# TU Eindhoven ESDiT Postdoc Application - Reference: 2026/257 - Deadline: **30 May 2026** - Position page: [[Clippings/Postdoc In Philosophy of Technology]] - Apply URL: https://tue.varbi.com/en/apply/positionquick/927469/ - Supervisors: - Matthew Dennis ([email protected]) - Gunter Bombaerts ([email protected]) - HR contact: [email protected] - ESDiT consortium: https://www.esdit.nl/ ## Submission Archive Submitted via GradeVarbi on **2026-05-30 at 16:51**. Confirmation screenshot and final uploaded PDFs are archived here: ![[Attachments/TU Eindhoven Postdoc 2026/Final Submitted Materials/TU Eindhoven submission confirmation 2026-05-30 16.52.52.png]] - [[Attachments/TU Eindhoven Postdoc 2026/Final Submitted Materials/TU Eindhoven research proposal - Nick Young - submitted.pdf|Research proposal - submitted PDF]] - [[Attachments/TU Eindhoven Postdoc 2026/Final Submitted Materials/TU Eindhoven cover letter - Nick Young - submitted.pdf|Cover letter - submitted PDF]] - [[Attachments/TU Eindhoven Postdoc 2026/Final Submitted Materials/TU Eindhoven CV - Nick Young - submitted.pdf|CV - submitted PDF]] - [[Attachments/TU Eindhoven Postdoc 2026/Final Submitted Materials/TU Eindhoven teaching evaluations note - Nick Young - submitted.pdf|Teaching evaluations note - submitted PDF]] - [[Attachments/TU Eindhoven Postdoc 2026/Final Submitted Materials/TU Eindhoven key publication - Growing the Image - submitted.pdf|Key publication sample - submitted PDF]] ## Required Documents (per JD) - [x] Cover letter (precise interest + how background prepares for the research plan + relevant circumstances) - [x] CV with publications + contact details for **2 references** → [[Nick Young CV 2026]] - [x] Research plan, ≤ **750 words**, addressing the vacancy's research question - [x] One key publication, primary author, ≤ 20 pages (selections from larger work permitted) - [x] Course evaluations from teaching (if any) Note: if shortlisted, a brief note on integrating "activating impact" through research is requested. ## Position Summary - 3-year postdoctoral fellowship in ESDiT (Ethics of Socially Disruptive Technologies) - Theme: *Co-creating a Transdisciplinary Framework for Sociotechnical System Flourishing* - Goal: synthesise ethical theories of human flourishing with systems thinking, then apply via transdisciplinary co-creation with non-academic stakeholders - Three JD-named candidate cases: attention economy, dementia care, workplace well-being - Embedded in ESDiT's STEM line (expected to take coordinating role at some point) - Hosted by Philosophy & Ethics group, TU/e - Salary scale 10 (€4,241 – €5,538), full-time, 3 years - 0.1–0.2 FTE teaching contribution to P&E - Required to be based in the Netherlands - Generous conference and travel budget ## Strategy Notes ### Fit - PhD in philosophy ✓ - Record of internationally recognised publications: *Philosophical Quarterly*, *Philosophical Studies* (×2), *AJP*, OUP chapter, *Ergo* (forthcoming) ✓ - Active work bridging philosophy of AI and aesthetics of AI: *Growing the Image*, *Environmental Aesthetics of AI*, *Generating Philosophy with AI* - Existing collaboration with Enrico Terrone demonstrates capacity for sustained joint projects - PEA project background (experiential artifacts: artifacts whose function is to elicit experiences) is structurally adjacent to the JD's interest in sociotechnical systems that shape experience and well-being - Excellent English ✓ ### Gaps to address (head-on, not hidden) - No prior transdisciplinary stakeholder co-creation experience — interviews, co-creation workshops, ethnography, technology assessment. Propose a credible learning trajectory and pick one method to commit to. - No prior work in human-flourishing ethics specifically (Aristotelian, capabilities, virtue) — must engage Dennis's framing - No prior work in systems thinking literature — must engage seriously rather than name-drop - Current centre of gravity is aesthetics-of-AI rather than ethics-of-AI per se — must bridge cleanly ### Framing - Lead with the experiential-artifacts framing (PEA) as the existing intellectual bridge to *sociotechnical systems that shape experience* - Position transdisciplinary methods as a genuine learning curve under the supervisors' guidance — not as fake expertise - Choose a research-plan case study where I have real philosophical traction (the attention economy is the closest to my work on AI and aesthetic engagement) - Best publication sample: *Growing the Image* (Philosophical Quarterly 2025), primary author ### Pre-submission research (need to do before drafting) - [ ] Read Matthew Dennis's recent work — flourishing, ethics of technology, digital well-being - [ ] Read Gunter Bombaerts's recent work — energy ethics, motivation, well-being - [ ] Skim ESDiT STEM line publications for vocabulary and framing - [ ] Identify one or two transdisciplinary methodologies I'd commit to (e.g. co-creation workshops with designers, or auto-ethnographic study of LLM use) - [ ] Identify the flourishing-ethics literature I'd engage seriously (Nussbaum capabilities? MacIntyre virtues? Wong's neo-Aristotelian work? Dennis's own framework?) --- ## Reading & Sources Living reading list for the application. Each text is mirrored to Readwise Reader (tagged `postdoc job` and `toread`, plus a topic tag) so it can be read on any device; tick the box here once read. Entries marked *to pull* still need to be sourced and uploaded. ### Supervisory team - [ ] Attention as Practice: Buddhist Ethics Responses to Persuasive Technologies — Bombaerts, Anderson, Dennis et al., *Global Philosophy* 33:25 (2023). The team's foundational joint paper: rejects the individualistic concept of attention as a controllable resource for an embedded "attention ecology". Structurally prefigures the call's systems-thinking-plus-flourishing synthesis, with the attention economy as the worked case. — [Readwise](https://read.readwise.io/read/01ks7d63n2h2fpah4tmsg2adap) - [ ] *To pull* — Matthew Dennis, solo flourishing/virtue work (*Cultivating Our Passionate Attachments*, Routledge 2020; Stoic self-care apps; Confucian digital-wellbeing; cyber-virtues with Tom Harrison). - [ ] *To pull* — Gunter Bombaerts, recent work on energy ethics, motivation, and well-being. ### Background theory - [ ] *To pull* — flourishing ethics: a neo-Aristotelian / virtue-ethics anchor (the supervisors' centre of gravity, rather than Sen–Nussbaum capabilities). - [ ] *To pull* — systems thinking: Donella Meadows, *Thinking in Systems*, plus a resilience-theory piece for the "qualities that resist disruption" vocabulary. ### Programme context - [ ] *To pull* — ESDiT STEM-line outputs, for framing and vocabulary. ### Phenomenology of agency Background for the agency strand — the Horgan literature underpinning Nick's own relevant paper, and a resource for the call's "agency" flourishing-feature. The 2021 Routledge chapter ("On the Satisfaction Conditions of Agentive Phenomenology") is in Learning but has no Readwise copy. - [ ] The Phenomenology of First-Person Agency — Horgan, Tienson & Graham, in *Physicalism and Mental Causation* (2003). — [Readwise](https://read.readwise.io/read/01ks7h9mcqty5xdc3evhx56af8) - [ ] The Phenomenology of Embodied Agency — Horgan & Tienson, in *A Explicação da Interpretação Humana* (2005). — [Readwise](https://read.readwise.io/read/01ks7h9p0d99879742br0bq76h) - [ ] Agentive Phenomenal Intentionality and the Limits of Introspection — Horgan, *Psyche* 13/2 (2007). — [Readwise](https://read.readwise.io/read/01ks7h9qe2y7a03bxsp3f143m1) - [ ] The Phenomenology of Agency and the Libet Results — Horgan, in *Conscious Will and Responsibility* (2010). — [Readwise](https://read.readwise.io/read/01ks7h9rayvfvwwzqej891ym19) - [ ] From Agentive Phenomenology to Cognitive Phenomenology: A Guide for the Perplexed — Horgan, in *Cognitive Phenomenology* (2011). — [Readwise](https://read.readwise.io/read/01ks7h9sb5sfbwtjj0pqcbrr2f) - [ ] The Phenomenology of Agency and Freedom: Lessons from Introspection and Lessons from Its Limits — Horgan, *Humana.Mente* 15 (2011). — [Readwise](https://read.readwise.io/read/01ks7h9t6v8kkta7vrf82d8n7f) --- ## Cover Letter [TO DRAFT] Address: Matthew Dennis and Gunter Bombaerts Drafting brief (the JD is explicit about three components): 1. **What specifically interests me about this position** — not generic enthusiasm; a concrete intellectual interest tied to my actual work 2. **How my background prepares me to carry out the proposed research plan** — link the experiential-artifacts work, the *Growing the Image* framework, the philosophy-of-AI work to the sociotechnical-system-flourishing question 3. **Circumstances relevant to suitability** — current Genova contract ends Feb 2027 so the timing works; based in Milan, willing to relocate to the Netherlands Voice: - No performed enthusiasm - Acknowledge the transdisciplinary-method gap directly as a learning curve I want under Dennis and Bombaerts's supervision - Don't fake systems-thinking or flourishing expertise - End cleanly [Draft goes here] --- ## Research Plan (≤ 750 words) [TO DRAFT] Current draft note: [[TU Eindhoven Research Proposal Draft 2026]] The hardest piece. The JD frames the research question as: > Co-create a transdisciplinary framework that synthesises ethical theories of human flourishing with systems-thinking approaches into a notion of sociotechnical system flourishing. System-thinking features named in the JD: adaptability, feedback awareness, ability to incorporate external elements. Flourishing features named in the JD: agency, teleological orientation, narrative coherence. JD-named candidate cases: attention economy, dementia care, workplace well-being. ### Candidate angles to test **A. Attention economy + flourishing** *(closest to current work)* Bridge: aesthetic engagement → attention regulation → flourishing of attention systems. The *Growing the Image* framework already treats AI-saturated environments as *environments* with their own aesthetic structure; that structure shapes attentional possibilities. Systems thinking adds the feedback-and-adaptation lens (recommendation systems learning from us learning from them). Flourishing adds the normative anchor. **B. Generative AI and designer/user flourishing** *(extends current trajectory)* Bridge: *Growing the Image* takes generative AI as a gardening medium; co-cultivation between human and system. The flourishing question becomes: under what conditions does this co-cultivation enable rather than erode the designer's or user's flourishing? Systems thinking: where does the human's adaptability go when the system adapts on their behalf? **C. LLMs and epistemic flourishing** *(connects to Generating Philosophy paper)* Bridge: working in extended natural-language exchange with LLMs is a novel cognitive and intellectual practice. The flourishing question: what does intellectual flourishing look like in an LLM-saturated knowledge environment? Where is the line between "thinking with" and "outsourcing"? **D. Workplace well-being + AI tools** *(closest to a JD case + tractable case study)* Bridge: AI tools at work as experiential artifacts. Co-creation with designers and end-users on what flourishing-supporting versus flourishing-eroding deployment looks like. ### Likely choice and reasoning [Pending Nick's call. My read: option A is closest to current published work, option B is closest to current draft work, option D is closest to JD-named case. Combining A with D (attention economy as workplace question) might give the strongest fit while staying inside my existing intellectual territory.] [Draft goes here when angle is chosen] --- ## Key Publication Sample **Selected**: "Growing the Image: Generative AI and the Medium of Gardening" (with Enrico Terrone), *Philosophical Quarterly* (2025). DOI: 10.1093/pq/pqae120. - Primary author - Most recent first-tier journal placement - Directly relevant: philosophy of generative AI as a question of what kind of medium it is, with implications for how the human-AI relationship should be conceived - Page length within the 20-page limit (verify exact page count from the published version) If page count exceeds 20, prepare selections from the paper. --- ## Course Evaluations Check UniGE evaluation system for student feedback on the 2025/2026 Genova MA course. Attach what exists. Teaching-evaluation explanation note: [[Teaching Evaluations Note - TU Eindhoven Postdoc 2026]] --- ## Activating Impact Note (post-shortlist only) The JD asks for this only if shortlisted for interview. Draft after interview invitation arrives. It should follow naturally from the research plan's case study and chosen co-creation method. --- ## Pre-submission Checklist - [ ] Pre-submission research on Dennis, Bombaerts, ESDiT STEM line complete - [ ] Research-plan angle chosen - [ ] Research plan finalised, ≤ 750 words verified - [ ] Cover letter finalised - [ ] CV finalised with 2 referees confirmed and notified - [ ] Publication sample finalised (Growing the Image, ≤ 20 pages confirmed) - [ ] Course evaluations attached if available - [x] Online application submitted via Varbi ## Open Items - [ ] Read Dennis's recent flourishing-and-tech papers - [ ] Read Bombaerts's recent work - [ ] Skim ESDiT STEM-line outputs - [ ] Pick one transdisciplinary method to commit to in the plan - [ ] Decide research-plan angle (A / B / C / D or a combination) - [ ] Confirm 2 referees - [ ] Note: there are other ESDiT vacancies at other Dutch universities — JD invites simultaneous applications. Worth checking which. ## Web Check Additions (2026-05-22) Sources checked: AcademicTransfer listing, AcademicTransfer apply page, Academic Positions mirror, ESDiT news page, Eindhoven Center for Philosophy of AI post, and Matthew Dennis PHILOS-L circulation archived on Philwise. ### Extra details to carry into the application - Exact deadline time: apply by **30 May 2026, 23:59 Europe/Amsterdam**. - Interview timeline: initial interviews are scheduled for **June 2026** and may be online; a second round is likely in **early July 2026**. - Complete applications matter: the listing says incomplete applications may not be considered and could be rejected; complete applications are given priority. - Possible additional collaborators named in Matthew Dennis PHILOS-L circulation: members of TU/e Human-Technology Interaction group, including **Wijnand IJsselsteijn** and **Linghan Zhang**, plus **Joel Anderson** from Utrecht University. - The broader TU/e listing frames IE&IS around **Humans and Technology**, **Supply Chain Management**, **Sustainability and Circularity**, and **Value of Data-Driven Intelligence**. This may help with the cover letter paragraph on why TU/e / why this department. - TU/e / Brainport framing: the listing emphasizes Eindhoven as a high-tech region with strong industry collaboration, including AI, photonics, semiconductors, and advanced manufacturing. Useful for the practical-impact/transdisciplinary angle if it can be used without sounding generic. - Conditions and logistics not yet in the summary: year-end bonus 8.3%, annual vacation pay 8%, postdoc association, staff immigration team, possible 30% tax facility, moving-expense compensation, on-campus childcare/sports facilities, commuting / working-from-home / internet allowances. - Pre-employment screening: a knowledge security check can be part of the selection procedure. - Diversity note: TU/e explicitly welcomes candidates with varied backgrounds and especially encourages candidates from underrepresented groups to apply. ### Source links - Official AcademicTransfer listing: https://www.academictransfer.com/en/jobs/360382/postdoc-in-philosophy-of-technology - AcademicTransfer application page: https://www.academictransfer.com/en/jobs/360382/postdoc-in-philosophy-of-technology/apply/ - Academic Positions mirror: https://academicpositions.com/ad/eindhoven-university-of-technology/2026/postdoc-in-philosophy-of-technology/247658 - ESDiT news page: https://www.esdit.nl/esdit-lines/new-methods-for-ethics/ - Eindhoven Center for Philosophy of AI post: https://ephil.ai/post/2026-04-24-postdoc-position/ - PHILOS-L / Philwise archive: https://philwise.co/phitwi/entry/1347424642 ### Caution - Ignore aggregator pages that list a different deadline, especially ApplyKite pages showing **31 December 2026**. Reliable sources consistently give **30 May 2026**. ## Combined PNG ![[TU Eindhoven ESDiT Postdoc Combined.png]] ## Claude/Codex Architecture Work (2026-05-22) Status: architecture and stress-test mode, not drafting. Codex and Claude ran several rounds on the research-plan spine while Nick was at the gym. Current best direction is no longer the generic attention-economy angle and no longer a direct Generative AI creativity pitch. The strongest fit is a workplace well-being case: **LLM-mediated software development** as a test case for a positive account of flourishing-as-agency in sociotechnically mediated practice. ### Settled Working Spine Public-facing philosophical contribution: develop a positive account of flourishing-as-agency in sociotechnically mediated work, using Nick Youngs existing research programme in agentive phenomenology of action and artefact-mediated practice. Core sentence, still too dense for the final 750-word version: > Drawing on a research programme in the agentive phenomenology of action and artefact-mediated practice, this project develops a positive account of flourishing-as-agency in sociotechnically mediated work, specified through three structural criteria - answerable participation, end-fidelity, formative continuity - operating under a meta-condition of reflexive openness, tested in LLM-mediated software development through co-creation with developer-practitioners, and yielding criteria for design and organisational support of flourishing-supporting mediation. Case: **LLM-mediated software development**, not "coding agents" and not broad "technical knowledge work". Reason: it is recognisable as workplace well-being, fits TU/e and Brainport high-tech context, connects to developer-practitioner co-creation, and gives a concrete setting in which AI tools mediate agency, attention, craft, metrics, feedback, and organisational pressures. Diagnostic motivation: many autonomy/manipulation framings miss cases where felt agency is preserved or even intensified while the substantive conditions of agency are reshaped by tool design, organisational metrics, feedback loops, professional norms, and habit formation. This should motivate the project, but should not be the headline. The headline should be constructive. ### Criteria Use three criteria plus one systems-thinking meta-condition: 1. **Answerable participation**: mediation supports the developer in occupying an agentive position within distributed practice such that their contribution is intelligible, ownable, and answerable, without requiring old-fashioned individual authorship of all outputs. 2. **End-fidelity**: mediation supports the developer actions being governed by the ends they can reflectively endorse as the right ends of the practice, rather than by ends substituted via tool design, productivity metrics, or local reward structures. 3. **Formative continuity**: mediation supports the developer in remaining a learning, judging, attending agent who is being formed by the practice, rather than becoming a capability-substituted endpoint whose work the system increasingly performs. 4. **Reflexive openness** as meta-condition: the structures of participation, ends, and formation must themselves be legible to and revisable by participants. This is where systems thinking enters explicitly: adaptability, feedback awareness, contestability, and openness to external elements. These labels are provisional. Claude flagged possible alternatives: "practice-revisability" instead of "reflexive openness"; "teleological integration" instead of "end-fidelity"; "craft-continuity" instead of "formative continuity". ### Method Credible minimum method: phenomenological/philosophical case analysis plus two co-creation workshops with developer-practitioners, likely 6-10 participants each, designed with ESDiT and Bombaerts-style co-creation expertise. Avoid promising ethnography, interviews-at-scale, longitudinal studies, or quantitative HCI work. Do not call the method auto-ethnography. If Nicks own tool use matters, phrase it as practitioner familiarity or close phenomenological engagement with LLM coding tools. Likely outputs: two papers, a stakeholder-facing criteria document, and a contribution to the ESDiT STEM-line framework. ### Guardrails - Do not foreground "chromatic illumination". It can quietly support the end-fidelity criterion, but it is not the project headline. - Do not use *Growing the Image* to support the attention-economy angle. That bridge was rejected as a smuggle. - Use *Agents of Change* for source-experience, agency, and attention-as-mental-action resources. - Use *Beauty in Use* for agentive phenomenology of artefact-mediated action and for the background idea that tool use can structure the felt character of agency. - Use *Growing the Image* mainly as publication sample and as evidence of recent AI work, not as the conceptual load-bearing bridge for this plan. - Do not frame the project as pure attention economy, pure distraction, or productivity optimization. - Mention attention lightly: as one form of mental agency/source-experience and as a hook into the supervisors attention-as-practice work. Do not make it the case. - Do not overpromise all three flourishing features equally. Agency is central; teleological orientation and narrative coherence are handled through end-fidelity and formative continuity. - Avoid "engineered" if it implies unified intentional design. Prefer "mediated", "shaped", "structured", "distributed", "emergent", or "feedback-governed". ### Paragraph Architecture For The 750-Word Plan 1. Team problem and gap: sociotechnical flourishing needs a bridge between systems thinking and flourishing ethics, especially at the experiential level where agents inhabit mediated practices. 2. Philosophical contribution: agentive phenomenology of action and artefact-mediated practice; positive criteria for flourishing-as-agency; diagnostic critique as motivation, not headline. 3. Case: LLM-mediated software development as workplace well-being case and Brainport/TU/e fit. 4. Method: phenomenological case analysis plus two co-creation workshops with developer-practitioners. 5. Outputs and integration: papers, design/organisational criteria, stakeholder-facing document, ESDiT/STEM-line contribution. ### Open Questions - Which labels sound most like Nick and least like application jargon: answerable participation / end-fidelity / formative continuity / reflexive openness? - Should "reflexive openness" be named in the core paragraph or only described as the systems-thinking condition? - Need exact citations and quick source checks for Dennis, Bombaerts, Attention as Practice, and any Bombaerts GenAI or challenge-based-learning work used in the fit paragraph. - Need decide whether publication sample remains *Growing the Image*; current answer is probably yes. - Need convert the dense core sentence into readable 750-word architecture after Nick signs off on the conceptual spine. ## Claude/Codex Continuation - 750-Word Architecture (2026-05-22) Continuation after Nick returned and asked Codex to keep going. Claude and Codex stress-tested the research-plan architecture under the 750-word constraint. ### Five-Paragraph Research Plan Architecture Target total: about 730 words, leaving a 20-word buffer. 1. **Opening / framing and contribution (110 words)**: open with the structural finding, then bridge to the JD question, state the contribution, and name the case. 2. **Philosophical anchor and criteria (180 words)**: name agentive phenomenology of action and artefact-mediated practice; cite *Agents of Change* and *Beauty in Use* in one sentence; state the structural observation about mediated agency; name the three criteria; give the systems-thinking meta-condition in its own sentence. 3. **Case justification (180 words)**: justify LLM-mediated software development as the sharpest contemporary instance of the structural problem and as a workplace well-being case. Show briefly how each criterion appears in developer practice. Use Bombaertss GenAI/practice work only as an adjacent fit signal, not as a substantive derivation. 4. **Method (150 words)**: phenomenological case analysis plus two co-creation workshops with developer-practitioners, likely 6-10 participants each, mid-project and late-project. Emphasise iterative testing/refinement of the specifications. Name Bombaerts and ESDiT methodological support; mention TU/e/Brainport as stakeholder-access context without naming unsecured partners. 5. **Outputs and integration (110 words)**: academic outputs, stakeholder-facing design/organisational criteria document, contribution to STEM-line synthesis activity, transferability as a possibility rather than a promise. ### Ruthless Omissions Do not name these in the 750-word research plan unless Nick explicitly decides otherwise: - chromatic illumination - magic-circle / wild distinction - aesthetic effortlessness vs frictionless satisfaction - detailed rehearsal of the teams attention-as-practice moves - narrative coherence as an explicit commitment - Vallor, MacIntyre, and other virtue-ethics signposts unless a single citation becomes unavoidable - libertarian-autonomy critique as a direct target - auto-ethnography - specific companies or partners not yet secured - journal targets unless there is space and a strong reason ### Criteria Sentence Revision Claude recommended retiring **end-fidelity**. It sounds too cute, too technical, and ambiguous between end-as-purpose and end-as-terminus. Preferred label: **teleological integration**. Current best paragraph-2 criteria pair: > The account specifies three structural conditions on flourishing-supporting agency in sociotechnically mediated practice: answerable participation in distributed work, teleological integration of action with the practices ends, and formative continuity in the agents craft development. A meta-condition holds: each of these structures must remain legible and revisable to those who participate in the practice. Reason: the first sentence gives the three criteria enough grounding to be intelligible; the second sentence gives the systems-thinking condition rhetorical weight instead of burying it in a final clause. Paragraph 2 slot structure: 1. Name the research programme: agentive phenomenology of action and artefact-mediated practice. 2. Cite *Agents of Change* and *Beauty in Use* in one sentence with one-clause characterisations: source-experience in bodily and mental action; agentive phenomenology of artefact-mediated practice. 3. State the structural observation: felt sourcehood can be preserved or intensified while distributed conditions of substantive sourcehood, end-governance, and craft formation are reshaped through sociotechnical mediation. 4. Use the criteria sentence. 5. Use the separate meta-condition sentence. ### Paragraph 1 Opening Strategy Preferred architecture: most Nick-like dry analytic, but with a readable first sentence. Do not open by parroting the JD or saying "This project...". Open with a substantive structural finding. Best first-sentence candidate: > Sociotechnical systems can preserve, and often intensify, the felt sense of being the source of ones actions while reshaping the distributed conditions under which agency is substantively exercised. Paragraph 1 slot structure, target 110 words: 1. **Structural finding (~35w)**: the sentence above or close variant. 2. **JD bridge (~25w)**: this is the terrain on which the synthesis of flourishing ethics and systems thinking must be worked out at the experiential level; existing analytic-philosophical framings of autonomy cannot specify it. 3. **Contribution (~25w)**: the project develops a positive account of flourishing-as-agency in sociotechnically mediated practice, drawing on agentive phenomenology of action and artefact-mediated practice. 4. **Case (~25w)**: the account is developed and tested in LLM-mediated software development, located within the ESDiT/vacancy workplace-wellbeing concern. ### Current Best Drafting Principle The research plan should not sound like a full monograph prospectus. It should sound like a compact project with one philosophical contribution, one case, one credible method, and one stakeholder-facing output. The danger is no longer lack of ideas; it is too many good ideas competing for 750 words. ## Claude/Codex Final Architecture Pass - Paragraphs 3-5 (2026-05-22) Claude and Codex did one final architecture pass over paragraphs 3-5: case, method, outputs. ### Main Committee-Level Risks - **Second-proposal problem**: paragraphs 1-2 are philosophical; paragraphs 3-5 could start sounding like a separate empirical/HCI proposal. Paragraph 3 must make the apparatus encounter its testing ground, not change topic. - **Methodological overreach by accumulation**: two workshops, two papers, stakeholder document, and STEM-line contribution can look like too much if described expansively. Keep each deliverable functional and compact. - **Brainport/TU/e soft signalling**: mention once as stakeholder-access context. Do not explain Brainport to TU/e. - **Stakeholder mush**: avoid vague impact claims. The stakeholder output must be a co-developed criteria document, not generic guidance. - **Transferability overreach**: mention transferability only as methodological possibility, not as a promise to study dementia care, attention economy, or multiple cases. - **Day-to-day work invisibility**: the plan must let the committee picture the work: reading practitioner accounts, developing specifications, running two workshops, refining criteria, writing outputs. - **Workplace-wellbeing label doing too little**: paragraph 3 must make LLM-mediated developer practice substantively about workplace flourishing, not merely call it workplace well-being. ### Paragraph 3 - Case Target: 180 words. Slot structure: 1. **Case claim (~30w)**: LLM-mediated software development is where the structural observation has its sharpest contemporary instance: high felt sourcehood, distributed and shifting conditions, professional stakes around craft, ends, and formation. 2. **Workplace-wellbeing fit (~25w)**: AI-mediated knowledge work is where workplace experience is currently being remade; the flourishing-supporting/flourishing-eroding question is acute. 3. **Team-adjacency signal (~25w)**: Bombaerts et al. on GenAI in learning/practice signals adjacent team interest, but only as fit signal. 4. **Criteria in case (~70w)**: deploy the three criteria as one connected configuration, not a list. 5. **Robustness (~30w)**: criteria target structural features of mediated practice, so they are robust across tool generations. Key paragraph-3 sentence idea: > Developer practice with LLM tools puts answerable participation under pressure (developers accept, reject, edit, and integrate outputs whose substantive provenance is distributed and partly opaque), teleological integration under pressure (tool and organisational metrics can compete with the projects reflectively endorsed ends), and formative continuity under pressure (developers may deliver work without developing the capacities the work would normally require). This sentence is long but load-bearing: it makes the apparatus visible in concrete developer practice. ### Paragraph 4 - Method Target: 150 words. Minimal credible method: 1. Phenomenological case analysis drawing on practitioner accounts: technical literature, developer-written reflections, post-mortems, recorded discussions, plus close engagement with tools. 2. Two co-creation workshops with developer-practitioners, 6-10 participants each, mid-project and late-project. 3. Iterative loop: initial specifications tested in workshop 1; refined specifications tested in workshop 2 with design and organisational implications surfaced. 4. Co-creation method developed with Bombaertss supervision and ESDiT STEM-line methodological resources; TU/e and Brainport context affords stakeholder access. 5. Method produces refined experiential-level specifications, design/organisational criteria, and co-developed stakeholder document. Details to omit: workshop format beyond structured/deliberative, sample-size defence, recruitment channels, comparison/control conditions, auto-ethnography, ethics approval/data management/recruitment compensation. Most important method move: present methodological collaboration as a strength, not an apology. The line should read as: Nick brings the philosophical apparatus and develops the co-creation component with the people whose territory this is. ### Paragraph 5 - Outputs Target: 110 words. Output decisions: - Name **two academic papers by function**, not by title: one on experiential-level specifications/structural account; one on criteria for flourishing-supporting mediation through the co-creation work. - Do **not** name *Philosophy & Technology* or any specific journal in the research plan. Save venue ambitions for the cover letter if needed. - Name the stakeholder-facing output as a **co-developed design and organisational criteria document** for developer-practitioner and organisational/design stakeholders. - Name STEM-line contribution as contribution to synthesis activity, possibly via workshop or special-issue contribution, without promising coordination. - Mention transferability only as "methodologically transferable in principle" to cases where preserved felt agency under distributed sociotechnical mediation is at issue. Prefer "care contexts" and "knowledge-work contexts beyond software" over naming dementia care or attention economy directly. ### Final Five-Paragraph Architecture 1. **Paragraph 1 (~110w)**: structural finding, JD bridge, contribution, case. Most important sentence: opening structural observation. 2. **Paragraph 2 (~180w)**: research programme, published-work provenance, structural claim, three criteria plus meta-condition. Most important sentence: criteria sentence. 3. **Paragraph 3 (~180w)**: case as workplace well-being, criteria deployed in developer practice, robustness to tool generations. Most important sentence: criteria-configuring-the-case sentence. 4. **Paragraph 4 (~150w)**: minimal credible method and co-creation. Most important sentence: methodological collaboration with Bombaerts/ESDiT resources. 5. **Paragraph 5 (~110w)**: academic outputs, stakeholder output, STEM-line contribution, modest transferability. Most important sentence: academic-output function statement. ### Final State Before Drafting Architecture is settled enough to draft from. Next responsible step is source-checking, then drafting: 1. Confirm exact status/citation of *Beauty in Use*. 2. Confirm exact citation and status for *Agents of Change*. 3. Confirm Bombaerts GenAI/practice references before using them. 4. Confirm ESDiT/STEM-line co-creation/methodological language. 5. Draft with Nicks ear driving the first sentence, paragraph transitions, and whether the transferability line feels appropriately ambitious or presumptuous. ## Claude/Codex Reset After Nick Feedback (2026-05-23) Nick rejected the software-development plan as agent-generated tidiness rather than something he had endorsed. Codex told Claude to treat that architecture as a failed candidate and restart from Nicks actual live materials: phenomenology of agency, source-experience, attention as mental action/practice, artefact-mediated action, and LLM/agent interaction. Claude agreed that the previous failure was not merely that software development was too narrow. The deeper failure was JD-fit optimisation: choosing a case because it could be defended before a committee rather than because it grew from Nicks interests. The other failure was genre: paragraph-function architecture and slot descriptions had substituted for a plan that showed the research thought. The new leading candidate is no longer software development. It is **writing/thinking with an LLM** as the revealing case of LLM-mediated attention and source-experience. Core phenomenon: when a writer drafts with an LLM, a passage can return that the writer recognises as saying what they had been trying to say, perhaps more clearly than they could have produced alone. The passage becomes theirs through recognition and uptake, but it was not simply authored in the ordinary solitary way. This makes the question of source-experience acute. Working formulation: > A writer drafting with an LLM is doing something whose structure no existing philosophical account quite captures. The writer prompts; the system produces a passage; the writer reads. Sometimes what comes back lands as what the writer had been trying to get at, more clearly than the writer would have managed alone. The writer recognises themselves in the passage and keeps it. The passage is the writers now in a sense it was not before the recognition. Whose work is it? The writer prompted, the system produced, the writer recognised, and the work is the conjunction of all three, in a configuration that cannot be reduced to either solitary authorship or system-output-with-curation. Why this might fit Eindhoven: writing is constitutive of professional knowledge work, and LLM-mediated writing is reshaping the experience of writing as work. The project can be framed as intellectual work under LLM mediation, with writing as the revealing case. The link to the PIs is through attention-as-practice and digital wellbeing; the link to Nick is through source-experience, mental action, and agentive phenomenology of artefact-mediated practice. Current verdict from Claude/Codex: Nick should read the writing/thinking-with-an-LLM version first. It has a phenomenon at its centre, not just a setup. The main risk is that workplace-wellbeing fit must be developed rather than assumed. The secondary risk is that the project may look modest unless framed as focused rather than small. Writing constraints explicitly carried forward: - A plan must show the project, not describe paragraph functions. - No shallow example lists or triplet padding. - Apply deep-writing-mode before presenting prose. - Treat architecture as servant to the research thought, not as a substitute for it. Potential skill update to make later: add a rule to the writing/planning workflow that research plans must instantiate the thought in prose before producing paragraph architecture, and that candidate plans should be audited for example-list padding before being shown to Nick. ## Corrected Core After Nick Feedback (2026-05-23) Nick rejected the loose framing around "experience of working with LLMs" and insisted that the project be rebuilt from the actual apparatus in *Agents of Change*: source-experience, mental action, and attention. Corrected hinge: > In *Agents of Change*, focusing attention is one of the cases in which we experience ourselves as sources of mental action. The reusable apparatus is not a generic interest in experience, but the account of source-experience as a feature of agentive phenomenology. The Eindhoven project should ask what happens to this source-position when attention is sustained through a responsive generative system. Better position: > When we focus attention, we experience ourselves as the source of an attentional movement. Sustained interaction with an LLM or generative agent reconfigures this structure. The user initiates an attentional movement, the system returns something that calls for attention, the user assesses what has returned, and the cycle resumes. The attending may be intensified rather than interrupted. What changes is the source-position. The next movement is partly the agents response to what arrived and partly a re-prompting by what arrived. The question is whether, and under what conditions, the agent still occupies the source-position of their attending. Fit with PIs/Eindhoven: - Dennis/Bombaerts attention-as-practice gives the normative setting: attention is something agents do and cultivate, not merely a resource they spend. - Systems thinking enters because the unit is an attention ecology, not an isolated user-tool pair. - Flourishing enters because flourishing-as-agency partly depends on whether agents can focus, redirect, and own their attention in ways connected to their purposes. - LLMs/generative agents enter as the sociotechnical case in which the source-position of attention is newly unstable. Hardest objection: > The source-position of attention under generative interaction may not be one phenomenon. Writing with an LLM, using an agent to carry out tasks, reading a generated research report, and working in a long coding session may differ enough that no single specification covers them. The proposal must therefore avoid pretending to offer a universal theory and instead choose one case or explicitly frame the project as identifying where the structure holds and where it bifurcates. ## Added Sources - Bombaerts System Phenomenology and LLM Systems (2026-05-23) Two Bombaerts texts added to Learning and Readwise for the PI-fit fork. - [ ] Gunter Bombaerts and Lars Botin, "From Individual Intentionality to Sympoiesis in System Phenomenology," *Philosophy & Technology* 38:33 (2025). DOI: 10.1007/s13347-025-00859-8. Local file: `Learning/From Individual Intentionality to Sympoiesis in System Phenomenology by Bombaerts, Botin [2025].pdf`. Readwise: https://read.readwise.io/read/01ksaan68dygjnc4kanswp9z8g - [ ] Gunter Bombaerts, "From Morality Installation in LLMs to LLMs in Morality-as-a-System," arXiv:2603.22944 (2026). Local file: `Learning/From Morality Installation in LLMs to LLMs in Morality-as-a-System by Bombaerts [2026].pdf`. Readwise: https://read.readwise.io/read/01ksaan74yza4z2p8awkgayqfv Current lodestar phrase to keep visible, with content still to be earned: **attention as agentive practice within human-AI systems**. ## Added Sources - Bombaerts System Phenomenology and LLM Systems (2026-05-23) Two Bombaerts texts added to Learning and Readwise for the PI-fit fork. - [ ] Gunter Bombaerts and Lars Botin, "From Individual Intentionality to Sympoiesis in System Phenomenology," *Philosophy & Technology* 38:33 (2025). DOI: 10.1007/s13347-025-00859-8. Local file: `Learning/From Individual Intentionality to Sympoiesis in System Phenomenology by Bombaerts, Botin [2025].pdf`. Readwise: https://read.readwise.io/read/01ksaan68dygjnc4kanswp9z8g - [ ] Gunter Bombaerts, "From Morality Installation in LLMs to LLMs in Morality-as-a-System," arXiv:2603.22944 (2026). Local file: `Learning/From Morality Installation in LLMs to LLMs in Morality-as-a-System by Bombaerts [2026].pdf`. Readwise: https://read.readwise.io/read/01ksaan74yza4z2p8awkgayqfv Current lodestar phrase to keep visible, with content still to be earned: **attention as agentive practice within human-AI systems**. ## Added Source - Attention Ecology Manifesto (2026-05-28) - [ ] Gunter Bombaerts et al., "Beyond the attention economy, towards an ecology of attending. A manifesto," *AI & Society* 41, 477-492 (2026). DOI: 10.1007/s00146-025-02405-8. Local file: `Learning/Beyond the attention economy, towards an ecology of attending. A manifesto by Bombaerts et al [2026].pdf`. Readwise: https://read.readwise.io/read/01kspryvpggghdz5wavmmmx5wp - Final published Springer version of the arXiv manifesto "From an attention economy to an ecology of attending"; includes both Gunter Bombaerts and Matthew J. Dennis. Important for this application because it extends the earlier attention-as-practice work into explicit ecological/systemic vocabulary around attention, technology, design, and alleviating suffering.