Reality-Contact Is Accountable Mediation

Reality-contact is not purity of access.

This is a working note from the Reality-Contact Resilience Explorer project. It is part of an exploratory civic-resilience model, not a formal research paper or validated index.

Start with a book.

A book can seem like a straightforward case of contact with another mind. There is a text. There is a reader. There is something to be understood.

But even this apparently simple case is mediated. We read closely, skim, forget, misremember, hear about, quote, summarize, discuss, and sometimes perform knowledge of books we barely know. The point is not that reading is impossible or that all forms of contact are the same. The point is that contact comes in degrees, and those degrees matter.

Pierre Bayard is useful here as a diagnostic figure, not as a core pillar of the Reality-Contact Resilience framework. In How to Talk About Books You Haven't Read, he distinguishes books one does not know, books one has skimmed, books one has heard about, and books one has forgotten.¹ His terms name a familiar problem: even literary culture, one of our great technologies of contact, includes partial, remembered, reputational, and socially mediated encounters. Non-reading is not the same as reading. But our ways of knowing need a more honest vocabulary.

The Mistaken Ideal

It is tempting to define reality-contact as direct access: seeing for oneself, reading the whole source, touching the real thing, being physically present, getting past mediation and finally meeting reality without distortion.

There is truth in that desire. Direct encounter matters. Bodies matter. Places matter. Sources matter. Institutions and arguments should remain open to correction by the world they describe.

But human beings rarely live by direct access alone. We live through language, memory, trust, maps, books, expertise, institutions, records, summaries, arguments, and inherited forms of attention. Most of what we know is mediated. That is not a defect to eliminate. It is a condition to discipline.

The danger is not mediation itself. The danger is mediation that loses accountability.

Accountable Mediation

Reality-contact is not purity of access. It is accountable mediation.

A mediated claim remains reality-tethered when there are paths back to the object, text, person, event, institution, data, or world being invoked. Those paths do not have to be perfect. They do have to exist and remain usable.

Often the practice is ordinary. Reread the passage. Check the quotation. Compare the summary to the source. Ask whether a metric measures the thing being claimed, or only a proxy for it. Distinguish memory from verification. Let disagreement correct performance. Admit when something is known through reputation, a review, classroom use, or AI synthesis rather than direct engagement.

Accountable mediation does not demand expertise on everything. That would be impossible. It asks for proportion. The more public, central, or consequential the claim, the stronger and more explicit the contact should be.

Bayard and Books as a Diagnostic Case

Bayard helps because books make the problem visible without making it new. His categories of unknown, skimmed, heard-about, and forgotten books describe common forms of partial literary contact. He also develops terms such as the screen book, inner book, phantom book, collective library, inner library, and virtual library.

The larger point is not the terminology by itself. Books are not only objects we read or fail to read. They are remembered, positioned, socially performed, and reconstructed. A book can function as an object in conversation, a private object shaped by memory, a shared cultural reference, or part of the social space where people discuss what they have and have not encountered.

For this project, Bayard's vocabulary is useful as a prompt for thinking about mediated contact. It is not a new model variable, and it is not proof of the Reality-Contact Resilience framework. It is also not a license to bluff. “I read it” should not carry the same force as “I know its reputation,” “I read the chapter on this question,” “I remember its argument imperfectly,” or “I have only seen summaries.” These are different kinds of contact. They can all be useful. They should not all be weighted the same.

If this essay speaks of an honest taxonomy of contact, that phrase is this project's adaptation of Bayard's diagnostic vocabulary, not Bayard's own framework.

Wittgenstein and Language Games

“I have read this” does not do the same work in every setting.

In a seminar, it may imply close engagement with the text. In a professional briefing, it may mean “I reviewed the relevant sections.” In a casual conversation, it may mean “I know enough to follow the reference.” In a literary review, it carries a heavier obligation. In an AI-generated summary, it may not mean that any human reader has encountered the underlying work at all.

This can be read in a Wittgensteinian way: meaning depends on use, context, audience, and responsibility. The point is an application, not a claim that Wittgenstein analyzed Bayard or book-talk in this form.²

The healthy practice is not to ban compression. Compression is necessary. Name it.

Rosa and Resonant Residue

Hartmut Rosa offers another useful angle. Books are not merely information containers. Read through Rosa's language of resonance, a book can be interpreted as a resonant object: something that changes what we notice, fear, admire, resist, or consider possible.

What this project might call “resonant residue” is a useful phrase for that after-effect. It should be used carefully. Resonance is not nostalgia or mere emotional reaction. It names a responsive relation to the world, a way something outside the self can answer, unsettle, or re-tune attention.³

That kind of contact is real, but it is not citation-level knowledge. Someone may be deeply formed by a book and still need to check the passage before quoting it.

AI Intensifies an Old Human Condition

AI did not invent mediated knowledge. People have always relied on summaries, reputations, inherited judgments, indexes, reviews, classrooms, briefings, and conversation. Human culture has never been a society of isolated direct encounters.

[Inference] AI appears to change the scale, speed, confidence, and friction of mediated knowledge. It can make orientation faster and more fluent. It can also make partial contact sound complete, or make a summary feel more authoritative than the contact behind it warrants.⁴

That does not make AI-mediated contact useless or automatically false. It can be helpful for orientation, triage, comparison, drafting, and translation. The discipline is to tag it honestly and return to sources when the claim matters. An AI summary is contact with a synthesis. It is not automatically contact with the underlying source, and stronger claims require source checking.

Synthetic Contact and Synthetic Drift

The project can make a modest distinction here. The useful contrast is not human versus machine, or direct versus mediated. It is synthetic contact versus synthetic drift. These are project terms, not Bayard's terms.

Synthetic contact is partial, mediated, compressed, or AI-assisted contact that remains answerable to a source, person, event, dataset, institution, or world. It may begin with a summary, but it preserves a path back.

Synthetic drift is fluent, self-reinforcing talk that loses corrective contact. Reputation replaces reading. Performance replaces understanding. Summary replaces source. Fluency replaces contact.

This matters for Reality-Contact Resilience because a resilient culture does not eliminate mediation. It disciplines mediation. It keeps summaries, institutions, models, and technologies answerable to what they claim to represent.

Contact-Type Tagging

One practical habit is to name the contact behind the claim. Claim-type tagging asks what kind of claim is being made. Contact-type tagging asks what kind of contact supports it.

This is not a compliance system. It is a small habit of discipline. Strong prose should not launder weak contact.

The tags do not need to appear in every published paragraph. Often they belong in drafts, notes, margin comments, or editorial passes. Their value is the habit: before a claim becomes polished, ask what kind of contact supports it.

Closing

A reality-contact society is not one without screens, summaries, institutions, models, or mediation. It is one that keeps paths of return open: back to the text, back to the body, back to the data, back to the place, back to the other person, back to the institution, back to the shared world.

The practical rule is simple: before leaning on a claim, ask what kind of contact supports it. Did I read the source? Did I read part of it? Do I know it through memory, reputation, summary, conversation, data, experience, or AI synthesis?

That question will not solve every epistemic problem. But it creates a small discipline against a large temptation: letting fluent language sound more grounded than it is.