Abstract
Over the past half century, psychology has repeatedly produced findings that achieve rapid popular and policy influence only to face serious empirical challenge. The marshmallow test, self-efficacy theory, the illusion of explanatory depth, and growth mindset interventions are prominent examples. This article argues that these are not isolated failures of rigour or communication but symptoms of a deeper structural and ontological problem. Drawing on Living Value Theory, the paper contends that psychology operates from a mistaken foundational ontology: it treats the individual as a container of discrete cognitive properties that can be isolated, measured, and modified. This framework systematically mischaracterises human life, which is better understood as ongoing, multimediated coordination across five irreducible domains—embodiment, being-with, dwelling, multimateriality, and multisymbolisation—operating at different levels of recursivity and through fundamentally different types of process (non-recursive, self-recursive, and interrecursive). Applying this diagnosis to four major paradigms reveals a consistent pattern: complex, distributed, interrecursive phenomena are reduced to individual cognitive variables; measurement instruments constitute rather than neutrally record the phenomena they claim to study; and findings that locate the sources of unequal outcomes inside individuals achieve disproportionate influence because they align with prevailing governance frameworks. Methodological reforms such as improved replication offer limited correction because the core problem is ontological, not merely procedural. The most influential claims of popular psychology are therefore not simply imprecise but systematically misleading. Wider awareness of this pattern, rather than any single methodological fix, is presented as the necessary first response.
Introduction
Something has gone seriously wrong with the relationship between psychology, popular culture, and public policy. Over the past half century, a succession of psychological findings have passed from laboratory to bestseller list to government initiative with remarkable speed and almost no scrutiny. The marshmallow test became a theory of human development. Self-efficacy became the master variable of health behaviour. The illusion of explanatory depth became a universal account of human irrationality. Growth mindset became an educational revolution. Each of these found, in Malcolm Gladwell or some equivalent transmission mechanism, a vehicle that detached the finding from its evidential basis and delivered it to millions of readers as established truth. Each subsequently faced serious empirical challenge, replication failure, or outright demolition. And in each case the popular cultural residue of the original claim has proved essentially immune to the corrective record.
This article argues that the problem is not one of poor communication, insufficient scientific literacy, or the inevitable simplification that attends popularisation. The problem is structural and ontological. Psychology as a discipline is constitutively prone to producing findings that are compelling, policy-relevant, individually actionable, and fundamentally misconceived about the nature of the phenomena they claim to describe. The misconception is not random. It follows a consistent and diagnosable pattern, because psychology begins from an ontology of the human that is wrong in specific, identifiable ways, and because that wrong ontology generates a methodology that cannot in principle capture what it claims to study.
The argument is developed in three movements. The first establishes the ontological diagnosis: what psychology gets wrong about what human life is, and why that wrongness is not correctable by better methods within the existing framework. The second applies that diagnosis to four paradigm cases: the marshmallow test, self-efficacy, the illusion of explanatory depth, and growth mindset. The third examines the machinery through which these misconceived findings achieve popular cultural and policy authority, and considers what follows from recognising that the problem has no particular solution, only a first necessity: much wider awareness of just how unreliable and biased these claims actually are.
I. What Psychology Gets Wrong About Human Life
Psychology's foundational ontological commitment is to the individual as the primary unit of analysis. Human behaviour, in this framework, is produced by properties of individuals: cognitive states, personality traits, neural architectures, motivational dispositions, beliefs, and capacities. These properties are located inside the individual, causally prior to the individual's interactions with the world, and in principle measurable by instruments that can isolate them from the contexts in which they are expressed. The individual is the container. Behaviour is the output. The experimental method that psychology has borrowed from the natural sciences is precisely designed to locate and measure the properties of this container by holding everything else constant.
This ontology is wrong about what human life is, and wrong in a way that is not correctable by methodological refinement. Living Value Theory proposes a different starting point. Human life is organised not through the properties of individual containers but through ongoing coordination between living beings and the world they inhabit. This coordination is what is always ultimately at stake. What makes life go well or badly, what constitutes value or its absence, what enables flourishing or produces harm, is always ultimately a matter of whether coordination is being maintained, threatened, or repaired. The individual human being is not the container of coordination. The individual is constituted, moment by moment, through coordinative processes that exceed any individual's cognitive states and that cannot be captured by instruments designed to isolate the individual from the world they are coordinating with.
This coordination is multimediated. Living Value Theory identifies five irreducible mediational domains through which coordination occurs, and through which alone vitality can be understood. Embodiment is the mediation through which coordination is carried in, by, and as a living body: its skills, habits, rhythms, fatigue, accumulated physical history, and ongoing metabolic engagement with the world. Being-with is the mediation through which coordination is structured by the presence, absence, anticipation, and memory of others: the recursive attunement through which others become part of the structure of one's own coordination, not as external influences on an internal state but as constitutive of what one is doing and capable of doing. Dwelling is the mediation through which coordination is structured by place, space, and the built and natural environment: the spatial and temporal infrastructure within which embodied and relational coordination becomes possible or impossible. Multimateriality is the mediation through which coordination is structured by the physical world of things, substances, food, infrastructure, and technology: the material conditions that either sustain or disrupt the other four mediational domains. Multisymbolisation is the mediation through which coordination is structured by symbols, language, categories, numbers, images, and institutional forms: the symbolic dimension of coordination itself, through which coordination becomes available to naming, transmission, and transformation.
These five mediations are always co-present in any living coordination. They are analytically distinguishable but practically inseparable. The child who is asked to wait for a second marshmallow is not a cognitive agent exercising or failing to exercise an internal capacity called self-control. The child is a body with a particular hunger state, in a relational situation with a researcher whose trustworthiness they are continuously assessing through the full interrecursive sensitivity of being-with, in a room whose institutional character and spatial organisation carry specific implications for what kinds of behaviour are appropriate, surrounded by objects whose significance depends on a material history of scarcity or abundance, and oriented by a symbolic framework that gives waiting its meaning or its pointlessness. All five mediations are simultaneously operative. The experimental isolation of the child in the test room does not remove the other four mediations. It constitutes a specific configuration of them, a configuration with its own properties, and then treats those properties as if they were properties of an individual cognitive capacity.
Psychology not only has no understanding of the irreducible multimediality of the mesocosm. It actively suppresses it. The experimental method requires holding conditions constant, which means treating the mediational configuration within which behaviour occurs as background noise to be controlled rather than as constitutive of the behaviour being studied. The more successfully the experiment controls for context, the more thoroughly it eliminates the actual substance of what it is studying. This is not a limitation that better experimental design can overcome. It is a consequence of beginning from an ontology that places the unit of analysis inside the individual rather than in the coordinative field that constitutes the individual.
The second dimension of psychology's ontological error concerns the levels of recursivity at which human coordination operates. Living Value Theory distinguishes five recursivity levels. At L1, coordination proceeds seamlessly, without reflection or deliberation: the level of the practised body, the established relationship, the inhabited place, the reliable material infrastructure, the shared symbolic form that functions as background condition rather than as explicit object of attention. This is the level at which the most valuable coordinations operate, and it is the level at which they are invisible, generating no data and producing no signal that an experimental instrument designed to detect articulated responses could register. At L2, coordination becomes unsettled: something feels wrong, a hesitation arises, attention is recruited. L2 is the level at which vitality stakes first become registerable as disruption. At L3, something is symbolically named, made available as a describable entity. At L4, abstraction and generalisation produce stable concepts, institutional categories, and rule-like structures. At L5, the symbolic systems themselves become objects of reflection.
Psychology's experimental methodology operates at L3 and above. It asks participants to produce verbal reports, numerical ratings, explicit judgments, and articulable responses. What it then interprets as measures of cognitive properties are in fact measures of L3 symbolic performance: of the capacity to produce appropriate verbal outputs in the specific social context of the experimental situation. The enormous range of human coordination that operates at L1 and L2, the seamless embodied expertise, the relational attunement, the spatial orientation, the material competence, the shared symbolic ease of communities whose coordination is working, is entirely invisible to an instrument that can only register what participants can say. Psychology has spent a century measuring L3 outputs and calling them measures of human capacity. The gap between what people can do and what they can say is not a cognitive bias. It is the structural relationship between L1 embodied competence and L3 articulatory capacity, and it is a constitutional feature of human expertise in every domain.
The third and most consequential dimension of psychology's ontological error concerns the failure to distinguish between fundamentally different types of recursive process. Living Value Theory distinguishes three recursivity domains that require entirely different analytical and methodological approaches. Non-recursive processes are those in which the elements of the process do not respond to expectations about them: the pharmacokinetics of a drug, the mechanics of a physical system, the chemical composition of a substance. These processes are indifferent to beliefs about them and can in principle be studied by experimental methods that isolate and control variables, because the phenomenon is not constituted by the relational situation in which it is observed. Self-recursive processes are those in which the organism's relationship to its own states is part of what shapes those states: pain experience, anxiety, hunger, and fatigue are all modulated by the organism's own monitoring of them. Experimental methods can illuminate self-recursive processes, though not without acknowledging that the observation changes the observed. Interrecursive processes are those in which multiple parties recursively affect each other's expectations, orientations, and behaviours, producing outcomes that cannot be specified independently of the recursive loop itself: the therapeutic relationship, the classroom dynamic, the family negotiation, the community response to crisis, the health behaviour embedded in a network of social accountability.
The overwhelming majority of the phenomena that psychology most influentially addresses are interrecursive. Health behaviour, child development, educational achievement, social cognition, and therapeutic outcome are all constitutively interrecursive: what one party does depends on what others do, which depends on what the first party does, in loops that cannot be captured by experiments that isolate individuals and hold relational conditions constant. Applying the experimental methodology of non-recursive natural science to interrecursive social phenomena does not produce imprecise or approximate results. It produces systematically misleading ones, because the methodology constitutes a specific interrecursive configuration and then misidentifies that configuration's properties as properties of an individual internal state. Psychology will continue to produce seemingly compelling but entirely misleading results for as long as it begins from an ontology that cannot distinguish between these three fundamentally different types of process, and applies a single methodology designed for the first to phenomena that belong to the third.
II. The Marshmallow Test: Self-control as Misidentified Interrecursion
Walter Mischel's delayed gratification experiments, conducted at Stanford's Bing Nursery School in the late 1960s and 1970s, placed children before a marshmallow and offered them a choice: eat it now, or wait for the researcher to return and receive two marshmallows. The finding that waiting time predicted later life outcomes became one of developmental psychology's most celebrated results. The popular version claimed that self-control is the master variable of human flourishing, measurable in early childhood and predictive across decades.
The empirical foundations were weaker than almost any element of their reception acknowledged. The original studies followed children attending a nursery school on the Stanford University campus, children of Stanford faculty, staff, and graduate students: one of the most socioeconomically privileged and homogeneous samples imaginable for a study claiming to reveal universal features of human development. The longitudinal correlations were modest, the follow-up methodology relied substantially on parental questionnaires, and the samples available for the most dramatic claims were small enough that the effect sizes could not support the interpretive weight placed on them.
When Tyler Watts, Greg Duncan, and Haonan Quan replicated the study in 2018 with approximately nine hundred children representative of the American population in terms of socioeconomic background, the predictive effect of waiting time shrank dramatically. Once family background and early cognitive ability were controlled for, the effect was negligible. The marshmallow test had been measuring socioeconomic stability, not individual self-control. Children from stable, well-resourced environments with reliable adults waited longer because they had good reason to trust that the second marshmallow would appear. Children from unstable environments with less reliable adult figures ate the marshmallow sooner because their experience had accurately taught them that adult promises are often not kept.
Celeste Kidd's earlier work had already shown this directly. When children were first exposed to a researcher who proved unreliable, failing to deliver on a prior promise, they ate the marshmallow significantly faster than children who had experienced a reliable researcher. The effect attributed to internal self-control capacity was substantially produced by the interrecursive ecology of trust and reliability within which the test was administered. This is the ontological diagnosis in miniature: the experiment was studying an interrecursive process, the relational configuration of trust between child and institution, and misidentifying it as a non-recursive individual property called self-control.
The multimediational analysis confirms the diagnosis at every level. In embodiment, the child's hunger state, their metabolic relationship to food, their physical history of scarcity or abundance, all determine what waiting means as a bodily experience. In being-with, the child's relational history with adults, their accumulated experience of whether promises from institutional figures are honoured, their capacity to read the specific researcher's trustworthiness, all determine whether waiting is a rational strategy or a predictable loss. In dwelling, the child's spatial environment, their familiarity or unfamiliarity with institutional test rooms, their comfort or discomfort with the specific configuration of the experimental situation, all shape the behaviour the test claims to be measuring as a context-free cognitive property. In multimateriality, the child's material history with food, their experience of whether desired things that are present remain available if they delay taking them, shapes their assessment of the risk of waiting. In multisymbolisation, the child's understanding of what the researcher's promise means in the symbolic framework of institutional authority, their capacity to interpret the social situation they are in, shapes their response to the offer. All five mediations are operating simultaneously. The experiment claims to be measuring one cognitive property. It is actually producing a reading of a specific mediational configuration and calling that reading a measure of self-control.
The transition from these fragile findings to sweeping policy influence is itself an interrecursive story. The marshmallow test achieved popular cultural authority because it told a story that governance frameworks needed: that the source of unequal outcomes lies inside children, in a cognitive property that can be trained. This is cheaper and more politically acceptable than addressing the structural conditions, the housing instability, the food insecurity, the degraded schools, the exhausted parents, that the test was actually measuring. The finding was not selected for policy uptake because of its evidential quality. It was selected because of its ideological fit with governance frameworks that prefer individual-level explanations of structural inequalities.
III. Self-efficacy: The Individual Container That Does Not Exist
Albert Bandura's 1977 proposal that an individual's belief in their capacity to execute the behaviours required to produce a specific outcome is the primary cognitive mediator between knowledge, motivation, and action became the most pervasively applied psychological construct in the history of health research. More than fifty thousand PubMed results. Nursing curricula, health promotion guidelines, app design logic, clinical training manuals. The appeal was immediate: self-efficacy was measurable via Likert-scale questionnaire, modifiable through four specified mechanisms, and applicable across any domain in which human beings did or failed to do things.
Every element of the foundational claim is wrong, and each wrongness maps onto one of the three ontological errors identified in Section I. First, self-efficacy is not located inside the individual. The capacity to manage a chronic illness is distributed across the patient's body and its current physiological state, their relationships with healthcare providers and family members, the material environment of their home and neighbourhood, the symbolic frameworks through which illness management is understood and valued, and the institutional configurations that make relevant knowledge, medication, and support available or unavailable. The self-efficacy questionnaire captures some signal from this entire mediational configuration and reduces it to a number. The field then treats the number as the cause of the configuration it imperfectly reflects. This is not a measurement problem. It is a container ontology applied to a phenomenon that is constitutively distributed across five mediations and irreducible to any of them individually.
Second, self-efficacy research systematically conflates non-recursive, self-recursive, and interrecursive processes, applying a single explanatory framework to phenomena that operate through entirely different mechanisms. In non-recursive domains, the oncology patient's confidence has no direct causal effect on the pharmacology of their chemotherapy. At most it may influence behaviours that are themselves interrecursive. In self-recursive domains, pain experience and anxiety, the organism's relationship to its own states does genuinely shape those states, and self-efficacy research captures something real here, though not through the mechanism Bandura described. In interrecursive domains, which constitute the overwhelming majority of health behaviour contexts, what the patient believes and declares affects what others believe and do, which affects what the patient subsequently does, which affects what others subsequently believe, in a loop that can only be understood at the level of the coordinative ecology rather than the individual mind. By failing to distinguish among these three types of process, the self-efficacy framework consistently overclaims causal power in non-recursive domains, misidentifies interrecursive effects as cognitive ones, and cannot explain why its interventions work when they do.
Third, the measurement instrument does not passively record a pre-existing internal state. It constitutes a new social fact. When a patient rates their confidence in their capacity to manage their condition on a numbered scale in the presence of a healthcare provider, they are performing a formal declaration of future orientation in a socially significant context. The South Asian philosophical tradition names this precisely: the sankalpa, a declaration that does not express a pre-existing state but creates a new one by reorganising the interrecursive ecology within which subsequent action will occur. The provider now holds an expectation. The patient now inhabits a world in which they have said something that can be confirmed or disconfirmed. Self-efficacy research has been studying ritual interrecursive commitment while believing it was measuring individual cognitive states. The evidence for this interpretation is embedded in findings the field has consistently failed to theorise: self-efficacy questionnaires predict better when completed immediately before performance, in public rather than private contexts, and in the presence of credible authorities. Each finding is inexplicable on the assumption that self-efficacy is a stable internal property. Each is entirely explicable on the account of the rating as an interrecursive social act whose constitutive force varies with the conditions of its production.
The political economy of self-efficacy theory is inseparable from its ontological architecture. A framework that locates the determinant of health behaviour in the individual's internal cognitive state systematically conceals structural determinants of health. The person without safe spaces to exercise, affordable healthy food, stable employment, or reliable healthcare access cannot adjust their self-efficacy upward and overcome these barriers. The framework does not merely fail to help such people. It misidentifies their problem as an internal cognitive deficit and thereby obstructs the identification of actual solutions. This is not a side effect. It is the direct consequence of applying a container ontology to phenomena that are constitutively distributed across mediational configurations that no individual-level instrument can capture.
IV. The Illusion of Explanatory Depth: Misreading the L1/L3 Gap
Rozenblit and Keil's 2002 paper introduced the illusion of explanatory depth: the finding that people believe they understand complex phenomena with greater precision and depth than they actually do. Participants rate their understanding, attempt a detailed explanation, and re-rate. The rating drops. This is interpreted as evidence of systematic overconfidence: a universal cognitive bias revealing that the human mind misestimates its own explanatory competence.
The experiment fails at a level prior to any question of interpretation. It never specifies what would count as a correct or sufficient explanation in any domain it tests, and the failure to specify this criterion is not a methodological oversight but an ontological impossibility. Consider the experiment's most tractable case: the toilet. The immediate answer is: the siphon mechanism, the ballcock valve, the cistern. But a toilet works with gravity. Gravity is not a detail that can be assumed away. It is the mechanism through which water descends, through which the siphon creates the pressure differential that produces the flush. A complete explanation therefore requires an account of gravity. But gravity defeated Newton's explanatory ambitions in a famous sense: he could describe its mathematics precisely while explicitly refusing to hypothesise about its underlying nature. Einstein reconceived it as spacetime curvature. Quantum gravity remains unresolved. The toilet, on minimal inspection, opens into one of the deepest unsolved problems in physics. The experimenters stop asking at the point where the account feels intuitively satisfying to them. That stopping point is arbitrary, unjustified, and variable. The experiment presents itself as measuring understanding against an objective standard. It is measuring participants' accounts against an unstated experimenter intuition.
This is the LVT diagnosis from the perspective of multimediational irreducibility: mesocosmic coordination across the five mediations is inexhaustible under symbolic rendering. Any verbal account of any mesocosmic process captures a slice while leaving the rest untouched. The toilet involves embodied practices of use, relational norms of sanitation, dwelling configurations of plumbing infrastructure, material supply chains of ceramic and plastic and metal, and symbolic frameworks of hygiene and privacy that together constitute a coordinative reality of which any verbal explanation is necessarily a radical reduction. The question of whether an explanation is correct or sufficient cannot be answered from within the symbolic register, because the phenomenon being explained exceeds any symbolic register in which it could be captured.
The consistent finding that felt competence outruns articulatory performance reflects not cognitive bias but the structural relationship between L1 embodied coordination and L3 symbolic articulation. The cyclist who cannot explain gear ratios rides perfectly. The surgeon whose hands know what her language cannot say performs with excellence. No one can articulate embodied coordination better than they can perform it. This is the correct relationship between non-symbolic and symbolic mediations. It is a constitutional feature of human expertise in every domain where skill is acquired through practice. The experiment pathologises this relationship. It takes the structural gap between L1 coordination and L3 articulation and calls it overconfidence. The correct conclusion is not that people think they know more than they do. It is that people can do more than they can say, and that this is as it should be.
The experiment compounds this error by treating ontologically heterogeneous domains as equivalent objects of understanding. A toilet is a non-recursive physical system with a determinate causal mechanism that operates independently of beliefs about it. Immigration policy is an interrecursive social process, historically sedimented, politically contested, and symbolically saturated across all five mediations simultaneously. There is no fact of the matter about how immigration works in the way there is a fact of the matter about how a siphon works, because immigration is constitutively altered by the declarations, policies, and symbolic contests that surround it. Policy declarations about immigration alter immigration. Toilet declarations do not alter toilets. Running both through the same experimental paradigm and producing a single effect size for the illusion of explanatory depth tells us nothing about cognition. It demonstrates that the experimenters had no understanding of the ontological difference between non-recursive physical systems and interrecursive social processes, and no methodology for distinguishing what understanding consists in across these different domains.
V. Growth Mindset: The Belief That Cannot Do What It Claims
Carol Dweck's growth mindset framework proposed that people hold one of two fundamental beliefs about intelligence: that it is fixed and innate, or that it can be developed through effort and learning. The finding generated an enormous applied literature and became the basis for educational interventions in thousands of schools across the world. Teaching children that their brains are like muscles that grow with exercise would, the framework claimed, improve their academic outcomes regardless of their starting conditions.
The empirical record is substantially more ambiguous than the popular reception suggested. A 2018 meta-analysis by Sisk and colleagues examining 273 effect sizes found that the relationship between growth mindset and academic achievement was weak, and that growth mindset interventions produced only small effects with significant heterogeneity across studies. A large pre-registered replication by Foliano and colleagues in the UK found no significant effect on GCSE attainment. The National Study of Learning Mindsets, a large-scale randomised controlled trial in the United States, found a small average effect concentrated in schools where the climate was already supportive, meaning the intervention worked where conditions were already good and made little difference where they were most challenging.
The ontological diagnosis follows the same structure as the previous cases. Growth mindset locates the causal variable for educational achievement inside the individual child, in the form of a belief about the nature of intelligence. This belief is presented as the lever that, once adjusted, will alter the trajectory of learning outcomes. The multimediational analysis immediately reveals the inadequacy of this account. In embodiment, the child's physical state, their nutrition, their sleep, their health, their sensory and motor development, all shape what learning is possible regardless of what they believe about intelligence. In being-with, the quality of the relationship between teacher and student, the relational dynamics of the classroom, the peer culture that either rewards or punishes academic effort, the family support or undermining of educational engagement, all determine what effort produces regardless of the child's mindset. In dwelling, the stability and quietness of the home environment, the availability of space for study, the fit between the child's spatial rhythms and the institutional rhythms of school, all shape learning outcomes independently of beliefs about intelligence. In multimateriality, the resources of the school, the availability of books and equipment, the class sizes and infrastructure of the institution, all determine what learning is possible in ways that no mindset intervention can substitute for. In multisymbolisation, the cultural frameworks through which academic achievement is understood and valued by the child's community, the symbolic stakes attached to school performance in their social world, all shape what the growth mindset belief means in practice and what it can achieve.
A child who believes their intelligence can grow but attends an under-resourced school with demoralised teachers, returns home to an unstable environment, and belongs to a peer culture that treats academic effort as betrayal of social identity, is not going to achieve substantially better outcomes because their mindset has been adjusted. The belief is embedded in a mediational configuration that the belief alone cannot alter. Growth mindset interventions work, to the extent they work at all, primarily in environments where the other four mediations are already supportive enough to make the belief actionable. This is precisely what the National Study found. The intervention is most effective where it is least needed and least effective where it is most needed, because the underlying mediational configurations are doing the real work and the belief is at best a modest supplement to conditions that are already sufficient.
The recursivity level analysis adds a further dimension. What children need in order to learn is not primarily a belief held at L3 or L4. It is L1 coordinative ease: the embodied, relational, spatial, material, and symbolic conditions under which learning proceeds without friction, without the cognitive overhead of managing disruption across multiple mediational domains simultaneously. A growth mindset intervention operates at L3 to L4: it attempts to install a new belief through explicit instruction and verbal persuasion. It cannot reach L1. It cannot substitute for the seamless coordinative conditions that make learning possible as a background activity rather than a struggle. The enormous resources directed toward growth mindset interventions in schools serving disadvantaged children would produce vastly more learning if directed toward the L1 mediational conditions, the material resources, the relational stability, the dwelling security, that the belief cannot create.
VI. The Ontological Machine and Its Products
These four cases are not independent failures of scientific rigour. They are outputs of a single ontological machine that operates consistently and whose products are predictable from its inputs. The machine begins from an ontology that places the unit of analysis inside the individual and defines the relevant causal variables as internal cognitive properties. It applies an experimental methodology designed for non-recursive natural science to phenomena that are constitutively interrecursive and multimediated. It uses measurement instruments that constitute rather than record the phenomena they purport to measure, creating interrecursive social events and then misidentifying those events as readouts of internal states. And it operates within a political economy that rewards findings of exactly this type: individually targeted, cheaply scalable, compatible with governance frameworks that do not want to engage with structural determinants of unequal outcomes.
The machine's products are predictable. They will always identify an individual cognitive property as the causal variable for outcomes that are constitutively distributed across mediational configurations. They will always be measurable by questionnaire or experimental procedure, because the methodology can only detect what participants can report or perform in controlled conditions. They will always be modifiable through individual-level interventions, because the ontology requires the cause to be inside the individual and therefore in principle directly accessible. They will always be scalable, because individual cognitive properties can in principle be modified in any individual regardless of their social, material, and relational context. And they will always be compatible with governance frameworks that want individual-level explanations for structural inequalities, because the ontology systematically excludes structural and mediational explanation from the start.
The machine's methodology is not incidentally inadequate to the phenomena it addresses. It is constitutively inadequate. The experimental isolation of individuals cannot capture multimediated coordination because it constitutes a specific mediational configuration and treats that configuration's properties as properties of an individual. The quantified measurement of self-reported states cannot capture L1 coordination because L1 coordination is by definition unreflective and leaves no trace that an instrument requiring L3 verbal report could detect. The classification of recursive processes into a single experimental paradigm cannot capture the ontological differences between non-recursive, self-recursive, and interrecursive processes because it applies the same method to phenomena that operate through fundamentally different mechanisms. These are not limitations that better experimental design can overcome. They are consequences of starting from the wrong ontology, and they will persist for as long as the ontology remains unchallenged.
The popular culture transmission mechanism that carries these products from laboratory to bestseller list to policy framework operates through different but complementary logic. A finding achieves wide popular circulation when it is counterintuitive enough to be interesting, confirmatory enough of existing intuitions to feel true, individually actionable by the reader, optimistic about human potential, and expressible in a concrete memorable image. The marshmallow test produces the image of the child and the marshmallow. Self-efficacy produces the image of the patient who believes they can manage their condition and thereby does. The illusion of explanatory depth produces the image of the confident person who discovers they cannot explain a zipper. Growth mindset produces the image of the brain as a muscle that grows with exercise. Each image travels. Each carries the abstract finding into popular memory in a form that is immune to corrective information, because it is held not as a scientific claim susceptible to revision but as a story about human nature that resonates with existing cultural frameworks.
Malcolm Gladwell is the most visible instance of a transmission mechanism that would exist without him. The popular science genre requires a counterintuitive insight, a memorable story that instantiates it, and a generalisation that makes the insight applicable to the reader's life. These requirements are structurally at odds with the actual character of psychological findings, which are probabilistic, bounded, contested, and embedded in specific methodological assumptions that substantially qualify their interpretation. The translation from scientific finding to popular wisdom is not a simplification of a complex truth. It is a substitution of a narrative that satisfies the genre's requirements for a finding that satisfies scientific ones. The substitution happens so smoothly that neither author nor reader notices that it has occurred, because the genre's requirements and the ontological machine's outputs are perfectly matched: both produce individual-level, actionable, optimistic, memorable accounts of human behaviour.
VII. Why Methodological Reform Is Not Enough
The open science movement within psychology has pursued genuine and important reforms. Pre-registration of studies, larger and more representative samples, stronger effect size requirements, adversarial collaboration, and mandatory replication before publication have all been advanced as correctives to the patterns this article describes. These reforms are genuine improvements and they should be supported. They will reduce the most egregious instances of the pattern. They will not dissolve it, because the pattern is not primarily produced by methodological sloppiness. It is produced by ontological error, and methodological improvements within a mistaken ontology produce more carefully executed versions of the same mistake.
Making experimental samples larger and more representative, as the Watts et al. marshmallow replication did, can reveal that an effect was a sampling artefact. It cannot reveal that the construct being measured does not exist as described: that self-control is not a unitary internal property but a distributed interrecursive and multimediated achievement, that waiting time in a specific social situation is not a readout of an individual cognitive variable but a response to a specific mediational configuration of trust, hunger, institutional context, and material history. The latter conclusion requires not better methods but a different ontology, and a different ontology requires different training, different conceptual vocabularies, and a different relationship between psychological science and the qualitative, ethnographic, and phenomenological traditions that have been most attentive to the multimediated and interrecursive character of human coordination.
The collective efficacy and social capital literatures emerged as attempts to correct the excessive individualism of self-efficacy and related frameworks. Each repeated the original error at a different scale. Collective efficacy located the relevant cognitive property in the group rather than the individual. Social capital converted relational life into an accumulating stock. Growth mindset remained focused on the individual child's belief. Each correction remained within the framework of the container ontology, merely adjusting the size of the container. When you do not know that the error is the localisation of distributed coordination inside a container, your corrections will always remain within the framework of containers and will merely vary their parameters. The pattern of correction-that-reproduces-the-error is itself a feature of paradigms that lack the ontological vocabulary to name what is wrong.
The institutional mechanisms that sustain these paradigms are also resistant to methodological reform. Once a construct achieves L4 institutional stabilisation, it becomes self-validating. It generates its own research programme, its own journals, its own training pipelines, its own policy frameworks, and its own popular cultural representations. The construct is no longer primarily in the business of accurately describing the phenomena it was originally designed to address. It is in the business of sustaining the institutional ecosystem that has grown around it. Critiques are evaluated by reviewers embedded in that ecosystem. Replications are designed by researchers whose careers depend on the construct's validity. Policy evaluations are commissioned by institutions whose programmes are built on the construct's assumptions. The feedback loops all run toward preservation. This is not scientific fraud. It is what Living Value Theory calls runaway L4: abstraction that has become self-validating and no longer checks itself against the coordination it originally named.
VIII. No Particular Solution, But One Necessity
The honest conclusion from this analysis is that there is no particular solution. The structural features that produce the pattern are deeply embedded in the institutional organisation of academic psychology, in the political economy of research funding and policy commissioning, in the genre conventions of popular science, and in the cultural demand for simple, optimistic, individually actionable accounts of human potential and failure. None of these is easily changed. The problem has no particular solution because its causes are distributed across institutional, economic, cultural, and ontological domains simultaneously, and no intervention at any single level can dissolve a pattern generated by all of them together.
What follows from this is not resignation. It is clarity about what the situation actually requires. Psychology will continue to produce seemingly compelling but entirely misleading results for as long as it begins from a completely wrong or misrecognised ontology. The wrong ontology is: human behaviour is produced by individual cognitive properties, measurable by instruments that isolate the individual from their mediational context, and causally prior to the interrecursive processes through which those properties are constituted and expressed. The right ontology, or closer to it, is: human coordination is multimediated, operating simultaneously across embodiment, being-with, dwelling, multimateriality, and multisymbolisation in ways that are irreducible to any individual cognitive property; it operates across recursivity levels from seamless L1 background coordination to explicit L5 meta-reflection, with the most valuable and the most invisible coordination concentrated at L1; and it operates through fundamentally different recursive mechanisms in non-recursive, self-recursive, and interrecursive domains that require fundamentally different analytical approaches.
Until psychology develops adequate concepts for multimediality, adequate distinctions between recursivity levels, and adequate differentiation between types of recursive process, it will remain constitutively incapable of accurately describing the phenomena it most influentially addresses. Its experimental findings in these domains will continue to be, not sometimes but systematically, ontologically misconceived renderings of phenomena that the methodology cannot in principle capture. Its popular cultural products will continue to be ideologically convenient stories that locate the sources of structural inequality inside individuals and thereby protect the structural conditions that produce inequality from serious analytical attention. And its policy influence will continue to direct resources away from the mediational configurations that actually determine outcomes, toward the individual cognitive interventions that its ontology permits it to see.
The one necessity that follows from this is the widest possible awareness of just how unreliable and biased the most influential claims of popular psychology actually are. Not the measured academic awareness that acknowledges limitations while affirming the research programme. Not the popular science awareness that celebrates replication failures as science self-correcting. But a more fundamental awareness: that the findings of experimental psychology in the domains of health behaviour, child development, educational achievement, and social cognition are not merely imprecise or incomplete. They are, when applied to the most consequential questions about human flourishing, systematically wrong in ways that follow necessarily from beginning with the wrong ontology. The marshmallow test did not measure self-control. Self-efficacy questionnaires did not measure confidence. The illusion of explanatory depth did not measure overconfidence. Growth mindset interventions did not modify the determinants of educational achievement. Each of these claims measured something real. Each systematically misidentified what it was measuring, because the ontological framework within which it was designed made the misidentification inevitable. And each, in being misidentified, was recruited into policy frameworks that used the misidentification to avoid confronting the structural conditions that the correct identification would have required addressing.
Living coordination is multimediated and in continuous flow. It cannot be frozen in a laboratory, isolated in an individual, reduced to a number on a scale, or captured by an instrument that must stop asking questions at the point where the experimenter's intuitions are satisfied. The most valuable coordinations, those that maintain L1 vitality across all five mediational domains, are the least visible because they are the most successful. They produce no data. They generate no signal. They appear in the psychological research record only through their absence. A discipline that can only see what it can measure will systematically undervalue what works and overattend to what fails. A discipline that can only study individuals will systematically attribute to individual cognitive properties outcomes that are constituted by mediational configurations no experiment can hold constant. A discipline that cannot distinguish between non-recursive, self-recursive, and interrecursive processes will systematically apply the methodology of the first to the phenomena of the third and produce findings whose confidence is unearned.