# can GCD(0,8)≠1 be proven purely by lattice laws?

Triggered by previous question, can one prove GCD(0,8)≠1 purely by lattice laws?

Brute force Prover9/Mace4 assertions

x ^ y = y ^ x.
(x ^ y) ^ z = x ^ (y ^ z).
x ^ (x v y) = x.
x v y = y v x.
(x v y) v z = x v (y v z).
x v (x ^ y) = x.

1 v x = x.
1 ^ x = 1.
0 ^ x = 1.


exhibit no [finite] model, which is indication that the system is inconsistent. I have trouble, however, understanding how to elevate this intuition into a formal proof (there is no goal).

-
Do you have x ^ x = x from your laws ? –  mercio Mar 18 '11 at 21:20
@chandok: Yes, it is redundant. –  Tegiri Nenashi Mar 18 '11 at 21:28
Is 0 v x = 1 correct? –  Myself Mar 19 '11 at 1:24
That is what the original thread suggested. It has been refuted together with another option: 0 v x = 1. The correct version is 0 v x = x and, predictably, Mace4 generates 2 element model. I'm interested, however disproving the wrong assertion 0 v x = 1. I tried putting its negation into Prover9 goal, still Prover9 doesn't seem to be able to derive it. –  Tegiri Nenashi Mar 19 '11 at 1:36
@Tegiri: You either have things backward, or you are viewing the divisibility order backwards, with $a|b$ meaning $b\leq a$... –  Arturo Magidin Mar 19 '11 at 2:00

Note $\rm\ x = 0\$ in $\rm\ x \wedge (x \vee y)\ =\ x\ \ \Rightarrow\ \ 0\wedge (0\vee y)\ =\ 0\ \$ contra $\rm\ \ 0\wedge x\ =\ 1\ \$ (presuming $\rm\ 0 \ne 1\:$).

Alternatively, recall that the idempotent laws follows from the absorption laws, viz.

$$\rm x\wedge x\ =\ x\wedge (x\vee (x\wedge x))\ =\ x$$

Hence $\rm\ \ 0\wedge 0\ =\ 0\ \$ contra $\rm\ \ 0\wedge x\ =\ 1\:.$

-
Just to clarify: you are using $0$ and $1$ to denote the corresponding integers, and not to denote the infimum and supremum of the lattice, correct? –  Arturo Magidin Mar 19 '11 at 21:21
No, the above is in the equational theory that the OP gave. –  Bill Dubuque Mar 19 '11 at 21:37
@Bill: Right: so that $0$ denotes the usual $0$, not the least element of the lattice. (His theory includes, e.g., $1\lor x = x$, which of course means that "1" does not represent the maximum, as is standard in 01-lattices). –  Arturo Magidin Mar 19 '11 at 21:39
@Arturo: Possible models are not relevant to the above equational proof (other than the assumption that $\rm\ 0 \ne 1\:$). –  Bill Dubuque Mar 19 '11 at 21:57
I tried absorption and even distributivity, but this rather simple proof escaped me. Thank you. –  Tegiri Nenashi Mar 21 '11 at 19:21

Solved the remaining bit of the puzzle: GCD(0,8)≠0.

1 v x = x ⇒  1 v 0 = 0
0 ^ x = 0 ⇒  1 ^ 0 = 0
1 v 0 = 1 ^ 0 ⇒  1 = 0

-