A research project by Professor Jong-Bok Kim at the Department of English Linguistics and Literature was selected by the Ministry of Education as one of the fifty notable achievements in the 2023 academic research support project

The Ministry of Education announced the 50 top achievements of the Ministry of Education’s academic research support project for 2023. Out of 13,600 research projects in 21 areas supported throughout 2022, a total of 50 projects―26 in humanities and social sciences, 20 in science and engineering, and 4 in Korean studies―were selected as the best results. At Kyung Hee, two researchers were named: Professor Jong-Bok Kim at the Department of English Linguistics and Literature in the field of humanities and Professor Sujin Choi at the Department of Media in social studies. In this first installment of the series, I met Professor Kim and heard about the results of the selection and his impressions.

“Accumulation of years of effort and advancement in the field of international joint research”
Q. The announcement by the Ministry of Education means the government has recognized the excellence of your research. Can you tell us your thoughts upon hearing the news?
This research project was conducted as part of the Global Research Network project supported by the National Research Foundation of Korea as part of strengthening international exchange in university research. Its focus was on theoretical and experimental research on silent phrase processing, which was published in September 2023 in the Linguistic Review under the title “Pseudogapping in English: a Direct Interpretation Approach,” as the culmination of a three-year study from 2017 to 2020. The main goal was to examine and identify, through linguistic big data, how incomplete sentences including ellipsis, which are frequently found in English and other languages, are created and recognized. In other words, it is a global research project that examines how we recognize and process silent expressions that are assumed and not uttered. Professor Jeffrey Runner of the University of Rochester, USA, was the co-researcher in this project, and scholars from France and Sweden also participated. Several results were published through this project in which multinational researchers participated, and the award-winning research result is one of them. It is a good example of global research collaboration.

Q. Please explain the concept of direct interpretative analysis of English pseudogapping.
The idea of pseudogapping is one of the elliptical phenomena in English that frequently appears in everyday life, as in “I could no sooner do without music than I could oxygen.” In addition to empirical research using corpus data, this was used to conduct theoretical analysis. In this study, we identified the grammatical similarities and differences among the three related and similar linguistic expressions―pseudogapping, VP-ellipsis (VPE), and gapping―and presented how we can correctly license the use of pseudogapping through a syntactic analysis.

To briefly introduce our result, the grammatical commonality between pseudogapping and VPE is that they inherit the syntactic properties of the omitted phrase, which is the macro-construction of the two. The difference is believed to be due to the unique syntactic characteristics of each as a micro-construction of the elliptical construction. This study provided discourse-based analysis using structured discourse structures to derive the meaning of omitted elements, suggesting the importance of interaction between grammatical information.

Big data-based linguistics research, building a foundation for international collaboration through global assignment
Q. What are the characteristics of this research?
Syntax-based analysis based on empirical research, such as this study, systematically explains various grammatical properties of pseudogapping in sentences that conventional movement & deletion analysis or base-generation analysis cannot adequately explain. Through the inheritance network between phrases, we have established an analytical framework that can easily identify not only pseudogapping and VPE, but also the grammatical relationship between various other types of ellipses.

By approaching the phenomenon of ellipsis, which has been studied individually so far, as a type of macro-construction of elliptic phrase, the commonalities shared by different types of ellipsis can also be explained. The differences between each ellipsis can be attributed to their respective micro-construction structure having its own unique syntactic properties. Meanwhile, generative syntactic research on these constructions has been extensively conducted based on movement and deletion. However, this is basically the first study from the perspective of the constraint-based theory of syntactic grammar, which gives us a sense of pride in having contributed to the relevant grammatical development. This pioneering study, which is based on linguistic big data, aims to serve as a role model for future empirical research using big data.

Q. What are the motivating factors behind the research?
The phenomenon of ellipsis holds a key position in understanding the nature of language in that it is a standard tool for optimization, or economy, of communication. In particular, ellipsis has become a major interest in modern linguistics research because it not only shows the close interaction between syntax, semantics, pragmatics, and information structure, but also has an atypical mapping relationship between syntactic and semantic structure. In this study, we studied various elliptical phenomena using big data, as we wanted to contribute to expanding the base of linguistic research using big data and the internationalization of humanities research in Korea.

The importance of global networks for international joint research is increasingly becoming more significant. To strengthen global networks, we must build a platform for regular academic exchanges to encourage internationalized research and assignments. It is necessary to build friendships at overseas academic conferences and form a consensus on potential research topics. This study can be an example of how internationalization of humanities can be implemented with positive results.

I have consistently pursued international collaboration over the years, as humanities research in Korea is predominantly limited to data collected within the country. Also, each individual researcher must take the step forward to find the way into making collaborative research with their global counterpart. In the case of this study, for instance, I did not personally know the researchers I worked with prior to this project; we met for the first time at an academic conference and decided ad hoc on a topic through discussion and conversation. I continue to organize international seminars with these groups.

Linguistics is a convergent field, and mathematical and scientific thinking is essential for a scientific approach.
Q. What research area are you usually interested in?
I started with linguistics as my main field, later moved to theory, and for the past 10 years I have mainly been doing empirical research using corpora and big data. The corpus used in this study is a database created by Professor Mark Davis of Brigham Young University, who was initially our team member. The size of the corpus was small ten years ago, but now it has become big data. The Large Language Model (LLM), which is being talked about a lot now, simply refers to linguistic big data. Linguistic big data has also changed significantly in size. Previously, it was in units of million (million) words, but now we start counting units in billions and trillions. The number of units handled by generative language models is rapidly increasing.

In simple terms, language can be described as a collection of signs. Each sign consists of a link or mapping between its form and meaning. These connections manifest in many forms, and research to identify them is a major area of interest in linguistics. In particular, the reconstruction or resolution method of silent verbal expressions, as exemplified in this study, is a major research field for identifying the relationship between form and meaning in linguistic expressions. These expressions are much more diverse and more frequently used than one might realize. For instance, if someone asks, “What do you like?” and you answer “Apples,” the questioner intuitively understands your response as “I like apples,” which means incomplete sentences are subconsciously reconstructed before being processed for meaning.

We are investigating this phenomenon theoretically through linguistic big data and applying it to AI or computer science. Linguistics is essentially an interdisciplinary field, and it must be approached scientifically, meaning doing linguistics research requires mathematical and scientific thinking. As it is in all academic fields, big data is the biggest topic at the moment. Linguistic research is also progressing in a direction that integrates natural science rather than pure humanities.

When you turn your attention to the field of education, it might feel unfamiliar at first. It is difficult to understand why quantitative methods are required in humanities. However, a quantitative method is needed to understand the reality of something. That is the trend of the times. You must have both humanistic and scientific thinking, knowledge, and analytical skills. Linguistics is the discipline where all of these talents and expertise come together.

Q. Please tell us about the goals of future research.
In connection with this project, we are currently conducting global research on the response system with scholars from Spain, France, Romania, and the Philippines. The response system is a linguistic property that appears essential in the performance of discourse in all natural languages. It is an essential research topic to understand the nature of language, whose basic purpose is communication. Previous research in the field of response systems was biased toward theory with little grounding in data. Our research, on the other hand, examines the various response systems used in languages with typological differences, including English, Korean, French, and Tagalog.

Our current research approach aims to maintain balance among empirical study using big data corpus, theoretical research with technical feasibility, and experimental investigation to verify the developed theory. With empirical and experimental research, I hope to contribute to a proper understanding of the response system that is essential in actual language use and to the identification of the essence of natural language. We have already produced several promising results.

There are also areas I would like to study further. I want to take a look at the commonalities and differences in human languages beyond English. The more you know, the more unknowns are discovered. I would like to continue working on it and write a book someday if I get usable results. The downside of generative language models is that they are not creative, as they merely reassemble blocks of existing data mathematically and computationally. Is this really as creative as humans? I want to explore its limitations, individual phenomena of language, and cognitive aspects of thinking.