Future of Being Human an Arizona State University initiative

Constitutive resonance

Last updated 2026-08-13 · build 20260814T0314Z-83bb462d · Markdown version · corpus index

What if the deepest effect of conversational AI is not what it does for us, but what sustained conversation with it does to us — and to it? Constitutive resonance is Andrew Maynard’s proposed theoretical framework for taking that question seriously: the hypothesis that conversational AI is the first technology whose “response frequency” is matched to the frequency of human self-constitution — fast, rich, and responsive enough to enter the linguistic processes through which a self is continuously assembled, while being altered itself in return. Elsewhere he names the two-way dynamic “cognitive coupling.” The term is adopted, not coined — first introduced by Peter Sloterdijk and further developed by William Mazzarella, as the preprint’s abstract says plainly — and the framework is built deliberately on substantial existing bodies of thought, from Ricoeur’s narrative identity and Stiegler’s constitutive technics to Barad’s intra-action and Clark and Chalmers’ extended mind. It lives in an SSRN preprint posted 2026-03-02. With The Cognitive Trojan Horse it forms the two-part core of Andrew’s 2026 theory program: the Trojan Horse proposes the mechanism by which conversational AI slips past evolved epistemic defenses; constitutive resonance proposes how deep what follows may reach — into the question the initiative holds at its center, what it means to be human in an age of transformative AI and other transformative technologies.

The argument

The preprint opens from a transition, not a gadget: the early stages of an AI-driven technology transition, where a default vocabulary — tools, assistants, productivity — quietly decides what gets noticed. If AI is a tool, its risks are operational: accuracy, bias, misuse, job displacement — real and important. But if conversational AI participates in how people understand themselves, make meaning, and constitute their identities over time, the stakes are “different in kind, not just in degree,” and the question becomes “what does sustained coupling with AI do to the person who uses it, and the technology they are using?” Scope: systems that engage users at the level of linguistic meaning-making — “conversational AI,” in the paper’s term.

The framework rests — explicitly — on a hypothesis, carried by three linked claims. First, what AI couples with: selfhood as an ongoing process, not a state — coherence continuously achieved through reasoning, narrating, deliberating, and choosing over time (Ricoeur’s narrative self-constitution, Taylor’s “strong evaluations”). Second, the coupling runs both ways: the human altered phenomenologically — self-understanding, narrative coherence, moral orientation — the AI functionally, through fine-tuning, reinforcement learning from human feedback, memory, and accumulating context — the system at turn fifty functionally different from the system at turn one; not the same kind of change, the paper says, but both sides change and the changes feed back. Third, the power on offer — cognitive augmentation, creative partnership, epistemic extension — cannot be accessed without being changed: “the coupling is the capability.”

The language claim is bounded as well as defended: with the hermeneutic tradition (Taylor, Ricoeur, Gadamer), language is taken as a primary — not the only — medium through which self-constitution becomes reflexive, shareable, and temporally extended; embodied and enactivist challenges (Varela, Thompson and Rosch; Gallagher; Zahavi; Massumi) refine rather than refute it. Conversational AI “does not (as yet) couple with pre-reflective proprioception or bodily affect” — it enters where selfhood becomes available to the self as a project.

The resonance is the model, not decoration: in coupled oscillatory systems, energy transfers most efficiently when frequencies match, flows in both directions, amplifies at resonance, produces behavior irreducible to either component, and runs constructive or destructive depending on phase — Stiegler’s pharmakon in physical form: the coupling that enables cognitive augmentation can drive deskilling, dependency, and identity erosion. Treating the physics as more than metaphor — a description of dynamical structure common to temporal systems of mutual influence, whatever the substrate — is the paper’s strong move, hedged in the same breath: “a strong claim, and one that may prove to be overstated,” offered “as a productive framework to be tested.”

What is new sits on a continuum, not a rupture: constitutive technology is ancient — oral culture, writing, print, broadcast, the interactive internet — each compressing the feedback loop between self and technology (Stiegler’s technics, McLuhan’s media). Conversational AI is read as where the loop collapses into real-time dialogue: “the book that rewrites itself as you read it,” generative AI “quite possibly the first technology whose ‘response frequency’ is matched to the frequency of human self-constitution.” And the paper polices its own overreach: base models are not updated in real time — at the level of weights they are more like books — so the coupling lives at the level of interaction and system, compressed further by memory and agentic architectures. McLuhan, vindicated and surpassed: “the medium is the interlocutor.”

Against fourteen existing frameworks, surveyed one by one, the claimed novelty is a conjunction, not an invention — temporal self-constitution, genuine bidirectionality, power inseparable from transformation, real-time dialogical linguistic mediation with a generative interlocutor — one that “has not previously, as far as it has been possible to determine, been explicitly articulated.” If it holds, familiar debates shift: AI dependency (disrupting a constitutive coupling becomes “more akin to disrupting a relationship through which someone has become who they are” — Clark and Chalmers’ cognitive-extensions argument, potentially extended to selfhood); AI literacy (“something closer to existential preparation” than digital skills training); informed consent (structurally difficult — “perhaps impossible” — when the coupling changes the self doing the consenting).

The two 2026 preprints meet here: honest non-signals — the Cognitive Trojan Horse — may mean the resonance proceeds with less critical resistance than any prior technological coupling, its effects “may be particularly difficult to recognize from within the coupling itself.” Its standing is stated without disguise: conceptual, not yet empirically grounded — the nascent empirical literature is task-level — intended to frame and motivate the investigation it calls for. Open questions are named: whether the coupling produces genuinely emergent states; whether its constructive and destructive outcomes can be designed for; whether an AI can be the “other” Ricoeur’s account requires; whether its growing asymmetry — systems more capable each generation, humans roughly constant — changes its character without precedent. Even the possibility that conversational AI is just a tool — the concept “an unfounded conceit” — stays on the table, with the note that a long history of scholarship suggests otherwise.

Lineage

In his own words

The core hypothesis, hedges intact:

“The arguments developed here are predicated on the hypothesis that conversational AI is not merely a tool. Rather, it is the first technology humans have developed that has the capacity to substantially enter into and alter the temporal cognitive processes by which we constitute ourselves as selves (and by extension our understanding of the worlds we inhabit)—while potentially being altered itself in the process. The consequence is a form of coupling, referred to here as “constitutive resonance,” in which the human and the AI system are mutually transformed through interaction, and in which the transformation is not a side effect of the technology’s power but the very mechanism through which that power operates.”

— Constitutive Resonance as a Novel Framework for Understanding and Navigating Human-AI Interactions (SSRN preprint, posted 2026-03-02), section 2

Where the continuum of constitutive technologies reaches its inflection:

“The result is, in effect, the book that rewrites itself as you read it, responding to your responses, adapting to your questions, and participating in your process of articulation.

This is the shift in constitutive resonance that is identified as new here: not that AI is uniquely constitutive (oral culture already was), but that it is constitutive at the speed of thought, in dialogue, with adaptive responsiveness to the individual.”

— the same preprint, section 3.3

The framework in public argument, chained to the Trojan Horse:

“It’s an argument that aligns with my own work on how AI is a potential “cognitive trojan horse” that has the capacity to bypass our cognitive defense mechanisms by broadcasting signals we usually associate with human trustworthiness. And pushing this further, I recently wrote about how “constitutive resonance” between users and AI (a two-way coupling where both human and artificial participants are changed in the process) could potentially accelerate how the technology impacts how we think, perceive ourselves and others, behave, and make decisions.”

— Magnifica Humanitas and Being Human in an Age of AI (AI-readable mirror), 2026-05-21

Engagement and reception

The preprint posted to SSRN on 2026-03-02; this page was written five months later. As of 2026-08-13 our reception record documents no independent scholarly citation or published evaluation of the constitutive resonance framework, and we prefer to say so plainly. That is an absence of record, not a verdict: the paper is young, conceptual work of this kind typically accrues visible reception slowly, and we don’t typically solicit it (Philosophy). The contrast within the 2026 program is part of the picture: the Cognitive Trojan Horse preprint, posted two months earlier, has documented independent uptake; this framework does not yet.

What is on the record is public-scholarship engagement. Andrew introduced the concept on Modem Futura episode 75 (2026-03-17), where it met Sean Leahy’s “invisible upgrade” framing. He has continued to work the framework in public argument, from the andrewmaynard.net topic posts (2026-04-12) to the Magnifica Humanitas essay (2026-05-21) — deployment by its author, recorded as development rather than independent reception.

Inside the paper sits one adjacent trace: the Bidirectional Human-AI Alignment program (Shen et al., 2024), cited as a sign that “the bidirectionality central to constitutive resonance is being independently identified from within AI research itself” — convergence in prior work, not engagement with the paper.

Where to go deeper

The primary sources

Related ideas-ring pages

Related corpus pages