Creativity, Invention and Linguistic Analysis.

Can we systematically analyse the linguistic features of literary texts that employ highly inventive language? How can existing tools for automated text analysis and computational linguistics be adapted when resources designed for natural language fall short? Traditional approaches often prioritize close reading over quantitative analysis, particularly for texts marked by extreme creativity. This study proposes an alternative method, using automated text analysis to examine the language and style of Giovanni Testori (1923-1993), a 20th-century Italian author. Testori’s works challenge standard linguistic frameworks, merging Italian, Milanese dialect, word deformations, and code-mixing across stylistic levels. This approach assumes that even highly creative literary languages must maintain a degree of comprehensibility to achieve effective communication, which means relying on iterative and systematic deformation techniques. By identifying and formalizing these patterns through regular expressions, inventive languages can be treated analytically as natural ones, allowing their features to be statistically examined. While the approach involves subjective interpretation and risk errors in manual labelling, it demonstrates how computational tools can quantify and represent an author’s style even in extreme cases. Integrating qualitative insights with quantitative analysis offers a deeper and more reliable understanding of the author’s poetics, reducing the biases and limitations of close reading alone.