[This blog post is written by Debarun Sarkar and Cheshta Arora]
In the last blog post, we stressed on principles that drove the project. Yet, at a more practical level, we were driven by the desire to intervene and play with semantic web technologies such as OWL2. This intervention we believe has produced interesting results, whose learnings are relevant to the disciplines of anthropology/sociology (we consider these as twin disciplines) as well as fields such as digital humanities. This blog post hopes to cover a question we are often asked: “Why OWL2?”. After all, behind the Audit4SG tool is the relational AI ethics ontology (RelAIEO).
In the field of anthropology/sociology, OWL2 and semantic web technologies with their open-world assumption can make significant contributions to scientific communication and knowledge-sharing of values and representations of reality in a machine-readable format. This can help the disciplines debate the objects or values under consideration via different re/presentations of knowledge. Such a move is not necessarily out of place as anthropologists/sociologists do not just study technologies but also use them. The use of audio records, personal computers, word processors, printers, spreadsheets, qualitative and quantitative data analysis software, reference management systems, and e-books have all shaped how we practice and present these disciplines. Tables and figures used to express ideas and representations in these disciplines could be communicated through semantic web technologies which could allow for the rapid development of ontologies that build on each other in however partial, consensual, and antagonistic ways. We explore and argue this in detail in a forthcoming publication.
The Audit4SG tool’s design can aid one to think in similar ways, wherein every user of the tool is a potential anthropologist/sociologist or a science and technology studies researcher who can look at different components and connections informing each ethical provocation. Dislodging the expert after all is a key ambition of this tool.
Our belief in using OWL2 was also strengthened by having encountered a range of digital humanities (DH) projects across the world which often communicated knowledge through images, videos and texts without any easy way to retrieve those bits of knowledge in a database form. This has been a major irritant for us. This dismal state of DH as a field was surprising to us because one of the members (first author of this blog post) of the project was a DH dropout and had first learnt semantic web technologies while in a course of digital humanities at an Indian university when studying text encoding initiative (TEI) and working with Wikidata. The image of DH that we had was one of strong interest in building interoperable knowledge outputs that can build from each other. We do not want to take names of DH projects, but relatively well-funded DH projects across the world have been engaging in digitally native output without any seeming interest in presenting their digitally native works in a manner that can allow smooth retrieval and interoperability. Merely making a website to present research findings in non-academic-non-journal-and-entertaining-formats is no different than uploading a fancy looking ppt file.
As we note elsewhere, “To suggest that ontologies and knowledge graphs have still not witnessed wider uptake in social sciences isn’t an exaggeration compared to the industrial, biological, and medical fields. This, even though semantic web languages such as OWL2 lend themselves particularly well to non-standardized expressions and the open-world assumption of the language means that the fundamental insecurity of underdetermination by social scientists is accounted for. Interestingly it is precisely due to this fluidity and openness of OWL2 that logicians who value logical rigour have been arguing for the development of stricter first-order languages.”
Let’s go mad with OWL2 and flood the field of the semantic web with subjective knowledge! Let it live up to its name of being semantic!
Leave a Reply