Interpretative skill cannot be digitized because it is embodied know-how

Evidence

But these cultures of interpretation cannot be digitalized and made accessible in one click through the magic of Google. We may try to describe them, understand them, explain them, and even formalize a number of their mechanisms – but, since they are rooted in practical abilities and not only in information, in knowing-how (to make sense) and not only knowing-where-to find-the-right-answer, it is up to each individual to incorporate them through habit, repetition, trial and error, through exercises, memorization and continual refinement. — Yves Citton, The Ecology of Attention, 478

Categorization

• XREF: Connects to the ‘grafting as physical skill’ theme in the Trees-as-infrastructure essay — embodied practice that resists codification — and to Polanyi’s tacit knowledge generally. — connects to: Embodied practices make linguistic resignification alone unable to destabilize norms — Saba Mahmood, Politics of Piety: The Islamic Revival…, 1130; Sexual experience cannot be conveyed through instruction but only through lived embodiment — Simone de Beauvoir, The Second Sex, 729; Interaction between two minds cannot be modeled by formal deduction alone — Thomas C. Schelling, The Strategy of Conflict, 368 (+1 more)
• QUESTION: How does this bear on claims that AI can solve interpretation? If interpretation is embodied know-how, then LLMs merely simulate the searchable surface of it.

Broader Connections

• Sexual experience cannot be conveyed through instruction but only through lived embodiment #exemplifies (0.85) — Simone de Beauvoir, The Second Sex, 729
• Embodied practices make linguistic resignification alone unable to destabilize norms #corroborates (0.78) — Saba Mahmood, Politics of Piety, 1130
• Interaction between two minds cannot be modeled by formal deduction alone #corroborates (0.75) — Thomas C. Schelling, The Strategy of Conflict, 368
• Visual competence is learned and more universally distributed than linguistic competence #corroborates (0.72)