From Close to Distant and Back: How to Read with the Help of Machines

Abstract

In recent years a common trend characterised by the adoption of text mining methods for the study of digital sources emerged in digital humanities, often in opposition to traditional hermeneutic approaches. In our paper, we intend to show how text mining methods will always need a strong support from the humanist. On the one hand we remark how humanities research involving computational techniques should be thought of as a three steps process: from close reading (identification of a specific case study, initial feature selection) to distant reading (text mining analysis) to close reading again (evaluation of the results, interpretation, use of the results). Moreover, we highlight how failing to understand the importance of all the three steps is a major cause for the mistrust in text mining techniques developed around the humanities. On the other hand we observe that text mining techniques could be a very promising tool for the humanities and that researchers should not renounce to such approaches, but should instead experiment with advanced methods such as the ones belonging to the family of deep learning. In this sense we remark that, especially in the field of digital humanities, exploiting complementarity between computational methods and humans will be the most advantageous research direction.

Humanistic interpretation and machine learning

“…Many digital humanities scholars have been decidedly critical about the idea that computational methods could mitigate human interpretive involvement in text analysis (Ramsay 2005;Clement 2013;Gibson and Ermus 2019). Instead, this debate has focused on articulating how these "reading machines" (Ramsay 2011) relate to more traditional close reading of texts (Bonfiglioli and Nanni 2016). The literature studies scholar Earhart (2015) has argued that tools for distant reading texts might introduce their own kinds of biases into the reading process, in addition to the "cultural biases" that traditionally influence the work of the literary critic.…”

“…Further, while modelers in the social sciences are typically interested in studying social phenomena for which text data provide evidence, digital humanities often investigates textual patterns in themselves. Such work focuses, for instance, on tracking the prevalence of certain keywords over time (e.g., Ramsey and Pence 2016) or on predicting the authorship of unfamiliar texts not included in the literary canon (Bonfiglioli and Nanni 2016). In such cases, issues concerning the relationship of computational results to the substantive phenomena studied are not similarly pressing.…”

Extração Automática De Informações: Uma Leitura Distante Do Dicionário Histórico-Biográfico Brasileiro (Dhbb

“…No entanto, apesar destas interações levarem a um bom número de projetos de pesquisa conjunta com sucesso, também é perceptível que estas colaborações podem ser difíceis de conduzir (Crymble, 2015), pois diferentes formações, abordagens e expectativas precisam ser constantemente focados em um objetivo comum. Não à toa, durante a última década esta lacuna de conhecimento em métodos computacionais tem levado muitos humanistas digitais a preferirem estudos exploratórios realizados a partir de toolboxespor exemplo, o Mallet, que não demanda conhecimento em programação para obter uma modelagem de tópicos -, ao invés de investigações apoiadas em hipóteses quantitativas (Bonfiglioli, 2015).…”

“…Elas incorporam os métodos e questões desenvolvidos pelas ciências humanas e sociais, ao mesmo tempo que mobilizam as ferramentas e perspectivas únicas abertas pela tecnologia digital (Schnapp et al, 2009). Na área mais intimamente ligada à linguagem e literatura, temos observado de forma crescente o uso de métodos quantitativos de mineração de texto para analisar grandes coleções digitais (Moretti, 2013;Bonfiglioli & Nanni, 2015;Wiedemann & Nielker, 2017).…”

“…Inúmeros acadêmicos abraçaram as ideias de Moretti, apontando que métodos computacionais poderiam representar uma alternativa sólida às abordagens hermenêuticas tradicionais, não apenas nos estudos literários, mas também na pesquisa histórica que precisa lidar com grandes volumes de fontes em formato digital (Jockers, 2013;Bonfiglioli, 2015).…”