Our new GLOBALCIT website is under construction. In the meantime, please use the current website as before.

How liberal are citizenship tests? - Sara Wallace Goodman: Lost and Found: An Empirical Foundation for Applying the ‘Liberal Test’

I thank Christian Joppke as well as fellow contributors for what has turned out to be a very interesting and relevant discussion. I find asking ‘how liberal are citizenship tests’ to be a worthwhile endeavor, particularly for understanding the context of integration and how states view newcomers. However, I am skeptical that this political theory question should be the starting point of inquiry. What is missing is a clearer picture of the wider practice of citizenship tests, civic integration courses, and language assessment. This empirical view of the range and diversity of civic practices must necessarily precede case selection for the application of a ‘liberal test’ and the subsequent wielding of metrics to determine which cases pass, as Joppke describes, the ‘threshold of the impermissible.’ By looking at the landscape of cases in Europe and beyond, comparativists and political theorists alike can ask necessarily different questions about tests but with adequately articulated scope conditions. To engage in Kostakopoulou’s o exercise of “zooming out” reveals that the most obvious violation of a test’s liberality in practice is not in its content (Michalowski) or existence (Carens), but it its purposeful connections to immigration policy and fusion with objectives of control.

“Zooming out” is required across three dimensions. First, we need to see the wider picture of which states have adopted citizenship tests. This practice extends far beyond the samples at the center of the forum debate. Second, we need to see what other requirements apart from civic knowledge complement citizenship tests, like language skills and demonstrating a commitment to a set of core values. Citizenship tests do not exist in vacuums. They are required among a configuration of material criteria, and these in combination determine whether the practice of citizenship is ‘illiberal’ and ‘discriminatory’ or ‘liberal’ and ‘inclusionary.’ Third, it is worthwhile to examine which states ask newcomers to demonstrate civic knowledge, language proficiency, and values at stages other than citizenship acquisition. One of the arguments for interpreting tests as illiberal is that they set a discriminatory double-standard between knowledge for naturalizing newcomers and knowledge for those with citizenship by birth. But another potential source of illiberalism is that it asks different standards at different stages of legal status, including entry and settlement. A wider, empirical view across these three dimensions is therefore imperative and prior to asking normative, theoretical questions.

This forum has applied the ‘liberal test’ to tests in the United States, Austria, Germany, the Netherlands and the UK. However, we also see citizenship tests that assess country or constitutional knowledge in places including Denmark, Estonia, Hungary, Latvia, Lithuania, Moldova, and Romania, [1] as well as further afield in Mexico and South Korea. Is it assumed or expected that tests in these other European and non-European countries violate liberalism in content, application, or outcome? Do we only expect Western European tests to be liberal, holding them up to a problematic double standard? Discussing only a few selected Western European countries and the US provides us with a biased view.

To select a case almost at random, Latvia has the same exemptions from naturalization tests as previously examined cases (obtaining a certain degree of education in Latvian, age, and disability). The cost of citizenship (28 Lats, or approximately €40) is far lower than the Netherlands (€230 for the test alone, €380 for the entire citizenship process, both of which are much lower than the $675 price-tag for the liberal exemplar of US citizenship). The exam assesses knowledge of principles of the constitution, the text of the national anthem, and history, asking content questions such as “What countries share Latvia’s land borders?” and a number of question on the country’s relationship with Soviet Russia. In 2008, almost one-third (28%) of applicants failed the language part of the exam and 18% failed the country knowledge component. [2] These figures are comparable to the 75% pass rate of the ‘Life in the UK’ test for citizenship. Finally, Latvia assesses national language at an average level—the B1 level of the Common European Framework of Reference for Languages. This is the same level of assessment as in Germany and the UK, but higher than that of the Netherlands (A1) and Austria (A2). Do these comparisons render the Latvian testas liberal as the Western European ones? Or at least equally subject to the same test of liberality? What are the scope conditions for liberal assessment? Should we not apply the same standard to new and old EU states?

Aside from civics tests, states also require the aforementioned language proficiency.. A far greater number of countries require sufficiency in a national language than do country knowledge in a citizenship test. Language is assessed through a requirement of certification of the completion of an exam or language education (Austria, Denmark, Estonia, Finland, Latvia, Moldova, Portugal, the UK), an oral interview (Croatia, the Czech Republic, France, Germany, Luxembourg, Malta, Slovakia, Spain, Turkey), or just a test itself (Iceland, Netherlands, Slovenia and, in the case of Hungary, indirectly through completion of the civics test). Practices in federated states like Switzerland and Germany necessarily vary, with some cantons/Länder requiring written tests and others oral interviews. The level of assessment also varies, but none assess beyond a B2 level of usage (which is required in Denmark), described as the “vantage” stage of “independent users”. Not one EU country requires language at the “proficiency” level (C1 or C2). In fact, Denmark has the highest standard of language assessment in the EU. All other countries require language (in descending order of difficulty) at the B1 “threshold” level for independent users (Czech Republic, Estonia, Finland, Germany, Hungary, Latvia, Spain, Switzerland, UK), the A2 “waystage” level for basic users (Austria, Iceland, Netherlands, Portugal) or at the most minimal standard of an A1 “breakthrough” level for basic users (France). And, finally, some countries do not have a language requirement at all (Ireland, Italy and Sweden). The consensus among the contributors to this debate rightly is that requiring language does not push citizenship tests across the threshold of discriminatory, illiberal practices. And to re-iterate Hansen’s words, “differential effects do not necessitate unfair unequal treatment.” But I think the more probing question is not whether language tests are liberal, but why, with the same variation in assessment and same expectation of learning and accumulated knowledge, is a language test safely within the bounds of liberalism while a country knowledge test teeters closer to the edge of illiberalism. Assuming that neither tests ‘inner dispositions’ and ‘morality’ (and I believe that it needs to be distilled from how a question is worded to determine whether it assesses replicated knowledge as opposed to acceptance of political values and attitudes), language and country knowledge seem equally fair conditions for sovereign states to require of potential members.

Third, turning to oaths and ceremonies, this is a requirement that can most precariously rub against the consensually illiberal practice of assessing morality and behavior, yet none of the contributors pays much attention to this aspect. Only the UK and the Netherlands have now a citizenship ceremony similar to the US model. Several require an oath or pledge of fidelity, which is sworn in front of a court, magistrate, or immigration official in contrast to the shared experience of a ceremony. These countries include Austria, the Czech Republic, Germany, Greece, Ireland, Italy, Lithuania, Romania and Slovakia. Other countries require a written commitment to values and principles, including Belgium, Denmark, Estonia, Finland, Latvia, Malta, and Spain. Can we hold oaths and value commitments to the same standard of tests? Applicants are asked to verbally commit to a set of principles that they do not need to hold personally. Just as you can reproduce knowledge, you can read a script. For example, naturalizing Americans swear to “absolutely and entirely renounce and abjure all allegiance and fidelity to any foreign prince, potentate, state, or sovereignty of whom or which I have heretofore been a subject or citizen” as an Oath of Allegiance to the United States during the US citizenship ceremony. Yet there is no attempt to prevent dual citizenship by enforcing renunciation of a previous one. And surely, during the British citizenship ceremony, a person can publically swear to be “faithful and bear true allegiance to Her Majesty Queen Elizabeth the Second, her Heirs and Successors” while holding privately republican views. I do not think that this subverts individual autonomy in a discriminatory way; instead, it detaches the end product (citizenship as a form of membership) from the means (naturalization as a ‘naturalizing’ process). This is certainly a problem, but not one of liberalism.

The final empirical dimension worth considering about civic tests before asking questions about how ‘fair’ they are is looking at where states require them. The real innovation of the European variant of civic testing—in contrast to the American model—is that it promotes ‘citizen’ values and skills not only among applicants for citizenship but also for various earlier gates of membership, including those where migrants seek permanent settlement and territorial admission. In fact, it is this extension of civic requirements to ostensibly non-traditional stages of civic status that raises major concerns and warrant special attention. Germany, Denmark, the Netherlands, Austria and the UK not only have civic integration requirements for citizenship but also for permanent residence (e.g., integration courses with tests) and pre-entry or immigration (with the exception of Austria, where there is currently only a language test for immigrant settlement). [3]

It is in this outward expansion of requirements where I think the assessment of liberalism is most relevant. By attaching such as ‘citizenship’ tests to the process of immigration and settlement, content that, in principle, promotes membership and inclusion becomes mixed up with immigration policy and ambitions of control. While states have autonomy in determining immigration policy, exclusionary rules that require knowledge for some groups and not for others are an explicit violation of equal treatment. Selectivity along grounds of race, religion, gender is universally regarded as illiberal in both immigration policy and, a fortiori, in naturalization policy. [4] Yet, by design or effect, pre-entry requirements do just that. By design, we see a pre-entry test in the Netherlands, for example, that exclusively targets certain nationalities over others (exempting citizens from the EU, Australia, Canada, Japan, New Zealand, and the United States). The Basic Civic Integration Exam (or immigration test) also explicitly targets religious leaders. By effect, the requirement that family-forming migrants are required to demonstrate knowledge or language before entry produces a clear bias against immigrants from Turkey and Morocco, which constitute the largest national groups that immigrate for family-formation or -unification purposes, as well as against female migrants, which in absolute numbers of entries are comparable to men, but make up the majority of family-based migrants. [5] This immigration test is onerous: as a condition for receiving a temporary residence permit, or MVV, a potential migrant must sit a 15-minute, 50-question oral Dutch language exam assessing listening and speaking skills as well as a separate Knowledge of Dutch Society examination, in which the applicant answers 30 questions on Dutch life and society. It is taken over the phone and administered at the Dutch embassy in the applicant’s country of origin. Yet it is not because of its difficulty that it is illiberal, it is because of the discriminatory effect it has on certain populations. Denmark is applying a similar immigration test. Germany and the UK (by 2011) only assess language at point of origin through a test, while France has the least difficult version by offering a cost-free interview to test the potential migrant’s expected level of French proficiency. [6]

In conclusion, a wider view of citizenship tests not only shows a diversity of countries that use such tests, but also that there are complementary material conditions for citizenship access as well as civic assessment at earlier stages which is perhaps the most clearly illiberal dimension of ‘citizenship tests’. With this fuller empirical foundation, a richer exploration of liberalism in immigration and citizenship policy becomes possible.

A final note on the United States as the primary example of a liberal citizenship test invoked by several contributions to this forum discussion: I find little resemblance between the US naturalization test and its European progeny in any way other than content. It is a 10-question civics exam where the applicant orally answers questions selected by an immigration officer from a total set of 100. Content ranges from “What did Martin Luther King, Jr. do?” to “Who is your U.S. Representative?” The applicant must answer 6 out of 10 questions correctly, and the test need not necessarily be conducted in English. Aspiring citizens can also take the exam multiple times if they fail. A 2008 redesign of the test content was for the purposes of shifting knowledge from the ‘trivial’ to the ‘useful.’ Moreover, this is the only civic barrier encountered by an immigrant from point of entry to citizenship acquisition, in contrast to the manifold barriers erected in a handful of European cases. But if the US is going to be the standard from which we deduce a liberal test, as opposed to inductively identifying concepts and selecting reflective, context-free standards for measurement, then it seems we are measuring cases against an outlier, not an average. This should not, in turn, render all European cases deviant and ‘illiberal’, but perhaps force us to identify new conditions to describe what we see as a common practice, rather than an exception, among democratic states.


[1] Sara Wallace Goodman, “Naturalization Practices Across Europe,” EUDO-Citizenship Working Paper, forthcoming 2010.

[2] Kristine Kruma, “Latvia: Country Report,” p. 13. Available at


[Accessed 28, February 2010].

[3] Sara Wallace Goodman, “Integration Requirements for Integration’s Sake? Identifying, Categorising and Comparing Civic Integration Policies,” in Journal of Ethnic and Migration Studies, forthcoming 2010.

[4] Thanks to Rainer Bauböck for this comment.

[5] Informatie-en Analyses Centrum (INDIAC) (2008). Monitor Inburgeringsexamen Buitenland. Den Haag: Immigratie- en Naturalisatiedienst.

[6] For more on pre-entry integration requirements, see Sara Wallace Goodman, “Integration Before Entry? Immigration Control through Language and Country Knowledge Requirements,” West European Politics, forthcoming.