Exploring the Correlations Between Picasso’s Artworks and his Personal Contacts

The digital humanities are at a turning point where we can digitize and analyze materials using computers with relative ease. The use of digital and computer-based methodologies can make these approaches seem innovative to researchers in the humanities, not necessarily because they are digital but because of the conclusions that can be reached. However, it is clear that we should make a departure from the traditional analyses used in other disciplines, devising our own methods to address our research questions
(Orlandi, 2019).

For some time now, our research has taken us through different approaches to analyze Picasso’s artistic legacy
(Meneses et al., 2008a), his poetry
(Meneses et al., 2008b), its semantic domains
(Meneses and Mallen, 2017) and its distinctions
(Meneses and Mallen, 2018). We also used computer vision to identify the graphic elements in Picasso’s poetry
(Meneses and Mallen, 2020). However, there are many facets of Picasso as an artist that are still unexplored.

Many authors have argued that there is a correlation between the person Picasso was with at a certain time and what he painted then, hence the so-called Olga Period, Marie-Thérèse Years, Dora Maar Years, etc.
(Freeman, 1994)
(Müller et al., 2002). We propose to verify if this correlation exists by examining how frequently specific individuals found in Picasso’s biography are mentioned with respect to each other, and then correlate those frequencies with specific techniques, artworks, places, periods, etc. For this purpose, we will use the entries found on the extensive Online Picasso Project biographical narrative.

The Online Picasso Project
(Mallen, 2018) originated from an emphasis on an interactive, digitally encoded art publication, moving away from an understanding of art criticism as predominantly stative and print-based. It consists of a complex system of interrelated databases which include both texts and images pertaining to Pablo Picasso. As a result of close collaboration between art scholars and computer scientists, the Picasso Project has adopted an innovative architecture with three major objectives: First: to facilitate the maintenance of a large collection of artworks along with its associated critical narratives. Second, to overcome the limitations of printed art publications; and third, to provide new ways for composing, browsing, and exploring the artworks in ways not possible with printed versions. The biographical entries included in the Online Picasso Project are particularly important since they provide a historical framework that is crucial to the understanding of the artist’s legacy which is tightly bound to the experiences he had in his life. Our study aims to employ this historical framework to highlight the hidden connections between Picasso’s life and his artistic output.

Our methodology is based on the taxonomy-based approach that we have used to identify the semantic domains in Picasso’s poetry
(Meneses and Mallen, 2019), but with three differences. First, we analyze
the term frequency-inverse document frequency (
TF-IDF) for keywords in
each biographical entry in the databases of the Online Picasso Project. TF-IDF is a statistical measure that evaluates how relevant a word is to a document in a collection of documents. More specifically, from a list of relationships/friends/dealers, we examine who are mentioned the most and how they are related to each other. For instance, are certain dealers mentioned when a certain model is also mentioned? Second, we analyze how the frequencies in those mentions correlate with specific years/seasons/months. If so, might there be a demonstrable correlation not only with a certain model, but also with a certain dealer or collector in specific periods in his career? Finally, are certain techniques (oil, watercolor, gouache, etc) more frequent in some of those periods? If so, could they be related to the people he is in contact with? Our computational analysis aims to objectively confirm correlations that until now have been simply stipulated by art scholars.

To summarize, we propose to explore and verify the correlations between Picasso’s personal contacts at certain times and his artistic output. This correlation will also provide
an alternate method for exploring the existing document collection by using a curated list of keywords and metadata as an ordering criterion. In this sense, the presentation of our results draws upon our previous work that utilized interactive calendars and timelines, as it also involves utilizing appropriate user interfaces to visualize the results. In the end, the methodological contribution of our proposal is to exemplify the departure from traditional analysis while devising our own methods that are based on other disciplines (in this case statistics and computer science) to address research questions that are unique to the digital humanities.