Spenser’s Spell: Archaism and Historical Stylometrics

Abstract

In the present essay, on Spenser’s orthography, we address the question of whether the orthography of Spenser’s texts, in early modern editions, is salient and therefore worth preserving in modern editions. Even when we face the bibliographic fact that Spenser’s texts seem not to have been regarded as so valuably idiosyncratic that early modern printers preserved them from adjustment, we may still seek to know whether the first editions of Spenser’s works are demonstrably and articulably idiosyncratic, whether orthographically, lexically, inflectionally, or syntactically. This essay formulates a method for statistically probing Spenser’s orthographic profile against what we demonstrate to be the variant, but coherent background of early printed English. We build a model of early modern orthographic change based on letter n-grams extracted from the 60,000 texts in the EEBO-TCP corpus. The n-grams yield some 30,000 features; we concentrate on a subset of the 200 most variant features, reducing the dimensionality of the data by means of principal component analysis (PCA). Situating Spenser’s texts against contemporary and near-contemporary texts, we demonstrate that, in general, Spenser’s corpus is not orthographically distinctive and propose that the impression of the linguistic distinctiveness of Spenser’s poetry is concentrated elsewhere. We conclude both with proposals for new tests that might enable us to isolate that distinctiveness and with a brief assessment of appropriate editorial responses to our investigations.