Library Types

Require Export Smallstep.

More Automation


Time to take the gloves off...

auto and eauto


The auto tactic solves goals that are solvable by any combination of
  • intros
  • apply (used on some local hypothesis)
  • reflexivity
The eauto tactic works just like auto, except that it uses eapply instead of apply.

Using auto is always "safe" in the sense that it will never fail and will never change the proof state: either it completely solves the current goal, or it does nothing.

Here is a contrived example:

Lemma auto_example_1 : forall P Q R S T U : Prop,
  (P -> Q) ->
  (P -> R) ->
  (T -> R) ->
  (S -> T -> U) ->
  ((P->Q) -> (P->S)) ->
  T ->
  P ->
  U.
Proof.
  auto. Qed.

By default, auto and eauto consider just the hypotheses in the current context. But we can also give it a list of lemmas and constructors to consider in addition to these, by writing auto using ....
Lemma auto_example_2 : forall P Q R : Prop,
  Q ->
  (Q -> R) ->
  P \/ (Q /\ R).
Proof.
  auto using conj, or_introl, or_intror. Qed.

There are some constructors and lemmas that are applied very often in proofs. We can tell auto to *always* consider these, instead of mentioning them explicitly each time. This is accomplished by writing

Hint Resolve l.

where l is a top-level lemma theorem, or a constructor of an inductively defined proposition (i.e., anything whose type is an implication). As a shorthand, we can write

Hint Constructors c.

to tell Coq to add *all* of the constructors from the inductive definition of c to the database used by auto.

It is also sometimes necessary to add

Hint Unfold d.

where d is a defined symbol, so that auto knows to expand uses of d and enable further possibilities for applying lemmas that it knows about.

Hint Constructors and.
Hint Constructors or.

Lemma auto_example_3 : forall P Q R : Prop,
  Q ->
  (Q -> R) ->
  P \/ (Q /\ R).
Proof.
  auto. Qed.

Warning: It is easy to overuse auto and eauto and wind up with proofs that are impossible to understand later!

Proof with (tactic)


If you start a proof with "Proof using (tactic)" then writing ... instead of . after a tactic in the body of the proof will try to solve all generated subgoals with tactic (and fail if this doesn't work).

One common use of this facility is "Proof with auto" (or eauto). We'll see many examples of this later in the file.

solve by inversion


Here's another nice automation feature: it often arises that the context contains a contradictory assumption and we want to use inversion on it to solve the goal. We'd like to be able to say to Coq, "find a contradictory assumption and invert it" without giving its name explicitly.

Unfortunately (and a bit surprisingly), Coq does not provide a built-in tactic that does this. However, it is not too hard to define one ourselves. (Thanks to Aaron Bohannon for this nice hack.)

Tactic Notation "solve_by_inversion_step" tactic(t) :=
  match goal with
  | H : _ |- _ => solve [ inversion H; subst; t ]
  end
  || fail "because the goal is not solvable by inversion.".

Tactic Notation "solve" "by" "inversion" "1" :=
  solve_by_inversion_step idtac.
Tactic Notation "solve" "by" "inversion" "2" :=
  solve_by_inversion_step (solve by inversion 1).
Tactic Notation "solve" "by" "inversion" "3" :=
  solve_by_inversion_step (solve by inversion 2).
Tactic Notation "solve" "by" "inversion" :=
  solve by inversion 1.

Again, the details of how the Tactic Notation definition works are not important. All we need to know is that doing solve by inversion will find a hypothesis that can be inverted to solve the goal, if there is one. The tactics solve by inversion 2 and solve by inversion 3 are slightly fancier versions which will perform two or three inversions in a row, if necessary, to solve the goal.

try solve


Doing the same thing to all subgoals is very useful. But notice, in the proof as it now stands, that many of the subcases are solved by either reflexivity or solve by inversion. It would be really nice to deal with all of these once and for all. I.e., what we want is a way to do the same thing to MOST of the subgoals! The try solve ... tactical gives us this ability.

If t is a tactic, then try solve [t] is a tactic that
  • if t solves the goal, behaves just like t, or
  • if t cannot completely solve the goal, does nothing.
More generally, try solve [t1 | t2 | ...] will try to solve the goal by using t1, t2, etc. If none of them succeeds in completely solving the goal, then try solve [t1 | t2 | ...] does nothing.

The fact that it does nothing when it doesn't completely succeed in solving the goal means that there is no harm in using try solve ... in situations where it makes no sense. In particular, if we put it after a ;, it will solve as many subgoals as it can and leave the rest for us to solve by other methods.

Doing try solve [reflexivity | solve by inversion] on each of the subgoals generated by the induction... inversion... at the top of the determinism proof removes a lot of silly subgoals and leaves us free to concentrate on the more interesting ones.

Typed arithmetic expressions


To motivate the discussion of types, let's begin with an extremely simple language of arithmetic and boolean expressions -- a language just barely complex enough to have the possibility of programs "going wrong" because of runtime type errors.

Syntax


Inductive tm : Set :=
  | tm_true : tm
  | tm_false : tm
  | tm_if : tm -> tm -> tm -> tm
  | tm_zero : tm
  | tm_succ : tm -> tm
  | tm_pred : tm -> tm
  | tm_iszero : tm -> tm.

Inductive bvalue : tm -> Prop :=
  | bv_true : bvalue tm_true
  | bv_false : bvalue tm_false.

Inductive nvalue : tm -> Prop :=
  | nv_zero : nvalue tm_zero
  | nv_succ : forall t, nvalue t -> nvalue (tm_succ t).

Definition value (t:tm) := bvalue t \/ nvalue t.

Hint Constructors bvalue nvalue.
Hint Unfold value.

Small-step reduction


(* Predeclare the notation t1 ~~> t2 so that we can use it in
   the definition of step.  (Note the where clause below.) *)

Reserved Notation "t1 '~~>' t2" (at level 40).

Inductive step : tm -> tm -> Prop :=
  | ES_IfTrue : forall t1 t2,
      (tm_if tm_true t1 t2) ~~> t1
  | ES_IfFalse : forall t1 t2,
      (tm_if tm_false t1 t2) ~~> t2
  | ES_If : forall t1 t1' t2 t3,
      t1 ~~> t1' ->
      (tm_if t1 t2 t3) ~~> (tm_if t1' t2 t3)
  | ES_Succ : forall t1 t1',
      t1 ~~> t1' ->
      (tm_succ t1) ~~> (tm_succ t1')
  | ES_PredZero :
      (tm_pred tm_zero) ~~> tm_zero
  | ES_PredSucc : forall t1,
      nvalue t1 ->
      (tm_pred (tm_succ t1)) ~~> t1
  | ES_Pred : forall t1 t1',
      t1 ~~> t1' ->
      (tm_pred t1) ~~> (tm_pred t1')
  | ES_IszeroZero :
      (tm_iszero tm_zero) ~~> tm_true
  | ES_IszeroSucc : forall t1,
       nvalue t1 ->
      (tm_iszero (tm_succ t1)) ~~> tm_false
  | ES_Iszero : forall t1 t1',
      t1 ~~> t1' ->
      (tm_iszero t1) ~~> (tm_iszero t1')

where "t1 '~~>' t2" := (step t1 t2).

Tactic Notation "step_cases" tactic(first) tactic(c) :=
  first;
  [ c "ES_IfTrue" | c "ES_IfFalse" | c "ES_If"
  | c "ES_Succ" | c "ES_PredZero" | c "ES_PredSucc" | c "ES_Pred"
  | c "ES_IszeroZero" | c "ES_IszeroSucc" | c "ES_Iszero" ].

Hint Constructors step.

Normal forms and values


The first interesting thing about the step relation is that the progress theorem as we stated it before fails! That is, there are terms that are normal forms (they can't take a step) but not values -- i.e., we have not included them in our definition of possible "results of evaluation." Such terms are said to be stuck.

Notation step_normal_form := (normal_form step).

Definition stuck (t:tm) : Prop :=
  step_normal_form t /\ ~ value t.

Hint Unfold stuck.

Exercise: 2 stars, optional (some_tm_is_stuck)

Example some_tm_is_stuck :
  exists t, stuck t.
Proof.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Fortunately, things are not *completely* messed up: values and normal forms are not the same in this language, but at least the former set is included in the latter (i.e., we did not accidentally define things so that some value could still take a step).

Exercise: 3 stars, optional (value_is_nf)

Lemma value_is_nf : forall t,
  value t -> step_normal_form t.
Proof.
  (* Hint: You will reach a point in this proof where you
     need to use an induction to reason about a term that is
     known to be a numeric value.  This induction can be
     performed either over the term itself or over the
     evidence that it is a numeric value.  The proof goes
     through in either case, but you will find that one way
     is quite a bit shorter than the other.  For the sake of
     the exercise, try to complete the proof both ways. *)

  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Exercise: 3 stars, optional

Using value_is_nf, we can show that the step relation is also deterministic...
Theorem step_deterministic:
  partial_function step.
Proof with eauto.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Typing


Critical observation: although there are stuck terms in this language, they are all "nonsensical", mixing booleans and numbers in a way that we don't even *want* to have a meaning. We can easily exclude such ill-typed terms by defining a "typing relation" that relates terms to the types (either numeric or boolean) of their final results.

Inductive ty : Set :=
  | ty_bool : ty
  | ty_nat : ty.

(* The typing relation *)
Inductive has_type : tm -> ty -> Prop :=
  | T_True :
       has_type tm_true ty_bool
  | T_False :
       has_type tm_false ty_bool
  | T_If : forall t1 t2 t3 T,
       has_type t1 ty_bool ->
       has_type t2 T ->
       has_type t3 T ->
       has_type (tm_if t1 t2 t3) T
  | T_Zero :
       has_type tm_zero ty_nat
  | T_Succ : forall t1,
       has_type t1 ty_nat ->
       has_type (tm_succ t1) ty_nat
  | T_Pred : forall t1,
       has_type t1 ty_nat ->
       has_type (tm_pred t1) ty_nat
  | T_Iszero : forall t1,
       has_type t1 ty_nat ->
       has_type (tm_iszero t1) ty_bool.

Tactic Notation "has_type_cases" tactic(first) tactic(c) :=
  first;
  [ c "T_True" | c "T_False" | c "T_If" | c "T_Zero" | c "T_Succ"
  | c "T_Pred" | c "T_Iszero" ].

Hint Constructors has_type.

Examples


It's important to realize that the typing relation is a _conservative_ (or _static_) approximation: it does not calculate the type of the normal form of a term.

Example has_type_1 :
  has_type (tm_if tm_false tm_zero (tm_succ tm_zero)) ty_nat.
Proof.
  apply T_If.
    apply T_False.
    apply T_Zero.
    apply T_Succ.
      apply T_Zero. Qed.

Example has_type_nonexample :
  ~ has_type (tm_if tm_false tm_zero tm_true) ty_bool.
Proof.
  intros Contra. inversion Contra. inversion H4. Qed.

Exercise: 1 star

Example succ_hastype_nat__hastype_nat : forall t,
  has_type (tm_succ t) ty_nat ->
  has_type t ty_nat.
Proof.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Progress


The typing relation enjoys two critical properties. The first is that well-typed normal forms are values...

Exercise: 2 stars (finish_progress)

Theorem progress : forall t T,
  has_type t T ->
  value t \/ exists t', t ~~> t'.
Proof with eauto.
  intros t T HT.
  has_type_cases (induction HT) (Case)...
  (* the cases that were obviously values, like T_True and
     T_False, were eliminated immediately by auto *)

  Case "T_If".
    right. destruct IHHT1.
    SCase "t1 is a value". destruct H.
      SSCase "t1 is a bvalue". destruct H.
        SSSCase "t1 is tm_true".
          exists t2...
        SSSCase "t1 is tm_false".
          exists t3...
      SSCase "t1 is an nvalue".
        solve by inversion 2. (* on H and HT1 *)
    SCase "t1 can take a step".
      destruct H as [t1' H1].
      exists (tm_if t1' t2 t3)...
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Exercise: 1 star

Quick review. Answer TRUE or FALSE...
  • In this language, every normal form is a value.
  • Every value is a normal form.
  • The single-step evaluation relation is a partial function.
  • The single-step evaluation relation is a partial function.
  • The single-step evaluation relation is a TOTAL function.

Type preservation


The second critical property of typing is that when well-typed terms take a step, the result is also a well-typed term.

This theorem is also sometimes called the "subject reduction" property, because it tells us what happens when the "subject" of the typing relation is reduced.

Exercise: 2 stars (finish_preservation)

Theorem preservation : forall t t' T,
  has_type t T ->
  t ~~> t' ->
  has_type t' T.
Proof with eauto.
  intros t t' T HT HE.
  generalize dependent t'.
  (has_type_cases (induction HT) (Case));
         (* every case needs to introduce a couple of things *)
         intros t' HE;
         (* and we can deal with several impossible
            cases all at once *)

         try (solve by inversion).
    Case "T_If". inversion HE; subst.
      SCase "ES_IFTrue". assumption.
      SCase "ES_IfFalse". assumption.
      SCase "ES_If". apply T_If; try assumption.
        apply IHHT1; assumption.
    (* FILL IN HERE (and delete "Admitted") *) Admitted.

Exercise: 3 stars, optional (preservation_alternate_proof)

Theorem preservation' : forall t t' T,
  has_type t T ->
  t ~~> t' ->
  has_type t' T.
Proof with eauto.
  (* Now prove the same property again by induction on the
     EVALUATION derivation instead of on the typing
     derivation.  Begin by carefully reading and thinking
     about the first few lines of the above proof to make
     sure you understand what each one is doing.  The set-up
     for this proof is similar, but not exactly the same. *)

  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Type soundness


Putting progress and preservation together, we can see that a well-typed term can _never_ reach a stuck state.

Notation stepmany := (refl_step_closure step).

Notation "t1 '~~>*' t2" := (stepmany t1 t2) (at level 40).

Corollary soundness : forall t t' T,
  has_type t T ->
  t ~~>* t' ->
  ~(stuck t').
Proof.
  intros t t' T HT P. induction P; intros [R S].
  destruct (progress x T HT); auto.
  apply IHP. apply (preservation x y T HT H).
  unfold stuck. split; auto. Qed.

Indeed, in the present -- extremely simple -- language, every well-typed term can be reduced to a value: this is the normalization property. In richer languages, this property often fails, though there are some interesting languages (such as Coq's Fixpoint language, and the simply typed lambda-calculus, which we'll be looking at next) where all _well-typed_ terms can be reduced to normal forms.

Additional exercises


Exercise: 2 stars (subject_expansion)

Having seen the subject reduction property, it is reasonable to wonder whether the opposity property -- subject EXPANSION -- also holds. That is, is it always the case that, if t ~~> t' and has_type t' T, then has_type t T? If so, prove it. If not, give a counter-example.

FILL IN HERE...

Exercise: 2 stars (variation1)

Suppose we add the following two new rules to the evaluation relation:
      | ES_PredTrue :
           (tm_pred tm_true) ~~> (tm_pred tm_false)
      | ES_PredFalse :
           (tm_pred tm_false) ~~> (tm_pred tm_true)


Which of the following properties remain true in the presence of these rules? For each one, write either "remains true" or else "becomes false." If a property becomes false, give a counterexample.
  • Determinacy of step
  • Normalization of step for well-typed terms
  • Progress
  • Preservation

Exercise: 2 stars (variation2)

Suppose, instead, that we add this new rule to the typing relation:
      | T_IfFunny : forall t2 t3,
           has_type t2 ty_nat ->
           has_type (tm_if tm_true t2 t3) ty_nat


Which of the following properties remain true in the presence of this rule? (Answer in the same style as above.)
  • Determinacy of step
  • Normalization of step for well-typed terms
  • Progress
  • Preservation

Exercise: 2 stars (variation3)

Suppose, instead, that we add this new rule to the typing relation:
      | T_SuccBool : forall t,
           has_type t ty_bool ->
           has_type (tm_succ t) ty_bool


Which of the following properties remain true in the presence of this rule? (Answer in the same style as above.)
  • Determinacy of step
  • Normalization of step for well-typed terms
  • Progress
  • Preservation

Exercise: 2 stars (variation4)

Suppose, instead, that we add this new rule to the step relation:
      | ES_Funny1 : forall t2 t3,
           (tm_if tm_true t2 t3) ~~> t3


Which of the above properties become false in the presence of this rule? For each one that does, give a counter-example.

Exercise: 2 stars (variation5)

Suppose instead that we add this rule:
      | ES_Funny2 : forall t1 t2 t2' t3,
           t2 ~~> t2' ->
           (tm_if t1 t2 t3) ~~> (tm_if t1 t2' t3)


Which of the above properties become false in the presence of this rule? For each one that does, give a counter-example.

Exercise: 2 stars (variation6)

Suppose instead that we add this rule:
      | ES_Funny3 :
          (tm_pred tm_false) ~~> (tm_pred (tm_pred tm_false))


Which of the above properties become false in the presence of this rule? For each one that does, give a counter-example.

Exercise: 2 stars (variation7)

Suppose instead that we add this rule:
      | T_Funny4 :
            has_type tm_zero ty_bool


Which of the above properties become false in the presence of this rule? For each one that does, give a counter-example.

Exercise: 2 stars (variation8)

Suppose instead that we add this rule:
      | T_Funny5 :
            has_type (tm_pred tm_zero) ty_bool


Which of the above properties become false in the presence of this rule? For each one that does, give a counter-example.

Exercise: 3 stars, optional (more_variations)

Make up some exercises of your own along the same lines as the ones above. Try to find ways of selectively breaking properties -- i.e., ways of changing the definitions that break just one of the properties and leave the others alone.

Exercise: 1 star (remove_predzero)

The evaluation rule E_PredZero is a bit counter-intuitive: we might feel that it makes more sense for the predecessor of zero to be undefined, rather than being defined to be zero. Can we achieve this simply by removing the rule from the definition of step?

FILL IN HERE...

Exercise: 3 stars, optional (prog_pres_bigstep)

Suppose our evaluation relation is defined in the big-step style. What are the appropriate analogs of the progress and preservation properties?

FILL IN HERE...

Digression: Some useful facts about identifiers


Here we pause to record a few useful lemmas about ids. They are all similar to facts that we proved about numbers -- re-stating them for id avoids some unfolding in proofs.

Theorem beq_id_eq : forall n m,
  true = beq_id n m -> n = m.
Proof.
  intros n m H. destruct n as [n]. destruct m as [m].
  unfold beq_id in H. apply beq_nat_eq in H.
  auto. Qed.

Theorem beqid_false_not_eq : forall n m,
  false = beq_id n m -> n <> m.
Proof.
  intros n m H. destruct n as [n]. destruct m as [m].
  unfold beq_id in H. apply beq_false_not_eq in H.
  unfold not in *. intros Heq. inversion Heq.
  auto. Qed.

Hint Resolve beq_id_eq beqid_false_not_eq.

Theorem not_eq_false_beqid : forall n n' : id,
     n <> n'
  -> false = beq_id n n'.
Proof.
  intros n n' Hneq.
  remember (beq_id n n') as beq. destruct beq; auto.
  destruct Hneq. auto. Qed.

Theorem beq_id_refl : forall n : id,
  true = beq_id n n.
Proof.
  intros n. destruct n.
  unfold beq_id. apply beq_nat_refl. Qed.

The Simply Typed Lambda-Calculus


The simply typed lambda-calculus (often written STLC for short) can be described as the simplest possible functional programming language -- it contains anonymous functions, application, variables, and nothing else.

Since it is essentially a small fragment of Coq's built-in functional language, we could informally use the same notation for programs. For example,

fun x:A => x

is the identity function at the type A, and

fun x:A => fun y:B => x

is a function that takes an argument of type A, throws it away, and returns the constant-x function (of type B->A).

However, this "fun ... => ..." notation is a little verbose and heavy, especially for writing on the blackboard. So when we discuss terms informally, we'll use another common way of writing function abstractions in this language: \x:A.x, where \ stands for the greek letter "lambda". (This explains the name of the calculus.)

To keep things simple, we'll start with the "pure" simply typed lambda-calculus, where base types like A, B, etc. are completely uninterpreted -- i.e., there are no constants belonging to base types and no primitive operations over base types. At the end of this file, we'll see how to extend the pure system to something closer to a real programming language by adding base types like numbers.

The main new technical issues that we'll have to face in this section are dealing with _variable binding_ and _substitution_.
  • We've already seen how to formalize a language with variables (IMP). There, however, the variables were all global. In STLC, we need to handle bound variables -- function parameters.
  • Moreover, instead of just looking up global variables in the store, we'll need to reduce function applications by substituting arguments for parameters in function bodies.

Acknowledgement: This treatment of the simply typed lambda-calculus (using named variables and "capture-incurring substitution") is based on a development by Andrew Tolmach.

Syntax and operational semantics


Module STLC.

Types


The function types of the STLC are just like the function types of Coq. We begin with a collection of base types, which we write A, B, etc., in informal examples (formally, they are designated by the constructor ty_base together with a number). Then, for every pair of types T1 and T2, we construct the arrow type T1->T2 (formally, ty_arrow T1 T2).

Inductive ty : Set :=
  | ty_base : id -> ty
  | ty_arrow : ty -> ty -> ty.

(* Some base types for use in examples... *)
Notation A := (ty_base (Loc 0)).
Notation B := (ty_base (Loc 1)).
Notation C := (ty_base (Loc 2)).

Terms


The terms of the STLC are just variables, function application, and function abstraction. We use identifiers as variable names.

Inductive tm : Set :=
  | tm_var : id -> tm
  | tm_app : tm -> tm -> tm
  | tm_abs : id -> ty -> tm -> tm.

One important thing to note here is that an abstraction \x:T.t (formally, tm_abs x T t) is always annotated with the type (T) of its parameter. This is in contrast to Coq (and other full-blown functional languages like ML, Haskell, etc.), where we can write either
        fun x : A => t

or
        fun x => t

and trust Coq to fill in an appropriate annotation in the latter case.

Tactic Notation "tm_cases" tactic(first) tactic(c) :=
  first;
  [ c "tm_var" | c "tm_app" | c "tm_abs" ].

(* Some variables for examples... *)
Notation x := (Loc 0).
Notation y := (Loc 1).
Notation z := (Loc 2).

(* Some identity functions on various types... *)
(*   id_A = \x:A. x *)
Definition id_A :=
  tm_abs x A (tm_var x).
(*   id_AarrowA = \x:A->A. x *)
Definition id_AarrowA :=
  tm_abs x (ty_arrow A A) (tm_var x).
(*   id_AarrowA_arrow_AarrowA = \x:(A->A)->(A->A). x *)
Definition id_AarrowA_arrow_AarrowA :=
  tm_abs x (ty_arrow (ty_arrow A A) (ty_arrow A A)) (tm_var x).

(* A function that takes an A and returns a constant function
   of type B->A: *)

(*   k = \x:A. \y:B. x *)
Definition k := tm_abs x A (tm_abs y B (tm_var x)).

Values


When we come to defining the values of the STLC, we have to think a little.

First, an application is clearly NOT a value.

For abstractions, on the other hand, we have a choice:
  • We can say that \x:A.t is a value only of t is a value -- i.e., only if the function's body has been reduced (as much as it can be without knowing what argument it is going to be applied to).
  • Or we can say that \x:A.t is always a value, no matter whether t is one or not -- in other words, we can say that reduction stops at abstractions.
Coq makes the first choice -- for example,
         Eval simpl in (fun x:bool => plus 3 4)
yields fun x:bool => 7. But most real functional programming languages make the second choice -- reduction of a function's body only begins when the function is actually applied to an argument. We also make the latter choice here.

Having decided this, we don't need to worry about whether variables are values, since we'll always be reducing programs "from the outside in," and that means the step relation will always be working with closed terms (ones with no free variables).

Inductive value : tm -> Prop :=
  | v_abs : forall x T t,
      value (tm_abs x T t).

Hint Constructors value.

Substitution


Now we come to the most interesting part of the definition: the operation of substituting one term for the free occurrences of some variable in another term. This operation will be used in a moment to define the operational semantics of function application, where we will need to substitute the argument term for the function parameter in the function's body.

Informally, substitution is often written [s/x]t, pronounced "substitute s for x in t".

For example:
  • [(\y:A.y)/x]x = \y:A.y
  • [(\y:A.y)/x]z = z
  • [(\y:A.y)/x](\z:B. x) = \z:B. \y:A.y
  • [(\y:A.y)/x](\z:(A->A)->(A->A). z x) = \z:(A->A)->(A->A). z ((\y:A.y))
  • [(\y:A.y)/x](\x:B.x) = \x:B.x
  • [(\y:A.y)/x](x (\x:B.x)) = (\y:A.y) (\x:B.x)
Note that, in the last two examples, we do NOT substitute for the x in the body of an abstraction whose bound variable is named x.

Here is the formal definition:

Fixpoint subst (x:id) (s:tm) (t:tm) {struct t} : tm :=
  match t with
  | tm_var x' => if beq_id x x' then s else t
  | tm_abs x' T t1 => tm_abs x' T (if beq_id x x' then t1 else (subst x s t1))
  | tm_app t1 t2 => tm_app (subst x s t1) (subst x s t2)
  end.

Technical note: Substitution is actually a little tricky to define if we consider the case where s, the term being substituted in, may itself contain free variables. Since we are only interested in defining the step relation on closed terms here, we can avoid this extra complexity.

Reduction


The small-step reduction relation for STLC follows the same pattern as the ones we have seen before. Intuitively, to reduce a function application, we first reduce its left-hand side until it becomes a literal function; then we reduce its right-hand side (the argument) until it is also a value; and finally we substitute the argument for the bound variable in the body of the function. This last rule, written informally as
        (\x:T.t12) v2 ~~> [v2/x]t12

is traditionally called "beta-reduction".

Reserved Notation "t1 '~~>' t2" (at level 40).

Inductive step : tm -> tm -> Prop :=
  | ST_AppAbs : forall x T t12 v2,
         value v2
      -> (tm_app (tm_abs x T t12) v2) ~~> (subst x v2 t12)
  | ST_App1 : forall t1 t1' t2,
         t1 ~~> t1'
      -> tm_app t1 t2 ~~> tm_app t1' t2
  | ST_App2 : forall v1 t2 t2',
         value v1
      -> t2 ~~> t2'
      -> tm_app v1 t2 ~~> tm_app v1 t2'

where "t1 '~~>' t2" := (step t1 t2).

Tactic Notation "step_cases" tactic(first) tactic(c) :=
  first;
  [ c "ST_AppAbs" | c "ST_App1" | c "ST_App2" ].

Notation "t1 '~~>' t2" := (step t1 t2) (at level 40).

Notation stepmany := (refl_step_closure step).
Notation "t1 '~~>*' t2" := (stepmany t1 t2) (at level 40).

Hint Constructors step.

Examples


Hint Constructors refl_step_closure.
Hint Unfold id_A id_AarrowA id_AarrowA_arrow_AarrowA k.

Lemma step_example1 :
  (tm_app id_AarrowA id_A) ~~>* id_A.
Proof.
  unfold id_AarrowA. unfold id_A.
  eapply rsc_step.
    unfold id_AarrowA. apply ST_AppAbs.
    unfold id_A. apply v_abs.
  simpl.
  apply rsc_refl. Qed.

Lemma step_example2 :
  (tm_app id_AarrowA (tm_app id_AarrowA id_A)) ~~>* id_A.
Proof.
  unfold id_AarrowA. unfold id_A.
  eapply rsc_step.
    apply ST_App2. auto.
    apply ST_AppAbs. auto.
  eapply rsc_step.
    apply ST_AppAbs. simpl. auto.
  simpl. apply rsc_refl. Qed.

Exercise: 2 stars (step_example3)

Lemma step_example3 :
       (tm_app (tm_app id_AarrowA_arrow_AarrowA id_AarrowA) id_A)
  ~~>* id_A.
Proof.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Typing


Contexts

Question: What is the type of the term "x tm_true"?

Answer: It depends on the type of x!

I.e., in order to assign a type to a term, we need to know what assumptions we should make about the types of its free variables.

This leads us to a three-place "typing judgment", informally written Gamma |- t : T, where Gamma is a "typing context" -- a mapping from variables to their types.

Definition context := id -> (option ty).

Definition empty : context := (fun _ => None).

Definition extend (Gamma : context) (x:id) (T : ty) :=
  fun x' => if beq_id x x' then Some T else Gamma x'.

Typing relation


Inductive has_type : context -> tm -> ty -> Prop :=
  | T_Var : forall Gamma x T,
      Gamma x = Some T ->
      has_type Gamma (tm_var x) T
  | T_Abs : forall Gamma x T11 T12 t12,
      has_type (extend Gamma x T11) t12 T12 ->
      has_type Gamma (tm_abs x T11 t12) (ty_arrow T11 T12)
  | T_App : forall T1 T2 Gamma t1 t2,
      has_type Gamma t1 (ty_arrow T1 T2) ->
      has_type Gamma t2 T1 ->
      has_type Gamma (tm_app t1 t2) T2.

Tactic Notation "typing_cases" tactic(first) tactic(c) :=
  first;
  [ c "T_Var" | c "T_Abs" | c "T_App" ].

Hint Constructors has_type.

Examples


Example typing_example_1 :
  has_type empty (tm_abs x A (tm_var x)) (ty_arrow A A).
Proof.
  apply T_Abs. apply T_Var. reflexivity. Qed.

(* Note that since we added the has_type constructors to the hints
   database, auto can actually solve this one immediately *)

Example typing_example_1' :
  has_type empty (tm_abs x A (tm_var x)) (ty_arrow A A).
Proof. auto. Qed.

Hint Unfold beq_id beq_nat extend.

Example typing_example_2 :
  has_type empty
    (tm_abs x A
       (tm_abs y (ty_arrow A A)
          (tm_app (tm_var y) (tm_app (tm_var y) (tm_var x)))))
    (ty_arrow A (ty_arrow (ty_arrow A A) A)).
(* I.e.:
     empty |- \x:A. \y:A->A. y (y x)) 
           : A -> (A->A) -> A.
*)

Proof with auto.
  apply T_Abs.
  apply T_Abs.
  eapply T_App. apply T_Var. unfold extend. unfold beq_id. unfold beq_nat. reflexivity.
  eapply T_App. apply T_Var. unfold extend. unfold beq_id. unfold beq_nat. reflexivity.
  apply T_Var. unfold extend. unfold beq_id. unfold beq_nat. reflexivity. Qed.

Exercise: 2 stars

(* Prove the same result without using autoeauto, or eapply *)
Example typing_example_2_full :
  has_type empty
    (tm_abs x A
       (tm_abs y (ty_arrow A A)
          (tm_app (tm_var y) (tm_app (tm_var y) (tm_var x)))))
    (ty_arrow A (ty_arrow (ty_arrow A A) A)).
(* I.e.:
     empty |- \x:A. \y:A->A. y (y x)) 
           : A -> (A->A) -> A.
*)

Proof.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Exercise: 2 stars (typing_example_3)

Example typing_example_3 :
  exists T,
    has_type empty
      (tm_abs x (ty_arrow A B)
         (tm_abs y (ty_arrow B C)
            (tm_abs z A
               (tm_app (tm_var y) (tm_app (tm_var x) (tm_var z))))))
      T.
(* i.e., 
    empty |- (\x:A->B. \y:B-->C. \z:A.
                y (x z)) 
          : T.
*)

Proof with auto.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Example typing_nonexample_1 :
  ~ exists T,
      has_type empty
        (tm_abs x A
            (tm_abs y B
               (tm_app (tm_var x) (tm_var y))))
        T.
(*
  ~ exists T,
      empty |- (\x:A. \y:B, x y) : T.
*)

Proof.
  intros C. destruct C.
  (* The clear tactic is useful here for tidying away bits of
     the context that we're not going to need again. *)

  inversion H. subst. clear H.
  inversion H5. subst. clear H5.
  inversion H4. subst. clear H4.
  inversion H2. subst. clear H2.
  inversion H5. subst. clear H5.
  compute in H1. inversion H1. Qed.
(* CJC : this last use of inversion also isn't obvious without 
  compute (simple isn't enough, at least in 8.1) *)


Exercise: 2 stars (typing_nonexample_2)

Example typing_nonexample_2 :
  ~ exists T,
      has_type empty
        (tm_abs x (ty_arrow A A)
           (tm_abs y B
              (tm_app (tm_var x) (tm_var y))))
        T.
(* I.e.,
  ~ exists T,
      empty |- (\x:A->A, \y:B, x y) : T.
*)

Proof.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Exercise: 3 stars (typing_nonexample_3)

Example typing_nonexample_3 :
  ~ (exists S, exists T,
        has_type empty
          (tm_abs x S
             (tm_app (tm_var x) (tm_var x)))
          T).
(* I.e.
  ~ (exists S, exists T,
        empty |- (\x:S. x x) : T).
*)

Proof.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Exercise: 1 star (typing_statements)


Which of the following propositions are provable?
  • y:B |- \x:A.x : A->A
  • exists T, empty |- (\y:B->B. \x:B. y x) : T
  • exists T, empty |- (\y:B->B. \x:B. x y) : T
  • exists S, x:S |- (\y:B->B. y) x : S
  • exists S, exists T, x:S |- (x x x) : T

Exercise: 1 star (more_typing_statements)


Which of the following propositions are provable? For the ones that are, give witnesses for the existentially bound variables.
  • exists T, empty |- (\y:B->B->B. \x:B, y x) : T
  • exists T, empty |- (\x:A->B, \y:B-->C, \z:A, y (x z)):T
  • exists S, exists U, exists T, x:S, y:U |- (\z:A. x y z) : T
  • exists S, exists T, x:S |- \y:A. x (x y) : T
  • exists S, exists U, exists T, x:S |- x (\z:U. z x) : T

Properties


Free variables


Inductive appears_free_in : id -> tm -> Prop :=
  | afi_var : forall x,
      appears_free_in x (tm_var x)
  | afi_app1 : forall x t1 t2,
      appears_free_in x t1 -> appears_free_in x (tm_app t1 t2)
  | afi_app2 : forall x t1 t2,
      appears_free_in x t2 -> appears_free_in x (tm_app t1 t2)
  | afi_abs : forall x y T11 t12,
        y <> x
     -> appears_free_in x t12
     -> appears_free_in x (tm_abs y T11 t12).

Tactic Notation "afi_cases" tactic(first) tactic(c) :=
  first;
  [ c "afi_var" | c "afi_app1" | c "afi_app2" | c "afi_abs" ].

Hint Constructors appears_free_in.

Definition closed (t:tm) :=
  forall x, ~ appears_free_in x t.

Substitution


Lemma free_in_context : forall x t T Gamma,
   appears_free_in x t ->
   has_type Gamma t T ->
   exists T', Gamma x = Some T'.
Proof.
  intros. generalize dependent Gamma. generalize dependent T.
  (afi_cases (induction H) Case); intros.
  Case "afi_var".
    inversion H0; subst. exists T. assumption.
  Case "afi_app1".
    inversion H0; subst. eapply IHappears_free_in. apply H4.
  Case "afi_app2".
    inversion H0; subst. eapply IHappears_free_in. apply H6.
  Case "afi_abs".
    inversion H1; subst.
    apply IHappears_free_in in H7. unfold extend in H7.
    apply not_eq_false_beqid in H. rewrite <- H in H7.
    assumption. Qed.

Exercise: 2 stars, optional

Corollary typable_empty__closed : forall t T,
    has_type empty t T
 -> closed t.
Proof.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Lemma context_invariance : forall Gamma Gamma' t S,
     has_type Gamma t S
  -> (forall x, appears_free_in x t -> Gamma x = Gamma' x)
  -> has_type Gamma' t S.
Proof with eauto.
  intros.
  generalize dependent Gamma'.
  (typing_cases (induction H) Case); intros.
  Case "T_Var".
    apply T_Var. rewrite <- H0...
  Case "T_Abs".
    apply T_Abs.
    apply IHhas_type. intros x0 Hafi.
    (* tricky step...the Gamma' we use to instantiate is 
       extend ty Gamma x T1 *)

    unfold extend. remember (beq_id x x0) as e. destruct e...
  Case "T_App".
    apply T_App with T1... Qed.

Lemma substitution_preserves_typing : forall Gamma x U v t S,
     has_type (extend Gamma x U) t S
  -> has_type empty v U
  -> has_type Gamma (subst x v t) S.
Proof with eauto.
  intros Gamma x U v t S Ht Hv.
  generalize dependent Gamma. generalize dependent S.
  (tm_cases (induction t) Case); intros S Gamma H;
    (* in each case, we'll want to get at the derivation of H *)
    inversion H; subst; simpl...
  Case "tm_var".
    rename i into y. remember (beq_id x y) as e. destruct e.
    SCase "x=y".
      apply beq_id_eq in Heqe. subst.
      unfold extend in H2. rewrite <- beq_id_refl in H2.
      inversion H2; subst. clear H2.
      eapply context_invariance... intros x Hcontra.
      destruct (free_in_context _ _ S empty Hcontra) as [T' HT']...
      inversion HT'.
    SCase "x<>y".
      apply T_Var. unfold extend in H2. rewrite <- Heqe in H2...
  Case "tm_abs".
    rename i into y. apply T_Abs.
    remember (beq_id x y) as e. destruct e.
    SCase "x=y".
      eapply context_invariance...
      apply beq_id_eq in Heqe. subst.
      intros x Hafi. unfold extend.
      destruct (beq_id y x)...
    SCase "x<>y".
      apply IHt. eapply context_invariance...
      intros z Hafi. unfold extend.
      remember (beq_id y z) as e0. destruct e0...
      apply beq_id_eq in Heqe0. subst.
      rewrite <- Heqe... Qed.

Preservation


Theorem preservation : forall t t' T,
     has_type empty t T
  -> t ~~> t'
  -> has_type empty t' T.
Proof with eauto.
  remember empty as Gamma.
  intros t t' T HT. generalize dependent t'.
  (typing_cases (induction HT) Case); intros t' HE; subst Gamma.
  Case "T_Var".
    inversion HE.
  Case "T_Abs".
    inversion HE.
  Case "T_App".
    (step_cases (inversion HE) SCase); subst...
    (* The ST_App1 and ST_App2 cases are immediate by induction, and
       auto takes care of them *)

    SCase "ST_AppAbs".
      apply substitution_preserves_typing with T1...
      inversion HT1... Qed.

Exercise: 2 stars (subject_expansion_stlc)

An exercise earlier in this file asked about the subject expansion property for the simple language of arithmetic and boolean expressions. Does this property hold for STLC? That is, is it always the case that, if t ~~> t' and has_type t' T, then has_type t T? If so, prove it. If not, give a counter-example.

FILL IN HERE...

Progress


Theorem progress : forall t T,
     has_type empty t T
  -> value t \/ exists t', t ~~> t'.
Proof with eauto.
  intros t T Ht.
  remember empty as Gamma.
  (typing_cases (induction Ht) Case); subst Gamma...
  Case "T_Var".
    inversion H.

  Case "T_App".
    right. destruct IHHt1...

    SCase "t1 is a value".
      destruct IHHt2...
      SSCase "t2 is a value".
        (* since t2 is a value and has an arrow type, it must be an abs.
           Sometimes, this is proved separately and called a "canonical 
           forms" lemma *)

        inversion H; subst. exists (subst x t2 t)...
      SSCase "t2 steps".
        destruct H0 as [t2' Hstp]. exists (tm_app t1 t2')...

    SCase "t1 steps".
      destruct H as [t1' Hstp]. exists (tm_app t1' t2)...
Qed.

Exercise: 3 stars, optional

Show that progress can also be proved by induction on terms instead of types.
Theorem progress' : forall t T,
     has_type empty t T
  -> value t \/ exists t', t ~~> t'.
Proof.
  intros t.
  (tm_cases (induction t) Case); intros T Ht.
  (* FILL IN HERE (and delete "Admitted") *) Admitted.

Uniqueness of types


Exercise: 3 stars, optional

Another pleasant property of the STLC is that types are unique: a given term (in a given context) has at most one type. Formalize this statement and prove it.

(* FILL IN HERE *)

Additional exercises


Exercise: 1 star

Without peeking, write down the progress and preservation theorems for the simply typed lambda-calculus.

Exercise: 2 stars (stlc_variation1)

Suppose we add the following new rule to the evaluation relation of the STLC:
         | T_Strange : forall x t,
                          has_type empty (tm_abs x A t) B


Which of the following properties of the STLC remain true in the presence of this rule? For each one, write either "remains true" or else "becomes false." If a property becomes false, give a counterexample.
  • Determinacy of step
  • Progress
  • Preservation

Exercise: 2 stars (stlc_variation2)

Suppose we remove the rule ST_App1 from the step relation. Which of the three properties in the previous exercise become false in the absence of this rule? For each that becomes false, give a counterexample.

End STLC.

STLC with arithmetic


To see how the STLC might function as the core of a real programming language, let's extend it with a concrete base type of numbers and some constants and primitive operators.

Module STLCArith.

(* Syntax and operational semantics *)

To types, we add a base type of natural numbers (and remove the other base types, for brevity)
Inductive ty : Set :=
  | ty_arrow : ty -> ty -> ty
  | ty_nat : ty.

To terms, we add natural number constants, along with successor, predicate, and zero-testing...
Inductive tm : Set :=
  | tm_var : id -> tm
  | tm_app : tm -> tm -> tm
  | tm_abs : id -> ty -> tm -> tm
  | tm_nat : nat -> tm
  | tm_succ : tm -> tm
  | tm_pred : tm -> tm
  | tm_mult : tm -> tm -> tm
  | tm_if0 : tm -> tm -> tm -> tm.

Tactic Notation "tm_cases" tactic(first) tactic(c) :=
  first;
  [ c "tm_var" | c "tm_app" | c "tm_abs" |
    c "tm_nat" | c "tm_succ" | c "tm_pred" | c "tm_mult" | c "tm_if0" ].

Exercise: 4 stars (STLCArith)

Finish formalizing the definition and properties of the STLC extended with arithmetic. Specifically:
  • Copy the whole development of STLC that we went through above (from the definition of values through the Progress theorem), and paste it into the file at this point.
  • Extend the definitions of the subst operation and the step relation to include appropriate clauses for the arithmetic operators.
  • Extend the proofs of all the properties of the original STLC to deal with the new syntactic forms. Make sure Coq accepts the whole file.
(* FILL IN HERE *)

End STLCArith.

Version of 4/28/2009