Learning to read is the most important outcome of primary schooling.  It brings knowledge, employment, and prosperity to an individual; high levels of literacy at a societal level also contribute to poverty reduction and economic growth.  This PhD studentship will focus on some aspect of how children learn to read but will be tailored to the student’s interests and expertise.  Projects may be focused on English but may also focus on writing systems that differ from English in important ways (e.g. Chinese, Arabic, Hebrew).  Likewise, while projects may be focused on learning to read in high-income, largely monolingual settings such as England, it would be possible to develop a project tackling the special challenges of learning to read in a low- and middle-income country, particularly where a child may be learning to read in a language that they do not know.  This project will be conducted with appropriate partner organisations (e.g. schools, literacy charities, international development agencies), and the second supervisor will be selected according to the particular focus of the project.

The last five years has seen an explosion of progress in building large language models such as Alexa and Siri able to mimic aspects of human language behaviour.  However, there has been very limited interaction between the engineers and computer scientists building these models and psychologists studying aspects of human language understanding. This project will focus on some aspect of the interface between large language models (typically used for engineering applications) and human language understanding but will be tailored to the student’s interests and expertise.  One avenue would be for students to uncover the similarities and differences between large language models and human language understanding; for example, how do humans and large language models generalise their knowledge to unfamiliar situations?  The other avenue would be for students to investigate how large language models can be used to improve the accessibility of text for individuals with low literacy or for children who struggle to learn to read; for example, through the use of summarisation tools or by parsing text automatically into meaningful elements.  These would both be cross-disciplinary projects conducted in collaboration with experts from the Department of Computer Science.  Students would require very strong quantitative and programming skills. 

The intuitions of organisations managing adolescent student exchanges are that this experience leads to later promotion of diversity and inclusivity, with supportive anecdotal evidence. Our initial longitudinal study suggests a more complex picture. The project will be informed by our initial findings and use our connections with such organisations to examine the process of change throughout and after the experience, focusing on value and behaviour change, using longitudinal surveys and possibly interviews and intervention studies. Excellent achievements in prior statistics training is required, and fluency in another European language is an advantage.

Individuals learn language from their environment. Therefore, individual differences in the social environment can influence linguistic skills. Indeed, prior research has shown that individuals with larger social networks have superior linguistic skills, such as better understanding of speech in noise (Lev-Ari, 2018), better comprehension of product reviews (Lev-Ari, 2016), better ability to communicate with others (Lev-Ari & Sebanz, 2020) etc. Individuals’ social networks, however, might change during their life-time. For example, individuals’ social networks often shrink upon retirement or for health reasons. It is unknown whether such changes can lead to deterioration in linguistic skills or whether accumulated experience throughout one’s lifetime can guard against deterioration. This project will use longitudinal and experimental methods to explore how changes in social network structure influence linguistic skills. 

Social norms are behavioural guidelines providing expectations for individual choices and inferences about others’ likely future behaviour. This way, social norms can be thought of as simplifying heuristics to help decision-makers to navigate the world. Some norms are strictly social and regulate how individuals behave in social settings (e.g., dress code); others refer to a moral or religious dimension and rely on societal beliefs (e.g., life is sacred and humans should be used as an end and not a means); still others represent more practical advice about what should be done that turns out to have some biologically rational (e.g., food prohibition). Adherence to these norms can be motivated by societal expectations or enforced through governmental measures. The prevailing literature primarily concentrates on treating social norms as a collective entity or delving into specific facets of these norms. This project makes a distinctive contribution by investigating the dynamics of norm adoption within a social group from an individual perspective.

The general question is: What does it take for a social norm to get established?

Different aspects are relevant to this question: How beneficial should a behaviour be for a social group to be adopted by its group members? How easy should its adoption be for the social members to be able to adopt it? How many social members need to agree with adopting that behaviour for it to become the norm? Is a belief change required for a behaviour change or does a belief change simply facilitate its establishment in a society?

This project will be both theoretical and experimental, will involve the development of experimental tasks in combination with computational modelling analyses, and will investigate the questions above in relation to group norms in general and in particular (e.g., environmental concerns and behaviour change for more sustainable living).