Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

Is there an automorphism of $\mathbb{R}^n$ (here it is seen as a vector space) that is not a linear mapping?

share|cite|improve this question
The answer is "no", but the way to prove that depends on your definition of automorphism, $\mathbb{R}^n$ and "linear mapping". – akkkk Apr 16 '12 at 13:57
You need to explain this better. Do you mean that "every automorphism of the topological group $\mathbb{R}$ is a linear mapping"? An automorphism in the category of vector spaces and linear mappings is by definition a linear mapping. – Xabier Domínguez Apr 16 '12 at 13:57
@XabierDomínguez, I can't even understand what you said. My qustion was more in the sense of "The exponential function is an isomorphism between $(\mathbb{R},+)$ and $(\mathbb{R}^+,\times )$, but if I add the hypotheses that this isomorphism is an automorphism and the spaces are a vector space with usual sum, will those automorphisms be necessarily linear?". – Marra Apr 16 '12 at 14:07
The exponential example was to show that not every isomorphism is linear. – Marra Apr 16 '12 at 14:10
up vote 10 down vote accepted

An automorphism of a vector space is, by definition, an invertible linear mapping. So no.

share|cite|improve this answer
"...invertible linear mapping from the space to itself" perhaps? – Ragib Zaman Apr 16 '12 at 13:57
It's in the definition of automorphism that it is from the space into itself. – Marra Apr 16 '12 at 14:08

This is an answer to the OP's question interpreted as "Is there an automorphism of the group $(\mathbb{R}^n,+)$ that is not $\mathbb{R}$-linear?"

The answer is yes. Consider $\mathbb{R}$ as a vector space over $\mathbb{Q}.$ Note that as a $\mathbb{R}$-vector space, $\mathbb{R}$ has dimension 1, but as a $\mathbb{Q}$-vector space its dimension is infinite (actually it is the continuum). Let $\{e_i\}$ and $\{v_i\}$ be two different basis of $\mathbb{R}$ as a $\mathbb{Q}$-vector space. Define $f:\mathbb{R}\to \mathbb{R}$ as follows: $f(\sum r_i e_i)=\sum r_i v_i.$ This is an isomorphism of the vector space $\mathbb{R}$ (over $\mathbb{Q}$) onto itself because it maps one basis onto another. In particular it is an automorphism of the additive group $\mathbb{R}.$ But most of these maps are not linear over $\mathbb{R}.$ For example, you can choose $e_1=1,$ $e_2=\pi,$ and complete the basis $\{e_i\}$ from there (note that these two are independent over $\mathbb{Q}$). Now choose $v_1=f(e_1)=1,$ and as $v_2=f(e_2)$ pick any irrational number different from $\pi,$ and complete the basis $\{v_i\}$ from there. This $f$ cannot be of the form $x\mapsto \lambda x,$ i. e. it cannot be $\mathbb{R}$-linear.

share|cite|improve this answer
Can you give me an example? – Marra Apr 16 '12 at 14:13
I can't because these maps are not constructible. They depend on the Axiom of Choice. – Xabier Domínguez Apr 16 '12 at 14:14
Are you talking about finite-dimension vector spaces? I think I should have said that. – Marra Apr 16 '12 at 14:15
I'll try to ellaborate my answer a bit further. – Xabier Domínguez Apr 16 '12 at 14:16
Ok, I'll make my question again: Consider $\mathbb{R}^n$ as the classical vector space over $\mathbb{R}$ (which was my obvious intent since the beggining). Then this must not hold. – Marra Apr 16 '12 at 16:04

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.