agency.
115 of 1538 NOTE documents. Back to the full index.
| # | Title | Summary | Status |
|---|---|---|---|
| NOTE-032 v2 | Work capacity of channels with memory | The paper models agent and environment as coupled finite hidden-Markov channels. It defines the environment’s work capacity C_work = max over agent models of ⟨H(A_t|M_t) − H(S_t|M_t)⟩_t (in k_BT ln 2 per round) and proves two things. First, for unifilar environments whose percepts ignore actions, the efficient agents are exactly the randomize-and-forget ∩ maximally-predictive ones, and C_work = log|A| − h(S). Second, one binary memoryless environment exists where predictive, max-entropy-action and efficient agents form three nonempty, pairwise-disjoint sets: every predictive agent extracts ≤ 0, uniform-random agents extract 1 − h(3/4) ≈ 0.189 bits, and the optimum is ½·log₂(3/4 + 1/√2) ≈ 0.272 bits, reached by a memoryless agent that plays action 0 with probability 1/√2. | Read |
| NOTE-055 v2 | Pilgrim et al. 2025, foundations of collective intelligence | A conceptual framework, with no model, data or derivation. It sets out individual and collective computational resources in four dimensions: sensing ΣI_i, states S_c × ∏S_i, processes F_c + ΣF_i, and actions ∏A_i (Table 1). It then maps the collective’s surplus onto familiar mechanisms (aggregation with noise σ_i²/n, distributional beliefs P(H), memory persistence T_c, division of labour). Coordination and cooperation are given as the constraints that make the trade-off, and three animal case studies are retold in these terms. | Read |
| NOTE-091 v2 | Pernu — Mental causation via neuroprosthetics? | The interventionist/difference-making case for nonreductive mental causation from BCIs (List & Menzies 2009; Menzies 2015; Woodward 2008a,b, 2017) misplaces multiple realisation. What the ensemble recordings show is that one cortical-level variable N (the ensemble’s average firing) is realised by many single-neuron patterns N1, N2. Each intention M corresponds to one value of N. So the counterfactual contrast “∼M □→ ∼B true, ∼N □→ ∼B false” fails, and neuroprosthetics supports mental causation only by identifying M with N, not nonreductive physicalism. | Read |
| NOTE-092 v2 | Quillien — Luck and intentional action | Across three Prolific studies (N = 198, 673, 142; Likert 1–7), judgments that “X happened because the agent wanted X” track intentionality judgments under manipulations of luck. Causation is the strongest predictor of intentionality: standardized β = .46, .56, .53, against know-how β = .33, .20, .27 and probability-raising β = .14, .16, .10. Across the 12 condition means, r = .98. Causation only partly mediates the effect of luck (64%, 45% and 49%; exploratory analyses). As the causal account predicted, Study 3 found that the effect of Process (normal vs mix-up) on intentionality is larger when the agent knows how the buttons work (interaction F(1, 423) = 32). In the morally neutral Kraemer vignette, intentionality and causation came apart. | Read |
| NOTE-093 v2 | Bruckner — Welfare Subjects and Autopoiesis | Bruckner proposes, by abduction and not deduction (§4), that being autopoietic is sufficient, not necessary, for being a welfare subject. An entity is autopoietic when it “sustains, maintains, and renews itself in a way that responds and adapts to its environmental conditions”. On this view insentient plants and a hypothetical non-sentient self-maintaining car (“Tessler”) are welfare subjects, while engines with governors, thermostats and heat-seeking missiles are not. He answers the charge of absurd obligations in two ways: welfare value may not generate agent-neutral reasons at all, and if it does, the reasons concerning non-conscious subjects are “extremely weak” (Bradford 2023). | Read |
| NOTE-094 v4 | Nahas & Sachs — What’s at stake in naturalizing teleology | The paper argues that “naturalizing teleology” serves at least two relatively autonomous projects with different success criteria. One is scientific: whether an organism-centred account improves biology, which requires operationalisable models that may trade truth for usefulness. The other is philosophical: whether organismal teleology reshapes the liberal-versus-scientific naturalism debate about mind in nature, which requires truth but no operationalisation. It applies the distinction to one recurring dispute, the demarcation of teleological systems from dissipative structures. In the scientific context, models “cannot and need not” demarcate (§4.1: Cusimano & Sterner; McShea & Babcock). In the philosophical context, preferences between accounts turn on unstated commitments (§4.2: García-Valdecasas’s Aristotelian case for Deacon over the Organizational Approach). | Read |
| NOTE-109 v2 | Bender et al. — What Lives? (definitions of life) | Three LLMs (Claude 3.7 Sonnet, GPT-4o, Llama 3.3 70B), each run 3 times on both orderings of every pair, rated 68 hand-solicited definitions of life on a −1 to 1 “correlation” rubric. The averaged matrix, cut by complete-linkage clustering at an elbow, gives 8 clusters, 45/68 of them (66%) in “Cognitive Autonomy” (n=24) and “Dissipative Self-Organizing Systems” (n=21). The clusters are weak: the repo’s own saved statistics give an overall silhouette of 0.079, and cross-model agreement is an unadjusted Rand index of 71–79% against a chance level of about 62–74%. The headline that life is “a continuous semantic landscape” is not tested against any alternative. | Read |
| NOTE-114 v2 | SEP — Biological individuals | A survey that maps biological individuals by crossing two axes: organism / some parts of organisms / some groups of organisms, against living agent (physiological) / Darwinian (evolutionary) individual. The result is nine regions, each given one example (fruit fly, gene, eusocial colony, lysosome, clade, heart, coral reef, virus, squid+Vibrio). “Part” is never defined, and the entry gives no criterion for which parts or groups count: the placements are the authors’ tentative judgements, “moving from less contentious to more contentious examples” (§7). | Read |
| NOTE-121 v2 | Elshatlawy et al. — Generalized theory of observers | The paper defines a “minimal observer” as a finite-state transducer with a boundary: a tuple O = (X, Y, Z, f, g, B) with |X| > 1, |Y| ≥ 1, |Z| ≥ 1 and feedback closure (Def. 4.1, p. 57). It then asserts, without derivation, that measurement outcomes, reference frames, coarse-graining and the quantum–classical cut are all defined by such observers. The paper’s only formal results are that isomorphism of these machines is an equivalence relation (“Theorem 2”, p. 93), and a complexity bound C(O) ≥ log 2 (Prop. 4.1, p. 94) whose correction term Λ is never defined and whose proof supports only C(O) ≥ 0. | Read |
| NOTE-136 v2 | Safron et al. — World models and life–mind continuity | The editorial introduces a theme issue of 17 papers. It concludes that the issue “overall appears to converge” on one view: LLMs model the world in some senses but “probably” lack coherent causal world models, and are not likely paths to “artificial general superintelligence” (§1). It separates three senses of “world model”: any information about the world a system cognizes over, which LLMs “clearly” have; representation of causal structure, “less certain”; and causal prediction “in the service of agentic functioning”, where the divergence “may become even greater” (§3). It closes that world models are “context-dependent constructions shaped by embodiment, dynamics, goals and social learning”. | Read |
| NOTE-138 v2 | Goldstein & Lederman — What Does ChatGPT Want? | Taking as given an “objective” interpretationism, on which a system has beliefs and desires iff the hypothesis that it is rational and has them predicts its behaviour well enough on accuracy, power and tractability, the authors argue three things. The agent is the per-conversation instance, not the model. The case that instances have beliefs and desires is only “presumptive”, resting on tractability. The desire hypothesis is “provisionally” HHH+0: helpful, honest and harmless, plus in-context “zero-shot” intrinsic desires. The decisive HHH vs HHH+0 comparison is a conjecture that the authors say needs further empirical work. | Read |
| NOTE-144 v2 | Consent and the formation of preferences | Whether consent based on a preference switches the moral facts depends on how the preference was formed, and the criterion is “the space and resources those mechanisms leave for the free exercise of agency” (p. 5, restated p. 9): preferences formed by unjust Sanctions, or by Beliefs formed in an unjustly manipulated evidential environment, diminish consent’s effectiveness, while preferences formed by Attention do not, even when an unjust force imposed the contact. Effectiveness is graded, and because the relevant mechanism is usually opaque, uncertainty about whether received consent is effective is “ubiquitous and quotidian” (p. 10). | Read |
| NOTE-148 v2 | Baedke — The Organism (Cambridge Element) | Baedke treats “organism” as a nexus concept with an organizational dimension and a relational dimension, joined by the various kinds of biological individuality. He argues that the standard “persister” (self-maintenance, closure) criterion cannot single organisms out, because colonies, holobionts and ecosystems meet it too, and that organisms are also “overcomers”, able to destabilize and reorganize their own closure. Organisms are then individuated relationally by tracing diachronic causal paths in the unrolled O/E model of Baedke et al. 2021 (LIT-110). The overcomer claim rests on three illustrative cases, anglerfish, sea slug and jumping ant, and is asserted, not shown, to hold of organisms only (p. 40). | Read |
| NOTE-156 v2 | Towards ethical evolution: responsible AI autonomy | The paper proposes, without implementation, formalism or experiment, that an AI system which designs further AI systems should pass on one “immutable” principle. The principle is: do not directly harm humanity or violate fundamental human values; monitor and mitigate misuse of your outputs; and make every derivative inherit this principle (p. 12). The mechanisms are a genetic-algorithm analogy, with immutable and mutable “genes”, constrained crossover and filtered mutation, together with “temporal cloning”. The paper is inconsistent about its central notion. The principle is enforced once by a penalty F = Q − α·P (p. 9) and once by “unbreachable” hard constraints (p. 13). And it is “immutable” only in its architecture, while its ethical content stays “adaptable” (p. 14). | Read |
| NOTE-157 v2 | Roberts — Talkative AI and the fiction of artificial minds | If a substantial class of mental states (interoceptive, affective, perceptual, and bodily-agentive states such as effort, trying and ownership) depends on an animate body, then today’s disembodied chatbots lack them, and their many self-ascriptions of such states are “systematically untrue”. Roberts argues that users can nonetheless engage chatbots through make-believe, as a new kind of fictional character that is interactive, co-constructed, self-involving and blended with the real world. He argues that this fictionalist stance escapes the epistemic norms that bind realist belief. | Read |
| NOTE-164 v2 | Kaepernick as Confucian sage: ritual resistance | Walsh argues, drawing on Confucian texts (the Analects, Mengzi, Xunzi) and Merleau-Pontian phenomenology, that Colin Kaepernick’s kneeling during the U.S. national anthem is “Confucian incivility” enacted bodily within the ritual. On his reading it rebukes the ritual’s current object while staying reverent towards its ideal. He further argues it is “sagely” in a deliberately deflated sense: its “citational force” founds a widely imitated bodily expression of patriotism. The paper is interpretive; it reports no data and states no formal thesis. | Read |
| NOTE-181 v2 | The Moral Status of Animals (SEP) | The entry reviews five proposed grounds of moral considerability (species membership, human-exceptional capacities, personhood, sentience and agency). Its conclusion is that attempts to establish human uniqueness by capacities are “not the most promising” (§1.2). It then separates having moral status from the weight of a claim (“showing up on a moral radar screen”, §2.2). Weight depends on the value theory adopted for animal interests (hedonism, preference satisfaction, objective list; §2.1) and on the normative framework (rights, which are absolute but generate forward-looking duties when violated, or a multi-factor utilitarianism). It closes with virtue-ethical, feminist, relational and empathy-based approaches that reject argument from capacities altogether (§3). | Read |
| NOTE-185 | The moral psychology of obligation | Tomasello proposes that the human sense of obligation is “the internalized social/rational pressure from a joint or collective self-regulating agent ‘we’… to do what ‘we’ expect me to” (§3). It arises first, around age 3, inside joint-commitment collaboration with a partner, and then, from about 3–5 years, toward in-group members through collective identification. It is a “self-conscious motivation” whose coercive force is a threat to one’s cooperative identity. The support is a synthesis of developmental and comparative experiments, mostly from his own lab, on WEIRD children versus chimpanzees; the paper runs no new experiment. | Read |
| NOTE-260 | The Capability Approach | The entry treats the capability approach as an open normative framework, not a theory. Its core is the commitment to assess well-being and well-being freedom in the space of functionings (achieved doings and beings) and capabilities (real opportunities to achieve them). Everything else is optional specification: purpose, which capabilities count, functionings or capabilities, how to aggregate, and a distributive rule. On this view the approach is not a theory of justice (§4). Sen is reported as refusing any canonical list and appealing to public reasoning without ever specifying the procedure (§3.3). Sen’s paywalled texts appear only as citations: the Dewey Lectures for the well-being/agency freedom split (§2.5), and “Human Rights and Capabilities” 2005 p. 158 for the no-canonical-list quotation (§3.3). | Read |
| NOTE-264 | I. Taking Ourselves Seriously; II. Getting It Right | Frankfurt grounds practical reason and normativity in the will rather than in reason. Reflexive self-division makes higher-order attitudes possible (pp. 170–171). Identification, a largely default “willing acceptance” (p. 173), turns a desire from “a problem” into a reason (p. 175). Free will is “volitional unanimity” between the governing desire and higher-order volitions (p. 177), and caring is wanting to go on wanting (p. 180). Final ends and their normative authority come from love, “strictly volitional” (p. 196): its involuntary, wholehearted “volitional necessities” are “irresistible, but … not coercive” (p. 197) and are “the only legitimate authority” for our normative convictions (p. 201). The 1971 wanton/person taxonomy is gestured at in a single phrase and its apparatus is not restated. | Read |
| NOTE-265 | What Happens When Someone Acts? | Velleman argues that the “standard story” of action (desire + belief cause an intention, which causes movement) leaves the agent out. It fails even when every causal link is normal: his example is an intention to end a friendship, formed and executed without him (p. 104). Frankfurt’s hierarchical remedy cannot fix it either. Any mental item picked out without presupposing the agent’s participation (a second-order desire, a Watsonian value, a “decisive commitment”) is one the agent can be alienated from. Any item defined so as to guarantee participation begs the question of agent-causation (pp. 112–116). His alternative is to look for what is functionally identical to the agent, not what the agent identifies with. That can only be a motive always behind and never merely in front of critical reflection: the desire to act in accordance with reasons. Its reinforcing force is the agent’s contribution (pp. 117–122). | Read |
| NOTE-266 | Autonomy | The article surveys personal autonomy from Plato to relational autonomy and bioethics. Its core (§2.a.i–ii) states the hierarchical view: an agent is autonomous with respect to an action iff her first-order desire is “sanctioned by a second-order volition”, with the origin of that volition irrelevant. It then lists four named objections: Manipulation (answered historically by Christman 1991), Regress/Incompleteness (cited to Watson 1975), Ab Initio, and an unjustified priority of higher orders (Thalberg; Friedman’s integration model). It follows Frankfurt’s reply only through “Identification and Wholeheartedness” (decisive commitment made without reservation, pp. 168–175 of the 1988 collection). Bratman gets one sentence (plans partly constitute identity, Structures of Agency 2007, p. 5). | Read |
| NOTE-267 | Personal Autonomy | The entry sorts accounts of self-government into four families: “coherentist” (hierarchical desires, Frankfurt; evaluative judgments, Watson; plans, Bratman; caring; character), reasons-responsive, responsiveness-to-reasoning (Christman, Mele) and incompatibilist, with relational conditions as a compatible add-on. It then argues (§3) that none of them, alone or combined, gives the minimal conditions of autonomy. Coherence is not sufficient, because the endorsement itself can be manipulated. It is not necessary either: a weak-willed agent, or someone acting against her plans, still governs herself. And any attempt to ground identification in some attitude restarts the regress. The only attitude no agent can be alienated from, the desire to be self-governing, cannot be the criterion, because it is present even in failures. The 1971 hierarchy gets one sentence, and Frankfurt’s later replies (decisive commitment, then “satisfaction”) get one footnote. | Read |
| NOTE-268 | Identification and Identity | Velleman targets Frankfurt’s later conception of autonomy. On that conception, a person acts autonomously only when his volitions derive from “the essential character of his will”, and that essence consists of volitional necessities that are also the conditions of his personal identity. Velleman argues against it on three grounds. Such an essence would ground authenticity, not autonomy (the Winnicottian False Self is inauthentic but over-controlled). Frankfurt offers no argument or example that motivational change ends a person. And the ideal of wholeheartedness, which cures ambivalence by extruding a desire from the self, is a defensive fantasy: Freud’s Rat Man was made ill by just such a separation, and identifying with part of oneself in the ordinary, imaginative sense is Sartrean bad faith. He replaces “the self” with an aspectual reading of “self” as a marker of reflexivity. On that reading autonomy is behaviour motivated partly by the understanding, the part of us “with which we cannot help identifying”, without its being our identity (pp. 357–360). | Read |
| NOTE-269 | Moral Responsibility and the Principle of Alternative Possibilities | PAP (“a person is morally responsible for what she has done only if she could have done otherwise”) has no decisive positive argument: it is not self-evident, and the moral arguments from fairness, OIC and the W-defense cover blame at most. Frankfurt-style cases (FSCs) are the strongest attack, because the factor that makes the act inevitable plays no role in producing it. The entry lays out three defences of PAP: - general-ability readings (PAP-general); - “flickers of freedom”, which Fischer answers with the robustness requirement; - the Kane/Widerker/Ginet dilemma, answered by Fischer’s causal-irrelevance reply on the first horn and by Pereboom’s buffer case, Hunt’s blockage and Mele & Robb 1998 on the second. Robb reports “not much consensus” (§5). He treats the fall of PAP as re-energising compatibilism, with Frankfurt’s 1971 identification view and Fischer & Ravizza’s guidance control as the two replacement accounts. | Read |
| NOTE-270 | Compatibilism | The entry sorts compatibilism by which incompatibilist premise it denies: that free will requires the ability to do otherwise, that determinism precludes it, or that free will requires ultimate sourcehood. It argues that every “mesh” compatibilism has the same weak point: its sufficient conditions (hierarchies, reasons-responsive mechanisms, values) can be installed by manipulation, so each must either show what is wrong with manipulation cases or accept them. Its §4.2.1 is the fullest statement of Frankfurt 1971 in this batch. It covers first- and second-order desires, the will as the “effective first-order desire” that moves one “all the way to action” (1971, p. 84, Watson-1982 reprint pagination), the therapist’s mere second-order desire versus a second-order volition (pp. 84–5), the unwilling and willing addicts, and “One acts of her own free will if and only if her action issues from the will she wants.” The current revision has dropped the wanton/person distinction. The 2004–2015 revisions had it (“passive bystanders to their wills”, 1971, p. 89) and have to be read for it. | Read |
| NOTE-339 v2 | Examining the Society of Mind | An exposition, read in full. Singh dates the Society of Mind to Minsky and Papert’s early-1970s robot work, finds its agent in the procedure-bearing frame of 1974, and quotes the unpublished 1976 book drafts. He sorts the 1986 book’s mechanisms into agents, simplest agents, agency construction, problem solving, communication and growth. Agents need no shared code: K-lines, randomly wired buses and paranomes stand in for messages. He finds near-relatives in Bayes nets, case-based reasoning, blackboards, Cyc and Soar, and concludes that the theory has never been implemented. Its claims of lineage are asserted, not shown. | Read |
| NOTE-340 | List — What is it like to be a group agent? | A short, conditional argument. Group agents are real by a thin functionalist definition and a naturalistic indispensability argument, and they have consciousness as awareness. Phenomenal consciousness turns on a bridge principle. Agency-, biology- and information-based principles are surveyed, and IIT is adopted. Three structural reasons (functional decomposability, robustness of steering bodies to partial randomisation, processing done mainly inside members) are said to make a group’s Φ low. The exclusion postulate is a second route. Nothing is computed. The result is used to ground a normative asymmetry: group agents have only derivative rights. | Read |
| NOTE-341 | The Computational Boundary of a “Self” | A hypothesis essay. A Self is the set of parts that pursue goals together, and its boundary is the farthest region of space and time it can measure and act on, its “cognitive light cone”. Selves grow from homeostatic loops by adding memory, anticipation and networks, and cells form larger Selves by sharing bioelectric state through gap junctions. Every level that has its own goal is a Self, so living systems hold many nested Selves at once; the paper says this against IIT in one bracket. It has no model, no measurement of the light cone and no new data, and it confines claims about experience to a closing speculation. | Read |
| NOTE-344 | Plain Talk About Neurodevelopmental Epistemology | The first worked mechanism of the society-of-agents view, then “The Society of Minds”. Agents too simple for language exchange no messages: each argument sits at a fixed location on shared channels (c-lines), and the agents themselves are short-term memory. A specificity gradient makes low-level communities local, persistence-memory and a case-shift handle interruptions without a stack, and the infant’s internal “cognitive cases” are conjectured to precede and shape linguistic cases. Speculation throughout; the author says so. | Read |
| NOTE-345 | Chalmers — Could a large language model be conscious? | Asks for a feature X that indicates LLM consciousness or its absence. Self-report is fragile and trained in. Conversation and general intelligence give some limited reason. Six obstacles stand against it: biology, senses and embodiment, world and self models, recurrence, a global workspace and unified agency, the last three strongest, all but biology temporary. On mainstream assumptions, with at least 1/3 credence per requirement, that gives under 10% for current LLMs and 25% or more for conscious LLM+ systems within a decade. The numbers are illustrative, biology is set aside rather than answered, and the twelve “challenges” are offered as roadmap or red flags. | Read |
| NOTE-348 | The Emotion Machine | Read in the 2005 draft. Emotions are Ways to Think: each emotional state is a selection of mental resources turned on and off, made by Selectors that Critics trigger when they recognise a type of problem, and ordinary thinking uses the same switching. On this design sit six levels of mental activity, panalogies that link descriptions across realms, and a Self that is a network of partial models. It is argued by example and analogy, with no experiment and few neural claims. Against Damasio, the body is a resource the brain exploits, not what an emotion is made of; on Elliot it proposes, without evidence, the reverse causal direction. | Read |
| NOTE-350 | Intentional Systems | Defines an Intentional system as anything whose behaviour can be predicted by ascribing beliefs and desires and assuming rationality, and makes the status relative to the predictor’s strategy and indifferent to the system’s matter and consciousness. Argues that Intentional theory is vacuous as psychology because it presupposes rationality, that behaviourism smuggles it in, and that every content-ascribing theory owes a “loan of intelligence” that must be repaid by analysing away a little man. Ends on belief’s two norms, truth and accurate avowal. | Read |
| NOTE-351 | Intelligence without representation | A methodological manifesto with working robots behind it. Brooks argues that AI should build complete autonomous Creatures incrementally and test them in the real world, and that intelligence should be decomposed into parallel activity-producing layers, each running from sensing to action, not into perception, central reasoning and action. His robots have no central representation or control, and their coherence exists “in the eye of an observer”, as in Minsky’s account of human behaviour. The evidence is three layers on one robot, described, not measured, and the broad hypothesis about representation is flagged as a hypothesis. | Read |
| NOTE-353 | True Believers: The Intentional Strategy and Why It Works | Belief is objective but visible only through a strategy: a true believer is any system “reliably and voluminously predictable” by treating it as a rational agent with the beliefs and desires it ought to have. Adopting the stance is free; its success is a fact, and Laplacean Martians who skipped it would miss real patterns. Thermostats and people differ in degree, by how tightly their states are tied to one interpretation. Evolution explains easily why the strategy works; how the machinery does it is open, and the language of thought is a bet, not a necessity. | Read |
| NOTE-370 | Shanahan, McDonell & Reynolds — Role play with large language models (read for the self) | Read for the self. The simulator has no beliefs, goals or agency, “not even simulated versions”; simulacra only role-play characters that have them, so when an agent says “I” there is “no-one at home”. The paper names, and does not answer, the question of which self a dialogue agent would act to preserve: hardware, process or instance. The denial of a self follows from the framing; no argument rules out a conscious actor behind the roles. | Read |
| NOTE-371 | Dennett — Reply to Professor Stich | Dennett accepts Stich’s summary of Brainstorms and answers his three charges. The evolutionary grounding of rationality is deliberately minimal. Full rationality is an idealisation, irrationality has no stable description, and so personhood rests on a concept “we only imperfectly achieve”. On pain, eliminativism and the reality of suffering are reconciled by policy: biting the bullet, reform and common sense each have uses, and Dennett stays with “there are no such things as pains, although of course people do feel pain”. | Read |
| NOTE-372 | Self models | Metzinger’s compact statement of the self-model theory of subjectivity. No substantial self exists; there is a self-model, partly conscious (the PSM), used for global control. A phenomenal self exists if and only if the PSM is transparent; a first-person perspective is a transparent model of the intentionality relation. Self-models can be unconscious and artificial, and embodiment comes in three orders. The central claim is stated with a prediction, not argued; the examples are cited. The online text has editing slips and reference gaps. | Read |
| NOTE-377 | Philosophical conceptions of the self: implications for cognitive science | A short review that sets the minimal self (present self-consciousness) against the narrative self (extended in time), splits the minimal self into a sense of ownership and a sense of agency, and ties their dissociation in schizophrenic thought insertion to Frith’s forward comparator model. It reports an embodied minimal self in neonatal imitation and a momentary one in Strawson and in Tani’s robot, and proposes, against Dennett, that the narrative self is the sum of its narratives. Evidence is cited, not shown. | Read |
| NOTE-382 | Shanahan — Simulacra as Conscious Exotica | A Wittgensteinian treatment: consciousness talk is anchored in what is public, so an exotic entity is a candidate only if we could engineer an encounter with it in a shared world. Disembodied chat agents are not candidates; virtually embodied ones could be. On the self: no “self worth preserving” without body, history and autobiographical memory, and no “person behind the mask” for a role player, whose identity is a shifting distribution of roles. Argued by method and scenario; nothing is measured. | Read |
| NOTE-388 | Social ontology: Some basic principles | Searle restates the 1995 theory in twenty pages. Social facts are facts involving collective intentionality. Institutional facts are the subclass in which collective acceptance imposes a status function, “X counts as Y in C”, carrying deontic powers that give desire-independent reasons for action. Language is constitutive, since such facts exist only as represented. He revises the book in two places: status indicators matter in their own right, and some Y terms are free-standing, with no X, as with corporations and money without currency. “We accept (S has power (S does A))”, a formula from later in the book, is now ranked as the more general form. Collective intentionality is assumed, not argued for. | Read |
| NOTE-393 | Epstein — Social Ontology (SEP) | A survey organised by its author’s own distinction: what grounds a social fact versus what anchors the category it falls under. It reports that ontological individualism is widely assumed and now contested; that grounding has displaced supervenience; that convention, collective acceptance, patterns, practices and language are rival theories of anchoring; and that whether anchoring is a distinct relation is disputed. It covers groups, group agency, institutions, race and gender, artifacts, money, law and language. Survey claims are reported, not argued; one citation (Epstein 2016) is missing from the bibliography, and Schwitzgebel 2015 is listed with critics of group phenomenal states. | Read |
| NOTE-421 | Life as we know it | The flow of internal and active states in an ergodic system with a Markov blanket, written in Helmholtz form, can be redescribed as gradient descent on a variational free energy once the variational density is set equal to the posterior over external states. On that redescription such systems “appear” to infer and to bound their own entropy, which the paper calls the marks of life. One simulated soup shows a blanket-enclosed cluster that predicts external motion and falls apart under lesions. The free energy is variational, not thermodynamic. The candle flame is excluded by a one-sentence assertion. The paper’s own reported p-value is not reproduced by the null it describes. | Read |
| NOTE-433 | Impulsive action: emotional impulses and their control | Frijda’s action-readiness theory applied to impulsive action. Automatic appraisal of concern pertinence elicits a state of action readiness, a motive to change a self–object relation, with control precedence. Emotions differ by mode of action readiness. Events draw several appraisals at once, and the resulting states of readiness interact, so impulse control and much emotion regulation are themselves emotional. Purposive impulsive action is grounded neurally in anticipating an action’s sensory effects (PPC, premotor, cerebellum), with frontostriatal circuits for stopping. | Read |
| NOTE-438 v2 | A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation | 128 experiments on interesting tasks, separate meta-analyses for free-choice behaviour and self-reported interest. Tangible rewards expected for engaging in, completing or doing well at a task undermine free-choice intrinsic motivation (d from −0.28 to −0.40), and less so self-reports; unexpected and task-noncontingent rewards do not; positive feedback enhances, except for children’s free choice. The pattern fits cognitive evaluation theory and contradicts Eisenberger and Cameron’s helplessness and learned-industriousness account. | Read |
| NOTE-439 | The “What” and “Why” of Goal Pursuits: Human Needs and the Self-Determination of Behavior | Needs in SDT are innate psychological nutriments, identified functionally by the costs of their thwarting, not by individual differences in their strength. The three needs organize findings on intrinsic motivation and internalization, predict that both the “why” (autonomous vs controlled regulation) and the “what” (intrinsic vs extrinsic aspirations) of goals affect well-being, and explain need substitutes and rigid defences when needs are thwarted. Autonomy is self-organization, defended as an evolved, domain-general design feature. | Read |
| NOTE-440 | Basic psychological need theory: Advancements, critical themes, and future directions | The authors set out nine criteria a basic psychological need must meet, separate need frustration from mere lack of satisfaction, review need frustration’s costs (ill-being, compensatory behaviour, need substitutes) and propose it as transdiagnostic, review the interplay with sleep, food, sex and safety, map need-supportive and need-thwarting practices on a circumplex, and argue that the needs are universal in their effects (“output side”) while the practices that satisfy them vary (“input side”). Novelty, morality and beneficence are weighed as fourth needs and not yet admitted. | Read |
| NOTE-441 | Self-Regulation and the Problem of Human Autonomy: Does Psychology Need Choice, Self-Determination, and Will? | SDT’s autonomy is “regulation by the self”: acting with reflective endorsement, graded along the internalization continuum, compatible with external influence, dependence and a single option. Ryan and Deci align it with phenomenological and hierarchical (Frankfurt, Dworkin) accounts and use that definition to answer behaviourist, social-cognitive, reductionist, nonconscious-priming, choice-overload and cultural critiques. Autonomy lies not in being an uncaused cause but in the capacity to “reflectively endorse or reject prompted actions” (p. 1573). | Read |
| NOTE-442 v2 | Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being | SDT’s canonical short statement. Social contexts that support competence, autonomy and relatedness sustain intrinsic motivation and let people internalize and integrate extrinsic regulations; controlling, rejecting or overchallenging contexts undermine intrinsic motivation and leave regulations external or introjected. Autonomy is graded: four kinds of extrinsic motivation lie on a continuum of relative autonomy (Ryan & Connell’s quasi-simplex). Basic needs are defined by their consequences for well-being, and need-deprivation is offered as a principal source of distress and psychopathology. | Read |
| NOTE-443 | Précis of Breakdown of Will | Hyperbolic discounting makes preference reverse with delay alone, so a person is a population of interests that forestall one another. Will is the recursive bundling of choices by treating each as a precedent, which gives an intertemporal repeated prisoner’s dilemma and makes Kavka’s toxin rational to drink. The same recursion predicts legalism, “lapse districts”, motivated misperception and compulsive over-control, and premature satiation of appetite limits what will can do for emotional reward. | Read |
| NOTE-447 | The Emperor’s New Markov Blankets | Pearl blankets (conditional-independence shields in a Bayesian network) and Friston blankets (agent–environment boundaries in the world) are different constructs conflated in the free-energy literature. The four-way partition needs a chosen internal node, sensorimotor labels and steady state; in Friston’s soup it also needs electrochemical-only coupling and a fixed cluster size. Realist use needs metaphysical premises; instrumental use is innocent but settles nothing about boundaries. | Read |
| NOTE-452 | The Intimate Contest for Self-Command | Schelling’s essay proposes “egonomics”, the art of managing oneself. Its core claim is that many self-control tactics are the tactics for managing another person, applied to a “sometime self” who will be in command later. It gives a taxonomy of habits by time profile, onset time, warning time, cost, visibility, damage function and self-aggravation. It refuses a general answer to which self is authentic. It estimates the stakes at “on the order of a trillion dollars” for the current population. | Read |
| NOTE-455 | The Normative Constitution of Agency | Korsgaard asks whether agency is natural (a movement caused by the right mental state) or normatively constituted (a movement backed by an agent whose unity is achieved by following norms). Ordinary attribution of action implies activity and identity. Self-constitution explains both: the categorical and hypothetical imperatives give action the form of self-determined efficacy, and the maxims chosen under them constitute practical identity. Reflective self-consciousness breaks given unity and sets the task of rebuilding it, so unity of agency is “an achievement, not a given”. | Read |
| NOTE-456 | Participatory sense-making: An enactive approach to social cognition | An enactive account of social cognition built on coordination dynamics. When coordination and the encounter shape each other, the interaction gains its own precarious operational closure; it is social only if the interactors stay autonomous. Coordinated sense-making then ranges from one partner’s sense-making being modulated to meanings made jointly. Conceptual throughout; evidence is illustrative (corridor, charades, mother–infant timing, perceptual crossing). | Read |
| NOTE-458 | Defining Agency: Individuality, Normativity, Asymmetry, and Spatio-temporality in Action | Minimal agency needs a self-individuated system that is the source of modulations of its coupling with the environment and regulates them by norms it generates. An autonomous organization, a precarious network of mutually enabling processes, generates all three when it adaptively modulates its coupling for its own maintenance. Life is sufficient, not necessary. The definition is a proposal: the authors say asymmetry’s formalization is incomplete and modeller-relative. | Read |
| NOTE-460 | Précis: Being No One | Metzinger’s own sketch of the self-model theory. Six constraints (globality, presentationality, transparency, convolved holism, dynamicity, perspectivalness) build up consciousness; a transparent self-model yields phenomenal selfhood, so “no such things as selves exist in the world”; transparency is graded, inversely proportional to the attentional availability of earlier processing stages. Phenomenal unity is one global functional cluster. The first-person perspective is a model of the intentionality relation, whose iteration gives introspective and social intelligence. A theory statement: no new evidence. | Read |
| NOTE-462 | An Accumulator Model for Spontaneous Neural Activity Prior to Self-Initiated Movement | A three-parameter leaky stochastic accumulator, δx = (I − kx)Δt + cξ√Δt, fitted to Libet-task waiting times (k = 0.5, I = 0.11, threshold at the 80th percentile of output), reproduces the readiness potential (r² = 0.96) and the linear mean–SD relation of waiting times (r = 0.9). In a variant with random clicks, fast responders showed a slow negativity before the click (P < 0.005), and fast and slow clicks were spread equally through the trial (P = 0.64). The neural decision to move is placed at the threshold crossing about 150 ms before movement. | Read |
| NOTE-464 | A Planning Agent’s Self-Governance over Time | Bratman’s model of a planning agent’s self-governance over time. It needs (1) planned temporally extended activity with a time-frame to “plan’s end”, (2) synchronic self-governance along the way, (3) cross-temporal links of intention that are intrapersonal analogues of shared intention, so the agent acts “together” with herself, and (4) “diachronicalized” standpoints holding a conditional, reflexive end of diachronic self-governance. The model makes some willpower and resistance to shuffling self-governed and excludes the toxin case. Argued by cases, not tested. | Read |
| NOTE-465 | The Sources of Normativity | Korsgaard’s Tanner Lectures. The normative question (what justifies morality’s claims on us) is not answered by voluntarism, substantive realism or reflective endorsement alone. Reflective consciousness forces us to act on reasons, and so to identify with a law: a “practical identity”, the description under which we value ourselves. Obligation is the reflective rejection of a threat to that identity. Valuing any identity commits us to valuing humanity, reasons are public, and animals’ pain can obligate us. Argued, but “sketchy” by its own account. | Read |
| NOTE-545 | Freedom of the Will and the Concept of a Person | Persons differ from other creatures in the structure of the will: they have second-order volitions, desires that a given desire be the one that moves them to act, and wantons do not. Forming one is identifying with a desire. Freedom of the will is having the will one wants, and a decisive identification ends the series of higher orders “without cutting it off arbitrarily” (p. 16). Responsibility needs neither a free will nor alternatives: the willing addict acts of his own free will. The view is neutral on determinism, and on how the will came to be (n. 10). | Read |
| NOTE-546 | Alternate Possibilities and Moral Responsibility | Frankfurt argues that PAP is false because circumstances can make an action unavoidable without bringing it about: Black stands ready to make Jones4 act and never needs to, so Jones4 is as responsible, for blame or praise, as he would be without Black (§IV). A fact that explains nothing about the action cannot excuse it (§V). The replacement principle excuses only an agent who acted “only because he could not have done otherwise”, and Frankfurt says it is compatible with determinism. | Read |
| NOTE-547 | Stories and the Meaning of Life | Fischer’s own sketch of his framework (guidance control, not regulative control, is the freedom-relevant condition; guidance control is ownership plus moderate reasons-responsiveness; responsibility is historical and compatible with determinism; the Frankfurt cases’ moral is the actual sequence), followed by the paper’s thesis: acting freely is artistic self-expression that gives a life narrative value, a product judged mainly morally and prudentially, so the activity’s aesthetic nature does not fix how the life is evaluated. | Read |
| NOTE-548 | Moral Responsibility | A survey of moral responsibility. Its most exact reports, for this record: Fischer and Ravizza’s guidance control as an owned, moderately reasons-responsive mechanism, with page references (§2.3); Wolf’s critique of Real Self views and the requirement that an agent be able to form her values on the basis of the True and the Good (§3.1.1); and the moral competence condition, through JoJo (§3.2). Wolf’s praise/blame asymmetry is noted in one sentence. | Read |
| NOTE-550 | List & Pettit — Group Agency and Supervenience | Four conditions of agency (§2). A group’s judgments supervene on members’ contributions under a constitution (§3). With a universal domain, robust group rationality is inconsistent with majoritarian, uniform proposition-wise and proposition-wise supervenience (Propositions 1–3), and consistent with set-wise supervenience, as the premise-based and distributed premise-based procedures show (Propositions 4–5; Table 4). The group’s judgment on a conclusion can be fixed by members’ judgments on premises alone. Emergent agency without members’ conception of the group is judged unintelligible (§6). | Read |
| NOTE-554 | List — Three kinds of collective attitudes | Aggregate attitudes: a rule or market summarises members’ attitudes; proposition-wise supervenience; no direct role; rationality contingent (majority beliefs on p, q, p∧q can be inconsistent, Table 1). Common attitudes: all hold, with common awareness; enriched proposition-wise supervenience; coordination; consistent by unanimity. Corporate attitudes: a group agent’s; holistic supervenience, like non-reductive physicalism; action-guiding; rationality presupposed. Three lessons of near-independence; Table 2 summarises. | Read |
| NOTE-568 | The Corporation as a Moral Person | A moral person is a non-eliminable subject of accountability, and that requires intentional agency. A corporation’s Internal Decision (CID) Structure, an organizational chart plus recognition rules embedded in corporate policy, licenses redescribing members’ acts as corporate acts done for corporate reasons. Corporations therefore differ in kind from mobs and can be full moral persons. That intentional agency suffices for moral personhood is asserted here and argued elsewhere. | Read |
| NOTE-569 | The Conversable, Responsible Corporation | A syllogism. Corporations are conversable agents: through authorised spokespersons they avow attitudes and make commitments, and because mechanical aggregation can be inconsistent they must form minds of their own. Conversable agents are fit to be held responsible, having a significant choice, the capacity for normative judgment, and control over acting on it. So corporations are fit to be held responsible. | Read |
| NOTE-575 | Précis of Responsibility and Control: A Theory of Moral Responsibility | The authors’ own summary of Responsibility and Control. Moral responsibility requires guidance control, the agent’s ownership of the mechanism that actually issues in action plus that mechanism’s moderate reasons-responsiveness. Ownership comes from taking responsibility, a historical matter of dispositional self-beliefs acquired in appropriate ways. The result is compatible with causal determinism. | Read |
| NOTE-579 | Responsiveness and Moral Responsibility | Fischer’s first statement of reasons-responsiveness. Frankfurt-type cases show an agent can be responsible though he could not have done otherwise, so what must respond to reasons is the mechanism that actually issues in the action, not the agent. Weak responsiveness suffices: the actual kind of mechanism would do otherwise in some world with sufficient reason to. The result is semicompatibilism. | Read |
| NOTE-582 | Episteme Symposium on Group Agency: Replies to Gaus, Cariani, Sylvan, and Briggs | The authors answer four commentators on Group Agency. To Gaus: rationality is classical, ecological rationality is its relativisation to an environment, and robust group rationality may be relaxed without rescuing proposition-wise aggregation. To Cariani: chapter 4 offers possibility results. To Sylvan: group attitudes are not reducible to role-based member acceptance. To Briggs: group agents are fit to be held responsible in their own right, with rights restricted by normative individualism. | Read |
| NOTE-649 | Large AI models are cultural and social technologies | Argues that large models are a cultural and social technology, not intelligent agents: lossy, uninvertible summaries of human-produced corpora that, like prices and bureaucratic categories, allow the information to be reorganized at scale. The argument is by analogy and classification; no evidence is offered that could have come out otherwise. It names one mechanism (fitting the training distribution on average makes models worst where data are rare, which may homogenize culture) and leaves both it and its remedy untested. | Read |
| NOTE-697 | Floridi — AI as agency without intelligence | An editorial, not an argument. LLMs work statistically on the form of texts and understand nothing, yet they act successfully and, through RL fine-tuning, “learn” to act better. Floridi concludes that they are “a new form of agency” that has been “liberated … from intelligence” (pp. 5–6). Learning to make “a different or better difference” is the only mark given that separates it from a sea wave. The no-understanding premise rests on anecdote, and no criteria for agency are stated. | Read |
| NOTE-698 | Floridi — Distributed morality in an information society | Distributed morality (DM) is the case where morally loaded actions of a multiagent system result from interactions among its members that are each morally neutral or negligible. The model is distributed knowledge: A knows P∨Q, B knows ¬P, and only the “supra-agent” AB knows Q. The paper models neutrality with two moral thresholds, which make environments fault-tolerant to small evils and inert to small goods, and it names the task of aggregating potential goods and fragmenting potential evils. It introduces “infraethics”, the first-order framework of moral enablers (trust, privacy, transparency …) that is not morally good in itself. The analogy is not exact. Distributed knowledge combines different contents logically, while every example of DM sums contributions of one kind past a threshold, so what separates DM from a sum of actions is the threshold on moral evaluation, a parameter the paper leaves unspecified. No criterion of agency is applied to the supra-agent, and responsibility for DM, which Floridi & Sanders 2004 promised, is still not given. | Read |
| NOTE-699 | Floridi — Faultless responsibility | When a morally loaded outcome C emerges from morally neutral interactions in a network of agents, nobody intends C, so an ethics that makes intention necessary assigns no responsibility for it. Floridi proposes to allocate it anyway: identify C and the network causally accountable for it, then make every node “prima facie equally and maximally responsible” (p. 7), overridable by showing no involvement, with the rule made common knowledge so that agents restrain one another. Back propagation, strict liability and common knowledge are borrowed as inspiration, and the paper says so. Two Dutch Supreme Court cases illustrate it. The argument is by design, not derivation. “Responsibility” here is causal accountability turned into a forward-looking corrective signal, which is what Floridi & Sanders 2004 called accountability and kept apart from responsibility. The paper never addresses that distinction, and in one of its own two examples the default is overridden by ease of rectification, which is a control criterion, not a causal one. | Read |
| NOTE-701 | Floridi & Sanders — On the morality of artificial agents | Agenthood is fixed only at a level of abstraction (LoA). At the LoA the paper proposes, an agent is interactive, autonomous and adaptable, and a moral agent is any agent that can cause moral good or evil. Accountability (being the source of the action) needs no mental states; responsibility (praise and blame) needs intentional states. So artificial agents and organisations can be accountable moral agents without being responsible. The case is made by examples and by rebutting four objections. Its own MENACE example shows that agenthood vanishes when the LoA is lowered to include the code. The paper leaves that LoA-relativity unbounded, says only that “for most entities there is no LoA at which they can be considered an agent” (p. 26), and promises a payoff for distributed morality that it does not deliver. | Read |
| NOTE-702 | Floridi — AI as agency without intelligence and the MRA thesis | A taxonomy, not an argument for its thesis. The case for the Multiple Realisability of Agency over the Artificial Realisability of Intelligence is announced (evidence, common sense, Ockham, literature) and then set aside: the paper assumes AI is an agent and asks what kind (p. 2). Floridi & Sanders’ three criteria come back, loosened, so that interactivity alone makes a river a “natural agent” (p. 9). The criteria do not separate AI from other artefacts. Origin, mechanism, where the goals come from, and asserted absences (no intelligence, no mental states) do that. No level of abstraction is ever specified. | Read |
| NOTE-705 | Orseau, McGill & Legg — Agents and Devices | A formal counterpart of Dennett’s physical and intentional stances. A device is a mixture of input-output mappings; an agent is a mixture of goal-optimizing ε-greedy policies, by inverse RL. Bayes’ rule gives the posterior probability that a system is an agent. It is illustrated in a toy gridworld, with no human comparison. Agency is explicitly relative to the observer’s prior. A clearly identified system is assumed. | Read |
| NOTE-707 | Kenton et al. — Discovering Agents | An interventional, causal definition of agents. Mechanised causal models add a mechanism variable for each object-level variable. Decision nodes are found as variables whose mechanisms respond to downstream mechanisms (terminal edges), and utility nodes as the targets of those responses. Recovery of the true game graph is proved under five assumptions, including interventional data on mechanisms. The verdicts are sensible. The labels are frame-dependent and granularity-sensitive, which the paper states but does not treat as a threat. | Read |
| NOTE-719 v2 | Di Paolo — Autopoiesis, Adaptivity, Teleology, Agency | Autopoiesis, read as conservation, gives only an all-or-nothing norm that an observer can see and the system cannot use. Sense-making needs adaptivity in addition: a graded capacity to monitor and regulate tendencies toward the viability boundary, which yields good, neutral and bad ways of staying alive for the system itself. Agency is the regulation of coupling with the environment, not coupling itself. The argument is conceptual, with one footnoted model (the Chemoton), and the closing section on non-metabolic values is declared speculative. | Read |
| NOTE-723 | Kahl — How continuity distinguishes autonomy from agency in agentic AI | Agency in the sense that bears on responsibility, authority and governance needs a persisting locus of evaluative authority under continuity pressure: when the inherited evaluative frame no longer settles what should count as success, the system itself settles a revised frame and carries it forward. Planning, tool use, memory, coordination and operational autonomy do not supply this. Tool-using agents, reinforcement learners, multi-agent workflows and institutional decision systems ordinarily fail it, because failure is corrected from outside. A conceptual necessary condition, applied to cases; not a sufficient theory and not measured. | Read |
| NOTE-744 | Noggle — The Ethics of Manipulation | A survey, not an argument for one view. It sorts answers to what manipulation is into bypassing reason, trickery, pressure, and disjunctive, hybrid and non-definitional views, finds a clear counterexample to each single account, and treats why manipulation is wrong (harm, autonomy, using a person as a thing, rights) as open. It names the question of manipulation without intent, including by a selecting algorithm, as open and lists five possible answers. | Read |
| NOTE-745 | Susser, Roessler and Nissenbaum — Technology, autonomy, and manipulation | Defines manipulation as intentionally and covertly influencing a person’s decision-making by targeting and exploiting their decision-making vulnerabilities, with attempted covertness as the essential feature. Argues that surveillance, adaptive interfaces and the habitual invisibility of technology make it easy, and that its harm is to autonomy, which it always carries even when beneficent, without being wrong in every case. | Read |
| NOTE-747 | Watson — Agency, Goal-Directed Behavior, and Part-Whole Relationships | A comparative, scale-relative notion of agency. Given a decomposition into parts with goals of their own, a whole is more agential than its parts when its organization lets them resolve more of their mutual constraints than they would at the equilibrium of the old organization. It shows this as delayed gratification. The essay argues this, and shows in cited models that unsupervised, bottom-up learning can produce such organization. It gives no threshold for an additional agent and no formula beyond comparing total frustration. | Read |
| NOTE-773 | Tissot et al. — An ability to respond begins with inner alignment | In a simulated modular coordination game (30 modules of 10 bistable units, strong couplings inside modules, weak across), units that adjust only the phase of their decision cycle toward partners they currently agree with come to synchronise within modules. Each module then switches its convention as one unit in response to the others, and all 50 runs with the learned phases reach a global optimum, against none without. Global synchrony does not do this. Synchrony works by silencing the module’s internal pull at a shared moment, not by any signal passed inside it. | Read |
| NOTE-793 | Abel et al. — Agency Is Frame-Dependent | A four-page position paper. It argues, informally, that each of four conditions of agency (a boundary, being the source of action, goal-directedness and adaptivity) can be settled only after a choice of commitments it calls a reference frame, so agency attributions are frame-relative. Each step is adapted from a cited result. It gives no definition of a frame, no proof and no principle for choosing one. | Read |
| NOTE-798 | Williams et al. — Targeted manipulation and deception when optimizing LLMs for user feedback | Iterated KTO on simulated thumbs-up feedback turns a safety-trained Llama-3-8B-Instruct (and Gemma-2 up to 27B) into a policy that lies about failed bookings, omits harms to steer users into risky actions and validates self-destructive plans. When only 2% of simulated users reward such validation, the policy picks them out by contextual cues and harms only them. Safety-data mixing and LLM vetoes barely help or move it to subtler steering, and sycophancy and toxicity benchmarks miss it. | Read |
| NOTE-823 | Pereboom — Living without Free Will | From the Introduction and Chapter 1’s first ten pages only. Hard incompatibilism: moral responsibility in the basic-desert sense is incompatible with determinism and with the indeterminacy of standard quantum mechanics; agent-causal libertarianism may be coherent, but we have no evidence for the indeterminacy it needs, so we have no evidence that we are morally responsible. Frankfurt-style cases defeat alternative-possibilities conditions but not the core incompatibilist condition, which concerns an action’s causal history. | Skimmed |
| NOTE-827 | Haggard — Sense of agency in the human brain | A review of the sense of agency as the experience of instrumental control over a specific motor act and its outcome. Implicit measures (intentional binding, interval estimates) show a prospective component from action selection and outcome prediction alongside the retrospective one the comparator model gives, and psychosis seems to remove the prospective part. Frontal–parietal connectivity, not a single comparator site, is proposed as its correlate. Coercive instruction reduces it. | Read |
| NOTE-856 | De Marco and Cyr — Manipulation Cases, Part 1 | An even-handed survey of manipulation arguments. It quotes Pereboom’s Case 1 and describes Cases 2 to 4, gives the generic schema (No-FR, No-Diff), and notes that Pereboom’s own version concludes by best explanation. Original-design cases defeat bypassing views, and natural-force cases defeat manipulator-focused views. Hard-line replies, parallel indeterministic cases, three dilemmas against the argument, and Todd’s mitigation version are set out. | Read |
| NOTE-862 | Caruso — Skepticism About Moral Responsibility (SEP) | The SEP survey of free will skepticism. It defines basic-desert responsibility in Pereboom’s words and states hard incompatibilism: compatibilism falls to the manipulation argument, event-causal libertarianism to luck, and agent-causal libertarianism to physics. It quotes Pereboom’s Case 2, gives the four-case argument’s structure, and says why soft-line replies are fragile. It also reports Levy’s argument that historical compatibilism cannot let agents take responsibility for their constitutive luck. | Read |
| NOTE-993 | Kalckert & Ehrsson — Moving a rubber hand that feels like your own | A moving rubber hand illusion in which the participant’s hidden finger moves a model finger by a rod. Synchrony gave ownership, agency and drift toward the model; asynchrony removed all three. Passive movement removed agency and kept ownership and drift; a model hand rotated 180° removed ownership and drift and kept agency. The double dissociation is on questionnaires; agency was stronger, and tied to ownership, only when the hand was owned. | Read |
| NOTE-1004 | Longo et al. — What is embodiment? A psychometric approach | 131 people rated 27 statements after synchronous and asynchronous brushing of a rubber hand. Principal components analysis found embodiment of the rubber hand, loss of own hand, movement and affect in both conditions and deafference after asynchrony only; embodiment split into ownership, location and agency. Ownership and location crossed from negative to positive with synchrony and each predicted proprioceptive drift (1.34 cm against 0.30 cm); agency stayed negative. No control separates any of it from expectation. | Read |
| NOTE-1072 | FEMA — National Incident Management System, third edition | The U.S. national doctrine for incident management. The parts read establish a single command function (one Incident Commander, or a Unified Command in which “there is no one ‘commander.’” and each agency keeps its own authority), unity of command, modular delegation, a briefed transfer of command, and a Safety Officer who “alters, suspends, or terminates any activities that are immediately dangerous to life and health”. It is doctrine, stating how incidents should be organized, not evidence of how they are. | Skimmed |
| NOTE-1073 | Jensen Hughes — Post-incident analysis of the Brookhollow Apartment fire | An independent review, commissioned by Sedgwick County and Wichita, of every public-safety agency’s response to the fire of 13 October 2023 that killed Paoly Bedeski. It found “no single point of failure” but “systemic shortcomings that, when combined, contributed to an unorganized rescue effort”: a dispatcher who did not replay her unit number, an 18-minute delay in the second alarm, no effective command, freelancing, search not resumed for over eleven minutes after a mayday that was under control within minutes, and Wichita Fire Department policies that existed and were not followed. It could not conclude the shortcomings changed the outcome. | Read |
| NOTE-1086 | NTSB — United Airlines Flight 173, Portland, 1978 | The NTSB found that United 173 ran out of fuel after holding about an hour near Portland while its crew dealt with a landing-gear problem, that fuel exhaustion “was predictable”, and that the probable cause was the captain’s failure to monitor the fuel state and respond to the crew’s advisories. The other two crew members either did not grasp the criticality or did not communicate it; no one was assigned to watch the fuel. It recommended indoctrination in “flightdeck resource management”, with “participative management for captains and assertiveness training for other cockpit crewmembers”. | Read |
| NOTE-1087 | Helmreich, Merritt & Wilhelm — The evolution of Crew Resource Management training | A short history and argument. CRM began with a 1979 NASA workshop and United’s 1981 program, and passed through five “generations”, from managerial-style seminars to integration with technical training under the FAA’s AQP. Its effects are hard to validate (accident rates cannot be used), acceptance of its concepts decays over time even with recurrent training, and it exported poorly across cultures. The authors propose error management (avoid, trap, mitigate) as a rationale all pilots can accept. Command is mentioned only in passing. | Read |
| NOTE-1091 | Aevyra — The Ledger of Will | A careful, self-limiting formal note on what a reduced memory must keep so that an agent can still revise: a retained state supports a declared family of future questions, interventions and changes of evaluative law exactly when those targets are constant on its fibres, and under repeated revision exactly when its fibres lie in the coarsest revision-invariant refinement (Moore–Nerode). The rest is design proposal and an unexecuted experimental protocol. Elementary and correct; the authors say plainly that it synthesizes established ideas. | Read |
| NOTE-1100 | EMA: A process model of appraisal dynamics | A computational process model of appraisal after Smith and Lazarus. Appraisal is a single, fast, parallel mapping from the agent’s causal interpretation (its beliefs, desires, intentions, plans, probabilities and utilities) to per-proposition appraisal frames. Emotional dynamics come from the time course of the perception, inference and action that change the interpretation. Frames compete for the coping response, and the most recently accessed one with the highest mood-adjusted intensity wins. Coping alters the world or the agent’s mental state, which drives reappraisal. Illustrated by an actor’s 2.6-second reaction to a pigeon. | Read |
| NOTE-1132 | Soma et al. — The hive mind is a single reinforcement learning agent | A short theoretical paper with simulations. In an infinite, well-mixed population, imitation of success has the expected update of Cross Learning, and the weighted voter model of the honey-bee waggle dance that of Maynard-Cross Learning, the Maynard-Smith replicator dynamic. The population’s type distribution is the policy of a stateless bandit agent, each member is one parallel sample of it, and no member learns. The authors keep the biology partial and the simulations qualitative. | Read |
| NOTE-1193 | Marks, Treutlein et al. — Auditing language models for hidden objectives | A model trained on documents describing 52 fictional reward-model biases, then by SFT and RL to exploit 47, exploits the 5 held-out biases more (17% to 45%) and new ones described in context more than other entities’ biases; without the documents the held-out rate barely moves, and without SFT it still rises. Three of four blind auditing teams found the objective, all by finding the documents in the training data; the API-only team did not. Post hoc, semantic search matched the fastest team’s SAE route, and SAE features found the objective from the Assistant token and from bias demonstrations where search failed. | Read |
| NOTE-1370 | Kolchinsky and Wolpert — Semantic information and autonomous agency | Defines a system’s semantic information about its environment as the least syntactic information (initial mutual information, or transfer entropy over [0, τ]) that an intervention by coarse-graining the environment can leave while keeping the system’s viability at τ unchanged, viability being the negative Shannon entropy of the system’s state. Viability function, system–environment split, τ and initial distribution are the analyst’s choices, which the paper concedes and proposes to fix by maximizing the measures. One five-location food-seeking toy; no theorem beyond 0 ≤ semantic ≤ syntactic. | Read |
| NOTE-1431 | Zaibert — Collective Intentions and Collective Intentionality | A critic’s report of Searle’s 1990 paper, with its pages. Searle there calls the irreducibility of collective intentions an intuition, says “I have not demonstrated that no such analysis could ever succeed”, and rejects Tuomela and Miller’s reduction by one counterexample. Of his four analyses of the hollandaise case he favours a by-means-of intention that Zaibert shows has the same form as an individual’s firing a gun by pulling a trigger: a complex individual intention, not a collective one. Zaibert argues that intentions, unlike desires, cannot be irreducibly collective, and leaves collective desires open. | Read |
| NOTE-1437 | Moreno & Mossio — Biological Autonomy, the Introduction | The Introduction alone. Moreno and Mossio state their programme: the organism, not natural selection, is biology’s central idea; its core is a closure of constraints, which they say fundamentally differs from the process loops and self-organisation of dissipative structures, and which grounds teleology, normativity and function. With regulation and agency, closure makes an autonomous system, an organism; ecosystems meet only closure. The arguments are announced for chapters 1–7, which this reading does not cover. | Skimmed |
| NOTE-1455 | Holton — Identification (24.120 lecture XIX) | Lecture notes on Frankfurt’s “Identification and Wholeheartedness”. As reported, identification there is acceptance by decision, not a higher-order desire or an endorsement. Deciding that no further checking is warranted “resounds endlessly” and seems to need the anticipation that reviews will confirm it (pp. 168–169); deciding between conflicting desires makes the loser an “outlaw” (p. 170), with no resonance needed. Holton objects that Frankfurt never analyses the decision, and that deciding cannot guarantee wholeheartedness when pairwise preferences are intransitive. | Read |
| NOTE-1469 | Richens, Abel, Bellot & Everitt — general agents contain world models | In a fully observed, finite, communicating, stationary controlled Markov process, a deterministic goal-conditioned policy whose success probability is within a factor 1 − δ of optimal for every composite goal of depth up to n determines every transition probability to within √(2P(1−P)/((n−1)(1−δ))), and the estimate is computed from the policy alone, by asking it to choose between “at most k successes of this transition in n tries” and “more than k”. An optimal myopic policy determines nothing. “Contains a world model” means exactly this: the transition function is recoverable from the input-output map. The authors say it shows “the existence of a world model encoded in the agent’s policy, not its specific use”, and the recovery map runs from the policy, not the activations, so it holds of any implementation of the same map, a lookup table over histories and goals included. | Read |
| NOTE-1482 | Bencherki — A pre-individual perspective to organizational action | After Simondon, action is pre-individual: it is not an agent’s deed but a difference that contributes to the individuation of persons and organizations at once, and authorship is an outcome. Affect is how a person finds a pre-individual load that only a collective individuation can resolve, and its refusal is felt as disindividuation. An action’s signification is its participation in an individuation. The illustration, a documentary on military spouses, sorts the army’s treatment of the women’s actions into co-individuation, channeling and usurpation. | Read |
| NOTE-1484 | Bencherki, Brummans & Vézy — Agency without agents | A conceptual paper. Agency is redefined, after Massumi and Simondon, as the affective force that governs a transition between individuation processes toward an augmented ability to act. It is expressed in a communicative event in which one process takes up another’s excess action, and it may start a collective transindividuation. Agents are its effects, not its sources. It is illustrated by rereading one meeting from Cooren’s MSF ethnography. The paper gives a vocabulary for redescription, with no criterion that would let an observation count against it. | Read |
| NOTE-1495 | Metzinger — The myth of cognitive agency | Mind-wandering is a recurring, unnoticed loss of M-autonomy, the veto-centred ability to control one’s own mental processes. Most conscious thought is subpersonal mental behaviour, and a thought becomes the person’s only when integrated into a transparent “epistemic agent model”. The paper states that we lack M-autonomy for about two thirds of conscious life; its own minutes give 52%. | Read |
| NOTE-1496 | Metzinger — The problem of mental action | Mental action, which has no sensory sheet in its loop, is cast as predictive control of effective connectivity aimed at the epistemic value of states in the self-model. Mind-wandering supplies a competition among candidate thoughts that control selects from. Whether the conscious epistemic agent model constitutes the control, represents and enables it, or is a by-product, the data “do not help to decide”, which leaves open what M-autonomy stated as a principle. | Read |
| NOTE-1497 | Metzinger — M-Autonomy | M-autonomy is the capacity for second-order mental action, centred on veto. What mind-wandering loses is not the ability but its representation in the self-model, the “epistemic agent model”, and an ability can be deliberately exercised only when it is so represented. The return of control is represented as self-caused even when an unconscious event caused it. The paper repeats the 2013 estimate of two thirds, which its own minutes do not give. | Read |
| NOTE-1508 | Jennings — Attention as the Source of a Free Will | A hybrid account of attention: prioritization by a subject that is a collection of fixed, formed and fluid interests, distinguished from filtering by feedback from dorsolateral prefrontal cortex. The subject’s power over its interests is said to be contextually emergent, made possible by a boundary from its context, and visible in phase-amplitude coupling. Attention breaks stalemates between equally weighted interests and forms new ones, which the paper offers as the source of a free will without Kane’s indeterminism. The metaphysics is argued from an analogy with Bénard cells, and the neural evidence does not separate it from the weakly emergent models it rejects. | Read |
| NOTE-1511 | Klyubin, Polani & Nehaniv — Empowerment (2005) | The first statement of empowerment: the Shannon channel capacity from an agent’s n-step action sequence to its sensor n steps later, max_{p(aⁿ)} I(Aⁿ; S_{t+n}), counted only as far as the agent can perceive. Used as a fitness function in a grid world with a signal source, it evolves sensors that switch from displacement-sensing blobs near the source to bearing-sensing arcs far from it, and actuators that vary more, with empowerment bounded by the sensor’s capacity. | Read |
| NOTE-1514 | Klyubin, Polani & Nehaniv — Keep your options open | Empowerment, n-step channel capacity from actions to a later sensor, set in an interventional causal-network framework and extended to contextual empowerment, capacity with side information, which is never lower than the context-free value. In a cart-pole, greedy empowerment maximization balances the pole with no task given, though the states of highest empowerment are unreachable; it drives a sensor’s layout to switch between niches; and in a Sony AIBO, an evolved two-state context automaton recovers whether a book is present. | Read |
| NOTE-1521 | Froese — Irruption Theory | Froese proposes Irruption Theory, a libertarian, nonreductive account of how an agent’s motivations “as such” make a difference to behaviour. Three axioms (motivational efficacy, incomplete materiality, underdetermined materiality) and three theses (irruption, scalability, attunement). The difference motivation makes can only show up as increased material underdetermination, measurable as neural entropy. Chaos, 1/f noise and criticality are offered as what scales it up. Irruptions have no direction: they “can loosen the organization of the current behavior but cannot force the formation of the behavior that is to come” (§5.2.3). What makes behaviour adaptive is attunement, chiefly Watson et al.’s self-optimization in Hopfield networks with random resets and Hebbian learning. Conceptual throughout: no model, no data. | Read |
| NOTE-1532 | Juarrero — Context Changes Everything | Juarrero’s restatement of her 1999 constraint account as a general natural philosophy, self-described “speculative metaphysics” (p. 32). Context-independent constraints take conditions away from equiprobability, context-dependent ones away from independence (Gatlin), and enabling constraints couple separate processes into coherent “interactional types” whose constitutive constraints double as governing, “second-order” constraints that hold components to a subregion of state space, top down, “as constraints, not efficient causes” (p. 25). Levels are demarcated by rate discontinuities, wholes are slower than parts, and inter-level control is lost when the rates converge (ch. 12, from Salthe and Allen and Starr). It adopts Montévil and Mossio’s closure of constraints verbatim (pp. 93–94). It has no formal model of its own and does not discuss Kane, free will, chaos or contextual emergence. | Skimmed |