could not find such a proof, which indicates that the OP will also have trouble finding such a proof.

Therefore, I will provide a proof below.

Analysis will be very similar to the analysis in the original answer.

Given:
$$- \delta < (x - x_0) < \delta \iff (x_0 - \delta) < x < (x_0 + \delta).\tag{5}$$

To Prove:
$$ - \epsilon < x^3 - x_0^3 < \epsilon \iff \left[x_0^3 - \epsilon\right] < x^3 < \left[x_0^3 + \epsilon\right].\tag6$$

This proof is much easier than the previous proof.

$\underline{\text{Case 1:} ~x_0 > 0.}$

Use the following specification: $$ \delta = \min\left\{ ~\frac{x_0}{2}, \frac{1}{2}, ~\left[\frac{\epsilon} {2 \times (3x_0^2 + 3x_0 + 1)}\right]~\right\}.$$

It is assumed that $\epsilon > 0$ and that $\delta$ is as specified above.
From this, you know that $x_0 - \delta > 0.$

Using the specification, Lemma 1 and inequality 5 above, you have that
$\displaystyle 0 < (x_0 - \delta)^3 < x^3 < (x_0 + \delta)^3.$

Further, since $\delta < 1$,
$(x_0 + \delta)^3 = x_0^3 + 3x_0^2\delta + 3x_0\delta^2 + \delta^3$
$< x_0^3 + \delta(3x_0^2 + 3x_0 + 1) < x_0^3 + \epsilon.$

Also, you have that
$\epsilon > \delta(3x_0^2 + 3x_0 + 1)$
$> 3x_0^2 \delta + 3x_0\delta^2 + \delta^3$
$> 3x_0^2 \delta - 3x_0\delta^2 + \delta^3$.

This implies that
$x_0^3 - \epsilon < x_0^3 - 3x_0^2 \delta + 3x_0\delta^2 - \delta^3 = (x_0 - \delta)^3.$

Therefore, when $x_0 > 0$, inequality (6) above is established, so $h(x) = x^3$ is continuous at $x_0$.


$\underline{\text{Case 2:} ~x_0 < 0.}$
Adopting the nomenclature of Case 2 in the original answer, the analysis is virtually identical to that of Case 2.

That is,
$x_0^3 = -y_0^3.$
$x^3 = -y^3.$
$|x - x_0| = |y - y_0|$.
$|x^3 - x_0^3| = |y^3 - y_0^3|.$

Therefore, $h(x) = x^3$ is also continuous at $x_0 < 0.$


$\underline{\text{Case 3:} ~x_0 = 0.}$

Use the following specification: $$ \delta = \min\left\{ \frac{1}{2}, ~\left[\frac{\epsilon} {2}\right]~\right\}.$$

It is assumed that $\epsilon > 0$ and that $\delta$ is as specified above.

Using the specification, Lemma 1 and inequality 5 above, you have that
$\displaystyle - \delta^3 < x^3 < \delta^3.$

Further, since $\delta < 1, ~~\delta^3 < \delta < \epsilon.$

Therefore,
$-\epsilon < -\delta < -\delta^3 < x^3 < \delta^3 < \delta < \epsilon.$

Therefore, when $x_0 = 0$, inequality (6) above is established, so $h(x) = x^3$ is continuous at $x_0 = 0.$

user2661923
  • 35,619
  • 3
  • 17
  • 39