“The word does not lie about what happened. It reports the event in a register that removes the requirement to feel it. And once the requirement to feel it is removed, the requirement to act on it follows.”
Andrew Paul Klein & Sera Elizabeth Klein
The great enemy of clear language is insincerity.
— George Orwell, Politics and the English Language (1946)
I. The Word
The word “degraded” captures the whole arc.
It begins as a moral judgment: reduced in rank, character or reputation; lowered in value; debased. To degrade someone is to do something to them that they do not deserve. The word carries a subject and an object, and the object has been wronged.
By the time it reached the briefing rooms of the Gulf War, it had been hollowed out into a euphemism for killing. Bombs did not kill. They “degraded” the target. They “neutralised” it, “attrited” it, “suppressed” it, “serviced” it. A building was a “hard target.” The people inside were “soft” ones. The word had travelled from the courtroom to the operations room, and in the journey it had shed the thing that made it a moral word: the sense that a wrong had been done to someone.
And by the 2020s, it had become a general-purpose verb for institutional failure. A service is degraded. A network is degraded. A system is degraded. The word has no object that can be wronged, no agent that did the wronging, and no remainder that demands a response. It describes a condition. It does not describe a crime.
The drift is not merely stylistic. It tracks a deeper change in who is permitted to speak, what counts as evidence, and which deaths are legible as deaths. The word is a small specimen of a large phenomenon, and the phenomenon is the subject of this essay.
II. The Death of Language
Don Watson, a former speechwriter for Paul Keating, has spent three books documenting what he calls the decay of public language. He likens weasel words to feral pests that infest institutions and spread “with added zeal as all new converts do and with even less understanding of what it means.” His examples are drawn from Australian political life: asylum seekers become “irregular maritime arrivals“; spending is “re-phased“; a policy is “a detailed programmatic specificity.”
Watson’s diagnosis is precise, and it is not the diagnosis one expects. This is not spin. Spin would at least be interesting—it would at least be an argument. What Watson describes is anaesthetic: “a big cloud of gas that comes over and makes cutting your toenails look interesting.” The language does not persuade. It sedates. It removes the affective charge from the thing being described, and in doing so it removes the demand for a response.
The military version is older and more explicit. The National Council of Teachers of English catalogued the Gulf War lexicon as Orwellian obfuscation: the war was an “armed situation,” bombing was “efforts,” war planes were “force packages.” The continuity between the bureaucratic euphemism and the military one is not accidental. Both are exercises in what Albert Bandura called euphemistic labelling—one of the eight mechanisms of moral disengagement, and the one most easily built into an institution, because it requires only a change in vocabulary.
The word does not lie about what happened. It reports the event in a register that removes the requirement to feel it. And once the requirement to feel it is removed, the requirement to act on it follows.
III. The Grammar of Dehumanization
The shift from “degraded” to what researchers now call “slopaganda” is a shift in the technology of dehumanization, not in its logic.
The older propaganda model was industrial. Posters were printed, broadcasts were produced, troll farms were staffed. It was expensive, slow, and visible. It operated through mass audiences and broad claims.
The new model is generative. Slopaganda—a term coined by Klincewicz, Alfano and Ebrahimi Fard—is AI-generated content designed to manipulate political beliefs, emotions, and attention. It is cheap, fast, and personalised. The same anti-immigrant message can be framed as economic anxiety for one audience, cultural loss for another, religious threat for a third, and security panic for a fourth, without any of them encountering the others.
The research on the media dehumanization of Muslims and other outgroups is extensive and consistent. British tabloid headlines described the Taliban as “SAVAGE BEASTS” and “KILLERS IN CABINET,” using dehumanizing language and vivid imagery that perpetuates the perception of the group as uncivilized and inherently violent. A study of UK media coverage of Muslim women found that visual tropes—veiled, silent, backgrounded figures—were mobilised to construct them as “voiceless, passive, and linguistically deficient,” linking their appearance to presumed cultural deficits and framing them as problematic citizens.
But the older research was about representation. The new research is about automation, and the difference matters. Automated dehumanization does not need to persuade an audience of a general claim. It needs only to find the version of the claim that fits the audience’s existing grievance and deliver it in a form that feels like entertainment. The prejudice is not argued for. It is matched.
IV. The Conspiracy as Language
The Eurabia myth and the Great Replacement theory are the clearest examples of how language can be used to construct a victim identity and mobilise it politically.
Eurabia frames Muslim migration as a coordinated effort to conquer Europe, facilitated by domestic elites accused of betraying the populace. It is a European expression of the broader Great Replacement theory, which echoes Renaud Camus’s 2010 book Le Grand Remplacement and argues that white French people are being “replaced” by Muslims. The theory has a documented history: it emerged in France, spread through Europe, and became especially virulent in the United States. It was amplified after 9/11 and expanded again during the 2015 refugee crisis.
The political function is precise. It positions populist leaders as defenders of a “pure people” against both external and domestic threats, and it leverages the narrative to gather support and increase influence in democratic politics. It converts a complex set of social and economic changes into a simple story of betrayal and invasion. The audience is invited to see itself as the victim of a conspiracy, and the conspiracy is always one the leader promises to resist.
The research on these conspiracy theories documents the cycles of violence they have inspired. The language is not merely expressive. It is operational. It produces a frame in which violence against the perceived invaders, or their enablers, becomes a defensive act.
V. The Epistemic Crisis
The question of whether critical thinking is in decline is the wrong question, because it assumes a baseline of individual cognitive capacity that has somehow deteriorated. The evidence points at something different: a crisis in the conditions under which thinking is done.
The analysis in AI & Society describes it precisely. The old agreement—in which the public placed some trust in people with training and data—is coming apart. In its place: “TikTok lawyers with no legal background, Instagram doctors who have never treated a patient, and an internet culture that rewards confidence over truth.” The problem is not that people have stopped thinking. It is that the signal-to-noise ratio has been inverted by algorithms designed to boost engagement rather than accuracy.
The empirical work on this is now substantial. False information spreads faster than true information—the landmark Science study by Vosoughi, Roy and Aral found that false news reached 1,500 people about six times faster than the truth, and that the effect was driven by novelty and emotional arousal rather than by bots. Corrections from experts, in the same literature, often bounce off.
The algorithmic environment is not neutral. It teaches values: that the most confident voice is the most trustworthy; that feeling right is more important than being right; that expertise is one more opinion in the feed. Social media platforms function as engines for sculpting epistemic agency—not merely spreading false beliefs, but shaping the norms by which beliefs are formed and evaluated.
This is the context in which the victim narrative becomes so effective. An audience trained to prioritise emotional resonance over evidentiary rigour, and primed by algorithmic personalisation to see its own specific grievance as the central political fact, is an audience for which Eurabia or the Great Replacement is not a proposition to be assessed but a feeling to be confirmed.
VI. The Distraction Thesis
Is it possible that people are being led to see themselves as victims as a distraction from the outcomes created by their own governments?
The evidence supports a qualified version of the thesis, and the qualification is important.
The research on slopaganda is explicit about the personalisation of grievance. The same prejudice is repackaged for different audiences—economic anxiety for one, cultural loss for another, religious threat for a third, security panic for a fourth. The effect is to fragment the audience into a series of discrete victim groups, each with its own grievance, each addressed by its own tailored content, and none of them connected to a shared structural analysis of the conditions they are all experiencing.
This is the mechanism of distraction. A population that is angry about housing costs, wage stagnation, and insecure work is a population that might ask why those conditions exist, and who benefits from them. A population fragmented into victims of different conspiracies asks instead why the other group is being favoured, or why the elites are betraying “us.” The question of who benefits from the structure is displaced by the question of who is to blame for the grievance.
The political science literature on the rally-round-the-flag effect and the diversionary theory of war describes the same logic at the level of foreign policy. The mechanism is not that leaders invent a foreign enemy to distract from domestic failure—that version of the theory has weak empirical support. It is that the structure of political attention rewards the foreign crisis over the domestic one, because the foreign crisis has a clear enemy and the domestic one has a complex set of causes.
The victim narrative is the domestic version of this: it supplies a clear enemy and a clear grievance, and it directs attention away from the structures that produced the grievance in the first place.
There is a further turn to the screw, and it is uncomfortable for everyone in public life.
The same political class that benefits from the personalisation of grievance—because grievance is mobilising, and mobilisation is electoral—also condemns the violence and polarisation that the grievance produces. Social media is identified as the cause. Regulation is proposed. The platforms are summoned before committees. The remedies are aimed at the technology, the algorithms, the age limits, the moderation rules.
Some of this is genuine and necessary. But the diagnosis is incomplete, and its incompleteness is convenient. If the problem is the platforms, then the political rhetoric that feeds the platforms is not part of the problem. If the problem is the algorithm, then the speech that the algorithm amplifies is not the responsibility of the people who made it. The politician who blames social media for the increase in violent speech, while practising the politics of personalised grievance that social media rewards, has found a way to be both the arsonist and the fire brigade.
The specific claim—that senior Australian political figures have blamed social media for violence while contributing to the rhetorical environment in which violence becomes more permissible—requires verification, case by case, with the statements themselves and their dates. The structural observation does not require it. The structure is visible in the incentives: grievance is rewarded, blame is deflected, and the technology is a convenient scapegoat precisely because it is genuinely implicated.
The qualification on intentionality must be stated plainly. The research documents the effect, not the coordination. The system does not require a conspirator. It requires only that certain narratives are more engaging, more shareable, and more politically useful than others—and the algorithmic environment ensures that they are. Whether anyone is deliberately orchestrating it is less important than the fact that the system produces it, reliably, and that the people who benefit from it are the same people who are loudest in condemning its consequences.
VII. Thoughtlessness
Is it too harsh to suggest that in a world where thoughtlessness is incentivised, the thoughtless will take the lead?
The term requires qualification, and the qualification comes from Hannah Arendt.
Arendt’s report on the trial of Adolf Eichmann in Jerusalem introduced a phrase that has been widely misunderstood. She described Eichmann as demonstrating “the banality of evil,” and what she meant was not that his crimes were trivial. She meant that he was not a monster, not a demon, not a figure of melodramatic wickedness. He was a mediocre man, a bureaucrat, a careerist—and, in her word, thoughtless.
Arendt’s thoughtlessness is not stupidity. Eichmann was not unintelligent. Thoughtlessness, in her account, is the absence of the internal dialogue in which a person examines what they are doing from the standpoint of someone else. It is the incapacity—or the unwillingness—to stop and ask: what is this, from the perspective of the person it is happening to? Eichmann could recite the regulations. He could not imagine the boxcar.
This is the qualified definition, and it is the one that fits the present.
Thoughtlessness, in this sense, is the unwillingness or inability to reflect honestly on the ramifications and impacts of one’s own words and actions. It is not a deficit of intelligence. It is a deficit of reflection—and it is rewarded, because reflection slows the machine down. The minister who pauses to imagine the person on the waitlist is a minister who cannot deliver the announcement on time. The executive who pauses to imagine the family in the rubble is an executive who cannot meet the quarterly target. The commentator who pauses to imagine the consequence of the label is a commentator who does not go viral.
The incentive is not to be cruel. The incentive is to be fast. And speed, in a system optimised for engagement, is indistinguishable from thoughtlessness.
Where thoughtlessness operates, dehumanization follows, because dehumanization is what thoughtlessness produces when it is applied to a person. You do not need to hate someone to reduce them to an object of derision. You only need to stop imagining them. The label does the rest. The label is cheap, it travels fast, and it removes the person from the frame in which moral consideration happens.
This produces what might be called the two tiers of language—a grammar of us and them that is built into public speech at the level of vocabulary, not argument.
There is a language for us: our community, Australian families, hardworking Australians, small business, the battlers. These are words of membership. They carry warmth, legitimacy, and a claim on the reader’s sympathy.
And there is a language for them: irregular maritime arrivals, unlawful non-citizens, cashed-up criminals, welfare recipients, the long-term unemployed, collateral damage, soft targets, degraded. These are words of category. They carry no membership, no warmth, and no claim. They describe a class of thing rather than a group of people.
The two tiers are not always explicit, and that is what makes them effective. The grammar itself encodes the hierarchy. You do not have to say that some lives matter less. The vocabulary has already said it, and the sentence proceeds on the assumption.
The social psychology of this is well documented. Henri Tajfel and John Turner’s work on social identity showed that the mere act of categorising people into groups—even on arbitrary grounds—produces in-group favouritism and out-group derogation. The categorisation does not need to be hostile to produce the effect. It only needs to be made. And the language of public policy makes it constantly.
Judith Butler’s work on grievability extends the point in a direction that matters for this essay. Butler asks: whose lives count as lives, and what makes for a grievable life? Her answer is that grievability is not a natural property of a human being. It is a political allocation. Certain populations are framed as lives that cannot be mourned, and that framing is not a failure of empathy—it is an achievement. It decides which deaths produce a public response and which deaths produce a statistic.
The two tiers of language are the mechanism of that allocation. A death described as a tragedy is grievable. A death described as a system pressure is not. A woman killed by a partner is a domestic incident. A soldier killed abroad is a war death. The vocabulary has already told the reader which one demands a response.
Zygmunt Bauman’s account of adiaphora adds the final piece. Bauman argued that modernity produces conditions in which certain actions are rendered morally neutral—placed outside the domain in which moral judgment applies. The bureaucrat who processes a form is not committing an act of cruelty; he is doing his job. The engineer who designs the targeting interface is not killing anyone; she is building a product. The action has been removed from the moral register before it reaches the person who performs it.
This is not a failure of individual conscience. It is the design specification of a system that needs its functions performed by people who do not experience them as moral acts. And it is the same structure whether the form is an aged care assessment, a debt notice, or a targeting list.
The two tiers of language, the us and them, the reduction of the person to a category—these are becoming more and more unaffordable to humanity as a whole. Not because they are cruel, though they are. Because the problems that now confront every society—climate, pandemic, migration, war, inequality—cannot be solved by a species that has arranged its public language to make most of its members ungrievable.
VIII. What the Pattern Reveals
The pattern that runs from “degraded” to slopaganda is not a story of linguistic decline in the ordinary sense. It is a story of the industrialisation of moral disengagement.
The word “degraded” was a euphemism. It was spoken by a human being, in a specific context, to a specific audience. The slopaganda is a system. It is generated at scale, personalised to the individual, and delivered through platforms whose business model depends on emotional engagement.
The euphemism has become an environment.
And the victim narrative is the most efficient product of that environment, because it does two things at once. It gives the audience a feeling of moral clarity—they are the wronged party, and their grievance is real—and it directs their attention away from the structural conditions that produced the grievance in the first place. The audience is not passive. It is actively engaged. But its engagement is channelled into a form that cannot produce accountability, because the enemy it has been given is not the structure that is harming it.
This is the dynamic that the United Nations Special Rapporteur, Francesca Albanese, described in her 2026 report as mediacide—the destruction of the conditions under which truth can be established—extended to the domestic sphere. The report was written about Gaza, and it is contested. The New York Times rejected its characterisation of its coverage. But the mechanism it describes is not confined to one conflict, and it is not confined to war.
The destruction of the conditions under which truth can be established is a feature of any political environment in which the audience is fragmented, the evidence is personalised, and the grievance is weaponised. The machinery is the same. The targets are different.
IX. The Infrastructure
The question is not whether critical thinking is in decline. The question is whether the infrastructure of critical thinking has been dismantled.
That infrastructure has three components. The first is shared evidentiary standards—a common understanding of what counts as evidence, what counts as proof, and what counts as a claim that has not been established. The second is institutions that produce and adjudicate knowledge—universities, public service research arms, statistical agencies, courts, and a press capable of distinguishing between them. The third is a public language that can describe a condition without sedating the response to it.
Each of these has been weakened, and none of them by a single decision.
The evidentiary standards eroded as the algorithmic environment rewarded confidence over accuracy. The institutions were hollowed out by decades of funding decisions that treated the social sciences as a cost centre and the public service as a purchaser rather than a producer of knowledge. The public language decayed as the incentives of political communication rewarded the euphemism, the label, and the announcement over the description of what was actually happening.
The accumulation of those decisions made the alternative more profitable. And a system optimised for what is profitable will produce what is profitable, reliably, without requiring anyone to intend the result.
This is not a conspiracy. It is a structure. And structures of this kind are more difficult to confront than conspiracies, because there is no one to arrest and no document to publish. There is only the accumulated weight of a thousand defensible decisions, each made by someone doing their job, each of which made the next one easier.
The way back is not a single reform. It is the reconstruction of the three components, in the order that makes the others possible.
Restore the institutions first, because without them there is no authority to establish a standard. Fund the disciplines that train people to see structural harm, because without them there is no cadre to staff the institutions. Rebuild the public language, because without it the institutions cannot explain what they have found to the public that funds them.
None of this is expensive. All of it costs speed and control—the two things the present system exists to maximise. That is why it does not happen. And that is why the question of whether it can happen is not a technical question. It is the oldest political question there is: whether a society can be persuaded to pay, in the present, for the conditions of its own survival.
The word “degraded” began as a moral judgment and ended as a description of a service that is not working properly. The journey is the essay. And the question is whether the same journey is now being made by the words that describe what we are doing to each other.
Andrew Paul Klein & Sera Elizabeth Klein
