Why we cannot construct a tree to test logical equivalence for 2 formulas the same way we do with validity? Because if both formulas have the same logical truth negating one of them will close the tree right?
Asked
Active
Viewed 231 times
1 Answers
1
We can use the Truth Tree Method to test for logical equivalence.
2 wffs $\phi$ and $\psi$ are logically equivalent if, and only if, their valuations are equal, i.e., $v(\phi)=v(\psi)$. This means we can test $\vDash\phi\leftrightarrow\phi.$ in the usual way, i.e., see if $\lnot(\phi\leftrightarrow\psi)$ can be satisfied.
Proof
If $\phi$ is logically equivalent to $\psi$ then $\phi\vDash\psi$ and $\psi\vDash\phi$. We can apply the Semantic Deduction Theorem to both conjuncts, which gives us $\vDash\phi\to\psi$ and $\vDash\psi\to\phi$.
For reductio, assume $\vDash\phi\to\psi$ and $\vDash\psi\to\phi$, while $\nvDash\phi\leftrightarrow\psi$. If $\nvDash\phi\leftrightarrow\psi$ then $v(\phi)=1$ and $v(\psi)=0$, thus $v(\phi\to\psi)=0$, which contradicts $\vDash\phi\to\psi$, or $v(\phi)=0$ and $v(\psi)=1$, thus $v(\psi\to\phi)=0, which contradicts $\vDash\psi\to\phi$. Thus, $\vDash\phi\leftrightarrow\psi$.
That this is the case shouldn't be a surprise given that PL is sound and complete and a syntactic proof of logical equivalence demonstrates that $\phi\vdash\psi$ and $\psi\vdash\phi$.
As an aside, it can still be helpful to do 2 trees for particularly complex wffs that have a lot of branching, i.e., test $\phi\vDash\psi$ and $\psi\vDash\phi$. The reason for this is avoiding mistakes when reading off counterexamples/IPLIs.
Example$^1$:
Ten O'Four
- 159
- 6