I am studying Linear Algebra Done Right, chapter 2 problem 6 states:

Prove that the real vector space consisting of all continuous real valued functions on the interval $[0,1]$ is infinite dimensional.

My solution:

Consider the sequence of functions $x, x^2, x^3, \dots$ This is a linearly independent infinite sequence of functions so clearly this space cannot have a finite basis. However this prove relies on the fact that no $x^n$ is a linear combination of the previous terms. In other words, is it possible for a polynomial of degree $n$ to be equal to a polynomial of degree less than $n$. I believe this is not possible, but does anyone know how to prove this? More specifically, could the following equation ever be true for all $x$?

$x^n = \sum\limits_{k=1}^{n-1} a_kx^k$ where each $a_k \in \mathbb R$


Then the polynomial $\displaystyle x^n-\sum_{k=0}^{n-1}a_kx^k$ would have infinitely many roots, but it can have $n$, at most.

Another way of dealing with this problem is based upon defining polynomials (in one variable $x$) as expressions of the type $a_0+a_1x+a_2x^2+\cdots+a_nx^n$, where $n\in\{0,1,2,\dots\}$ and each $a_n$ is real. Under this definition, the polynomial $a_0+a_1x+a_2x^2+\cdots+a_nx^n$ is equal to the polynomial $b_0+b_1x+a_2x^2+\cdots+b_nx^n$ if and only if the coefficients are equal, that is, if and only if $a_0=b_0$, $a_1=b_1$, and so on. Under this definition, the problem discussed here is trivial.

What did I prove above then? Well, for each $P(x)\in\mathbb{R}[x]$, there is a corresponding polynomial function from $\mathbb R$ into $\mathbb R$. What I proved above is that this correspondence is one-to-one — when we are dealing with $\mathbb R$. It is still one-to-one if we are dealing with any field with charactristic $0$, such as $\mathbb Q$ or $\mathbb C$. But this is not true in general. For instance, if our field is $\mathbb{F}_2$, then $x$ and $x^2$ are distinct polynomials. But they correspond to the same polynomial function.

  • 2
    $\begingroup$ "Equations" can have infinitely many solutions, but nonzero polynomials cannot. $\endgroup$ – Bill Dubuque Aug 6 '17 at 14:11
  • 1
    $\begingroup$ @Joe A reference? Why? Isn't my proof clear? $\endgroup$ – José Carlos Santos Aug 6 '17 at 14:14
  • 1
    $\begingroup$ @Joe See this answer for a proof of the root bound for polynomials. $\endgroup$ – Bill Dubuque Aug 6 '17 at 14:16
  • 1
    $\begingroup$ @Joe You can find several proofs of the Fundamental theorem of algebra. $\endgroup$ – Sahiba Arora Aug 6 '17 at 14:18
  • 3
    $\begingroup$ @ZubinMukerjee Jokes aside, I really do mean what I wrote. We should expect statements of results to be precise, but not names of results (else we'd be stuck with monstrous names like "the root bound for nonzero formal univariate polynomials over a commutative integral domain"). $\endgroup$ – Bill Dubuque Aug 6 '17 at 14:59

By definition, $x^n + \sum\limits_{k=0}^{n-1}a_k x^k$ is a polynomial of degree $n$. A polynomial of degree $n$ can have at most $n$ zeros in any field, and $[0,1]$ has more than $n$ elements.

  • 4
    $\begingroup$ nonzero polynomai .... $\endgroup$ – Bill Dubuque Aug 6 '17 at 14:13
  • 7
    $\begingroup$ @BillDubuque pedantic, but you are right $\endgroup$ – Zubin Mukerjee Aug 6 '17 at 14:18

Two identical functions have identical derivatives. Differentiate your equality $n$ times and you will see the two sides cannot be identical.

  • $\begingroup$ Good point, but I have not studied derivatives on a rigorous level yet... so I can't really use that as part of my proof. $\endgroup$ – Joe Aug 6 '17 at 15:01

Suppose $x^n = \sum\limits_{k=1}^{n-1} a_kx^k$. Consider what happens when $x$ is much larger than $|a_1|+\ldots+|a_{n-1}|$. By the triangle inequality, you will get a contradiction. Essentially, you are showing that $x^n$ grows faster than any degree $n-1$ polynomial...


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.