This paper proposes Synthetic Legibility as the designed capacity of a corpus to remain coherent across human reading and machine processing. Contemporary research is increasingly encountered first through computational systems: search engines, repository crawlers, citation graphs, indexing bots and large language models. Under these conditions, visibility is insufficient; scholarly objects must become traversable. The paper examines identification, metadata, semantic recurrence, dataset architecture, graph integration and interface as layers of legibility infrastructure. It argues that metadata is not administrative aftercare but part of the public architecture of knowledge.
Anto Lloveras (Sat,) studied this question.