Profile & Ambition
The main goal of the LLI research group is to study the linguistic representation of information. The main linguistic sub-domains to which the work in the group belongs are semantics and computational linguistics, with applications in human language technology (HLT) and cross-overs with artificial intelligence and cognitive science. The deeper understanding of the various ways in which natural language encodes information is assumed to contribute to more effective applications of HLT and to better methodological frameworks for the processing of language as social and cultural data. There is a strong focus on multidisciplinary collaboration, which increases the potential for generating insights that have impact beyond academia.
To achieve the aforementioned goal, the research in the group takes a computational and logical perspective. Two themes underlie the main research effort – lexical meaning and the meaning/use interface. The group embraces an integrative approach to natural language understanding from both a grammatical perspective (meaning) and an extra-grammatical perspective (use and context). As part of its research efforts, the group develops linguistic resources and tools supporting language technology. These tools and resources are embedded in international and national research infrastructures, which are also managed and maintained by group members.
The study of lexical semantics in the group focuses on how information is organized in natural language lexicons and the contribution of lexical meanings for communication. This includes work on multi-word expressions (Odijk), spatial meaning (Matushansky, Zwarts, Winter), proper names (Bloothooft, Matushansky) and lexical meanings in compositional semantics (Moortgat, Winter). The motivation of understanding thought and communication also underlies much theory-driven research on Semantics & Pragmatics. This includes semantic and pragmatic work on degree expressions (Nouwen), determiner systems (Le Bruyn, de Swart, Nouwen, Matushansky, Zwarts), tense and aspect (de Swart), quantification and the syntax-semantics interface (Dimitriadis, Moortgat, Ruys, Winter) and use of language in social media (Monachesi). This research involves meaning representations and their variation across languages, which are studied using experimental methods, corpus analysis and computational modelling.
One target group of the work in the LLI research team is international specialist peers in semantics and pragmatics. Typical ways to address these peers are publications in highly qualified international journals, refereed volumes and books, (partially) refereed conference proceedings, as well as organization of workshops and colloquia that present the group’s output.
Another target group of the LLI team includes researchers, practitioners and users of natural language technology, in Linguistics and other parts of the Humanities. In addition to academic publications as stated above, the team addresses these groups through computational tools and data resources which are communicated through collaborated work in research infrastructures and which enable the reuse of data and the uptake of computational tools for the processing of language materials.
Yet another target group of the LLI team consists of a range of societal partners and stakeholders. This target group is served by tools and data resources, as well as by knowledge and insights gained by applying HLT to their data collections. The societal partners and stakeholders in this target group include heritage institutions such as the Dutch National Library, the Netherlands Institute of Sound and Vision, and other members of the Galleries, Libraries, Archives and Museums (GLAM) sector. In addition, the group also forms links to commercial companies with an interest in technology for the automated extraction of information from textual data (text mining).
Since its formation in 2012, the group’s ambition has been to contribute to the interdisciplinary study of information representation in language, at the cross-roads of linguistics, philosophy and artificial intelligence. An underlying guideline in this ambition is the use of logical methods. Both in the theoretical analysis of information and in its computational modelling, the group aims to accommodate logical principles as a common core of computational tools for treating meaning and information. This use of logic is further geared towards generating linguistic resources and tools. These outputs are employed for information processing, providing feedback to linguistic theory, and to academic and societal parties. The latter consist of partners and stakeholders are interested in developing and improving human language technologies: the data science industry, digital humanists who work with heritage data, and developers of support tools for language learners or people with speech disabilities.
The group seeks to materialize this ambition using international publications and big research infrastructures, which are also relevant for the (Digital) Humanities in general. Publications include scientific books, journal articles and reviews, book chapters, edited volumes and special issues, all addressing the target groups of international peers. Through big research infrastructures (RIs), the LLI group makes substantial collaborative contributions to the design, implementation and management of computational tools, datasets and technologies, which embed research outputs using linguistic resources, and guarantee that these outputs are made available to a large research community. These RIs serve the needs of researchers, but also play a crucial role in the education of the next generation of researchers and the promotion of digital scholarship. It is also to the benefit of society that LLI students are offered the opportunity to develop skills that are on high demand for work in the data science industry.