Studying Political Communication and Media in East Asia

2026.05.12

Course Recap｜Introduction to Political Communication: From the Communication Model and Framing to Doing Discourse Analysis

This was the first session of an intensive mini-course on political communication, taught by a scholar of Chinese studies who has written a textbook on qualitative research methods. He noted that many books on political language are really about philosophy and theory and never answer the practical question of how one is actually supposed to do the analysis; his aim in both the book and the course is to make research methods clearer and more usable. The course stresses method over theory alone, and across four sessions covers political language, visual communication, political space and mass events (Tiananmen Square, military parades, the Osaka World Expo), and social media and digital politics; a Friday public lecture addresses how misinformation, disinformation, and AI are eroding the public sphere. This session focused on political language: the first half built the conceptual foundations of communication and discourse, and the second half put discourse analysis into practice.

Why Study Political Communication? The Three P’s of Politics

The professor began with the “three P’s” of politics: politics (the process of negotiation, debate, and contestation), polity (the institutions—parliaments, government departments, even buildings), and policy (the outcomes—rules and regulations). All three are deeply embedded in communicative processes, yet political scientists often overlook this dimension, which he sees as a missed opportunity.

Political communication is in fact everywhere: terrorist attacks are fundamentally about producing fear and sending a message; gestures are themselves political (a German chancellor kneeling for WWII, how Trump and Xi Jinping shake hands); demonstrations (such as the Sunflower Movement) and symbolic images (the Obama administration watching a raid from the situation room) all communicate an attitude. He also introduced the power/counterpower frame, while cautioning against it: reducing PRC politics to “the state versus the people” is inaccurate, since a great deal of negotiation and collaboration is actually going on.

The Basic Model of Communication: Message, Medium, and Meaning

The traditional model runs: producer/sender → message → recipient/receiver. It is often criticized for being too rigid and for treating the message as too unified, but the professor finds it still useful: we can examine any single component, and we can also examine the processes of production and of reception/decoding themselves. The key missing element is the medium—the channel and container that carries, transmits, or stores the message and data.

This leads to the distinction between data and information: data are the raw materials that have been collected, whereas information is data that has been worked on, organized, given meaning, and presented to someone. In other words, data become information only once someone interprets them (a pile of numbers becomes information only when someone identifies it as, say, “voting behaviour from 2022”). He also cited Marshall McLuhan’s famous line, “the medium is the message”: in the early television era McLuhan argued the medium mattered more than the content. The professor is uncomfortable with the strong version, but accepts the core insight—the same message delivered via television, a newspaper, ChatGPT, or a manga takes on different meanings and possibilities.

Framing and Priming: Guiding How We Think

Agenda setting tells you what to think about; framing builds a structure of reference by bringing in certain concepts and connections, guiding how we understand something. The professor stressed that it is almost impossible not to frame—we are framing the whole time we speak—but the problem is manipulative framing.

His favourite example is migration: it can be framed as an opportunity, especially in low-birth-rate societies (who will sustain the social systems, who will do the work), or as “vermin, filth, threat”—the latter steering perception in a disturbing way toward particular policy conclusions. This kind of reactionary right-wing framing is visible across East Asia, Europe, and the United States. He also mentioned priming, another powerful tool from political psychology.

Discourse, Power, and Construction: Who May Legitimately Speak

Using the “tree falling in the forest / sound” thought experiment, the professor returned to the distinction between data and information, and between signal and sign—the heart of semiotics: a sound can be a signal, but it becomes a meaning-bearing sign only when someone attributes meaning to it.

This opens onto the debate between radical constructivism and critical realism: the physical world exists and continues independently of us (a frisbee really does fly), but it acquires political meaning only through us, because only humans do politics. He highlighted an important political twist: constructivism was originally a left-wing tool (Latour, for instance, criticized science as not purely objective and showed how resources and power shape knowledge production, in order to make science and society fairer and more open), but it has since been appropriated by the alt-right as a “blueprint for domination.” Latour later clarified that criticizing science was never the same as denying it—what is debatable is the physics of why the frisbee flies, not whether it flies at all.

Discourse also involves institutionalization: when does a “truth” established through language, thought, and social practice crystallize into an institution with lasting power? Moreover, “common sense” is itself a political act—it is not automatically true but is collectively constructed, and the very making of common sense deserves scholarly attention. This is also why feminism and critical race theory rely heavily on discourse analysis to unmask how dominant power operates (he cited the gendered histories of medical authority, early computing, and gynaecology). In sum, doing discourse analysis means asking who has the power to make statements about an issue and what those statements do, and who may legitimately speak about a topic.

How to Do Discourse Analysis: Two Traditions of Content Analysis

After the break the session turned practical. The professor first distinguished the American and the continental-European traditions of “content analysis”: the American approach is quantitative, big-data, and word-frequency based (e.g., counting positive/negative coverage of parties, sentiment analysis) and, under American academic influence, is popular in East Asia; the European approach (at least in his context) is qualitative, where even a close reading of a single text counts as content analysis. The two can in fact be mixed.

Here “content” means not only language but also visual art, photography, moving images, and other media content. And “analysis” must meet three conditions: it must be systematic, empirical, and transparent. This is the shared academic foundation of the humanities, social studies, and the hard sciences alike. The professor even argued that the social “sciences” should really be called social studies, but stressed that what makes academic work academic is having procedures and protocols and being transparent about them, so that others can check and replicate the work.

Designing the Study: Question, Time, and Layer

The next step is to formulate an operationalizable and relevant research question. The professor recommended narrowing things down: most people cannot work across centuries of material the way Foucault did; it is far more feasible to focus on a specific actor, a specific issue, and a specific moment (e.g., how a particular leader talks about one issue at one point in time).

In terms of time, he distinguished synchronic from diachronic: for student-level work, a synchronic (single point in time) study is usually enough; if going diachronic, it is best to anchor the analysis to a concrete “discursive event” (Fukushima, the Russia–Ukraine war, the COVID pandemic, the Iraq War, the Strait of Hormuz) so as to work within a manageable frame. One then chooses a layer: academic, news, and public/everyday discourse are different social layers that influence one another, but multi-layer analysis is very time-consuming, so focusing on a single layer is usually advisable. Finally, one must settle the materials and digitization: is the text already digitized, does it need transcribing? Speech-to-text tools are now quite good, but the output still needs checking (interview transcripts are often only 70–80% accurate), and copyright and ethics must be considered. Every choice should be justified transparently.

Coding, Quantification, and the Machine: Practical Trade-offs

A core systematic step in qualitative method is coding: going through the material by hand and labelling themes, issues, and short phrases at the level of paragraph, sentence, or word—which requires first defining a coding framework, i.e., which elements you are identifying. Topic modelling asks an algorithm to find clusters of frequently co-occurring words; but how many topics to extract is a human decision with no objective answer, and researchers often adjust it repeatedly until the result seems “meaningful,” still needing to check sample sentences before labelling a cluster (e.g., the “security” topic).

The professor’s position: he does not oppose using ChatGPT or machine learning as aids, but argues for doing it by hand first and understanding the method itself, so as to retain the capacity for critical reflection and avoid prematurely “cognitively offloading” the work to a machine. He is also wary of purely quantitative methods: sentiment analysis on its own (say, 70% negative, 30% positive) is of limited interest; the more fruitful move is to use quantification as a starting point and then dig into the qualitative—for instance, asking how exactly those 30% are positive, and whether there are different kinds of positivity (such as “positive energy” in CCP ideology).

Case Study One: How the US and China Talk About the Internet (2010)

The professor compared two 2010 statements—one from the United States and one from the People’s Republic of China—on the political meaning of the internet (a moment with its own specific historical context). Using WordSmith, he built a concordance of the keyword “government” to examine the words surrounding it. In Hillary Clinton’s speech the plural “governments” dominates, with frequent references to “foreign governments” and “authoritarian governments,” while she refers to her own side as “our government / our own government”; the Chinese version shows a striking contrast.

He then moved to linguistic features and rhetoric: watching for passive constructions (when political actors remove agency and imply that things “just happen”) and for different word fields—of war, disease, or liberal democracy. Clinton’s text shows a clear us/them narrative and a tricolon-style rhetorical escalation, placing “violent extremists, criminal cartels, sexual predators, and authoritarian governments” in a single category. The professor found this rhetorically deft but highly manipulative: however one criticizes China’s authoritarian government, it is not the same as a sexual predator or a terrorist; using conjunctions like “and” and “but” to manufacture equivalences and contrasts that do not logically follow is precisely the kind of rhetorical strategy worth spotting.

Case Study Two: The “Asia vs. the West” Pandemic Discourse

The second case was an article on COVID arguing that East Asia “won” the pandemic, attributing this to ancient philosophies—Confucianism, Buddhism, Taoism—set against “the West” and “big government”; it uses words like “fast, immediate, voluntary” for Asia and panic-tinged language (e.g., “hysterical”) for the West, and mentions not a single Asian failure. Students observed that the author implies Asia’s “big government” is not bad for its people and that some freedom can be sacrificed for the more important value of life—and that the word choices are already heavily framing.

The professor’s critique had two levels. First, “Asia vs. the West” is a false and dangerous binary: the responses of Vietnam, Thailand, Taiwan, and China were in fact very different and cannot be lumped together, and “the West” is just as muddled (Europe is not the United States; where do Japan and Taiwan fit?). Such Orientalizing narratives are tied to colonialism and imperialism and have historically been abused to justify violence; the surge of anti-Asian hatred in Europe during the pandemic was one result. Second, the framing misfocuses attention: parts of Asia handled the pandemic well not because of Confucian philosophy but because of the experience of SARS, institution-building, political trust, and good leadership—yet none of these decisive factors appears in the article at all.

Reflection｜Discourse Analysis Asks “How” and “Why,” Not Just “What Was Said”

The professor closed by stressing that a beginner’s instinct is to summarize what the text says. That is a necessary step, but not the endpoint of discourse analysis. Discourse analysis goes further to ask how they say it, even why they say it, and—through the specific words and concepts being established—what politics this implies. Identifying passive voice, binary oppositions, word fields, rhetorical escalation, and framing is exactly the way to see through how a discourse works and then to critique it. And because every step involves the researcher’s choices, no two discourse analyses are ever identical; the best way to stay academically sound is to be transparent about why those choices were made.

———————————————————————————————————–

2026.05.14

Course Recap｜How Do Political Images Work? Visual Communication, Semiotics, and a Forensic Reading of Newspaper and Magazine Covers

This was the second session of the intensive course on political communication, moving from political language to visual communication. The professor noted that many people who do discourse and framing analysis do not take visual or multimodal communication seriously—they may make a token gesture toward an image (“yes, this is an advertisement, there’s a big picture”) but in practice only analyze the headline; and those who study film often come from literature studies, so they use the same tools they use for Shakespeare to analyze, say, The Godfather, treating film as a “text.” That has some value, but it overlooks everything else happening in visual analysis. The goal of this session was to supply that missing toolkit: first grounding the discussion in the theory of semiotics and perception, then working through a “forensic method” step by step, and finally taking apart the front pages of newspapers and magazines. The professor stressed that theory and method must speak to each other—do only one and you miss the other.

Why Take Visual Communication Seriously

We live in hypermediated societies, under a constant barrage of advertising, images, and film. For exactly this reason, the professor argued, we must train ourselves to see better—to bring critical sensibility and empirical care to identifying how political images are built and how they work, rather than just glancing past them. One scholar frequently cited in cultural studies and visual communication is Roland Barthes, who did extensive work, using advertising as his example, on how language and image relate to each other. The professor also explained that the previous session drew mostly on French influences, while this one would turn quickly to the visual-communication tools developed by Italian, American, and German scholars.

The Forensic Method: Describe, Explain, Interpret

The professor proposed a “forensic method” that breaks looking at an image into three deliberately separated steps: first describe, then explain, then interpret. This is of course artificial—we don’t do it this way in daily life—but it is very useful: if we jump straight to “what it all means,” we tend to miss the small components that led us to that conclusion, and we miss how the image’s makers are guiding us toward it.

The visual can also frame our perception. The professor gave an example: Germany’s first Green Party foreign minister had been an anti-nuclear activist in his youth, and a right-wing tabloid used a photo from a protest to frame him as “attacking police officers.” But zoom out and you see he was not holding a stick—he was holding the rope separating demonstrators from the police; he was doing exactly what was permitted, inside the safe protest zone. The paper changed the meaning of the whole event simply by choosing that one frame. He also showed a Transport for London clip urging drivers to watch for cyclists, to illustrate that only by looking frame-by-frame can you catch the manipulations you miss while watching casually.

Semiotics: How Meaning Is Made

The professor then returned to philosophy to introduce semiotics—the science of the sign, the study of how things become meaningful and what the “vehicle” of meaning is. He admires semiotics because it genuinely counts as a communication science: it is used in computer science and in biology (to study how ants or bees communicate), since humans are not the only ones who transmit meaning. One of his favorite authors is the Italian Umberto Eco, who wrote The Name of the Rose (adapted into a famous film starring Sean Connery).

At the heart of semiotics is the semiotic triangle: the object (the actual apple), the sign (the picture of an apple I show you), and the interpretant (the meaning you associate with it). Meaning is generated when the three come together. We can also move the corners of the triangle closer or farther apart, which yields different kinds of signs.

Three Types of Signs: Icon, Symbol, Index

This gives Peirce’s three classes of signs. First, the icon: it closely resembles the object, like a realistic photograph of an apple—which is exactly what people imply when they call something “iconic.” Second, the symbol: it does not resemble the object and is purely conventional, like the flag of the People’s Republic of China (nothing on the flag “looks like” China), and indeed all language is symbolic—no word actually “sounds like” the thing it denotes (even onomatopoeia like “kaboom” only imitates, it isn’t the thing). Third, the index: it points to the object through an existential or causal link—smoke points to fire, a bullet hole points to gunfire, a fever points to something wrong in the body—which is also where semiotics overlaps with the natural sciences, such as medicine.

The professor stressed a commonly misunderstood point: it may seem that only symbols require cultural background, but in fact all three sign types require learning and cultural context. Someone who has never seen fire and smoke may not connect smoke to fire; someone who doesn’t know what a fever is cannot infer illness from temperature. Even the seemingly most “natural” photograph is no exception—small children looking at photos or screens don’t know it is just a “window” onto somewhere else; he shared a meme of a three-year-old FaceTiming grandma on an iPad who, after dropping it, cried out “Grandma, are you okay?”—because the child hadn’t yet learned that convention.

Connotation, and the Biology of Perception

Where culture becomes most visible is at the level of connotation. Semiotics stacks a second triangle on top of the first: the first triangle clarifies “what it is” (denotation), the second clarifies “what it means” (connotation). The same apple may, for a Christian familiar with the Bible, connect to Adam and Eve, temptation, and original sin; for someone unfamiliar with that story, it carries no such meaning. He can also hold up an iPad and say “Apple”—it has nothing to do with the fruit, yet anyone who knows the company understands.

But the professor cautioned that perception itself is strongly biological, not necessarily cultural. He gave several examples: the blind spot on the retina (your brain automatically “fills in” the gap where there are no photoreceptors, so you don’t see a black hole in your vision); the yellow lights in long-term parking lots that make a red car no longer look red at night, so people lose their cars; and the famous “blue-and-black or gold-and-white” dress—half of people see one, half the other, purely because each person’s eyes were “calibrated” over their lifetime to expect a different lighting model (the correct answer is blue and black, but seeing gold and white is perfectly normal). From this he drew the distinction between a conceit (a willingly accepted illusion) and a deceit (deception with ill intent): looking at an oil painting or going to the cinema, we knowingly and willingly enter the illusion—that is a conceit; reframing the foreign minister’s photo as that of a violent extremist is a deceit. He therefore argued that red is not “seen fundamentally differently by Chinese and Europeans”—biologically we see the same red; what differs are the learned cultural codes and associations (imperial red and gold, the red of communism).

Recoding as a Form of Resistance

Since meaning comes from encoding, we can also apply a different code to reinterpret something deliberately, thereby exposing the artifice of the original code. The professor described an exercise by a colleague who admires Derrida: take the PRC’s 60th-anniversary military parade and flag-raising on Tiananmen Square (1 October 2009)—covered in Hu Jintao posters, encoded to carry a very particular nationalist message—and willfully “misread” it as a gay Pride parade, so that the whole thing becomes the story of “Hu Jintao coming out at Pride.” That is obviously not the intent, yet it fits surprisingly well, is very funny, and precisely lays bare the artificiality of the code.

A more profound real-world case is the film Avatar. It was a huge hit in mainland China and, unexpectedly, became a “political film”: the original is essentially about European colonists trampling Native Americans, but many Chinese viewers at the time recoded it through their own shared experience—being forcibly evicted while their homes, Beijing’s hutongs, and old Shanghai were bulldozed to make way for malls and high-rises. To them Avatar became the story of “the city inspectors coming after us,” and watching or discussing the film became a way to push back against the state and the Party. The professor used this to suggest that scholars should sometimes do this exercise too—place something in a completely different context and see what it looks like. He also introduced multimodality: the same message can be layered across modes simultaneously—writing, speech, gesture, even smell—so when analyzing a film, it is worth examining how these different modes work together to generate meaning.

China’s Three Newspaper Types: One Conglomerate, Different Political Roles

After the break the session turned practical. The professor began with three 2009 anniversary covers from different Chinese newspapers, to explain their distinct roles in the Chinese media. At the core is the “daily” (the People’s Daily and provincial dailies): this is the Communist Party’s propaganda paper, the best place to learn what policy is being promoted and how; its reporting almost always follows the three-stage structure his Singaporean colleagues joke about—“our leaders are very busy, everything in China is great, everything abroad is terrible.” Almost no one buys these papers; they are heavily subsidized, printed, and handed out to state organs and Party cadres to read, and they lose money.

The money is made by the “metro / morning tablet”: brightly colored, with bold headlines, the kind you buy at a kiosk on your commute—hence its visual style differs sharply from the static, traditional daily. The third type is the “weekend / weekly”: with a full week to work, its journalists produce long-form investigative reporting and are more intellectual. The professor noted that Southern Weekend (Nanfang Zhoumo) was at the time the face of Chinese investigative journalism—people pushing back against mainstream narratives from inside a state media enterprise—a reminder not to treat Chinese media as a monolithic propaganda machine. He cited the 2013 Southern Weekend New Year’s message incident, when Party officials inside the group swapped out the front page overnight, prompting journalist protests and a strike. He added that this space has shrunk dramatically since Xi Jinping (the era of “positive energy” and “telling China’s story well”), which is why he had to go all the way back to 2009 for his example.

Case One: The Human-Rights Argument “Hidden” in a 2009 Anniversary Cover

Reading that 2009 cover with the forensic method reveals many deliberate clues. The figure is not climbing up toward the light—he is abseiling down, into a cave, away from the light; the characters are rendered muddy, broken, and rusty; western China is dark, as if “switched off,” with growth concentrated in Shanghai; in the original the clouds are darker than they appear on the projector, and the character “六” (six) forms a barrier across the image. Together these signs imply that China is in fact in decline.

The article’s title was “Let Every Chinese Person Stand Up”—deliberately distinct from Mao’s “The Chinese people have stood up,” emphasizing every individual. The first half is long-winded nationalist language (most of the professor’s students in Leiden stopped reading there, finding it dull), but midway the tone shifts to what China still lacks—human rights. In other words, a state-run newspaper, using this image sandwiched between nationalist language, was quietly making the case for human rights in China—and it passed the censors. Why? Either the censors didn’t notice (so it passed), or they did and faced a dilemma: pulling it would let everyone know something had been censored and make them want to know what, which is risky; so it was easier to let it through. Add ambiguous imagery and language, plus the fact that the censorship apparatus is not monolithic (it involves hundreds of thousands of people, some of whom hold liberal views and privately dislike what is going on), and there is room for plausible deniability—the journalists stay safe, the censors stay safe, and an intriguing story gets made at the same time.

Case Two: Describe–Explain–Interpret on Covers (Time / Germany / Shanghai)

Then came the cover exercise. The first was Time’s cover on the Fukushima nuclear disaster. The professor used it to demonstrate the difference between description and interpretation: when a student said this was “a crying woman,” “crying” had already crossed from description into interpretation—we can’t be sure she is crying rather than wiping away sweat (she wears a glove and holds something like a towel against her face). A more neutral description would be: a profile, close-up view showing only the head; a woman who appears middle-aged and East Asian, in a green (work) shirt, wiping around her eyes with a towel or tissue. At the “explain” step, one can note that she seems to be laboring, cleaning up; and “why a woman” matters—it ties to gendered associations of cleaning and tidying the home. At the top is white text, “Japan’s meltdown: earthquake. tsunami. nuclear disaster. resilient,” plus the magazine’s name.

The second was one of Germany’s best-known news magazines (a social-democratic-left publication) on the pandemic: the image shows someone who could be right out on the street, wearing earphones / holding an iPhone just like any of us, yet masked, with an eerie, foreboding sense that something is off—framing the pandemic as “a Chinese thing handed to us along with our iPhone.” The professor found this fairly racist, and an illustration of how such visuals can shift policy attitudes toward nuclear energy and pandemic preparedness.

The third was a Shanghai news magazine. Its Chinese headline used words like “go to all-out war against the pandemic / strike the virus”—and the war language matters greatly. The two buildings in the image look like the colonial concession architecture of Shanghai’s Bund, but are in fact in Wuhan (an old customs / trade building strongly resembling the Bund style); a Shanghai reader would instantly know it refers to Wuhan, whereas his Dutch students often mistook it for the United States. Juxtaposing an old building with a newer one neatly echoes China’s discourse of “5,000 years of continuous civilization” (strictly it’s about four-thousand-something, but “5,000” sounds better), implying “we are one great civilization, experienced, and we know how to handle this.” The figures in white are the volunteers known during the pandemic as “Dabai” (the cover also corresponds to reports of Shanghai’s medical staff rushing to help Wuhan).

Most crucial is the posture: unlike the suited man on the previous cover, these figures look resolute—like warriors, like soldiers, about to “do battle”—matching the whole register of war language. The professor explained that this posture has a lineage, called the “socialist realist gaze,” drawn from Mao-era propaganda art: the figures are usually faceless, unidentifiable, hard to tell male from female, because they are not identifiable individuals but “the masses.” Even if people no longer much like Mao today, the symbolism still works—just as in many European contexts you need not be Christian to recognize, and even fear, Christian symbols. The image also contains viral particles—the enemy in this “war.” But that is the problem: a war needs an enemy, and the virus is invisible, so many people began searching for “who is my enemy,” and the answer was steered toward “foreigners”—in Guangzhou there were even incidents of African / Black residents being shunned and attacked as supposed virus carriers. From an us/them framing and the language of war, it is only a small step to xenophobic narratives.

Reflection｜Similar Tools, Different Politics

The professor concluded that the three news magazines seen today each communicate very different politics, yet all reach for similar strategic tools. Visual imagery is never neutral—it is carefully constructed; learning to slow down, stop, and take apart its components one by one, reading with a critical and empirical eye, is exactly the skill this course aims to give us. The same method can be extended to moving images (film). Tomorrow’s public lecture will go on to discuss AI and digital politics, and the question of “deceitful media.”

———————————————————————————————————–

2026.05.19

Course Recap｜Space as Politics: Analyzing the “Politics of Space” from Tiananmen Square to the World Expo

This was the third session of the intensive course on political communication, turning to “space and mass events.” Professor Schneider noted that this session is the stepping stone between last week’s discourse/visual topics and Thursday’s digital, structural questions: space is a “place” of communication, something we communicate about, and something we inhabit. So how do we analyze it? The session began with the physicality and surveillance of Tiananmen Square, then introduced a “matrix” for making sense of space, and used the Shanghai Expo (and the Osaka Expo from the students’ assignment) as the main cases, running throughout a key methodological reminder: what the organizers claim something means is not the same as how people actually experience it or what it actually does in practice.

Tiananmen Square: Start with “Physicality” and Surveillance

Before discussing how Tiananmen Square is represented, the professor insisted on first looking at its physicality. Students who had been there described it as enormous, cold and windy in winter, and—because it is so vast and empty—oddly lonely; the buildings are very tall, an odd hybrid of “Chinese-style” and “Soviet-style” architecture. Historically, Mao demolished a whole section of the old city to build this monolithic administrative center, breaking up and reinventing Beijing’s north–south axis—an axis that runs straight to Tiananmen (the gate), with the Forbidden City behind it.

The strongest impression is “securitization”: strict checks to enter (one student waited three hours), visible cameras everywhere (even in the underpasses), guards, and heavy commercialization (souvenir shops, food). The professor stressed that the map is, in effect, “wrong”—you cannot walk onto the square from just anywhere; metal barriers surround it, and you can only enter along Chang’an Avenue, through specific underpasses, after an “Apple-level” security check. The cameras being deliberately visible is itself a choice: China may be among the most surveilled places on earth, but unlike London or The Hague—equally camera-covered, yet hidden (students who counted found hundreds on a five-minute walk from the station)—in China, showing the cameras is part of the power, announcing “you are now in the panopticon, you are being watched.” He mentioned Shenzhen’s drone camera-robots that patrol intersections and broadcast at violators (“you are parked illegally, move along”), and big screens that name jaywalkers in real time (“Mr. Wang is crossing on red”) to shame them—dystopian and creepy, in his view.

Paternalistic Governance: Guiding Behavior through the Built Environment

This is of a piece with how compliance is organized in mainland China—highly paternalistic: the Party is the father or mother of the nation, while the laobaixing and the dazhong (the people) are treated as children who must be taught. So (more densely than anywhere else in Asia) you find people at intersections waving flags to tell you when to walk, even though there are already traffic lights; these are usually unemployed retirees, almost a make-work program. Some intersections even have odd reward schemes: with few enough incidents, an intersection can earn a “civilized intersection” plaque, with the local Party and neighborhood committees responsible for keeping things “civilized.” The professor raises this in a political-communication course because there is a slippage here between communication/discourse/representation on one hand and social interaction/social practice on the other—hard to separate, and sometimes we shouldn’t. Using the built environment to guide behavior is itself part of political communication.

A Matrix for Analyzing Space: Place vs. Space

The professor then introduced the “matrix for making sense of space” from his book (using an Escher image to show how crazy space can get). One axis is the three dimensions of space: material space (what we experience crossing the road), representations of space (when I show you a photo or map of the street), and spaces of representation (when you later tell a friend what crossing the street was like). The other axis is: absolute physical space, relative space (things related to one another—one building higher, another lower; one near, another far; or in motion), and relational space (people moving around within these spaces). Examining each overlap in the matrix lets you take apart what is going on in a given place. He also drew the readings’ distinction between place and space: as opposed to the more abstract, geometric notion of space, place refers to the lived experience of different people within it.

A student used a train station as an example: a huge, very tall yet very empty station with no benches in the main hall, so people sit on the ground, which in turn makes the crowd hard to move through. Why no benches? They reasoned that it is to keep people moving—“go catch your train”—since you only get a bench after scanning your ticket and reaching the platform, effectively holding people captive. A clear case of using spatial design to guide behavior.

The Shanghai Expo: Staging “Chinese Modernity”

The main case was the 2010 Shanghai Expo. Even the “geopolitical arrangement” of the pavilions is telling: Asia (Japan, Korea, India, etc.), a Europe section (Germany, Poland, Switzerland…, with Turkey and Greece side by side—who, in reality, don’t always get along), Australasia (Australia), and the Americas (US, Canada); as for Africa, most countries got only a small slot in one big trade hall (a few, like South Africa, had standalone pavilions). The other side held mostly industry and city pavilions (Hamburg, Rotterdam, etc.) showcasing how to “manage a city.”

From this the professor opened a discussion of modernity: only by starting from the Opium War (Britain’s attack on China) do you get the “hundred years of humiliation” (the 19th century, the First Opium War—the Chinese way of counting). History is, after all, “made up.” “Modernity” is traditionally dated from Britain’s Industrial Revolution, though some trace it to 15th–16th-century Europe (the Enlightenment, the late Renaissance, the first stirrings of capitalism, the colonial age, the emergence of the nation-state, and the moment the “masses” gained political agency); some even argue for an early modernity in Song-dynasty China. The less contested factors are industrial development, the building of nation-states, and the globalization of violence (WWI, WWII, the atomic bomb). Later communication technologies and globalization are mostly called “high modernity” rather than “postmodernity”—we haven’t really left modernity; each society just develops its own version. The Party is among the most “modernist” organizations on earth, because Marxism (like fascism, like liberalism) is itself a modernist ideology; it firmly believes in Marxist historical materialism—the world advancing through stages via innovation and development.

So in the China pavilion you see exactly a performance of “Chinese modernity”: how an ordinary Shanghai living room evolved from the 1970s through the 1980s, 1990s, and 2000s—the unspoken subtext being “all thanks to the Communist Party.” Tellingly, there are no 1960s and no Cultural Revolution: the narrative begins at the tail end of the Cultural Revolution and then moves forward “naturally, organically.” It is also a story told through consumer products (sewing machine → refrigerator → computer → flat-screen TV), so every ordinary Chinese person can relate to what “being modern and developed” looks like; and all of it is set on the frame of “5,000 years of continuous history” (never mind the Yuan dynasty). He noted the exhibit was built heavily around the Qingming Scroll—presented as evidence that China “appreciated the urban environment” very early on.

Modernity and the Nation: Invented National Narratives

From modernity the talk flowed into nation and nationalism. The professor returned to Benedict Anderson’s “imagined communities,” noting that this school is called “modernism”: it argues that the nation (and the nation-state built on it) are modern inventions, arising roughly after 1600–1700, after the Peace of Westphalia; the nation is justified retroactively—you have to “make France happen” and “make Germany happen” (invoking, say, the Sudeten Germans), often through great violence, standardizing names and measures and forcing a single language. For constructivist-leaning IR scholars this connects easily: analyze the identities politicians present for the nation-state, which national mythologies they draw on, and how they justify “America,” “China,” or “Taiwan”—well suited to discourse or visual analysis set in historical context.

He cited James Wertsch’s Voices of Collective Remembering on Russian nationalism: Russia has a recurring moral narrative whose pattern is almost invariant—innocent Russians minding their own business, an evil outside force attacks, they have no choice but to rise and fight, they win gloriously and defend the motherland, and then everything settles into peace. From Napoleon to Hitler, to the Americans in the Cold War, to NATO today, it’s the same story; grasp the pattern and you understand how Putin convinced Russians the war in Ukraine is “justified.” The same holds for China—the “hundred years of humiliation” and “5,000 years” are all constructed, but assembled from real materials, so they work once connected. Every nation-state does this: Germany is especially interesting—most Germans will say “we Germans have no nationalism (because of WWII and the Nazis),” yet “we Germans have no nationalism” is itself nationalistic, since it still says “we Germans.” Taiwan is interesting precisely because the KMT and the DPP hold very different understandings of what the nation is and where Taiwan fits—you can take apart the identity versions each political actor constructs.

Hegemonic Discourse vs. Lived Experience: Don’t Be Led by the Organizers’ Story

Some colleagues read the Shanghai Expo as a “hegemonic discourse”—a stable, harmonious utopia fusing global capitalism with Chinese civilization; another colleague went further, calling it “the only future, one that welcomes no protest”—after all, the Expo’s theme was the future of sustainable cities, oriented toward future utopias. Such readings often inherit postmodern writing (the “the World Fair is everywhere… simultaneously nowhere and now here” kind of word-play). Here the professor inserted a writing warning: don’t write like this—it is the show-off “it’s both this and that” style of too much French philosophy; unless you’re a top-tier writer, don’t do it in your MA thesis. The style descends from Jean Baudrillard’s “simulacra”—“signs of the real substituting for the real itself,” Matrix-level philosophy (The Matrix was built on his work, and Baudrillard himself was unhappy about the film; he was actually writing about Disneyland—he felt the “real America” had vanished because Disneyland was more real than the real thing).

But the professor made clear he does not buy reading the Shanghai Expo as one all-encompassing simulacrum or hegemonic machine manipulating us. Once you actually walk the grounds, too much fails to fit the story. First, the officially organized culture was more complicated: there was a contest for children to draw the “future city”—some clearly Party-approved modernity, but many drawings with no clear meaning, some quite dark. Second, many official theme pavilions were outsourced to foreign organizations (Party-approved, yet left to outsiders to interpret)—for example, a German company handling environmentalism. Third, and most important, the lived experience: people dragging crying kids and pushing grandma in a wheelchair, saying “I’m so hot, where’s the bathroom, where’s water, it’s 50 degrees outside”—nobody was paying attention to the propaganda narrative.

He interviewed the developer of the Australian pavilion, who said one of the cleverest things he’s ever heard in an interview: visitors come out the far end understanding that this was Australia, not Austria, and the job is done—“the best we can do is show people cute animals and beaches,” not aiming to impart more. By contrast, the German and Spanish pavilions carried a strong didactic intent, feeling a moral obligation to “teach the Chinese what a good person and a good city look like”—which the professor found very condescending, very Eurocentric and Enlightenment-ish. The Spanish pavilion was especially telling: run by marketing and political-communication experts who knew little about China, whose national myth is “overcoming authoritarianism” (Spain was once fascist, then democratized in the 1970s–80s)—a narrative close to South Korea’s and Taiwan’s; they noticed China’s authoritarian everyday (someone waving a flag at every intersection) and decided to “teach the Chinese a life beyond authoritarianism”—by designing “non-hierarchical uniforms” (made by Zara), expecting Chinese visitors to think “oh, it can be done differently.” The professor found it both baffling and intriguing that they truly thought this would work.

Consider too the “Expo passport”: visitors could collect stamps from each pavilion. Many scholars wrote papers claiming this “indoctrinated people into the imagined international community of nation-states,” making Chinese visitors “simulate being national citizens.” But that’s not what happened—people sprinted between pavilions just to collect stamps and then sold them on Taobao (even a black market emerged); eventually many pavilions moved the stamping outside so stamp-hunters wouldn’t disrupt those actually wanting to see the exhibits. If this is a “simulation,” it is a simulation of radically commercialized capitalism, with almost nothing to do with the border-crossing experience of a citizen. Hence the core lesson: the organizers telling you “what this wonderful thing does” does not mean people experienced it that way, nor that it worked that way in practice. Keep an open mind and ask “what else might be going on here”; this is also why, when studying space, he strongly recommends adding an ethnographic component if you can be on the ground.

More Spatial Cases: Singapore, Subway Art, the Osaka Expo

He offered further cases from the book. Singapore’s Marina Bay Sands has a rooftop infinity pool overlooking the whole city, at three to four hundred dollars a night; but “who took that beautiful photo?”—just behind a glass wall is a crowd of tourists jostling with selfie sticks. Would you really want to be in that pool? It shows precisely how space is organized, and that “normal” is constructed. Another is a subway art intervention in Japan: using art to disrupt people’s sense of the “normal”—why should it be normal to be “packed in a box with a crowd for two hours, all ignoring each other”? It also echoed a period of stagnation, ennui, and everyday meaninglessness in Japan that the art sought to stir. Running in parallel, though, is an ethnocentric cultural nationalism—for instance, Halloween on the Yamanote line (mostly celebrated by foreigners) being repeatedly pressured to stop on the grounds that “it isn’t Japanese.”

Finally, the Japan pavilion at the Osaka Expo, which the students’ assignment addressed: a wooden, ring-shaped structure, the timber read as cyclical, recyclable, and tied to the environment; the panels are not fully closed, so you can see inside from outside and walk in, symbolizing “openness to the world.” But the students also offered a critique: you can see the “ring” and enter it, yet the ring never opens and there is no real internal crossing between its parts—as if Japan “displays” its culture for the world to view and enter, but has no intention of truly “opening” or changing it. There are also themes of craftsmanship (doing one simple thing supremely well) and sustainability; and Yumeshima (reclaimed land), where the Expo sits, will be dismantled afterward and its materials reused—so what gets built there next? The professor again drove home the same point: what the organizers claim it means is not necessarily how people experience it, nor what it actually does.

Afterword｜The Assignment

Thursday’s final session will extend this discussion of architecture and space into “cyberspace / digital architecture,” asking what might be hidden inside our digital tools. On the assignment: the deadline is two weeks after the last session—by Tuesday, 5 June—submitted into folders the professor can access; he will comment on each one with feedback and assign a grade. There are three criteria. First, content: ask yourself what you really want to do and what question you want to ask, and whether the scope you pick is viable for such a small assignment (neither too big nor too small), with a satisfying answer reachable in two weeks. Second, analysis: whether you do appropriate empirical analysis (text or image analysis—anything taught in the course can be the foundation). Third, presentation: clarity of language and visuals—write clear, straightforward English, with no convoluted sentences, passive clauses, or words no one understands.

In form, it is “a single A4 page”: not a long essay, but a small topic you genuinely care about—ideally tied to your thesis or research—dug into a little using the tools from class, presented on one side of one page. Include a title, your name and date in a corner, an opening sentence stating “what this one-pager does and what it asks,” a large central section for the analysis, and a little space at the end for a conclusion telling the reader “why this matters and what you think.” Use visuals (even an infographic) if helpful, but don’t cram the page into a tiny wall of text; it should pass the “arm’s-length test”—printed out and held at arm’s length, someone should be able to tell at a glance what it’s about. The point of the exercise is that it is genuinely hard, but also extremely useful: whether you become a journalist, an analyst, or a policy advisor, you will constantly need to sum something up in limited space in a way everyone—not just experts—can understand.

———————————————————————————————————–

2026.05.21

Course Recap｜The Net as Space: Social Network Analysis, Digital Bias, and “Xi Jinping’s App”

The final session of the intensive course on political communication, extending the “space” lens into cyberspace and digital architecture. It covers social network analysis (nodes, ties, centralities), then how digital bias is produced, and closes with Xi Jinping’s “Xuexi Qiangguo” app to demonstrate two methods of dismantling it: actor-network analysis and functional resonance. A core reminder runs throughout: numbers, visuals, and systems are never neutral, and as researchers our every step is a choice and a political statement.

Recap: Student Image-Analysis Presentations

Picking up the previous day’s group presentations. The first group analyzed a YouTube video painting an extremely positive picture of China, but found it “artificial”: locals mob and welcome the foreign streamer as if they knew he was coming, and the comments are suspicious (a “new account” posing as an ordinary user urges people to visit Xinjiang to debunk the “genocide” lie); related coverage traces back to media effectively controlled by Alibaba. Their conclusion: interactions that look “natural and non-political” may conceal a deeper-than-assumed link between the public and the government. The second group disagreed, reading the streamer as someone being legitimized—he went viral precisely by doing “rogue” things Chinese people can’t. The third group summarized the same Chongqing video: staging infrastructure (trains through buildings), overcoming the “mountain city” terrain, an energetic and diverse city.

The professor used these cases to probe the gap between a creator’s intent and the audience’s lived reception, endorsing the path of “pick one topic and see how different people talk about it.” In offering directions he also reminded students: material can go in an appendix, but don’t bury the argument there; an Economist cover “can only tell you how the Economist sees Taiwan, not how the world does.”

Don’t Rush to Judge: Participant Observation and “Follow the Money”

The professor criticized certain bestsellers (the kind arguing the iPhone is evil) for offering almost no evidence—one cites the author’s daughter ignoring her at dinner to scroll, but she may just be a teenager who doesn’t want to hang out with mom; there are countless reasons. We see everyone on the subway staring at screens and conclude “how disconnected,” yet they may be talking to friends or parents. Participant observation often corrects this bias: younger generations are often deeply engaged and social with phones, while the ones truly bad at tech are the parents taking business calls at the table. He noted American teens (under sixteen, can’t drive, no public transit) are quite isolated; malls are securitized and unwelcoming, so they gather on social media—rather than blaming social media, ask why adults created no other spaces for kids.

His other method is “follow the money,” the source of his very negative view of contemporary digital tech. He doesn’t oppose the technology itself; the problem is that “our” social media is Meta, X—corporations profiting from anxiety, anger, and fear—which political-economy analysis can expose (read the earnings reports, how they address shareholders). Documents show Meta long knew young women suffer from how women are represented on its platforms yet did nothing, because it drives viral traffic. He cited Weibo’s “skinny challenges” (hiding your waist behind A4 paper, lining coins on your collarbone, hiding your knees behind an iPhone): posts earn huge followings while creating a toxic environment. But Instagram didn’t “invent” the obsession with thinness—advertising is the bigger culprit, and advertising mines cultural tropes already present in society; those ultimately responsible are “usually not the most visible ones.”

Social Network Analysis: Nodes, Ties, and “Centrality”

Anthropologists once mapped networks “by hand” (asking people how many friends they had, then asking those friends), and ethics alone made this very hard. The concepts come into their own with masses of mutually-referencing digital data: third-party tools can extract which accounts an account links to, expanding layer by layer. He recommends the open-source tool Gephi, which auto-visualizes from an Excel sheet or CSV.

The two core concepts are node and tie (edge); ties can be directionless, unidirectional, or bidirectional. Then come degree (how many ties a node has) and centrality. Three common centralities: (1) degree centrality—the most-connected node, the most direct but rarely truly most important; (2) betweenness centrality—concerned with “bridges,” which node sits between the most others (not so important on social media, but decisive in transport: one bridge connecting two halves of a city may be its most important road); (3) eigenvector centrality (Google’s PageRank)—not just how many links you have, but how many links the nodes linking to you have (twenty followers who are all Obama-level may make you more important).

The professor stressed these graphs are full of “choices”: sizing nodes by value, using “gravity” to highlight clusters—but “it’s all made up,” a social abstraction, not the situation on the ground. He colors Party nodes red, but showing a “big block of red” to an anti-communist European already presets a value judgment. Every visualization demands the question: am I guiding others’ understanding inaccurately, and am I transparent enough? Using the London Underground, most intuitively say King’s Cross matters most—and these high-betweenness stations were exactly the 2000s terror-attack targets; a paper argued the terrorists “must have computed these numbers,” but he thinks one glance at the map suffices, no calculation needed. He also reminded that the numbers only mean something compared within a single network and have no unit in themselves. Further reading: John Scott’s book on social network analysis.

Digital Bias: Search Engines, Geolocation, and “We’re All Biased”

The professor first clarified: we are all necessarily biased—unable to distinguish things by value, we couldn’t function; and the sheer volume of data forces a filtering mechanism. Search engines “are supposed to be biased”; the real questions are whether the bias is useful and fair. Results are almost always personalized (log into Google/Gmail and your search differs completely from anyone else’s), so “the standard Google search” doesn’t exist. This is where digital research collides with reproducibility: you can use a VPN, stay logged out, delete history, and run a clean “research browser” to stay neutral, but that creates an artificial situation “no one actually searches that way.”

Bias reveals things through: (1) materiality—which computer or browser you use, where your IP is. Some YouTube videos play in the Netherlands but need a VPN in copyright-strict Germany, all within a supposedly unified EU; illegal downloading can cost thousands in Germany but likely goes unpunished in the Netherlands or Taiwan. (2) Social bias—even without (or after deleting) a Facebook account, browsing sites with a “Like” button lets Meta track you. A famous case: a journalist testing Facebook’s political bias began liking everything and was quickly steered into fascist/Nazi content, which was also pushed onto his friends’ feeds; Facebook eventually told him to stop because he was “breaking the algorithm”—concluding not “our algorithm privileges fascist content” but “you did something wrong.” (3) Subjective bias—we feed the system our preferences hourly, and its whole point is to recommend “more of what you already like,” which is why targeted advertising works. He recommends the book If Then, on 1960s “Simulmatics”—a precursor to Cambridge Analytica, a “data broker” selling your data points to advertisers and political actors.

A demonstration: feeding “Tiananmen” into Google and Baidu. Google returns masses of 1989 information (about 60% of results relate to the protests and massacre)—itself a bias, since the system “assumes” people searching Tiananmen also sought these before; Baidu has no “Tank Man,” no protests—here the bias is “less about the algorithm than state intervention” (the state works with Baidu to adjust filters). DeepSeek is similar: ask about Tibet or Taiwan and you get “what the Party put in,” but it answers non-political questions quite well. He recounted that when his wife asked “can I buy Taiwanese calligraphy supplies in Beijing,” the system suddenly replied in English that “Taiwan is part of China”—though she only asked where to buy paint.

He also discussed “monopoly/ecosystem bias” (top results are usually the company’s own or endorsed content) and “banal bias”—borrowing Billig’s “banal nationalism”: like the flag in a classroom no one notices, it is precisely “normal and backgrounded” that makes it influential. Searching “Diaoyu” on Baidu pops up a weather app and flights (because it’s geocoded as a “real place”), with autocomplete ensuring you know it means “China’s Diaoyu,” even classifying it as an “urban district” of Fujian; for users long told “Diaoyutai is a Chinese place with weather,” it all “naturally lines up.” On Hudong Baike, registration asks which province you’re from and presets that province’s landmark as your avatar, so anyone not from these provinces becomes “someone from across the sea.”

Another bias produced by “looping behavior”: studying Nanjing and Diaoyu topics, he found a site that should be “military history” stuffed with pornographic and sensational content (uncensored in strict China, making him suspect “money changed hands somewhere”). Behind it is a self-fulfilling loop: post porn → sensationalism “works,” people click → the dashboard shows porn gets the most clicks → so post more to earn ad revenue. The same logic appears under a Nanjing Massacre report, where a hateful comment is pushed to the top for having “the most engagement.” So searching “Japan” on the Chinese web yields two extremes: sushi, kawaii, and “ultra-thin girls,” or hatred, the far right, and Yasukuni visits. To be heard you must pick one register—yet behind that register hide all kinds of left and right positions; people have simply internalized that “to be heard you must use this language.” This in turn shapes how the Party itself talks about Japan: hard to make friendly gestures unless it simultaneously signals to nationalists “we are a tough China.”

Representation and “Influencer Urbanism”: Hotspots, Selfies, and Cookie-Cutter “Individuality”

The professor opened with his own photo of the Wukang Mansion in Shanghai: twenty years ago almost empty, now packed with selfie-stick crowds and surrounded by guards (mainly because crowds risk injury, not protest). Such scenes, driven by Xiaohongshu/TikTok hotspots, have become a “formula”: the visual grammar of self-presentation is always the same, repeated endlessly. Interviewing Fudan colleagues: during cherry-blossom season the campus is overrun, and entering many Chinese universities now requires “showing ID”—which, he was told, is not a pandemic legacy but something “the students themselves” want, fed up with tourists flooding the campus.

He finds this “bottom-up, middle-class, conservative” incentive structure intriguing: people themselves call for “guards, control this place,” and the Party happily “goes with the flow”—something it already wanted now becomes “at the people’s request,” and even cadres probably sincerely believe they’re doing “the right thing the people want.” This is a case of discourse, mechanism, architecture, and lived experience “aligning”; you needn’t go to China to see it (Paris and Amsterdam are also barely walkable from tourist overload). Using airport security as an example of “how discourse embeds into technology and daily practice and becomes hard to reverse”: post-9/11 body scanners and the “no liquids over 100 ml” rule have little good reason, yet sustain a whole lobbying and mini-bottle industry—security is a fascinating research subject; ask “who actually benefits.”

He doesn’t want to single out Xiaohongshu—Instagram, London, Paris are the same. The “Selfie City” project of about fifteen years ago collected thousands of selfies, already an archive where “everyone looks the same”: identical gestures, angles, grammar—supposedly celebrating individuality, yet showing none. For Xiaohongshu and the “Wanghong” phenomenon he recommends Calvin Morris and team’s research—the point is not just individuals but “how cities become influencer hotspots.” He cites Düsseldorf, Germany: a city he finds “unremarkable” that, through its Little Tokyo, K-pop, and bubble-tea shops and restaurants opened by Chinese migrants, developed its own “influencer urbanism” and markets itself that way. So who holds “influence”—Xiaohongshu, the tourists, the entrepreneurs, or the city government? To some degree “all of the above.” Further reading: Guy Debord’s Society of the Spectacle, who noted as early as the 1960s–70s that we increasingly fetishize “the image that comes with the commodity”; as Mad Men shows, advertising no longer sells “it cleans well” but sells identity markers like “this is what a good wife uses.”

“Xi Jinping’s App”: Xuexi Qiangguo and Actor-Network Analysis

Closing with the “Xuexi Qiangguo” app. Its most interesting feature is that for some it is “mandatory”—unlike ordinary social media, it is “pushed” onto masses of users, many of whom just keep it installed. Why? Xi read much theory, wrote down his socialist vision, and amended the constitution to enshrine related principles, so the app offers an “easier entry” into this “new nationalism.” It is indeed a good way to learn “what’s in these books” (like the “Bible” on a hotel nightstand). English media called it the “Little Red App” (playing on “Mao had the Little Red Book, Xi has the Little Red App”), but it’s more of a “super-app” (like WeChat): you can watch almost any CCTV programming (many use it to binge dramas), plus skills and language learning. It’s well-designed because “they hired Alibaba”—without paying, you can’t get good IT talent.

He cited Leiden colleague Rogier Creemers’ “strategic nexus”: a “mutually understood shared interest” between private firms and China’s leadership. This is why he dislikes focusing only on symbols like the “red scarf”—lots of money is usually involved, and “pointing out the Communist Party component” is not the most interesting aspect. Portraying the state as “a monolith against the public” is common, but in reality you see extensive collaboration with commercial interests, often “gamified” (badges, rankings, XP)—scholars call this “gamification from above.”

The professor used this to demonstrate drawing networks “qualitatively”—not the earlier network analysis, but Latour’s “actor-network,” clarifying “who did what to whom, and what technical elements exist.” Latour holds that “objects can have agency,” which many social scientists dispute; he understands the concern (in writing, find human actors and use active verbs), but denying that “organizations can have agency” is also strange. To go finer, open the “Alibaba” box: what are its finance department, design department, and internal “Party committee” doing? At the simplest level you could just draw “platform influences users,” but if unsatisfied with treating the platform as one box, split it into: content and services (Xinhua, People’s Daily material), the “points mechanism” (earning XP through interaction), and the “leaderboard.” One key point: some “work units” actually folded leaderboard scores into annual reviews, tying bonuses and raises to “how well you performed on the app”—not a central-Party requirement, later abolished, but for a time people really were evaluated by their app performance.

He then introduced the “functional resonance analysis model (FRAM),” borrowed from psychology: instead of drawing a box for each “actor,” draw one for each “activity/function,” each shaped by six things (time, control, preconditions, resources, input), with input transformed into output. Political scientists should find this familiar (policy processes are often drawn as “input → output”), and it fits AI and social media especially well as things “literally mathematical functions.” Treating one process’s “output” as another’s “input” reveals “looping behavior”—once an output is “tuned” to a value, it can trigger unpredictable chain effects until the whole thing suddenly blows up. In his research on digital nationalism he argues nationalism is itself such a “function,” and the Party can’t reliably manage it—sometimes it “just erupts”—precisely because the process has so many feedback loops.

In Xuexi Qiangguo, these loops explain why people “game the system”: leaving CCTV videos playing in the background while doing other things, just to farm points, climb the leaderboard, and earn better bonuses—entirely contrary to the system’s intent. This is also why the Party later “called a halt”: too many distortions emerged, alienating people from their work; the app meant to make work more efficient instead led many to go through the motions because “it felt wrong.” He says this is what many now slowly realize about AI: do too much with generative AI and work becomes soulless—can you imagine a future job that is just “feed it into a program, get output, copy-paste”? The Party’s response was “very CCP”: admit the method didn’t work and dial it back—keep some uses, drop others—because “our social governance doesn’t actually work that way.”

This becomes a window onto where “digital governance” and “digital capitalism” meet: behind it is building Taylorism’s efficiency and optimization logic into how the Party operates. Frederick Taylor—the man with a stopwatch beside the conveyor belt, trying to shave off milliseconds—held that an “interlocking” system achieves perfect efficiency, and that all of society should run this way. This explains why the Party (though Taylor was a radical capitalist) rather admires him: they imagine “systems” the same way—build a good system that nudges all incentives in the right direction and you “needn’t force anyone to do anything.” Hence we should examine the feedback loops and resonance in digital systems, and ask: is the human still in the loop? At what point does the system develop a “self-running, no-input-needed” logic and spin out of control—especially when it comes to AI?

Note taken by Peter Tkach