VC-THICK Thick vs thin value concepts

STELA: a community-centred approach to norm elicitation for AI alignment · Stevie Bergman; Nahema Marchal; John Mellor; Shak… · 2024

“The majority of rules in this ruleset reflect a desire for impartiality, factuality, thoughtfulness, and proper contextualization from AI systems. By and large, study participants were not looking to have their beliefs validated or contested by LLMs; rather, they sought reasonable, helpful, and nuanced answers that showed consideration for the complex reality of socio-political issues.”

why coded: Community rules are thick: impartiality, factuality, contextualization - with reasons attached · unit #7, pp. 10

Philosophical Investigations into AI Alignment: A Wittgensteinian Framework · José Antonio Pérez-Escobar; Deniz Sarikaya · 2024

“the later Wittgenstein's philosophy of language and mathematics, substantially focused on rule-following, is relevant to understand and improve on the Artificial Intelligence (AI) alignment problem: his discussions on the categories that influence alignment between humans can inform about the categories that should be controlled to improve on the alignment problem when creating large data sets [...] as well as when introducing hard coded guardrails for AI models.”

why coded: Rule-following underdetermination: rules don't fix their application - alignment rests on practice/forms of life · unit #1, pp. 1

How to measure value alignment in AI · Martin Peterson; Peter Gärdenfors · 2024

“If we use conceptual spaces for representing ethical principles, the distance between two points represents how similar the cases are, and the area (or volume) covered by an ethical principle corresponds to its generality.”

why coded: Geometric representation preserves conceptual structure (similarity, generality) that scalar measures flatten · unit #3, pp. 1499

Beyond Preferences in AI Alignment · Tan Zhi-Xuan; Micah Carroll; Matija Franklin; Hal… · 2024

“'preference' is a thin concept because it does not encode richer semantic information beyond the bare notion of 'betterness'. [...] But why exactly are some options preferred over others? In virtue of what reasons do people make these preference judgments? Without answering these questions, we are unlikely to model how someone's preferences generalize to novel options in ways they would endorse. To do so, we must go beyond preferences as the fundamental unit of analysis, and understand how preferences are computed and constructed from our reasons and values.”

why coded: Preference as thin concept; values/reasons as what preferences are constructed from · unit #2, pp. 1822

“they are thick evaluative concepts—concepts that comprise both descriptive and normative elements—such as beauty, humor, or health. As Blili-Hamelin and Hancox (2023) point out, even the concept of intelligence so central to AI is thick in this way.”

why coded: Thick evaluative concepts defined; even 'intelligence' is thick · unit #3, pp. 1823

Full-Stack Alignment: Co-Aligning AI and Institutions with Thick Models of Value · Joe Edelman; Tan Zhi-Xuan; Ryan Lowe; Oliver Klin… · 2025

“By thick models of value, we refer to a broad class of structured approaches to modeling values and norms that meet the above desiderata. In invoking the term 'thick', we draw upon the distinction between thick and thin evaluative concepts and between thick and thin descriptions in anthropology. [...] Similarly to how a grammar constrains language or a type system constrains code, TMV places constraints on what can count as a value – while also remaining open about which values any person or community should endorse.”

why coded: TMV defined via Williams thick concepts + Geertz thick description - grammar-like constraints on valuehood · unit #3, pp. 4

A matter of principle? AI alignment as the fair treatment of claims · Iason Gabriel; Geoff Keeling · 2025

“Stakeholders should be able to present concerns in their own voice and in a relatively un-mediated manner, so long as they are directed at a general audience. This will often involve making claims directly about rights, interests, fairness or well-being in a language with which they are already familiar. [...] the notion of fair process needs to be scoped dynamically on a case-by-case basis [...] leading to what Archon Fung refers to as 'overlapping circles of inclusion'.”

why coded: Own-voice claims = thick, familiar evaluative language admitted into deliberation (tentative) · unit #9, pp. 1961

Why human-AI relationships need socioaffective alignment · Hannah Rose Kirk; Iason Gabriel; Chris Summerfiel… · 2025

“Understanding how to align AI in practice requires moving from narrow, assumption-ridden or 'thin' specifications of alignment towards what anthropologist Geertz (1973) terms [...] a 'thick' description: one that examines the deeper contexts and layers of meaning in which AI systems operate.”

why coded: Explicit Geertzian thin/thick framing of alignment specifications · unit #2, pp. 2

On AI Alignment and the Later Wittgenstein: a Response To Sorin Bangu · José Antonio Pérez-Escobar; Deniz Sarikaya · 2025

“[Commentary clarifying the Wittgensteinian alignment framework against Sorin Bangu's two objections to Pérez-Escobar & Sarikaya (2024).]”

why coded: Annex to the rule-following thread (tentative) · unit #1, pp. 1

Sincerity as ethical alignment to reconstruct the moral foundation of AI ethics · Toru Iwao; Yusuke Nemoto; Nico Surantha; Kenji Su… · 2026

“a meta-ethical enabling condition of coherence and sincerity, defined as the alignment of truth, intention, action, and trust, is missing. [...] Sincerity is treated as a regulatory orientation for human and institutional actors who design, deploy, and oversee AI systems. [...] the proposed SBEF serves as a revision regulator that clarifies when FAT-style mechanisms should be reconsidered and how such revisions can be publicly justified.”

why coded: Sincerity as a thick coherence concept beneath thin FAT principles (tentative) · unit #1, pp. 1

Beyond Preference-based Value-alignment (IEAI Research Brief Q2 2026) · Julia Li · 2026

“In philosophy, thin concepts such as 'bad' or 'good' carry evaluative force with minimal descriptive content, while thick concepts such as 'cruel' or 'honest' track features of the world and evaluate them [...] (Väyrynen, 2013; Williams, 1985). [...] According to Edelman et al. (2025), thin values are descriptively superficial, easily personalizable, and provide little to no justification for why they should exist as norms. [...] Alignment to thin values runs the risk of setting norms that arise from manipulation, addiction and power imbalances.”

why coded: Thick/thin values distinction imported from Williams/Väyrynen via Edelman · unit #8, pp. 4