In its Judgment of 24 September 2015 in Italy v Commission, joined cases T-124/13 and T-191/13, EU:T:2015:690 (funnily enough, not available in English or German), the General Court of the Court of Justice of the European Union (GC) ruled that EPSO, the Commission’s recruitment office, is breaking its own rules by forcing applicants to use English, French, or German (see short comment in English here).
Beyond its implications in the way EU Institutions recruit their staff, which are certainly not minor, the case is very important in at least two other ways: firstly, it is important from a legal-technical perspective because it follows up on the string of case law concerned with language requirements to access employment at the EU institutions started in the CJEU Judgment of 27 November 2012 in Italy v Commission, C-566/10 P, EU:C:2012:752 and severely limits the possibilities to justify those requirements under the framework that the CJEU had created (thus, crying out loud for an appeal of the GC's Judgment by the Commission); and, secondly, the case is relevant from a legal-functional (or policy) perspective because of its broader implications in terms of the tension between the EU's languages policy and the workability of its institutions [for discussion of the language policy element of the tension, see Stefaan van der Jeught's remarks here].
On the first point, from a technical perspective, I find the reasoning of the GC remarkably narrow-minded and disappointing. In the 2012 Italy v Commission CJEU Judgment, a reasonable framework was created by recognising the discretion of the EU Institutions to establish a balance between language requirements linked to the interest of the service (ie avoiding them becoming a sad 21st-century reincarnation of Babel's Tower) and the limitation in the selection of the best candidates should they not command the languages specifically chosen. The general principles of that framework were as follows:
88 ... it is apparent ... that the interest of the service may be a legitimate objective that can be taken into consideration. In particular, [relevant EU law] authorises limitations on the principles of non-discrimination and proportionality. Those interests of the service must however be objectively justified and the required level of knowledge of languages must be proportionate to the genuine needs of the service (see, to that effect, Case 79/74 Küster vParliament  ECR 725, paragraphs 16 and 20, and Case 22/75 Küster v Parliament  ECR 1267, paragraphs 13 and 17).
93. In so far as a legitimate objective of general interest may be relied upon and be shown to be genuine, it should be noted that a difference in treatment on the grounds of language must also observe the principle of proportionality, that is to say, it must be appropriate for attaining the objective pursued and must not go beyond what is necessary to achieve it (see, to that effect, Joined Cases C-453/03, C-11/04, C-12/04 and C-194/04 ABNA and Others  ECR I-10423, paragraph 68).
94 ... the recruitment of officials is to be directed to securing for the institution the services of officials of the highest standard of ability, efficiency and integrity. Since that objective can best be achieved when the candidates are allowed to sit the selection tests in their mother tongue or in the second language of which they think they have the best command, it is, in that regard, for those institutions to weigh the legitimate objective justifying the limitation of the number of languages of the competition against the objective of identifying the most competent candidates.
97 ... it is therefore a matter for the institutions to weigh the legitimate objective justifying the limitation of the number of languages of the competitions against the opportunities for recruited officials of learning, within the institutions, the languages necessary in the interest of the service (C-566/10 P, paras 88, 93-94 and 97, emphasis added).
Given this framework, which the GC recognises in its 2015 Italy v Commission Judgment, the decision it reaches is really surprising because the Commission went out of its way to justify very precisely the reasons why it insisted on candidates commanding one of either English, French or German as a second language (including those whose native tongue was one of them). In paras 74 and ff of its Judgment, the GC details how the Commission justified the discretionary choice to impose those language requirements and, in fact, concludes that 'the contested calls certainly contain a motivation aimed at justifying the requirement that candidates must have a satisfactory knowledge of German, English or French, languages to which their choice of the second language for the selection process is limited. Therefore, its author, EPSO, cannot be found to have breached the obligation to state reasons. The issue of the justification for this motivation is different, and will be discussed separately' (T-124/13 and T-191/13, para 83, own translation from Spanish).
Precisely in that analysis of the justification provided to motivate the imposition of the language requirements is where, in my view, the GC goes astray and engages in a sort of analysis that nobody acquainted with the way in which EU Institutions work could consider realistic or reflective of reality. This is particularly clear in this passage:
110 The [Commission's] claim that the three languages mentioned above "remain the most widely used languages" in view, specifically, of "the practice already firmly established in the EU institutions with regard to the languages used for internal communication" occupies a key position in this reasoning. However, it must be said that this is a vague statement, which is not supported by specific indications.
111 Indeed, this alleged (sic) practice of the EU institutions with regard to the languages used for internal communication is not explained in any way. In particular, [the Commission] does not specify if it involves parallel use of these languages as languages of internal communication in all services of all the institutions affected by the contested calls or, rather, some services use one of these languages and some another. In the latter case, there is a risk that services which may be interested in candidates who have passed the controversial oppositions do not use either of the three languages mentioned above as the language of internal communication, which would challenge the reasonableness and proportionality of the limitation, to these three languages, of the choice of a second language for the controversial selection process. Indeed, in that case, either some candidates that have passed the selection process will not be contracted, or the services in question will be forced to appoint, in part, candidates who do not speak the language of internal communication, in which case the a question of the meaning and utility of the above limitation may be legitimately raised.
112 The Commission has provided some details in his writings in this regard and submitted additional evidence. However, their analysis does not dispel the serious doubts raised by the above statements contained in the contested calls (T-124/13 and T-191/13, paras 110-112, own translation from Spanish and emphasis added).
In my view and based on my experience of interaction with the European Commission, this is a formal analysis with no grounding on reality. In fact, in my experience, the Commission fundamentally works only in English, and I would think that anyone familiar with the working of the Institutions would have a similar experience in views. Thus, the GC seems to have been chasing ghosts and imposed a burden on the Commission to justify that "alleged" language practice despite the fact that it is vox populi.
The GC is also very dismissive of all statistics and arguments submitted by the Commission to try to justify that practice (paras 113-144), which in my view exceeds the level of adequate substantive judicial review established in Art 263 TFEU, and concludes 'that the limitation ... to German, English and French of the choice of the second language for the selection process ... is not objectively justified or proportionate to its aim, which, according to the Commission, is to select officials and agents are immediately operational' (T-124/13 and T-191/13, para 145, own translation from Spanish).
In my opinion, by engaging in such a tough and dismissive analysis of the reasons provided by EPSO to justify the language requirements it saw fit to ensure the needs of the service, the GC went too far and emptied the analytical framework created by the CJEU in 2012 of any meaning by actually dismissing the important point that it is 'for those institutions to weigh the legitimate objective justifying the limitation of the number of languages of the competition against the objective of identifying the most competent candidates' and 'a matter for the institutions to weigh the legitimate objective justifying the limitation of the number of languages of the competitions against the opportunities for recruited officials of learning, within the institutions, the languages necessary in the interest of the service' (C-566/10 P, paras 94 & 97). Reading the GC's Judgment, it seems clear that the GC has substituted EPSO's discretion with its own, which does not seem to me to be compatible with the CJEU's approach.
Generally then, on the second point of policy, it seems clear that the GC (and the CJEU to a more limited extent, at least for now) is not ready to support the workability of the European Union and its institutions by decoupling issues of language recognition and support under cultural policies from issues of operability and efficiency of the institutional architecture of the EU. Entire books have been dedicated to these issues [eg T.J.M. van Els, 'The European Union, its Institutions and its Languages: Some Language Political Observations' (2001) 2(4) Current Issues in Language Planning 311-360], but very limited advances have been attained (other than sporadic highlights, such as the acceptance of languages restrictions for the purposes of the Single European Patent; see here). Thus, the problem remains unsolved and, by the looks of it, it will only grow more and more difficult to sort out...