Speech remains the method of choice for human communication, and as a consequence live, recorded and synthetic speech is increasingly used to deliver information. However, current speech output technologies lack an essential element of human interaction, namely the ability to listen while talking. Talkers react and adapt to both the environment and the listener, and make modifications to their speech in order to promote intelligible and fluid communication. For example, talkers modify fundamental frequency, spectral tilt and duration in response to noisy conditions, and produce more prototypical speech sounds when speaking clearly, resulting in increased intelligibility. Speech output technology, in contrast, does not listen while talking, so there is no guarantee that the message is understandable in a given environment and by specific listeners. The purpose of LISTA -- the LIStening TAlker -- is to develop scientific foundations for spoken language technologies based on human communicative strategies. LISTA will investigate how talkers react to changes in the environment, measure the relative success of these strategies using behavioural studies and models of speech intelligibility/quality and study algorithms for rapid characterisation of the listening context. The central objective of LISTA is to apply this information to develop novel techniques for spoken output generation of artificial and natural speech. LISTA will have an impact in all situations where synthetic speech is embedded in devices such as computers and PDAs, mobile and fixed telephones, public information kiosks and information systems, in addition to the domains where live speech is employed e.g. public address systems and teleconferencing. In all these applications, the ability to increase intelligibility in noise has an immediate value, not only in terms of message comprehension but in reducing overall time-of-interaction, noise pollution and individual noise exposure.