If I have a set endowed with an addition operation (say a general group) and I know it contains 1, can I always define a multiplication operation so that it distributes over the addition? By this I don't require that the multiplication to even be associative but I wish for it to satisfy the condition that 1*a=a*1=a. Does it matter if addition was defined to be commutative?

  • $\begingroup$ If you have addition, then there is $0$. I guess that then $a\cdot b=0$ (trivial multiplication) is distributive. I know, zou are asking for non-trivila multiplication. $\endgroup$ – Janko Bracic Jan 23 '15 at 6:48
  • $\begingroup$ well I didn't think of that, but I am still interested in nontrivial ones, maybe a proof that nontrivial ones exist? $\endgroup$ – davik Jan 23 '15 at 6:49
  • $\begingroup$ wait what if I required the identity property to hold $\endgroup$ – davik Jan 23 '15 at 6:51
  • $\begingroup$ You mean, there is $1\ne 0$ in the additive group, such that $1\cdot a=a$ for all $a$? Then it will be even harder to find non-trivial multiplication. $\endgroup$ – Janko Bracic Jan 23 '15 at 6:53
  • $\begingroup$ I have restated the question this way, but I wouldn't mind if you gave a proof of a nontrivial multiplication that didn't satisfy this, I just wanted a way to get rid of the trivial case :) $\endgroup$ – davik Jan 23 '15 at 6:58

My idea is to steal a multiplication from an algebraic structure which already has distributive multiplication. More precisely, let our additive group be $(G,+)$. Assume that there exist a ring $(R,+,\cdot)$ with a nontrivial multiplication and an injective additive map $\phi: G \to R$. Then one can define $\phi^{-1}:\phi(G) \to G$ and this is again an additive map. If $\phi(G)$ is closed for the multiplication in $R$, then one can define a multiplication on $G$ by $$ a\cdot b:=\phi^{-1}(\phi(a)\phi(b))\quad (a, b\in G).$$ This will be a distributive multiplication. If $R$ has an identity $1$ and $1\in \phi(G)$, then $\phi^{-1}(1)$ is the identity for the multiplication in $G$.

Of course, in the above reasoning, it is not necessary that we have a ring, it is enough that $(R,+)$ is an additive group with additional non-trivial binary operation which is distributive.

Of course, the above idea is not a solution of the posed problem. It is just a reformulation of the question.

One more comment, not related to the above thinking. If the multiplication in $(G,+)$ is defined by $a\cdot b=a$, then we have distributivity from one side: $$ (x+y)\cdot z=x+y=x\cdot z+y\cdot z\quad (x,y,z\in G). $$


For non-commutative multiplication, you cannot. Every element's order must divide the order of $1$: if the order of $1$ is $n$, then $$ \underbrace{a+\dots+a}_{n\text{ times}} = a\cdot (\underbrace{1+\dots+1}_{n\text{ times}}) = a\cdot 0 = 0. $$ (And $a\cdot 0$ has to be $0$ since $a\cdot 0 = a\cdot (0+0) = a\cdot 0 + a\cdot 0$.)

Now consider $S_3$: its elements have orders $1,2$ and $3$, so whatever element is chosen to be $1$, there would be elements such that their order does not divide its order.


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.