Give a signature to express an strict lineal order on a set and show that the minimal element is the minimum.

39 Views Asked by At

Give a signature to express an strict lineal order on a set and show that the minimal element is the minimum.

What I've done is set our signature as $\Sigma = \{ A, <\mid_2, a \}$, where $A$ is the set we'll be working with, $<$ is a $2$-arity predicate and $a$ is a constant representing the minimal element of $A$. Then the axioms of a strict order (antireflective and transitivity) would read as follows: $$\forall x \forall y \forall x ((x < y) \land (y < z) \rightarrow (x<z))$$ and $$\forall x \neg (x < x)$$ Now I want to formalize that each two distinct elements can be compared. I have set that property as $$\forall x \forall y ((x<y) \lor (y<x))$$ I had some doubts regarding this last one since I didn't not mention at all that $x$ and $y$ are not equal, but that case is excluded from the fact that $<$ is antireflective and hence, we cannot have $x < x$ on the first place. Finally, I want to express that $a$ is minimal, that is, there is no element smaller than $a$, which reads as $$\forall x \neg(x<a)$$ And finally, to prove that $a$ is the minimum of the set, I would write that as follows $$\forall x (a < x)$$ Now, in order to prove the assertion, negate the conclusion, that is, $\neg \forall x (a <x)$. From that I can say that $\neg (a<b)$ from some $b$. Now, since $\forall x \neg(x < a)$ I can say that $\neg (b <a)$ for some $b$. Now, I can compare $a$ and $b$ and we must have that $a <b$, which contradicts $\neg (a <b)$, or $b < a$, which contradicts $\neg (b < a)$. Then, the set of all premises along with the conclusion negated is unsatisfiable, an hence, the minimal element is the minimum, as desired.

Is this proof okay? Thank you in advance!

1

There are 1 best solutions below

4
On BEST ANSWER

You're right to doubt $\forall x\forall y(x<y\lor y<x)$. In first-order logic there's nothing that prevents two nested quantifiers from being instantiated to the same element of the domain, so that axiom would be demanding, among other things, that $c<c\lor c<c$. So your theory is inconsistent.

In order to complete your task here, you need a way to speak about equality of elements. In some developments of first-order logic, $=$ is actually considered a part of the logic, with automatic rules to go with it, and in that case you can just write $$\forall x\forall y(x<y\lor y<x \lor x=y)$$

Otherwise you need manufacture a $=$ yourself. In that case you cannot actually get all the way to having your theory speak about "a strict linear order on a set", but need to make do with having a theory that describes a strict linear order on equivalence classes of indistinguishable elements of the universe. But in practice that is often considered good enough -- you can derive all the same theorems under that interpretation, for example.

Your options are then basically either

  • Add a $=$ symbol to your signature, and provide axioms that force it to behaves like equality ought to behave -- in particular, enough to guarantee that $=$ is an equivalence relation, and let us derive $\phi(x)$ from $x=y$ and $\phi(y)$ for all formulas $\phi(\cdots)$.

  • Declare that $x=y$ is an abbreviation for $\neg(x<y \lor y<x)$ and again provide axioms that guarantee this will actually behave like equality.