Tips on Transcription Practices for Linguistic Anthropological Analyses

Some considerations on transcription from Alessandro Duranti's Transcription: From Writing to Digitized Images:

We must keep in mind that a transcript of a conversation is not the same thing as the conversation; just as an audio or video recording of an interaction is not the same as the interaction. But the systematic inscription of verbal, gestural, and spatio-temporal dimensions of interactions can open new windows on our understanding of how human beings use talk and other tools in their daily interactions." (Duranti, 1999, p. 161)

A transcript is a technique for the fixing (e.g. on paper, on a computer screen) of fleeting events (e.g. utterances, gestures) for the purpose of detailed analysis.Transcripts are inherently incomplete and should be continuously revised to display features of an interaction that have been illuminated by a particular analysis and allow for new insights that might lead to a new analysis. (See Alessandro Duranti Linguistic Anthropology, Cambridge University Press, 1997: ch. 5)

There are different kinds of transcripts. Some transcripts are designed to only represent talk. Other ones try to integrate information about talk and gestures. Some other ones might focus exclusively on non-verbal interaction. Linguistic ethnographers often produce an annotated transcript, that is, a text where the representation of talk is enriched by contextual information that is relevant to talk or makes it meaningful.

Duranti's main points:

(i) transcription is a selective process, aimed at highlighting certain aspects of the interaction for specific research goals;

(ii) there is no perfect transcript in the sense of a transcript that can fully recapture the total experience of being in the original situation, but there are better transcripts, that is, transcripts that represent information in ways that are (more) consistent with our descriptive and theoretical goals;

(iii) there is no final transcription, only different, revised versions of a transcript for a particular purpose, for a particular audience;

(iv) transcripts are analytical products; that must be continuously updated and compared with the material out of which they were produced (one should never grow tired of going back to an audio tape or video tape and checking whether the existing transcript of the tape conforms to our present standards and theoretical goals);

(v) we should be as explicit as possible about the choices we make in representing information on a page (or on a screen);

(v1) transcription formats vary and must be evaluated vis-a-vis the goals they must fulfil;

(vii) we must be critically aware of the theoretical, political and ethical implications of our transcription process and the final products resulting from it;

(viii) as we gain access to tools that allow us to integrate visual and verbal information, we must compare the result of these new transcription formats with former ones and evaluate their features;

(ix) transcription changes over time because our goals change and our understanding changes (hopefully becomes "thicker,") that is, with more layers of signification.

Recall this example from Marcyliena Morgan's book Speech Communities, p. 94

  Morgan transcription

Morgan chose to highlight only a few non-verbal signs to make her argument about the meanings of this conversation that were indirectly implied rather than directly stated (eg., noting instances of laughs, elongated vowels, eye contact, head shaking, and loud talking. Meanwhile, the numbered lines make it easier to refer readers to specific parts of the interaction.


To Try Your Hand at Transcribing:

Make sure to number your lines (either by using your word processor's number lines feature, or by manually numbering the utterances). For ease of reading, make sure to break up the interaction by speaker and utterance. Start a new speaker on a new line, and new utterances of one speaker on a new line. Select your transcription symbols from this chart adapted from Jefferson and Gumperz and Berlenz:




[ text ]


Indicates the start and end points of overlapping speech.


Timed Pause

A number in parentheses indicates the time, in seconds, of a pause in speech.



A brief pause, usually less than 0.2 seconds.



Indicates falling pitch.


Question Mark

Indicates rising pitch.



Indicates a temporary rise or fall in intonation.



Indicates an abrupt halt or interruption in utterance.



Indicates that the enclosed speech was delivered faster than usual for the speaker. (G and P)



Indicates that the enclosed speech was delivered more slowly than usual for the speaker. (G and P)



Indicates that the enclosed speech was delivered at a higher pitch than usual for the speaker. (G and P)



Indicates that the enclosed speech was delivered at a lower pitch than usual for the speaker. (G and P)

* word


Indicates prominence or emphasis on the syllable (G and P)

** word


Indicates extra prominence or emphasis on the syllable (G and P)

~ word


Indicates fluctuating intonation over a word or syllable (G and P)

° word

Degree symbol (option 0)

Indicates whisper or reduced volume speech.


Capitalized text

Indicates shouted or increased volume speech.



Indicates prolongation of a syllable.



Audible exhalation



Audible inhalation

( text )


Speech which is unclear or in doubt in the transcript.

(( italic text ))

Double Parentheses

Annotation of non-verbal activity.