Understanding the relationship between speaker's attitudes and linguistic behaviour across speech communities

There are in excess of 60 regional/minority languages in Europe alone, spoken by over 50 million people. Most suffer from some degree of endangerment, and recurring issues of language maintenance arise in different yet comparable language communities. Central to these issues is the need to monitor and understand linguistic attitudes, which are known to be a potential predictor of linguistic vitality.
However, the reliability with which linguistic attitudes can predict vitality depends on the type of measurement employed. Attitudes measured explicitly (e.g. self-reports) can be unreliable, and may only have an additive role compared to attitudes elicited via implicit methods where participants are unaware of the research aims. Evidence is also emerging that implicit methods as traditionally defined may not tap implicit attitudes as reliably as previously assumed.

This project applies interdisciplinary, cross-methodological measurements to gain a deeper understanding of linguistic attitudes across three different communities with varying degrees of language recognition, namely Lombardy (Italy), Eifel (Belgium) and Wales (UK). As the success of language maintenance hinges on understanding the link between speakers' attitudes and speakers' behaviour, we investigate the reliability of different attitudinal measurements in predicting speakers' language choices. Employing a range of methodologies varying from declarative (self-reports) to fully unconscious (neurophysiological data), via behavioural measurements with low participant awareness (reaction time tasks), we investigate a hitherto unstudied spectrum of attitudinal domains. To achieve this, the following data will be collected:

(i) Fully explicit attitudes, using a declarative method (self-report questionnaire).
(ii) Semi-implicit attitudes, using an established method where participants are only partly aware of the linguistic aims. As the implicitness of this method has been called into question, these data serve as a reference point of comparison with the innovative methods in (iii) and (iv) below.
(iii) Implicit attitudes using modern behavioural measurements (reaction time tasks) with low levels of participant awareness.
(iv) Fully implicit attitudes, measured via unconscious neurophysiological responses.
(v) To establish the reliability of each method (i)-(iv) in predicting linguistic behaviour, participants will also take part in a behavioural choice task which requires them to choose between using their regional/minority language and their majority language.

While some previous studies measured linguistic attitudes using more than one method - typically some variation of (i) and (ii) above - the full range of measurements described in (i)-(iv) is unprecedented in this field, as is the analysis of results in relation to behavioural choice. This project is also the first to apply the technique in (iv) to the investigation of linguistic attitudes.
To cater for different sociolinguistic contexts, we collect data from three communities with different degrees of language recognition: (1) Lombardy, where Lombard has minimal recognition as a regional cultural expression, but no state recognition or official use; (2) Belgian Eifel, where Franconian has some indirect recognition via the administrative and educational use of German (a related language) but no direct recognition or institutional use; and (3) Wales, where Welsh has equal status to English in public life (Welsh Language Act 1993) and benefits from a host of policies for its inclusion in all aspects of life.
The cross-methodological design, the inclusion of behavioural choice measures and the involvement of three different speech communities allow us to achieve a deeper understanding of the feedback loop between socio-political recognition and linguistic attitudes, and provide novel insights on the role that different types of linguistic attitudes may play in predicting actual linguistic behaviour.