I knew - but I did know that I had crossed
The border. Everything I loved was lost
But no aorta could report regret.
A sun of rubber was convulsed and set;
And blood-black nothingness began to spin
A system of cells interlinked within
Cells interlinked within cells interlinked
Within one stem. And dreadfully distinct
Against the dark, a tall white fountain played.From Pale Fire by Vladimir Nabokov
This essay can be read “stand-alone”. That said, it is the second essay in a series about the role of beliefs as signaling mechanisms in intra-elite status games.
Introduction
People learn what to desire by observing others. Through observation, we learn both what to want and the ways of being that will help us to get what we want. We end up desiring both the outcome and the means to achieve it. Our desires become bundles of means and ends, transcending the simple object fixations of infancy.
Signals, Hierarchies, and Social Order
The underlying qualities that we come to admire through this process: competence, intelligence, networks, and discipline, can never be perfectly perceived. Instead, they are inferred through legible traits: patterns of behavior, appearance, possessions, affiliations, and other cues that can be interpreted within given social contexts.
These legible traits function as signals. These signals form the basis of hierarchies, which in turn give rise to the complex social systems that underpin our civilization.
Over time, this signaling process organizes behavior from small groups to whole civilizations, structuring how coordination occurs.
In high-functioning societies, social signals remain at least somewhat tethered to the underlying capacities they are meant to represent: competence, resources, trustworthiness, and the ability to contribute to shared endeavors.
This relationship is never fixed. It is always dynamic. It can and does drift over time.
When Signals Detach from Reality
In a world where signaling is pervasive, it becomes disconcertingly easy for entire systems of signals to become self-referential.
As Jean Baudrillard argued, signs can detach from reality and begin referring only to other signs: forming closed systems that no longer point to anything outside of themselves.
Signals propagate when they are recognized, rewarded, and replicated, which is not necessarily the same thing as being true.
As many users of Large Language Models have learned, signals do not necessarily require accuracy to propagate. The user must enforce accuracy as a requirement on the system; it will not do so automatically. It cannot.
Here lies a critical point: reality itself is not socially constructed, but the demand for accuracy in our models of reality is.
Different individuals, societies, and historical moments place varying emphasis on whether their symbolic systems remain grounded in the realities they claim to represent.
Over time, signals can inflate, much like currencies. They can lose contact with the underlying conditions that made them meaningful in the first place.
When that divergence grows too large, a correction may eventually come. But these corrections are slow, uneven, and often arrive only after prolonged distortion.
These processes play out over long horizons.
As I argued in a past essay, in our current cultural moment, there is little doubt that we have entered a highly self-referential symbolic order. Increasingly, reality is treated as something that can be constructed mostly through language and belief. The past is reframed not as a set of constraints to be understood, but as an embarrassment or inconvenience to be overwritten.
The Truth About Truth
When signaling systems become self-referential, the consequences do not remain confined to status or culture. They begin to affect how truth itself is understood.
Truth is neither absolute nor arbitrary, and confusion on this point has driven significant civilizational drift in the past few decades.
Across modern science, especially in physics, a consistent insight has emerged: what we call truth depends on the relationship between the observer, the observed, and the conditions under which observation occurs.1
This does not mean that truth is optional. It means that truth is constrained.
It also means that truth cannot be fully captured by abstract, universal forms, as Plato proposed. Truth exists at different degrees of specificity, at different points in space and time.
This does not render truth meaningless as a concept.
The common misreading of the observer-dependent nature of truth is to treat it as a license for total subjectivity: the idea that all claims are equally valid. This is the foundation of ideas like “my truth,” where assertions about reality are grounded primarily in perspective rather than in shared methods of verification.
Taken literally, the notion of “my truth” collapses the shared distinction between true and false.
The notion that truth is arbitrarily and socially constructed is easier to sustain in domains that are already abstract and self-referential. The persistence of empirically false beliefs, such as flat-earth theories, illustrates the consequences of this drift.
Scientific practice, especially in the “hard” sciences, with their rigorous methods and traditions for testing reality, cannot operate this way. They rely on constraints.
Rather than treating truth as absolute, science treats it as a bounded approximation. Models of reality hold under specific conditions, and at particular levels of abstraction.
Boyle’s law, for example, describes the relationship between pressure and volume in a gas. It works at the macro level, where the complexity of individual molecules resolves into stable patterns. It cannot predict the motion of a single molecule, nor does it need to: it is a useful and stable guide to reality at the macro scale.
Truth, in this sense, is not arbitrary. But neither is it universal in scope. It is always tied to the level at which a system becomes legible.
The universe may be irreducible at its most fundamental levels, but stable patterns do emerge. These patterns are what we call rules. These rules not absolute, they are not the source code of the universe as we once thought, but they are reliable within the conditions in which we operate.
That’s not nothing.
All models are wrong, but some are useful.
George Box, 1976
I wrote about this conflict in more depth recently. The difficulty is that we no longer agree on how to distinguish between useful models and invalid ones. This has produced a fragmentation across intellectual traditions, from postmodernism to technocracy to various forms of traditionalism, each operating with different assumptions about what constitutes truth.
A Practical Example
These dynamics are not confined to theory. They are visible in systems that rely heavily on signals to function.
Modern labor markets provide a clear example. Credentials, experience, and institutional affiliations are all signals meant to represent underlying qualities such as intelligence, competence, judgment, and the ability to create value. In principle, they allow organizations to make decisions under conditions of uncertainty, using proxies to infer capability.
Over time, however, these signals become targets in their own right.
Degrees are pursued not only for the knowledge they provide, but for the status they confer. Career paths are shaped to maximize legibility rather than effectiveness. Individuals learn to present themselves in ways that align with institutional expectations, optimizing for selection within the system rather than for performance outside of it.
There is some evidence that this shift carries real consequences. Firms led by highly credentialed managers, particularly those trained in financial and managerial disciplines, have been shown in many contexts to favor short-term optimization over long-term value creation. In parallel, functions such as marketing have become increasingly dominated by metrics and performance frameworks that are legible within the system, but not always well aligned with the underlying drivers of demand.2
The pattern is not universal, but it is suggestive. As signals become targets for imitators, the behaviors they were meant to represent can begin to erode.
As more people learn to play this game, the signals themselves begin to degrade. They become easier to acquire, more widely distributed, and less informative. What was once a useful proxy for capability becomes a noisy indicator and, in some cases, a misleading one.
I discussed this in more detail in the previous essay in this series.
The result is not simply inefficiency. It is a breakdown in coordination.
Organizations struggle to identify genuine talent. Individuals struggle to distinguish meaningful paths from performative ones. The system continues to function, but with increasing friction, as more effort is required to interpret signals that no longer reliably point to the underlying reality.
This is what it looks like when a signaling system begins to lose its anchor.
When Coordination Breaks
What does this mean for signals and their role in human society?
As signals circulate, they increasingly refer to one another.
This tendency is not accidental. Signals propagate through recognition and reinforcement, not through direct verification against underlying reality. As they spread, individuals respond to the signals of others rather than to the conditions those signals were originally meant to represent.
Under conditions of intensified elite competition, this process accelerates. As more individuals compete for status within the same symbolic space, the pressure within those spaces shifts from accurately reflecting reality to producing signals that are legible, repeatable, and rewarded within the system itself. Speed, visibility, and recognizability begin to dominate over accuracy.
As I have previously written, those conditions are being met now, which means we can expect social networks and hierarchies to become increasingly self-referential.
Over time, this creates a self-reinforcing loop in which signals are calibrated against other signals rather than against the underlying conditions of the world. The system begins to close. Meaning is no longer anchored in material conditions, but in the internal logic of the signaling system itself.
This is creates a practical problem for human societies.
When signals drift too far from reality, they stop coordinating behavior effectively. They no longer help individuals identify competence, trustworthiness, or capacity. Instead, they become tools for navigating status within a system that is increasingly detached from the conditions that sustain it.
At that point, the system does not simply mislead. It begins to fail.
We will begin at this point in the next essay in this series.
I always appreciate reader comments, questions, and observations. I try to reply to all of them.
Related Content:
Signals Without Anchors
This tension has long been explored in science fiction.
In The Last Question, Asimov traces the pursuit of a final, complete answer across cosmic timescales, only to suggest that such an answer may lie beyond the horizon of knowability.
In The Hitchhiker’s Guide to the Galaxy, Adams treats the idea of an ultimate explanation for “life, the universe and everything” as inherently absurd, offering “42” as a deliberately meaningless answer to an ill-posed question.
Both works arrive at a similar insight. There is no view from nowhere.
What matters, then, is not the elimination of perspective, but models that respect the constraints imposed on us by the universe.
Computational Irreducibility and the New Science
Wolfram’s “New Science” contains difficult but important concepts (especially computational irreducibility) that I alluded to in this essay and discussed in the essay below; I will revisit this topic in future essays.
![What the Cat Knows [and we don't]](https://substackcdn.com/image/fetch/$s_!spj3!,w_140,h_140,c_fill,f_webp,q_auto:good,fl_progressive:steep,g_auto/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F07a1f5f1-e662-4670-8138-e961f16320ec_1024x1024.png)
See Marianne Bertrand and Antoinette Schoar, “Managing with Style,” Quarterly Journal of Economics (2003); Neil Bendle and Charan Bagga, “The Metrics That Marketers Muddle,” MIT Sloan Management Review (2016); and Jerry Z. Muller, The Tyranny of Metrics (2018), for evidence that managerial incentives and metric-driven systems can drift from underlying value creation.