Talkbank Schema version 1.1.2

The current version defines two types of documents: the global elements below... The global types are available for reusing through schema type extension/restriction. The most up to date document definition is CHAT, it is also the richest in structure. Ideally, each group should develop a schema module defining the structure of their specific (class of) annotations, this schema should be an assembly of their definitions.

Developed by Romeo Anghelache, from the CHAT specifications, released under the GNU Public License, 2001

Imported schemata:

• http://dublincore.org/schemas/xmls/simpledc20020312.xsd from namespace http://purl.org/dc/elements/1.1/

Global Elements in main module

• <CHAT>
  • (
    • <Metadata>? {Metadata}
    • <comment>* {comment}
    • <Participants> @Participants; a structure enumerating the beings participating
      • (
        • <participant>+ {participant}
        )
    • [+
      • <comment> {comment}
      • <tcu> a group of utterances having something in common, usually the speaker
        • (
          • <u>+
            • << {utterance} >>
              • who : {xs:string}
          • <media>? {media}
          )
        • who : {xs:string}
      • <u>
        • << {utterance} >>
          • who : {xs:string}
      ]
    • <media>? {media}
    )
  • Id : {xs:string}
  • Version :
    • >> {xs:string}<<
      pattern:1.1.[0-9]{1,}
  • Date : {xs:date} 31 March 1999 is formatted as 1999-03-31
  • Duration : {xs:duration} this work might be done in an extended interval of time; a duration of 1 year, 2 months, 3 days, 10 hours, and 30 minutes, one would write: P1Y2M3DT10H30M
  • Coder : {xs:string}
  • Comment : {xs:string}
  • Situation : {xs:string}
  • Corpus : {xs:string}
  • Media : {xs:anyURI}
  • Lang : {xs:language}
  • Script : {xs:language} e.g. when you write russian words using english characters, then Lang="ru" and Script is "en"
  • Warning : {xs:string}

• <AGSet>
  • (
    • <Metadata> {Metadata}
    • <Timeline>*
      • (
        • <Metadata>? {Metadata}
        • <Signal>*
          • (
            • <Metadata>? {Metadata}
            )
          • id : {xs:ID}
          • mimeClass : {xs:string}
          • mimeType : {xs:string}
          • encoding : {xs:string}
          • unit : {xs:string}
          • href : {xs:string}
          • track : {xs:NMTOKEN}
        )
      • id : {xs:ID}
    • <AG>*
      • (
        • <Metadata>? {Metadata}
        • <Anchor>*
          • id : {xs:ID}
          • offset : {xs:string}
          • unit : {xs:string}
          • signals : {xs:IDREFS}
        • <Annotation>*
          • (
            • <Feature>*
              • << {xs:string} >>
                • name : {xs:NMTOKEN}
            )
          • id : {xs:ID}
          • type : {xs:NMTOKEN}
          • start : {xs:IDREF}
          • end : {xs:IDREF}
        )
      • id : {xs:ID}
      • type : {xs:NMTOKEN}
      • timeline : {xs:IDREF}
    )
  • id : {xs:ID}
  • version : {xs:string}

Global Types in main module

• Metadata : administrative descriptions, reused from Dublin Core

  • {dc:elementsGroup}

• annotation : {free text}

  • (?
    • <media> {media}
    )
  • lang : {xs:language}
  • script : {xs:language}
  • flavor : {xs:string} allows semi structured extensions to the current set of annotations:
  • type : these are the (legacy) dependent tiers, %mor line is, now, <morphemics> element
    • >> {xs:string}<<
      SALT %def; on the main line, not recommended
      actions %act
      addressee %add
      alternative %alt
      cohesion %coh; cohesion tier
      comments %com;[% text]; comments by investigator
      completion [0 text]; an omitted word
      english translation %eng
      errcoding %err; error coding
      excluded material [%exc ...]
      explanation %exp; [= text]
      extension %cod; general purpose coding
      extension arbitrary annotations, CHAT postcodes, intended as an extension mechanism
      facial %fac
      flow %flo
      gesture %gpx
      intonation %int
      language %lan
      paralinguistics %par:
      phonetic %pho:
      phonomodel %mod:
      replacement [: text]
      situation %sit
      speech act %spa
      target gloss %gls
      time stamp %tim

• comment :

  • << {xs:string} >>
    • type :
      • >> {xs:string}<<
        Activities
        Background
        Coder
        Coding
        Date
        Education
        Generic
        Language
        Location
        New Episode
        Room Layout
        Script
        Situation
        Stimulus
        Tape Location
        Time Duration
        Time Start
        Transcriber
        Warning

• event : non verbal happenings

  • (
    • [*
      • <k> {markers} scoped symbols
      • <a> {annotation} the place to add research content
      • <tone> {tone}
      ]
    • <media>? {media}
    )
  • type :
    • >> {xs:NMTOKEN}<<
      action 0
      breath
      cough
      inbreath .hhh in CA
      knock
      laugh
      noise a not yet clearly cathegorized noise
      outbreath hhh in CA
      rustle
      smack smacking lips noise
      squeak
      swallow
      throat clearing throat noise
      unintelligible xxx;yyy
      untranscribed www

• form : {free text} intended as a feature of a word, see also the CHAT conventional notations

  • (
    • <p>* {prosody}
    • <media>? {media}
    )
  • type :
    • >> {xs:string}<<
      UNIBET @u
      approximation @ap
      babbling @b
      child-invented @c
      completion ()
      cue @cue
      dialect @d
      ellipsis 00word
      family-specific @f
      filled pause @fp
      fragment &; phonological fragment
      generic @g
      incorrect omission 0*word
      informal sign @inf
      informal sign with speech @ins
      interjection @i
      letter @l
      neologism @n
      no voice @nv
      omission 0word
      onomatopoeia @o
      phonology consistent @p
      phrasal repetition @pr
      schwa @sc
      second language @s
      sign speech @sas
      signed language @sl
      test @t
      unintelligible xx/yy
      word play @wp
      words to be excluded @x

• group : structure used to let annotations to belong to more than one word, can be recursive, although unnecessary: one can attach more than one annotations to a word, group of words, or whole utterances

  • (
    • [+
      • <g> {group}
      • <w> {word} a word
      • <wn> {wordnet} a construct formed by words linked through clitic or compound e.g. once+and+for+all
      • <tone> {tone}
      • <pause> {pause}
      • <e> {event} a reference to a point/portion of a mute/action signal, e.g. 0
      • <s>? {separator} semicolon , clause delimiter [c];
      ]
    • [*
      • <k>* {markers} scoped symbols
      • <a> {annotation} the place to add research content
      • <x> {translation} one may attach a translation of a word;
      ]
    • <media>? {media}
    )
  • who : {xs:string}

• linker : utterance initiators or linkers; they indicate the way to fit the current utterance with an earlier one, the CHAT conventional symbols are listed too

  • type :
    • >> {xs:string}<<
      lazy overlap mark +<
      other completion ++
      quick uptake +^
      quoted utterance next +"
      self completion +,

• markers : inlined annotations, the conventional CHAT symbols are listed too

  • (?
    • <media> {media}
    )
  • lang : {xs:language}
  • script : {xs:language}
  • times : {xs:positiveInteger}
  • type :
    • >> {xs:string}<<
      allegro quicker tempo, no CHAT equivalent, used in CA
      best guess [?] in CHAT, ( text ) in CA
      contrastive stressing [!!]
      error [*]
      false start [/-]
      forte larger volume, louder, no CHAT equivalent, used in CA
      lento slower tempo, no CHAT equivalent, used in CA
      overlap follows [>]
      overlap precedes [<]
      piano lower volume, no CHAT equivalent, used in CA
      quotation ["]
      retracing [/] in CHAT
      retracing reformulation [///] in CHAT
      retracing unclear [/?]
      retracing with correction [//] in CHAT, - in CA
      stressing [!]

• media : a pointer to a selection in a video/audio file

  • type :
    • >> {xs:NMTOKEN}<<
      application
      audio
      image
      text
      video
  • href : {xs:anyURI}
  • start : {xs:float}
  • end : {xs:float}
  • unit :
    • >> {xs:NMTOKEN}<<
      b byte
      c character
      f frame
      s second

• morphemic_marker : morphemes

  • << {string} >>
    • type :
      • >> {xs:string}<<
        enx english translation, CHAT equivalent is =
        ioafx incorrectly omitted affix, CHAT equivalent is -0*
        mc morphological cathegory, CHAt equivalent is :, when used after the stem
        oafx omitted affix, CHAT equivalent is -0
        pfx prefix marker, CHAT equivalent is #
        sfx suffix marker, CHAT equivalent is -
        sfxf suffix fusion marker, CHAT equivalent is &;

• morphemics : the unit of a %mor line corresponding to a word (this element belongs to a word element, but, if the precise correspondence is not yet established, these elements will be present at the utterance level (contained in an utterance);

  • (+
    • {morphemic_unit}
    • (*
      • <wk>? {wordnet_marker}
      • {morphemic_unit}
      )
    )
  • lang : {xs:language}
  • script : {xs:language}

• participant : the beings along with their characteristics (age, sex...)

  • id : {xs:string}
  • role : {xs:string}
  • name : {xs:string}
  • age : {xs:duration}
  • ageTo : {xs:duration}
  • group : {xs:string}
  • sex :
    • >> {xs:NMTOKEN}<<
      female
      male
  • SES : {xs:string}
  • birthday : {xs:string}

• pause : the # symbol, pause between words

  • [*
    • <a> {annotation} the place to add research content
    • <k> {markers} the place to add research content
    ]
  • length : {xs:float}

• prosody :

  • (?
    • <media> {media}
    )
  • type :
    • >> {xs:string}<<
      accented nucleus //
      blocking ^
      contrastive stress ///
      creaky *text* in CA
      drawl :
      pause ::
      stress /

• separator :

  • type :
    • >> {xs:string}<<
      clause delimiter [c] clause-delimiter;
      comma
      semicolon

• string : a nonempty string

  • >> {xs:string}<<

• syntactic : {free text} syntactic structure

  • [*
    • <AGR> {string}
    • <CAT> {string}
    • <COUNTER> {string}
    • <DUMMY> {string}
    • <ELLIPSIS> {string}
    • <INDEX> {string}
    • <MODAL> {string}
    • <MOOD> {string}
    • <NEGATION> {string}
    • <NUM> {string}
    • <POSS> {string}
    • <REL-PRO> {syntactic}
    • <ROOT> {string}
    • <VFORM> {string}
    • <SUBCAT> {string}
    • <ADJUNCT> {syntactic}
    • <AUX> {syntactic}
    • <COMMUNICATOR> {syntactic}
    • <COMP> {syntactic}
    • <CONJ> {syntactic}
    • <DET> {syntactic}
    • <INF-PARTICLE> {syntactic}
    • <MULTIPLE> {syntactic}
    • <MODIFIER> {syntactic}
    • <OBJECT> {syntactic}
    • <PP-OBJ> {syntactic}
    • <PREDICATE> {syntactic}
    • <POSSESSOR> {syntactic}
    • <QUANTIFIER> {syntactic}
    • <QUANT> {syntactic}
    • <QUANT1> {syntactic}
    • <QUANT2> {syntactic}
    • <RECIPIENT> {syntactic}
    • <RELATIVE-CLAUSE> {syntactic}
    • <SUBJ> {syntactic}
    • <TAG-AUX> {syntactic}
    • <TAG-PRO> {syntactic}
    • <TITLE> {syntactic}
    • <XCOMP> {syntactic}
    • <VOCATIVE> {syntactic}
    ]

• terminator : period, question, exclamation; basic utterance terminator;

  • type :
    • >> {xs:string}<<
      e
      interruption +/.
      interruption question +/?
      p
      q
      question exclamation +!?
      quotation next line +"/.
      quotation precedes +".
      self interruption +//.
      self interruption question +//?
      trail off +...
      trail off question +..?

• tone :

  • type :
    • >> {xs:string}<<
      fall rise -,.
      falling final -.
      falling nonfinal -_
      final exclamation -!
      lengthened -:
      level nonfinal -,
      low level -
      rise fall -'.
      rising final -?
      rising nonfinal -'
      tag question ,,

• translation :

  • << {xs:string} >>
    • lang : {xs:language}

• utterance :

  • (
    • <linker>* {linker}
    • <a>* {annotation} the place to add research content
    • [+
      • <w> {word} a word
      • <g> {group} a group of words
      • <wn> {wordnet} a construct formed by words linked through clitic or compound e.g. once+and+for+all
      • <tone> {tone}
      • <pause> {pause}
      • <e> {event} a reference to a point/portion of a mute/action signal, e.g. 0
      • <s> {separator} semicolon, clause_delimiter [c];
      ]
    • [*
      • <k>* {markers} scoped symbols
      • <a>* {annotation} the place to add research content
      • <mor>? {morphemics}
      • <syn>? {syntactic}
      • <t>? {terminator}
      • <media>? {media}
      ]
    )
  • date : {xs:string}
  • lang : {xs:language}
  • script : {xs:language}

• word : {free text}

  • [*
    • <p>* {prosody} stress, blocking etc.
    • <f>* {form} equivalent of CHAT symbol @;
    • <mor>? {morphemics}
    • <tone>? {tone}
    • <a>* {annotation} the place to add research content
    • <k>? {markers} scoped symbols
    • <x>? {translation} one may attach a translation of a word;
    • <mk>? {morphemic_marker} an optional suffix
    • <media>? {media}
    ]
  • who : {xs:string}
  • lang : {xs:language}
  • script : {xs:language}

• wordnet :

  • (
    • <w> {word} a word
    • (+
      • <wk> {wordnet_marker} clitic or compound markers, may be used in morphemics and wordnet
      • <w> {word} a word
      )
    • <f>* {form} equivalent of CHAT symbol @;
    • <a>* {annotation} the place to add research content
    • <k>? {markers} scoped symbols
    • <x>? {translation} one may attach a translation of a word;
    • <mk>? {morphemic_marker} an optional suffix
    • <media>? {media}
    )
  • who : {xs:string}

• wordnet_marker :

  • type :
    • >> {xs:string}<<
      cli clitic, CHAT ~
      cmp compound, CHAT +

Useful groups in main module

• morphemic_unit :

  • (
    • <pfx>? the whole morpheme is actually a prefix, CHAT equivalent is ~#
    • <pos>
      • (
        • <err>?
        • <c>{xs:string} category
        • <s>*{xs:string} subcategory
        )
    • <stem>{xs:string}
    • <err>?
    • <mk>* {morphemic_marker}
    )

Schema viewer developed by Romeo Anghelache, copyleft GNU GPL, 2001, version 1.01