In Linear Algebra Done Right, it gives an example of linear maps as follow:

multiplication by $x^2$

Define $T\in L(P(R),P(R))$ by $$(Tp)(x) = x^2p(x)$$ for $x \in \mathbb{R}$

My attempt: $$(Tp)(x) = x^2p(x)$$ $$(Tp)(x+y) = x^2p(x+y) = x^2p(x) + x^2p(y) = (Tp)(x) + (Tp)(y)$$ I am not sure whether I can do $p(x+y) = p(x) + p(y)$. It seems I cannot. Right?

$$(Tp)(\lambda x) = x^2p(\lambda x) = \lambda x^2p(x) = \lambda(Tp)(x)$$ I am also not sure whether I can do $p(\lambda x) = \lambda p(x)$.

  • 1
    $\begingroup$ What's $P(R)$? Polynomials with coefficients in $\Bbb R$? $\endgroup$ – user608030 Jan 12 at 1:03
  • $\begingroup$ @ZacharySelk I think it is. $\endgroup$ – JOHN Jan 12 at 1:04
  • 1
    $\begingroup$ If so your attempt is very wrong. $\endgroup$ – user608030 Jan 12 at 1:04
  • 6
    $\begingroup$ You're on a wrong track. You have to think about $(T(p+q))(x)$, not $(Tp)(x+y)$. The linearity is about the addition and scalar multiplication in the space of polynomials, not in the real numbers where you are evaluating the polynomials. $\endgroup$ – Ethan Bolker Jan 12 at 1:05

Remember, you are adding polynomials, not independent variables like $x$ and $y$. If you want to check additivity, you need to verify that, given polynomials $p$ and $q$, that $$T(p + q) = Tp + Tq.$$ Now, $T(p + q), T(p)$, and $T(q)$ are all polynomials themselves, so if you want to show that they are equal, you can show that they are equal at some arbitrary point $x$. That is, show $$(T(p + q))(x) = (Tp)(x) + (Tq)(x).$$ Note that, by definition of $T$, we have that $(Tp)(x) = x^2 p(x)$ for all $x$, and similar for $q$. So, equivalently, $$x^2(p + q)(x) = x^2 p(x) + x^2 q(x).$$ All I've done is use the definition of $T$. Then, we need to use the definition of $+$ for polynomials, which is $(p + q)(x) = p(x) + q(x)$. Note the $+$ on the left is the $+$ being defined, and adds functions. The $+$ on the right is the usual addition between real scalars, as $p(x)$ and $q(x)$ are actually just real numbers (although, unknown real numbers, since we don't know $x$, or $p$ and $q$). Therefore, we have $$x^2 (p + q)(x) = x^2(p(x) + q(x)) = x^2p(x) + x^2q(x),$$ using the definition of polynomial addition, and the distributive law of real numbers.

Have another go at scalar multiplication. Remember the definition of polynomial scalar multiplication: $$(\lambda \cdot p)(x) = \lambda \cdot p(x)$$ Again, on the left is the operation being defined, between a real number and a real valued polynomial. The operation on the right is ordinary multiplication between real numbers.


You are mistaken on the definition of $T$. Since $T$ is an application from $P(R)$ into $P(R)$ (I assume that $P(R)$ stands for the vector space of polynomials with real coefficients), the input is a polynomial $p$ and the output is also a polynomial function $T(p)$, usually written as $Tp$.

What you have to show in order to prove that $T$ is linear is: 1. for any $p$ and $q$ in $P(R)$, $T(p+q)=T(p)+T(q)$. 2. for any $p$in $P(R)$ and any real number $c$, $T(cp)=cT(p)$.

Then... what is that $x$ in the definition of $T$? Since $f=T(p)$ is itself a function, it is defined by expliciting its evaluation $f(x)$ at every real number $x$.

To sum up, $T$ is an application (= a function) that transforms a polynomial function $p$ into a new ploynomial application $f=T(p)$.

Let's prove the point $1$ above (point $2$ is proved similarly). Let $p$ and $q$ be two polynomials. Then, we need to check the equality of functions $T(p+q)=T(p)+T(q)$. For, consider a real number $x$. We have

$$T(p+q)(x)=x^2(p+q)(x)=x^2(p(x)+q(x))=x^2p(x)+x^2q(x) = T(p)(x) + T(q)(x)$$

that is $T(p+q)(x) = (T(p)+T(q))(x)$. The claim $T(p+q)=T(p)+T(q)$ is proved! Make sure you understand each step of the computation above, this kind of game with definitions/formulas is important for abstract algebra.

Note: the notation $T(p)(x)$ means that the function $T(p)$ is evaluated at $x$. If you let $f=T(p)$, then it is just $f(x)$.

Note 2: in general, polynomials are not linear applications, so we don't have $p(x+y)=p(x)+p(y)$ and $p(cx)=cp(x)$.

Note 3: for simplicity, I identified polynomials and polynomial functions, though they are not exactly the same. It is fine because we are considering real coefficients here.


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.