Digital history has been booming over the past decade due to an increasing availability of digital-born and digitised source materials. The digitisation of parliamentary records, for instance, has opened prospects for renewing political history by combining quantitative and qualitative approaches to the language of politics. It also enables moving beyond limited nation-state contexts to long-term, comparative, transnational, and global perspectives in new ways. However, there are also downsides to this digital history industry if parliamentary debates are subjected to highly technical text mining without understanding the peculiar features of the genre. This is a point that Jo Guldi makes in her book The Dangerous Art of Text Mining: A Methodology for Digital History (2023). The 2025 volume Writing Conceptual Histories, edited by myself and Jani Marjanen, also discusses it extensively.

Milan van Lange’s Emotional Imprints of War: A Computer-Assisted Analysis of Emotions in Dutch Parliamentary Debates, 1945-1989, initially accepted as a PhD dissertation in history at Utrecht University in 2021, represents both sides of this development. Van Lange makes a promising attempt at innovation by introducing a combination of the history of emotions and quantitative computational text analysis to postwar Dutch political history. Previous research by Remieg Aerts and Ido de Haan has pointed at a shift towards recognising the role of emotions in war experiences since the 1960s, while Carla Hoetink has interpreted politicians as having distanced themselves from emotional expressions in the 1980s. Van Lange’s book challenges such arguments quantitatively. Contrary to the retrospective narrative of postwar emotional silence, this study highlights emotional engagement in parliamentary debates as early as the immediate postwar period. No significant increase in emotional language over time is observed; when peaks do occur, they are more evident in the late 1940s and 1950s than in the 1970s. The intensity of emotions varied by topic, with debates on war criminals, collaborators, and postwar justice eliciting stronger negative emotions such as anger, fear, and disgust. Additionally, there were qualitative shifts – from expressions of empathy to more abstract, strategic uses of emotion and their increasing politicisation.

Since Dutch historians can better judge the book’s contribution to our understanding of Dutch memory politics (Geschichtspolitik) regarding the German occupation during the Second World War, in this review I will focus on its lessons for political historians applying digital approaches. The definition, identification, and analysis of trends in emotions and their impact on national politics is one of the most demanding research topics a digital historian can undertake as argued by Risto Turunen and Ilari Taskinen in Digital Handbook of the History of Experience (2024). For digital history – as for conceptual history traditionally – researchers usually focus on ‘emotion words’. Van Lange aims to assess the role of emotions in Dutch parliamentary debates, particularly in relation to war memories and related legislation. His goal is to provide more precise data and arguments about changes in war-related emotions. Unfortunately, the data are not very precise, and the arguments only emerge during the analysis. The book would have benefited from a clearer overall argument in its introduction.

Deeper source-critical observations on parliamentary debates as a genre and the potential impact of parliamentary agenda-setting on text-mining results would have been helpful. It turns out that the analysed 180 debates have been selected based on their official subjects and keyword searches, rather than on text mining the total dataset of millions of words as context for the analysis of emotions – Chapter 5 being an exception here. The role of random samples from two national newspapers and royal speeches as additional sources remains rather marginal in the overall analysis. As for previous research, emotions arising from wartime occupation are discussed within the confines of methodological nationalism, complemented with occasional references to English-language literature, without relating the research to comparable historiographies from European countries, which would have been too numerous to list here.

Van Lange correctly recognises the need to move between quantitative and qualitative research (37), or to use ‘mixed methods’ (87). He succeeds in the general contextualisation of debates on war experiences and in locating parliamentary interventions that express emotions related to the war, subjecting them to thick description rather than an analysis of speech acts. However, not all readers may find separate text-mining chapters convincing contributions to the historiography of Dutch memory politics. A more integrated presentation of distant-reading results and close-reading descriptions would have increased clarity regarding how the two relate to each other and how certain conclusions were reached.

In response to potential traditionalist opposition to digital history, the book contains numerous apologetical, unfocused, simplistic, and repetitive justifications of digital history. Basic explanations of data processing (ocr, lemmatisation, removing stop words) that might not be necessary in the 2020s are also provided. By contrast, Chapter 4 on operationalisation has been written for digital humanities experts rather than other historians and would have benefited from concrete examples on the research results each introduced statistical method may lead to. While ‘the text-mining techniques are primarily a means rather than a goal in this investigation’ (113), this is not the impression the reader gets after 142 pages of introductory chapters. With professional copy-editing by the publisher, the same could have been expressed much more concisely.

My major methodological criticism concerns the application of the machine-translated Word-Emotion Association Lexicon, based on Anglophone annotations from the 2000s, as ‘the pre-eminent external (re)source’ (65) that is ‘independent of the context’ (69). The author does recognise semantic change but plays down its relevance (71-72). Greater sensitivity to semantic change and cultural-specific variation in the vocabulary of emotions is needed. Semantic change in emotion words could have been traceable through word embeddings, collocations, co-occurrences, and n-grams derived from historical data, for example, but these alternatives are not discussed. Van Lange rather prioritises a focus on relatively rare words as defining the distinctiveness of a text and on the relative emotional exceptionality of a subset of documents (97, 101).

When implemented in Chapter 5, for example, text mining leads to rather undramatic results. No major quantitative changes over time are traceable in any of the consulted genres, leading to the conclusion that no evidence on ‘emotionalisation’ can be found. Some changes in frequencies of emotion words were observed (205) but whether this is proof of the validity of emotion mining (128) remains an open question. As the case studies, too, suggest a rather consensual political discourse lacking strong expressions of emotions, a more ideologically divisive theme might have worked better for emotion mining.

Van Lange’s book reports creative testing of various digital humanities methods but with rather meager results. Readers do not get the kind of convincing fusion of digital methods with political history they might have expected. Even if mixed methods were declared the point of departure, digital humanities approaches overshadow systematic close reading that would have sufficiently considered the rhetorical functions of language.

The author does not hide the limitations of emotion mining, recognising the impact of parliamentary procedure, agenda-setting, and other debated topics on his text-mining results towards the end of the book. He also reiterates the importance of close reading to control distant-reading results. The role of rhetoric now emerges as a result, while it is universally known by parliamentary historians. These features should have been emphasised in the introduction and implemented throughout the analysis. All in all, no quantitative evidence for ‘emotionalisation’ was found (230), a general ‘politicisation of emotions’ in the 1970s is observed (217-218, 236), and an abstractivisation of emotions slightly differing from previous knowledge is noted (222, 237). The conclusion includes good source-critical observations and sufficient links to ‘traditional’ historiography, but with limited added value from text mining.