| TalkBank | 
| Ground Rules |
Researchers who use TalkBank data and programs are asked to abide by these rules:
Citation: When using TalkBank corpora in published articles, you need to cite your sources by following these rules.
Contributing: If you collect new corpora, please contribute them to TalkBank or CHILDES and urge your colleagues to contribute new corpora. This is key to the growth of the database. Make sure that your informed consent forms and IRB review forms explicitly mention that the data will be included in TalkBank.
Copyright: Except where otherwise indicated, the use of TalkBank data is governed by the Creative Commons CC BY-NC-SA 3.0 copyright license. This license precludes the use of the data in commercial products, including systems such as large language models (LLMs) such as ChatGPT. Commercial enterprises that have been allowed access to password-protected data can use the data for the development of algorithms, such as for automatic speech recognition or clinical assessment, but the data themselves cannot be included in models. In addition, users cannot upload TalkBank data to web-based systems unless those systems include assurance that they will not keep the data. In particular, data can only be uploaded for processing by OpenAI, Rev-AI and similar systems if the upload is done through an API which explicitly allows for non-storage of data.
Ethics: You must read and follow the TalkBank Code of Ethics. which deals with issues of respect for participants and contributors and maintenance of confidentiality.
Password Access: Access to the password-protected clinical databases (AphasiaBank, DementiaBank, RHDBank, TBIBank) is restricted to faculty members. Students who wish to use these data should ask their faculty advisors to become members to provide them with access. Once that is done, the faculty advisor becomes responsible for making sure that the advisees read and follow the Ground Rules. Password-protected data cannot be posted on other websites or shared with anyone who does not already have password access.
HomeBank: Access to password-protected data in HomeBank is even more tightly restricted. For details on how to gain permission to use these data, consult this page .
Membership: You should become a member of the TalkBank system by contacting Brian MacWhinney (macw@cmu.edu) and sending him your academic affiliation, postal address, and email address. Also, please describe how you plan to use the data and whether you will eventually be able to make your own contribution to the database. As noted above, students who wish to use password-protected data must be supervised by their faculty supervisors who will join as members.
Google Groups: We maintain separate membership lists and datasets by interest groups. You can find information on how to subscribe here.
Teaching: Researchers who abide by the above rules are welcome to use the programs and the non-password-protected database freely in their classes and teaching. The use of password-protected data in teaching must be carefully monitored, since those data cannot be further distributed. Because they were student projects, the files in the "CMU" folder in CABank can be used for teaching purposes, but not for research purposes.
Local Installation: The programs and non-password-protected data can be freely installed on university computer systems. However, password-protected data cannot be installed in this way or posted elsewhere.