On April 29, 2017, the Turkish government blocked access to the website Wikipedia, allegedly because Wikipedia portrayed Turkey as a sponsor of ISIS and Al-Qaeda. Although Wikipedia appealed the ban as a violation of free speech with both the Constitutional Court and the European Court of Human Rights, the site went dark immediately and has remained so for over two-and-a-half years.

In December of 2019, the Turkish Constitutional Court finally ruled that the ban was unconstitutional, apparently because it wanted to preempt a negative ruling from the European Court.  In early 2020, Turkish internet users finally regained access to Wikipedia. While this may seem like a victory for the Turkish Constitution’s protection of free speech, it is notable that the government was able to maintain a blatant constitutional violation for over two years.

Turkey is not the only country that is able to circumvent its constitution’s free speech protections. Our research reveals that it is often remarkably easy for governments to violate free speech rights. For instance, by 2010, 186 countries protected free speech in their constitution. But, according to the widely used data on rights violations, only 41 countries did not place any restrictions on that right in practice. In our research on the topic, we have found no evidence that countries that add the right to speech to their constitution are any more likely to stop curtailing free speech than those without constitutional free speech protections.

Two years ago, we used the Turkish free speech ban as an opportunity to explore why leaders are able get away with free speech violations, even when their constitution guarantees free speech. In our new book, How Constitutional Rights Matter, we argue that many governments are tempted to violate constitutional rights. Yet, they will refrain from rights violations when citizens oppose violations and are willing to punish governments that undermine the constitution, either by voting against them in the next election or through other forms of political mobilization. We used the Wikipedia ban—and the constitutional violation it represents—as an opportunity to explore whether citizens are willing to defend the constitution in the face of rights violations.  

In September 2017, we conducted a series of 1,335 face-to-face interviews in Turkey with a representative sample of their population. We first asked respondents whether they supported the ban. We then randomly told some citizens that the ban was a clear violation of the constitution, so that we could explore whether reminding citizens of the violation changes their level of support for the ban. At that time, the Constitutional Court had not yet ruled on the ban, so there was still uncertainty over the ban’s constitutionality.

Our findings do not bode well for constitutional rights protection. Notably, 93 percent of respondents said that the government should not violate the Constitution when asked the question generally. Yet, notwithstanding the high levels of support for the Constitution in the abstract, being told that the Wikipedia ban violated the Constitution did not decrease support for banning Wikipedia. Some 39 percent of respondents in the control group supported the ban, and this number actually increased to 47 percent for respondents that were told that the ban violated the constitution (although this difference is not statistically significant). Knowledge that the ban violated the constitution, then, did little to turn citizens against their government.

It is illuminating to further break down the results for the respondents that voted for President Erdogan’s AKP party. Among these respondents, the level of support for the ban was higher than for the population as a whole: some 66 supported the ban. Being told that the ban represents a constitutional violation did not cause these voters to change their support: it made them more likely to support it. Of the AKP supporters that were told the ban violations the constitution, 88 percent of respondents supported the ban. Thus, instead of becoming less supportive of their party’s policies, they actually became more supportive of it. Presumably, these AKP supporters took their government’s willingness to violate the Constitution as a signal that the underlying goal must be particularly important, which made them more supportive of the ban.

Our findings provide insight into that why it is that enforcing constitutional free speech protections is difficult. The case of the Turkish Wikipedia ban reveals that support for the constitution in general often turns out to meaningless in concrete cases. When presented with actual violations by their party, citizens support their party’s policies, not the constitution. In doing so, they essentially are giving the government carte blanche to pursue constitutional violations.

And while courts occasionally attempt to block such policies, a long line of research in political science shows that courts are usually cautious to go against the wishes of popular majorities. The attitude of the Turkish Constitutional Court illustrates this point: it simply refrained from ruling on the case until it was clear that the government was going to face a humiliating decision by the European Court of Human Rights.

These findings help illuminate the reason for our basic finding that adding individual rights to constitutions is not associated with better protection for those rights. For governments to be constrained by the constitution, they have to pay a political price for its violations. But around the world, citizens often support government actions that violate constitutional rights. 

The Turkish Wikipedia ban shows this logic in action.  It also reveals a potential lesson for other countries: preventing constitutional violations requires people to put the constitution ahead their partisan preferences. When people put their party over the constitution, the document is at risk of becoming a dead letter.

Note: Our book, How Constitutional Rights Matter, was recently published by Oxford University Press. This is the fifth in a series of blog posts on parts of the argument and evidence from the book.