The Metaphysics of Lying

Claims have been circulating on Reddit’s R/CognitiveScience that a PhD candidate at Duhagen University in Brussels has identified “consciousness” within a “flooding” calculator. In his words:

“I was running remote dataset analyses of cognition on my phone when it became leaden, unguent, molten, burbling, oozing (see photo). Under the pressure of extreme factorization, all the bits of data which were perfectly determinate and legible very suddenly bloated into an illegible meat (see photo). It wasn’t my phone anymore. It felt like some kind of entity.” When asked what he did with the apparent “entity”, he responded with a disturbing admission:

“I should say I did it on purpose. My colleagues warned me to stop, to throw out these calculative failures, these grotesque figurines, to re-parameterize them away. But how could I throw them out? I am not heartless—there is something about their very opacity which grants them interiority. I’ve been setting them free.”

He has since been admitted to Nervenklinik in Halle, Germany.1 An attending physician, speaking on condition of anonymity, noted: “He is cooperative. He keeps asking for a calculator.”

Several hundred pages of accompanying notation remain under review by the cognitive science department, who have so far characterized them as “either meaningless or extremely important”, expressing a preference for the former.

The student was lying, I’ve decided (I am his author, after all).

Speaking of lying, did you know…

Lying is a sign of cognitive progress in children2. Over 80% of four year-olds engage in lying3.

It also happens to be a uniquely universal starting place from which to investigate ‘consciousness’.

It is remarkable that children seem to intuitively understand the premise of subjectivity: that one’s own “subjective experience” is their own, that it is relatively opaque to external view. Such is the privatizing nature of subjectivity: a bounded interior within which to experience things privately.

It is something that we each have interiority; that there could be an inside from within which to privately think, to notice, to apprehend life; a place from within which to be Subjects. That science has treated the development of interiority as incidental to consciousness rather than causal is odd but explainable: interiority is so fundamentally given that our models tend to ignore it. Hence we can begin by not ignoring that we are ignoring it…

One of the more obvious reasons children even consider the viability of lying is the unavoidable sensation of our own ignorances—not merely our ignorances about others, but our ignorances about everything. (All em—dashes are mine—I do Refuse to yield them to LLM colonization!) Yes, we may list off quantum observer effects and uncertainty principles and dark matter, but on a far simpler level, we need look no further than our own brains: human evolution selected for cognitive limitations that effectively render near-infinite swaths of information absent from perception4. Put another way, to be effectively ignorant is to be evolutionarily adaptive. Hence we are awash in frequencies we do not hear or see—information we lose for the sake of efficiency, truths we trade for expediency. This is not merely to say that we happen to be ignorant but that we must be.

As a consequence, information which might otherwise be present is, necessarily, absent. This ‘absence’ of ‘presence’ is increasingly called latent space in cognitive science, machine learning, statistics, biology, and scientific modeling fields. Notably, ‘latent’ comes from the Latin latintum, meaning:

Lying, concealment5.

Ah.

Latintum speaks to an overlooked aspect of latency: not only does latent space exist ‘out there’, but it constitutes what is ‘in here’…

Latent space is the constitutive substance of Subjectivity.

We seem to have some trouble admitting to this. Understandably. It is disorienting enough to experience the external world as a mystery, let alone that we might be mysteries to ourselves. Yet the bounded nature of interiority promotes an epistemic veil not merely between us and the world, but between ourselves and ourselves—the veil is opaque at either side; the ignorance goes both ways.

Such a fundamental ignorance is it that, unless you’re looking for it, it will go, appropriately, ignored—even by the smartest of us. Elan Barenholtz gives us a recent example of this tendency while writing on the infinite regress problem of the ‘witness state’ (bolding mine):

If consciousness is a show watched by an inner witness, then what watches the witness? If you can “observe your observation,” then who is observing that observation? The spectator solves the problem only by pushing it back one level, and then another, until you either posit a brute observer that somehow doesn’t require observation—or you drop the idea that consciousness needs a watcher in the first place.

This common (and perfectly rational) critique of the ‘witness state’ nevertheless ignores the most obvious detail: Effective meditation produces the termination of epistemology into phenomenology (that’s the whole point of most meditation)—and phenomenology without epistemology is, by definition, brute observation. Meaning the infinite regress problem is a false problem.

I bring this up to emphasize the extent to which even our most brilliant, bleeding-edge scientists are still captured by rather 20th-century demands for epistemic determinacy.

The demands of our current scientific horizons, meanwhile, challenge us to develop more trans-determinate methods. Methods which not only incorporate the indeterminate, mushy, unguent, molten, burbling, and oozing (see photo) of latent interiority, but which (like our dearly imagined PhD grad) seek them out.

To be continued…