theory Ind_Examples

imports Main

begin

section{* Examples of inductive definitions *}

text {*

\label{sec:ind-examples}

In this section, we will give three examples showing how to define inductive

predicates by hand and prove characteristic properties such as introduction

rules and an induction rule. From these examples, we will then figure out a

general method for defining inductive predicates, which will be described in

\S\ref{sec:ind-general-method}. This description will serve as a basis for the

actual implementation to be developed in \S\ref{sec:ind-interface} -- \S\ref{sec:ind-proofs}.

It should be noted that our aim in this section is not to write proofs that

are as beautiful as possible, but as close as possible to the ML code producing

the proofs that we will develop later.

As a first example, we consider the \emph{transitive closure} @{text "trcl R"}

of a relation @{text R}. It is characterized by the following

two introduction rules

\[

\begin{array}{l}

@{term "trcl R x x"} \\

@{term [mode=no_brackets] "R x y \<Longrightarrow> trcl R y z \<Longrightarrow> trcl R x z"}

\end{array}

\]

Note that the @{text trcl} predicate has two different kinds of parameters: the

first parameter @{text R} stays \emph{fixed} throughout the definition, whereas

the second and third parameter changes in the ``recursive call''.

Since an inductively defined predicate is the \emph{least} predicate closed under

a collection of introduction rules, we define the predicate @{text "trcl R x y"} in

such a way that it holds if and only if @{text "P x y"} holds for every predicate

@{text P} closed under the above rules. This gives rise to a definition containing

a universal quantifier over the predicate @{text P}:

*}

definition "trcl \<equiv> \<lambda>R x y.

  \<forall>P. (\<forall>x. P x x) \<longrightarrow> (\<forall>x y z. R x y \<longrightarrow> P y z \<longrightarrow> P x z) \<longrightarrow> P x y"

text {*

\noindent

Since the predicate @{term "trcl R x y"} yields an element of the type of object

level truth values @{text bool}, the meta-level implications @{text "\<Longrightarrow>"} in the

above introduction rules have to be converted to object-level implications

@{text "\<longrightarrow>"}. Moreover, we use object-level universal quantifiers @{text "\<forall>"}

rather than meta-level universal quantifiers @{text "\<And>"} for quantifying over

the variable parameters of the introduction rules. Isabelle already offers some

infrastructure for converting between meta-level and object-level connectives,

which we will use later on.

With this definition of the transitive closure, the proof of the (weak) induction

theorem is almost immediate. It suffices to convert all the meta-level connectives

in the induction rule to object-level connectives using the @{text atomize} proof

method, expand the definition of @{text trcl}, eliminate the universal quantifier

contained in it, and then solve the goal by assumption.

*}

lemma trcl_induct:

  assumes trcl: "trcl R x y"

  shows "(\<And>x. P x x) \<Longrightarrow> (\<And>x y z. R x y \<Longrightarrow> P y z \<Longrightarrow> P x z) \<Longrightarrow> P x y"

  apply (atomize (full))

  apply (cut_tac trcl)

  apply (unfold trcl_def)

  apply (drule spec [where x=P])

  apply assumption

  done

(*<*)

lemma "trcl R x x"

  apply (unfold trcl_def)

  apply (rule allI impI)+

(*>*)

txt {*

\noindent

The above induction rule is \emph{weak} in the sense that the induction step may

only be proved using the assumptions @{term "R x y"} and @{term "P y z"}, but not

using the additional assumption \mbox{@{term "trcl R y z"}}. A stronger induction rule

containing this additional assumption can be derived from the weaker one with the

help of the introduction rules for @{text trcl}.

We now turn to the proofs of the introduction rules, which are slightly more complicated.

In order to prove the first introduction rule, we again unfold the definition and

then apply the introdution rules for @{text "\<forall>"} and @{text "\<longrightarrow>"} as often as possible.

We then end up in a proof state of the following form:

@{subgoals [display]}

The two assumptions correspond to the introduction rules, where @{term "trcl R"} has been

replaced by @{term "P"}. Thus, all we have to do is to eliminate the universal quantifier

in front of the first assumption, and then solve the goal by assumption:

*}

(*<*)oops(*>*)

lemma trcl_base: "trcl R x x"

  apply (unfold trcl_def)

  apply (rule allI impI)+

  apply (drule spec)

  apply assumption

  done

(*<*)

lemma "R x y \<Longrightarrow> trcl R y z \<Longrightarrow> trcl R x z"

  apply (unfold trcl_def)

  apply (rule allI impI)+

(*>*)

txt {*

\noindent

Since the second introduction rule has premises, its proof is not as easy as the previous

one. After unfolding the definitions and applying the introduction rules for @{text "\<forall>"}

and @{text "\<longrightarrow>"}, we get the proof state

@{subgoals [display]}

The third and fourth assumption corresponds to the first and second introduction rule,

respectively, whereas the first and second assumption corresponds to the premises of

the introduction rule. Since we want to prove the second introduction rule, we apply

the fourth assumption to the goal @{term "P x z"}. In order for the assumption to

be applicable, we have to eliminate the universal quantifiers and turn the object-level

implications into meta-level ones. This can be accomplished using the @{text rule_format}

attribute. Applying the assumption produces two new subgoals, which can be solved using

the first and second assumption. The second assumption again involves a quantifier and

implications that have to be eliminated before it can be applied. To avoid problems

with higher order unification, it is advisable to provide an instantiation for the

universally quantified predicate variable in the assumption.

*}

(*<*)oops(*>*)

lemma trcl_step: "R x y \<Longrightarrow> trcl R y z \<Longrightarrow> trcl R x z"

  apply (unfold trcl_def)

  apply (rule allI impI)+

  proof -

    case goal1

    show ?case

      apply (rule goal1(4) [rule_format])

      apply (rule goal1(1))

      apply (rule goal1(2) [THEN spec [where x=P], THEN mp, THEN mp,

	OF goal1(3-4)])

      done

  qed

text {*

\noindent

This method of defining inductive predicates easily generalizes to mutually inductive

predicates, like the predicates @{text even} and @{text odd} characterized by the

following introduction rules:

\[

\begin{array}{l}

@{term "even (0::nat)"} \\

@{term "odd m \<Longrightarrow> even (Suc m)"} \\

@{term "even m \<Longrightarrow> odd (Suc m)"}

\end{array}

\]

Since the predicates are mutually inductive, each of the definitions contain two

quantifiers over the predicates @{text P} and @{text Q}.

*}

definition "even \<equiv> \<lambda>n.

  \<forall>P Q. P 0 \<longrightarrow> (\<forall>m. Q m \<longrightarrow> P (Suc m)) \<longrightarrow> (\<forall>m. P m \<longrightarrow> Q (Suc m)) \<longrightarrow> P n"

definition "odd \<equiv> \<lambda>n.

  \<forall>P Q. P 0 \<longrightarrow> (\<forall>m. Q m \<longrightarrow> P (Suc m)) \<longrightarrow> (\<forall>m. P m \<longrightarrow> Q (Suc m)) \<longrightarrow> Q n"

text {*

\noindent

For proving the induction rule, we use exactly the same technique as in the transitive

closure example:

*}

lemma even_induct:

  assumes even: "even n"

  shows "P 0 \<Longrightarrow> 

             (\<And>m. Q m \<Longrightarrow> P (Suc m)) \<Longrightarrow> (\<And>m. P m \<Longrightarrow> Q (Suc m)) \<Longrightarrow> P n"

  apply (atomize (full))

  apply (cut_tac even)

  apply (unfold even_def)

  apply (drule spec [where x=P])

  apply (drule spec [where x=Q])

  apply assumption

  done

text {*

\noindent

A similar induction rule having @{term "Q n"} as a conclusion can be proved for

the @{text odd} predicate.

The proofs of the introduction rules are also very similar to the ones in the

previous example. We only show the proof of the second introduction rule,

since it is almost the same as the one for the third introduction rule,

and the proof of the first rule is trivial.

*}

lemma evenS: "odd m \<Longrightarrow> even (Suc m)"

  apply (unfold odd_def even_def)

  apply (rule allI impI)+

  proof -

    case goal1

    show ?case

      apply (rule goal1(3) [rule_format])

      apply (rule goal1(1) [THEN spec [where x=P], THEN spec [where x=Q],

	THEN mp, THEN mp, THEN mp, OF goal1(2-4)])

      done

  qed

(*<*)

lemma even0: "even 0"

  apply (unfold even_def)

  apply (rule allI impI)+

  apply assumption

  done

lemma oddS: "even m \<Longrightarrow> odd (Suc m)"

  apply (unfold odd_def even_def)

  apply (rule allI impI)+

  proof -

    case goal1

    show ?case

      apply (rule goal1(4) [rule_format])

      apply (rule goal1(1) [THEN spec [where x=P], THEN spec [where x=Q],

	THEN mp, THEN mp, THEN mp, OF goal1(2-4)])

      done

  qed

(*>*)

text {*

\noindent

As a final example, we will consider the definition of the accessible

part of a relation @{text R} characterized by the introduction rule

\[

@{term "(\<And>y. R y x \<Longrightarrow> accpart R y) \<Longrightarrow> accpart R x"}

\]

whose premise involves a universal quantifier and an implication. The

definition of @{text accpart} is as follows:

*}

definition "accpart \<equiv> \<lambda>R x. \<forall>P. (\<forall>x. (\<forall>y. R y x \<longrightarrow> P y) \<longrightarrow> P x) \<longrightarrow> P x"

text {*

\noindent

The proof of the induction theorem is again straightforward:

*}

lemma accpart_induct:

  assumes acc: "accpart R x"

  shows "(\<And>x. (\<And>y. R y x \<Longrightarrow> P y) \<Longrightarrow> P x) \<Longrightarrow> P x"

  apply (atomize (full))

  apply (cut_tac acc)

  apply (unfold accpart_def)

  apply (drule spec [where x=P])

  apply assumption

  done

(*<*)

lemma accpartI: "(\<And>y. R y x \<Longrightarrow> accpart R y) \<Longrightarrow> accpart R x"

  apply (unfold accpart_def)

  apply (rule allI impI)+

(*>*)

txt {*

\noindent

Proving the introduction rule is a little more complicated, due to the quantifier

and the implication in the premise. We first convert the meta-level universal quantifier

and implication to their object-level counterparts. Unfolding the definition of

@{text accpart} and applying the introduction rules for @{text "\<forall>"} and @{text "\<longrightarrow>"}

yields the following proof state:

@{subgoals [display]}

Applying the second assumption produces a proof state with the new local assumption

@{term "R y x"}, which will then be used to solve the goal @{term "P y"} using the

first assumption.

*}

(*<*)oops(*>*)

lemma accpartI: "(\<And>y. R y x \<Longrightarrow> accpart R y) \<Longrightarrow> accpart R x"

  apply (unfold accpart_def)

  apply (rule allI impI)+

  proof -

    case goal1

    note goal1' = this

    show ?case

      apply (rule goal1'(2) [rule_format])

      proof -

        case goal1

        show ?case

	  apply (rule goal1'(1) [OF goal1, THEN spec [where x=P],

            THEN mp, OF goal1'(2)])

	  done

      qed

    qed

end