This paper returns to the relationship of “narrative versus database” (an argument originally made by Lev Manovich in 2001) as one that can be further addressed. A specific issue persists in text analysis research in the digital humanities: the difficulty of representing the figurative meaning of narratives through digital tools. Towards an accommodation, this paper adopts a narratological framework in order to propose alternative models of content management and organization that more closely resemble figurative meaning making in human language. These alternative models therefore better allow for the computational representation of figurative elements that N. Katherine Hayles describes as “the inexplicable, the unspeakable, the ineffable” of narrative literature. This paper argues that the construction of figurative meaning through paradigmatic substitution (as part of an imaginary vocabulary that is drawn from in the process of meaning making) is difficult to account for in the relational database—arguably still the most culturally prominent database model. By focusing on NoSQL (“no” or “not only” Structured Query Language) databases, this paper explores how layers of figurative meaning can be represented together through these flexible and non-relational models. In particular, the ability of non-relational databases to group together multiple values—encouraging their association, comparison, and juxtaposition—can be analyzed as a computational albeit imprecise counterpart to the formation of paradigmatic and figurative meaning. Thus, towards accounting for a word, image, or idea’s layers of meaning as expressed in literature, this paper offers a study of the limitations of digital tools and their critical negotiation with humanities research and reflection.