1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274
|
.. _sprop:
SProp (proof irrelevant propositions)
=====================================
.. warning::
The status of strict propositions is experimental.
In particular, conversion checking through bytecode or native code
compilation currently does not understand proof irrelevance.
This section describes the extension of Coq with definitionally
proof irrelevant propositions (types in the sort :math:`\SProp`, also
known as strict propositions) as described in
:cite:`Gilbert:POPL2019`.
Use of |SProp| may be disabled by passing ``-disallow-sprop`` to the
Coq program or by turning the :flag:`Allow StrictProp` flag off.
.. flag:: Allow StrictProp
This :term:`flag` enables or disables the use of |SProp|. It is enabled by default.
The command-line flag ``-disallow-sprop`` disables |SProp| at
startup.
.. exn:: SProp is disallowed because the "Allow StrictProp" flag is off.
:undocumented:
Some of the definitions described in this document are available
through ``Coq.Logic.StrictProp``, which see.
Basic constructs
----------------
The purpose of :math:`\SProp` is to provide types where all elements
are convertible:
.. coqtop:: all
Theorem irrelevance (A : SProp) (P : A -> Prop) : forall x : A, P x -> forall y : A, P y.
Proof.
intros * Hx *.
exact Hx.
Qed.
Since we have definitional :ref:`eta-expansion-sect` for
functions, the property of being a type of definitionally irrelevant
values is impredicative, and so is :math:`\SProp`:
.. coqtop:: all
Check fun (A:Type) (B:A -> SProp) => (forall x:A, B x) : SProp.
In order to keep conversion tractable, cumulativity for :math:`\SProp`
is forbidden.
.. coqtop:: all
Fail Check (fun (A:SProp) => A : Type).
We can explicitly lift strict propositions into the relevant world by
using a wrapping inductive type. The inductive stops definitional
proof irrelevance from escaping.
.. coqtop:: in
Inductive Box (A:SProp) : Prop := box : A -> Box A.
Arguments box {_} _.
.. coqtop:: all
Fail Check fun (A:SProp) (x y : Box A) => eq_refl : x = y.
.. doesn't get merged with the above if coqdoc
.. coqtop:: in
Definition box_irrelevant (A:SProp) (x y : Box A) : x = y
:= match x, y with box x, box y => eq_refl end.
In the other direction, we can use impredicativity to "squash" a
relevant type, making an irrelevant approximation.
.. coqdoc::
Definition iSquash (A:Type) : SProp
:= forall P : SProp, (A -> P) -> P.
Definition isquash A : A -> iSquash A
:= fun a P f => f a.
Definition iSquash_sind A (P : iSquash A -> SProp) (H : forall x : A, P (isquash A x))
: forall x : iSquash A, P x
:= fun x => x (P x) (H : A -> P x).
Or more conveniently (but equivalently)
.. coqdoc::
Inductive Squash (A:Type) : SProp := squash : A -> Squash A.
Most inductives types defined in :math:`\SProp` are squashed types,
i.e. they can only be eliminated to construct proofs of other strict
propositions. Empty types are the only exception.
.. coqtop:: in
Inductive sEmpty : SProp := .
.. coqtop:: all
Check sEmpty_rect.
.. note::
Eliminators to strict propositions are called ``foo_sind``, in the
same way that eliminators to propositions are called ``foo_ind``.
Primitive records in :math:`\SProp` are allowed when fields are strict
propositions, for instance:
.. coqtop:: in
Set Primitive Projections.
Record sProd (A B : SProp) : SProp := { sfst : A; ssnd : B }.
On the other hand, to avoid having definitionally irrelevant types in
non-:math:`\SProp` sorts (through record η-extensionality), primitive
records in relevant sorts must have at least one relevant field.
.. coqtop:: all
Set Warnings "+non-primitive-record".
Fail Record rBox (A:SProp) : Prop := rbox { runbox : A }.
.. coqdoc::
Record ssig (A:Type) (P:A -> SProp) : Type := { spr1 : A; spr2 : P spr1 }.
Note that ``rBox`` works as an emulated record, which is equivalent to
the Box inductive.
Encodings for strict propositions
---------------------------------
The elimination for unit types can be encoded by a trivial function
thanks to proof irrelevance:
.. coqdoc::
Inductive sUnit : SProp := stt.
Definition sUnit_rect (P:sUnit->Type) (v:P stt) (x:sUnit) : P x := v.
By using empty and unit types as base values, we can encode other
strict propositions. For instance:
.. coqdoc::
Definition is_true (b:bool) : SProp := if b then sUnit else sEmpty.
Definition is_true_eq_true b : is_true b -> true = b
:= match b with
| true => fun _ => eq_refl
| false => sEmpty_ind _
end.
Definition eq_true_is_true b (H:true=b) : is_true b
:= match H in _ = x return is_true x with eq_refl => stt end.
Definitional UIP
----------------
.. flag:: Definitional UIP
This :term:`flag`, off by default, allows the declaration of non-squashed
inductive types with 1 constructor which takes no argument in
|SProp|. Since this includes equality types, it provides
definitional uniqueness of identity proofs.
Because squashing is a universe restriction, unsetting
:flag:`Universe Checking` is stronger than setting
:flag:`Definitional UIP`.
Definitional UIP involves a special reduction rule through which
reduction depends on conversion. Consider the following code:
.. coqtop:: in
Set Definitional UIP.
Inductive seq {A} (a:A) : A -> SProp :=
srefl : seq a a.
Axiom e : seq 0 0.
Definition hidden_arrow := match e return Set with srefl _ => nat -> nat end.
Check (fun (f : hidden_arrow) (x:nat) => (f : nat -> nat) x).
By the usual reduction rules :g:`hidden_arrow` is a stuck match, but
by proof irrelevance :g:`e` is convertible to :g:`srefl 0` and then by
congruence :g:`hidden_arrow` is convertible to `nat -> nat`.
The special reduction reduces any match on a type which uses
definitional UIP when the indices are convertible to those of the
constructor. For `seq`, this means a match on a value of type `seq x
y` reduces if and only if `x` and `y` are convertible.
Such matches are indicated in the printed representation by inserting
a cast around the discriminee:
.. coqtop:: out
Print hidden_arrow.
Non Termination with UIP
++++++++++++++++++++++++
The special reduction rule of UIP combined with an impredicative sort
breaks termination of reduction
:cite:`abel19:failur_normal_impred_type_theor`:
.. coqtop:: all
Axiom all_eq : forall (P Q:Prop), P -> Q -> seq P Q.
Definition transport (P Q:Prop) (x:P) (y:Q) : Q
:= match all_eq P Q x y with srefl _ => x end.
Definition top : Prop := forall P : Prop, P -> P.
Definition c : top :=
fun P p =>
transport
(top -> top)
P
(fun x : top => x (top -> top) (fun x => x) x)
p.
Fail Timeout 1 Eval lazy in c (top -> top) (fun x => x) c.
The term :g:`c (top -> top) (fun x => x) c` infinitely reduces to itself.
Debugging |SProp| issues
------------------------
Every binder in a term (such as `fun x` or `forall x`) caches
information called the :gdef:`relevance mark` indicating whether its type is
in |SProp| or not. This is used to efficiently implement proof
irrelevance.
The user should usually not be concerned with relevance marks, so by
default they are not displayed. However code outside the kernel may
generate incorrect marks resulting in bugs. Typically this means a
conversion will incorrectly fail as a variable was incorrectly marked
proof relevant.
.. warn:: Bad relevance
This is a developer warning, which is treated as an error by default. It is
emitted by the kernel when it is passed a term with incorrect relevance marks.
This is always caused by a bug in Coq (or a plugin), which should thus be reported and
fixed. In order to allow the user to work around such bugs, we leave the
ability to unset the ``bad-relevance`` warning for the time being, so that the
kernel will silently repair the proof term instead of failing.
.. flag:: Printing Relevance Marks
This :term:`flag` enables debug printing of relevance marks. It is off by default.
Note that :flag:`Printing All` does not affect printing of relevance marks.
.. coqtop:: all
Set Printing Relevance Marks.
Check fun x : nat => x.
Check fun (P:SProp) (p:P) => p.
|