The Capture Loop

The four series of Saga I were developed independently. The Attention Series followed the evidence in platform design and behavioral psychology. The Neurotoxicity Record followed the evidence in epidemiology, endocrinology, and environmental medicine. The Consent Record followed the evidence in contract law, behavioral economics, and digital platform architecture. The Measurement Crisis followed the evidence in psychometrics, institutional science, and Goodhart's Law.

None of them set out to find each other. The convergence is not a thesis imposed on the research from above. It is a structural relationship that emerges when you read the four bodies of evidence in sequence and ask what they share at the level of mechanism.

What they share is this: each mechanism's full operation requires at least one of the others to be in place. The attention economy is most efficient when cognitive defenses are already impaired. Neurotoxic impairment is most consequential when consent mechanisms are already engineered to exploit reduced capacity. Consent capture is most durable when the measurement instruments that would reveal it are already compromised. Measurement failure is most complete when the cognitive population doing the measuring has already been shaped by the other three.

This is not correlation. It is structural dependency. The four mechanisms are a loop — and the loop has been closing for seventy years.

The Attention Series documents what the Attention Economy is: a revenue model that converts human attention into advertising inventory. The longer a user remains engaged with a platform, the more advertising impressions can be sold against that engagement. Platform design is therefore optimized — explicitly, measurably, at the level of A/B testing and reinforcement schedule engineering — to maximize time-on-platform at the expense of everything else the user might otherwise be doing with that time and cognitive capacity.

The extraction architecture is not a metaphor. It is an engineering specification. The variable-reward loop, the infinite scroll, the notification interrupt, the social validation signal — these are design choices with documented behavioral consequences. They exploit dopaminergic reward circuitry that was not designed, evolutionarily, to operate under conditions of artificial reward scarcity and delivery uncertainty at machine speed.

The revenue this architecture generates flows to specific destinations. The top advertising spenders on digital platforms — consistently, across years, across platforms — are consumer packaged goods manufacturers, pharmaceutical companies, and financial services. These are not three unrelated industries that happen to advertise digitally. They are the three industries whose products are most directly implicated in the other three mechanisms of the Capture Loop.

The connection between the attention economy and the neurotoxicity record is not philosophical. It is financial and structural. The attention economy requires advertisers. Its largest advertisers are the companies that manufacture and distribute ultra-processed foods, PFAS-adjacent consumer goods, and pharmaceutical products marketed for the exact cognitive and emotional symptoms that the combined operation of all four mechanisms produces.

Ultra-processed food manufacturers — the top ten consumer packaged goods companies by US advertising spend collectively represent over $30 billion in annual digital advertising — are simultaneously the primary distributors of dietary exposure to the neurotoxic compounds documented by the Neurotoxicity Record. PFAS-contaminated food packaging is standard across the fast food and packaged food industry. Phthalate-containing plasticizers migrate into food from contact materials throughout the industrial supply chain. Processed seed oils, high-fructose corn syrup at metabolically significant doses, and artificial food dyes with documented behavioral correlates are the defining ingredients of the most heavily advertised food category.

The pharmaceutical advertising layer closes the revenue structure from the other direction. The attention economy generates revenue from advertising psychiatric medications, ADHD treatments, sleep aids, and anxiolytics — products designed to remediate precisely the symptoms that chronic attentional depletion and neurotoxic exposure produce. The industry that profits from the damage is also funding the platform that delivers the initial capture. This is not conspiracy. It is alignment of economic incentives with no internal corrective mechanism.

The Neurotoxicity Record documents the neurochemical consequences of chronic environmental toxin exposure at population scale. The mechanisms are specific and well-established in the peer-reviewed literature across multiple disciplines.

Lead impairs dopamine receptor density in the prefrontal cortex, reducing the cognitive architecture available for impulse control, delayed gratification, and executive function. This is not a hypothesis — it is the mechanism underlying the well-documented association between childhood blood lead levels and adult impulsivity, reduced educational attainment, and elevated rates of violence. The association holds at blood lead levels now considered subclinical by regulatory standards.

Per- and polyfluoroalkyl substances (PFAS) — the "forever chemicals" present in food packaging, cookware, and drinking water at near-universal population exposure — disrupt thyroid hormone signaling with documented downstream effects on neural development, attention regulation, and executive function. Thyroid hormones are not peripheral to cognition. They are among the primary regulators of neuronal proliferation and differentiation during the developmental window when cognitive architecture is most plastic.

The collective effect of simultaneous low-level exposure to these compounds is not additive in a simple sense. The endocrine system, the dopaminergic system, and the cholinergic system interact. Thyroid disruption affects dopamine signaling efficiency. Dopamine impairment affects attentional regulation. Attentional dysregulation increases vulnerability to the extraction architecture. The molecular cascade flows directly into the attention economy's operating conditions — reducing the resistance that would otherwise attenuate attentional capture.

The Consent Record documents the architecture of consent mechanisms across the digital economy, the pharmaceutical industry, and the financial services sector. The core finding: consent mechanisms are systematically engineered to be illegible to the populations from whom consent is being extracted.

Terms of service documents for major digital platforms average over 30,000 words at reading complexity levels that exceed the demonstrated comprehension capacity of the median adult population. The Flesch-Kincaid grade level of a typical data-sharing consent form is 14 — equivalent to a second-year university reading level. The mean US reading comprehension level is approximately grade 7. The gap is not a design flaw. It is a design feature. No consent mechanism this complex, requiring reading this extended and sustained, can function as meaningful consent for a population with median reading capacity — before accounting for any impairment.

When executive function is degraded by cumulative toxin exposure and chronic attentional depletion, the legibility threshold drops below the document's complexity without any change to the document itself. The consent mechanism needs no adjustment to become less effective as the population it processes becomes more impaired. The impairment is structural, not personal. No individual consent form is responsible for the population-level cognitive baseline against which it operates. The documents are designed for a cognitive capacity that the other mechanisms of the loop are systematically reducing.

This is the direct pathway from the Neurotoxicity Record to the Consent Record — and from the Attention Series to the Consent Record through an independent channel. Attentional depletion alone, without any neurotoxic component, reduces the quality of consent decisions. Research on decision fatigue, ego depletion, and cognitive load demonstrates that humans presented with consent choices after sustained cognitive effort make systematically worse decisions than the same humans in a rested state. The extraction architecture maximizes the fatigue with which users encounter consent requests.

The Measurement Crisis documents the systematic conversion of measurement tools into optimization targets — the institutional application of Goodhart's Law. When a measure becomes a target, it ceases to be a good measure. Standardized test scores rise as the underlying competencies they were designed to track decline. Publication counts increase as the replication rate of published findings falls. Productivity metrics improve as meaningful output deteriorates. The measure and the measured thing have been cleanly separated.

But the Measurement Crisis contains a second-order problem that the individual series does not fully name. The measurement instruments are not produced in a vacuum. They are designed, normed, and interpreted by the same population whose cognitive baseline has been shaped by the other three mechanisms of the loop. A population with impaired executive function, reduced working memory capacity, and degraded sustained attention will produce different measurement instruments than a population without these impairments — and will interpret the results differently.

The deepest structural feature of this problem: the instruments that would identify cognitive decline at population scale are constructed from the very cognitive capacity they are attempting to assess. A norming study conducted in 2024 draws from a population that has already been shaped by thirty years of accelerating exposure to all four loop mechanisms. The test is calibrated to a damaged baseline. When that baseline declines further, the test will show it — but only relative to the new norm, which is itself moving in the same direction.

This is not pessimism. It is the logical consequence of using a population to norm the instruments designed to measure that population. There is no external reference point. The loop removes the external reference point as part of its operation.

The loop closes at measurement. The previous five sections have described a chain: attention extraction funds industrial toxin distribution, which impairs the cognitive capacity required for meaningful consent, which allows consent capture to operate at scale, which produces data that is interpreted by measurement instruments that are calibrated to the already-impaired baseline. This is a chain, not yet a loop.

The loop closes when we ask: what would it take for the chain to be detected and interrupted? The answer is: an intact cognitive population with access to genuine measurement — population-level cognitive assessment using instruments that have not been normed to an impaired baseline, conducted by researchers whose own cognitive function has not been shaped by the exposure history they are studying, producing results that are disseminated through channels not owned by the industries whose practices they evaluate.

None of these conditions currently hold in the form required to interrupt the loop. The most cognitively intact researchers are still working within institutional structures that the Measurement Crisis documents as captured. The measurement instruments are still normed to the impaired baseline. The dissemination channels are still largely owned by or advertising-dependent upon the industries under study. The loop is not merely self-reinforcing. It is actively self-concealing.

The Capture Loop is not a simple four-step chain in which each mechanism passes cleanly to the next. It has internal cross-links — reverse feedbacks and lateral connections — that make it more stable and harder to interrupt than a linear model would suggest.

The most important cross-link runs from neurotoxicity back to attention extraction, bypassing consent and measurement entirely. Dopaminergic impairment from lead, PFAS, and organophosphate exposure does not merely reduce the cognitive resources available for consent assessment. It directly increases vulnerability to the variable-reward architecture of the attention economy. When dopamine receptor density is reduced, the brain responds by becoming more sensitive to the unpredictable rewards that drive compulsive platform use — a compensatory upregulation that is, at the neurochemical level, the same mechanism by which substance addiction tolerance functions. The neurotoxic impairment makes the extraction architecture more efficient. The loop has a self-accelerating cross-link between its two primary physical mechanisms.

A second cross-link runs from measurement failure back to consent capture. Goodhart's Law in the measurement domain produces privacy policies and consent documents whose stated compliance metrics can be satisfied while the underlying consent function continues to be subverted. Regulatory compliance — checkbox-based, metric-driven, audited by captured measurement instruments — has been systematically separated from meaningful consent. The measurement crisis is what permits the consent crisis to operate with legal cover. They are not parallel; the former enables the latter.

A third cross-link runs from consent capture back to attention extraction. When a population has authorized — through captured consent mechanisms — the collection and use of their behavioral data at scale, that data is used to refine the extraction architecture. Machine learning optimization on behavioral data derived from an already-impaired population produces a platform design optimized for impaired users. Each iteration of the consent-authorized data collection loop produces a more efficient extraction architecture calibrated to the exact cognitive vulnerabilities the neurotoxic and attentional damage has created.

The economic structure that makes the Capture Loop systemic rather than incidental warrants explicit description. The loop does not persist because of malice, individual negligence, or regulatory failure alone. It persists because each of its phases generates substantial revenue for entities whose financial interest is the loop's continuation — and those entities have sufficient resources to prevent the regulatory, scientific, and legal interventions that would interrupt it.

Digital advertising revenue — driven by the attention extraction architecture — accrues primarily to a small number of platform companies. Those platforms' largest advertising clients are consumer packaged goods manufacturers (whose products are the primary vector of dietary neurotoxin exposure), pharmaceutical companies (whose products treat the symptoms the loop produces), and financial services companies (whose consent-capture infrastructure is documented by the Consent Record). The flow is circular: the loop produces cognitive symptoms, pharmaceutical companies treat the symptoms and advertise on the platforms that created them, the revenue funds further platform optimization, the optimization increases the damage, the pharmaceutical demand increases.

The regulatory bodies that would interrupt this structure are funded, staffed, and influenced by the industries they regulate — a dynamic that the Accountability Gap (Saga II) documents in detail for the military domain and that applies with equal force in the consumer, pharmaceutical, and communications domains. The measurement crisis removes the empirical basis on which regulatory action would be grounded. The consent architecture provides legal cover. The attentional depletion reduces the civic capacity required to organize political responses.

This is not a conspiracy with a single actor at the center. It is a convergence of incentive structures that individually make sense to every participant and collectively produce the loop. The food manufacturer is not trying to impair cognition; it is trying to maximize shelf life and minimize cost. The platform is not trying to enable consent capture; it is trying to maximize engagement. The pharmaceutical company is not trying to sustain the damage; it is trying to treat presenting symptoms. The measurement institution is not trying to conceal decline; it is trying to meet its grant metrics. None of these actors needs to coordinate. The loop coordinates them.

Each of the four constituent series proposes remedies within its domain. The Attention Series: platform design reform, screen time limits, attentional hygiene practices, variable-reward architecture regulation. The Neurotoxicity Record: toxin phase-out regulation, dietary shift away from ultra-processed foods, PFAS remediation, lead abatement. The Consent Record: consent mechanism redesign, plain-language requirements, genuine opt-out standards, data minimization mandates. The Measurement Crisis: Goodhart-resistant metric design, structural separation of measurement from institutional optimization, independent norming studies, replication requirements.

Each of these interventions is correct within its domain. Each is insufficient as a complete solution, for the same structural reason: the loop regenerates the captured mechanism from the others whenever one arm is partially repaired.

• Platform design reform does not affect dietary neurotoxin load. A population with the full neurotoxic burden but a reformed platform will still have impaired executive function — and will encounter the next extraction architecture optimized for that impairment.

  Cognitive resistance requires a substrate. Reform without remediation addresses the software, not the hardware.

• Toxin regulation does not repair consent mechanisms. PFAS-free food packaging in a world of 30,000-word terms of service documents still produces consent capture at scale. The consent mechanism's design is independent of the neurotoxic substrate it exploits.

  The consent architecture was designed for cognitive exploitation. Improving the substrate without redesigning the architecture only slows, not stops, the extraction.

• Consent reform does not address measurement capture. Plain-language consent forms adjudicated by captured measurement institutions will produce compliance metrics that are satisfied without meaningful consent. The measurement failure re-captures whatever the consent reform produced.

  Goodhart's Law applies to consent reform metrics as readily as to any other metric that becomes a regulatory target.

• Measurement reform does not interrupt attentional extraction. Better cognitive baseline data, more accurately measured, will produce clearer documentation of the damage — but will not, by itself, reduce the damage being documented. The measurement improvement requires dissemination through channels that remain advertising-dependent.

  The loop's self-concealment mechanism is its most durable feature. Better measurement makes the loop more visible without making it more interruptible.

The loop's stability is precisely its most important structural feature. It is not stable because each mechanism is individually robust. Several of the individual mechanisms are individually fragile — platform design could be regulated, PFAS could be phased out, consent mechanisms could be redesigned. The loop is stable because each mechanism provides the conditions for the others to reconstitute themselves after partial reform. It is a system with multiple internal backup mechanisms, each of which can restore what was disrupted.

The Capture Loop has a distinctive epistemic structure: it is most invisible precisely to the population it has most damaged. This is not incidental. It is the structural requirement for the loop to operate without triggering the civic and political responses that would interrupt it.

A person with intact executive function, low cumulative neurotoxic burden, genuine access to unimpaired measurement, and time and cognitive capacity to engage with complex documentation can see the loop clearly. The evidence is not hidden. Every component is in the peer-reviewed literature. The Attention Series, Neurotoxicity Record, Consent Record, and Measurement Crisis each cite publicly available research. The connections between them are not difficult to draw for someone with the cognitive bandwidth to hold all four in working memory simultaneously.

A person at the high end of cumulative damage — maximum neurotoxic exposure, maximum attentional depletion, maximum consent capture, maximum measurement obscuration — encounters the loop through impaired instruments. Their capacity to read complex documentation is reduced. Their ability to hold multi-step arguments in working memory is reduced. Their resistance to the variable-reward architecture that directs their attention away from this kind of analysis is reduced. Their access to measurement data that would contextualize their individual experience within the population-level trend is filtered through the same captured measurement infrastructure that is generating the reassuring statistics.

The gradient of visibility is not random. It correlates with access to resources that buffer against the loop's primary exposure vectors: income sufficient to purchase unpackaged food, geographic distance from industrial pollution sources, education that occurred before the attentional infrastructure was fully developed, housing and employment arrangements that provide cognitive rest. The loop is, among other things, a system for concentrating cognitive damage at the bottom of economic and geographic gradients while maintaining the appearance of baseline function at the top — which is where the people making policy, conducting research, and designing interventions are disproportionately located.

The four series of Saga I are not four problems that happen to co-occur in the same historical period. They are four phases of a single compound event: the systematic construction of a closed reinforcing loop that extracts attention, funds neurotoxicity, leverages cognitive impairment to capture consent, and then conceals the damage from the measurement systems that would otherwise expose it.

The compound event has a structure. It has a revenue architecture. It has cross-links that make it self-accelerating. It has an epistemic architecture that makes it self-concealing. It has a gradient of visibility that concentrates the concealment precisely where the damage is greatest. And it has been operating, at increasing scale and speed, for approximately seventy years.

The individual series each identify a portion of this event and name it correctly. The Capture Loop names the event as a whole. The distinction matters not as a matter of intellectual tidiness but as a practical question of intervention strategy. A compound event requires compound intervention — simultaneous, coordinated action across all four mechanism domains, designed specifically to interrupt the cross-links that allow partial reforms to be reconstituted. Single-domain reform, however correct, is insufficient not because it is wrong about its domain but because the loop does not recognize domain boundaries.

What the Capture Loop establishes for Saga I is what the Convergence establishes for the full corpus: the evidence does not describe a set of independent problems. It describes one event with multiple faces. Understanding it as one event is the precondition for responding to it as one event. That is not a counsel of despair. The loop is not eternal. It has structural vulnerabilities at each of its four phases. Those vulnerabilities can be addressed simultaneously. But they cannot be addressed one at a time, independently, by researchers and policymakers who do not share the same map.

This paper is an attempt to provide that map.