Major claim: "[Vector space models] call for distinct modes of humanistic interpretation and explication that are related to but distinct from those that may have been used on the original source texts."
— James E. Dobson (@jeddobson) October 11, 2021
Surprisingly perhaps, it turns out that the hermeneutical theories of nineteenth-century theologian and philosopher Friedrich Schleiermacher are perhaps the most useful frame for understanding word embeddings.
— James E. Dobson (@jeddobson) October 11, 2021
This essay builds on my recent DHQ essay "Interpretable Outputs: Criteria for Machine Learning in the Humanities" as the conditions for interpretabilityhttps://t.co/3j4IEffP86 #ml
— James E. Dobson (@jeddobson) October 11, 2021
