next up previous contents
Next: Transcription practices Up: Evaluation of the Text Previous: Evaluation of the Text

Elements in a text


While, in the TEI Guidelines, contextual information about a text is provided in the header, in French's system this kind of information is found in the text ID which contains contextual details such as the date, location and manner of the recording as well as information about the participants and the topic

Since the TEI header is more detailed, Payne (1992:12) suggests that: This is an area in which it may be necessary in the short term to regard the TEI conventions as a target [...] Users will need to exercise their discretion within the context of the long-term aim of full conformity.

Units intermediate between the text and the utterance can be marked as divisions in the TEI Guidelines. No explicit provision for this subdivision is found in French's conventions.

Basic structural elements:
TEI distinguishes seven structural elements in spoken texts: utterances, pauses, vocal, kinesic, events, writing and shifts. They are defined as follows (Sperberg-McQueen - Burnard, Eds., 1994):
a stretch of speech usually preceded and followed by a pause or by a change of speaker.
a perceived pause within or between utterances.
a vocalized but not necessarily lexical phenomenon (e.g. voiced pauses).
any communicative phenomenon, not necessarily vocalized (e.g. a gesture).
any phenomenon or occurrence, not necessarily vocalized or communicative (e.g. incidental noises).
a passage of written text revealed to participants in the course of a spoken text.
marks the point at which there is a change in some paralinguistic feature.

Each basic structural element is identified by a tag that has various attributes:

TEI utterances tend to correspond with speaker turns, while French defines functional sentences, according to grammatical, semantic and pragmatic criteria. Straightforward procedures for conversion between both representations are suggested by Payne (1992:16ff).

French's functional sentences can be related to TEI segments.

Tone units:
According to Payne (1992:28-29), TEI conventions should be enlarged to allow the simultaneous transcription of syntactic and intonation units. NERC conventions for the delimitation of tone units can be translated to TEI standards.

Changes in paralinguistic features -- voice quality, loudness, pitch range and speech rate -- can be specified in the TEI conventions by a specific tag for shifts with associated feature values. In French's system this will be signalled by means of the transcriber comments.

However, it is worth mentioning that in Payne's opinion, the detailed transcription of changes in paralinguistic features might not be useful for a large variety of users: instead, he favours a way of linking the orthographic transcription to a recording of the particular spoken text, with a more accessible lexicon of descriptive texts (Payne, 1992:30).

TEI provides guidelines to define the temporal relationship of points in an utterance using different strategies, while French's system uses the tape counter. Conversion between both systems is possible. Payne (1992:32ff) notes that when accurate timing is necessary the alignment should be made between the digitized speech signal and units on the CD-ROM on which speech is stored.

Simultaneous events:
In the TEI Guidelines, simultaneous events are represented referring to an external alignment map, while in French's conventions only speaker overlaps are coded at a detailed level of transcription. Suggestions for improvements in the proposed system can be found in Payne (1992:36-7).

Mechanisms for conversion between TEI and NERC systems for annotating pause can be easily implemented; some suggestions are presented by Payne (1992:17ff)

In the TEI Guidelines, vocal refers to vocalized non-lexical -- burp, click, cough, giggle, laugh, sneeze, sob, yawn -- and semi-lexical events; in French's system these would appear as non-verbal. A quite straightforward conversion is possible when vocals are produced by the same speaker of the utterance.

Kinesic refers to non-vocalized communicative events such as gestures or facial expressions. Since the distinction between vocals and kinesic is not made in French's system, automatic conversion from NERC's adopted system to TEI standards would require a previous classification of French's non-verbal elements.

In a TEI transcription, a non-vocalized non-communicative event that affects communication -- e.g. a sudden noise -- can be described with a specific tag. These events would appear in the transcriber comments in French's system.

Writing revealed to participants during the course of a spoken text can be marked with a TEI tag, while this fact would be included in the transcriber comments in French's system. Payne (1992:26) suggests the need for more flexibility in the TEI system of representing the disclosure of written text.

next up previous contents
Next: Transcription practices Up: Evaluation of the Text Previous: Evaluation of the Text