0. Bridge to the Trilogy

Before introducing Hammond's Cognitive Continuum, it is worth pinning down exactly what gap it comes to fill — because without this explicit tie to the two preceding essays, the entry of a second theoretical apparatus risks reading as a digression rather than a continuation of the same argument.

Article 1 established a shift: from the epistemology of the response to the epistemology of the consequence. Judging an AI system by the plausibility of its output — coherence, apparent factual accuracy, fluency — is insufficient; the relevant test is whether that output, once inserted into the world, produces the distal state it was meant to produce. Although Article 1 did not formalize it as such, this is already a thesis about ecological validity in Brunswik's precise sense: the correlation between a proximal cue (the generated text) and a real distal criterion (the outcome in the environment).

Article 2 applied that same logic reflexively — not to the AI system, but to the mechanism of human supervision itself. In a closed cyber-physical loop, the supervisor is also a Brunswikian judge: their proximal cues are dashboards and alerts; the distal criterion is the real state of the physical system those dashboards are meant to represent. What was formalized there, through Donabedian's output/outcome distinction, is a two-term architecture that corresponds structurally to the one the Lens Model Equation would go on to specify algebraically in Article 5 — output as proximal judgment (analogous to $Y_s$, in the notation fixed there), outcome as distal criterion (analogous to $Y_e$). It also identified a concrete mechanism by which that correspondence erodes — the institutional void, deliberately preferred over any appeal to active resistance on grounds of explanatory parsimony — and tested that hypothesis against empirical evidence (Citigroup 2022, the Iberian blackout of 2025, cybersecurity breakout times in 2026).

What the two articles share, and also what they leave unresolved, is this: both evaluate a judgment — the AI's, the human supervisor's — after that judgment has already been produced, comparing it to the distal criterion after the fact. Neither offers a vocabulary for characterizing, before the judgment is made, the task itself that is being judged: whether that task, by its structure, is better served by an intuitive mode of cognition, an analytical one, or some point in between; whether the same task demands the same cognitive mode when performed by a human and when performed by an AI system; and, above all, which objective properties of the task — not merely performance observed after the fact — allow that prediction to be made in advance. Article 2 comes close to touching the problem in passing, noting that the distribution of authority between human and AI runs into Bainbridge's paradox, without, however, proposing a criterion for deciding a priori what authority architecture a given task should have. Without that vocabulary, Article 4 would have no non-arbitrary basis on which to classify design architectures, and Article 5 would be auditing judgments with no way of knowing, in advance, what kind of judgment it should be observing.

It is precisely this gap that Kenneth Hammond's Cognitive Continuum fills — and it does so not as an import of a theoretical apparatus foreign to the lineage already established, but as its direct continuation. Hammond was a student of Brunswik, edited the 1966 commemorative volume dedicated to his work, and founded the Brunswik Society in 1983; his Social Judgment Theory was explicitly born as an extension of the Brunswikian lens model to the analysis of human judgment, and it is on that same foundation that he builds Cognitive Continuum Theory (CCT): the thesis that intuition and analysis do not form a dichotomy but a continuum, that judgment tasks define a parallel continuum, and that cognitive performance depends on the degree of match between the mode of cognition invoked and the properties of the task at hand. Where the lens model — in Articles 1 and 2, and later in Article 5 — describes how a judge, human or, by extension, artificial, combines cues into a judgment, CCT describes what the task, by its own structure, demands of that judge. That is the one piece, and only that piece, still missing from the trilogy at this point.

Article 3 introduces it in five steps. Section 1 recovers the functional structure of the Cognitive Continuum as Hammond formulated it. Section 2 examines the eleven task properties that allow a task to be located on that continuum, in the formulation of Doherty & Kurz, and chooses between two competing formulations. Section 3 maps those eleven properties, property by property, from the classic human task to the cyber-physical task of an AI-IoT system. Section 4 adapts the resulting Task Continuum Index to that same domain. Section 5 illustrates the result conceptually, and Section 6 identifies what deliberately remains open, setting up the entry into Article 4. The purpose of this theoretical foundation is not to replace the ecological validity established in Articles 1 and 2, but to precede it logically: before asking whether a judgment — human, artificial, or the chain of the two — achieved ecological validity, one must know what kind of judgment that task demanded in the first place.

(Bibliographic note: the original formulation of Cognitive Continuum Theory is found in Hammond, K. R. (1996). Human Judgment and Social Policy: Irreducible Uncertainty, Inevitable Error, Unavoidable Injustice. New York: Oxford University Press [hardcover, ISBN 0195097343, LCCN 95035319 — confirmed in an academic library catalog and in a WorldCat/MARC record]. The citation "Hammond (2000)," which appears in some later reviews of the theory, refers to the paperback reissue of the same work [ISBN 0195143272], not to a second edition with distinct content — there is, therefore, no real ambiguity regarding the date of the original formulation. The formulation of the eleven task properties used in this article follows Doherty & Kurz (1996), already verified in the trilogy's core bibliography.)