It will depend on virtually every aspect of the logical system you have developed. Both the method of proving things (as different systems will have different notions of proof) and the semantics of your system (how you have defined what it means for a sentence in your system to be "true")
In other words, you prove it for any set of assumptions, including the empty set (tautologies/valid statements)
@Michelle The best person to ask these types of questions to is @user21820 :) As he literally holds logic classes in here and will lock you in here until he is convinced you understand it better than he does ;)
He's in here every day, should be logging in within the next 3 hrs or so
He is familiar with virtually every logical system there is and will likely be able to give you better feedback suited to the specific system you are using
@TungNguyen They are not equal. They just so happen to be equivalent. Let's put it another way. "1=2" is equivalent to "1=3", not because their logical forms are inherently equivalent, but simply because we can prove both false. Similarly, there are some sentences that are equivalent not because they are instances of some general equivalent pairs, but just because we can prove them equivalent.
@LeakyNun I was indeed here about an hour ago, but haven't come to this room yet.
@TungNguyen: Did you also read and understand my response starting here?
@Michelle In general, a system that involves rules typically means that each rule can be thought of as an operation on valid statements; when you apply the rule to existing statements, it gives more statements. The theorems of the system are then generated by applying the rules to all the axioms, and keep applying them to all newly generated statements, and so on until the 'end'. Mathematically we simply say that the theorems are the closure of the axioms under the rules.
In general soundness of a system means that every theorem it proves is true in every model of that system. For rule-based systems I just described, it suffices (but may not be necessary) to prove that each rule is sound, namely when you apply it to true statements it gives only true statements. We call such rules truth-preserving.
So you can summarize the proof of the soundness theorem by saying that you prove that every rule is truth-preserving, and hence the closure of the axioms under the rules only includes true statements, since every statement in the closure is generated by finitely many applications of truth-preserving rules.
@LeakyNun @DavidReed: Did I miss out any important point?
@DavidReed Oh okay I missed that. In many textbooks they would just use axiom schemas to capture "generalization", and so MP would be the only rule. Convenient for meta-theorems, but terrible for actual human use.
I'm not sure if we can get away with using the complex projective space C* = C ⋃ {∞}, where 1/0 = ∞ and 1/∞ = 0 and z+∞ = ∞ for any z in C* and c·∞ = ∞ for any nonzero c in C*, but 0·∞ is undefined.
@AlessandroCodenotti i thought so too. the definition is that if $A$ is a closed formula in the form $\exists x B$ then there is a constant $c$ s.t $T \vdash \exists x B \to B_x[c]$
Ok, so pick a sentence $A$, if $\text{Th}(M)\vdash \neg \exists x B$ the Henkin condition is satisfied, what if $\text{Th}(M)\vdash\exists x B$? (I'm using the fact that $\text{Th}(M)$ is complete here)
@AlessandroCodenotti sorry didnt saw you answered. if $T \vdash A$ then for all $\sigma :Vars \to \{ \Bbb T , \Bbb F\}$ ,$\sigma(A) = \Bbb T$ so there is $a\in |M|$ s.t $\sigma(B) = \Bbb T$
i want to say this implies that $T\vdash A \to B_x[c_a]$
(@AlessandroCodenotti you also need to assume $A$ is closed)
i needed to write "if $M \vDash A$ " not "$T \vdash A$"
That theory is complete so any two models are elementarily equivalent, if you have $A,B\vDash T$, then for all sentences $\varphi$, $A\vDash\varphi\iff B\vDash\varphi$
A lot of nonsensical comments on that Meta thread have been deleted by the moderators, but many of those from George were recorded in my post, and one more later one here.
I feel it is necessary to warn you about all this, since you may not be aware of what has happened. Moreover, note that the claim by George that Godel's sentence is self-referential is false and clearly reveals that he in fact does not understand the incompleteness theorem at all. Note the irony that he tells people to give Godel's work a finishing blow, despite not having understood it at all.
@user21820 I'm curious as to how you would solve $f'(x) = f(x)$ (rigorously)
if you want the formal statement: find all possible functions $f:\Bbb R \to \Bbb R$ that is differentiable everywhere such that $f'(x) = f(x)$ for all $x \in \Bbb R$