W3C

Speech Recognition Grammar Specification Version 1.0

W3C Recommendation 16 March 2004

This version:
http://www.w3.org/TR/2004/REC-speech-grammar-20040316/
Latest version:
http://www.w3.org/TR/speech-grammar/
Previous version:
http://www.w3.org/TR/2003/PR-speech-grammar-20031218/
Editors:
Andrew Hunt, ScanSoft
Scott McGlashan, Hewlett-Packard
Contributors:
See Acknowledgements

Please refer to the errata for this document, which may include some normative corrections.

See also translations.

Copyright © 2004 W3C® (MIT, ERCIM, Keio), All Rights Reserved. W3C liability, trademark, document use and software licensing rules apply.

Abstract

This document defines syntax for representing grammars for use in speech recognition so that developers can specify the words and patterns of words to be listened for by a speech recognizer. The syntax of the grammar format is presented in two forms, an Augmented BNF Form and an XML Form. The specification makes the two representations mappable to allow automatic transformations between the two forms.

Status of this Document

This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the W3C technical reports index at http://www.w3.org/TR/.

This document has been reviewed by W3C Members and other interested parties, and it has been endorsed by the Director as a W3C Recommendation. W3C's role in making the Recommendation is to draw attention to the specification and to promote its widespread deployment. This enhances the functionaility and interoperability of the Web.

This specification is part of the W3C Speech Interface Framework and has been developed within the W3C Voice Browser Activity (activity statement) by participants in the Voice Browser Working Group (W3C Members only).

The design of SRGS 1.0 has been widely reviewed (see the disposition of comments) and satisfies the Working Group's technical requirements. A list of implementations is included in the SRGS 1.0 implementation report, along with the associated test suite.

Comments are welcome on www-voice@w3.org (archive). See W3C mailing list and archive usage guidelines.

The W3C maintains a list of any patent disclosures related to this work.

Table of Contents

1. Introduction

This document defines the syntax for grammar representation. The grammars are intended for use by speech recognizers and other grammar processors so that developers can specify the words and patterns of words to be listened for by a speech recognizer.

The syntax of the grammar format is presented in two forms, an Augmented BNF (ABNF) Form and an XML Form. The specification ensures that the two representations are semantically mappable to allow automatic transformations between the two forms.

Both the ABNF Form and XML Form have the expressive power of a Context-Free Grammar (CFG). A grammar processor that does not support recursive grammars has the expressive power of a Finite State Machine (FSM) or regular expression language. For definitions of CFG, FSM, regular expressions and other formal computational language theory see, for example, [HU79]. This form of language expression is sufficient for the vast majority of speech recognition applications.

This W3C standard is known as the Speech Recognition Grammar Specification and is modelled on the JSpeech Grammar Format specification [JSGF], which is owned by Sun Microsystems, Inc., California, U.S.A.

1.1 Grammar Processors and User Agents

A grammar processor is any entity that accepts as input grammars as described in this specification.

A user agent is a grammar processor that accepts user input and matches that input against a grammar to produce a recognition result that represents the detected input.

As the specification title implies, speech recognizers are an important class of grammar processor. Another class of grammar processor anticipated by this specification is a Dual-Tone Multi-Frequency (DTMF) detector. The type of input accepted by a user agent is determined by the mode or modes of grammars it can process: e.g. speech input for "voice" mode grammars and DTMF input for "dtmf" mode grammars.

For simplicity, throughout this document references to a speech recognizer apply to other types of grammar processor unless explicitly stated otherwise.

A speech recognizer is a user agent with the following inputs and outputs:

1.2 Scope

The primary use of a speech recognizer grammar is to permit a speech application to indicate to a recognizer what it should listen for, specifically:

Speech recognizers may also support the Stochastic Language Models (N-Gram) Specification [NGRAM]. Both specifications define ways to set up a speech recognizer to detect spoken input but define the word and patterns of words by different and complementary means. Some recognizers permit cross-references between grammars in the two formats. The rule reference element of this specification describes how to reference an N-gram document.

The grammar specification does not address a number of other issues that affect speech recognition performance. Most of the following capabilities are addressed by the context in which a grammar is referenced or invoked: for example, through VoiceXML 2.0 [VXML2] or through a speech recognizer API.

1.3 Grammar Conversions

The ABNF Form and XML Form are specified to ensure that the two representations are semantically mappable. It should be possible to automatically convert an ABNF Form grammar to an XML Form grammar (or the reverse) so that the semantic performance of the grammars are identical. Equivalence of semantic performance implies that:

  1. Both grammars accept the same language as input and reject the same language as input
  2. Both grammars parse any input string identically

The XSL Transformation document in Appendix F demonstrates automatic conversion from XML to ABNF. The reverse conversion requires an ABNF parser and a transformational program.

There are inherent limits to the automatic conversion to and From ABNF Form and XML Form.

1.4 Semantic Interpretation

A speech recognizer is capable of matching audio input against a grammar to produce a raw text transcription (also known as literal text) of the detected input. A recognizer may be capable of, but is not required to, perform subsequent processing of the raw text to produce a semantic interpretation of the input.

For example, the natural language utterance "I want to book a flight from Prague to Paris" could result in the following XML data structure. To perform this additional interpretation step requires semantic processing instructions that may be contained within a grammar that defines the legal spoken input or in an associated document.

  <book-flight>
    <depart>Prague</depart>
    <arrive>Paris</arrive>
  </book-flight>

The Speech Recognition Grammar Specification provides syntactic support for limited semantic interpretation. The tag construct and the tag-format and tag declarations provide a placeholder for instructions to a semantic processor.

The W3C Voice Browser Working Group is presently developing the Semantic Interpretation for Speech Recognition specification [SEM]. That specification defines a language that can be embedded in tags within SRGS grammars to perform the interpretation process. The semantic processing is defined with respect to the logical parse structure for grammar processing (see Appendix H). Other tag formats could be used but are outside the scope of the W3C activities.

For examples of semantic interpretation in the latest working draft see [SEM].

The output of the semantic interpretation processor may be represented using the Natural Language Semantics Markup Language [NLSML]. This XML representation of interpreted spoken input can be used to transmit the result, as input to VoiceXML 2.0 [VXML2] processing or in other ways.

The semantic interpretation carried out in the speech recognition process is typically characterized by:

It is this restricted form of semantic interpretation that this approach is intended to support. A VoiceXML application that receives a speech result with semantic interpretation will typically process the user input to carry out a dialog. The application may also perform deeper semantic analysis, for example resolving deictic or anaphoric references.

1.5 Embedded Grammars

The Speech Recognition Grammar Specification is designed to permit ABNF Form and XML Form grammars to be embedded into other documents. For example, VoiceXML 1.0 [VXML1] and VoiceXML 2.0 [VXML2] permit inline grammars [VXML2 §3.1.1.1] in which an ABNF Form grammar or XML Form grammar is contained within a VoiceXML document.

Embedding an XML Form grammar within an XML document can be achieved with XML namespaces [XMLNS] or by incorporating the grammar XML Schema definition or DTD into to enclosing document's schema or DTD.

An ABNF Form grammar may be embedded into any XML document as character data. ABNF grammars will often contain angle brackets which require special handling within XML. A CDATA section [XML §2.7] or the escape sequences of "&lt;" and "&gt;" may be required to create well-formed XML. Note: angle brackets ('<' and '>') are used in ABNF to delimit any URI, media type or repeat operator.

1.6 Terminology

Requirements terms
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119]. However, for readability, these words do not appear in all uppercase letters in this specification.

URI: Uniform Resource Identifier
A URI is a unifying syntax for the expression of names and addresses of objects on the network as used in the World Wide Web. A URI is defined as any legal 'anyURI ' primitive as defined in XML Schema Part 2: Datatypes [SCHEMA2 §3.2.17]. The XML Schema definition follows [RFC2396] and [RFC2732]. The syntax representation of a URI differs between the ABNF Form and the XML Form. Any relative URI reference must be resolved according to the rules given in Section 4.9.1.
  • ABNF URI: in the ABNF Form of this specification a URI is delimited by angle brackets ('<' '>'). For example, <http://www.example.com/file-path>
  • XML URI: in the XML Form of this specification any URI is provided as an attribute to an element; for example the ruleref and lexicon elements.

Media Type
A media type (defined in [RFC2045] and [RFC2046]) specifies the nature of a linked resource. Media types are case insensitive. A list of registered media types is available for download [TYPES]. In places where a URI can be specified a media type may be provided to indicate the content type of URI.

[See Appendix G for information on media types for the ABNF and XML Forms of the Speech Recognition Grammar Specification.]

  • ABNF URI with Media Type: in the ABNF Form a media type may be attached as a postfix to any URI. The media type is delimited by angle brackets ('<' '>') and the URI and media type are separated by a tilde character ('~') without intervening white space. For example,
    <http://example.com/file-path>~<media-type>
  • XML URI with Media Type: in the XML Form any element that carries a URI attribute may carry a type attribute.

Language identifier
A language identifier labels information content as being of a particular human language variant. Following the XML specification for language identification [XML §2.12] a legal language identifier in ABNF Form grammars and XML Form grammars is identified by an RFC 3066 [RFC3066] code. A language code is required by RFC 3066. A country code or other subtag identifier is optional by RFC 3066. A grammar's language declaration declares the language of a grammar. Additionally a legal rule expansion may be labeled by its language content.

White space
White space consists of one or more space (#x20) characters, carriage returns, line feeds, or tabs. The normative definition for both ABNF and XML follows the XML white space definition [XML §2.3]. ABNF processors must also follow the end-of-line handling of XML 1.0 [XML §2.11].

DTMF
DTMF (Dual Tone Multiple Frequency) is an ITU standard for telephony signaling. ITU Recommendation Q.23 defines DTMF generation. ITU Recommendation Q.24 [Q24] defines DTMF reception. A grammar processor that accepts DTMF input should implement Q.24.

2. Rule Expansions

A legal rule expansion is any legal token, rule reference, tag, or any logical combination of legal rule expansions as sequence, alternatives, repeated expansion or language-attributed expansion.

A rule expansion is formally a regular expression (see, for example, [HU79]).

A rule definition associates a legal rule expansion with a rulename.

2.1 Tokens

A token (a.k.a. a terminal symbol) is the part of a grammar that defines words or other entities that may be spoken. Any legal token is a legal expansion.

For speech recognition, a token is typically an orthographic entity of the language being recognized. However, a token may be any string that the speech recognizer can convert to a phonetic representation.

Token Content: In both the XML Form and ABNF Form any unmarked text within a rule definition, except example phrases (XML only) or tag content, is token content. The unmarked text is delimited by any syntactic construct of the grammar form (see below for details on the ABNF Form and XML Form). For each token content span in a grammar the grammar processor applies the following tokenization, white space normalization, token normalization and pronunciation lookup processes. All token content in both the XML Form and ABNF Form is treated as Characters in [XML]. (informative: XML specifies Characters by reference to ISO/IEC 10646 [ISO/IEC 10646] and Unicode [Unicode].)

Tokenization behavior: Text spans containing token sequences are delimited as follows:

Token type Form Example
Single unquoted token ABNF & XML hello
Single unquoted token: non-alphabetic ABNF & XML 2
Single quoted token: including white space ABNF & XML "San Francisco"
Single quoted token: no white space ABNF & XML "hello"
Two tokens delimited by white space ABNF & XML bon voyage
Four tokens delimited by white space ABNF & XML this is a test
Single XML token in <token> XML Only <token>San Francisco</token>

White Space Normalization: White space must be normalized when contained in any token delimited by a <token> elements or by double quotes. Leading and trailing white space characters are stripped. Any token-internal white space character or sequence is collapsed to a single space character (#x20). For example, the following are all normalized to the same string, "San Francisco".

  "San Francisco"
  " San Francisco "
  "San 
  Francisco"
  " San   Francisco "

Because the presence of white space within a token is significant the following are distinct tokens.

  "San Francisco"
  "SanFrancisco"
  "San_Francisco"

Token Normalization: Other normalization processes are applied to the white space normalized token according to the language and the capabilities of the speech recognizer.

Grammar processors may assume Early Uniform Normalization as defined in the Character Model for the World Wide Web 1.0 [CHARMOD §4].

Pronunciation Lookup: To match spoken (audio) input to a grammar a speech recognition must be capable of modelling the audio patterns of any token in a grammar. Speech recognizers employ a diverse set of techniques for performing this key recognition process. The following is an informative description of techniques that a speech recognizer may apply based on conventional large vocabulary speech recognition technology.

A large vocabulary speech recognizer converts each normalized token to a phoneme sequence or a set of possible phoneme sequences. Conversion of an orthographic form (token) to the spoken form (phonemes) is a highly language-specific process. In many cases the conversion is even specific to a national variant, regional dialect or other variant of the language. For example, for some tokens Parisian French, Quebec French and Swiss French will each convert to different pronunciations.

The text-to-phoneme conversion in a large vocabulary speech recognizer may involve some or all of the following sub-processes.

Any language is likely to have other specialized processes for determining a pronunciation for a token. For example, for Japanese special techniques are required for Kanji and each Kana form.

For any language and recognizer there may be variation in coverage and completeness of the language's tokens.

When a grammar processor handles a grammar containing a token that it cannot convert to phonemic form or otherwise use in the speech recognition processing of audio it should inform the hosting environment.

Limitations of token handling: the following is informative guidance to grammar developers.

The Pronunciation Lexicon activity [LEX] of the W3C Voice Browser Working Group will provide guidance on the token-handling processes outlined above.

Token handling will vary between recognizers and will vary between languages.

Grammar authors can improve document portability by avoiding characters and forms in tokens that do not have obvious pronunciations in the language. For English, the following are ways to handle some orthographic forms:

ABNF Form

Any plain text within a rule definition is token content. The ABNF Syntax (Appendix D) normatively defines the token parsing behavior.

A language attachment may be provided for any token. When attached to a token the language modifies the handling of that token only.

Informative

The rule expansion of a rule definition is delimited at the start and end by equals sign ('=') and semicolon (';') respectively. Any leading plain text of the rule expansion is delimited by ('=') and similarly any final plain text is closed by semicolon.

Within a rule expansion the following symbols have syntactic function and delimit plain text.

Within plain text regions delimited by these characters the tokenization, white space normalization, token normalization and pronunciation lookup processes described above apply.

XML Form

Any token element explicitly delimits a single token as described above. The token element may include an optional xml:lang attribute to indicate the language of the contained token.

Any other character data within a rule element (rule definition) or item element is token content. Note that character data within tag or example is not token text.

2.2 Rule Reference

Any legal rule reference is a legal rule expansion .

Rulenames: Every rule definition has a local name that must be unique within the scope of the grammar in which it is defined. A rulename must match the "Name" Production of XML 1.0 [XML §2.3] and be a legal XML ID. Section 3.1 documents the rule definition mechanism and the legal naming of rules.

This table summarizes the various forms of rule reference that are possible within and across grammar documents.

Note: an XML Form grammar document must provide one and only one of the uri or special attributes on a ruleref element. There is no equivalent constraint in ABNF since the syntactic forms are distinct.

Reference type ABNF Form XML Form
2.2.1: Explicit local rule reference $rulename <ruleref uri="#rulename"/>
2.2.2: Explicit reference to a named rule of a grammar identified by a URI $<grammarURI#rulename> <ruleref uri="grammarURI#rulename"/>
2.2.2: Implicit reference to the root rule of a grammar identified by a URI $<grammarURI> <ruleref uri="grammarURI"/>
2.2.2: Explicit reference to a named rule of a grammar identified by a URI with a media type $<grammarURI#rulename>~<media-type> <ruleref uri="grammarURI#rulename" type="media-type"/>
2.2.2: Implicit reference to the root rule of a grammar identified by a URI with a media type $<grammarURI>~<media-type> <ruleref uri="grammarURI" type="media-type"/>
2.2.3: Special rule definitions $NULL
$VOID
$GARBAGE		 <ruleref special="NULL"/>
<ruleref special="VOID"/>
<ruleref special="GARBAGE"/>

2.2.1 Local References

When referencing rules defined locally (defined in the same grammar as contains the reference), always use a simple rulename reference which consists of the local rulename only. The ABNF Form and XML Form have a different syntax for representing a simple rulename reference.

ABNF Form

The simple rulename reference is prefixed by a "$" character.

$city
$digit

XML Form

The ruleref element is an empty element with a uri attribute that specifies the rule reference as a same-document reference URI [RFC2396]: that is, the attribute consists only of the number sign ('#') and the fragment identifier that indicates the locally referenced rulename.

<ruleref uri="#city"/>
<ruleref uri="#digit"/>

2.2.2 External Reference by URI

References to rules defined in other grammars are legal under the conditions defined in Section 3. The external reference must identify the external grammar by URI and may identify a specific rule within that grammar. If the fragment identifier that would indicate a rulename is omitted, then the reference implicitly targets the root rule of the external grammar.

Any externally-referenced rule may be activated for recognition. That is it may define the top-level syntax of spoken input. For instance, VoiceXML [VXML2] grammar activation may explicitly reference one or more public rules (see Section 3.2) and/or implicitly reference the root rule (see Section 4.7).

A URI reference is illegal if the referring document and referenced document have different modes. For instance, it is illegal to reference a "dtmf" grammar from a "voice" grammar. (See Section 4.6 for additional detail on modes).

A resource indicated by an URI reference may be available in one or more media types. The grammar author may specify the preferred media-type via the type attribute (XML form) or in angle braces following the URI (ABNF form).When the content represented by a URI is available in many data formats, a grammar processor may use the preferred media-type to influence which of the multiple formats is used. For instance, on a server implementing HTTP content negotiation, the processor may use the preferred media-type to order the preferences in the negotiation.

The resource representation delivered by dereferencing the URI reference may be considered in terms of two types. The declared media-type is the asserted value for the resource and the actual media-type is the true format of its content. The actual media-type should be the same as the declared media-type, but this is not always the case (e.g. a misconfigured HTTP server might return text/plain for an application/srgs+xml document). A specific URI scheme may require that the resource owner always, sometimes, or never return a media-type. The declared media-type is the value returned by the resource owner or, if none is returned, the preferred media type given in the grammar. There may be no declared media-type if the resource owner does not return a value and no preferred type is specified. Whenever specified, the declared media-type is authoritative.

Three special cases may arise. The declared media-type may not be supported by the processor; this is an error. The declared media-type may be supported but the actual media-type may not match; this is also an error. Finally, there may be no declared media-type; the behavior depends on the specific URI scheme and the capabilities of the grammar processor. For instance, HTTP 1.1 allows document introspection (see RFC 2616, section 7.2.1), the data scheme falls back to a default media type, and local file access defines no guidelines. The following table provides some informative examples:

HTTP 1.1 request
Local file access
Media-type returned by the resource owner text/plain application/srgs+xml <none> <none>
Preferred media-type appearing in the grammar Not applicable; the returned type takes precedence application/srgs+xml <none>
Declared media-type text/plain application/srgs+xml application/srgs+xml <none>
Behavior if the actual media-type is application/srgs+xml Error; the declared and actual types do not match The declared and actual types match; success if application/srgs+xml is supported by the grammar processor; otherwise an error Scheme specific; the grammar processor might introspect the document to determine the type.

See Appendix G for a summary of the status for media types for ABNF Form and XML Form grammars.

ABNF Form

In ABNF an external reference by URI is represented by a dollar sign ('$') followed immediately by either an ABNF URI or ABNF URI with media type. There must be no white space between the dollar sign and the URI.

// References to specific rules of an external grammar
$<http://grammar.example.com/world-cities.gram#canada>
$<http://grammar.example.com/numbers.gram#digit>

// Implicit reference to the root rule of an external grammar
$<../date.gram>

// References with associated media types
$<http://grammar.example.com/world-cities.gram#canada>~<application/srgs>
$<../date.gram>~<application/srgs>

Note: the media type of "application/srgs" has been requested for ABNF Form grammars. See Appendix G for details.

XML Form

An XML rule reference is represented by a ruleref element with a uri attribute that defines the URI of the referenced grammar and rule within it. If a fragment identifier is appended then the identifier indicates a specific rulename being referenced. If the fragment identifier is omitted then the reference is (implicitly) to the root rule of the referenced grammar.

The optional type attribute specifies the media type of the grammar containing the reference.

<!-- References to specific rules of an external grammar -->
<ruleref uri="http://grammar.example.com/world-cities.grxml#canada"/>
<ruleref uri="http://grammar.example.com/numbers.grxml#digit"/>

<!-- Implicit reference to the root rule of an external grammar -->
<ruleref uri="../date.grxml"/>

<!-- References with associated media types -->
<ruleref uri="http://grammar.example.com/world-cities.grxml#canada"
         type="application/srgs+xml"/>
<ruleref uri="../date.grxml" type="application/srgs+xml"/>

Note: the media type "application/srgs+xml" has been requested for XML Form grammars. See Appendix G for details on media types for grammars.

2.2.3 Special Rules

Several rulenames are defined to have specific interpretation and processing by a speech recognizer. A grammar must not redefine these rulenames.

In the ABNF Form a special rule reference is syntactically identical to a local rule reference. However, the names of the special rules are reserved to prevent a rule definition with the same name.

In the XML Form a special rulename is represented with the special attribute on a ruleref element. It is illegal to provide both the special and the uri attributes.

NULL
Defines a rule that is automatically matched: that is, matched without the user speaking any word.

ABNF Form: $NULL
XML Form: <ruleref special="NULL"/>

VOID
Defines a rule that can never be spoken. Inserting VOID into a sequence automatically makes that sequence unspeakable.

ABNF Form: $VOID
XML Form: <ruleref special="VOID"/>

GARBAGE
Defines a rule that may match any speech up until the next rule match, the next token or until the end of spoken input. A grammar processor must accept grammars that contain special references to GARBAGE. The behavior GARBAGE rule is implementation-specific. A user agent should be capable of matching arbitrary spoken input up to the next token but may treat GARBAGE as equivalent to NULL (match no spoken input).

ABNF Form: $GARBAGE
XML Form: <ruleref special="GARBAGE"/>

Informative example: given suitable definitions of US cities and states, a speech recognizer may implement the following ABNF and XML rule definitions to match "Philadelphia in the great state of Pennsylvania" as well as simply "Philadelphia Pennsylvania". 
$location = $city $GARBAGE $state;

<rule id="location">
  <ruleref uri="#city"/>
  <ruleref special="GARBAGE"/>
  <ruleref uri="#state"/>
</rule>

2.2.4 Referencing N-gram Documents (Informative)

The W3C Voice Browser Working Group has released a Working Draft for the Stochastic Language Models (N-Gram) Specification [NGRAM]. These two specifications represent different and complementary ways of informing a speech recognizer of which words and patterns of words to listen for.

A speech recognizer may choose to support the Speech Recognition N-Gram Grammar Specification in addition to the speech recognition grammar defined in this document.

If a speech recognizer supports both grammar representations it may optionally support references between the two formats. Grammars defined in the ABNF Form or XML Form may reference start symbols of N-Gram documents and vice versa.

The syntax for referencing an N-Gram is the same as referencing externally defined ABNF Form or XML Form grammar documents. A media type is recommended on a reference to an N-gram document. The Working Group has not yet applied for a type on N-gram documents so no example is given. The fragment identifier (a rulename when referencing ABNF Form and XML Form grammars) identifies a start symbol as defined by the N-Gram specification. If the start symbol is absent the N-Gram, as a whole, is referenced as defined in the N-Gram specification.

ABNF Form

URI references to N-Gram documents follow the same syntax as references to other ABNF or XML Form grammar documents. The following are examples of references to an N-Gram document via an explicit rule reference and an implicit reference to the root rule.

$<http://grammar.example.com/ngram.xml#StartSymbol>
$<http://grammar.example.com/ngram.xml>

XML Form

URI references to N-Gram documents follow the same syntax as reference to other ABNF Form and XML Form grammar documents. The following are examples of references to an N-Gram document via an explicit rule reference and an implicit reference to the root rule.

<ruleref uri="http://grammar.example.com/ngram.xml#StartSymbol"/>
<ruleref uri="http://grammar.example.com/ngram.xml"/>

2.3 Sequences and Encapsulation

A sequence of legal rule expansions is itself a legal rule expansion.

The sequence of rule expansions implies the temporal order in which the expansions must be detected by the user agent. This constraint applies to sequences of tokens, sequences of rule references, sequences of tags, parentheticals and all combinations of these rule expansions.

Both the ABNF Form and XML Form provide syntax for encapsulating any expansion. This is used, for example, to attach a repeat operator, a language identifier or to ensure correct precedence in parsing (ABNF only).

ABNF Form

A sequence of legal expansions separated by white space is a legal expansion.

A legal expansion surrounded by parentheses ('(' and ')') is a legal expansion.

this is a test           // sequence of tokens
$action $object          // sequence of rule references
the $object is $color    // sequence of tokens and rule references
(fly to $city)           // parentheses for encapsulation
Special cases

An empty parenthetical is legal as is a parenthetical containing only white space; e.g. '()' or '( )'. Both forms are equivalent to $NULL and a grammar processor will behave as if the parenthetical were not present.

// equivalent sequences
phone home
phone ( ) home

XML Form

A sequence of XML rule expansion elements ( <ruleref>, <item>, <one-of>, <token> <tag>) and CDATA sections containing space separated tokens must be recognized in temporal sequence. (The only exception is where one or more "item" elements appear within a one-of element.)

An item element can surround any expansion to permit a repeat attribute or language identifier to be attached. The weight attribute of item is ignored unless the element appears within a one-of element.

<!-- sequence of tokens -->
this is a test

<!--sequence of rule references-->
<ruleref uri="#action"/> <ruleref uri="#object"/>

<!--sequence of tokens and rule references-->
the <ruleref uri="#object"/> is <ruleref uri="#color"/>

<!-- sequence container -->
<item>fly to <ruleref uri="#city"/> </item>
Special cases

An empty item element is legal as is an item element containing only white space. Both forms are equivalent to a NULL reference and a grammar processor will behave as if the item were not present.

<!-- equivalent sequences -->
phone home
phone <item/> home
phone <item></item> home
phone <item>    </item> home

2.4 Alternatives

Any set of alternative legal rule expansions is itself a legal rule expansion. For input to match a set of alternative rule expansions it must match one of the set of alternative expansions. A set of alternatives must contain one or more alternatives.

Any set of alternatives may be labeled with a language attachment. In the XML Form an xml:lang attribute is present on the one-of element. In the ABNF Form to ensure correct precedence the set of alternatives must be delimited by parentheses with the ABNF language attachment immediately following.

2.4.1 Weights

A weight may be optionally provided for any number of alternatives in an alternative expansion. Weights are simple positive floating point values without exponentials. Legal formats are "n", "n.", ".n" and "n.n" where "n" is a sequence of one or many digits.

A weight is nominally a multiplying factor in the likelihood domain of a speech recognition search. A weight of 1.0 is equivalent to providing no weight at all. A weight greater than "1.0" positively biases the alternative and a weight less than "1.0" negatively biases the alternative.

[JEL98] and [RAB93] are informative references on the topic of speech recognition technology and the underlying statistical framework within which weights are applied.

Grammar authors and speech recognizer developers should be aware of the following limitations upon the definition and application of weights as outlined above.

ABNF Form

A set of alternative choices is identified as a list of legal expansions separated by the vertical bar symbol. If necessary, the set of alternative choices may be delimited by parentheses.

Michael | Yuriko | Mary | Duke | $otherNames
(1 | 2 | 3)

A weight is surrounded by forward slashes and placed before each item in the alternatives list.

/10/ small | /2/ medium |  large
/3.1415/ pie | /1.414/ root beer | /.25/ cola
Special Cases

It is legal for an alternative to be a reference to $NULL, an empty parenthetical or a single tag. In each case the input is equivalent to matching $NULL and as a result the other alternatives are optional.

// Legal
$rule1 = word | $NULL;
$rule2 = () | word;
$rule3 = word | {TAG-CONTENT};

An empty alternative (white space only) is not legal.

// ILLEGAL
$rule1 = a | | b;
$rule2 = | b;
$rule3 = a |;

The following construct is interpreted as a single weighted alternative.

// Legal
$rule1 = /2/ word;
$rule2 = /2/ {TAG-CONTENT};
$rule3 = /2/ $NULL;

XML Form

The one-of element identifies a set of alternative elements. Each alternative expansion is contained in a item element. There must be at least one item element contained within a one-of element. Weights are optionally indicated by the weight attribute on the item element.

<one-of>
  <item>Michael</item>
  <item>Yuriko</item>
  <item>Mary</item>
  <item>Duke</item>
  <item><ruleref uri="#otherNames"/></item>
</one-of>

<one-of><item>1</item> <item>2</item> <item>3</item></one-of>

<one-of>
  <item weight="10">small</item>
  <item weight="2">medium</item>
  <item>large</item>
</one-of>

<one-of>
  <item weight="3.1415">pie</item>
  <item weight="1.414">root beer</item>
  <item weight=".25">cola</item>
</one-of>
Special cases

A one-of element containing a single item is legal and requires that input match the single item. The single item may be optionally weighted.

<one-of>
  <item>word</item>
</one-of>

<one-of>
  <item weight="2.0">word</item>
</one-of>

Is it legal for an alternative to be a reference to NULL, an empty item or a single tag. In each case the input is equivalent to matching NULL and as a result the other alternatives are optional.

<one-of>
  <item>word</item>
  <item/>
</one-of>
<one-of>
  <item>word</item>
  
  <item> <ruleref special="NULL"/> </item>
</one-of>
<one-of>
  <item>word</item>
  <item> <tag>TAG-CONTENT</tag> </item>
</one-of>

2.5 Repeats

Any repeated legal rule expansion is itself a legal rule expansion.

Operators are provided that define a legal rule expansion as being another sub-expansion that is optional, that is repeated zero or more times, that is repeated one or more times, or that is repeated some range of times.

ABNF Form
Example		 XML Form
Example		 Behavior
<n>
<6>		 repeat="n"
repeat="6"		 The contained expansion is repeated exactly "n" times. "n" must be "0" or a positive integer.
<m-n>
<4-6>		 repeat="m-n"
repeat="4-6"		 The contained expansion is repeated between "m" and "n" times (inclusive). "m" and "n" must both be "0" or a positive integer and "m" must be less than or equal to "n".
<m->
<3->		 repeat="m-"
repeat="3-"		 The contained expansion is repeated "m" times or more (inclusive). "m" must be "0" or a positive integer. For example, "3-" declares that the contained expansion can occur three, four, five or more times.
<0-1>
[...]		 repeat="0-1" The contained expansion is optional.
Common Repeats

As indicated in the table above, an expansion that can occur 0-1 times is optional. Because optionality is such a common form the ABNF syntax provides square brackets as a special operator for representing optionality.

A repeat of "0-" indicates that an expansion can occur zero times, once or any number of multiple times. In regular expression languages this is often represented by the Kleene star ('*') which is reserved but not used in ABNF.

A repeat of "1-" indicates that an expansion can occur once or any number of multiple times. In regular expression languages this is often represented by the positive closure ('+') which is reserved but not used in ABNF.

Although both ABNF and XML support a grammar that permits an unbounded number of input tokens it is not the case that users will speak indefinitely. Speech recognition can perform more effectively if the author indicates a more limited range of repeat occurrences.

Special Cases

Where a number of possible repetitions (e.g. <m-> or <m-n> (n > 0) but not <0>) is expressed on a construct whose only content is one or more tag elements, the behavior of the grammar processor is not defined and will be specific to individual implementations.

Any number of non-optional repetitions (e.g., <m-n>; m>0) of VOID is equivalent to a single VOID.

The behavior of a grammar processor in handling any number of repetitions of NULL is not defined and will be specific to individual implementations.

If the number of repetitions for any expansion can be only zero (i.e. <0> or <0-0>) then the expansion is equivalent to NULL.

2.5.1 Repeat Probabilities

Any repeat operator may specify an optional repeat probability. The value indicates the probability of successive repetition of the repeated expansion.

A repeat probability value must be in the floating pointing range of "0.0" to "1.0" (inclusive). Values outside this range are illegal. The floating point format is one of "n", "n.", "n.nnnn", ".nnnn" (with any number of digits after the dot).

Note: repeat probabilities and weights are different logical entities and have a different impact upon a speech recognition search.

Informative example: A simple example is an optional expansion (zero or one occurrences) with a probability — say "0.6". The grammar indicates that the chance that the expansion will be matched is 60% and that the chance that the expansion will not be present is 40%.

When no maximum is specified in a range (m-) the probabilities decay exponentially.

Grammar authors and speech recognizer developers should be aware of the following limitations upon the definition and application of repeat probabilities as outlined above.

Useful references on statistical models of speech recognition include [JEL98] and [RAB93].

ABNF Form

The following are postfix operators: <m-n> <m-> <m>. A postfix operator is logically attached to the preceding expansion. Postfix operators have high precedence and so are tightly bound to the immediately preceding expansion (see Section 2.8).

Optional expansions may be delimited by square brackets: [expansion]. Alternatively, an optional expansion is indicated by the postfix operator "<0-1>".

The following symbols are reserved for future use in ABNF: '*', '+', '?'. These symbols must not be used at any place in a grammar where the syntax currently permits a repeat operator.

// the token "very" is optional
[very]
very <0-1>

// the rule reference $digit can occur zero, one or many times

$digit <0->

// the rule reference $digit can occur one or more times

$digit <1->

// the rule reference $digit can occur four, five or six times
$digit <4-6>

// the rule reference $digit can occur ten or more times
$digit <10->

// Examples of the following expansion
//   "pizza"
//   "big pizza with pepperoni"
//   "very big pizza with cheese and pepperoni"
[[very] big] pizza ([with | and] $topping) <0->

Repeat probabilities are only supported in the range form. The probability is delimited by slash characters and contained within the angle brackets: <m-n /prob/> and <m- /prob/>.

// the token "very" is optional and is 60% likely to occur
// and 40% likely to be absent in input
very <0-1 /0.6/>

// the rule reference $digit must occur two to four times 
// with 80% probability of recurrence
$digit <2-4 /.8/>

XML Form

The item element has a repeat attribute that indicates the number of times the contained expansion may be repeated. The following example illustrates the accepted values of the attribute.

<!-- the token "very" is optional -->

<item repeat="0-1">very</item>

<!-- the rule reference to digit can occur zero, one or many times -->

<item repeat="0-"> <ruleref uri="#digit"/> </item>

<!-- the rule reference to digit can occur one or more times -->

<item repeat="1-"> <ruleref uri="#digit"/> </item>

<!-- the rule reference to digit can occur four, five or six times -->
<item repeat="4-6"> <ruleref uri="#digit"/> </item>

<!-- the rule reference to digit can occur ten or more times -->
<item repeat="10-"> <ruleref uri="#digit"/> </item>

<!-- Examples of the following expansion -->
<!--   "pizza" -->
<!--   "big pizza with pepperoni" -->
<!--   "very big pizza with cheese and pepperoni" -->

<item repeat="0-1"> 
   <item repeat="0-1"> very </item>
   big 
</item> 
pizza
<item repeat="0-">
   <item repeat="0-1">
      <one-of>
         <item>with</item>
         <item>and</item>
      </one-of>
   </item>
   <ruleref uri="#topping"/>
</item>

The repeat-prob on the item element carries the repeat probability. Repeat probabilities are supported on any item element but are ignored if the repeat attribute is not also specified.

<-- The token "very" is optional and is 60% likely to occur. -->
<-- Means 40% chance that "very" is absent in input -->
<item repeat="0-1" repeat-prob="0.6">very</item>

<-- The rule reference to digit must occur two to four times -->
<-- with 80% probability of recurrence. -->
<item repeat="2-4" repeat-prob=".8">
   <ruleref uri="#digit"/> 
</item>

2.6 Tags

A tag is a legal rule expansion (a tag can also be declared in the grammar header - see S4.1).

A tag is an arbitrary string that may be included inline within any legal rule expansion. Any number of tags may be included inline within a rule expansion.

Tags do not affect the legal word patterns defined by the grammars or the process of recognizing speech or other input given a grammar.

Tags may contain content for semantic interpretation. The semantic interpretation processes may affect the recognition result.

Language attachments have no effect upon tags.

The tag format declaration indicates the content type of all tags in a grammar.

Special Cases

It is legal to use a tag as a stand-alone expansion. For example, a rule may expand to a single tag and no tokens.

  $rule = {TAG-CONTENT};

  <rule id="rule"><tag>TAG-CONTENT</tag></rule>

ABNF Form

A tag is delimited by either a pair of opening and closing curly brackets — '{' and '}' — or by the following 3-character sequences which are considered very unlikely to occur within a tag — '{!{' and '}!}'. A tag delimited by single curly brackets cannot contain the single closing curly bracket character ('}'). A tag delimited by the 3-character sequence cannot contain the closing 3-character sequence ('}!}').

The tag content is all text between the opening and closing character sequences including leading and trailing white space. The contents of the tag are not parsed by the grammar processor.

Tag precedence is the same as for rule references and tokens. In the first example below there is a sequence of six space-separated expansions (3 tokens, a tag, a token and a tag). In the second example, the alternative is a choice between a sequence containing a token and a tag or a sequence containing a rule reference and a tag.

$rule1 = this is a {TAG-CONTENT-1} test {TAG-CONTENT-2};

$rule2 = open {TAG-CONTENT-1} | $close {TAG-CONTENT-2};

$rule3 = {!{ a simple tag containing { and } needs no escaping }!};

XML Form

A tag element can be a direct child of the item and rule elements. The content of tag is CDATA.

<rule id="rule1">this is a <tag>TAG-CONTENT-1</tag> test <tag>TAG-CONTENT-2</tag> </rule>

<rule id="rule2">
   <one-of>
      <item> open <tag>TAG-CONTENT-1</tag> </item>
      <item> <ruleref uri="#close"/> <tag>TAG-CONTENT-2</tag> </item>
   </one-of>
</rule>

2.7 Language

Any legal rule expansion that has an attached language identifier is itself a legal rule expansion. Both the ABNF Form and the XML Form permit a legal language identifier to be attached to any token, sequence or set of alternatives (Note that rule reference does not permit a language identifier to be attached). The syntax for the ABNF Form and for the XML Form are provided below.

The language declaration for a rule expansion affects only the contained content. Moreover, the language declaration affects only the handling of tokens in the contained content and does not affect tags or rule references. The application of language to token handling and particularly to pronunciation lookup is described in Section 2.1.

By default a grammar is a single language document with a language identifier provided in the language declaration in the grammar header (see Section 4.5). All tokens within that grammar, unless otherwise declared, will be handled according to the grammar's language.

In situations where applications target a multilingual user community, grammars that contain words in more than one language may be needed. For example, in response to a prompt such as: "Do you want to talk to André Prévost?" (a combination of an English sentence with a French name), the response may be either "yes" or "oui".

The Speech Recognition Grammar Specification permits one grammar to collect input from more than one language. The specification also permits multiple grammars each with a separate single language to be used in parallel. The specification also permits a single input utterance to contain more than one language. Finally, the specification permits any combination of the above: for example, parallel grammars each with multi-lingual capability.

Not all user agents are required to support all languages, or indeed any or all of the multi-lingual capabilities. The conformance requirements regarding multi-lingual support for XML Form grammar processors and ABNF Form grammar processors are the same and are laid out in Section 5.4 and Section 5.6 respectively.

There is a related challenge for multilingual applications that deal with proper names (people, streets, companies, etc.) that may be spoken with different pronunciations or accents depending upon the language of origin and the speaking language. It is often impossible to predict the language that users will use to pronounce certain tokens. In fact, users may actually use different languages for different words in the same sentence, and in unpredictable ways. For instance, the name "Robert Jones" might be pronounced by a French-speaking user using the French pronunciation for "Robert" but an English pronunciation for "Jones", whereas a mono-lingual English speaker would use the English pronunciation for both words.

Language scoping: language declarations are scoped locally to a document and to a rule definition. In XML terminology, the language attribute is inherited down the document tree. Where a language change encompasses a reference to another grammar, the referenced rule and its containing grammar define the language of the reference expansion. The language in effect at the point of the rule reference does not have any effect upon the referenced rule.

Language and results: The language used in the recognition of a token is not considered a part of the speech result even in the case that a language declaration is associated with a token.

ABNF Form

In the ABNF Form a language identifier may be right-attached to any legal rule expansion except rule reference. The attachment is an exclamation point character ('!') followed by a legal language identifier without intervening white space.

The language attachment has higher precedence than sequences or alternatives. To attach a language to these rule expansion types the expansion should be delimited by parentheses (see Section 2.3).

#ABNF 1.0 ISO-8859-1;

// Default grammar language is US English
language en-US;

// Single language attachment to tokens
// Note that "fr-CA" (Canadian French) is applied to only
//  the word "oui" because of precedence rules
$yes = yes | oui!fr-CA;

// Single language attachment to an expansion
$people1 = (Michel Tremblay | André Roy)!fr-CA;

// Handling language-specific pronunciations of the same word
// A capable speech recognizer will listen for Mexican Spanish and
//   US English pronunciations.
$people2 = Jose!en-US; | Jose!es-MX;

/**
 * Multi-lingual input possible
 * @example may I speak to André Roy
 * @example may I speak to Jose
 */
public $request = may I speak to ($people1 | $people2);

XML Form

XML 1.0 [XML §2.12] defines the xml:lang attribute for language identification. The attribute provides a single language identifier for the content of the element on which it appears. The xml:lang attribute may be attached to one-of , token and item. It applies the token handling of scoped tokens.

<?xml version="1.0" encoding="ISO-8859-1"?>

<!DOCTYPE grammar PUBLIC "-//W3C//DTD GRAMMAR 1.0//EN"
                  "http://www.w3.org/TR/speech-grammar/grammar.dtd">
 
<!-- the default grammar language is US English -->
<grammar xmlns="http://www.w3.org/2001/06/grammar"
         xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" 
         xsi:schemaLocation="http://www.w3.org/2001/06/grammar 
                             http://www.w3.org/TR/speech-grammar/grammar.xsd"
         xml:lang="en-US" version="1.0">
 
  <!-- 
     single language attachment to tokens
     "yes" inherits US English language
     "oui" is Canadian French language
  -->
  <rule id="yes">
    <one-of>
      <item>yes</item>
      <item xml:lang="fr-CA">oui</item>
    </one-of> 
  </rule> 
  
  <!-- Single language attachment to an expansion -->
  <rule id="people1">
    <one-of xml:lang="fr-CA">
      <item>Michel Tremblay</item>
      <item>André Roy</item>
    </one-of>
  </rule>
  
  <!--
     Handling language-specific pronunciations of the same word
     A capable speech recognizer will listen for Mexican Spanish 
     and US English pronunciations.
  -->
  <rule id="people2">
    <one-of>
      <item xml:lang="en-US">Jose</item>
      <item xml:lang="es-MX">Jose</item>
    </one-of>
  </rule>
  
  <!-- Multi-lingual input is possible -->
  <rule id="request" scope="public">
    <example> may I speak with André Roy </example>
    <example> may I speak with Jose </example>
  
    may I speak with
    <one-of>
      <item> <ruleref uri="#people1"/> </item>
      <item> <ruleref uri="#people2"/> </item>
    </one-of>
  </rule>
</grammar>

2.8 Precedence

This section defines the precedence of the ABNF rule expansion syntax. Because XML documents explicitly indicate structure there is no ambiguity and thus a precedence definition is not required. The precedence definitions for the ABNF Form are intended to minimize the need for parentheses.

ABNF Form

The following is the ordering of precedence of rule expansions. Parentheses may be used to explicitly control rule structure.

  1. A rule reference, a quoted token, an unquoted token or a tag.
  2. Parentheses ('(' and ')') for grouping and square brackets ('[' and ']') for optional grouping.
  3. Repeat operator (e.g. "<0-1>") and language attachment (e.g. "!en-AU") apply to the tightest immediate preceding rule expansion. (To apply them to a sequence or to alternatives, use `()' or `[]' for grouping.)
  4. Sequence of rule expansions.
  5. Set of alternative rule expansions separated by vertical bars ('|') with optional weights.

XML Form

None required. XML structure is explicit.

3. Rule Definitions

A rule definition associates a legal rule expansion with a rulename. The rule definition is also responsible for defining the scope of the rule definition: whether it is local to the grammar in which it is defined or whether it may be referenced within other grammars. Finally, the rule definition may additionally include documentation comments and other pragmatics.

The rulename for each rule definition must be unique within a grammar. The same rulename may be used in multiple grammars.

A rule definition is referenced by a URI in a rule reference with the rulename being represented as the fragment identifier.

3.1 Basic Rule Definition

The core purpose of a rule definition is to associate a legal rule expansion with a rulename.

A legal rulename in either the XML Form or ABNF Form is a character sequence that:

Defined rulenames must be unique within a grammar. The schema enforces this by declaring the rulename as an XML ID.

Rulenames are case-sensitive in both XML and ABNF grammars. Exact string comparison is used to resolve rulename references.

A legal rulename cannot be one of the special rules: specifically "NULL", "VOID" or "GARBAGE".

ABNF Form

The rule definition consists of an optional scoping declaration (explained in the next section) followed by a legal rule name, an equals sign, a legal rule expansion and a closing semicolon. The rule definition has one of the following legal forms:

$ruleName = ruleExpansion;
public $ruleName = ruleExpansion;
private $ruleName = ruleExpansion;

For example:

$city = Boston | "New York" | Madrid;
$command = $action $object;
Special Cases

An empty rule definition is illegal.

It is legal to define a rule that expands to empty parentheses or $NULL (equivalent forms). It is legal to define a rule that expands to a single tag.

// Legal
$rule = ();
$rule = $NULL;
$rule = {TAG-CONTENT};

// ILLEGAL
$rule = ;

XML Form

A rule definition is represented by the rule element. The id attribute of the element indicates the name of the rule and must be unique within the grammar (this is enforced by XML). The contents of the rule element may be any legal rule expansion defined in Section 2. The scope attribute is explained in the next section.

<rule id="city">
   <one-of>
      <item>Boston</item>
      <item>"San Francisco"</item>
      <item>Madrid</item> 
   </one-of>
</rule>
<rule id="command">
   <ruleref uri="#action"/>
   <ruleref uri="#object"/>
</rule>
Special Cases

It is not legal to define an empty rule element or a rule element that contains only white space CDATA.

It is legal to define a rule that expands to an empty item or to a single rule reference to NULL.

It is legal to define a rule that expands to a single tag element.

<!-- Legal -->
<rule id="rule"><item/></rule>
<rule id="rule"><ruleref special="NULL"/></rule>
<rule id="rule"><tag>TAG-CONTENT</tag></rule>

<!-- ILLEGAL -->
<rule id="rule"/>
<rule id="rule"></rule>
<rule id="rule">   </rule>

3.2 Scoping of Rule Definitions

Each defined rule has a scope. The scope is either "private" or "public". If not explicitly declared in a rule definition then the scope defaults to "private".

A public-scoped rule may be explicitly referenced (using the fragment identifier syntax of a URI) in the rule definitions of other grammars and in other non-grammar documents. A private-scoped rule cannot be so referenced and is directly accessible only within its containing grammar. A private rule may be explicitly referenced only by other rules within the same grammar.

Informative: grammar authors may consider the following guidance when scoping the rules of a grammar.

ABNF Form

A rule definition may be annotated with the keywords "public" or "private". If no scope is provided, the default is "private".

$town = Townsville | Beantown;
private $city = Boston | "New York" | Madrid;
public $command = $action $object;

XML Form

The scope attribute of the rule element defines the scope of the rule definition. Defined values are public and private. If omitted, the default scope is private.

<rule id="town">
   <one-of>
      <item>Townsville</item>
      <item>Beantown</item> 
   </one-of>
</rule>
<rule id="city" scope="private">
   <one-of>
      <item>Boston</item>
      <item>"San Francisco"</item>
      <item>Madrid</item> 
   </one-of>
</rule>
<rule id="command" scope="public">
   <ruleref uri="#action"/>
   <ruleref uri="#object"/>
</rule>

3.3 Example Phrases

It is often desirable to include examples of phrases that match rule definitions along with the definition. Zero, one or many example phrases may be provided for any rule definition. Because the examples are explicitly marked, automated tools can be used for regression testing and for generation of grammar documentation.

ABNF Form

A documentation comment is a C/C++/Java comment that starts with the sequence of characters /** and which immediately precedes the relevant rule definition. Zero or more @example tags may be contained at the end of the documentation comment. The syntax follows the Tagged Paragraph of a documentation comment of the Java Programming Language [JAVA §18.4]. The tokenization of the example follows the tokenization and sequence rules defined in Section 2.1 and Section 2.3 respectively.

/**
 * A simple directive to execute an action.
 *
 * @example open the window
 * @example close the door
 */
public $command = $action $object;

XML Form

Any number of "example" elements may be provided as the initial content within a "rule" element. The tokenization of the example follows the tokenization and sequence rules defined in Section 2.1 and Section 2.3 respectively.

<rule id="command" scope="public">
    <!-- A simple directive to execute an action. -->
    <example> open the window </example>
    <example> close the door </example>
    <ruleref uri="#action"/> <ruleref uri="#object"/>
</rule>

4. Grammar Documents

A conforming stand-alone grammar document consists of a legal header followed by a body consisting of a set of legal rule definitions. All rules defined within that grammar are scoped within the grammar's rulename namespace and each rulename must be legal and unique.

It is legal for a grammar to define no rules. The grammar cannot be used for processing input since it defines no patterns for matching user input.

4.1 Grammar Header Declarations

A legal stand-alone grammar header consists of a number of required declarations and other optional declarations. In addition, the ABNF Form and XML Form each have additional requirements and capabilities of the header that are specific to each syntactic form. The ordering of header declarations is also specific to the two forms.

The table summarizes the information declared in a grammar header and the appropriate representation in the ABNF Form and XML Form.

Declaration Status ABNF Form XML Form
Grammar version Required §4.2: ABNF self-identifying header §4.3: version attribute on grammar element
XML Namespace Required (XML only) Not applicable §4.3: xmlns attribute on grammar element
Document type Recommended (XML only) Not applicable §4.3: XML DOCTYPE
Character encoding Recommended §4.4: ABNF self-identifying header §4.4: encoding attribute in XML declaration
Language Required in voice mode
Ignored in DTMF mode		 §4.5: language declaration §4.5: xml:lang attribute on grammar element
Mode Optional §4.6: mode declaration §4.6: mode attribute on grammar element
Root rule Optional §4.7: root declaration §4.7: root attribute on grammar element
Tag format Optional §4.8: tag-format declaration §4.8: tag-format attribute on grammar element
Base URI Optional §4.9: base declaration §4.9: xml:base attribute on grammar element
Pronunciation lexicon Optional. Multiple allowed. §4.10: lexicon declaration §4.10: lexicon element
Metadata Optional. Multiple allowed. §4.11.1: meta and http-equiv declarations §4.11.1: meta element
XML metadata Optional. (XML Only) Not applicable §4.11.2: metadata element
Tag Optional. Multiple allowed. §4.12: tag declaration §4.12: tag element

A grammar that complies to this specification must declare the version to be "1.0".

Note: the grammar version indicates the version of the specification implemented by the grammar and is not for versioning of the grammar content. A meta or metadata declaration may be used for content versioning.

ABNF Form: Header Summary

A legal header for a stand-alone ABNF document consists of a required ABNF self-identifying header including the grammar version and optional character encoding followed by these declarations in any order:

ABNF comments may appear between the declarations in the ABNF header after the ABNF self-identifying header.

The header declarations are followed by the rule definitions of the grammar.

The following are two examples of ABNF headers. Note that ordering of the declarations (except the ABNF self-identifying header) is unimportant.

#ABNF 1.0 ISO-8859-1;

language en;
mode voice;
root $myRule;
tag-format FORMAT-STRING;
base <http://www.example.com/base-file-path>;
lexicon <http://www.example.com/lexicon.file>;
lexicon <http://www.example.com/strange-city-names.file>~<media-type>;
meta "Author" is "Stephanie Williams";
http-equiv "Date" is "Fri, 10 Feb 2002 17:27:21 GMT";
{var x=1};

#ABNF 1.0;

// A French Canadian grammar
language fr-CA;

// It's a speech grammar
mode voice;

// Here's the root rule
root $QuebecCities;

XML Form: Header Summary

A legal stand-alone XML Form grammar document consists of:

  1. Legal XML Prolog
  2. Root grammar element with the following attributes
  3. grammar element containing any number of the following elements in any order:

Rule definitions follow the lexicon, meta, metadata and tag declarations.

The following are examples of XML Form grammars headers each including all declarations permitted on the grammar element and one with the DOCTYPE declaration.

<?xml version="1.0" encoding="ISO-8859-1"?>

<grammar version="1.0" 
         xml:lang="en"
         mode="voice"
         root="myRule"
         xmlns="http://www.w3.org/2001/06/grammar"
         xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" 
         xsi:schemaLocation="http://www.w3.org/2001/06/grammar 
                             http://www.w3.org/TR/speech-grammar/grammar.xsd"
         xml:base="http://www.example.com/base-file-path">

<?xml version="1.0" encoding="ISO-8859-1"?>

<!DOCTYPE grammar PUBLIC "-//W3C//DTD GRAMMAR 1.0//EN"
                  "http://www.w3.org/TR/speech-grammar/grammar.dtd">

<grammar version="1.0"
         xml:lang="fr-CA"
         mode="voice"
         root="QuebecCities"
         xmlns="http://www.w3.org/2001/06/grammar"
         xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" 
         xsi:schemaLocation="http://www.w3.org/2001/06/grammar 
                             http://www.w3.org/TR/speech-grammar/grammar.xsd"
         xml:base="http://www.example.com/another-base-file-path">

4.2 ABNF Self-Identifying Header

The ABNF self-identifying header must be present in any legal stand-alone ABNF Form grammar document.

The first character of an ABNF document must be the "#" symbol (x23) unless preceded by an optional XML 1.0 byte order mark [XML §4.3.3]. The ABNF byte order mark follows the XML definition and requirements. For example, documents encoded in UTF-16 must begin with the byte order mark.

The optional byte order mark and required "#" symbol must be followed immediately by the exact string "ABNF" (x41 x42 x4d x46) or the appropriate equivalent for the document's encoding (e.g. for UTF-16 little-endian: x23 x00 x41 x00 x42 x00 x4d x00 x46 x00). If the byte order mark is absent on a grammar encoded in UTF-16 then the grammar processor should perform auto-detection of character encoding in a manner analogous to auto-detection of character encoding in XML [XML §F].

Next follows a single space character (x20) and the required version number which is "1.0" for this specification (x31 x2e x30).

Next follows an optional character encoding. Section 4.4 defines character encodings in more detail. If present, there must be a single space character (x20) between the version number and the character encoding.

The self-identifying header is finalized with a semicolon (x3b) followed immediately by a newline. The semicolon must be the first character following the version number or the character encoding if is present.

For the remaining declarations of the ABNF header white space is not significant.

4.3 XML Form Prolog and Root Element

A legal stand-alone XML Form grammar document must have a legal XML Prolog [XML §2.8].

The XML prolog in an XML Form grammar comprises the XML declaration and an optional DOCTYPE declaration referencing the grammar DTD. It is followed by the root grammar element. The XML prolog may also contain XML comments, processor instructions and other content permitted by XML in a prolog.

The version number of the XML declaration indicates which version of XML is being used. The version number of the grammar element indicates which version of the grammar specification is being used — "1.0" for this specification. The grammar version is a required attribute.

The grammar element must designate the grammar namespace. This can be achieved by declaring an xmlns attribute or an attribute with an "xmlns" prefix. See [XMLNS] for details. Note that when the xmlns attribute is used alone, it sets the default namespace for the element on which it appears and for any child elements. The namespace for XML Form grammars is defined as http://www.w3.org/2001/06/grammar.

It is recommended that the grammar element also indicate the location of the grammar schema (see Appendix C) via the xsi:schemaLocation attribute from [SCHEMA1]. Although such indication is not required, to encourage it this document provides such indication on all of the examples:

<grammar version="1.0"
         xmlns="http://www.w3.org/2001/06/grammar"
         xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" 
         xsi:schemaLocation="http://www.w3.org/2001/06/grammar 
                             http://www.w3.org/TR/speech-grammar/grammar.xsd">
  ...
</grammar>

If present, the optional DOCTYPE must reference the standard DOCTYPE and identifier.

<!DOCTYPE grammar PUBLIC "-//W3C//DTD GRAMMAR 1.0//EN"
                  "http://www.w3.org/TR/speech-grammar/grammar.dtd">

The character encoding is defined on the XML declaration as defined by the XML specification. See Section 4.4 for detail.

The language is defined by the xml:lang attribute on the grammar element. See Section 4.5 for details.

The grammar mode is defined on the grammar element. See Section 4.6 for details.

The root rule is defined on the grammar element. See Section 4.7 for details.

The tag-format is defined on the grammar element. See Section 4.8 for details.

The base URI for the document is defined by the xml:base attribute on the grammar element. See Section 4.9 for details.

4.4 Character Encoding

The character encoding declaration indicates the scheme used for encoding character data in the document. For example, for US applications it would be common to use US-ASCII, UTF-8 (8-bit Unicode) or ISO-8859-1. For Japanese grammars, character encodings such as EUC-JP and UTF-16 (16-bit Unicode) could be used.

Except for the different syntactic representation, the ABNF Form follows the character encoding handling defined for XML. XML grammar processors must accept both the UTF-8 and UTF-16 encodings of ISO/IEC 10646 and may support other character encodings. This follows from an XML grammar processor being a compliant XML processor and thus required to support those character encodings. For consistency, ABNF grammar processor must also accept both the UTF-8 and UTF-16 encodings of ISO/IEC 10646 and may support other character encodings.

For both XML Form and ABNF Form grammars the declaration of the character encoding is optional but strongly recommended. XML defines behavior for XML processors that receive an XML document without a character encoding declaration. For consistency an ABNF grammar processor must follow the same behavior (with adjustments for the different syntax). (Note the character encoding declaration is optional only in cases where it is optional for a legal XML document.)

ABNF Form

The character encoding declaration is part of the self-identifying grammar header defined in Section 4.1 and is processed in combination with the byte order mark, if present, using the same procedure as XML 1.0 [XML §4.3.3].

The following are examples of ABNF self-identifying grammar headers with and without the character encoding declaration.

Note: the ABNF Form syntax does not provide a character reference syntax for entry of a specific character, for example, one not directly accessible fr