We wanted to give a little more information on how we are thinking about cyberinfrastructure, since this term has proven to be rather vague.

A cyberinfrastructure for linguistics:

* facilitates both the creation and sharing of data, where one linguist's analysis becomes the next linguist's data

* supports and connects all subfields of linguistics

* supports collaboration across space and time

* is not a single research project, nor a single set of tools, but rather technological and organizational infrastructure supporting massive interoperability, including standards, apis, and committee structures

* requires buy-in and participation from a broad cross-section of the field; "build it and they will come" won't work here

To build a cyberinfrastructure, we need to solve at least three large problems:

* the culture change problem: getting linguists on board with sharing data, verifying analyses against more data, and participating in the design of the infrastructure

* the funding model problem: finding a funding model which can incorporate fundamental research in computational linguistics, software engineering (including UI design), and long-term maintenance of standards and software tools

* the design problem: designing the cyberinfrastructure, including both committee structures and technical standards