An information session for University of Cambridge researchers interested in submitting a proposal for the Language Sciences Research Incubator Fund. Led by Professor Paula Buttery.

The Chaozhou dialect is a branch of Southern Min Chinese with eight tones and a wealth of tone sandhi. In this paper we explore whether there is a tone-sandhi effect on melodic construction and tone realisation in Chaozhou song, using a corpus analysis and observational study. Outcomes from the corpus analysis show a strikingly higher rate of tone-melody matching in Sandhi dataset than that in Citation dataset. In the observational study, we found significant differences between sandhi form and citation form concerning tones /53/ and /21/, but no significant difference for tones /35/ and /213/. Results suggest that falling tones in the final position of a phrase tended to exhibit a larger contoural range, and that tones in non-final positions may be more affected by the pitches of tones that precede or follow them.

Face masks can cause speech processing difficulties. However, it is unclear to what extent these difficulties are caused by the visual obstruction of the speaker’s mouth or by changes of the acoustic signal, and whether the effects can be found regardless of semantic context. In the present study, children and adults performed a cued shadowing task online, repeating the last word of English sentences. Target words were embedded in sentence-final position and manipulated visually, acoustically, and by semantic context (cloze probability). First results from 16 children and 16 adults suggest that processing language through face masks leads to slower responses in both groups, but visual, acoustic, and semantic cues all significantly reduce the mask effect. Although children were less proficient in predictive speech processing overall, they were still able to use semantic cues to compensate for face mask effects in a similar fashion to adults.

The evolution of the capacity for language remains a contested and important subject. Newer approaches consider language evolution to be a protracted and mosaic process, with selection operating in different contexts through multiple drivers. This research examines two potential factors in language evolution: cognitive-structuring functions of language, and non-arbitrariness (iconicity). It is based on the hypothesis that language exerts powerful facilitative effects on cognition, which may provide an additional adaptive advantage to early proto-language. Iconicity meanwhile may offer a foothold for the emergence of semantics by providing an innate link between sound and meaning. This study investigated both aspects through an online game in which participants had to categorise novel species of aliens. Results provided evidence for both cognitive structuring due to language, and iconicity, adding credence to the suggestion that both may have played important roles in language evolution.

We develop a web-application for practice of listening skills by learners of English, which allows users to listen to pre-recorded sound-files and respond to multiple-choice questions. They receive feedback as to the accuracy of their responses, and they are navigated through the set of items in one of two ways according to the group they are randomly assigned to. Members of a control group are guided from one item to the next depending on success or failure with each new item, and the difficulty ratings of the remaining items. For the experiment group, item selection is made in an adaptive fashion: selecting items through automatic predictions based on individual performance and observations of other students’ interactions with the platform, as well as known item attributes obtained through tagging. Based on the cognitive literature, we also provide listeners with the option of controlling the speed of presentation of the listening items.

Accents, along with other cultural features including shared place of origin, helped to increase the number of people with whom an individual could signal cooperative tendencies (Cohen, 2012). Yet as groups became larger and underwent continued fission and fusion, signals of group membership may have become more important to reduce the risk of infiltration (Foley, 2004). We would expect, as the risk of imposters grew along with group size, for signals of group membership to become more complex, and for true group members to become adept at recognising false signals (McElreath et al., 2003). Here we are exploring how well people who speak naturally in 7 specific regions of the British Isles detect mimicry of their own accent. Our findings suggest that individuals are better than chance at detecting accent-mimicry of their own native accents, supporting this evolutionary account.

'Documenting the endangered Neo-Aramaic dialects of Iraq and Iran’, Geoffrey Khan Faculty of Asian & Middle Eastern Studies, University of Cambridge Aramaic, a Semitic language, has a documented history of over three thousand years. The earliest inscriptions are datable to the beginning of the 1st millennium BCE and the language is still spo-ken today in several ‘language islands’ in various parts of the Middle East. The main focus of my research on vernacular Neo-Aramaic is on the subgroup known as North-Eastern Neo-Aramaic (NENA). This subgroup contains over one hundred dialects, which are spoken by Christian and Jewish minority groups in northern Iraq and western Iran. Most of the dialects are now highly endangered. They have been in contact with other languages in the region for many centuries, in particular Iranian languages such as Kurdish and Go-rani. The NENA dialects exhibit a fascinating convergence with the Iranian languages. This reflects not only processes of language change but also the history of the speech communities. 'Language under the shadow of another language: implications and revitalisation strategies for Runyakitara and So languages', Fridah Katushemererwe (Makerere University, Uganda) Eberhard, et al, (2021) considers levels of language endangerment as a continuum. On the one hand, there are languages which are categorized as vigorous and may even be expanding in terms of numbers of speakers or functional areas of use, but exist under the shadow of dominant language(s). On the other hand, there are those languages that are on the verge of extinction because of loss of speakers. In between these two extremes are many degrees of greater or lesser language vitality. Although languages are endangered differently, current efforts in documenting, preserving and revitalising endangered languages of the world have been dedicated more to languages which are on the verge of extinction, giving limited attention to more vigorous languages. Based on this observation, this study presents evidence of varying degrees of language endangerment from two languages of Uganda namely 'Runyakitara' which is categorised as educational, and the 'So' language, which is categorised as moribund. The study answers three questions. What does it mean for a language to exist “under the shadow” of another language? What is the likely impact to a language that survives under the shadow of another bigger language? What approaches to language revitalisation are appropriate for such languages?

'Dyslexia as a Window into Language', Maria Teresa Guasti Università di Milano-Bicocca Recent research has uncovered deficits with rhythmic processing in children and adults with developmental dyslexia (DD) and an association between these deficits and reading. Other studies show a link between rhythm perception and grammar, and that children with DD often display language problems beyond weaker phonological skills. The literature also reveals that children with DD experience fine and gross motor problems. This research reveals a comorbidity of different deficits in individuals with DD, which seem to have a common thread. Reading, language, and motor activities are all activities that unfold in time and in which the single acts are interdependent. As such, they all involve “co-articulation” in a broad sense; that is, what one does at time N is influenced or somehow linked by what one has to do at time N+1. In other words, to co-articulate, one needs to be in an anticipatory or predictive modality; that is, she must be ready to act in the single precise moment WHEN it is required. But one can anticipate/predict only if rhythmic principles regulate the behaviour. In our view, rhythm is a key to understanding what goes awry in individuals with dyslexia. We propose that a deficit in the anticipatory mechanism impairs reading, some motor activity, handwriting, rhythmic processing and language. We provide evidence that individuals with dyslexia have anticipation/prediction deficits, explaining why these deficits affect reading. Anticipation skills allow us to deal with timing and require a hierarchical organization, as we find in language and motor activities. We speculate that anticipation is a mechanism that has been recruited by language to linearize our internal thoughts, which must be hierarchically organized. 'Beyond the label: A transdiagnostic approach to understanding cognitive difficulties in childhood', Duncan Astle, MRC Cognition & Brain Sciences Unit, Cambridge As our cognitive skills develop differences gradually emerge between individuals. Some of these differences can act as barriers to learning, for example phonological processing skills and executive functions have both been linked to difficulties in reading and maths, respectively. But how much are our conclusions influenced by who we study? The traditional approach to studying neurodevelopment difficulties is to recruit children with a diagnosis, or screen children according to a particular diagnostic standard. What could be learned with a more inclusive recruitment strategy? We collected a transdiagnostic sample that aimed to capture the broad mixed population of children in the community at neurodevelopmental risk. Using a simple machine learning approach we were able to identify the different cognitive profiles that exist within this cohort, including a large subgroup of children with more selective difficulties in phonological processing. These difficulties extend to all tasks that require phonological decoding – such as phonological decoding, verbal short-term or working memory. These cognitive differences generalise. Their parents also report substantial difficulties with the structural elements of language, like syntax. However, this data-driven mapping has no overlap with formal diagnostic status. We next explored the neural mechanisms that give rise to these differences in cognitive profile, by creating structural neuroimaging to explore study different features of macroscopic brain organisation. We subsequently developed a computational framework, using generative network modelling (GNM), to model these emergent differences in brain organisation. Relatively subtle changes within the wiring rules of this computational framework give rise to differential developmental trajectories, because of small biases in the preferential wiring properties of different brain regions. Finally, we were able to use this GNM to implicate the molecular and cellular processes that govern these different growth patterns.

There is no doubt that learning new languages is infuriatingly difficult, especially at the later stages of life. As the world becomes "smaller" through globalisation, certain languages begin to increase in utility and start taking precedence over others, resulting in the extinction of the less "useful" languages. According to McWhorter (2009), in the next 100 years, the 6,000 languages in use today will be reduced to about 600. Whether and how to save these endangered languages is an important question plaguing the language sciences community. We are now in the age of information and artificial intelligence. All the data we need is available in the palm of our hands. Mobile applications like Babbel and Duolingo lower the barriers to entry when it comes to learning a new language. So why is language learning still so difficult? Haven't the plethora of philosophical thought experiments, cognitive theories and neuroscience research combined with the scale and reach of modern technology enabled us to make language learning as easy and intuitive as playing a video game? Can we not use this technology to then increase the number of speakers for endangered languages? The answers and further questions lie in the history of personalised learning and the underlying principles and paradigm shifts that have shaped it over the centuries. The nature of knowing represented through contemporary theories of learning such as behaviourism, cognitivism, and constructivism have provided some insight into the question of how we learn. In this talk, I will walk through a brief history of learning that will shed some light on not only the question of whether artificial intelligence can save endangered languages, but also whether it can play a role in making language learning less difficult.

Over time, new words enter the language, others become obsolete, and existing words acquire new meanings. In ancient Greek, the meaning of the Persian loanword paradeisos expanded from ‘garden’ to the Jewish-Christian ‘paradise’ in the Greek translation of the Old Testament and in the New Testament. In Classical Latin passio meant ‘emotion’ and later on referred to the suffering and death of Christ and the martyrs. The English word chill originally meant ‘to cool’ and has metaphorically been extended to ‘to relax’. Follow only acquired the social media sense of staying informed about someone’s postings after the launch of Twitter. The phenomenon of lexical semantic change, with its fascinating complexities grounded in cognitive, social and contextual factors, has important implications not just for linguistic theory and historical linguistics. It can shed new light into how we understand long-term and short-term changes in our cultural history and in our society. It is also a fundamental aspect of dictionary-making and it is important to keep automatic language processing systems up to date with the constant changes in language. The recent digitization efforts have now made it possible to access and mine digital collections of historical texts using automatic methods and investigate the question of semantic change over centuries. Easy access to very large born-digital collections from the web also allows us to study changes in contemporary language data spanning short time periods. In this talk I will present my research on developing models for semantic change drawing on state-of-the-art computational linguistics methods relying on distributional semantics principles, Bayesian learning and embedding technologies. I will share my experience of working at different scales and in a range of interdisciplinary projects, from Ancient Greek and Latin to Charles Darwin’s letters, web archives, Twitter and emoji.

There is a pressing clinical demand for tools to predict individual patients' disease trajectories for schizophrenia and other conditions involving psychosis, however to date such tools have proved elusive. Behaviourally and cognitively, psychosis expresses itself by subtle alterations in language. Recent work has suggested that Natural Language Processing markers of transcribed speech might be powerful predictors of later psychosis (Mota et al 2017, Corcoran et al 2018), for example, Corcoran et al 2018 used quantitative markers of semantic coherence collected at baseline from individuals at clinical high risk for psychosis, to predict transition to psychosis with 79% accuracy. However, it remains unclear which NLP measures are most likely to be predictive, how different NLP measures relate to each other and how best to collect speech data from patients. In this talk, I will discuss our research tackling these questions, as well as the wider challenges of translating this type of approach to the clinic. Ultimately, computational markers of speech have the potential to transform healthcare of mental health conditions such as schizophrenia, since they are relatively easy to collect and could be measured longitudinally to quickly identify changes in patients' disease trajectories.