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

Let $M$ denote the additive group of $ \mathbb{C} $. Why is $ \mathbb{C}$, as a ring, not isomorphic to $\mathrm{End}(M)$, where addition is defined pointwise, and multiplication as endomorphisms composition?


share|cite|improve this question
End(M) is huge. Really huge. Try to find a large class of elements it contains. – Qiaochu Yuan Feb 6 '11 at 21:07
Why would you expect it to be isomorphic? – Mariano Suárez-Alvarez Feb 6 '11 at 21:38
@Mariano - Well, $\mathbb{Z}$ is isomorphic to $End(\mathbb{Z},+)$, and a similar result holds for $\mathbb{Q}$. – Pandora Feb 6 '11 at 22:11
@Mariano: I see good sense in Pandora's expectation. Namely, for any commutative ring $R$, $R$ is isomorphic to the ring of $R$-module endomorphisms of $R$. When $R = \mathbb{Z}$, an $R$-module endomorphism is the same as an abelian group endomorphism, by definition. The same happens to hold for $\mathbb{Q}$. For most other rings it doesn't hold, but it's a good question... – Pete L. Clark Feb 6 '11 at 23:42
@Pete @Pandora: I did not say that it is a bad question, nor do I think so! It is a great working hypothesis to try on examples---and, as Pete says, for most rings it won't work. – Mariano Suárez-Alvarez Feb 7 '11 at 2:40
up vote 17 down vote accepted

Method 1: $\operatorname{End}(M)$ is not an integral domain.

Method 2: Count idempotents. There are only 2 idempotent elements of $\mathbb{C}$, but lots in $\operatorname{End}(M)$, including for example real and imaginary projections along with the identity and zero maps.

Method 3: Cardinality. Assuming a Hamel basis for $\mathbb{C}$ as a $\mathbb{Q}$ vector space, there are $2^\mathfrak{c}$ $\mathbb{Q}$ vector vector space endomorphisms of $\mathbb{C}$, but only $\mathfrak{c}$ elements of $\mathbb{C}$.

share|cite|improve this answer
Method 0: $\mathrm{End}(M)$ is not commutative. – user26857 Apr 19 at 20:41

The endomorphism ring of $M$ is its endomorphism group as a vector space over $\mathbb{Q}$ (since any additive function on $\mathbb{C}$ is in fact $\mathbb{Q}$-linear, and any $\mathbb{Q}$-linear map is necessarily additive).

How many endomorphisms does a vector space over $\mathbb{Q}$ of dimension $\kappa\geq\aleph_0$ have? Turns out it has $2^{\kappa}$ endomorphisms. Any endomorphism is completely determined by the image of a basis. Fix a basis $\beta$ for $\mathbf{V}$ over $\mathbb{Q}$, which is of cardinality $\kappa$.

Each endomorphism corresponds to a function $\beta\to\mathop{\oplus}\limits_{b\in\beta}\mathbb{Q} = \mathbb{Q}^{(\beta)}$. So the cardinality of the endomorphism ring is the cardinality of $(\mathbb{Q}^{(\beta)})^{\beta}$.

The cardinality of $\mathbb{Q}^{(\beta)}$ is $|\beta|=\kappa$, so the entire thing has cardinality $$|\mathrm{end}(M)| = \left|\mathbb{Q}^{(\beta)}\right|^{|\beta|} = \kappa^{\kappa} = 2^{\kappa}.$$ (For a proof that if $\kappa$ is infinite and $2\leq\lambda\leq 2^{\kappa}$, then $\lambda^{\kappa} = 2^{\kappa}{}{}$, see this previous answer).

In the case of $M$ and $\mathbb{C}$, since $M$ is of dimension $\mathfrak{c}=2^{\aleph_0}$ as a $\mathbb{Q}$-vector space, you have that $|\mathrm{End}(M)| = 2^{\mathfrak{c}}\gt \mathfrak{c}=|\mathbb{C}|$. So there aren't enough elements in $\mathbb{C}$ to give you all the endomorphisms.

(The elements of $\mathbb{C}$ correspond to the $\mathbb{C}$-linear maps, of course, by simple multiplication; but here you want a lot more maps).

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.