Events represented in the transcription of spoken texts

 

The list presented here is a first attempt to put together the events that are symbolized in the transcription of spoken language. This is the result of a partial survey of different transcription systems used in conversation and discourse analysis (see Edwards, 1992, 1993, 1995; Edwards & Lampert (Eds.), 1993; Ochs, 1979). The aim of the list is to provide a preliminary set of items in order to start a discussion of the type of events and labels that should be recommended for use in the symbolic transcription of spoken language. However, it should be noted that there is a certain degree of overlap between categories and that this section is to be regarded as a compilation of current practices organized according to the classical levels used in linguistic analysis. The interested reader can find more details in some of the sources for this compilation (Atkinson - Heritage (Eds.), 1984; Blanche-Benveniste et al., 1991; Coulthard & Montgomery (Eds.), 1981; Crowdy, 1994, 1995; Du Bois, 1991; Du Bois et al., 1993; Gumperz - Berenz, 1993; MacWhinney, 1991; Nelson, 1995; Payne,1995; Peppé, 1995; Stenström, 1994; Tusón,1995).

Segmental level:

Lengthening produced by the speaker, phonetic quality (when different from the standard one in the language or variant transcribed) and stress are usually marked. Also, segments that do not appear in the speaker's production but would appear in the standard form of the language are sometimes registered.

Syllabic level:

Syllable boundaries and syllable lengthening are marked in detailed transcriptions of spoken language.

Word level:

Word boundaries, truncated words, non-standard forms, unfamiliar words, onomatopoeic forms, spelt-out words, acronyms and abbreviations are usually marked up in transcription.

Prosodic aspects of the word such as changes in intonation from the beginning to the end of the word and word stress are transcribed in some systems. Auditorily perceived pauses within a word and between words can be also transcribed.

Utterance level:

Utterance boundaries and utterance modality can be coded. Also, breaks and interruptions in the utterance -- with or without pauses -- are signalled in transcription systems. Utterance boundaries can be also marked.

Suprasegmental level:

Intonational units:

Intonation unit and embedded sub-unit boundaries, incomplete or truncated tone units, resets, junctures or break indices can be marked.

Final contours of terminated and non-terminated tone units are transcribed according to basic patterns (falling, rising, level) and combinations of these.

Pitch:

Changes in pitch over the course or part of an utterance, pitch level, range, register and pitch on word and on phrase together with transitional continuities can be part of a prosodic markup

Stress:

Stress in any part of an utterance is usually signalled; syllable weight, pitch accent and different levels of stress can also be introduced in the transcription.

Prominence, emphasis or contrastively stressed syllables can be also indicated.

In some transcription systems, indications about the tension and the rhythmic qualities of the utterance are introduced.

Intensity or loudness:

Relative or absolute intensity of parts of the utterance can be marked in subjective terms.

Speech rate or tempo:

In spoken language transcriptions, speech rate can be timed or untimed. In this latter case, absolute or relative subjective assessments can be found. Irregularities in rhythm are also registered.

Pauses:

Both silent and vocalized pauses are usually marked in spoken language transcription. Audible breathings can also be indicated in the transcription. Pauses can be timed or untimed, depending on the accuracy and aims of the transcription.

Paralinguistic events or vocalized semi-lexical or non-lexical phenomena:

Semi-lexical - such as aha, erm, mm - and non-lexical vocalized events are represented in transcription, together with their position relative to the lexical stretch. Voice quality and other vocal events - such as shouting or singing - can be also coded in the transcription.

Speaker turns:

Speaker turns are signalled in conversation analysis. Speaker identity, the nature of the transition between utterances, its sequential relation, the type of overlapping and latching of speaker's utterances are usually marked.

Contextual comments on the transcription:

There should be means to include in the transcription any additionalinformation provided by the researcher: non-vocalized non-communicative phenomena, non-vocalized communicative phenomena (kinesic information), information about the type of text being transcribed or background noise.

Transcription difficulties:

Researchers have devised ways of codifying and noting the difficulties found in the process of transcribing recorded spoken language. These difficulties may be related to the performance of the speaker or to technical problems in the recordings. Ways to assess the degree of accuracy of the transcription have also been developed.