Discourse Functional Transcription (DFT) is a system for transcribing natural language discourse developed by John W. DuBois (Department of Linguistics, University of California, Santa Barbara). It consists of two components:
-
a data format for representing transcripts in human- and computer-readable form
-
a set of transcription conventions for representing various aspects of speech and its context
This repository contains specifications for formatting data in DFT, and the set of transcription conventions it uses.
DFT is the successor to two earlier versions of this system—DT1 and DT2 (where DT = Discourse Transcription).
This repository contains specifications for formatting data in the DT1, DT2, and DFT systems, and the set of transcription conventions used by each. It uses a form of semantic versioning to track changes to the DFT specification, where DT1 is considered v1.0, DT2 is v2.0, and DFT is v3.0. Each new version release may be viewed on the releases page.
| System | Version |
|---|---|
| DT1 | v1.0 |
| DT2 | v2.0 |
| DFT | v3.0+ |
For more information on DT1 and DT2, see the following sources:
This project uses Zenodo to publish the code in this repository with a citable Digital Object Identifier (DOI). Click the DOI link below to cite this repository.
To cite the latest version of the data format specifications in this repository, you may use the following bibliographic model:
John W. DuBois & Daniel W. Hieber. (2017, December 30). digitallinguistics/DFT. Zenodo. https://doi.org/10.5281/zenodo.1134007
You can also cite specific versions of the specification (if you want to refer to the DT1 format, for instance), by selecting the version on Zenodo and copying its citation:
To cite data from the Santa Barbara Corpus (SBC), use the citation guidelines found here.
If you see any issues in the specifications, or have any questions, please open an issue.
Please see the license for this repository to view the licenses for different parts of this project.