Recently, fifty international experts on nineteenth-century music criticism gathered in Lucca for a three-day conference (link). One of the returning issues at this conference was the methodological problem of interpreting the music critics’ writings: not as truthful accounts isolated from their context, but as products of a multifaceted, biased historical practice. Different papers critically addressed or incorporated this issue – such as Katherine Ellis in her keynote – others showed how ingrained this research practice actually still is. Especially the enormous amount of anonymous critics continuously causes interpretative challenges.
Besides the undeniable importance of historical contextualisation, I want to propose another approach to tackle this methodological problem. As a Junior Researcher within CREATE I learned the benefits of de-contextualising or, more accurately, re-contextualising the reviews of nineteenth-century anonymous music critics. For this project I have been using ‘text mining’ tools to digitally explore concert reviews in the Dutch musical journal Caecilia: Algemeen muzikaal tijdschrift van Nederland (1844-1917). This methodology has been described as ‘distant reading’ – in contrast to ‘close reading’ – for which textual data is appropriated from its original context and modulated to fit different types of enquiries.
What kind of results can researchers of nineteenth-century music criticism expect from digitally re-contextualised texts of anonymous reviewers? A relatively simple, but highly effective way in which mining a dataset, such as Caecilia’s, can be useful is by looking at the frequency of mentions of specific (groups of) composers in the concert reviews. The mentions in the reviews are reflections on the performed repertoire and, as such, give a strong indication of shifts in repertory and can aid studying processes of canonisation. Furthermore, calculating the correlation between composers and specific conductors helps locating transformations (and the obstructions of these) in the repertoire.
Besides plotting shifts of occurrences of composers and conductors and the relations between them, digital tools can help counting patterns in the dataset on a semantic level by looking at often used words. This can, for example, be used to identify changes in critical vocabulary and ideological orientations – a topic I discussed with my paper at the Nineteenth-Century Music Criticism conference. Moreover, the mapping of words can also be used in a more associative manner. By searching for often used words in correlation with articles for specific years or locations, it can help forming a quick understanding of the text’s content and establish interaction between distant and close reading as an aid for the researcher. By further processing the dataset, it’s possible to specify the search for all these tools to, for example, different cities, venues, orchestras, conductors, and composers.
While scholars experiment with more advanced text mining techniques (for example the Stanford Literary Lab and the Topic Modeling project Mining the Dispatch), the above-mentioned research applications can relatively easily be incorporated to complement more traditional research methodologies. In order to do so, we hope to publish an interactive database of concert reviews from Caecilia at the beginning of 2016 – it will offer to investigate the texts of nineteenth-century anonymous music critics in various new contexts.
Written by: Thomas Delpeut