(conversational) games
A category of pragmatic meso-level annotation, encompassing all utterances following an initiating move up to the point where the goal of the initiating move has been fulfilled or abandoned.

activity type
A type of communicative activity with its own conventions of dialogue, e.g. negotiation, instruction, problem solving, etc.

A grammatical element which is relatively peripheral to the clause or sentence in which it occurs, often optional and moveable, and expressing one or a range of meanings such as time, place, manner, purpose, and reason.

A type of dysfluency which takes the form of an ill-formed syntactic structure, beginning according to one structural plan, and ending according to another.

a) The act of adding additional types of linguistic information to the transcription (representation) of a text or discourse.
b) The material added to a corpus by means of (a): e.g. part-of-speech tags.

annotation tools
Specialised tools that allow the mark-up of corpora with annotation; e.g. an automatic part-of-speech tagger.

An adjective characterising a particular piece of dialogue in a corpus which has been sampled with the specific aim of using it in the context of some application development.

ASCII (American Standard Code for Information Interchange)
7-bit character encoding scheme

autosegmental/metrical phonology
An explicitly phonological approach to intonation. Tunes are characterised in terms of discrete elements, relatively high (H) or relatively low (L) tones, which map onto acoustic F0 targets. The tones are associated with either stressed syllables, lending prominence to a particular syllable or word, or with the edges of phrases, where they take on a junctural function.

A type of dialogue turn or utterance whose function is to encourage one’s interlocutor to continue talking; e.g. uh-huh, mhm

backward-looking communicative function
A communicative function that refers back or relates to an item that has previously been mentioned.

break indices
A tier in ToBI annotation indicating a perceived juncture between words transcribed on the orthographic tier.

channel characteristics
Information relating to the recording conditions for speech data, i.e. type of microphone, number of recording channels, etc.

Units into which speech can be segmented or according to which it is organised.

common ground unit (CGU)
An abstract category of pragmatic meso-level annotation, comprising all units of speech that are relevant to developing mutual understanding of a topic in a dialogue.

communicative status
A pragmatic utterance tag that indicates whether an utterance is intelligible or complete.

A word whose function is to signal a link between syntactic units of equivalent status; conjunctions are normally subdivided into coordinating (and, or, etc.) and subordinating (if, because, etc.) types.

An electronic collection of texts, comprising either written or spoken data or both

DAMSL (Dialog Act Mark-up in Several Layers)
An annotation scheme for marking up dialogue corpora, partly developed upon a consensus agreement reached at the first Discourse Resource Initiative workshop.

A type of discourse taking place between two or more human participants or between human participants and a computer.

dialogue act
a) A type of speech act that contributes towards achieving a specific sub-goal or goal in a dialogue. It may consist of part of an utterance, a complete utterance or even a set of utterance
b) A stretch of speech which is normally identified by the criteria that it is spoken by one person and has an identifiable function. The term "utterance" is used by different researchers for different purposes, and is liable to ambiguity.

dialogue corpus
A corpus consisting of a collection of dialogues.

discourse marker
A word (or fixed phrase) which is loosely attached to a larger structure in a stretch of speech and which has discoursally defined role such as indicating a change in the direction of the discourse, or signalling the speaker's stance towards what has been said, e.g. well, right,

A variable defining the type of dialogue according to its subject-matter, e.g. travel, transport, appointment scheduling, etc.

DRI (discourse resource initiative)
A group of researchers working in the area of dialogue annotation that is attempting to create standards in dialogue annotation based on consensus ideas reached at annual workshops.

DTD (Document Template Definition)
An SGML or XML template outlining and constraining the structure of SGML or XML documents.

dysfluency phenomena
Unintended phenomena that occur in spoken language as a result of on-line production pressures, such as hesitations, false starts, some kinds of repetition, and filled pauses.

dysfluent repetition
A dysfluency phenomenon that consists of the speaker repeating part of an utterance.

A variant of ToBI, developed for American English, and applied to Southern Standard British and Standard Australian English..

The fundamental frequency of an utterance, representing its intonation contour and corresponding roughly to perceived pitch. An acoustic measurement, rather than a perceptual category.

false start
A dysfluency phenomenon where the speaker interrupts an utterance he or she has already begun and corrects/reformulates it.

forward-looking commmunicative function
A communicative function that establishes the background for verbal or non-verbal action that is to follow or constrains it.

functional utterance
A classification of an utterance based on its functional content, rather than its structural, syntactic properties.

fundamental frequency
see F0.

A variant of ToBI, developed for the transcription of Glaswegian English

A variant of ToBI, developed for the transcription of German

A part of a computer text file that precedes the textual material itself, and includes information applying to the text as a whole: e.g. its origin, content, structure, and information about speakers.

information level
A pragmatic utterance tag that indicates the semantic content of the utterance and how it relates to the task at hand.

information status
A pragmatic utterance tag that indicates whether the content of an utterance is old or new.

integrated resources
Resources such as corpora, lexicons and tools which relate to be spoken and written language data.

An exclamatory part of speech that can stand on its own or can be loosely attached to an utterance, conveying emotive or attitudinal content rather than referential meaning; e.g. Ah, Oops. The most typical interjections are phonetically simple and have phonetic peculiarities not found in the regular vocabulary of a language. E.g., Phew in English begins with a bilabial fricative. In a broader sense, interjection can refer to a range of word types, including discourse markers such as well and response forms such as yeah, no.

intermediate phrase
A prosodic unit in autosegmental/metrical phonology that corresponds roughly to a minor tone group. The smaller of of two possible levels of phrase.

intonation phrase
A prosodic unit in autosegmental/metrical phonology that corresponds roughly to a major tone group. The larger of two possible levels of phrase.

A unit of pragmatic macro-level annotation that seeks to capture the relationship between CGUs. I can stand for either informational or intentional.

A variant of ToBI, developed for the transcription of Japanese

kinesic features
Extralinguistic communicative behaviour of a dialogue participant involving bodily movement: e.g. gestures, eye contact.

A mode of vibration of the vocal cords that can produce the effect of creaky voice.

London-Lund Corpus
A large (computer) corpus of spoken British English containing orthographic transcription and prosodic annotation.

macro-level annotation
A type of pragmatic annotation that seeks to capture the highest-level structures in a dialogue.

Map Task corpus
A corpus developed at the Human Communication Research Centre at the University of Edinburgh.

MARSEC (MAchine Readable Spoken English Corpus)
A time-aligned version of the Spoken English Corpus.

meso-level annotation
A type of pragmatic annotation that seeks to capture intermediary-level structures in a dialogue.

micro-level annotation
A type of pragmatic annotation that seeks to capture the low-level structures in a dialogue.

morphosyntactic annotation
Part of Speech tagging, i.e. a type of annotation that identifies the grammatical category of each word token in a corpus.


An umbrella term for a sequence which consists of more than one orthographically-defined word, but which constitute a minimal unit for inclusion in a lexicon or for some level of annotation. The clearest case is a complex preposition or conjunction such as in spite of or so that, which may be given a single part-of-speech tag, although it consists of two or three orthographic words.

non-verbal sounds
Background sounds in a recording that do not form part of the verbal interaction of the participants; e.g. the noise of a door closing.

orthographic sentence
A sentence according to the normal conventions of written language: i.e. a minimal segment of text which begins and ends with sentence-boundary markers (normally this means beginning with a capital letter and ending with a full-stop).

orthographic transcription
A written rendering of spoken discourse using the normal conventions of written language: i.e. using convention spellings and (sometimes) conventional punctuation.

orthographic word
A word according to the normal conventions of written language: i.e. a minimal segment of text which begins and ends with word-boundary markers (normally this means beginning and ending with a space character, sometimes accompanied by one or more punctuation marks.

paralinguistic features
Vocal but non-verbal, non-phonemic features accompanying speech, such as tone of voice, extra slow delivery, laughter, etc.

A person who takes part in a dialogue as speaker, addressee, etc.

perceived pauses
Pauses that the human listener perceives as the absence of sound when they may be realised as genuine pauses (absence of any acoustic signal) or simply be perceived as pauses owing to rhythmic factors such as final lengthening.

pitch movement
Raising or lowering of the fundamental frequency to achieve intonational effects.

pitch reset
Raising (or sometimes lowering) of the fundamental frequency at the beginning of a new structural unit, e.g. a sentence, paragraph, etc., which may have boundary-marking function.

POS (Part Of Speech)
A morphosyntactic/grammatical word, such as noun, verb, adjective, etc.

The act of assigning morphosyntactic/grammatical word tags to the elements of a text.

pragmatic particle (e.g. doch, ja)
A particle that may be used to express agreement or negation.

A phonetic/phonological term referring to features of (intonational) accentuation, increased loudness, increased stress or a combination of those factors. Perceived prominence is normally related to a a combined effect created by the above factors.

prosodic annotation
A level of annotation of spoken data which represents suprasegmental features such as pitch, and accent.

quasi-lexical vocalization
A term used for speech items such as uh-huh and erm which are on the boundaries of the lexicon of a language. They may be regarded as words, or alternatively as conventionalised noises used to fulfil some function in speech. Often their representation in writing is variable.

regularization tag
An SGML tag which indicates that a particular text segment has been changed from the original record, in order to regularize its representation in the text (e.g. it may be a correction of a typographical error).

Orthographic transcription of a written or spoken text.

retrace-and-repair sequence
An alternative term for false start (q.v.)

SAMPA (Speech Assessment Methods Phonetic Alphabet)
An ASCII encoding scheme for phonetic transcription.

SAMPROSA (Speech Assessment Methods Prosodic Alphabet)
An ASCII encoding scheme for prosodic transcription.

External circumstances governing the collection of data, such as channel or speaker characteristics, etc.

SEC (Spoken English Corpus)
A Corpus of spoken English developed at the University of Lancaster.

(a) Act of dividing up a stretch of speech into its phonetic constituents (segments).
(b) Sometimes also used to refer to the division of dialogues into different structural (or functional) units. It may be best to avoid this usage because of ambiguity.

The act of rephrasing an item of speech that the speaker has realised is incorrect or may need further explanation. A dysfluency phenomenon.

SGML (Standard Generalized Markup Language)
A type of document markup language, mainly used in publishing and research, which uses tags to identify structural elements of a text.

speaker characteristics
Features of a speaker such as age, sex, native language, etc., important in identifying factors that may aid in speech recognition.

speech management
The use of phenomena that help a speaker to structure/organise and correct utterances appropriately.

structural utterance
A unit of textual structure, rather than of functional nature.

SUSANNE corpus
A corpus of written American English morphosyntactically and syntactically annotated according to a detailed standard parsing scheme (see Sampson 1995).

syntactic annotation
The addition to a text or a corpus of annotations indicating syntactic structure. E.g., each sentence may be annotated with a parse tree.

syntactic blend
An alternative term for anacoluthon (q.v.).

Task-driven dialogues are dialogues in which the participants aim to solve a specific task, the nature of which is likely to influence the structure and content of the dialogue itself.

ToBI (Tones and Break Indices)
A phonological annotation scheme for marking up prosodic information

tone group
Also sometimes referred to as tone unit. Prosodically defined stretch of speech, comprising one or several words, containing at least one accent and ending with a prosodic break. Generally, a distinction is made between minor and major tone groups. The former often correlate with phrases or clauses at the syntactic level, while the latter tend to correlate with complete sentences. C.f. intermediate and intonation phrase in ToBI.

Topic spotting
An automatic procedure of trying to identify the topic of a dialogue(/text) by extracting and analysing key words.

A dialogue corpus developed at the University of Rochester, USA, containing dialogues dealing with plans to move trains or cargo.

Syntactically annotated corpus; more specifically: a collection of naturally occurring sentences annotated with parse trees, and typically used for training and testing automatic parsers.

TSM (Tonetic Stress Marks)
Symbols of prosodic annotation, in the form of diacritics preceding accented syllables and indicating falling, level or rising pitch.

A unit of (spoken) text, generally comprising everything said by one speaker up to the point where another speaker takes over.

A unit of (spoken) text, often loosely defined and used. On the structural level, utterances may correspond to phrases or sentences uttered by a speaker, whereas on the functional level, they may correspond to dialogue acts.

The acoustic/physical representation of a complex speech signal, based upon the amplitude of its component parts, usually displayed on a graph on which the x-axis represents amplitude and the y-axis time.

Wizard of Oz experiment/scenario
A type of experimental dialogue scenario in which a human subject believes himself/herself to be communicating with a machine, when in fact the responses are made by a hidden human being. The purpose of Wizard of Oz (WOZ) experiment is to find out as much as possible about how humans would behave in a human-machine dialogue situation.

A word whose articulation is begun but not completed, either because of interruption or because of dysfluency.

XML (eXtensible Markup Language)
A type of document markup language, which represents an extensible subset of SGML.

X-SAMPA (eXtended Speech Assessment Methods Phonetic Alphabet)
An extended ASCII encoding scheme for phonetic transcription.

Items for acronym section:

ATR (Advanced Telecommunications Research)

BNC (British National Corpus)

CES (Corpus Encoding Standard)

CREA (Corpus de Referencia del Español Actual)

CSCS (Corpus of Spoken Contemporary Spanish)

HCRC (Human Communication Research Centre)

INTSINT (International Transcription System for Intonation)

LE (Language Engineering)

MULTEXT (Multilingual Text Tools and Corpora)

TEI (Text Incoding Initiative)

UCREL (University Centre for Computer Corpus Research on Language)

URL (Uniform Resource Locator)