GRAPHOLOGICAL DEVIATION IN CONTEMPORARY POETRY: A COMPARATIVE STYLISTIC ANALYSIS OF HUMAN-WRITTEN AND AI-GENERATED POETRY

Abstract

This study aims to examine the graphological deviation in the poetry of the present time by making a comparative stylistic analysis between human-written and AI-generated poems. The present study uses the theory of foregrounding by Geoffrey Leech and the computational model of digital text generation by Nick Montfort to explore the role of typographical features. It shows how the graphological elements like punctuation, capitalization, lineation, stanza structure, and spatial arrangement are used in meaning-making and aesthetic effects in different poetic modes. The human corpus is made up of three poems by Rupi Kaur. She is selected for her characteristic graphological style. Three poems generated by ChatGPT (OpenAI, GPT-based large language model) form an AI corpus. These poems are generated under controlled prompting conditions to minimize researcher bias. This study reveals a distinction between two corpora through close reading and qualitative stylistic analysis. Kaur’s poetry expresses intentional deviation from conventional graphological norms, such as the absence of punctuation, lowercase lettering, and fragmented visual layout. On the other hand, ChatGPT-generated poetry follows conventional orthographic standards such as capitalization, regular punctuation, and structured visual layout with only limited incidental deviation. These findings demonstrate that graphological deviation acts as a stylistic indicator. It reflects the intentional creativity of poets to achieve salience and create aesthetic effects. The study presents empirical observation of how graphological deviation differentiates human and machine-generated poetry. Thus, it contributes to ongoing discussion in stylistics, artificial intelligence, and digital poetics.