\[\begin{split}\newcommand{\as}{\kw{as}} \newcommand{\Assum}[3]{\kw{Assum}(#1)(#2:#3)} \newcommand{\case}{\kw{case}} \newcommand{\cons}{\textsf{cons}} \newcommand{\consf}{\textsf{consf}} \newcommand{\Def}[4]{\kw{Def}(#1)(#2:=#3:#4)} \newcommand{\emptyf}{\textsf{emptyf}} \newcommand{\End}{\kw{End}} \newcommand{\kwend}{\kw{end}} \newcommand{\even}{\textsf{even}} \newcommand{\evenO}{\textsf{even}_\textsf{O}} \newcommand{\evenS}{\textsf{even}_\textsf{S}} \newcommand{\Fix}{\kw{Fix}} \newcommand{\fix}{\kw{fix}} \newcommand{\for}{\textsf{for}} \newcommand{\forest}{\textsf{forest}} \newcommand{\Functor}{\kw{Functor}} \newcommand{\In}{\kw{in}} \newcommand{\Ind}[4]{\kw{Ind}[#2](#3:=#4)} \newcommand{\ind}[3]{\kw{Ind}~[#1]\left(#2\mathrm{~:=~}#3\right)} \newcommand{\Indp}[5]{\kw{Ind}_{#5}(#1)[#2](#3:=#4)} \newcommand{\Indpstr}[6]{\kw{Ind}_{#5}(#1)[#2](#3:=#4)/{#6}} \newcommand{\injective}{\kw{injective}} \newcommand{\kw}[1]{\textsf{#1}} \newcommand{\length}{\textsf{length}} \newcommand{\letin}[3]{\kw{let}~#1:=#2~\kw{in}~#3} \newcommand{\List}{\textsf{list}} \newcommand{\lra}{\longrightarrow} \newcommand{\Match}{\kw{match}} \newcommand{\Mod}[3]{{\kw{Mod}}({#1}:{#2}\,\zeroone{:={#3}})} \newcommand{\ModA}[2]{{\kw{ModA}}({#1}=={#2})} \newcommand{\ModS}[2]{{\kw{Mod}}({#1}:{#2})} \newcommand{\ModType}[2]{{\kw{ModType}}({#1}:={#2})} \newcommand{\mto}{.\;} \newcommand{\nat}{\textsf{nat}} \newcommand{\Nil}{\textsf{nil}} \newcommand{\nilhl}{\textsf{nil\_hl}} \newcommand{\nO}{\textsf{O}} \newcommand{\node}{\textsf{node}} \newcommand{\nS}{\textsf{S}} \newcommand{\odd}{\textsf{odd}} \newcommand{\oddS}{\textsf{odd}_\textsf{S}} \newcommand{\ovl}[1]{\overline{#1}} \newcommand{\Pair}{\textsf{pair}} \newcommand{\plus}{\mathsf{plus}} \newcommand{\SProp}{\textsf{SProp}} \newcommand{\Prop}{\textsf{Prop}} \newcommand{\return}{\kw{return}} \newcommand{\Set}{\textsf{Set}} \newcommand{\Sort}{\mathcal{S}} \newcommand{\Str}{\textsf{Stream}} \newcommand{\Struct}{\kw{Struct}} \newcommand{\subst}[3]{#1\{#2/#3\}} \newcommand{\tl}{\textsf{tl}} \newcommand{\tree}{\textsf{tree}} \newcommand{\trii}{\triangleright_\iota} \newcommand{\Type}{\textsf{Type}} \newcommand{\WEV}[3]{\mbox{$#1[] \vdash #2 \lra #3$}} \newcommand{\WEVT}[3]{\mbox{$#1[] \vdash #2 \lra$}\\ \mbox{$ #3$}} \newcommand{\WF}[2]{{\mathcal{W\!F}}(#1)[#2]} \newcommand{\WFE}[1]{\WF{E}{#1}} \newcommand{\WFT}[2]{#1[] \vdash {\mathcal{W\!F}}(#2)} \newcommand{\WFTWOLINES}[2]{{\mathcal{W\!F}}\begin{array}{l}(#1)\\\mbox{}[{#2}]\end{array}} \newcommand{\with}{\kw{with}} \newcommand{\WS}[3]{#1[] \vdash #2 <: #3} \newcommand{\WSE}[2]{\WS{E}{#1}{#2}} \newcommand{\WT}[4]{#1[#2] \vdash #3 : #4} \newcommand{\WTE}[3]{\WT{E}{#1}{#2}{#3}} \newcommand{\WTEG}[2]{\WTE{\Gamma}{#1}{#2}} \newcommand{\WTM}[3]{\WT{#1}{}{#2}{#3}} \newcommand{\zeroone}[1]{[{#1}]} \end{split}\]

Basic notions and conventions¶

This section provides some essential notions and conventions for reading the manual.

We start by explaining the syntax and lexical conventions used in the manual. Then, we present the essential vocabulary necessary to read the rest of the manual. Other terms are defined throughout the manual. The reader may refer to the glossary index for a complete list of defined terms. Finally, we describe the various types of settings that Coq provides.

Syntax and lexical conventions¶

Syntax conventions¶

The syntax described in this documentation is equivalent to that accepted by the Coq parser, but the grammar has been edited to improve readability and presentation.

In the grammar presented in this manual, the terminal symbols are black (e.g. forall), whereas the nonterminals are green, italic and hyperlinked (e.g. term). Some syntax is represented graphically using the following kinds of blocks:

item?: An optional item.
item+: A list of one or more items.
item*: An optional list of items.
item+s: A list of one or more items separated by "s" (e.g. item₁ s item₂ s item₃).
item*s: An optional list of items separated by "s".
item₁item₂...: Alternatives (either item₁ or item₂ or ...).

Precedence levels that are implemented in the Coq parser are shown in the documentation by appending the level to the nonterminal name (as in term100 or ltac_expr3).

Note

Coq uses an extensible parser. Plugins and the notation system can extend the syntax at run time. Some notations are defined in the prelude, which is loaded by default. The documented grammar doesn't include these notations. Precedence levels not used by the base grammar are omitted from the documentation, even though they could still be populated by notations or plugins.

Furthermore, some parsing rules are only activated in certain contexts (interactive proof mode, custom entries...).

Warning

Given the complexity of these parsing rules, it would be extremely difficult to create an external program that can properly parse a Coq document. Therefore, tool writers are advised to delegate parsing to Coq, by communicating with it, for instance through SerAPI.

Lexical conventions¶

Blanks

Space, newline and horizontal tab are considered blanks. Blanks are ignored but they separate tokens.

Comments

Comments are enclosed between (* and *). They can be nested. They can contain any character. However, embedded string literals must be correctly closed. Comments are treated as blanks.

Identifiers

Identifiers, written ident, are sequences of letters, digits, _ and ', that do not start with a digit or '. That is, they are recognized by the following grammar (except that the string _ is reserved; it is not a valid identifier):

ident::=first_letter subsequent_letter*first_letter::=a .. zA .. Z_unicode_lettersubsequent_letter::=first_letterdigit'unicode_id_part

All characters are meaningful. In particular, identifiers are case-sensitive. unicode_letter non-exhaustively includes Latin, Greek, Gothic, Cyrillic, Arabic, Hebrew, Georgian, Hangul, Hiragana and Katakana characters, CJK ideographs, mathematical letter-like symbols and non-breaking space. unicode_id_part non-exhaustively includes symbols for prime letters and subscripts.

Numerals

Numerals are sequences of digits with an optional fractional part and exponent, optionally preceded by a minus sign. Hexadecimal numerals start with 0x or 0X. int is an integer; a numeral without fractional nor exponent parts. num is a non-negative integer. Underscores embedded in the digits are ignored, for example 1_000_000 is the same as 1000000.

numeral::=-? decnum . digit_+? eE +-? decnum?|-? hexnum . hexdigit_+? pP +-? decnum?int::=-? numnum::=decnumhexnumdecnum::=digit digit_*digit::=0 .. 9hexnum::=0x0X hexdigit hexdigit_*hexdigit::=0 .. 9a .. fA .. F

Strings

Strings begin and end with " (double quote). Use "" to represent a double quote character within a string. In the grammar, strings are identified with string.

Keywords

The following character sequences are keywords defined in the main Coq grammar that cannot be used as identifiers (even when starting Coq with the -noinit command-line flag):

_ Axiom CoFixpoint Definition Fixpoint Hypothesis Parameter Prop
SProp Set Theorem Type Variable as at cofix else end
fix for forall fun if in let match return then where with

The following are keywords defined in notations or plugins loaded in the prelude:

IF by exists exists2 using

Note that loading additional modules or plugins may expand the set of reserved keywords.

Other tokens

The following character sequences are tokens defined in the main Coq grammar (even when starting Coq with the -noinit command-line flag):

! #[ % & ' ( () ) * + , - ->
. .( .. ... / : ::= := :> :>> ; < <+ <- <:
<<: <= = => > >-> >= ? @ @{ [ ] _
`( `{ { {| | }

The following character sequences are tokens defined in notations or plugins loaded in the prelude:

** [= |- || ->

Note that loading additional modules or plugins may expand the set of defined tokens.

When multiple tokens match the beginning of a sequence of characters, the longest matching token is used. Occasionally you may need to insert spaces to separate tokens. For example, if ~ and ~~ are both defined as tokens, the inputs ~ ~ and ~~ generate different tokens, whereas if ~~ is not defined, then the two inputs are equivalent.

Essential vocabulary¶

This section presents the most essential notions to understand the rest of the Coq manual: terms and types on the one hand, commands and tactics on the other hand.

term

Terms are the basic expressions of Coq. Terms can represent mathematical expressions, propositions and proofs, but also executable programs and program types.

Here is the top-level syntax of terms. Each of the listed constructs is presented in a dedicated section. Some of these constructs (like term_forall_or_fun) are part of the core language that the kernel of Coq understands and are therefore described in this chapter, while others (like term_if) are language extensions that are presented in the next chapter.

Note

Many commands and tactics use one_term (in the syntax of their arguments) rather than term. The former need to be enclosed in parentheses unless they're very simple, such as a single identifier. This avoids confusing a space-separated list of terms or identifiers with a term_application.

type

To be valid and accepted by the Coq kernel, a term needs an associated type. We express this relationship by “$x$ of type $T$”, which we write as “$x:T$”. Informally, “$x:T$” can be thought as “$x$ belongs to $T$”.

The Coq kernel is a type checker: it verifies that a term has the expected type by applying a set of typing rules (see Typing rules). If that's indeed the case, we say that the term is well-typed.

A special feature of the Coq language is that types can depend on terms (we say that the language is dependently-typed). Because of this, types and terms share a common syntax. All types are terms, but not all terms are types:

type::=term

Intuitively, types may be viewed as sets containing terms. We say that a type is inhabited if it contains at least one term (i.e. if we can find a term which is associated with this type). We call such terms witnesses. Note that deciding whether a type is inhabited is undecidable.

Formally, types can be used to construct logical foundations for mathematics alternative to the standard "set theory": we call such logical foundations "type theories". Coq is based on the Calculus of Inductive Constructions, which is a particular instance of type theory.

sentence

Coq documents are made of a series of sentences that contain commands or tactics, generally terminated with a period and optionally decorated with attributes.

document::=sentence*sentence::=attributes? command .|attributes? num :? query_command .|attributes? toplevel_selector :? ltac_expr ....|control_command

ltac_expr syntax supports both simple and compound tactics. For example: split is a simple tactic while split; auto combines two simple tactics.

command

A command can be used to modify the state of a Coq document, for instance by declaring a new object, or to get information about the current state.

By convention, command names begin with uppercase letters. Commands appear in the HTML documentation in blue or gray boxes after the label "Command". In the pdf, they appear after the boldface label "Command:". Commands are listed in the Command index. Example:

Command Comments string*¶: This command prints "Comments ok" and does not change anything to the state of the document.

tactic

Tactics specify how to transform the current proof state as a step in creating a proof. They are syntactically valid only when Coq is in proof mode, such as after a Theorem command and before any subsequent proof-terminating command such as Qed. See Proof handling for more on proof mode.

By convention, tactic names begin with lowercase letters. Tactic appear in the HTML documentation in blue or gray boxes after the label "Tactic". In the pdf, they appear after the boldface label "Tactic:". Tactics are listed in the Tactic index.

Settings¶

There are several mechanisms for changing the behavior of Coq. The attribute mechanism is used to modify the behavior of a single sentence. The flag, option and table mechanisms are used to modify the behavior of Coq more globally in a document or project.

Attributes¶

An attribute modifies the behavior of a single sentence. Syntactically, most commands and tactics can be decorated with attributes (cf. sentence), but attributes not supported by the command or tactic will trigger This command does not support this attribute.

The order of top-level attributes doesn't affect their meaning. #[foo,bar], #[bar,foo], #[foo]#[bar] and #[bar]#[foo] are equivalent.

The legacy attributes (legacy_attr) provide an older, alternate syntax for certain attributes. They are equivalent to new attributes as follows:

Legacy attribute	New attribute
`Local`	`local`
`Global`	`global`
`Polymorphic`	`universes(polymorphic)`
`Monomorphic`	`universes(monomorphic)`
`Cumulative`	`universes(cumulative)`
`NonCumulative`	`universes(noncumulative)`
`Private`	`private(matching)`
`Program`	`program`

Attributes appear in the HTML documentation in blue or gray boxes after the label "Attribute". In the pdf, they appear after the boldface label "Attribute:". Attributes are listed in the Attribute index.

Warning This command does not support this attribute: ident.¶

This warning is configured to behave as an error by default. You may turn it into a normal warning by using the Warnings option:

Set Silent.

Set Warnings "unsupported-attributes".
#[ foo ] Comments.: Toplevel input, characters 0-18: > #[ foo ] Comments. > ^^^^^^^^^^^^^^^^^^ Warning: This command does not support this attribute: foo. [unsupported-attributes,parsing]

Flags, Options and Tables¶

The following types of settings can be used to change the behavior of Coq in subsequent commands and tactics (see Locality attributes supported by Set and Unset for a more precise description of the scope of these settings):

A flag has a boolean value, such as Universe Polymorphism.
An option generally has a numeric or string value, such as Firstorder Depth.
A table contains a set of strings or qualids.
In addition, some commands provide settings, such as Extraction Language.

setting_name::=ident+

Flags, options and tables are identified by a series of identifiers. By convention, each of the identifiers start with an initial capital letter.

Flags, options and tables appear in the HTML documentation in blue or gray boxes after the labels "Flag", "Option" and "Table". In the pdf, they appear after a boldface label. They are listed in the Flags, options and tables index.

Command Set setting_name intstring?¶

If setting_name is a flag, no value may be provided; the flag is set to on. If setting_name is an option, a value of the appropriate type must be provided; the option is set to the specified value.

This command supports the local, global and export attributes. They are described here.

Warning There is no flag or option with this name: "setting_name".¶: This warning message can be raised by Set and Unset when setting_name is unknown. It is a warning rather than an error because this helps library authors produce Coq code that is compatible with several Coq versions. To preserve the same behavior, they may need to set some compatibility flags or options that did not exist in previous Coq versions.

Command Unset setting_name¶

If setting_name is a flag, it is set to off. If setting_name is an option, it is set to its default value.

This command supports the local, global and export attributes. They are described here.

Command Add setting_name qualidstring+¶: Adds the specified values to the table setting_name.

Command Remove setting_name qualidstring+¶: Removes the specified value from the table setting_name.

Command Test setting_name for qualidstring+?¶

If setting_name is a flag or option, prints its current value. If setting_name is a table: if the for clause is specified, reports whether the table contains each specified value, otherwise this is equivalent to Print Table. The for clause is not valid for flags and options.

Error There is no flag, option or table with this name: "setting_name".¶: This error message is raised when calling the Test command (without the for clause), or the Print Table command, for an unknown setting_name.

Error There is no qualid-valued table with this name: "setting_name".¶
Error There is no string-valued table with this name: "setting_name".¶: These error messages are raised when calling the Add or Remove commands, or the Test command with the for clause, if setting_name is unknown or does not have the right type.

Command Print Options¶: Prints the current value of all flags and options, and the names of all tables.

Command Print Table setting_name¶: Prints the values in the table setting_name.

Command Print Tables¶: A synonym for Print Options.

Locality attributes supported by `Set` and `Unset`¶

The Set and Unset commands support the local, global and export locality attributes:

no attribute: the original setting is not restored at the end of the current module or section.
local (or alternatively, the Local prefix): the setting is applied within the current module or section. The original value of the setting is restored at the end of the current module or section.
export (or alternatively, the Export prefix): similar to local, the original value of the setting is restored at the end of the current module or section. In addition, if the value is set in a module, then Import-ing the module sets the option or flag.
global (or alternatively, the Global prefix): the original setting is not restored at the end of the current module or section. In addition, if the value is set in a file, then Require-ing the file sets the option.

Newly opened modules and sections inherit the current settings.

Note

We discourage using the global attribute with the Set and Unset commands. If your goal is to define project-wide settings, you should rather use the command-line arguments -set and -unset for setting flags and options (see By command line options).

Basic notions and conventions¶

Syntax and lexical conventions¶

Syntax conventions¶

Lexical conventions¶

Essential vocabulary¶

Settings¶

Attributes¶

Flags, Options and Tables¶

Locality attributes supported by Set and Unset¶

Locality attributes supported by `Set` and `Unset`¶