The Rearview Mirror Problem: A Catholic Psychological Response to AI and Literature

The argument in the rearview

Carlo Cordasco, in his Aeon essay 'Illegible benefits,' names a recurring structure in the history of disruptive technology. When a genuinely transformative practice arrives, its costs are immediately measurable against existing standards. Its benefits are not — because many of them depend on vocabulary, methods, and possibilities that the practice itself will eventually create. Cordasco calls these gains 'illegible benefits': real but not yet speakable in the language available at the moment of evaluation.

The sharpest example in Cordasco's essay is ether anesthesia. When ether appeared in the 1840s, experienced clinicians could document its costs with precision — lost patient feedback, disrupted surgical communication, the genuine risk of death from the agent itself. What they could not document was open-heart surgery, because open-heart surgery was not yet a conceivable procedure. The deepest benefit of ether was not simply 'less pain during amputation.' It was the opening of an entirely new surgical possibility space, and that space could not be described in 1846 because it did not yet exist.

Cordasco extends this analysis to artificial intelligence. Critics of AI can point to concrete present harms — labor displacement, misinformation, cognitive atrophy in certain skills. Proponents gesture at transformations that are, by definition, difficult to specify before they have occurred. The natural response to this asymmetry is either optimism ('the benefits will outweigh the costs, as they have before') or pessimism ('the costs are real and the benefits are speculative'). Cordasco's actual argument is more careful than either: the asymmetry is structural, not a reason for confidence or alarm in itself. Legibility is historically conditioned. Any evaluation made entirely within the present vocabulary is necessarily partial.

This is a serious epistemological observation. It is also, as it turns out, a problem the Catholic intellectual tradition has been working on for several centuries — not primarily as a feature of technology, but as a feature of the human person ordered toward an end that exceeds present comprehension.

What Aquinas already knew

The Thomistic account of teleology handles Cordasco's structural observation directly. Aquinas, working from Aristotle's Nicomachean Ethics and Physics, holds that the proper criterion for judging any action or practice is its finis — its end. A being ordered toward a genuine good participates in that good even when its immediate form is turbulent or costly. The acorn does not look like an oak. Evaluated by the standards appropriate to an acorn, the oak is an extravagant waste. Evaluated by the standards appropriate to an oak, the acorn is already a success in potentia.[^6]

This is not a license for wishful thinking. Aquinas is careful: the goodness of an end does not automatically sanctify the means, and not every disruption is ordered toward a genuine good. The question is whether there is a real ordo — a structural relationship between present action and true human flourishing — or merely an asserted one. Cordasco's historical cases suggest that ether genuinely was ordered toward an expansion of surgical possibility that constituted a genuine human good, even though that good was illegible in 1846.

The deeper Thomistic point is epistemological. Aquinas holds that every entity, insofar as it is capable of being known, is true; and insofar as it is capable of being desired, it is good.[^4] But knowability and actual knowledge are not the same thing. The intelligibility of a thing exceeds any particular knower's current access to it. What Cordasco calls 'illegible' is, in Thomistic terms, not yet adequately apprehended by the available cognitive instruments — a precise description, not a counsel of despair.

Roger Verneaux makes a related point about intelligibility and rational access: from the fact that being is intelligible, one cannot immediately deduce that it is fully explicable by the rational tools presently available, since reason is a strictly human function and cannot be erected into an absolute.[^7] The being precedes the explanation. Premature closure — declaring a practice harmful because its benefits cannot yet be articulated — is not rationality but a particular form of epistemological overreach.

The asymmetry of natural law application

Cordasco's argument carries a further implication that he does not develop but that the Catholic natural law tradition addresses directly. Zanotti, following Aquinas, observes that not all common principles of natural law can be applied in the same way to all persons and all historical circumstances, because the variety of human situations is too great.[^1] The principles — promote genuine human goods, avoid real harms, respect the dignity of persons — are stable. Their application to any specific transformative practice requires prudentia, the practical wisdom Aquinas describes as the virtue that bridges stable principles and historically conditioned circumstances.

Prudence takes time to develop. It learns from experience and calibrates itself to actual situations rather than theoretical models of them. This is precisely what Cordasco's historical argument calls for, even if he does not use the term: an evaluation of transformative technology that is epistemically humble about the limits of present legibility, without collapsing into either optimism or pessimism. That humility is the recognition that human persons are ordered toward ends that exceed what any given historical moment can fully articulate.

Where Cordasco's framework needs supplementing

Cordasco's essay operates almost entirely within a consequentialist register: costs versus benefits, legible versus illegible, measurable now versus measurable later. This framing handles the epistemological problem well. It handles the anthropological problem less well.

The deepest question raised by transformative technologies is not 'are the benefits large enough to outweigh the costs?' It is 'what kind of persons are formed by habitual engagement with this practice?' Aquinas argues that human beings are constituted by their acts — that repeated action forms habitus, stable dispositions of the soul that then shape subsequent perception, desire, and choice.[^6] The danger of any powerful new practice is not only that its costs may exceed its benefits in a ledger calculation, but that it may form persons who are less capable of the very excellences the practice was supposed to serve.

Character formation, like the illegible benefits Cordasco describes, unfolds slowly and becomes visible only after the fact. The Catholic tradition's insistence on virtue as the proper medium of human flourishing is, in this respect, an antidote to both the naïve optimism that dismisses costs and the naïve pessimism that cannot see future goods. It asks not only 'what will this technology produce?' but 'what will this technology make of us?'

Maritain's observation that Thomism functions as a living organism rather than a closed system is apt here: what is most remarkable in it is that the character of organism prevails over the systematic aspect.[^8] A living tradition does not evaluate new practices by matching them against a fixed checklist. It brings stable principles into genuine contact with new realities, allows the contact to generate new questions, and develops its understanding through that encounter.

The grammar that does not yet exist

Cordasco's framework predicts that the deepest benefits of any transformative practice will only become articulable after the practice has reshaped the conceptual space in which articulation happens. The same holds for the question of character formation under AI. The costs of certain habits of attention — what is built or lost through extended engagement with language models, algorithmic mediation, and distributed cognition — are real and already partially legible. The Catholic anthropological tradition, with its account of habitus and the way repeated acts shape the soul, takes those costs more seriously than a simple cost-benefit calculation allows.

But the full grammar for asking what AI will make of us has not yet been written. What Cordasco's structural observation adds to the Thomistic account is a historical patience: the recognition that the vocabulary adequate to evaluate a transformative practice is partly produced by the practice itself. That does not resolve the question of whether any particular engagement with AI orders the person toward genuine human flourishing. It means the question cannot be answered entirely in advance, and that prudence — patient, experience-calibrated, ordered to actual human goods — is the appropriate instrument for continuing to ask it.

The grammar for AI's deepest benefits may not yet exist. The grammar for asking what AI will make of us has been under development for centuries.

References

[^1]: Gabriel Zanotti, 'Economia de Mercado y Doctrina Social de la Iglesia,' noting that not all common principles of natural law can be applied in the same way to all persons because of the great variety of human situations.

[^4]: Gabriel Zanotti, 'Free Market Economy and Social Doctrine of the Church': every entity insofar as it is capable of being known is true; and insofar as it is capable of being desired, it is good.

[^6]: Gabriel Zanotti, 'Comentario a la Suma Contra Gentiles': everything through which man tends toward his natural end is naturally right, and what is contrary to it is naturally harmful; and on the Thomistic account of habitus, that repeated acts form stable dispositions of the soul.

[^7]: Roger Verneaux, Epistemologia General o Critica del Conocimiento: from the fact that being is intelligible it cannot be deduced that it is fully explicable by rational tools presently available, since reason is a strictly human function that cannot be erected into an absolute.

[^8]: Jacques Maritain, Distinguish to Unite, or The Degrees of Knowledge: what is most remarkable in Thomism is that while it is sovereignly one and linked together in its parts, the character of organism prevails over the systematic aspect.