“Information technology changes the way we perceive the world.”

Frédéric Kaplan, "digital humanities"… how did this adventure start?
“This term was coined about ten years ago in the USA to refer to a field at the crossroads between humanities and information science. Although the first interdisciplinary research in this field began over fifty years ago, developments of the first large-scale digitalization projects have amplified the trend. In fact, we are now entering a new phase of more ambitious projects involving large groups of researchers. These new projects seek to create powerful databases to manage massive amounts of information. This evolution is somewhat similar to the one experienced in the field of life sciences in recent few decades.
In Lausanne, the adventure began a few years back as an informal endeavor by a small group of researchers at the Federal Institute of Technology Lausanne (EPFL) and the University of Lausanne (UNIL). This lead to the creation of my laboratory at the EPFL - the Digital Humanities Laboratory (DHLAB) - and later of the UNIL's Cultural and Digital Humanities Laboratory (LADHUL). Our research is now internationally recognized. We are jointly organizing the world's largest conference in this field: Digital Humanities 2014, which will be held from July 7th to July 11th, 2014.”

You have just received a Courtesy Appointment from the School of Computer and Communication Sciences…
“I already work regularly with Sabine Süsstrunk on issues relating to digitalization. Last year, we co-authored part of a National Centre of Competence in Research (NCCR) proposal on the subject of digital humanities. Over the past few years, I have also interacted regularly with Pascal Fua. We currently discuss together algorithms to analyze the structure of building façades depicted in paintings and engravings. I also recently had very stimulating exchanges with Patrick Thiran and Matthias Grossglauser concerning analysis of social networks of the past and phenomena that enabled the propagation of epidemics such as the "Black Death" during the Middle Ages. There have also been discussions with Anastasia Ailamaki on the best way to create large databases to store "Big Data" (i.e. massive quantities of trace data from historical documents). And finally, I am planning to work with Babak Falsafi to explore the relevance of dedicated software and hardware architectures for text mining (i.e. the automatic extraction of information in texts), which is a major challenge for digital humanities. I am delighted by all of these interactions and look forward to those to come.”

You like to say that some of tomorrow’s information technology challenges can be found in the humanities. Please explain…
“When you try to create "Big Data" for humanities, the main scientific challenge is finding solutions to qualify, quantify and represent uncertainty. A historical document may contain imprecisions, errors (intentional or otherwise) or simply reflect an ideological bias that results in "facts" being presented in a certain way. In addition, the document may be damaged, partially illegible or indecipherable. And, if you lack a perfect understanding of the context, the document may be misinterpreted. The process of digitalization, transcription and indexing add an entirely different layer of possible errors. It is therefore crucial that approaches be developed to formalize not only the information contained in a given document, but also all of the metadata associated with the process of extracting this information. Then you need to find a way to compile information from potentially very different sources in order to build large databases.”

How do you intend to proceed?
“We have developed the concept of fictional spaces associated with a document or corpus of documents. A fictional space is the combination of all of the information that may be extracted from the document(s) in question. Various forms of uncertainty are attributed to this information. We then formulate a framework that enables these fictional spaces to be “joined” to form larger spaces containing even more information. Once these spaces have been joined, we are then able to detect possible inconsistencies (e.g. a person claimed to be involved in a series of events is already deceased at the time the events transpired). In all of these steps, it is nevertheless possible to turn back the clock by excluding, for instance, data coming from a given author or by using a specific digitalization technique. Ultimately, the aim is not to achieve a single reconstituted version of the past but actually to present multiple probable pasts that correspond to certain groups of documents and certain working hypotheses. This type of flexibility does not exist in current web-based semantic techniques. Moreover, efforts to compile "Big Data" from the past could lead to innovations that could then be redeployed to compile "Big Data" today.”

How do we then navigate through all of this information?
“Over the past decade, we have witnessed the spectacular development of tools to navigate through space. It seems to me that the main challenge we face today is how to add a temporal dimension to these tools. For example, is it possible to add a "slider" to Google Maps allowing us to see what a town looked like several hundred years ago? Are we able to reconstitute the social networks of the past with the same density of information as we can with current social networks? Are we able to create search engines that not only provide us with the best answer to our query but also allow us to see how these answers have evolved according to the information available to people in different time periods? These are the types of tools that we are trying to develop at the Digital Humanities Laboratory.”

In your opinion, what impact will information technology have on humanities?
“Humanities is not the only field being transformed by information technology. Information science tools and methods are having a profound impact on all sciences. Databases, simulations, learning algorithms may alter current practices and working processes. The possibility of launching large-scale projects changes the way in which teams are organized. However, there is also something that is even more profound: an epistemological transformation. Information technology changes the way we perceive the world. Computer scientists are always trying to create tools that tend towards a certain degree of generality, proposing models that can be adapted to multiple situations, finding common denominators in diversity. This trend towards generality and abstraction is opposed by a certain culture of specificity in humanities. By striking the right balance between these two trends, very interesting projects will emerge.”

Frédéric Kaplan works as a Professor at the EPFL’s College of Humanities CDH and runs the Digital Humanities Laboratory DHLAB.