The TalkBank Project at Carnegie Mellon University has received this certification as a CLARIN-K Knowledge Centre for Language Learning Analysis (CLARIN-Learn). We seek to provide resources, methods, and advice to assist researchers interested in using corpora to study language learning, recovery, and processing. We hope that, through this process, researchers will create new corpora and new ways of analyzing language.

This CLARIN blog post describes the overall functionality of TalkBank. Further information about other CLARIN K-Centres is available here.

The TalkBank centre is happy to provide these services:

  1. We can provide advice on the use of the corpora in the 14 component databanks in TalkBank, including AphasiaBank, BilingBank, ClassBank, CHILDES, CABank, DementiaBank, FluencyBank, HomeBank, PhonBank, RHDBank, SLABank, SamtaleBank, and TBIBank. The homepage for TalkBank provides links to each of these databanks, as well as other materials.
  2. Once you arrive at one of the databanks, you can click on "**Index to Corpora**" to locate particular corpora by language or you can click on "Browsable Database" to directly browse from file to file.
  3. Databanks such as CHILDES, CABank and ClassBank have completely open access. However, clinical databases such as AphasiaBank require passwords, which are readily given to researchers by sending a request to
  4. Because many of these corpora include transcripts linked to audio or video, we can also provide help on methods for multimedia and gesture analysis.
  5. We can provide pointers to relevant corpora outside of TalkBank, particularly including those at other CLARIN centres.
  6. We can provide methods for allowing researchers to contribute new data to any of the TalkBank archives.
  7. We can provide help in the use of analytic tools with a focus on use of the CLAN, Phon, Praat, and ELAN programs. These tools all operate well on data from TalkBank.
  8. We can provide help in the use of R and R scripts for corpus analysis using the TalkBankDB search engine.
  9. In collaboration with Yvan Rose at Memorial University, we can provide help regarding details of phonological analysis with PHON and its integration with Praat.


Help requests can be sent to Brian MacWhinney at