Earlier this month, Chinese Internet authorities commemorated the 25th anniversary of the end of the Tiananmen Square protests with disruptions to Google services, heightened keyword blocking on social media, and other measures to restrict discussion of the event. As usual, some people evaded the censorship with coded images and text.

Using Internet memes (for example, tanks replaced with “big yellow ducks” in the iconic Tank Man photo), coded language (May 35, shorthand for June 4, 1989, the date of the crackdown on protesters), and seemingly innocent actions (such as wearing black shirts or singing a song from Les Miserables) to sneak criticism or protest past authorities has become a tradition in China. Netizens have developed a kind of sixth sense for which topics are sensitive, and a Northwestern study shows that microbloggers begin developing code words or “morphs” in the form of puns, metaphors, and homophones even before they are targeted by censors. In recent years, foreign observers (including myself) have celebrated Chinese Internet users for their sense of irony, dirty humor, and political criticism in the face of restrictions on freedom of expression.

But though these evasive tactics certainly allow citizens to express their feelings when more direct language might be blocked, it’s debatable how effective they really are at defeating Communist Party censorship. If the goal is communication rather than expression, then one must marvel at how thoroughly the authorities have on the whole managed to stifle the spread of dissent and unwanted organizing online.

At times, censors and citizens are engaged in a sort of cat-and-mouse game. As the censors glom onto and block the keywords that Internet users develop, the users must create more codes in order to stay ahead. The mouse’s hiding spots become smaller and harder to get to with each iteration as the codes grow more and more obscure.

In the case of the Tiananmen Square anniversary, after mostly transparent morphs like “May 35” and “8964” were exhausted, newer codes like “VIIV” (6/4 in Roman numerals) and “six quatre” (6/4 in French) were invented. Over time, these various keywords gradually splinter the topic into separate clusters, making it harder for people to easily find and collect information and eliminating any cohesive hashtag that large numbers could rally around. In time, even these semitransparent keywords are caught by authorities as they become too well-known, and soon activists are down to the slimmest of cracks—for example, codes like 2的6次方 (“two to the sixth power,” i.e. 64, for June 4), which itself is now occasionally censored.

This process also divides potential audiences into haves and have-nots: those with the knowledge necessary to decode an obscured post, and others who are left out. For any activist who hopes to spread a message, finding a meme or coded language that is both capable of circumventing censorship and intelligible to large numbers of people is increasingly difficult.

As Chinese journalist Yang Xiao has observed, the tendency of social media to insulate people from information they disagree with intensifies this effect. “Subtle linguistic tricks are too superficial for the well-informed and too sophisticated for those who just don’t care,” he writes. For Yang, coded language may offer its users a feeling of valiantly fighting the system when they are merely shouting into the void—or, at best, speaking to an inner circle that already shares their views.

Obviously, censorship and the subsequent ignorance it breeds hinders the effectiveness of these anticensorship tricks. For instance, the enforced silence of state media and China’s education system no doubt leaves many young people unaware of what happened on June 4, 1989. In a recent informal survey of 100 college students, only 15 recognized the Tank Man photo. Without a shared base of knowledge, what use are memes like the big yellow duck for reaching the masses?

But apathy perhaps plays an even bigger role. Tools that circumvent the Great Firewall to provide unfettered access to foreign news sources are readily available to anyone motivated enough to seek them out. Yet Helen Gao, a Beijinger who recently completed university studies, reports that few of her peers in China use these tools; most simply accept censorship without question. As she writes in a recent New York Times op-ed, apathy seems to be the reason that many Chinese university graduates are unfamiliar with dissidents like Chen Guangcheng or Ai Weiwei.

All this forces people working on sensitive topics in China to make hard choices: do you communicate directly, risking suppression and harm to yourself, or do you try to speak around censorship, limiting your potential audience? Do you seek to encourage those already in the know or to convert the unenlightened? Do you take your diluted message to the masses on Weibo and WeChat, or do you abandon it all and speak openly to peers who have opted to jump the Great Firewall?

Online censorship is about more than just the technical means of scrubbing information away. Its power to hinder effective communication can truly demoralize those seeking to oppose the status quo. If memes and morphs are to become a tool for organizing, activists must engage and inspire young people before they can even think of educating them about the things being hidden from them, much less rousing them to action. Only then will the censors and their targets be playing on equal terms.

Jason Q. Ng is a research fellow at the University of Toronto’s Citizen Lab and author of Blocked on Weibo, a book on Chinese social media. He tweets at @jasonqng.