Tijdschrift voor Nederlandse Taal- en Letterkunde - Volume 135, Issue 4, 2019

This article argues in favour of studying processes of random cultural change in the humanities. Studies of cultural change are of limited value if they are not embedded in or contrasted with a formalized and systematic theory of random change. To situate the discussion, this article first provides a brief overview of historical and more recent accounts of random cultural change. While quantitative, cultural evolutionary studies have for long considered random copying to be fundamental to cultural change, it has received hardly any attention in the humanities literature. Accounts of cultural change in the humanities typically concentrate on informal, narrative descriptions of non-random factors of change, and downplay, reject or simply ignore random change as an alternative explanation. The paper reflects on several reasons pertaining to this lack of attention and highlights the benefits of an integrated quantitative framework of chance for the study of cultural change.

In this essay, we discuss distant reading as one of the various takes on reading that currently prevail in literary scholarship as well as the teaching of literature. We do so in the light of three concepts of reading, which for various reasons can be considered interrelated: close reading, surface reading and distant reading. We introduce a novel, twotier dissection of distant reading and demonstrate how this view relates distant reading to the abstract goals of artificial intelligence. Our analysis foregrounds the notable absence of readerresponse theory in current Digital Literary Studies. We argue that the field has much to gain from more explicit attempts at simulating the socalled ‘natural’ reading process of individual readers.

In this article, we assess the merits of Digital Humanities, and argue that this approach has advanced the field of Dutch linguistics considerably, but at the same time holds some dangers and misconceptions, like (i) the assumption that the new-fangled studies clash with older, more traditional research, (ii) a conceited but unwarranted disdain for current-day non-quantitative work, (iii) the idea that the early adoption of quantitative methods in linguistics created an unbridgeable gap with its neighbouring discipline, the literary studies, and (iv) the risk of the dialectics of lead. We then detect the weak spot of the Digital Humanities approach in linguistics. Though we acknowledge that great strides are currently being made and that earlier research grappled with the issue as well, what turns out to be frustratingly elusive is the whole field of ‘semantics’, especially given the insights from cognitive linguistics about its complicated nature, including encyclopaedic and contextual knowledge. The aporia of semantics is coped with differently in linguistics and literature. While linguistics has sought a rapprochement with the exact sciences, it steered clear from interpretation as much as possible. The field of literary studies, by contrast, remains loyal to its original ‘humanities’ mission, but is now no longer exclusively – or primarily – driven by critically interpreting the meaning of texts as such. Instead, its focus has shifted to offering a critical investigation of the field itself, and of society at large. Both approaches are in danger of derailing: the ‘semanticophobia’ of some linguists dismisses a real and fascinating part of language from its purview; the ‘critical discourse’ approach of some literary colleagues runs the risk of ignoring some useful text-based techniques for empirical research. The main task of Digital Humanities, in both fields, will be to navigate between these pitfalls.

The article aims to explore methodological tenets and options underpinning the ongoing development of computational narratology. So far, the digital turn in narrative studies seems to have been dominated by a return to early formalist and proto-structuralist accounts of story grammar (Propp). However, it is shown that aspects of discourse and style have been latent in approaches to distant reading outside of narrative studies. The application of digital humanities to narrative studies requires us to abandon a staple of literary theory: the alleged oppositions between between system and performance, between structure and pragmatics, autonomy and heteronomy etc. This means that aspects of the paratext and a multimodal take on the setting of narrative are likely to feed into the digital analysis of narrative, but also that categories hitherto considered to be formalist turn out to rely on interpretation. The article then takes a look at the purport of sentiment mining, network analysis and co-reference resolution for digital (literary) narrative studies. Finally, it is argued that the anti-computationalism of second-generation cognitive theory (currently dominant in narrative studies) is in fact a misnomer, as the tenets of embodied cognitionn are in fact highly amenable to the study of the actual reception and comprehension of narratives via digital tools.

‘By the time you’ve sorted out a complicated idea into little steps that even a stupid machine can deal with, you’ve certainly learned something about it yourself.’

– Douglas Adams, Dirk Gently’s Holistic Detective Agency (1987)

The Digital Humanities myths of being ‘revolutionary’ and preoccupied with ‘making’ and ‘building’ are in fact shibboleths that hamper the under-standing of Digital Humanities as a field that in reality is thoroughly grounded in theory. We argue that computational work in the Digital Humanities goes well beyond some utilitarian service to a perceived ‘humanities proper’. It is proper academic research in which computational models are expressions of theory. We explain that it is often hard for humanists to appreciate such forms of theory as they are indeed very differently expressed than more conventional abductive humanities reasoning. Nonetheless, thorough and viable theories result from Digital Humanities intellectual work, which includes the labor to produce software and computational models. We propose that all these forms of theory should not be separate from each other. Instead we strive for a strong reciprocal relation between new computational modeling and existing hermeneutics. This is a holistic approach inspired by the New Sociology of Art in which we relate the computational analysis of text immanent features to sociological processes and existing literary theory. We demonstrate this approach with an example from our current project ‘The Riddle of Literary Quality’. We explain how this is an approach of many small analytical steps carefully building upon each other. Finally we suggest that many of the unproductive misunderstandings between computational and conventional research approaches could be mitigated by increased attention in curricula syllabi for currently developing methods.

Traditionally, the critique of literary representation hinges on ‘symptomatic’ forms of reading in which textual elements are regarded as expressive of deeper ideological issues. In the Digital Humanities, textual approaches are operationalised that target measurable elements on the textual surface. For a digitally oriented critique of representation, this creates a tension between the text’s surface and its ‘hidden’ depths that reside between the lines. In this contribution, the me-rits and pitfalls of a Digital Humanities critique of representation are considered by focusing on character representation in literature. It is argued that symptoma-tic and surface reading are not mutually exclusive, but can potentially enhance one another, provided that one stays clear from problem solving rhetoric and neopositivist assumptions.

Both historical linguistics and Digital Humanities have made great progress in recent decades. However, technology and research have still remained separate worlds, and as a result the digital infrastructures and corpora are not optimally equipped for diachronic linguistic research and take too little account of the theoretical insights. In this article I show on the basis of a concrete research question what requirements the data, metadata and infrastructure should meet in order to make the best use of the theoretical insights, and how far advanced we are at this moment. I demonstrate which research possibilities the available infrastructures offer and what more is needed, and I conclude with a description of the ideal infrastructure for a historical linguist and how to achieve this.

As argued in this article, the digital analysis of literary texts could be enhanced by a closer connection between literary theories and digital techniques. Moretti tried to bridge the gap between the two, but did not fully succeed. Could a more systematic approach to what literary theory has to offer – explanatory hypotheses about the relations between textual structures, societal effects and intellectual shifts – and the requirements of analyses based on numbers (be it statistical patterns or other numeric patterns) help to make progress?

This paper argues that formal-theoretical – in particular generative – and quantitative-statistical approaches to language variation are highly complementary. The latter have sophisticated means of dealing with large and highly variable data sets, while the former introduce hypotheses and analyses that can guide the interpretation of the statistical results. We illustrate this claim on the basis of two case studies of morphosyntactic variation in Dutch dialects.

Dit artikel wenst op drie manieren bij te dragen aan het debat over Digital Humanities. Op het empirische niveau willen we het competentie profiel van de Letterenfaculteit van kuLeuven blootleggen zoals dat door haar studenten wordt gepercipieerd: op welke vaardigheden zetten opleidingen volgens studenten het meeste in, en welke plaats krijgt het adjectief digital in hun kijk op de humanities? Om deze vragen te beantwoorden, hanteren we een bottom-up, mixed-methodaanpak waar technieken gerelateerd aan Digital Humanities concreet worden ingezet, wat meteen onze methodologische bijdrage vormt. Met name verwerken we de input van een free response-vragenlijst via word space models (Turney & Pantel 2010; Lenci 2018). Uit de resultaten komt een duidelijk ‘humanistisch’ profiel naar voren. Een opvallende afwezige is de digitale component, maar net die komt herhaaldelijk terug in antwoorden van alumni op de vraag welke competenties ze vaak nodig hebben op de werkvloer, maar niet aangeleerd kregen tijdens de opleiding. Als beleidsmatige bijdrage zullen we tot slot stilstaan bij de verschillende manieren waarop opleidingen in de humane wetenschappen met de aangetroffen lacune kunnen omgaan.

Dit artikel wenst op drie manieren bij te dragen aan het debat over Digital Humanities. Op het empirische niveau willen we het competentie profiel van de Letterenfaculteit van kuLeuven blootleggen zoals dat door haar studenten wordt gepercipieerd: op welke vaardigheden zetten opleidingen volgens studenten het meeste in, en welke plaats krijgt het adjectief digital in hun kijk op de humanities? Om deze vragen te beantwoorden, hanteren we een bottom-up, mixed-methodaanpak waar technieken gerelateerd aan Digital Humanities concreet worden ingezet, wat meteen onze methodologische bijdrage vormt. Met name verwerken we de input van een free response-vragenlijst via word space models (Turney & Pantel 2010; Lenci 2018). Uit de resultaten komt een duidelijk ‘humanistisch’ profiel naar voren. Een opvallende afwezige is de digitale component, maar net die komt herhaaldelijk terug in antwoorden van alumni op de vraag welke competenties ze vaak nodig hebben op de werkvloer, maar niet aangeleerd kregen tijdens de opleiding. Als beleidsmatige bijdrage zullen we tot slot stilstaan bij de verschillende manieren waarop opleidingen in de humane wetenschappen met de aangetroffen lacune kunnen omgaan.

This article examines the impact of born-digital literature on editorial theory and genetic criticism, notably on the notion of the textual ‘version’. Born-digital works challenge the logic of the version. At first sight, it seems as if writings whose genesis is recorded with keystroke logging software force us to rethink genetic research so thoroughly that the notion of the version becomes obsolete. However, the research hypothesis of this article is that, on condition that the size of the textual unit to which the notion of ‘version’ applies is clearly defined, this concept does remain a valuable tool in the study of literary geneses. The case study to investigate this research hypothesis is a born-digital story by the Dutch author Ronald Giphart, written with the help of a writing bot.