Abstract
This study investigates the writing styles of identical twins to determine if their shared genetics result in identical writing. We used computational authorship analysis techniques on writing samples from three sets of identical twins, focusing on lexical, grammatical, and syntactic features. Our results show that machine learning can effectively distinguish between the writing styles of identical twins, primarily using function words, part of speech tags or dependency relations as features. We also found that twins’ writing styles are not necessarily more similar to each other than to unrelated individuals. While the sample is too small to make generalizable conclusions, these findings potentially highlight the complex interplay of genetics and environment in shaping writing style, with implications for authorship attribution, forensic linguistics, and our understanding of human individuality.