The Dice Do Not Need Loading: How Technology Steers What We Believe

How technology steers meaning in an age of information overload

Roll two ordinary six-sided dice often enough and seven will appear more frequently than two. Nobody needs to cheat. There are simply more combinations that produce seven. Communication operates through a related principle, except that the dice remember being thrown. Every time a message circulates, it can change the conditions under which it will circulate again. Repetition creates familiarity, familiarity invites recognition, recognition produces responses, and responses create further opportunities for repetition. The probability of recurrence changes as the system develops. Some messages become easier to encounter, recognise and reproduce, while others disappear into the enormous reservoir of things nobody has time to consider. This is selection: the differential continuation of messages through available pathways. It is not a judgment of truth, intelligence or merit. A message can be selected because it is useful, because it is entertaining, because it is outrageous, or because an extraordinary number of people have decided to explain why it is complete rubbish. The system registers the activity. Congratulations. You have participated.

Meaning, meanwhile, is not a little parcel that travels intact from one head to another. Consider the sentence, “Everything is under control.” Spoken by a pilot during turbulence, a politician during a crisis or a friend standing in a smoke-filled kitchen, the words remain identical while the meaning changes considerably. Meaning emerges through relationships between language, circumstance, expectation, memory and response. Technology accelerates the circulation of messages while frequently separating them from the circumstances necessary to understand them. A sentence becomes a headline. A fragment becomes evidence of an entire position. An image travels between communities, acquiring implications its original context never contained. Information overload makes this worse because attention is finite. People, platforms, journalists, advertisers and political organisations must continually decide what to notice, publish, rank and repeat. These decisions interact. Human reactions inform technical ranking, technical visibility influences editorial judgment, coverage produces political responses, and those responses generate further public reaction. Nobody needs to design the overall result. Billions of locally intelligible decisions can produce a collective arrangement of breathtaking stupidity. We have built machinery capable of circulating more information than any civilisation in history, then made attention the scarce resource around which much of it competes.

The consequences become especially interesting when complexity encounters the cost of explanation. Housing pressure involves planning, taxation, wages, investment, infrastructure, migration, demographics and decades of interacting decisions. Explaining those relationships takes time, context and a willingness to tolerate uncertainty. “They caused this” takes three words. It supplies an explanation, an adversary and an implied solution, with the added convenience of requiring almost no understanding of the problem. This is compression: reducing a complicated situation to a smaller, more portable representation. All explanation requires compression, but what gets excluded does not stop operating merely because it has disappeared from the story. Some explanations are easier to transmit because they connect with existing concerns, reproduce familiar distinctions and invite predictable responses. Repeated circulation reinforces those connections. A complicated problem can consequently become publicly organised around a narrow explanation, not because the explanation has survived rigorous examination, but because it has survived circulation. The distinction matters. Popularity is evidence that something has become popular. It is not a particularly sophisticated epistemology, although an impressive amount of contemporary civilisation appears to be running on it.

Digital technology turns this process into feedback. A message circulates, generates activity, produces measurements, influences ranking, acquires visibility and circulates again. Each passage can leave behind a larger audience, a searchable phrase, a familiar association, an established identity or a record of behavioural response. Communication does not merely travel through existing pathways. Recurrence can construct and reinforce the pathways through which subsequent recurrence becomes more probable. Platforms do not need to distinguish agreement from disgust to register activity. A furious denunciation can introduce a marginal claim to an entirely new audience. A correction can repeat the very association it seeks to dismantle. Media organisations report the resulting controversy, political institutions respond, commentators interpret those responses, and the whole arrangement acquires another opportunity to continue. This does not make correction pointless. It means that establishing whether a statement is true and understanding the communicative pattern sustaining it are different tasks. Disinformation is only one part of the problem. An accurate statement can mislead through omission, implication, timing or repetition. A genuine grievance can be attached to an inadequate explanation. A real event can be circulated until it appears representative of an entire population. Fact-checking can establish what happened. It cannot, by itself, explain why a particular interpretation keeps returning, what relationships sustain it or why the surrounding controversy has become more durable than the original event. The machinery has become exceptionally proficient at measuring participation without establishing what that participation means. Apparently, a million people arguing about something is sufficient evidence that another million should be invited.

Populist communication demonstrates how readily these mechanisms can organise political meaning. Its characteristic division between a supposedly unified people and a corrupt elite converts a complicated field of interacting problems into a recognisable conflict. The betrayed citizen, distant official, threatening outsider and fearless defender become recurring positions within the narrative. Once established, these roles can be occupied by different people and attached to different events without changing the underlying arrangement. Housing, migration, taxation, crime, education and cultural change become additional opportunities to reproduce the same relationships. The grievances can be genuine. People experience economic insecurity, institutional distance and political frustration. The simplification gives that distributed pressure a portable explanation and a familiar adversary. Supporters affirm the identity offered to them, critics occupy the opposing position, and institutional responses can be incorporated as further evidence of the conflict. Even failure need not interrupt the pattern. It can be interpreted as proof that the opposition is more extensive than previously imagined. This structure is not exclusive to any party, nation or political direction. It emerges wherever moral certainty, permanent adversaries and group identity become more important than correction and learning. The problem is not that political communication simplifies reality. It must. The problem begins when a simplification becomes incapable of recognising what it excludes, and every contradiction becomes another reason to repeat it more forcefully. At that point, an explanation has acquired the remarkable ability to survive being wrong without undergoing the inconvenience of changing.

None of this requires a conspiracy, although it provides an extraordinarily accommodating environment for anyone interested in manipulation. Communication is stochastic: individual encounters remain uncertain, while recurring patterns emerge across large populations. The probability of an interpretation recurring depends upon the audience, timing, medium, previous circulation and available connections. Change those conditions and the probabilities change. Money can purchase repetition. Advertising creates additional routes of exposure. Media ownership influences selection. Targeting places different expressions before different audiences. Coordinated accounts can create apparent momentum, while established influencers connect unfamiliar claims to existing relationships of trust. Artificial intelligence expands the available machinery by generating thousands of variations, adapting language to different communities and reproducing narratives through apparently independent voices. Steering does not require determining what every individual believes. It means changing the distribution of opportunities through which particular messages, interpretations and responses become more probable. The larger system can remain emergent while particular actors deliberately exploit its mechanisms. Different participants pursue different immediate objectives, collectively producing patterns that none entirely controls. We have consequently developed communication technologies that can make an interpretation familiar before it has been examined, socially significant before it has been understood, and politically consequential before anyone has established whether it addresses the problem. The dice do not need loading, but the communication environment can certainly be engineered.

The practical response is not to retreat into the comforting fiction that independent judgment means standing outside influence. Nobody stands outside language, culture, history, technology or other people. Independence requires recognising the relationships through which an interpretation has become available and preserving the capacity to revise it. What is actually being asserted? What evidence supports it? Which causes have been made visible, and which have disappeared? What remains recognisable when the wording, speaker or controversy changes? Who benefits from additional circulation, and does opposition interrupt the pattern or complete it? Most importantly, does the proposed response address the conditions producing the problem, or does it simply provide another opportunity for the same conflict to recur? A democratic society does not require agreement. It requires people capable of disagreement without converting every difference into an enemy, recognising genuine grievances without accepting every explanation attached to them, and changing their understanding without treating correction as humiliation. Repetition is not independent confirmation. Engagement is not agreement. Familiarity is not understanding. The central danger is that communication systems can become increasingly effective at reproducing interpretations while becoming progressively less capable of correcting them. Every recurrence can strengthen recognition, generate responses, reinforce pathways and alter the probabilities of what happens next. The dice do not need loading. The pathways remember. Between encountering a message and responding to it remains an interval in which its assumptions can be examined, its omissions recognised and a different response chosen. That interval is not a technological inefficiency to be eliminated. It is where judgment remains possible.