There is a version of political communication that seeks to persuade. And then there is a version that is not really about persuasion at all — that is, instead, about activation. About keeping a body of people in a heightened emotional state, alert to threat, certain of their tribe, and reliably present at the algorithm's next serving. The two look superficially similar. Both involve words, posted publicly, about consequential things. But they operate according to entirely different psychological logics, and the results they produce could not be more different.
Donald Trump's social media output — across Truth Social, X, and the broader media ecosystem that amplifies it — provides the most extensively documented example of activation-oriented communication in the digital era. The posts are the primary case study here. Not because this analysis is a verdict on Trump as a political figure — it is not — but because no other communicator at comparable scale has generated as much observable data about what specific rhetorical choices produce which specific audience responses. The volume and consistency of the output make it analytically legible in a way that more cautious political communication simply is not.
Understanding what that output actually does requires setting aside the assumption that it is simply the product of an undisciplined temperament. The evidence points elsewhere. What looks like chaos frequently turns out to be a system — one that is sophisticated, internally coherent, and demonstrably effective at the specific outcomes it appears designed to achieve. This piece maps that system, examining the four interlocking psychological mechanisms that together constitute what we are calling the Polarisation Engine. And it holds open, honestly, the question that makes the analysis genuinely interesting: is this architecture the product of deliberate design, or of instinct so refined it has arrived at the same place?
01Disintermediation & In-Group Identity
The first and foundational move is the bypass. Traditional political communication operated through a filter — a press conference was mediated by journalists, a speech was clipped and contextualised by editors. That mediation served, among other things, as a temperature regulator. It was difficult to sustain a state of perpetual crisis when access to the audience ran through people whose professional norms inclined them toward balance and a news cycle that demanded novelty.
Direct-to-audience platforms dissolved that filter. The political figure could now speak to millions without editorial intervention, setting tone, choosing words, and — crucially — selecting which conflicts to surface. This is disintermediation. But the strategic value of disintermediation goes beyond mere convenience. Its deeper function is to reshape the media environment itself.
When a post is sufficiently inflammatory, the out-group — established media, political opponents, institutional critics — cannot ignore it. They must respond. And in responding, they perform precisely the role that Social Identity Theory would predict: they become the external threat that consolidates the in-group. Henri Tajfel's foundational research on group psychology showed that inter-group hostility reliably strengthens intra-group solidarity. Every critical editorial, every outraged cable segment, every fact-check article becomes confirmation to the base that their identity is under attack. The very act of rebuttal functions as mobilisation.
Every polarising post is a loyalty test — political support transforms into a deeply personal identity, one that feels like a self that must be defended rather than a preference that might be revised.GlamBon Culture Editorial
The logical endpoint of this mechanism is that the political figure's supporters do not merely hold certain views. They come to feel that they are those views — that criticism of the movement is criticism of the self. At this point, conventional persuasion becomes almost irrelevant. The in-group is no longer reachable by argument, because argument has been reframed as attack.
02Emotional Dominance & Affective Polarisation
The second mechanism operates at the level of emotion selection. Research into viral content — work by Jonah Berger, Matthew Salganik, and the teams behind the 2018 Science study on misinformation spread — converges on a consistent finding: not all emotions are equally shareable. High-arousal emotions, particularly anger and anxiety, propagate dramatically faster through social networks than their low-arousal counterparts — satisfaction, sadness, contentment.
This creates a straightforward strategic imperative: if maximising reach and engagement is the goal, the optimal emotional register is anger. Not necessarily pure rage — that tends to alienate moderate supporters — but the particular species of righteous indignation that feels productive, that gives its bearer a sense of clarity and agency. A threat that has a villain and implies a necessary response.
The rhetorical techniques that reliably produce this state have been well catalogued. Superlatives do the work efficiently: 'total disaster,' 'the greatest ever,' 'complete fraud.' They flatten nuance — a complex policy position becomes a binary moral choice — while simultaneously signalling epistemic confidence. There is no ambivalence in this register, and ambivalence, as it happens, is not particularly shareable. The audience is offered certainty in exchange for complexity, and the exchange tends to be accepted.
The cumulative effect across a sustained body of communication is what researchers call affective polarisation: a deepening emotional distance between political groups that operates independently of policy disagreement. People on opposite sides of the affective divide do not merely disagree. They find each other increasingly unintelligible, increasingly threatening, increasingly easy to caricature. This state of affairs is, from an engagement standpoint, extraordinarily durable. It does not require new arguments. It only requires maintenance.
03The Performance Strategy
The third mechanism is the one most frequently misread as evidence of disorder: the use of humour, mockery, and what internet culture has come to call trolling. The temptation is to interpret this mode as unprofessional, unserious — the rhetorical equivalent of a playground taunt. This is a category error.
Mockery performs a precise argumentative function: it de-legitimises the opponent without engaging their actual position. If a critic raises a substantive concern and receives a nickname in response, they face an impossible rhetorical situation. If they respond to the nickname with a policy argument, they appear to be missing the joke — po-faced, elitist, incapable of reading the room. If they respond in kind, they have abandoned the substance of their critique. And if they do nothing, the nickname propagates.
The conventional response to mockery — careful, evidence-laden rebuttal — is precisely the wrong tool for the context. It is the response of someone who believes the game being played is a debate, when the game is actually about performance. The audience being addressed is not persuadable swing voters weighing arguments; it is a consolidated base being entertained.
While the media spends 48 hours anatomising an apparently absurd post, the policy shifts and legal manoeuvres happening in the background receive proportionally less scrutiny. The performance is also a screen.GlamBon Culture Editorial
The distraction function of this mechanism is perhaps its most underappreciated dimension. A post that generates sustained media outrage is a post that is consuming editorial bandwidth that would otherwise be allocated to accountability journalism. The ratio is not coincidental. It is, in the most literal sense, engineered.
04The Voice as Product
Here is the question that the previous three pillars quietly raise: who, exactly, is doing this?
Trump's posts have a recognisable linguistic fingerprint. The capitalisation functions as a kind of punctuation substitute — TOTAL DISASTER does the work that italics and an exclamation mark would do in edited prose, but lands differently. It feels typed, not composed. The superlatives are dense and unvarying: everything is the greatest, the worst, the most beautiful, the most corrupt. The syntax is loose in a specific way — clauses pile up, qualifications are absent, the grammatical looseness signals speed rather than care. Combined, these features produce writing that feels like it arrived directly from a mind to a screen without the intervention of an editor, a comms team, or a moment of reflection.
That feeling of directness is the product's core value proposition. The audience is being offered something that political communication almost never delivers: apparent unmediated access to what a powerful person actually thinks, in real time. It is an extraordinarily compelling offer. And it raises an immediate craft question: if a team of communications professionals were tasked with producing content that created this impression, what would they have to do?
The answer is that they would have to actively suppress their instincts. Professional communicators are trained to qualify, to hedge, to consider the secondary audiences for any public statement. They think about legal exposure, about the news cycle, about stakeholder reactions. All of that training produces writing that is measurably more careful — and measurably less viral. To write in the Trump register, a professional writer would need to deliberately de-skill: to drop the qualifications, flatten the complexity, and replicate not just the vocabulary but the specific texture of unfiltered thought. They would need to write badly in a very particular way, consistently and at scale.
The posts feel like the absence of a filter. But the absence of a filter is itself a filter — one that has to be maintained, replicated, and kept consistent across thousands of posts and years of output. That kind of consistency is not accidental.GlamBon Culture Editorial
Which brings us to the genuinely open question. Former White House staffers, journalists who have covered Trump for decades, and communications researchers who have analysed his output extensively do not agree on the answer. Some maintain that the posts are his — that the voice is authentic because it is the actual voice of an actual person who communicates this way. Others point to patterns of consistency and strategic timing that seem difficult to explain as pure impulsiveness. The truth may be somewhere between: a natural voice that has been refined and amplified rather than fabricated from scratch, shaped over time by feedback about what lands and what doesn't.
What is clear is that the voice, however it is produced, functions as a brand asset of extraordinary power. Billions of dollars in media attention have been generated by it. It has made the informal register of a social media post feel more authoritative, to its intended audience, than a prepared statement read from a teleprompter. The soup of words — the capitalisation, the nicknames, the non-sequiturs, the apparent errors — is the authenticity signal. Whether it is engineered or instinctive, it works in exactly the same way.
05The Permanent Campaign
The fourth mechanism is architectural rather than rhetorical. It concerns not the content of individual communications but the temporal structure of the communication strategy as a whole. Traditional political communication oscillated between modes — campaign mode, governing mode, crisis mode. The Polarisation Engine abolishes this rhythm.
By maintaining a constant level of perceived threat and urgency, the permanent campaign prevents something that is, from a mobilisation standpoint, genuinely dangerous: supporter fatigue. A base that feels safe will eventually disengage. Its members will start to think critically, to ask whether the promised outcomes have materialised, to evaluate performance against expectations. The solution to this risk is elegantly simple: never let the base feel safe.
Crisis framing — the persistent presentation of the world as a binary struggle between an in-group and its existential opponents — keeps supporters in a state of activation that is incompatible with the dispassionate evaluation of evidence. When every news cycle is the most important news cycle, disengagement feels like desertion. Loyalty is not a preference but a duty.
For a team producing this communication, the operational implication is continuous optimisation. Digital publishing tools allow real-time measurement of which grievances generate the most engagement, which framings produce the most sharing, which targets elicit the most intense responses. This data, used iteratively, produces a communication output that is progressively more attuned to the emotional architecture of the base — and progressively further from the rhetorical norms that the out-group would recognise as legitimate.
Engagement Heat Map
How each pillar of the Polarisation Engine maps to distinct emotional outcomes. Darker cells indicate stronger research-evidenced effect. Hover any cell for analysis.
—What the Engine Tells Us
The point of mapping these mechanisms is not to arrive at a verdict on the individuals who deploy them. It is to understand the environment they create — and the degree to which that environment is engineered rather than organic. The apparent chaos of contemporary political communication is, in significant measure, a designed output. The unpredictability is a feature; the outrage is a product; the algorithmic attention is the revenue.
Understanding this does not make the audience immune to the mechanisms. Social identity processes are not voluntarily suspended because one has read a description of them. But it does alter the terms of public conversation. Media organisations that treat each provocative post as breaking news are, by doing so, functioning as amplification infrastructure for a strategy that depends on exactly that response. Citizens who feel their political identity is under existential attack are, by feeling that, experiencing a psychologically induced state that was optimised into being by a team with access to engagement analytics.
None of this is entirely new. Political communication has always involved the strategic selection of enemies, the deployment of emotion, the cultivation of tribe. What is new is the feedback loop — the real-time data pipeline that allows the machine to learn, iterate, and optimise faster than any previous political communication apparatus. The Polarisation Engine is not a static design. It is a self-improving one.
The honest answer is that we do not know with certainty, and that uncertainty is itself analytically interesting. Former staffers and journalists who have covered Trump extensively do not agree. What is clear is that the posts carry a highly consistent linguistic fingerprint — the capitalisation patterns, superlative density, loose syntax, and specific nickname conventions — that has remained stable across years and platforms. Whether that consistency is the product of one person's authentic voice, a team trained to replicate it, or some combination of the two, the communication architecture functions identically in all three cases. The authenticity signal — the feeling that this is unmediated access to a real person's real thoughts — is the strategic asset. How it is produced is, from a psychological-effects standpoint, secondary.
Affective polarisation refers to the emotional distance between political in-groups and out-groups that develops independently of policy disagreement. When rhetoric consistently triggers high-arousal emotions such as anger or fear, audiences begin to identify with their political tribe on a deeply personal level. Research shows that opposing partisan groups increasingly view one another not merely as wrong but as morally threatening — a state of affairs that is extraordinarily resistant to conventional persuasion.
Platform recommendation algorithms are tuned to maximise time-on-platform, and high-arousal content — anger, outrage, fear — consistently outperforms measured analysis by every engagement metric. This creates a feedback loop: divisive posts receive wider distribution, receive more reactive attention, which generates more engagement signals, which further boosts distribution. The Polarisation Engine does not create this dynamic; it exploits one that the platforms have built and continue to maintain.
Disintermediation is the strategy of communicating directly with a base audience, bypassing traditional media filters entirely. By doing so, a political communicator controls framing, tone, and the selection of which conflicts to surface. Critically, when established media react to direct posts — as they structurally must — they perform the role of out-group attacker, which Social Identity Theory predicts will consolidate in-group loyalty. The critic is, paradoxically, doing the communicator's work for them.
Mockery de-legitimises opponents without requiring engagement with their substantive arguments. A serious rebuttal to a meme-style attack places the rebutting party in an impossible rhetorical position: respond with evidence and appear humourless and elitist; respond in kind and abandon the substance; say nothing and allow the framing to propagate. The performance also functions as a distraction mechanism — extended media analysis of an apparently absurd post consumes editorial bandwidth that would otherwise flow toward accountability journalism.