A formal proof of the monotoncity property of Natural Deduction

187 Views Asked by At

I have formalized meta-theories for first-order logic with the Coq Proof Assistant. I defined rule (${\forall}$I) by: $$ \frac {\Gamma \vdash \left[ y / x \right] A} {\Gamma \vdash \left( \forall x \right) A} $$ provided by $y \notin \mathrm{FV} \left( \Gamma \right)$ and $y \notin \mathrm{FV} \left( \left( \forall x \right) A \right)$; where $x$ is an individual variables and so is $y$, $A$ is a well-formed formula, $\Gamma$ is a set of well-formed formulae, $\mathrm{FV} \left( X \right)$ denotes the set of all free variables occur in object $X$, and $[t/x]\varphi$ is substitution of $t$ for $x$ in $\varphi$.

And then, I have tried to find a formal proof of the monotonicity property, i.e., $$ \Gamma \vdash C \, \mathrm{and} \, \Gamma \subseteq \Gamma^{\prime} \implies \Gamma^{\prime} \vdash C . $$ However, if $\Gamma^{\prime}$ is of a set but not a list, it is failed for me to find an individual variable $y^{\prime}$ such that $y^{\prime} \notin \mathrm{FV} \left( \Gamma^{\prime} \right)$ and $y^{\prime} \notin \mathrm{FV} \left( \left( \forall x \right) A \right)$, because $\mathrm{FV} \left( \Gamma \right) \subseteq \mathrm{FV} \left( \Gamma^{\prime} \right)$ .

Actually, letting $v_i$ denote the $i$-th individual variable, there is a counter example: $$ \emptyset \vdash \left( \forall v_1 \right) \left( v_1 = v_1 \right) \;\not\!\!\!\implies \Gamma \vdash \left( \forall v_1 \right) \left( v_1 = v_1 \right) $$ where $\Gamma := \left\{ v_i = v_i \, | \, i \in \mathbb{N} \right\}$.

So, I am considering a plan: when proving a meta-theory of first-order logic, I add the constraint on $\Gamma$ in the statement of the meta-theory, for which has countably infinite free-variables and is denoted by $\mathbf{P} \left( \cdot \right)$. If it is accetable, for an example, the statement of the monotonicity property will be $$ \Gamma \vdash C \, \mathrm{and} \, \Gamma \subseteq \Gamma^{\prime} \, \mathrm{and} \, \mathbf{P} \left( \Gamma^{\prime} \right) \implies \Gamma^{\prime} \vdash C . $$

However, I have not found any textbook in which such a constraint is defined or used. I have two questions:

  1. I wonder if the above statement is considered okay in terms of many mathematicians.
  2. Furthermore, my question is whether the plan is acceptable or not.

On Oct 9, 2023, the original question was:

Title: How can I prove the Monotonicity property of Natural Deduction?

I'm working on making another framework for 1st-order Logic in The Coq Proof Assistant.

I stated $\dot{\forall}$-intro rule: for any set of formula $\Gamma$ and any formula $C$ and any individual variable $x$ and $y$, if $\Gamma \vdash C \left[ y / x \right]$ then $\Gamma \vdash \dot{\forall} x , C$ provided by $y \notin \mathrm{FV}\left(\Gamma\right)$ and $y \notin \mathrm{FV} \left(\dot{\forall} x , C \right)$, where $\varphi \left[ t / x \right]$ denotes the result substituting $x$ for $t$ in $\varphi$.

To prove the Monotonicity property, I struggled with the condition $y \notin \mathrm{FV} \left( \Gamma \right)$. Assume $\Gamma \subseteq \Delta$ and $\Gamma \vdash C$. Then, to prove $\Delta \vdash C$, it is required to assume $\mathrm{FV} \left( \Gamma \right) = \mathrm{FV} \left( \Delta \right)$.

So I wonder that the $\dot{\forall}$-intro rule is correct. If so, how can I prove the property?

Or, how can I state again correctly if it is wrong? Since I've already proved the soundness theorem of Natural Deduction for the Tarski's semantics, I hope it correct.


Proving the following Extraction Lemma requires the Monotonicity property: Extraction Lemma where p⟦ x ↦ t ⟧ denotes $p \left[ t / x \right]$.