1) People Live in Their Imagined World

The Industrial Genesis of Mass Communication

Before the end of the 19th century, various forms of "communication" certainly had long existed among humanity, but that was not yet Communication Studies in the modern sense. Religious proselytizing, royal edicts, the spreading of rumors, and oral transmission always existed, yet they lacked a key prerequisite: a "mass influence mechanism" that could be operated at scale, reused, and present a statistically predictable effect. What the Industrial Revolution changed was not merely the mode of production, but the relationship between information and social structure. Urbanization compressed individuals, who were originally dispersed within acquaintance networks, into highly dense environments of strangers living together long-term. Newspapers, the telegraph, and cheap printing technology allowed information, for the first time, to synchronously cover populations numbering in the millions who did not know each other. This change did not simply mean "communication is faster," but introduced an entirely new problem space: who is controlling the production and distribution of information, who is being influenced, and why a single piece of information can trigger highly consistent emotions and reactions within a group.

The Shift from Context to Variable

Previous forms of communication were highly dependent on specific contexts for their influence. Whether a religious sermon was effective depended on the personal prestige of the clergy, the sanctity of the ritual environment, and the faith background of the listeners; whether a royal edict could be implemented relied on violence, hierarchy, and local power structures rather than persuasion itself; the diffusion paths of rumors were highly unstable, the speed uncontrollable, and the content constantly deformed during the transmission process. These forms were not without influence, but they were difficult to deconstruct into stable variables, and difficult to verify repeatedly across different scenes. You could not systematically answer: "If I use the same information structure to speak again in another city to another group of strangers, will it produce a similar effect?" Under conditions lacking this reproducibility, theorization was almost impossible to establish.

The Emergence of the "Mass" and the Independent Information

The change brought by industrialized society lies in the fact that the city was not just an expansion of population scale, but the first time strangers were continuously connected together at the same time and in the same information space. In rural society, people understood the world primarily through acquaintance relationships; in the city, a vast number of individuals who did not know each other would read the same newspaper, see the same poster, and be enraged or appeased by the same news on the same day. It was precisely under these conditions that an entirely new object began to emerge—the "Mass." It is neither a community nor an organization, but a collection of strangers that can be triggered synchronously at the psychological level.

The true significance of newspapers, the telegraph, and cheap printing was also not just improving communication efficiency, but making information exist independently of specific interpersonal relationships for the first time. A piece of news does not require you to know the journalist, nor does it require you to know other readers, yet it can influence thousands upon thousands of people at the same time. Under these conditions, communication began to present a certain regularity that could be observed and compared: similar information structures, narrative frames, and emotional cues often triggered similar group reactions in different locations. Fear, anger, nationalistic sentiment, and moral panics began to exhibit inter-individual synchronicity. It was precisely within this structural change that communication, for the first time, became an object that must be systematically studied.

The Birth of Communication Studies as Risk Management

The problems that appeared subsequently were not abstract philosophical puzzles, but realistic problems directly associated with governance, commerce, and war. If the masses can be triggered simultaneously, then who is setting the trigger conditions? Why are certain narratives able to diffuse rapidly while others are ignored? Why do the same facts, through different modes of presentation, trigger diametrically opposed social reactions? Communication Studies gradually took shape in the process of responding to these questions. Its original intention was not "to let people understand each other better," but to understand and manage an entirely new social risk.

In this sense, the historical starting point of Communication Studies lies not in "how long communication has existed," but in society for the first time clearly realizing: public opinion can be systematically shaped, and if this power is not understood, it might in turn threaten order itself. From this moment on, communication was no longer just a subordinate phenomenon of culture, but became a force that needs to be researched, modeled, and controlled. Communication Studies was thus born, and from the very beginning, it was closely linked with power and fear. This also explains why the starting point of Communication Studies is not romantic. It did not originate from "humanity finally wanting to understand each other," but originated from the deep anxiety of the state, capital, and ruling blocs toward the out-of-control masses. Strikes, riots, revolutions, and financial panics were repeatedly discovered to be highly correlated with information diffusion. Information was no longer just reflecting reality, but began to participate in manufacturing reality in a structural sense.

Edward Bernays and the Crystallization of Opinion

In this context, the appearance of Edward Bernays was no accident. Bernays was not a simple theorist, but a person who directly entered the sites of national and corporate opinion operations. He once participated in the U.S. government’s opinion mobilization system during World War I, and later, as a public relations consultant, participated in shaping the cognitive environment for external political events in the United States, including creating public opinion legitimacy for specific political actions regarding the Guatemala issue. His significance lay not in "one man overthrew a regime," but in demonstrating how the shaping of opinion became a part of modern power operations. His publication of Crystallizing Public Opinion in 1923 possesses clear significance as a historical watershed. This book was almost the first time public relations, as a profession, provided a systematic theoretical explanation for itself. In it, Bernays explicitly denied that public opinion is generated naturally, but proposed that public opinion needs to be "crystallized." The so-called crystallization is not fabricating facts, but organizing, focusing, and orienting dispersed, vague, and emotional social attitudes. In this book, he still maintained a technician-like restrained posture, describing public relations as professional labor that coordinates social relations, striving to maintain its neutrality and public interest orientation. This was also his mildest work.

Propaganda and the Engineering of Consent

Propaganda, published in 1928, significantly changed the tone. If the previous book was still explaining "what we are doing," this one directly admitted "this is the reality of democratic society." In the book, Bernays openly pointed out that modern democracy cannot exist without propaganda; the true problem lies not in whether there is manipulation, but in who is to manipulate and how to manipulate. He rationalized guidance and manipulation as necessary conditions for the operation of democracy. The danger of this book lies not in its moral stance, but in its high degree of frankness: communication here is explicitly understood as a technology of power, and is no longer an art of communication. The Engineering of Consent, published after the war, further pushed in this direction. At this stage, Bernays' focus was no longer on justifying legitimacy, but on systematic operation. The concept of the "Engineering of Consent" itself clearly indicates that public attitudes are regarded as objects that can be designed, tested, adjusted, and maintained long-term. He analogized the shaping of public opinion to engineering projects, emphasizing process, expert collaboration, and long-term planning. This marked communication’s upgrade from a strategic behavior to an institutional capability, and foreshadowed the comprehensive appearance of later political consultant systems, campaign machines, and corporate brand systems.

Walter Lippmann and the Pseudo-Environment

If Bernays represents execution and operation, then Walter Lippmann represents cold and sharp diagnosis. Lippmann was not a technical communication scholar, but a person standing at the front lines of news and political practice, observing how modern society loses stability within information. He was born in 1889, grew up in the era of rapid American industrialization and highly concentrated media, received systematic philosophical and political training, and long served as a political commentator for mainstream media. This point is crucial: he did not abstractly imagine the "public" in a study, but in real news production, political maneuvering, and war narratives, repeatedly saw how the public was led, mobilized, and even misled by information. World War I possessed decisive significance for him. During the war, the U.S. government mobilized public opinion on a large scale, manufacturing enemy-vs-us narratives, emotional consensus, and moral legitimacy. Lippmann saw with his own eyes how a democratic society that considered itself rational was led by a unified narrative in an extremely short time. This experience did not make him excited, but made him highly vigilant. From this, he realized that the complexity of modern society had already far exceeded the ability of ordinary individuals to understand the world through direct experience.

In Public Opinion, published in 1922, he proposed the "Pseudo-Environment" concept that has been repeatedly cited since. This concept was not accusing the media of lying, nor was it belittling public intelligence, but was pointing out a structural fact: reality itself has already become too large to be directly experienced by individuals. Politics, economy, war, finance, and international relations mostly happen outside the radius of individual life, but people still must make judgments on these affairs. Under these conditions, people can only live through a "mediated reality." News headlines, pictures, statistical figures, story frames, and moral labels are not reality itself, but the versions of reality after being compressed, trimmed, and formatted. People do not act according to the world itself, but act according to the imagination of the world, and this imagination is continuously supplied by the media and constantly reinforced. Lippmann called this psychological layer the "Pseudo-Environment," emphasizing that it is not equivalent to reality, yet in practice, it replaces the status of reality.

Stereotypes and the Logic of Power

The key lies in the fact that this process does not rely on conspiracy. Even if all journalists were absolutely honest, the pseudo-environment would still be unavoidable, because the form of news itself can only present fragments of reality, while public affairs require systematic understanding. This is a type of cognitive compression, rather than simple distortion. On this basis, Lippmann gave a cold but important judgment: stereotypes are not moral defects, but cognitive tools. Humans in a complex world must first categorize before they can quickly orient their stance. The problem lies not in why humans use stereotypes, but in who provides these stereotypes, how they are reinforced, and what kind of power structure they serve. Therefore, the public is not stupid, but is rationally adapting to a world that cannot be completely understood. The true risk to democracy lies not in the public lacking rationality, but in the institution pretending the public is making judgments about the "real world." On this point, the divergence between Bernays and Lippmann becomes clear. Lippmann cared about what this structural rupture meant, while Bernays cared about how power should operate under the prerequisite that the rupture is uneradicable. Bernays did not misread Lippmann, but translated his diagnosis into operational logic. In this sense, Bernays is Lippmann’s most faithful and most dangerous successor. Lippmann’s contribution lies not in providing solutions, but in continuously deconstructing the illusions of modern society: public rationality, media transparency, democracy operating naturally, and international order based on morality. Once these illusions are punctured, people will feel uneasy, but the institution may become more honest.

2) Communication in the Age of TikTok

The Evolution of Communication Models

If we pull Communication Studies from after World War I all the way to today, its evolution is not a history of knowledge following linear progress, but a process of being forcibly rerouted time and again by power needs, technological conditions, and media structures. Every time a new form of communication appears, it forces Communication Studies to rewrite what it is actually researching, whom it serves, and what kind of object it defaultingly treats humans as. From the end of WWI to WWII and the early Cold War, Communication Studies gradually formed a highly "engineered" research orientation within the American context. This stage did not possess a unified theory at the start, but under the push of needs for propaganda practice, public opinion management, and attitude research, it gradually focused on a core problem: whether information can be effectively deployed and produce predictable effects in a statistical sense. Propaganda analysis, attitude change research, and effect measurement constituted the most important research centers of gravity during this period. The masses were often assumed to be objects that could be stimulated, influenced, and regulated; the media was understood as a relatively neutral transmission pipe; and communication was deconstructed into separable, manipulable input and output stages. This mode of understanding, during WWII and the early Cold War, was highly compatible with national propaganda, psychological warfare, and domestic governance needs, and was thus continuously reinforced.

Harold Lasswell and the Focus on Effect

Within this research orientation, Harold Lasswell’s contribution lay not in proposing a specific theoretical model, but in the fact that he set a basic problem framework for communication research. The analysis formula "Who, through what channel, to whom, said what, and produced what effect" proposed in 1948 explicitly deconstructed communication into a process with "effect" as the endpoint. In this framework, communication is first a relationship of action, rather than a process of interaction; it cares not whether meaning is commonly understood, but whether influence happens and whether it is effective. In this setting, the audience only appears with the identity of the "one being acted upon," and its significance of existence is manifested in whether a measurable change ultimately occurs. Understanding is not the core variable; change is. Once communication is modeled this way, dialogue has already been marginalized at the research level. Because dialogue implies that both sides have the possibility of being changed in the interaction, while in this effect-oriented framework, change is defaultingly allowed to happen only unidirectionally.

The deeper problem lies in the fact that this model implicitly assumes meaning has already been completed before communication begins. Meaning is treated as an object that can be encoded, transmitted, and delivered, rather than a process that needs to be negotiated in interaction, depends on context, and may undergo deviation. As long as the sending end chooses the right symbols, channels, and rhythm, and triggers the expected reaction in the audience, the communication is regarded as a success. Negotiation, ambiguity, and uncertainty are not the research focus, but are variables that need to be compressed and managed. Therefore, the concept of "effect" itself carries strong normativity. Communication is understood as a type of intervention behavior with a predetermined goal, rather than an open public process. Under this logic, as long as the goal is regarded as legitimate, whether the audience truly understands is not important. This line of thought has extremely high operability in political propaganda, advertising, and psychological warfare, but it is not suitable for explaining any social process that requires public discussion, negotiation of meaning, and collective judgment.

The Medium is the Message

Entering the mid-20th century, this assumption encountered systematic challenge for the first time. Election research since the 1940s gradually showed that the masses are not completely passive, and information effects do not always happen directly. Interpersonal networks, opinion leaders, and social structures played significant mediating and buffering roles in the communication process. The research orientation subsequently summarized as "Limited Effects" was not denying the influence of communication, but pointing out that influence is not linear, instantaneous, and omnipotent. Communication Studies appeared more "mild" and more socialized on the surface, but its underlying goal did not undergo a fundamental transformation: it was not giving up on influence, but achieving influence through more indirect and structural modes. Communication shifted from "acting directly upon you" to "acting upon you through the people and environment you trust."

Almost during the same period, the medium's ontology began to enter the field of vision of Communication Studies. Marshall McLuhan proposed "The Medium is the Message" in the 1960s, constituting a fundamental challenge to existing communication assumptions. What he focused on was not specific content, but how the form of the medium itself acts as an environment, reshaping the human perceptual structure, temporal experience, and mode of social organization. The medium was no longer regarded as a neutral pipe, but was understood as a force that actively shapes social relations. Before McLuhan, Communication Studies often defaultingly assumed: the same content transmitted through different media is essentially the same information, only with different efficiency and coverage. McLuhan denied this prerequisite. He believed that different media are not "different packaging for the same information," but introduce entirely different perceptual structures. What truly, long-term, deeply, and irreversibly changes society is not what you said in the medium, but how the medium itself changes the way humans understand the world and relate to each other.

In this sense, "The Medium is the Message" is not denying the importance of content, but pointing out: the most critical influence brought by the medium happens before the content is understood. By changing the distribution of attention, the sense of rhythm, the mode of participation, and the sensory proportions, the medium in advance shapes how a person can understand and to what extent they can understand. By the time a person begins to judge "whether to believe" certain content, the perceptual structure has often already been set. This perspective makes McLuhan’s judgment appear particularly forward-looking in today's platform environment. When the medium is no longer just a tool, but becomes an environment that humans cannot escape, the power of communication no longer primarily manifests in specific persuasive content, but manifests in the medium's form itself and its ability to rewrite lifestyles.

Platform Power and Rhythmic Control

Entering the internet and platform era, Communication Studies has not completely abandoned existing theories, but its core assumptions have begun to clearly fail. Communication is no longer linear, but is networked, recursive, and highly feedback-driven. Theories such as agenda-setting, framing, and the spiral of silence did not just appear in this period, but they must be re-understood under new conditions: in an environment where platforms control visibility, ranking, and distribution, who possesses communication power is no longer a simple content problem. In the communication environment represented by short-video platforms, the research focus has further shifted. What communication faces is no longer just "how information persuades people," but "how attention is continuously shaped." Content is not actively chosen, but is systematically pushed; the subject of communication becomes blurred, and algorithms assume the role of distribution and screening, yet lack visibility and accountability. The unit of communication has shifted from "information" to "behavioral feedback": dwell time, re-watching, and interaction are more important than understanding itself.

In this structure, the communication goal shifts from attitude change to habit formation. The platform does not need you to believe in a certain stable explanatory framework; it only needs you to form a predictable path of usage. Power no longer primarily achieves itself through discourse consistency, but through rhythmic control—controlling when you watch, how long you watch, and in what emotional state you watch. From this, the main battlefield of communication slides from public reasoning toward attention management, and from meaning production toward time occupation. Different and even conflicting narratives can coexist in the same information stream, as long as they can all grab attention. Ideological consistency is no longer a necessary condition; occupying cognitive bandwidth itself is sufficient to maintain the operation of the system.

When communication operates through rhythm rather than meaning, public discussion then faces a structural dilemma. People are not conquered by a certain viewpoint, but are long-term reshaped by a mode of information intake. Viewpoints can be frequently replaced, but once attention patterns, patience thresholds, and participation rhythms are fixed, the range of what can be understood and seriously thought about will subsequently contract. Ultimately, occupying time will transform into shaping the boundaries of what is thinkable, and communication power subsequently completes the transfer from "persuading people what to believe" to "deciding how people exist within the media environment."

3) The Filtering System of TikTok

Content as a Behavioral Probe

Continuing with deeper analysis, one can understand TikTok as a real-time behavioral screening system, rather than just a "content platform." The key here lies not in how the platform evaluates the content itself, but in how it judges whether the content can stably trigger a certain type of user behavior. In this framework, video is closer to a probe. What it is used to test is: within a specific user structure, whether a predictable, repeatable behavioral reaction will appear. What the system cares about is not what viewpoint you expressed, but whether these expressions can statistically form a stable behavioral signal.

The Logic of Pressure Testing and Distribution

From the externally observable distribution mode, a new video on TikTok will usually first enter a user sample that is small in scale, but not random. This sample is more likely to be composed of users who have shown explicit reactions to similar content or stimulus types. The focus of this stage is not "recommendation," but is more like a pressure test: the system, through very early behavioral indicators, judges whether it is worth continuing to invest more attention resources. In this stage, what the system cares about are often the most basic, cold-start signals, for example, whether the user pauses at the moment they swipe to the video, or whether a brief continuation of watching appears. This type of signal is not equivalent to complete watching or interaction, but they decide whether the video possesses the minimum qualification to enter the next round of testing. If these early indicators are significantly lower than the baseline level, more complex data subsequently will no longer be focused on for examination.

From the distribution results, this process is closer to layer-by-layer clearing (release), rather than a comprehensive score. The system does not need to judge whether a video is "overall good," but continuously evaluates whether it satisfies the conditions to enter the next stage of testing. Once key behavioral signals are insufficient, distribution often will stop rapidly, rather than gradually weakening. This can explain why some videos, under the condition that early data is acceptable, will still suddenly lose diffusion. One direct consequence of this mechanism is that the weight of the early sample is significantly higher than later performance. The fate of a video depends more on the reaction of the initial small part of users, rather than "proving itself" slowly through time. Once the system judges that the stimulus is not statistically valuable for expansion, even if potential quality exists subsequently, it is very difficult to obtain new display opportunities.

Differentiation and Predictable Stimuli

After passing the preliminary test, the distribution logic will not simply expand the scale, but is more likely to turn toward differentiated testing. That is to say, the video will be deployed to user groups with different structures and higher thresholds, to observe whether the reaction still holds. What the system cares about at this time is not the quantity of reactions, but the transferability of the reaction across different user structures. If a certain stimulus is only effective among low-threshold populations, but rapidly fails among more picky or higher-behavior-cost populations, the system often will identify it as a locally established mode, thereby limiting its diffusion range. This type of content will not be cleared, but usually will be stably restricted within a narrow distribution band. Conversely, when content, under the condition that the scale is not large, can stably trigger high-cost behaviors, such as comments or shares with clear stances, the system often will treat it as a signal with explicit positioning value. This does not mean the content is "better," but means it has formed a stronger binding relationship with a certain type of behavior. For the platform, this type of stimulus structure that can be reused and predicted often possesses more engineering value than "overall acceptable" content.

At this level, the core the system cares about is not the content semantics itself, but whether the content stably triggers a certain state change. The platform does not need to completely understand the viewpoint logic; it only needs to identify: which type of user, under what kind of stimulus, transfers from a pre-behavioral state to another state. Emotion here is not the goal, but an intermediate variable that can be recorded and utilized. Once this "content—state transition—behavioral reaction" mapping relationship tends toward stability, the video is no longer treated as an object to be evaluated, but is closer to a tool with a known use. Thereafter, it will be continuously deployed to the population most likely to complete that behavior, and will rarely again be used to explore entirely different user structures. This phenomenon is not naturally formed echo chambers, but is closer to the result of precision deployment.

Account Stability and Rhythmic Logic

On an even longer time scale, the focus of the system's evaluation will gradually shift from a single piece of content toward the account as a whole. The account will be regarded as a relatively stable source of stimuli, while a single video is a sample of that source. The problem the system cares about is no longer "how is this video," but "what kind of behavioral reaction does this account usually manufacture." When the account statistically exhibits a stable output mode, distribution then will rely more on prediction rather than re-testing. Content that conforms to existing modes often starts faster and has higher fault tolerance; while content that clearly deviates from existing stimulus types may obtain fewer testing opportunities because the uncertainty is too high. This is also why account transformation in practice often has extremely high costs; it is not that the audience cannot accept it, but that the system needs to pay additional learning costs.

Under this logic, the platform does not naturally prefer so-called "positive energy" or "negative energy," but prefers content whose reaction curve is highly predictable. Accounts with clear stances, stable structures, and consistent output modes are more easily trusted by the system; while expressions that are vague, wavering, and frequently testing different stimulus types are more difficult to be modeled and amplified. The so-called reaction curve can be understood as a time sequence that can be recorded: at which second the user pauses, at what time they generate emotion, and at what time they take action. The system does not judge the value orientation of this curve, but judges whether it is sufficiently stable, and whether it can be accurately predicted the next time. Stability, rather than morality or the viewpoint itself, is the most core foundation of trust at the engineering level. From this perspective, TikTok is closer to a screener, rather than a stage. Creators are not unidirectionally "expressing," but are continuously undergoing the system's inspection of their stimulus stability. Content that can continuously diffuse is often not completely consciously catering, but has unconsciously formed a high degree of compatibility with this set of screening mechanisms.