Authors:

Sydney Sherman

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

A fundamental problem in cultural studies is the analysis of "fuzzy concepts". Fuzzy concepts are ideas and words that are highly contested. We used a Word2Vec model by Mikolov, Chen, Corrado, and Dean demonstrating a distinction in measuring synaptic and semantic word similarities to generate seed words for downstream analysis of the term "sublime". It argues a disposition — tracing how culture has reached for it, wrestled with it, worn it differently across time. This paradox of the scientific and the deeply humanist reveals itself in the trained model: words that have long been grasping at awe and terror, at vastness and overwhelm, without ever naming the sublime directly. The model surfaces what language has always been doing in its absence. Where centuries of aesthetic discourse have circled the concept without consensus, the DTM traces what language does when it attempts the sublime, mapping the themes and words that gather around it in the act of trying. TensorFlow Embedding Projector (Smilkov et al., 2016) instrument used to visualize the embedding space and clustering. We found that the 'sublime' is most associated with the words — disregarding the names of people ex. Kant, Burton, Wordsworth- "deceptions", "sensuous", "imaginative", "longing", and "impressionable", indicating these seed words, signal more than association — they illuminate it as a realm of impression, of felt encounter. It follows, then, to move outward into the literary corpus, gathering all texts that orbit these "fuzzy" conceptual territories, training the model on a larger, richer body of work toward a more complete picture of how the sublime has been rendered in text, image, and media.

Source:

University of Illinois Chicago

Topics:

No topics listed

Co-authors:

Sydney Sherman