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

Do there exist binary operators *, **, and *** on the real numbers, such that * distributes over **, ** distributes over ***, *** distributes over *, but not vice versa?

share|cite|improve this question
Likely there are such over a countable, perhaps even small finite set. Do you require anything else of the operators? Gerhard "Ask Me About System Design" Paseman, 2012.09.08 – Gerhard Paseman Sep 9 '12 at 4:11
These conditions can be expressed by first-order formulas, hence if such operators exist on some infinite set (or on arbitrarily large finite sets), then they exist on every infinite set. – Emil Jeřábek Sep 9 '12 at 9:27

I shall show that we can have commutative associative binary operations that satisfy the required properties.

Without loss of generality, we can replace $\mathbb{R}$ by any other infinite set. Consider the binary operations $+,\cdot,\uparrow$ on the interval [1,\infty] where $+,\cdot$ are just normal addition and multiplication and $x\uparrow y=\infty$ for all $x,y$. Every elementary school student should know the distributive property for addition and multiplication $(a+b)\cdot c=a\cdot c+b\cdot c$. Furthermore, we have

$$ (a\cdot b)\uparrow c=\infty=(a\uparrow c)\cdot(b\uparrow c)=\infty $$ and $$ (a\uparrow b)+c=\infty+c=\infty=(a+c)\uparrow(b+c). $$

On the other hand, the average kindergartner can verify that $(1\cdot 2)+3=5\neq 20=(1+3)\cdot(2+3)$. Therefore $+,\cdot,\uparrow$ are the required distributive operations.

One can easily modify the above example to get an algebra where all three reverse distributive properties fail. For instance take binary operations $*,**,***$ on $[1,\infty]^3$ where $$(a,b,c)*(x,y,z)=(a+x,b\cdot y,c\uparrow z)$$ $$(a,b,c)**(x,y,z)=(a\cdot x,b\uparrow y,c+z)$$ $$(a,b,c)***(x,y,z)=(a\uparrow x,b+y,c\cdot z).$$

In fact, it appears that you can get $n$ cyclic binary operations that only distribute on one side simply by letting $(+,\cdot,\uparrow,\uparrow,...,\uparrow).$

share|cite|improve this answer
I say instead "some desired" for "the required". I definitely like using the cube to propogate the failure of distributivity in the other direction. I suspect one can get a similar example on a three or four element algebra (perhaps even two). Gerhard "Wish I Thought Of It" Paseman, 2012.09.11 – Gerhard Paseman Sep 11 '12 at 16:17

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.