Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $V$ vector space of all polynomials $p(t) = a_0 + a_1t + \cdots + a_nt^n$,$\forall n \in\mathbb{N}$ and $a_0,\ldots,a_n \in\mathbb{R}$. How can I prove that $ \gamma = \{1,t,t^2,\ldots\}$ is a basis of $V$, and use it to find a linear transformation $T:V \rightarrow V$ such that $T$ is surjective but not injective. Is $T(x) = x^2$ an example of this transformation? Some help please.

share|improve this question
$T(x)=x^2$ is either not a linear map or not defined on all of $V$. You still have a finite dimensional vector space, so linear maps are injective if and only if they are surjective. If you took $V$ to be all polynomials (with no bound on the degree), then differentiation would give you the map you want. –  Jessica B Jul 8 '14 at 21:15
@kahen the question has been edited several times; the earlier comment no longer applies but was helpful before. –  Omnomnomnom Jul 8 '14 at 21:32

3 Answers 3

up vote 3 down vote accepted

You seem to have a fundamental misunderstanding about this question.

$T$ is a transformation from the set of polynomials on $t$ to the set of polynomials on $t$. So, the input to $T$ should be a polynomial, and the output should be some other polynomial. Two common linear transformations are differentiation and integration from $t=0$. Namely, we can describe differentiation operator $T(p) = \frac{dp}{dt}$ by saying that if $p(t) = a_0 + a_1 t + \cdots + a_n t^n$, then $$ T[p(t)] = a_1 + 2a_2 t + \cdots + na_n t^{n-1} $$ Similarly, we can describe the operator $T(p) = \int_0^t p(x)\,dx$ by saying that if $p(t) = a_0 + a_1 t + \cdots + a_n t^n$, then $$ T[p(t)] = a_0t + \frac {a_1}2 t^2 + \cdots + \frac {a_n}{n+1} t^{n+1} $$ Try to prove that the first operator is surjective, but not injective, while the second is injective, but not surjective.

As for your first question: you should look up the definition of a basis, and verify that $\gamma$ satisfies that definition.

Is it true that every member $v \in V$ can be written as a finite sum $$ v = \sum_{i=1}^n a_i v_i $$ where $v_i$ are elements of $\gamma$? Is it true that if $$ \sum_{i=1}^n a_i v_i = 0 $$ for some $a_i$, then all $a_i$ have to be equal to zero? If so, then $\gamma$ is a basis. If you show that this is the case, then you have proven $\gamma$ to be a basis of $V$.

share|improve this answer
The surjective part is easy. Given a polynomial $a_{0} + a_{1}t + a_{2}t^{2} + \cdots + a_{n}t^{n}$, which polynomial is being sent to this polynomial under the differentiation transformation $T$? Well, any polynomial of the form $C + a_{0}t + a_{1}\frac{t^{2}}{2} + \cdots + a_{n-1}\frac{t^{n}}{n} + a_{n}\frac{t^{n + 1}}{n + 1}$, where $C$ is any constant. Why? Differentiate $C + a_{0}t + a_{1}\frac{t^{2}}{2} + \cdots + a_{n-1}\frac{t^{n}}{n} + a_{n}\frac{t^{n + 1}}{n + 1}$ and –  Knight Jul 15 '14 at 4:37
you will get exactly $a_{0} + a_{1}t + a_{2}t^{2} + \cdots + a_{n}t^{n}$. Now, I said $C$ can be any constant, and that polynomial will still be sent to $a_{0} + a_{1}t + a_{2}t^{2} + \cdots + a_{n}t^{n}$. In particular, if $C_{1}$ and $C_{2}$ are two distinct constants, both $C_{1} + a_{0}t + a_{1}\frac{t^{2}}{2} + \cdots + a_{n-1}\frac{t^{n}}{n} + a_{n}\frac{t^{n + 1}}{n + 1}$ and $C_{2} + a_{0}t + a_{1}\frac{t^{2}}{2} + \cdots + a_{n-1}\frac{t^{n}}{n} + a_{n}\frac{t^{n + 1}}{n + 1}$ are being sent to the same polynomial, even though –  Knight Jul 15 '14 at 4:37
they are distinct polynomials (they differ by the constants $C_{1}$ and $C_{2}$). So this is an example of $T(a) = T(b)$ but $a \neq b$, which means $T$ is not injective. But i stuck when i trid to prove that the differentiation operator i surjective but not injective, some help please –  Knight Jul 15 '14 at 4:38
@Knight you've shown that the differentiation operator is surjective. In order to show that it is not injective, find two distinct polynomials with the same derivative. –  Omnomnomnom Jul 15 '14 at 16:25

Since every polynomial is by definition a linear combination of $1,t,t^2\dots$ to prove they form a basis you just need to show that the powers are linearly independent, i.e. $a_01+a_1t+\dots a_nt^n=0$ implies that all $a_i=0$. The simplest way is to differentiate the equality $n$ times, which eliminates all the terms except the last one, leading to $n!\,a_n=0$ and $a_n=0$. Repeating the process shows that $a_{n-1}=0$, $a_{n-2}=0$ and so on.

In finite dimensional spaces there can be no linear transformation which is surjective but not injective, it follows from the Rank-Nullity theorem. Your suggested transformation is not even linear. If you consider the space of all polynomials, which is infinite dimensional with basis $\{1,t,t^2...\}$, then such a transformation can be defined on the basis as $T(t^n):=nt^{n-1}$ for $n>0$, and $T(1):=0$. Then extend it by linearity to all polynomials by $T(p):=a_0T(1)+a_1T(t)+\dots a_nT(t^n)$. It is surjective since every basis element is in the range, but not injective since $0$ and $1$ are mapped into $0$. This transformation happens to be the derivative $T(p)=p'$.

share|improve this answer
The Fredholm alternative is overkill; normally I refer to the rank-nullity theorem –  Omnomnomnom Jul 8 '14 at 21:21
Furthermore, the Fredholm alternative doesn't apply in the case of $\mathbb R[x] \cong c_c$ since it's not complete. –  kahen Jul 8 '14 at 21:34
@Omnomnomnom Changed as suggested. It does apply to polynomials of degree at most $n$ that the original question asked about. –  Conifold Jul 8 '14 at 21:34

An alternative to the differentiation transformation is the shift,

$T(p(t)) = \dfrac{p(t)-p(0)}{t}$, which sends $a_0+a_1t+\cdots +a_nt^n$ to $a_1+a_2t+\cdots+a_nt^{n-1}$, i.e., it removes the constant term and decreases the other powers by $1$.

For injective but not surjective, there is the shift in the other direction,

$S(p(t)) = tp(t)$, which increases the powers by $1$. Or $S(p(t))=p(t^2)$, which doubles all the powers.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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