MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let, $V$ be a vector space over a field $K.$ Let, $T$ be a function from $V$ to $V$ such that $T(kX) = kT(X)$ for all $k \in K$ and for all $X \in V$ and also $T(k + X) = T(X)$ for all $k \in K$ and for all $X \in V$.

If $X = (x_1, x_2, \ldots, x_n)$, then $k + X = (k + x_1, k_ + x_2, \ldots, k + x_n)$.

I would like to know how to study the behavior of $T$ and its effect on the vector space $V$ be studied.

share|cite|improve this question
$k \in K$ is a scalar and $X$ is an element of your vector space $V$, so what does $k + X$ mean? – Pete L. Clark Sep 28 '10 at 11:42
This could use some retagging... «linear-algebra» maybe? – Mariano Suárez-Alvarez Sep 28 '10 at 11:51
If $X = (x_1, x_2, \ldots, x_n)$, then $k + X = (k + x_1, k_ + x_2, \ldots, k + x_n)$. – debapriyay Sep 28 '10 at 12:09
Vector spaces have no choosen basis. – Martin Brandenburg Sep 28 '10 at 12:22
@MB: indeed. Nor do they need to be finite-dimensional. – Pete L. Clark Sep 28 '10 at 12:37
up vote 2 down vote accepted

It seems like the question means to set $X=K^n$. The first condition means that $T$ is homogeneous, and the second that $T(k1+x)=T(x)$ for all $x\in X$ and $k\in K$, where $1=(1,\cdots,1)\in X=K^n$.

As rpotrie says, move to projective space $PK^{n-1}$. This is the set of lines through the origin, or the $K^n$ mod the equivalence relation that $x \sim kx$ for any $k\not=0$. Write the equivalece class of $(x_1,\cdots,x_n)$ as $[x_1,\cdots,x_n]$. As $T$ is homogeneous, it drops to a map $T:PK^{n-1}\rightarrow K^n$. The second condition is just that $T [x_1+k,\cdots,x_n+k] = T[x_1,\cdots,x_n]$ for any $k\in K$. This is equivalent to $T[0,x_2-x_1,\cdots,x_n-x_1] = T[x_1,\cdots,x_n]$.

So it seems to me that $T$ is completely determined by some map (which need satisfy no further conditions at all) $PK^{n-2}\rightarrow K^n$.

share|cite|improve this answer
Nice. I could't express this fact correctly. This defines exactly all possible maps. – rpotrie Sep 28 '10 at 12:47
$X = K'^{n}$, where $K'$ is a field and $K \subset K'$. This can also be the case. Now, you talked about an equivalence relation, one question is if $K'$ is a finite field, then can it be possible to enumerate all the equivalence classes, or the number of equivalence classes. – debapriyay Sep 28 '10 at 12:56
Well, sure, if K is finite then so is $X=K^n$, and so $PK^{n-2}$ is finite. It's just an exercise to write down explicit representatives. If you don't know how to do this, read an introduction to projective space. – Matthew Daws Sep 28 '10 at 13:42

If $X=(1,1,...,1)$ then $k+X= (k+1)X$ so we get that $T(X)=0$ (and this holds in the subspace generated by $X$), otherwise $(k+1)=1$ for every $k$.

You can proyectivize your function from the first hipothesis and get a function (not necesarilly continuous) $f: P(V) \to P(V)\cup \{0\}$ such that $f([(1,1,...,1)])=0$. Also, you get that $f([X+(1,...1)])= f([X])$ so you get that the function is constant under the orbit of adding $[(1,...1)]$. Since adding $(k,...,k)$ depends on the point I would guess that under some conditions on the field this implies that the function is constant in "circles" for $[X] \neq [(1,...,1)]$.

For example, if $T$ is continuous, and $V=\mathbb{R}^d$ over $\mathbb{R}$, I believe you get that $T=0$.

share|cite|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.