adamc@227
|
1 (* Copyright (c) 2009, Adam Chlipala
|
adamc@227
|
2 *
|
adamc@227
|
3 * This work is licensed under a
|
adamc@227
|
4 * Creative Commons Attribution-Noncommercial-No Derivative Works 3.0
|
adamc@227
|
5 * Unported License.
|
adamc@227
|
6 * The license text is available at:
|
adamc@227
|
7 * http://creativecommons.org/licenses/by-nc-nd/3.0/
|
adamc@227
|
8 *)
|
adamc@227
|
9
|
adamc@227
|
10 (* begin hide *)
|
adamc@230
|
11 Require Import DepList Tactics.
|
adamc@227
|
12
|
adamc@227
|
13 Set Implicit Arguments.
|
adamc@227
|
14 (* end hide *)
|
adamc@227
|
15
|
adamc@227
|
16
|
adamc@227
|
17 (** %\chapter{Universes and Axioms}% *)
|
adamc@227
|
18
|
adamc@228
|
19 (** Many traditional theorems can be proved in Coq without special knowledge of CIC, the logic behind the prover. A development just seems to be using a particular ASCII notation for standard formulas based on set theory. Nonetheless, as we saw in Chapter 4, CIC differs from set theory in starting from fewer orthogonal primitives. It is possible to define the usual logical connectives as derived notions. The foundation of it all is a dependently-typed functional programming language, based on dependent function types and inductive type families. By using the facilities of this language directly, we can accomplish some things much more easily than in mainstream math.
|
adamc@227
|
20
|
adamc@227
|
21 Gallina, which adds features to the more theoretical CIC, is the logic implemented in Coq. It has a relatively simple foundation that can be defined rigorously in a page or two of formal proof rules. Still, there are some important subtleties that have practical ramifications. This chapter focuses on those subtleties, avoiding formal metatheory in favor of example code. *)
|
adamc@227
|
22
|
adamc@227
|
23
|
adamc@227
|
24 (** * The [Type] Hierarchy *)
|
adamc@227
|
25
|
adamc@227
|
26 (** Every object in Gallina has a type. *)
|
adamc@227
|
27
|
adamc@227
|
28 Check 0.
|
adamc@227
|
29 (** %\vspace{-.15in}% [[
|
adamc@227
|
30 0
|
adamc@227
|
31 : nat
|
adamc@227
|
32
|
adamc@227
|
33 ]]
|
adamc@227
|
34
|
adamc@227
|
35 It is natural enough that zero be considered as a natural number. *)
|
adamc@227
|
36
|
adamc@227
|
37 Check nat.
|
adamc@227
|
38 (** %\vspace{-.15in}% [[
|
adamc@227
|
39 nat
|
adamc@227
|
40 : Set
|
adamc@227
|
41
|
adamc@227
|
42 ]]
|
adamc@227
|
43
|
adamc@227
|
44 From a set theory perspective, it is unsurprising to consider the natural numbers as a "set." *)
|
adamc@227
|
45
|
adamc@227
|
46 Check Set.
|
adamc@227
|
47 (** %\vspace{-.15in}% [[
|
adamc@227
|
48 Set
|
adamc@227
|
49 : Type
|
adamc@227
|
50
|
adamc@227
|
51 ]]
|
adamc@227
|
52
|
adamc@227
|
53 The type [Set] may be considered as the set of all sets, a concept that set theory handles in terms of %\textit{%#<i>#classes#</i>#%}%. In Coq, this more general notion is [Type]. *)
|
adamc@227
|
54
|
adamc@227
|
55 Check Type.
|
adamc@227
|
56 (** %\vspace{-.15in}% [[
|
adamc@227
|
57 Type
|
adamc@227
|
58 : Type
|
adamc@227
|
59
|
adamc@227
|
60 ]]
|
adamc@227
|
61
|
adamc@228
|
62 Strangely enough, [Type] appears to be its own type. It is known that polymorphic languages with this property are inconsistent. That is, using such a language to encode proofs is unwise, because it is possible to "prove" any proposition. What is really going on here?
|
adamc@227
|
63
|
adamc@227
|
64 Let us repeat some of our queries after toggling a flag related to Coq's printing behavior. *)
|
adamc@227
|
65
|
adamc@227
|
66 Set Printing Universes.
|
adamc@227
|
67
|
adamc@227
|
68 Check nat.
|
adamc@227
|
69 (** %\vspace{-.15in}% [[
|
adamc@227
|
70 nat
|
adamc@227
|
71 : Set
|
adamc@227
|
72 ]] *)
|
adamc@227
|
73
|
adamc@227
|
74 (** printing $ %({}*% #(<a/>*# *)
|
adamc@227
|
75 (** printing ^ %*{})% #*<a/>)# *)
|
adamc@227
|
76
|
adamc@227
|
77 Check Set.
|
adamc@227
|
78 (** %\vspace{-.15in}% [[
|
adamc@227
|
79 Set
|
adamc@227
|
80 : Type $ (0)+1 ^
|
adamc@227
|
81
|
adamc@227
|
82 ]] *)
|
adamc@227
|
83
|
adamc@227
|
84 Check Type.
|
adamc@227
|
85 (** %\vspace{-.15in}% [[
|
adamc@227
|
86 Type $ Top.3 ^
|
adamc@227
|
87 : Type $ (Top.3)+1 ^
|
adamc@227
|
88
|
adamc@227
|
89 ]]
|
adamc@227
|
90
|
adamc@228
|
91 Occurrences of [Type] are annotated with some additional information, inside comments. These annotations have to do with the secret behind [Type]: it Sreally stands for an infinite hierarchy of types. The type of [Set] is [Type(0)], the type of [Type(0)] is [Type(1)], the type of [Type(1)] is [Type(2)], and so on. This is how we avoid the "[Type : Type]" paradox. As a convenience, the universe hierarchy drives Coq's one variety of subtyping. Any term whose type is [Type] at level [i] is automatically also described by [Type] at level [j] when [j > i].
|
adamc@227
|
92
|
adamc@227
|
93 In the outputs of our first [Check] query, we see that the type level of [Set]'s type is [(0)+1]. Here [0] stands for the level of [Set], and we increment it to arrive at the level that %\textit{%#<i>#classifies#</i>#%}% [Set].
|
adamc@227
|
94
|
adamc@227
|
95 In the second query's output, we see that the occurrence of [Type] that we check is assigned a fresh %\textit{%#<i>#universe variable#</i>#%}% [Top.3]. The output type increments [Top.3] to move up a level in the universe hierarchy. As we write code that uses definitions whose types mention universe variables, unification may refine the values of those variables. Luckily, the user rarely has to worry about the details.
|
adamc@227
|
96
|
adamc@227
|
97 Another crucial concept in CIC is %\textit{%#<i>#predicativity#</i>#%}%. Consider these queries. *)
|
adamc@227
|
98
|
adamc@227
|
99 Check forall T : nat, fin T.
|
adamc@227
|
100 (** %\vspace{-.15in}% [[
|
adamc@227
|
101 forall T : nat, fin T
|
adamc@227
|
102 : Set
|
adamc@227
|
103 ]] *)
|
adamc@227
|
104
|
adamc@227
|
105 Check forall T : Set, T.
|
adamc@227
|
106 (** %\vspace{-.15in}% [[
|
adamc@227
|
107 forall T : Set, T
|
adamc@227
|
108 : Type $ max(0, (0)+1) ^
|
adamc@227
|
109 ]] *)
|
adamc@227
|
110
|
adamc@227
|
111 Check forall T : Type, T.
|
adamc@227
|
112 (** %\vspace{-.15in}% [[
|
adamc@227
|
113 forall T : Type $ Top.9 ^ , T
|
adamc@227
|
114 : Type $ max(Top.9, (Top.9)+1) ^
|
adamc@227
|
115
|
adamc@227
|
116 ]]
|
adamc@227
|
117
|
adamc@227
|
118 These outputs demonstrate the rule for determining which universe a [forall] type lives in. In particular, for a type [forall x : T1, T2], we take the maximum of the universes of [T1] and [T2]. In the first example query, both [T1] ([nat]) and [T2] ([fin T]) are in [Set], so the [forall] type is in [Set], too. In the second query, [T1] is [Set], which is at level [(0)+1]; and [T2] is [T], which is at level [0]. Thus, the [forall] exists at the maximum of these two levels. The third example illustrates the same outcome, where we replace [Set] with an occurrence of [Type] that is assigned universe variable [Top.9]. This universe variable appears in the places where [0] appeared in the previous query.
|
adamc@227
|
119
|
adamc@227
|
120 The behind-the-scenes manipulation of universe variables gives us predicativity. Consider this simple definition of a polymorphic identity function. *)
|
adamc@227
|
121
|
adamc@227
|
122 Definition id (T : Set) (x : T) : T := x.
|
adamc@227
|
123
|
adamc@227
|
124 Check id 0.
|
adamc@227
|
125 (** %\vspace{-.15in}% [[
|
adamc@227
|
126 id 0
|
adamc@227
|
127 : nat
|
adamc@227
|
128
|
adamc@227
|
129 Check id Set.
|
adamc@227
|
130
|
adamc@227
|
131 Error: Illegal application (Type Error):
|
adamc@227
|
132 ...
|
adamc@228
|
133 The 1st term has type "Type $ (Top.15)+1 ^" which should be coercible to "Set".
|
adamc@227
|
134
|
adamc@227
|
135 ]]
|
adamc@227
|
136
|
adamc@227
|
137 The parameter [T] of [id] must be instantiated with a [Set]. [nat] is a [Set], but [Set] is not. We can try fixing the problem by generalizing our definition of [id]. *)
|
adamc@227
|
138
|
adamc@227
|
139 Reset id.
|
adamc@227
|
140 Definition id (T : Type) (x : T) : T := x.
|
adamc@227
|
141 Check id 0.
|
adamc@227
|
142 (** %\vspace{-.15in}% [[
|
adamc@227
|
143 id 0
|
adamc@227
|
144 : nat
|
adamc@227
|
145 ]] *)
|
adamc@227
|
146
|
adamc@227
|
147 Check id Set.
|
adamc@227
|
148 (** %\vspace{-.15in}% [[
|
adamc@227
|
149 id Set
|
adamc@227
|
150 : Type $ Top.17 ^
|
adamc@227
|
151 ]] *)
|
adamc@227
|
152
|
adamc@227
|
153 Check id Type.
|
adamc@227
|
154 (** %\vspace{-.15in}% [[
|
adamc@227
|
155 id Type $ Top.18 ^
|
adamc@227
|
156 : Type $ Top.19 ^
|
adamc@227
|
157 ]] *)
|
adamc@227
|
158
|
adamc@227
|
159 (** So far so good. As we apply [id] to different [T] values, the inferred index for [T]'s [Type] occurrence automatically moves higher up the type hierarchy.
|
adamc@227
|
160
|
adamc@227
|
161 [[
|
adamc@227
|
162 Check id id.
|
adamc@227
|
163
|
adamc@227
|
164 Error: Universe inconsistency (cannot enforce Top.16 < Top.16).
|
adamc@227
|
165
|
adamc@227
|
166 ]]
|
adamc@227
|
167
|
adamc@228
|
168 This error message reminds us that the universe variable for [T] still exists, even though it is usually hidden. To apply [id] to itself, that variable would need to be less than itself in the type hierarchy. Universe inconsistency error messages announce cases like this one where a term could only type-check by violating an implied constraint over universe variables. Such errors demonstrate that [Type] is %\textit{%#<i>#predicative#</i>#%}%, where this word has a CIC meaning closely related to its usual mathematical meaning. A predicative system enforces the constraint that, for any object of quantified type, none of those quantifiers may ever be instantiated with the object itself. Impredicativity is associated with popular paradoxes in set theory, involving inconsistent constructions like "the set of all sets that do not contain themselves." Similar paradoxes result from uncontrolled impredicativity in Coq. *)
|
adamc@227
|
169
|
adamc@227
|
170
|
adamc@227
|
171 (** ** Inductive Definitions *)
|
adamc@227
|
172
|
adamc@227
|
173 (** Predicativity restrictions also apply to inductive definitions. As an example, let us consider a type of expression trees that allows injection of any native Coq value. The idea is that an [exp T] stands for a reflected expression of type [T].
|
adamc@227
|
174
|
adamc@227
|
175 [[
|
adamc@227
|
176 Inductive exp : Set -> Set :=
|
adamc@227
|
177 | Const : forall T : Set, T -> exp T
|
adamc@227
|
178 | Pair : forall T1 T2, exp T1 -> exp T2 -> exp (T1 * T2)
|
adamc@227
|
179 | Eq : forall T, exp T -> exp T -> exp bool.
|
adamc@227
|
180
|
adamc@227
|
181 Error: Large non-propositional inductive types must be in Type.
|
adamc@227
|
182
|
adamc@227
|
183 ]]
|
adamc@227
|
184
|
adamc@227
|
185 This definition is %\textit{%#<i>#large#</i>#%}% in the sense that at least one of its constructors takes an argument whose type has type [Type]. Coq would be inconsistent if we allowed definitions like this one in their full generality. Instead, we must change [exp] to live in [Type]. We will go even further and move [exp]'s index to [Type] as well. *)
|
adamc@227
|
186
|
adamc@227
|
187 Inductive exp : Type -> Type :=
|
adamc@227
|
188 | Const : forall T, T -> exp T
|
adamc@227
|
189 | Pair : forall T1 T2, exp T1 -> exp T2 -> exp (T1 * T2)
|
adamc@227
|
190 | Eq : forall T, exp T -> exp T -> exp bool.
|
adamc@227
|
191
|
adamc@228
|
192 (** Note that before we had to include an annotation [: Set] for the variable [T] in [Const]'s type, but we need no annotation now. When the type of a variable is not known, and when that variable is used in a context where only types are allowed, Coq infers that the variable is of type [Type]. That is the right behavior here, but it was wrong for the [Set] version of [exp].
|
adamc@228
|
193
|
adamc@228
|
194 Our new definition is accepted. We can build some sample expressions. *)
|
adamc@227
|
195
|
adamc@227
|
196 Check Const 0.
|
adamc@227
|
197 (** %\vspace{-.15in}% [[
|
adamc@227
|
198 Const 0
|
adamc@227
|
199 : exp nat
|
adamc@227
|
200 ]] *)
|
adamc@227
|
201
|
adamc@227
|
202 Check Pair (Const 0) (Const tt).
|
adamc@227
|
203 (** %\vspace{-.15in}% [[
|
adamc@227
|
204 Pair (Const 0) (Const tt)
|
adamc@227
|
205 : exp (nat * unit)
|
adamc@227
|
206 ]] *)
|
adamc@227
|
207
|
adamc@227
|
208 Check Eq (Const Set) (Const Type).
|
adamc@227
|
209 (** %\vspace{-.15in}% [[
|
adamc@228
|
210 Eq (Const Set) (Const Type $ Top.59 ^ )
|
adamc@227
|
211 : exp bool
|
adamc@227
|
212
|
adamc@227
|
213 ]]
|
adamc@227
|
214
|
adamc@227
|
215 We can check many expressions, including fancy expressions that include types. However, it is not hard to hit a type-checking wall.
|
adamc@227
|
216
|
adamc@227
|
217 [[
|
adamc@227
|
218 Check Const (Const O).
|
adamc@227
|
219
|
adamc@227
|
220 Error: Universe inconsistency (cannot enforce Top.42 < Top.42).
|
adamc@227
|
221
|
adamc@227
|
222 ]]
|
adamc@227
|
223
|
adamc@227
|
224 We are unable to instantiate the parameter [T] of [Const] with an [exp] type. To see why, it is helpful to print the annotated version of [exp]'s inductive definition. *)
|
adamc@227
|
225
|
adamc@227
|
226 Print exp.
|
adamc@227
|
227 (** %\vspace{-.15in}% [[
|
adamc@227
|
228 Inductive exp
|
adamc@227
|
229 : Type $ Top.8 ^ ->
|
adamc@227
|
230 Type
|
adamc@227
|
231 $ max(0, (Top.11)+1, (Top.14)+1, (Top.15)+1, (Top.19)+1) ^ :=
|
adamc@227
|
232 Const : forall T : Type $ Top.11 ^ , T -> exp T
|
adamc@227
|
233 | Pair : forall (T1 : Type $ Top.14 ^ ) (T2 : Type $ Top.15 ^ ),
|
adamc@227
|
234 exp T1 -> exp T2 -> exp (T1 * T2)
|
adamc@227
|
235 | Eq : forall T : Type $ Top.19 ^ , exp T -> exp T -> exp bool
|
adamc@227
|
236
|
adamc@227
|
237 ]]
|
adamc@227
|
238
|
adamc@227
|
239 We see that the index type of [exp] has been assigned to universe level [Top.8]. In addition, each of the four occurrences of [Type] in the types of the constructors gets its own universe variable. Each of these variables appears explicitly in the type of [exp]. In particular, any type [exp T] lives at a universe level found by incrementing by one the maximum of the four argument variables. A consequence of this is that [exp] %\textit{%#<i>#must#</i>#%}% live at a higher universe level than any type which may be passed to one of its constructors. This consequence led to the universe inconsistency.
|
adamc@227
|
240
|
adamc@227
|
241 Strangely, the universe variable [Top.8] only appears in one place. Is there no restriction imposed on which types are valid arguments to [exp]? In fact, there is a restriction, but it only appears in a global set of universe constraints that are maintained "off to the side," not appearing explicitly in types. We can print the current database. *)
|
adamc@227
|
242
|
adamc@227
|
243 Print Universes.
|
adamc@227
|
244 (** %\vspace{-.15in}% [[
|
adamc@227
|
245 Top.19 < Top.9 <= Top.8
|
adamc@227
|
246 Top.15 < Top.9 <= Top.8 <= Coq.Init.Datatypes.38
|
adamc@227
|
247 Top.14 < Top.9 <= Top.8 <= Coq.Init.Datatypes.37
|
adamc@227
|
248 Top.11 < Top.9 <= Top.8
|
adamc@227
|
249
|
adamc@227
|
250 ]]
|
adamc@227
|
251
|
adamc@227
|
252 [Print Universes] outputs many more constraints, but we have collected only those that mention [Top] variables. We see one constraint for each universe variable associated with a constructor argument from [exp]'s definition. [Top.19] is the type argument to [Eq]. The constraint for [Top.19] effectively says that [Top.19] must be less than [Top.8], the universe of [exp]'s indices; an intermediate variable [Top.9] appears as an artifact of the way the constraint was generated.
|
adamc@227
|
253
|
adamc@227
|
254 The next constraint, for [Top.15], is more complicated. This is the universe of the second argument to the [Pair] constructor. Not only must [Top.15] be less than [Top.8], but it also comes out that [Top.8] must be less than [Coq.Init.Datatypes.38]. What is this new universe variable? It is from the definition of the [prod] inductive family, to which types of the form [A * B] are desugared. *)
|
adamc@227
|
255
|
adamc@227
|
256 Print prod.
|
adamc@227
|
257 (** %\vspace{-.15in}% [[
|
adamc@227
|
258 Inductive prod (A : Type $ Coq.Init.Datatypes.37 ^ )
|
adamc@227
|
259 (B : Type $ Coq.Init.Datatypes.38 ^ )
|
adamc@227
|
260 : Type $ max(Coq.Init.Datatypes.37, Coq.Init.Datatypes.38) ^ :=
|
adamc@227
|
261 pair : A -> B -> A * B
|
adamc@227
|
262
|
adamc@227
|
263 ]]
|
adamc@227
|
264
|
adamc@227
|
265 We see that the constraint is enforcing that indices to [exp] must not live in a higher universe level than [B]-indices to [prod]. The next constraint above establishes a symmetric condition for [A].
|
adamc@227
|
266
|
adamc@227
|
267 Thus it is apparent that Coq maintains a tortuous set of universe variable inequalities behind the scenes. It may look like some functions are polymorphic in the universe levels of their arguments, but what is really happening is imperative updating of a system of constraints, such that all uses of a function are consistent with a global set of universe levels. When the constraint system may not be evolved soundly, we get a universe inconsistency error.
|
adamc@227
|
268
|
adamc@227
|
269 %\medskip%
|
adamc@227
|
270
|
adamc@227
|
271 Something interesting is revealed in the annotated definition of [prod]. A type [prod A B] lives at a universe that is the maximum of the universes of [A] and [B]. From our earlier experiments, we might expect that [prod]'s universe would in fact need to be %\textit{%#<i>#one higher#</i>#%}% than the maximum. The critical difference is that, in the definition of [prod], [A] and [B] are defined as %\textit{%#<i>#parameters#</i>#%}%; that is, they appear named to the left of the main colon, rather than appearing (possibly unnamed) to the right.
|
adamc@227
|
272
|
adamc@227
|
273 Parameters are not as flexible as normal inductive type arguments. The range types of all of the constructors of a parameterized type must share the same indices. Nonetheless, when it is possible to define a polymorphic type in this way, we gain the ability to use the new type family in more ways, without triggering universe inconsistencies. For instance, nested pairs of types are perfectly legal. *)
|
adamc@227
|
274
|
adamc@227
|
275 Check (nat, (Type, Set)).
|
adamc@227
|
276 (** %\vspace{-.15in}% [[
|
adamc@227
|
277 (nat, (Type $ Top.44 ^ , Set))
|
adamc@227
|
278 : Set * (Type $ Top.45 ^ * Type $ Top.46 ^ )
|
adamc@227
|
279
|
adamc@227
|
280 ]]
|
adamc@227
|
281
|
adamc@227
|
282 The same cannot be done with a counterpart to [prod] that does not use parameters. *)
|
adamc@227
|
283
|
adamc@227
|
284 Inductive prod' : Type -> Type -> Type :=
|
adamc@227
|
285 | pair' : forall A B : Type, A -> B -> prod' A B.
|
adamc@227
|
286
|
adamc@227
|
287 (** [[
|
adamc@227
|
288 Check (pair' nat (pair' Type Set)).
|
adamc@227
|
289
|
adamc@227
|
290 Error: Universe inconsistency (cannot enforce Top.51 < Top.51).
|
adamc@227
|
291
|
adamc@227
|
292 ]]
|
adamc@227
|
293
|
adamc@227
|
294 The key benefit parameters bring us is the ability to avoid quantifying over types in the types of constructors. Such quantification induces less-than constraints, while parameters only introduce less-than-or-equal-to constraints. *)
|
adamc@229
|
295
|
adamc@229
|
296 (* begin hide *)
|
adamc@229
|
297 Unset Printing Universes.
|
adamc@229
|
298 (* end hide *)
|
adamc@229
|
299
|
adamc@229
|
300
|
adamc@229
|
301 (** * The [Prop] Universe *)
|
adamc@229
|
302
|
adamc@229
|
303 (** In Chapter 4, we saw parallel versions of useful datatypes for "programs" and "proofs." The convention was that programs live in [Set], and proofs live in [Prop]. We gave little explanation for why it is useful to maintain this distinction. There is certainly documentation value from separating programs from proofs; in practice, different concerns apply to building the two types of objects. It turns out, however, that these concerns motivate formal differences between the two universes in Coq.
|
adamc@229
|
304
|
adamc@229
|
305 Recall the types [sig] and [ex], which are the program and proof versions of existential quantification. Their definitions differ only in one place, where [sig] uses [Type] and [ex] uses [Prop]. *)
|
adamc@229
|
306
|
adamc@229
|
307 Print sig.
|
adamc@229
|
308 (** %\vspace{-.15in}% [[
|
adamc@229
|
309 Inductive sig (A : Type) (P : A -> Prop) : Type :=
|
adamc@229
|
310 exist : forall x : A, P x -> sig P
|
adamc@229
|
311 ]] *)
|
adamc@229
|
312
|
adamc@229
|
313 Print ex.
|
adamc@229
|
314 (** %\vspace{-.15in}% [[
|
adamc@229
|
315 Inductive ex (A : Type) (P : A -> Prop) : Prop :=
|
adamc@229
|
316 ex_intro : forall x : A, P x -> ex P
|
adamc@229
|
317
|
adamc@229
|
318 ]]
|
adamc@229
|
319
|
adamc@229
|
320 It is natural to want a function to extract the first components of data structures like these. Doing so is easy enough for [sig]. *)
|
adamc@229
|
321
|
adamc@229
|
322 Definition projS A (P : A -> Prop) (x : sig P) : A :=
|
adamc@229
|
323 match x with
|
adamc@229
|
324 | exist v _ => v
|
adamc@229
|
325 end.
|
adamc@229
|
326
|
adamc@229
|
327 (** We run into trouble with a version that has been changed to work with [ex].
|
adamc@229
|
328
|
adamc@229
|
329 [[
|
adamc@229
|
330 Definition projE A (P : A -> Prop) (x : ex P) : A :=
|
adamc@229
|
331 match x with
|
adamc@229
|
332 | ex_intro v _ => v
|
adamc@229
|
333 end.
|
adamc@229
|
334
|
adamc@229
|
335 Error:
|
adamc@229
|
336 Incorrect elimination of "x" in the inductive type "ex":
|
adamc@229
|
337 the return type has sort "Type" while it should be "Prop".
|
adamc@229
|
338 Elimination of an inductive object of sort Prop
|
adamc@229
|
339 is not allowed on a predicate in sort Type
|
adamc@229
|
340 because proofs can be eliminated only to build proofs.
|
adamc@229
|
341
|
adamc@229
|
342 ]]
|
adamc@229
|
343
|
adamc@230
|
344 In formal Coq parlance, "elimination" means "pattern-matching." The typing rules of Gallina forbid us from pattern-matching on a discriminee whose type belongs to [Prop], whenever the result type of the [match] has a type besides [Prop]. This is a sort of "information flow" policy, where the type system ensures that the details of proofs can never have any effect on parts of a development that are not also marked as proofs.
|
adamc@229
|
345
|
adamc@229
|
346 This restriction matches informal practice. We think of programs and proofs as clearly separated, and, outside of constructive logic, the idea of computing with proofs is ill-formed. The distinction also has practical importance in Coq, where it affects the behavior of extraction.
|
adamc@229
|
347
|
adamc@229
|
348 Recall that extraction is Coq's facility for translating Coq developments into programs in general-purpose programming languages like OCaml. Extraction %\textit{%#<i>#erases#</i>#%}% proofs and leaves programs intact. A simple example with [sig] and [ex] demonstrates the distinction. *)
|
adamc@229
|
349
|
adamc@229
|
350 Definition sym_sig (x : sig (fun n => n = 0)) : sig (fun n => 0 = n) :=
|
adamc@229
|
351 match x with
|
adamc@229
|
352 | exist n pf => exist _ n (sym_eq pf)
|
adamc@229
|
353 end.
|
adamc@229
|
354
|
adamc@229
|
355 Extraction sym_sig.
|
adamc@229
|
356 (** <<
|
adamc@229
|
357 (** val sym_sig : nat -> nat **)
|
adamc@229
|
358
|
adamc@229
|
359 let sym_sig x = x
|
adamc@229
|
360 >>
|
adamc@229
|
361
|
adamc@229
|
362 Since extraction erases proofs, the second components of [sig] values are elided, making [sig] a simple identity type family. The [sym_sig] operation is thus an identity function. *)
|
adamc@229
|
363
|
adamc@229
|
364 Definition sym_ex (x : ex (fun n => n = 0)) : ex (fun n => 0 = n) :=
|
adamc@229
|
365 match x with
|
adamc@229
|
366 | ex_intro n pf => ex_intro _ n (sym_eq pf)
|
adamc@229
|
367 end.
|
adamc@229
|
368
|
adamc@229
|
369 Extraction sym_ex.
|
adamc@229
|
370 (** <<
|
adamc@229
|
371 (** val sym_ex : __ **)
|
adamc@229
|
372
|
adamc@229
|
373 let sym_ex = __
|
adamc@229
|
374 >>
|
adamc@229
|
375
|
adamc@229
|
376 In this example, the [ex] type itself is in [Prop], so whole [ex] packages are erased. Coq extracts every proposition as the type %\texttt{%#<tt>#__#</tt>#%}%, whose single constructor is %\texttt{%#<tt>#__#</tt>#%}%. Not only are proofs replaced by [__], but proof arguments to functions are also removed completely, as we see here.
|
adamc@229
|
377
|
adamc@229
|
378 Extraction is very helpful as an optimization over programs that contain proofs. In languages like Haskell, advanced features make it possible to program with proofs, as a way of convincing the type checker to accept particular definitions. Unfortunately, when proofs are encoded as values in GADTs, these proofs exist at runtime and consume resources. In contrast, with Coq, as long as you keep all of your proofs within [Prop], extraction is guaranteed to erase them.
|
adamc@229
|
379
|
adamc@229
|
380 Many fans of the Curry-Howard correspondence support the idea of %\textit{%#<i>#extracting programs from proofs#</i>#%}%. In reality, few users of Coq and related tools do any such thing. Instead, extraction is better thought of as an optimization that reduces the runtime costs of expressive typing.
|
adamc@229
|
381
|
adamc@229
|
382 %\medskip%
|
adamc@229
|
383
|
adamc@229
|
384 We have seen two of the differences between proofs and programs: proofs are subject to an elimination restriction and are elided by extraction. The remaining difference is that [Prop] is %\textit{%#<i>#impredicative#</i>#%}%, as this example shows. *)
|
adamc@229
|
385
|
adamc@229
|
386 Check forall P Q : Prop, P \/ Q -> Q \/ P.
|
adamc@229
|
387 (** %\vspace{-.15in}% [[
|
adamc@229
|
388 forall P Q : Prop, P \/ Q -> Q \/ P
|
adamc@229
|
389 : Prop
|
adamc@229
|
390
|
adamc@229
|
391 ]]
|
adamc@229
|
392
|
adamc@230
|
393 We see that it is possible to define a [Prop] that quantifies over other [Prop]s. This is fortunate, as we start wanting that ability even for such basic purposes as stating propositional tautologies. In the next section of this chapter, we will see some reasons why unrestricted impredicativity is undesirable. The impredicativity of [Prop] interacts crucially with the elimination restriction to avoid those pitfalls.
|
adamc@230
|
394
|
adamc@230
|
395 Impredicativity also allows us to implement a version of our earlier [exp] type that does not suffer from the weakness that we found. *)
|
adamc@230
|
396
|
adamc@230
|
397 Inductive expP : Type -> Prop :=
|
adamc@230
|
398 | ConstP : forall T, T -> expP T
|
adamc@230
|
399 | PairP : forall T1 T2, expP T1 -> expP T2 -> expP (T1 * T2)
|
adamc@230
|
400 | EqP : forall T, expP T -> expP T -> expP bool.
|
adamc@230
|
401
|
adamc@230
|
402 Check ConstP 0.
|
adamc@230
|
403 (** %\vspace{-.15in}% [[
|
adamc@230
|
404 ConstP 0
|
adamc@230
|
405 : expP nat
|
adamc@230
|
406 ]] *)
|
adamc@230
|
407
|
adamc@230
|
408 Check PairP (ConstP 0) (ConstP tt).
|
adamc@230
|
409 (** %\vspace{-.15in}% [[
|
adamc@230
|
410 PairP (ConstP 0) (ConstP tt)
|
adamc@230
|
411 : expP (nat * unit)
|
adamc@230
|
412 ]] *)
|
adamc@230
|
413
|
adamc@230
|
414 Check EqP (ConstP Set) (ConstP Type).
|
adamc@230
|
415 (** %\vspace{-.15in}% [[
|
adamc@230
|
416 EqP (ConstP Set) (ConstP Type)
|
adamc@230
|
417 : expP bool
|
adamc@230
|
418 ]] *)
|
adamc@230
|
419
|
adamc@230
|
420 Check ConstP (ConstP O).
|
adamc@230
|
421 (** %\vspace{-.15in}% [[
|
adamc@230
|
422 ConstP (ConstP 0)
|
adamc@230
|
423 : expP (expP nat)
|
adamc@230
|
424
|
adamc@230
|
425 ]]
|
adamc@230
|
426
|
adamc@230
|
427 In this case, our victory is really a shallow one. As we have marked [expP] as a family of proofs, we cannot deconstruct our expressions in the usual programmatic ways, which makes them almost useless for the usual purposes. Impredicative quantification is much more useful in defining inductive families that we really think of as judgments. For instance, this code defines a notion of equality that is strictly stronger than the base equality [=]. *)
|
adamc@230
|
428
|
adamc@230
|
429 Inductive eqPlus : forall T, T -> T -> Prop :=
|
adamc@230
|
430 | Base : forall T (x : T), eqPlus x x
|
adamc@230
|
431 | Func : forall dom ran (f1 f2 : dom -> ran),
|
adamc@230
|
432 (forall x : dom, eqPlus (f1 x) (f2 x))
|
adamc@230
|
433 -> eqPlus f1 f2.
|
adamc@230
|
434
|
adamc@230
|
435 Check (Base 0).
|
adamc@230
|
436 (** %\vspace{-.15in}% [[
|
adamc@230
|
437 Base 0
|
adamc@230
|
438 : eqPlus 0 0
|
adamc@230
|
439 ]] *)
|
adamc@230
|
440
|
adamc@230
|
441 Check (Func (fun n => n) (fun n => 0 + n) (fun n => Base n)).
|
adamc@230
|
442 (** %\vspace{-.15in}% [[
|
adamc@230
|
443 Func (fun n : nat => n) (fun n : nat => 0 + n) (fun n : nat => Base n)
|
adamc@230
|
444 : eqPlus (fun n : nat => n) (fun n : nat => 0 + n)
|
adamc@230
|
445 ]] *)
|
adamc@230
|
446
|
adamc@230
|
447 Check (Base (Base 1)).
|
adamc@230
|
448 (** %\vspace{-.15in}% [[
|
adamc@230
|
449 Base (Base 1)
|
adamc@230
|
450 : eqPlus (Base 1) (Base 1)
|
adamc@230
|
451 ]] *)
|
adamc@230
|
452
|
adamc@230
|
453
|
adamc@230
|
454 (** * Axioms *)
|
adamc@230
|
455
|
adamc@230
|
456 (** While the specific logic Gallina is hardcoded into Coq's implementation, it is possible to add certain logical rules in a controlled way. In other words, Coq may be used to reason about many different refinements of Gallina where strictly more theorems are provable. We achieve this by asserting %\textit{%#<i>#axioms#</i>#%}% without proof.
|
adamc@230
|
457
|
adamc@230
|
458 We will motivate the idea by touring through some standard axioms, as enumerated in Coq's online FAQ. I will add additional commentary as appropriate. *)
|
adamc@230
|
459
|
adamc@230
|
460 (** ** The Basics *)
|
adamc@230
|
461
|
adamc@230
|
462 (* One simple example of a useful axiom is the law of the excluded middle. *)
|
adamc@230
|
463
|
adamc@230
|
464 Require Import Classical_Prop.
|
adamc@230
|
465 Print classic.
|
adamc@230
|
466 (** %\vspace{-.15in}% [[
|
adamc@230
|
467 *** [ classic : forall P : Prop, P \/ ~ P ]
|
adamc@230
|
468
|
adamc@230
|
469 ]]
|
adamc@230
|
470
|
adamc@230
|
471 In the implementation of module [Classical_Prop], this axiom was defined with the command *)
|
adamc@230
|
472
|
adamc@230
|
473 Axiom classic : forall P : Prop, P \/ ~ P.
|
adamc@230
|
474
|
adamc@230
|
475 (** An [Axiom] may be declared with any type, in any of the universes. There is a synonym [Parameter] for [Axiom], and that synonym is often clearer for assertions not of type [Prop]. For instance, we can assert the existence of objects with certain properties. *)
|
adamc@230
|
476
|
adamc@230
|
477 Parameter n : nat.
|
adamc@230
|
478 Axiom positive : n > 0.
|
adamc@230
|
479 Reset n.
|
adamc@230
|
480
|
adamc@230
|
481 (** This kind of "axiomatic presentation" of a theory is very common outside of higher-order logic. However, in Coq, it is almost always preferable to stick to defining your objects, functions, and predicates via inductive definitions and functional programming.
|
adamc@230
|
482
|
adamc@230
|
483 In general, there is a significant burden associated with any use of axioms. It is easy to assert a set of axioms that together is %\textit{%#<i>#inconsistent#</i>#%}%. That is, a set of axioms may imply [False], which allows any theorem to proved, which defeats the purpose of a proof assistant. For example, we could assert the following axiom, which is consistent by itself but inconsistent when combined with [classic]. *)
|
adamc@230
|
484
|
adamc@230
|
485 Axiom not_classic : exists P : Prop, ~ (P \/ ~ P).
|
adamc@230
|
486
|
adamc@230
|
487 Theorem uhoh : False.
|
adamc@230
|
488 generalize classic not_classic; firstorder.
|
adamc@230
|
489 Qed.
|
adamc@230
|
490
|
adamc@230
|
491 Theorem uhoh_again : 1 + 1 = 3.
|
adamc@230
|
492 destruct uhoh.
|
adamc@230
|
493 Qed.
|
adamc@230
|
494
|
adamc@230
|
495 Reset not_classic.
|
adamc@230
|
496
|
adamc@230
|
497 (** On the subject of the law of the excluded middle itself, this axiom is usually quite harmless, and many practical Coq developments assume it. It has been proved metatheoretically to be consistent with CIC. Here, "proved metatheoretically" means that someone proved on paper that excluded middle holds in a %\textit{%#<i>#model#</i>#%}% of CIC in set theory. All of the other axioms that we will survey in this section hold in the same model, so they are all consistent together.
|
adamc@230
|
498
|
adamc@230
|
499 Recall that Coq implements %\textit{%#<i>#constructive#</i>#%}% logic by default, where excluded middle is not provable. Proofs in constructive logic can be thought of as programs. A [forall] quantifier denotes a dependent function type, and a disjunction denotes a variant type. In such a setting, excluded middle could be interpreted as a decision procedure for arbitrary propositions, which computability theory tells us cannot exist. Thus, constructive logic with excluded middle can no longer be associated with our usual notion of programming.
|
adamc@230
|
500
|
adamc@230
|
501 Given all this, why is it all right to assert excluded middle as an axiom? I do not want to go into the technical details, but the intuitive justification is that the elimination restriction for [Prop] prevents us from treating proofs as programs. An excluded middle axiom that quantified over [Set] instead of [Prop] %\textit{%#<i>#would#</i>#%}% be problematic. If a development used that axiom, we would not be able to extract the code to OCaml (soundly) without implementing a genuine universal decision procedure. In contrast, values whose types belong to [Prop] are always erased by extraction, so we sidestep the axiom's algorithmic consequences.
|
adamc@230
|
502
|
adamc@230
|
503 Because the proper use of axioms is so precarious, there are helpful commands for determining which axioms a theorem relies on. *)
|
adamc@230
|
504
|
adamc@230
|
505 Theorem t1 : forall P : Prop, P -> ~ ~ P.
|
adamc@230
|
506 tauto.
|
adamc@230
|
507 Qed.
|
adamc@230
|
508
|
adamc@230
|
509 Print Assumptions t1.
|
adamc@230
|
510 (** %\vspace{-.15in}% [[
|
adamc@230
|
511 Closed under the global context
|
adamc@230
|
512 ]] *)
|
adamc@230
|
513
|
adamc@230
|
514 Theorem t2 : forall P : Prop, ~ ~ P -> P.
|
adamc@230
|
515 (** [[
|
adamc@230
|
516 tauto.
|
adamc@230
|
517
|
adamc@230
|
518 Error: tauto failed.
|
adamc@230
|
519
|
adamc@230
|
520 ]] *)
|
adamc@230
|
521
|
adamc@230
|
522 intro P; destruct (classic P); tauto.
|
adamc@230
|
523 Qed.
|
adamc@230
|
524
|
adamc@230
|
525 Print Assumptions t2.
|
adamc@230
|
526 (** %\vspace{-.15in}% [[
|
adamc@230
|
527 Axioms:
|
adamc@230
|
528 classic : forall P : Prop, P \/ ~ P
|
adamc@230
|
529
|
adamc@230
|
530 ]]
|
adamc@230
|
531
|
adamc@230
|
532 It is possible to avoid this dependence in some specific cases, where excluded middle %\textit{%#<i>#is#</i>#%}% provable, for decidable propositions. *)
|
adamc@230
|
533
|
adamc@230
|
534 Theorem classic_nat_eq : forall n m : nat, n = m \/ n <> m.
|
adamc@230
|
535 induction n; destruct m; intuition; generalize (IHn m); intuition.
|
adamc@230
|
536 Qed.
|
adamc@230
|
537
|
adamc@230
|
538 Theorem t2' : forall n m : nat, ~ ~ (n = m) -> n = m.
|
adamc@230
|
539 intros n m; destruct (classic_nat_eq n m); tauto.
|
adamc@230
|
540 Qed.
|
adamc@230
|
541
|
adamc@230
|
542 Print Assumptions t2'.
|
adamc@230
|
543 (** %\vspace{-.15in}% [[
|
adamc@230
|
544 Closed under the global context
|
adamc@230
|
545 ]]
|
adamc@230
|
546
|
adamc@230
|
547 %\bigskip%
|
adamc@230
|
548
|
adamc@230
|
549 Mainstream mathematical practice assumes excluded middle, so it can be useful to have it available in Coq developments, though it is also nice to know that a theorem is proved in a simpler formal system than classical logic. There is a similar story for %\textit{%#<i>#proof irrelevance#</i>#%}%, which simplifies proof issues that would not even arise in mainstream math. *)
|
adamc@230
|
550
|
adamc@230
|
551 Require Import ProofIrrelevance.
|
adamc@230
|
552 Print proof_irrelevance.
|
adamc@230
|
553 (** %\vspace{-.15in}% [[
|
adamc@230
|
554 *** [ proof_irrelevance : forall (P : Prop) (p1 p2 : P), p1 = p2 ]
|
adamc@230
|
555
|
adamc@230
|
556 ]]
|
adamc@230
|
557
|
adamc@230
|
558 This axiom asserts that any two proofs of the same proposition are equal. If we replaced [p1 = p2] by [p1 <-> p2], then the statement would be provable. However, without this axiom, equality is a stronger notion than logical equivalence. Recall this example function from Chapter 6. *)
|
adamc@230
|
559
|
adamc@230
|
560 (* begin hide *)
|
adamc@230
|
561 Lemma zgtz : 0 > 0 -> False.
|
adamc@230
|
562 crush.
|
adamc@230
|
563 Qed.
|
adamc@230
|
564 (* end hide *)
|
adamc@230
|
565
|
adamc@230
|
566 Definition pred_strong1 (n : nat) : n > 0 -> nat :=
|
adamc@230
|
567 match n with
|
adamc@230
|
568 | O => fun pf : 0 > 0 => match zgtz pf with end
|
adamc@230
|
569 | S n' => fun _ => n'
|
adamc@230
|
570 end.
|
adamc@230
|
571
|
adamc@230
|
572 (** We might want to prove that different proofs of [n > 0] do not lead to different results from our richly-typed predecessor function. *)
|
adamc@230
|
573
|
adamc@230
|
574 Theorem pred_strong1_irrel : forall n (pf1 pf2 : n > 0), pred_strong1 pf1 = pred_strong1 pf2.
|
adamc@230
|
575 destruct n; crush.
|
adamc@230
|
576 Qed.
|
adamc@230
|
577
|
adamc@230
|
578 (** The proof script is simple, but it involved peeking into the definition of [pred_strong1]. For more complicated function definitions, it can be considerably more work to prove that they do not discriminate on details of proof arguments. This can seem like a shame, since the [Prop] elimination restriction makes it impossible to write any function that does otherwise. Unfortunately, this fact is only true metatheoretically, unless we assert an axiom like [proof_irrelevance]. With that axiom, we can prove our theorem without consulting the definition of [pred_strong1]. *)
|
adamc@230
|
579
|
adamc@230
|
580 Theorem pred_strong1_irrel' : forall n (pf1 pf2 : n > 0), pred_strong1 pf1 = pred_strong1 pf2.
|
adamc@230
|
581 intros; f_equal; apply proof_irrelevance.
|
adamc@230
|
582 Qed.
|
adamc@230
|
583
|
adamc@230
|
584
|
adamc@230
|
585 (** %\bigskip%
|
adamc@230
|
586
|
adamc@230
|
587 In the chapter on equality, we already discussed some axioms that are related to proof irrelevance. In particular, Coq's standard library includes this axiom: *)
|
adamc@230
|
588
|
adamc@230
|
589 Require Import Eqdep.
|
adamc@230
|
590 Import Eq_rect_eq.
|
adamc@230
|
591 Print eq_rect_eq.
|
adamc@230
|
592 (** %\vspace{-.15in}% [[
|
adamc@230
|
593 *** [ eq_rect_eq :
|
adamc@230
|
594 forall (U : Type) (p : U) (Q : U -> Type) (x : Q p) (h : p = p),
|
adamc@230
|
595 x = eq_rect p Q x p h ]
|
adamc@230
|
596
|
adamc@230
|
597 ]]
|
adamc@230
|
598
|
adamc@230
|
599 This axiom says that it is permissible to simplify pattern matches over proofs of equalities like [e = e]. The axiom is logically equivalent to some simpler corollaries. *)
|
adamc@230
|
600
|
adamc@230
|
601 Corollary UIP_refl : forall A (x : A) (pf : x = x), pf = refl_equal x.
|
adamc@230
|
602 intros; replace pf with (eq_rect x (eq x) (refl_equal x) x pf); [
|
adamc@230
|
603 symmetry; apply eq_rect_eq
|
adamc@230
|
604 | exact (match pf as pf' return match pf' in _ = y return x = y with
|
adamc@230
|
605 | refl_equal => refl_equal x
|
adamc@230
|
606 end = pf' with
|
adamc@230
|
607 | refl_equal => refl_equal _
|
adamc@230
|
608 end) ].
|
adamc@230
|
609 Qed.
|
adamc@230
|
610
|
adamc@230
|
611 Corollary UIP : forall A (x y : A) (pf1 pf2 : x = y), pf1 = pf2.
|
adamc@230
|
612 intros; generalize pf1 pf2; subst; intros;
|
adamc@230
|
613 match goal with
|
adamc@230
|
614 | [ |- ?pf1 = ?pf2 ] => rewrite (UIP_refl pf1); rewrite (UIP_refl pf2); reflexivity
|
adamc@230
|
615 end.
|
adamc@230
|
616 Qed.
|
adamc@230
|
617
|
adamc@230
|
618 (** These corollaries are special cases of proof irrelevance. Many developments only need proof irrelevance for equality, so there is no need to assert full irrelevance for them.
|
adamc@230
|
619
|
adamc@230
|
620 Another facet of proof irrelevance is that, like excluded middle, it is often provable for specific propositions. For instance, [UIP] is provable whenever the type [A] has a decidable equality operation. The module [Eqdep_dec] of the standard library contains a proof. A similar phenomenon applies to other notable cases, including less-than proofs. Thus, it is often possible to use proof irrelevance without asserting axioms.
|
adamc@230
|
621
|
adamc@230
|
622 %\bigskip%
|
adamc@230
|
623
|
adamc@230
|
624 There are two more basic axioms that are often assumed, to avoid complications that do not arise in set theory. *)
|
adamc@230
|
625
|
adamc@230
|
626 Require Import FunctionalExtensionality.
|
adamc@230
|
627 Print functional_extensionality_dep.
|
adamc@230
|
628 (** %\vspace{-.15in}% [[
|
adamc@230
|
629 *** [ functional_extensionality_dep :
|
adamc@230
|
630 forall (A : Type) (B : A -> Type) (f g : forall x : A, B x),
|
adamc@230
|
631 (forall x : A, f x = g x) -> f = g ]
|
adamc@230
|
632
|
adamc@230
|
633 ]]
|
adamc@230
|
634
|
adamc@230
|
635 This axiom says that two functions are equal if they map equal inputs to equal outputs. Such facts are not provable in general in CIC, but it is consistent to assume that they are.
|
adamc@230
|
636
|
adamc@230
|
637 A simple corollary shows that the same property applies to predicates. In some cases, one might prefer to assert this corollary as the axiom, to restrict the consequences to proofs and not programs. *)
|
adamc@230
|
638
|
adamc@230
|
639 Corollary predicate_extensionality : forall (A : Type) (B : A -> Prop) (f g : forall x : A, B x),
|
adamc@230
|
640 (forall x : A, f x = g x) -> f = g.
|
adamc@230
|
641 intros; apply functional_extensionality_dep; assumption.
|
adamc@230
|
642 Qed.
|
adamc@230
|
643
|
adamc@230
|
644
|
adamc@230
|
645 (** ** Axioms of Choice *)
|
adamc@230
|
646
|
adamc@230
|
647 (** Some Coq axioms are also points of contention in mainstream math. The most prominent example is the axiom of choice. In fact, there are multiple versions that we might consider, and, considered in isolation, none of these versions means quite what it means in classical set theory.
|
adamc@230
|
648
|
adamc@230
|
649 First, it is possible to implement a choice operator %\textit{%#<i>#without#</i>#%}% axioms in some potentially surprising cases. *)
|
adamc@230
|
650
|
adamc@230
|
651 Require Import ConstructiveEpsilon.
|
adamc@230
|
652 Check constructive_definite_description.
|
adamc@230
|
653 (** %\vspace{-.15in}% [[
|
adamc@230
|
654 constructive_definite_description
|
adamc@230
|
655 : forall (A : Set) (f : A -> nat) (g : nat -> A),
|
adamc@230
|
656 (forall x : A, g (f x) = x) ->
|
adamc@230
|
657 forall P : A -> Prop,
|
adamc@230
|
658 (forall x : A, {P x} + {~ P x}) ->
|
adamc@230
|
659 (exists! x : A, P x) -> {x : A | P x}
|
adamc@230
|
660 ]] *)
|
adamc@230
|
661
|
adamc@230
|
662 Print Assumptions constructive_definite_description.
|
adamc@230
|
663 (** %\vspace{-.15in}% [[
|
adamc@230
|
664 Closed under the global context
|
adamc@230
|
665
|
adamc@230
|
666 ]]
|
adamc@230
|
667
|
adamc@230
|
668 This function transforms a decidable predicate [P] into a function that produces an element satisfying [P] from a proof that such an element exists. The functions [f] and [g], plus an associated injectivity property, are used to express the idea that the set [A] is countable. Under these conditions, a simple brute force algorithm gets the job done: we just enumerate all elements of [A], stopping when we find one satisfying [P]. The existence proof, specified in terms of %\textit{%#<i>#unique#</i>#%}% existence [exists!], guarantees termination. The definition of this operator in Coq uses some interesting techniques, as seen in the implementation of the [ConstructiveEpsilon] module.
|
adamc@230
|
669
|
adamc@230
|
670 Countable choice is provable in set theory without appealing to the general axiom of choice. To support the more general principle in Coq, we must also add an axiom. Here is a functional version of the axiom of unique choice. *)
|
adamc@230
|
671
|
adamc@230
|
672 Require Import ClassicalUniqueChoice.
|
adamc@230
|
673 Check dependent_unique_choice.
|
adamc@230
|
674 (** %\vspace{-.15in}% [[
|
adamc@230
|
675 dependent_unique_choice
|
adamc@230
|
676 : forall (A : Type) (B : A -> Type) (R : forall x : A, B x -> Prop),
|
adamc@230
|
677 (forall x : A, exists! y : B x, R x y) ->
|
adamc@230
|
678 exists f : forall x : A, B x, forall x : A, R x (f x)
|
adamc@230
|
679
|
adamc@230
|
680 ]]
|
adamc@230
|
681
|
adamc@230
|
682 This axiom lets us convert a relational specification [R] into a function implementing that specification. We need only prove that [R] is truly a function. An alternate, stronger formulation applies to cases where [R] maps each input to one or more outputs. We also simplify the statement of the theorem by considering only non-dependent function types. *)
|
adamc@230
|
683
|
adamc@230
|
684 Require Import ClassicalChoice.
|
adamc@230
|
685 Check choice.
|
adamc@230
|
686 (** %\vspace{-.15in}% [[
|
adamc@230
|
687 choice
|
adamc@230
|
688 : forall (A B : Type) (R : A -> B -> Prop),
|
adamc@230
|
689 (forall x : A, exists y : B, R x y) ->
|
adamc@230
|
690 exists f : A -> B, forall x : A, R x (f x)
|
adamc@230
|
691
|
adamc@230
|
692 ]]
|
adamc@230
|
693
|
adamc@230
|
694 This principle is proved as a theorem, based on the unique choice axiom and an additional axiom of relational choice from the [RelationalChoice] module.
|
adamc@230
|
695
|
adamc@230
|
696 In set theory, the axiom of choice is a fundamental philosophical commitment one makes about the universe of sets. In Coq, the choice axioms say something weaker. For instance, consider the simple restatement of the [choice] axiom where we replace existential quantification by its Curry-Howard analogue, subset types. *)
|
adamc@230
|
697
|
adamc@230
|
698 Definition choice_Set (A B : Type) (R : A -> B -> Prop) (H : forall x : A, {y : B | R x y})
|
adamc@230
|
699 : {f : A -> B | forall x : A, R x (f x)} :=
|
adamc@230
|
700 exist (fun f => forall x : A, R x (f x))
|
adamc@230
|
701 (fun x => proj1_sig (H x)) (fun x => proj2_sig (H x)).
|
adamc@230
|
702
|
adamc@230
|
703 (** Via the Curry-Howard correspondence, this "axiom" can be taken to have the same meaning as the original. It is implemented trivially as a transformation not much deeper than uncurrying. Thus, we see that the utility of the axioms that we mentioned earlier comes in their usage to build programs from proofs. Normal set theory has no explicit proofs, so the meaning of the usual axiom of choice is subtlely different. In Gallina, the axioms implement a controlled relaxation of the restrictions on information flow from proofs to programs.
|
adamc@230
|
704
|
adamc@230
|
705 However, when we combine an axiom of choice with the law of the excluded middle, the idea of "choice" becomes more interesting. Excluded middle gives us a highly non-computational way of constructing proofs, but it does not change the computational nature of programs. Thus, the axiom of choice is still giving us a way of translating between two different sorts of "programs," but the input programs (which are proofs) may be written in a rich language that goes beyond normal computability. This truly is more than repackaging a function with a different type.
|
adamc@230
|
706
|
adamc@230
|
707 %\bigskip%
|
adamc@230
|
708
|
adamc@230
|
709 The Coq tools support a command-line flag %\texttt{%#<tt>#-impredicative-set#</tt>#%}%, which modifies Gallina in a more fundamental way by making [Set] impredicative. A term like [forall T : Set, T] has type [Set], and inductive definitions in [Set] may have constructors that quantify over arguments of any types. To maintain consistency, an elimination restriction must be imposed, similarly to the restriction for [Prop]. The restriction only applies to large inductive types, where some constructor quantifies over a type of type [Type]. In such cases, a value in this inductive type may only be pattern-matched over to yield a result type whose type is [Set] or [Prop]. This contrasts with [Prop], where the restriction applies even to non-large inductive types, and where the result type may only have type [Prop].
|
adamc@230
|
710
|
adamc@230
|
711 In old versions of Coq, [Set] was impredicative by default. Later versions make [Set] predicative to avoid inconsistency with some classical axioms. In particular, one should watch out when using impredicative [Set] with axioms of choice. In combination with excluded middle or predicate extensionality, this can lead to inconsistency. Impredicative [Set] can be useful for modeling inherently impredicative mathematical concepts, but almost all Coq developments get by fine without it. *)
|
adamc@230
|
712
|
adamc@230
|
713 (** ** Axioms and Computation *)
|
adamc@230
|
714
|
adamc@230
|
715 (** One additional axiom-related wrinkle arises from an aspect of Gallina that is very different from set theory: a notion of %\textit{%#<i>#computational equivalence#</i>#%}% is central to the definition of the formal system. Axioms tend not to play well with computation. Consider this example. We start by implementing a function that uses a type equality proof to perform a safe type-cast. *)
|
adamc@230
|
716
|
adamc@230
|
717 Definition cast (x y : Set) (pf : x = y) (v : x) : y :=
|
adamc@230
|
718 match pf with
|
adamc@230
|
719 | refl_equal => v
|
adamc@230
|
720 end.
|
adamc@230
|
721
|
adamc@230
|
722 (** Computation over programs that use [cast] can proceed smoothly. *)
|
adamc@230
|
723
|
adamc@230
|
724 Eval compute in (cast (refl_equal (nat -> nat)) (fun n => S n)) 12.
|
adamc@230
|
725 (** [[
|
adamc@230
|
726 = 13
|
adamc@230
|
727 : nat
|
adamc@230
|
728 ]] *)
|
adamc@230
|
729
|
adamc@230
|
730 (** Things do not go as smoothly when we use [cast] with proofs that rely on axioms. *)
|
adamc@230
|
731
|
adamc@230
|
732 Theorem t3 : (forall n : nat, fin (S n)) = (forall n : nat, fin (n + 1)).
|
adamc@230
|
733 change ((forall n : nat, (fun n => fin (S n)) n) = (forall n : nat, (fun n => fin (n + 1)) n));
|
adamc@230
|
734 rewrite (functional_extensionality (fun n => fin (n + 1)) (fun n => fin (S n))); crush.
|
adamc@230
|
735 Qed.
|
adamc@230
|
736
|
adamc@230
|
737 Eval compute in (cast t3 (fun _ => First)) 12.
|
adamc@230
|
738 (** [[
|
adamc@230
|
739 = match t3 in (_ = P) return P with
|
adamc@230
|
740 | refl_equal => fun n : nat => First
|
adamc@230
|
741 end 12
|
adamc@230
|
742 : fin (12 + 1)
|
adamc@230
|
743
|
adamc@230
|
744 ]]
|
adamc@230
|
745
|
adamc@230
|
746 Computation gets stuck in a pattern-match on the proof [t3]. The structure of [t3] is not known, so the match cannot proceed. It turns out a more basic problem leads to this particular situation. We ended the proof of [t3] with [Qed], so the definition of [t3] is not available to computation. That is easily fixed. *)
|
adamc@230
|
747
|
adamc@230
|
748 Reset t3.
|
adamc@230
|
749
|
adamc@230
|
750 Theorem t3 : (forall n : nat, fin (S n)) = (forall n : nat, fin (n + 1)).
|
adamc@230
|
751 change ((forall n : nat, (fun n => fin (S n)) n) = (forall n : nat, (fun n => fin (n + 1)) n));
|
adamc@230
|
752 rewrite (functional_extensionality (fun n => fin (n + 1)) (fun n => fin (S n))); crush.
|
adamc@230
|
753 Defined.
|
adamc@230
|
754
|
adamc@230
|
755 Eval compute in (cast t3 (fun _ => First)) 12.
|
adamc@230
|
756 (** [[
|
adamc@230
|
757 = match
|
adamc@230
|
758 match
|
adamc@230
|
759 match
|
adamc@230
|
760 functional_extensionality
|
adamc@230
|
761 ....
|
adamc@230
|
762
|
adamc@230
|
763 ]]
|
adamc@230
|
764
|
adamc@230
|
765 We elide most of the details. A very unwieldy tree of nested matches on equality proofs appears. This time evaluation really %\textit{%#<i>#is#</i>#%}% stuck on a use of an axiom.
|
adamc@230
|
766
|
adamc@230
|
767 If we are careful in using tactics to prove an equality, we can still compute with casts over the proof. *)
|
adamc@230
|
768
|
adamc@230
|
769 Lemma plus1 : forall n, S n = n + 1.
|
adamc@230
|
770 induction n; simpl; intuition.
|
adamc@230
|
771 Defined.
|
adamc@230
|
772
|
adamc@230
|
773 Theorem t4 : forall n, fin (S n) = fin (n + 1).
|
adamc@230
|
774 intro; f_equal; apply plus1.
|
adamc@230
|
775 Defined.
|
adamc@230
|
776
|
adamc@230
|
777 Eval compute in cast (t4 13) First.
|
adamc@230
|
778 (** %\vspace{-.15in}% [[
|
adamc@230
|
779 = First
|
adamc@230
|
780 : fin (13 + 1)
|
adamc@230
|
781 ]] *)
|