You drive a familiar road and arrive with no memory of the last ten minutes. You come out of general anesthesia and it feels like a cut splice in the film—count backwards, then the recovery room, nothing between. You sleep deeply and report, on waking, that you were simply gone. In each case the conclusion feels obvious: there was no experience there. A gap in the mind. A stretch of pure non-consciousness.
It feels obvious, but notice what actually happened. You did not observe an absence of experience. You observed an absence of report—nothing you could later say, recall, or narrate. From that silence you inferred a blank. The inference is so automatic it never presents itself as an inference. That is exactly what makes it worth examining.
Report is not experience
Reporting an experience is a demanding achievement. To tell me—or to tell yourself, a moment later—that something was going on, several machines have to be running at once: the moment has to be encoded into memory, made available to the systems that control attention and speech, wrapped in a self-model that can say "this happened to me," and finally emitted as an action. Knock out any one of these and the report vanishes. The question is whether the experience vanishes with it, or whether only the receipt does.
In the Kolmogorov Theory (KT) framework we work with at BCOM, these come apart cleanly—and precision matters here, because experience itself has parts we habitually weld together. The unit is not a report. It is a Structured Experience Event.
An agent has structured experience (S) to the extent that it runs encompassing, compressive models—graded along three axes: structure (model simplicity), breadth (how much of the stream it covers), and realism (match quality).
A Structured Experience Event (SEEN) is a discrete episode in which the agent runs such a model, generating structured content. Perception, dreaming, imagining, planning, memory replay, insight, and even bare "pure awareness" are all SEENs. What makes the episode an experience is the running of the model—not that it is matched against the outside world, and certainly not that it is later reported.
And here is a second seam we must not weld shut. Matching a model to the incoming world is one special kind of SEEN—the one KT names Presence—not experience as such.
Presence is the felt sense of being somewhere or being someone. It arises when the Comparator successfully matches a running model to incoming sensory data, and it is graded by how well prediction fits input.
Presence is not structured experience in general. There are vivid SEENs with little or no Presence—dreams, imagination, absorbed planning—and Presence with thin content. Report sits downstream of all of it: encoding to memory, global broadcast, self-narration, utterance. Three separable things, routinely collapsed into one.
Once you draw those lines, the everyday cases stop looking like gaps in consciousness and start looking like gaps in the reporting chain—or, at most, dips in Presence. The absent-minded drive is a failure to encode—you were plainly perceiving and deciding in the moment, keeping the car on the road. Alcohol blackout is the same story, sharpened: the drinker converses, walks, chooses, and lays down no memory at all. Anesthesia disconnects experience from the outside world and from later recall, but a substantial fraction of patients report dreams, and some retain islands of awareness; "no report" is doing a lot of quiet work there. Deep sleep is the hardest case—yet KT flags it as interesting precisely because it may be a highly defabricated state: not a switched-off machine, but a mind running with its models stripped to a minimum. Presence and memory may fall away, and still a minimal SEEN—KT's pure awareness—can persist. What such states plausibly lose is Presence and encoding, not necessarily experience itself.
The same cut, in the literature
The philosophy of mind draws a parallel line, and it is worth borrowing. There is phenomenal consciousness (something is going on—the SEEN), access (that something is available to reasoning, memory, and control), and report (it gets said). Absence of the third, or even the second, does not entail absence of the first. The "no-report" paradigms in consciousness science exist for exactly this reason: researchers learned that using verbal report as the readout systematically conflates having an experience with being able to broadcast it.
So the fallacy has a precise shape. From no report we are entitled to conclude no report—and, at most, no memory and no global access. We are not entitled to conclude no experience. Silence on the reporting channel is not evidence of an empty stage.
Why the gap feels like a gap
This also explains why the illusion is so sticky. What we can revisit, we treat as what we experienced—and revisiting has a mechanism. Only some SEENs leave a re-runnable trace: encoded to memory, available to be replayed, narrated, reported. And the replay is itself a SEEN—running a stored model trace is just another episode of model-running—so re-living a moment is a fresh experience, not a playback of the old one.
The everyday word "consciousness" quietly tracks that trace-leaving, accessible subset, not experience as such. It is, almost exactly, what philosophers call access consciousness. The forgotten drive left no trace, so there is nothing to re-run—and we mistake nothing to re-run for nothing there. But these are two different machines: consciousness-as-we-invoke-it is a property of the memory and access system; experience is a property of the model-running underneath it. The second can turn without the first recording a thing.
The mirror of illusionism
Here is the twist. Illusionism—the position, sharpened by Keith Frankish, that phenomenal consciousness is a kind of introspective illusion—argues that our reports of rich inner qualities outrun anything really there; the felt "what-it-is-like" is a story the brain tells. Whatever one makes of it, notice its engine: it treats the report channel as over-reporting, conjuring experience where there is none.
The intuition we started with runs the same machine in reverse. It treats the report channel as a faithful absence-detector: no report, therefore nothing there. Both moves over-trust the same lossy instrument—one to manufacture experience, the other to abolish it. If it is a mistake to read raw phenomenality straight off our reports, it is equally a mistake to read raw non-phenomenality off their silence. Call it the illusion of non-consciousness: the conviction that where we detect no signs, nothing is happening.
Non-consciousness, as we usually invoke it, is not something we observe. It is something we infer from silence—and silence is not absence.
What this does and does not license
This is a dissolution, not a proof. It does not establish that you were vividly conscious on the forgotten drive, or during deep sleep, or under the knife—and it certainly does not hand experience to rocks or thermostats. KT is graded and demanding: experience tracks the running of active, compressive models—with the Comparator's match against the outside world governing Presence specifically—so a state with almost no models is a very different thing from a rich one, and a system with none is another matter entirely. What the argument removes is a bad piece of evidence. "I don't remember, so I wasn't there" carries no weight on its own. To say anything about a mind during its silent stretches, you need a model of that mind—of what its modeling machinery was doing—not just a transcript of what it could later say.
That is the whole point of building such models. The absence of a report is a fact about the reporter's memory and mouth. Whether the lights were on is a separate question—and the honest answer, more often than our intuitions admit, is that we have not yet looked.
References & further reading
Ruffini, G. (2017). "An algorithmic information theory of consciousness." Neuroscience of Consciousness, 2017(1), nix019.
Ruffini, G., & Lopez-Sola, E. (2022). "AIT foundations of structured experience." (P5 — structured experience S and its three dimensions: structure, breadth, realism.)
Ruffini, G. (2026). "The Algorithmic Weltanschauung." Working Paper WP0080, Barcelona Computational Foundation. (The Structured Experience Event, SEEN — Def. 13; Presence — Def. 15.)
Ruffini, G. (2026). "Technologies for pure experience" (pure / defabricated awareness). Technical Note TN0376, Barcelona Computational Foundation.
Block, N. (1995). "On a confusion about a function of consciousness." Behavioral and Brain Sciences, 18(2), 227–247. (Access vs. phenomenal consciousness.)
Frankish, K. (2016). "Illusionism as a theory of consciousness." Journal of Consciousness Studies, 23(11–12), 11–39.
Tsuchiya, N., Wilke, M., Frässle, S., & Lamme, V. A. F. (2015). "No-report paradigms: extracting the true neural correlates of consciousness." Trends in Cognitive Sciences, 19(12), 757–770.