Disciplines, specialization and interdisciplinarity in the social sciences and humanities

Is the unity of the sciences threatened to be lost by increased specialization? Do we think too much in boxes? Is there a need for more interdisciplinarity? Although interdisciplinarity has occupied a central place in policy discourse for some time now, a parameter for it will soon also be added in the allocation model for Flemish research funding. The central idea put forward to substantiate this choice is that the complex problems we face as a society can only be solved when knowledge from different disciplines is combined. Ecological and social crises, such as human-accelerated climate warming, political extremism, or the multifaceted problems of a global pandemic, require an integrated approach, which is difficult to offer by a single discipline, it is said. In the first two chapters of this thesis, we first examine scientific specialization and the changing position of research disciplines in science in general. Due to an increased internal and external differentiation, the disciplinary structure of the scientific system is subject to change. For example, internal differentiation leads to the emergence of various specialisms, which in turn continuously explore the boundaries of disciplines. Within traditional disciplines, such as psychology or economics, we can note the emergence of countless subdisciplines and smaller-scale specialisms as a result of the enormous growth of the scientific system. These often arise around new subjects, theories or methods around which or with which research is carried out. But new specialisms are also created as a result of impulses external to the science system. The 'scientificization' of social novelties or emerging research into the rapidly developing social subsystems leads to the emergence of domains that have no direct or just several disciplinary predecessor(s) (e.g. new media studies, scientometrics, etc.). These specialties function as the primary reference group for researchers, and as numerous case studies demonstrate, they are interdisciplinary by definition. We also pay attention to transdisciplinarity. How does increased external differentiation lead to collaborations with actors outside the academic system? The main lesson that emerges from these introductory chapters relates to research on interdisciplinarity as well as research evaluation. We point out that a revaluation of the categories 'discipline' and 'specialization' is necessary within bibliometric research. History teaches us that increased specialization and differentiation makes the disciplinary-interdisciplinary opposition irrelevant. Following this, guidelines for research evaluation related to interdisciplinarity are summarized. The second part of this thesis deals with the bibliometric operationalization of specialties in the social sciences and humanities (SSH). Common methods are often based on citation analyses, but for the SSH this is a problematic approach in several respects. In search of an alternative, we therefore reconstructed a fine-grained research classification at the level of subdisciplines and specialisms using machine learning techniques and textual data from discipline-specific databases (Sociological Abstracts, EconLit, and ERIC – Education Resources Information Center). At the same time, we had experts classify a series of scientific publications from the same set and then studied the degree of consistency of the assigned categories. This showed that the implementation of a classification at the level of research specialties is a difficult task, both in the case of algorithmic approaches based on machine learning and for domain experts from the involved disciplines. In a third study, we look at the results of cluster analysis to group documents based on textual similarities. The conclusion is that recently introduced vectorization techniques based on neural networks can offer promising results. On the basis of a comprehensive bibliographic database for the Flemish SSH (VABB-SHW), we study disciplinarity and interdisciplinarity in the third and final part. The two empirical studies presented address the interdisciplinary mobility of researchers and identity of research specialisms, respectively. In accordance with what we suggest in the introductory chapters, we can only speak to a very limited extent (or none at all in the case of the specialisms) of strictly disciplinary identities or profiles for both the researchers and the specialties. Researchers in the SSH, for example, are increasingly publishing in a variety of disciplines. We also show that the same researchers do not only publish across discipline categories that are very similar in content. We also note that for those who publish across a large number of different categories, there is only a limited degree of cognitive mobility. It seems that this disciplinary mobile - or interdisciplinary group of researchers is in fact conducting hyper-specialized work. For the research specialisms in the SSH in turn, we note that no specialism is strictly disciplined. Based on contrasting the disciplinary diversity in terms of author affiliation and the classification of research output, we establish a typology of different forms of interdisciplinary specialisms. Concrete recommendations for bibliometric research into interdisciplinarity, more specifically in the context of the SSH are formulated, as well as a series of considerations that may be relevant for those in charge of research evaluation or policy.

Publication year:2022

Keywords:Doctoral thesis

Accessibility:Open