I. Introduction

II. Legal framework

III. Facts, national proceedings and the questions referred

IV. Assessment

A. The layers

1. The outer layer: it is always for the national court to decide

2. The deeper layer: should all questions really be referred?

B. CILFIT (and its progeny)

1. The reasons for the duty to refer

2. The exceptions to the duty

3. The subsequent application practice (of this Court)

C. The problems with CILFIT

1. Conceptual

2. Feasibility

3. Systemic coherence of EU law remedies

4. The evolution of EU law and the judicial system

5. Interim summary

D. The proposal

1. A general or generalisable issue of interpretation of EU law

2. More reasonably possible interpretations

3. No case-law of the Court of Justice

4. The duty to state reasons (and the open question of enforcement of the duty to refer)

V. Conclusion

1.

In contrast to national courts of last instance, I suspect that students of EU law have always rather liked the judgment in CILFIT and Others. ( 2 ) In the course of the last decade or two, the hearts of many EU law students were likely to pound with a sudden flush of joy and relief upon seeing ‘CILFIT’, ‘exceptions to the duty to refer’, and ‘discuss’ on their exam paper or essay assignment. Indeed, questioning the feasibility of the CILFIT exceptions to the duty to make a reference for a preliminary ruling, particularly the exception relating to the absence of any reasonable doubt on the part of the national court of last instance, is perhaps not the most demanding argumentative exercise. Are those courts really required to compare (all) the equally authentic language versions of EU law? How are they, in practical terms, supposed to determine whether the matter is equally obvious to the courts of other Member States and to the Court of Justice?

2.

For a number of years, the duty to refer a question for a preliminary ruling pursuant to the third paragraph of Article 267 TFEU, the exceptions to that duty, and above all its enforcement, have been the metaphorical sleeping dogs of EU law. We are all aware that they are there. We are all able to discuss or even write scholarly treatises about them. However, in real life, it is best that they be left undisturbed. Pragmatically (or cynically) speaking, the entire system of preliminary rulings functions because no one in fact applies CILFIT, certainly not to its letter. Often, the idea of a dog is better than having to deal with the living animal.

3.

For number of reasons set out in this Opinion, I suggest to the Court that it is time to revisit the CILFIT file. My proposal in this regard is rather simple: to adapt the duty to refer under the third paragraph of Article 267 TFEU and likewise the exceptions to it, so that they reflect the needs of the current EU law judicial system, and so that they can then be realistically applied (and, possibly one day, enforced).

4.

However, the process of adaptation suggested requires a major paradigm shift. The logic and the focus of the duty to refer and the exceptions to it ought to change from the absence of any reasonable doubt as to the correct application of EU law in an individual case, evidenced by there being subjective judicial doubt, to a more objective imperative of securing uniform interpretation of EU law across the European Union. In other words, the duty to refer a question for a preliminary ruling should not be focused primarily on the correct answers, but rather on identifying the right questions.

5.

Pursuant to Article 267 TFEU:

‘The Court of Justice of the European Union shall have jurisdiction to give preliminary rulings concerning:

Where such a question is raised before any court or tribunal of a Member State, that court or tribunal may, if it considers that a decision on the question is necessary to enable it to give judgment, request the Court to give a ruling thereon.

Where any such question is raised in a case pending before a court or tribunal of a Member State against whose decisions there is no judicial remedy under national law, that court or tribunal shall bring the matter before the Court.

…’

6.

Article 99 of the Rules of Procedure of the Court of Justice states:

‘Where a question referred to the Court for a preliminary ruling is identical to a question on which the Court has already ruled, where the reply to such a question may be clearly deduced from existing case-law or where the answer to the question referred for a preliminary ruling admits of no reasonable doubt, the Court may at any time, on a proposal from the Judge-Rapporteur and after hearing the Advocate General, decide to rule by reasoned order.’

7.

On 23 February 2006, Consorzio Italian Management and Catania Multiservizi SpA (‘the applicants’) were awarded, by Rete Ferroviaria Italiana SpA (‘RFI’), a contract for ‘services relating to the cleaning and maintenance of the decoration of the premises and other areas which are open to the public, as well as ancillary services in stations, plants, offices and workshops at various sites throughout the territory covered by the Direzione compartimentale movimento di Cagliari [(Cagliari Regional Operations Division, Italy)]’.

8.

That contract contained a clause limiting price review. During the performance of the contract, the applicants requested RFI to adjust the contract price due to an alleged increase in contractual costs arising from an increase in staff costs. RFI refused.

9.

The applicants brought an action against RFI before the Tribunale amministrativo regionale per la Sardegna (Regional Administrative Court, Sardinia, Italy; ‘the TAR’). By judgment of 11 June 2014, the TAR dismissed the action. It found that Article 115 of Legislative Decree No 163/2006 was not applicable to activities falling within the special sectors, such as cleaning services, when they form an essential part of the railway transport network. The TAR also found that price review was not mandatory under Article 1664 of the Codice Civile (Civil Code) and that the parties could derogate from that rule.

10.

The applicants lodged an appeal against that judgment before the Consiglio di Stato (Council of State, Italy). They argued that Article 115 of Legislative Decree No 163/2006 or, in the alternative, Article 1664 of the Civil Code were applicable. Moreover, in their view, the national rules infringed Directive 2004/17/EC ( 3 ) in so far as they resulted in the exclusion of the review of prices in the transport sector and in related cleaning contracts. Those rules gave rise to an unfair and disproportionate contractual imbalance, ultimately distorting the rules governing the functioning of the market. Were Directive 2004/17 to be interpreted as excluding the review of prices in all contracts concluded and implemented within the special sectors, it would then be invalid.

11.

Against that background, on 24 November 2016, the Consiglio di Stato (Council of State) referred the following two questions to the Court of Justice for a preliminary ruling:

12.

The Court replied by judgment of 19 April 2018. ( 4 ) With regard to the first question, the Court noted that the referring court did not provide any explanation as to the relevance of the interpretation of Article 3(3) TEU or Articles 26, 57 to 58 and 101 TFEU for the resolution of the dispute in the main proceedings. Article 16 of the Charter of Fundamental Rights of the European Union (‘the Charter’) is, similar to the other Charter provisions, addressed to the Member States only when they are implementing EU law, which was not the case in the main proceedings. The Court thus concluded that the first question was inadmissible with regard to those provisions. ( 5 )

13.

However, as regards Directive 2004/17 and the general principles underlying that directive, in particular the principle of equal treatment and the principle of transparency, the Court found that they were to be interpreted as not precluding national rules, which do not provide for periodic price review after a contract has been awarded in the sectors covered by that directive. In view of that answer, the Court considered that the second question was hypothetical.

14.

On 28 October 2018, following the delivery of the Court’s judgment, the applicants requested the Consiglio di Stato (Council of State), ahead of the public hearing taking place on 14 November 2018, to refer to the Court new questions for a preliminary ruling. In substance, the applicants maintain that, in its judgment, the Court did not take a position on whether cleaning services were functionally linked to the transport service. They point out that, in its judgment, the Court assumed that, as stated in the initial invitation to tender, the duration of the contractual relationship could not be extended. However, that does not reflect the situation in Italy where service contracts are frequently extended by public authorities, sometimes indefinitely, thus creating a situation of contractual imbalance. In support of their position, the applicants rely on recitals 9, 10 and 45 of Directive 2004/17, as well as Article 57 thereof.

15.

According to the Consiglio di Stato (Council of State), which is once again the referring court in the main proceedings, the applicants have thus raised further questions for a preliminary ruling. However, the referring court wonders whether, in the circumstances of the main proceedings, it must refer those questions to the Court under the third paragraph of Article 267 TFEU. It considers that the obligation on a court of last instance to refer a question for a preliminary ruling cannot be separated from a system of ‘procedural bars’ which incentivises the parties to submit ‘once and for all’ to the national court the aspects of national law that they consider to be inconsistent with EU law. Otherwise, the successive or continuous proposal of questions for a preliminary ruling – in addition to lending itself to possible misuse and, in extreme cases, a real ‘abuse of procedure’ – would ultimately, as a result of the obligation to make a request for a preliminary ruling, render the right to judicial protection nugatory and undermine the principle that legal proceedings must be concluded swiftly and effectively.

16.

It is within this factual and legal context that the Consiglio di Stato (Council of State) has referred the following questions to the Court of Justice for a preliminary ruling:

17.

Written observations were submitted by the applicants, RFI, the Italian Government and the European Commission. Those interested parties, together with the German and French Governments, all participated at the hearing which took place on 15 July 2020.

18.

In accordance with the Court’s request, this Opinion shall focus on the first question raised by the referring court. That question has two layers. On the one hand, on its text, the first question could perhaps be read as simply enquiring whether a national court of last instance is bound to make a reference in a specific factual scenario containing three elements: (i) where the question is raised by one of the parties; (ii) where the question is raised after a party has lodged its initial pleading; and (iii) even after a request for a preliminary ruling has already been made to the Court. To provide an answer to these questions is not difficult. The answers indeed flow from well-established case-law of the Court (as will be demonstrated in Section A below).

19.

On the other hand, I do not think that such an answer would do justice to the issues actually raised by the referring court. That jurisdiction notes that there are additional issues raised by the parties which can in fact be settled on basis of the answer already provided by the Court. However, the referring court also acknowledges that there are new issues that cannot be disposed of in this way. With regard to those additional issues, the referring court points out that, being itself a court of last instance under national law, it is bound to make a reference to the Court because a question of interpretation of EU law has been raised before it. ( 6 )

20.

At this juncture, the broader issue of the first question emerges fully: are any and all cases in which there are lingering doubts as to the correct application of EU law in that particular case, regardless of whether or not a previous request for a preliminary ruling in the same case has already been made, covered by the duty to refer which is incumbent upon the courts of last instance? What is the proper scope of the duty to refer a question for a preliminary ruling and the exceptions to it, particularly in view of cases such as the present one in the main proceedings?

21.

In suggesting an answer to that question, I start by setting out the current outlines of the duty to refer and the exceptions to it (B). Next, I shall examine the different problems posed by that case-law, with those problems subsequently being redefined as the reasons for which I suggest that the Court articulate the duty to make a reference in a different light (C). Finally, I shall close with a specific proposal in this regard (D).

22.

The elements invoked by the referring court relate to three issues: (i) the role of the parties in raising a question; (ii) the timing of the question in relation to the various stages that a national judicial procedure may be composed of; and (iii) the possibility of a second request for a preliminary ruling within the same proceedings. If those three issues were understood as three separate questions, and not as three facets of one and the same question, their answer could easily be inferred from the established case-law of the Court.

23.

First, whether a referring court considers a decision of this Court to be necessary in order for it deliver a judgment is exclusively for that court to decide. Certainly, a national court, like any other court, is likely to listen to the views of the parties regarding the opportunity to refer a question for a preliminary ruling. However, Article 267 TFEU instituted direct cooperation between the Court of Justice and the national courts by means of a procedure, which is completely independent of any initiative by the parties. ( 7 ) Thus, the mere fact that a party to the dispute in the main proceedings has raised certain issues of EU law does not oblige the court concerned to consider that a question has been raised within the meaning of Article 267 TFEU, rendering a reference mandatory. ( 8 ) Conversely, that also means that a national court may submit a request for a preliminary ruling of its own motion. ( 9 )

24.

Second, it is also for the national court alone to decide when it is appropriate to refer a question to the Court of Justice for a preliminary ruling. ( 10 ) The national court is in the best position to appreciate at what stage of the proceedings it requires a preliminary ruling from the Court of Justice. ( 11 ) The Court only requires the dispute to be pending at the time of the request for a preliminary ruling. ( 12 )

25.

Certainly, from the point of view of the Court, and its wish to assist the referring court as fully as possible, situations in which the national court decides to make a reference to the Court only after a dispute and all its implications have fully materialised before that court cannot but be welcomed. Thus, in some cases, the sound administration of justice might require that the Court be seised only after both parties to a dispute have been heard by the referring court. However, that is certainly not, in and of itself, a condition imposed by this Court. ( 13 )

26.

In similar vein, I do not think that the Court would be opposed to national rules or provisions that require or mandate the concentration of proceedings, in particular before appellate or supreme courts, by obliging the parties to bring forward new or additional arguments at a certain moment in time or stage of the judicial procedure. However, it is true that in the past, the Court has in general taken issue if such deadlines or concentration of proceedings were to make it effectively impossible for the national courts to raise questions of compatibility of national rules with EU law. ( 14 )

27.

The logic of the case-law of the Court in this area thus involves ensuring that national rules of procedure do not prevent points of EU law from being raised, with a request for a preliminary ruling potentially being made regardless of the stage of proceedings. Where that is likely to lead, in practical terms, is to the potential setting aside of any restrictive national rules. ( 15 )

28.

However, in the context of the present case, it would appear that the question of whether a potential duty to refer pursuant to Article 267 TFEU could arise ‘even where the question is submitted to [the national court] by one of the parties to the proceedings after that party has lodged its initial pleading, or even after the case has been set down for judgment for the first time’ is being raised with a different outcome in mind. In the present case, it seems that no limiting rule in fact exists in national law, with the parties apparently being allowed to start (re)discussing elements before the referring court that were already the subject of a request for a preliminary ruling. However, it would be very unusual, to say the least, if one were to start invoking the above recalled ‘enabling’ line of case-law that has always categorically insisted on national judges being entirely free to raise any point of EU law at any stage of proceedings in order to suddenly seek the opposite result: to shut down any further discussion permitted under national law following a ruling of this Court received at the national level.

29.

Third, that would also sit very uneasily with the traditional approach of the Court with respect to the third element singled out by the referring court: the possibility to ask again. In that regard, the Court has always insisted that national courts remain at liberty to bring a matter before the Court if they consider it appropriate to do so, and the fact that the provisions whose interpretation is sought have already been interpreted by the Court does not deprive the Court of jurisdiction to give a further ruling. ( 16 )

30.

The same applies within the same national proceedings. According to the Court, ‘such a procedure may be justified when the national court or tribunal encounters difficulties in understanding or applying the judgment, when it refers a fresh question of law to the Court, or again when it submits new considerations which might lead the Court to give a different answer to a question submitted earlier’. ( 17 ) Thus, a new reference is always possible, on the same provisions of EU law, but also on potentially different provisions or issues within the same proceedings.

31.

It follows that the short answer to the first question is that a request for a preliminary ruling may be made at any time, irrespective of a previous preliminary judgment of the Court issued within the same proceedings, as long as the referring court is of the view that the Court’s answer is necessary to enable it to deliver a judgment. That decision is always to be taken by the national court, in the light of any reasonable doubt that it may still have regarding the correct application of EU law in the case before it.

32.

In short, all this remains within the exclusive competence of the referring court and its assessment of the (subjective) need to seek further guidance from the Court. However, is that really the answer? Or is that statement rather the description of the problem? Indeed, should there still be a duty to refer in cases such as the present one?

33.

Viewed in that context, the referring court’s first question reaches another, much deeper layer. That is further emphasised by the fact that all the three elements discussed in the previous section as free-standing questions are present in the main proceedings simultaneously. The question thus becomes whether the referring court is, despite all these circumstances, still under the obligation to make a reference.

34.

Therefore, I do not believe that such a question can properly be answered in the manner outlined in the previous section. I also do not believe that it could be answered by suggesting that the present case concerns merely the binding force of and the respect for a previous judgment of the Court. It is certainly true that, already since Da Costa, ( 18 ) and as later consolidated in CILFIT, ( 19 ) the authority of an interpretation provided by the Court under Article 267 TFEU may deprive that obligation of its purpose and thus empty it of its substance. ( 20 ) That is a fortiori the case when the question raised is substantially the same as a question which has already been the subject of a preliminary ruling in the same national proceedings. ( 21 )

35.

However, the referring court clearly stated that it is, in part, faced with new, additional elements to the case not previously addressed by the Court. Thus, the issue does not concern a disregard for a previous decision of the Court.

36.

Within such a context, is it impossible to state that EU law would preclude the Italian practice of giving the parties the opportunity to make their submissions in relation to an answer that was received from a superior court concerning a request made as part of those proceedings. At the hearing, the Italian Government stated that this is not only the practice adopted with regard to an answer received from the Court of Justice, but also with regard to answers received from the national constitutional court following a question of constitutionality submitted by a national judge. In such circumstances, the parties to the main proceedings are given a say as to the consequences that ought to be drawn for their case in view of the guidance issued by a superior jurisdiction.

37.

In summary, unless the Court wished to restate the obvious without engaging with the deeper layer of this case, or unless it were to rather radically revisit some of the elements just outlined, I suggest that a discussion on the nature and the scope of the duty to refer a question for a preliminary ruling is warranted in the context of the present case. The interested parties presented different views on the matter.

38.

In their submissions, the applicants and the defendant in the main proceedings focus primarily on Questions 2 and 3. With regard to the first question, the applicants maintain that requests for a preliminary ruling are redundant where there is established case-law unless the relevant judicial precedents are too old, or new arguments have been raised before the referring court, as is the case in the main proceedings. The defendant is of the view that the referring court should not have referred new questions as the interpretation of the EU law rules at issue was clear and as the Court’s case-law already provided an answer.

39.

The German, French and Italian Governments, as well as the Commission, all address the first question in greater detail, presenting a range of views. The German Government and the Commission consider that there is no reason to revisit CILFIT in any way. The German Government points out that the CILFIT criteria have withstood the test of time and ought to be maintained.

40.

The Italian Government called for a better balancing between the duty to refer and the sound administration of justice. According to that government, courts of last instance would infringe the third paragraph of Article 267 TFEU only if they omit to take into consideration questions on EU law raised by the parties or declare them unfounded without a statement of reasons. That government is of the view that that statement of reasons is crucial. It could even mitigate the liability incurred by Member States for the failure of their courts of last instance to refer.

41.

The French Government suggested that the CILFIT criteria should be (re)interpreted in the light of the overall purpose of Article 267 TFEU and in the light of the current state of EU law, taking into account the structural changes that have occurred. The duty to refer should focus on important questions of interpretation and on questions that can give rise to divergent interpretations within the Union, not necessarily on individual cases within the Member States. Questions concerning the application of EU law should not trigger the duty to refer. Such duty ought to be maintained only for general questions or questions that, albeit more casuistic, require the Court to set out a general framework of analysis or the criteria of legal reasoning. Although courts of last instance may still decide to refer other types of question, the requirements of the sound administration of justice and of the delivery of judgments within a reasonable period of time should be taken into account by the national courts, in particular when the Court has delivered a first preliminary ruling within the same proceedings.

42.

Pursuant to the third paragraph of Article 267 TFEU, courts or tribunals of a Member State against whose decisions there is no judicial remedy under national law are under a duty to refer questions on the interpretation or on the validity of EU law to the Court of Justice.

43.

There is thus clearly, in the text of the Treaty, a duty on courts of last instance ( 22 ) to refer to the Court questions covered by the first paragraph of that article. However, as is the case with a number of primary law provisions, the rest is simply a case-law construct.

44.

First, the text of the Treaty does not distinguish between questions of interpretation and questions of validity in terms of the scope of the obligation to refer. However, as regards questions on validity, the Court has stated that all national courts, that is to say not only those of last instance, are under an unreserved obligation to refer to the Court such questions. National courts have no jurisdiction themselves to determine the validity of acts of the EU institutions. ( 23 ) The requirement of uniformity is particularly essential where the validity of an EU act is in question. Differences between courts of the Member States as to the validity of EU acts would be liable to jeopardise the essential unity of the EU legal order and undermine the fundamental requirement of legal certainty. ( 24 )

45.

Such considerations place questions concerning the validity of EU law acts on a different and separate track. Crucially, since the reason for the categorical duty on all national courts to refer any issue of validity to the Court of Justice is different from the reasons for the duty to refer questions of interpretation, the same is true as regards their exceptions. The CILFIT exceptions are not applicable to the duty to refer a question on validity. ( 25 )

46.

It ought to be emphasised already at this stage that the following discussion in this Opinion concerns exclusively requests for a preliminary ruling on the interpretation of EU law.

47.

Second, it is fair to admit that, on its text, the third paragraph of Article 267 TFEU sets out a categorical duty to make a reference which is incumbent on national courts of last instance, without any exceptions: ‘where any such question is raised … that court … shall bring the matter before the Court.’ In this regard, it is the case-law of the Court that has effectively pencilled in the CILFIT exceptions regarding questions on interpretation.

48.

I am certainly not raising this point in order to insinuate that such exceptions are incorrect or unlawful. On the contrary, they are, in fact, necessary. However, the reason for my wish to underline this as a preliminary point to the discussion that follows is to emphasise the fact that when it comes to the nature and the scope of the duty to refer or the exceptions to that duty, the argument that ‘this cannot be altered because it is written like that in the Treaty’ is somewhat peculiar. In terms of what is the exact scope of the duty to make a request for a preliminary ruling, Article 267 TFEU is a notably open text. In terms of the exceptions to that duty, the Treaty is completely silent. To be more precise, if one wished to be a nagging textualist, one could even suggest that the text of Article 267 TFEU precludes there being any exception to the duty to refer.

49.

In general terms, ‘the system set up by Article 267 TFEU … establishes between the Court of Justice and the national courts direct cooperation as part of which the latter are closely involved in the correct application and uniform interpretation of European Union law and also in the protection of individual rights conferred by that legal order’. ( 26 )

50.

However, the Treaty requires more from the national courts of last instance, covered by the third paragraph of Article 267 TFEU, than it does from any court or tribunal under the second paragraph of thereof. There must therefore be some further, structural reasons as to why, in addition to the discretion to refer open to all national courts or tribunals, last-instance courts are under a duty to refer.

51.

The structural reason for the obligation incumbent on courts of last instance was stated early on in Hoffmann-Laroche: ‘to prevent a body of national case-law [that is] not in accordance with the rules of EU law from being established in any of the Member States’. ( 27 ) In other words, the special purpose of the third paragraph of Article 267 TFEU is the need ‘to prevent the occurrence within the [Union] of divergences in judicial decisions on questions of [Union] law’. ( 28 ) The limitation of that obligation to last-instance courts is specifically justified by the fact that ‘a court adjudicating at last instance is by definition the last judicial body before which individuals may assert the rights conferred on them by EU law. Courts adjudicating at last instance have the task of ensuring at national level the uniform interpretation of rules of law’. ( 29 )

52.

Thus, besides the desire to assist the national court in the correct interpretation or application of EU law in the individual case, which appears to be the overarching logic of Article 267 TFEU, the logic of the obligation to make a reference has been stated in systemic, structural terms: to prevent divergence in EU case-law. Logically then, that aim is indeed best pursued at the level of those national courts that tend themselves to be entrusted with the aim of ensuring unity at the national level.

53.

However, it is fair to acknowledge that in the statements (and in the application practice) of this Court over the years, the reason for the duty to refer has not always been expressed in such consistent terms. Sometimes the terms used are uniform interpretation and application of EU law; ( 30 ) sometimes the term is proper or correct application and uniform interpretation; ( 31 ) sometimes even just uniform application. ( 32 )

54.

These may be merely unintentional alterations in the wording used. However, on other occasions, they are indicative of a deeper discrepancy. They hint at an ongoing tension in terms of how differently the obligation under the third paragraph of Article 267 TFEU should be construed from the possibility under the second paragraph.

55.

The possibility under the second paragraph, as well as the overall purpose of the preliminary rulings procedure, is no doubt to assist national courts in resolving individual cases involving elements of EU law. That case-focused ‘micro purpose’ certainly serves, in the long run, the more systemic ‘macro purpose’ of the preliminary rulings procedure. It gradually builds up a system of precedents (or, in the language of the Court, established case-law), which helps to ensure the application of EU law uniformly across the European Union.

56.

However, can the obligation to make a reference be understood as a simple extension of the possibility to make a reference, as an attempt to identify instances in which something that has as its foundation the option to refer suddenly becomes a structural obligation, with the individual national judge who might have ‘subjective doubt’ suddenly finding himself or herself in ‘objective need’ of the assistance of the Court?

57.

CILFIT concerned a dispute between wool importers and the Italian Ministry of Health regarding the payment of a health inspection levy in respect of wool imported from outside the (then) Community. The importers relied on a provision of a regulation prohibiting Member States from levying any charge having an effect equivalent to a customs duty on certain imported animal products. The Ministry of Health contended that wool was not included under the Treaty and thus was not subject to that regulation.

58.

Against that background, the Corte Suprema di Cassazione (Supreme Court of Cassation, Italy) referred a question on the interpretation of the third paragraph of the then Article 177 of the EEC Treaty (now Article 267 TFEU), enquiring essentially whether the obligation to refer is automatic or conditional on the prior finding of a reasonable interpretative doubt. According to the Ministry of Health, the interpretation of the regulation was so obvious as to rule out the possibility of there being any interpretative doubt, thereby obviating the need to refer the matter to the Court. The importers concerned maintained that, since a question concerning the interpretation of a regulation was raised before a supreme court, against whose decisions there is no judicial remedy under national law, that court could not escape the obligation to bring the matter before the Court of Justice.

59.

In its judgment, the Court first recalled the purpose of the obligation to refer, which was ‘established with a view to ensuring the proper application and uniform interpretation of Community law in all the Member States … More particularly, the third paragraph of Article 177 seeks to prevent the occurrence within the Community of divergences in judicial decisions on questions of Community law. The scope of that obligation must therefore be assessed, in view of those objectives …’ ( 33 )

60.

Thereafter, in view of those objectives, the Court held that the obligation to make a reference is not absolute. It laid down three exceptions to the duty to refer of courts of last instance.

61.

First, courts of last instance are not obliged to refer a question on interpretation ‘if that question is not relevant, that is to say, if the answer to that question, regardless of what it may be, can in no way affect the outcome of the case’. ( 34 )

62.

I agree that a discussion as to whether this is in fact an exception to a duty, or rather confirmation of the fact that no duty exists, is entirely possible. If there is no (relevant) issue of EU law, what then should or even could be asked? When the question referred lacks relevance for the outcome in the main proceedings, not only is there no duty to refer, but such a question would simply be inadmissible. ( 35 )

63.

However, I think that the first ‘exception’ ought properly to be read in its temporal context. This was the first time that the Court stated that, despite the wording of the third paragraph of Article 267 TFEU, the same condition found in its second paragraph would also apply. However and perhaps more importantly, by building a bridge between those two paragraphs of Article 267 TFEU, the first exception became structurally tied to the individual case and the micro purpose of the request for a preliminary ruling: to assist national courts (of first- and last-instance) in settling a particular dispute pending before them when an issue of EU law arises.

64.

Second, there is no duty to refer where ‘the question raised is materially identical with a question which has already been the subject of a preliminary ruling in a similar case’, ( 36 ) and in situations in which ‘previous decisions of the Court have already dealt with the point of law in question, irrespective of the nature of the proceedings which led to those decisions, even though the questions at issue are not strictly identical’. ( 37 )

65.

That second exception, described in colloquial terms as ‘acte éclairé’ (referred to as such in other languages too), covers situations in which a precedent has already been established by the Court. It came about as an extension of the scope of the Da Costa judgment, where the Court held that ‘the authority of an interpretation … already given by the Court may deprive the obligation [to refer] of its purpose and thus empty it of its substance’. ( 38 )

66.

Third, the Court finally noted that there is no duty to refer when ‘the correct application of [Union] law may be so obvious as to leave no scope for any reasonable doubt as to the manner in which the question raised is to be resolved’. ( 39 ) In essence, this is how (arguably) the most well-known exception was born – the ‘no reasonable doubt’ exception or, as also referred to in other languages, the ‘acte clair’ exception.

67.

The Court went on to list several requirements that a national court must be convinced have been met in order to arrive at the conclusion that there is no scope for any reasonable doubt as to how to interpret the EU law provision at issue. It may be open to debate whether the requirement that the national court or tribunal must be ‘convinced that the matter is equally obvious to the courts of the other Member States and to the Court of Justice’ ( 40 ) is one of those specific requirements, or whether it is in fact a sort of general criterion which provides further guidance as to the type of situation in which there will be no reasonable doubt.

68.

However, assuming that this is a ‘general’ criterion, the Court went on to list additional features characteristic of Community law that the national court ought to keep in mind before reaching such a conclusion. These include that: (i) ‘an interpretation of a provision of Community law … involves a comparison of the different language versions’; ( 41 ) (ii) ‘legal concepts do not necessarily have the same meaning in Community law and in the law of the various Member States’; ( 42 ) (iii) ‘every provision of Community law must be placed in its context and interpreted in the light of the provisions of Community law as a whole, regard being had to the objectives thereof and to its state of evolution at the date on which the provision in question is to be applied’. ( 43 )

69.

Over the years, there has been no shortage of scholarly ( 44 ) or even judicial ( 45 ) contributions discussing whether the CILFIT exceptions, and in particular the points made with regard to acte clair, should indeed be read as requirements, thereby laying down a checklist for the national courts of last instance, or whether they should be approached more in the terms of an ‘overall guidance system’ or a ‘tool kit’, ( 46 ) not to be taken verbatim.

70.

What is significant in this regard is the subsequent case-law of the Court itself on this matter. There are two points arising from that case-law which are particularly noteworthy. First, the Court never expressly revisited or nuanced CILFIT, even though it was repeatedly invited to do so by a number of Advocates General. Second, when it came to cases involving the actual application of the CILFIT criteria, one could say that the application practice of the Court has been somewhat varied, particularly in the course of the last decade.

71.

First, in Intermodal Transports, ( 47 ) the referring court sought clarification on whether it had to refer questions on the interpretation of the combined nomenclature (CN) in circumstances where a party to a dispute concerning classification in the CN of a certain product invokes a decision by a customs authority, laid down in a binding tariff information (BTI), issued to a third party in respect of a similar product, and where the referring court thinks that that BTI is at variance with the CN. Put simply, is a court considered to harbour reasonable doubt in circumstances where it has a different legal view than that of an administrative authority of a different Member State?

72.

The Court determined that the third CILFIT exception (acte clair) could still be satisfied despite the existence of a divergent interpretation of EU law by an administrative authority in another Member State. According to the Court, the fact that the customs authorities of another Member State issued a decision which appeared to reflect a different interpretation of the CN from that which the referring court considered it must adopt in respect of similar goods in question in that dispute, ‘most certainly must cause that court to take particular care in its assessment of whether there is no reasonable doubt as to the correct application of the CN’. ( 48 ) However, the existence of such a BTI could not, in itself, prevent the referring court from concluding that the correct application of a CN tariff heading was so obvious as to leave no scope for any reasonable doubt. ( 49 )

73.

Second, in X and van Dijk, ( 50 ) the referring court (Hoge Raad der Nederlanden (Supreme Court of the Netherlands)) asked whether, in spite of the fact that a lower court in the Netherlands had submitted a request for a preliminary ruling on the same issue, the Hoge Raad (Supreme Court) could still lawfully rule on a dispute without referring questions to the Court of Justice for a preliminary ruling and without awaiting the answers to the questions referred for a preliminary ruling by that lower court.

74.

The Court stated that the existence of interpretive doubt on the part of a lower court within one Member State does not preclude the existence of an acte clair for a last-instance court of that same State. It first emphasised that ‘it is for the national courts alone against whose decisions there is no judicial remedy under national law, to take upon themselves independently the responsibility for determining whether the case before them involves an “acte clair”’. ( 51 ) It further held that, ‘since the fact that a lower court has made a reference to the Court for a preliminary ruling on the same legal issue as that raised before the national court ruling at final instance does not in and of itself preclude the criteria laid down in the judgment in [CILFIT] from being met, with the result that the latter court might decide to refrain from making a reference to the Court and resolve the question raised before it on its own, nor is the supreme national court required to wait until the Court of Justice has given an answer to the question referred for a preliminary ruling by the lower court’. ( 52 )

75.

Third, in Ferreira Da Silva e Brito, ( 53 ) the applicants challenged, in an action for a declaration of non-contractual civil liability against the Portuguese State, the interpretation of ‘transfer of a business’ within the meaning of Directive 2001/23/EC ( 54 ) embraced by the Supremo Tribunal de Justiça (Supreme Court of Justice, Portugal). According to the applicants, the Supremo Tribunal de Justiça (Supreme Court of Justice) should have complied with its obligation to refer and should have submitted a question on that issue to the Court. The referring lower court asked the Court whether Article 267 TFEU was to be interpreted as meaning that, in the light of the fact that the lower national courts adjudicating on the case adopted contradictory decisions, the Supremo Tribunal de Justiça (Supreme Court of Justice) was under an obligation to refer to the Court for a preliminary ruling a question concerning the correct interpretation of the concept of a ‘transfer of a business’ within the meaning of Directive 2001/23.

76.

The Court stated that ‘in itself, the fact that other national courts or tribunals have given contradictory decisions is not a conclusive factor capable of triggering the obligation set out in the third paragraph of Article 267 TFEU. A court or tribunal adjudicating at last instance may take the view that, although the lower courts have interpreted a provision of EU law in a particular way, the interpretation that it proposes to give of that provision, which is different from the interpretation espoused by the lower courts, is so obvious that there is no reasonable doubt’. ( 55 ) However, unlike in X and van Dijk, on the very facts of the case, the Court considered that, as a result of conflicting lines of case-law at national level together with the fact that that concept frequently gives rise to difficulties of interpretation in the various Member States, a national court or tribunal against whose decisions there is no judicial remedy under national law must make a reference to the Court in order to avert the risk of an incorrect interpretation of EU law. ( 56 )

77.

Fourth, in Association France Nature Environnement, ( 57 ) the Court, while not expressly addressing the issue of reasonable doubt, excluded the application of the CILFIT exceptions (in casu both acte clair and acte éclairé). In the specific context where a court of last instance envisages making use of the exceptional power which enables it to decide to maintain certain effects of a national measure that are incompatible with EU law, the Court retained a strict approach to the obligation to refer.

78.

The Court first restated the criteria for the existence of an ‘acte clair’, ( 58 ) in order to note that ‘in respect of a case such as that in the main proceedings, therefore, in which the question of its being possible for a national court to limit in time certain of the effects of a declaration of illegality of a provision of national law adopted in disregard of the obligations provided for by [a directive], has not been the subject of another decision of the Court since the [first] judgment ( 59 ) and, moreover, in which such a possibility is exceptional in nature, … the national court against whose decisions there is no longer any judicial remedy under law must make a reference to the Court for a preliminary ruling when it has the slightest doubt as regards the interpretation or correct application of EU law’. ( 60 )‘In particular, given that the exercise of that exceptional power could adversely affect observance of the principle of the primacy of EU law, that national court could be relieved of the obligation to make a reference to the Court for a preliminary ruling only if it is convinced that the exercise of that exceptional power does not give rise to any reasonable doubt. In addition, it must be established in detail that there is no such doubt.’ ( 61 )

79.

If that case-law is viewed together, there are at least three elements worth noting.

80.

First, although all the cases refer to CILFIT, and some even invoke the exceptions stated therein, none of them actually applied the CILFIT criteria. It could be suggested that these are elements of facts to be applied by national courts. However, that is not entirely correct because, in these cases, it was in fact the Court itself that carried out the evaluation of whether or not in such circumstances the CILFIT exceptions, most frequently that of acte clair, could have been validly invoked by the referring court. Nevertheless, even if clearly carrying out that type of evaluation, the Court seems not to be applying its own test.

81.

Second, the case-law referred to above demonstrates very well the difficulty that ensues in the application practice as a result of the conceptual lack of clarity in the ‘subjective-objective’ nature as to the existence of ‘any reasonable doubt’. Is it purely subjective doubt, that is to say doubt arising in the mind of the national court? Or is that doubt only subjective as long as it is not rebutted by strong, objective circumstantial evidence? Or should the judge have in fact subjectively understood that there were objective doubts? The first two cases, Intermodal Transports and X and van Dijk, fall on the subjective and hence deference side (it is ultimately for the national judge who knows best). By contrast, the other two cases, Ferreira Da Silva e Brito and Association France Nature Environnement, come down heavily on the objective side (in view of the objective circumstances, the judge should have known better).

82.

Third, such fluctuation in terms of test naturally translates into somewhat different outcomes. Certainly, each case is different on its facts. However, in terms of overall approach, it is not immediately clear how to reconcile, for example, X and van Dijk and Ferreira Da Silva e Brito. One would assume that there are likely to be reasonable doubts (and thus absence of an acte clair) if there are objective contradictory views on the interpretation of the same provisions. It is true that if one were to add arguments relating to formal authority or the separation of powers into the mix, one could distinguish Intermodal Transports: even if there are divergent legal interpretations of the same piece of EU law across jurisdictions, these are ‘only’ the work of administrative authorities, they are not judicial pronouncements.

83.

However, in both X and van Dijk and Ferreira Da Silva e Brito, there were diverging judicial pronouncements on the same elements of EU law emerging from within the respective jurisdictions. Since both of those cases concerned different proceedings within those Member States, it could hardly be maintained that either of them concerned a ‘one-off mistake’ by a lower court, which was waiting to be corrected by a court of last instance. There were objectively different interpretations of the same legal provisions in different proceedings within the same Member State.

84.

Again, every case is different on its facts. One could thus identify potentially relevant factual differences between the two cases, justifying a different outcome. Perhaps the most intriguing one would be to suggest that whereas X and van Dijk concerned a recognised interpretative divergence solely within one and the same national legal system, in Ferreira Da Silva e Brito, that divergence was not only within one Member State, but also apparently across the Union. Indeed, as the Court noted in that latter judgment, ‘the question as to how the concept of a “transfer of a business” should be interpreted has given rise to a great deal of uncertainty on the part of many national courts and tribunals which, as a consequence, have found it necessary to make a reference to the Court of Justice. That uncertainty shows not only that there are difficulties of interpretation, but also that there is a risk of divergences in judicial decision within the European Union’. ( 62 )

85.

Was it indeed the intention of the Court suddenly to start taking CILFIT literally, that is to say that a national court ‘must be convinced that the matter is equally obvious to the courts of other Member States and to the Court of Justice’, ( 63 ) to the exclusion of contradictory decisions on matters of EU law coming from within the same Member State? Intriguingly, such a reading is possible. Nonetheless, there is a very important premiss hidden therein, namely that it is for the respective national supreme courts to ‘put their own house in order’ and unify national case-law. Only then, in the event of interpretative divergences across the Member States, would the duty to refer be triggered. ( 64 )

86.

However, the fact remains that, on the whole, it is not easy to reconcile the recent case-law of the Court outlined in this section. This fact has encouraged a number of scholars to speculate about whether the Court has relaxed, somewhat tacitly, the CILFIT acte clair criteria. ( 65 ) It would appear that certain national courts have also understood it in the same way. ( 66 )

87.

In summary, I would suggest that the diversity in the case-law outlined justifies already, in its own right, the intervention of the (Grand Chamber of the) Court in order to clarify exactly what, at present, is the scope of the duty to refer under the third paragraph of Article 267 TFEU and the possible exceptions to that duty.

88.

In this section, I seek to point out the problems (C) that ought to guide, in my view, the revisiting of CILFIT suggested in the following section (D). I place the problems into four categories: those that are conceptual and were present and inherent in CILFIT from the very beginning (1); those relating to the feasibility of the CILFIT criteria (2); those concerning the systemic inconsistency of the CILFIT criteria with other types of proceedings and remedies in EU law (3); and finally those which have arisen later due to the subsequent evolution of EU law and the EU judicial system (4).

89.

There are a number of flaws in the original CILFIT blueprint and design. This is why the CILFIT criteria were conceptually problematic from their very inception.

90.

First and foremost, there is what I would call the ‘Hoffmann-Laroche-CILFIT mismatch’. Put simply, the logic of the CILFIT exceptions does not correspond with the nature of the Hoffmann-Laroche duty to refer a question for a preliminary ruling.

91.

Ever since Hoffmann-Laroche in 1977, the Court has continued to insist that the purpose of the duty to refer is to prevent a body of case-law being established in a Member State which deviates from that of other Member States and also from that of the Court. ( 67 ) The logic behind such a duty to refer is clearly structural and focused on uniform interpretation across the European Union. Its triggering point and nature appear objective, focusing on the case-law in general, not merely on the individual case before the referring court: there shall be no divergence in the national case-law (‘la jurisprudence’; ‘die Rechtsprechung’; ‘guirisprudenza’; or ‘rechtspraak’ as it appeared in the other official languages at that time). ( 68 )

92.

However, around five years later, when the exceptions to that duty were set out in CILFIT, the logic guiding those exceptions became focused primarily on the individual case and the subjective judicial hesitations in that specific case. The only real exception, taking into account structural or systemic considerations, that asks the national judge to look beyond the individual file, is the exception relating to acte éclairé: the existence of a precedent. By contrast, the first exception (confirmation of the relevance of the question) is concerned exclusively with the specific case. Most importantly, the acte clair exception is conceived of as a strange combination of elements relating to the individual case, some of which are objective and general, but most of which are subjective.

93.

In my view, that is where the real problem lies. One would normally expect the exceptions to a duty to mirror the overall logic and purpose of that duty. In a way, they should represent the other side of the same proverbial coin. However, the exceptions are effectively detached from the duty they are supposed to be bringing about. Indeed, while CILFIT proclaimed that it wished to follow the Hoffmann-Laroche objectives, ( 69 ) it nevertheless construed the exceptions around a very different logic.

94.

Second, and as a side note, the genealogy of the third exception (acte clair), appears to confirm the doubts relating to the systemic appropriateness of that distinct legal transplant. In CILFIT, the Court essentially took the defendant’s argument, and apparently that of the referring court as well, and made it its own: the interpretation of the regulation at issue was ‘so obvious as to rule out the possibility of there being any interpretative doubt, [thereby obviating] the need to refer the matter to the Court’. ( 70 ) In its judgment, the Court largely endorsed that approach, although qualifying it slightly by adding ‘reasonable’ to characterise the type of interpretive doubt.

95.

The fact that that exception arose in the context of a particular dispute is in itself unproblematic. The more questionable element is the transplantability of what was a distinctively French doctrine, crafted in a very different context for a different purpose, into a sui generis EU-type of proceedings. ( 71 ) What became known as the ‘acte clair theory’ was applied in French law in different contexts, in particular in cases involving the interpretation of treaties. While it was in principle for the Minister for Foreign Affairs alone to interpret the treaties (the national courts being merely entrusted with applying that interpretation to the case), the French courts relied on that theory to strengthen judicial interpretive powers to the detriment of those of the executive. ( 72 ) From 1964, the Conseil d’État (Council of State, France) began to apply the theory, in the context of the duty to refer, in order to qualify the latter to its own benefit. ( 73 ) Presumably intending to bring the national use of the acte clair practice under control, the Court took as its own and transplanted into the EU legal order a device that had very different function in its system of origin.

96.

Third, even if one were to believe that such trans-functional legal transplants are indeed possible without endangering the health of the (recipient) patient, the fact remains that what had been transplanted into the (then Community) legal order had simply a different function. The CILFIT exceptions, in particular the acte clair exception, remained centred around the absence of any reasonable doubt as to the outcome of an individual case. The language remained one of correct interpretation and application of EU law in the specific case.

97.

However, a requirement to check the correct application of the law in each individual case is a very tall order. That ideal is very difficult to achieve even in national systems that are hierarchical in nature, relying on extensive individual monitoring of the correctness of individual decisions, and within which superior courts review thousands or rather tens of thousands of decisions annually. However, such an ideal and systemic logic is alien to more coordinated systems of judicial governance, that rely extensively on the force of precedents, and within which a single precedent may carry weight. ( 74 ) In my view, it might be debated how close the present EU law judicial system is to the latter, coordinated ideal. But it is, certainly, worlds apart from the former, hierarchical model.

98.

Finally, the CILFIT exceptions to the duty to refer have the tendency of blurring the line between interpretation and application of EU law, underlining the correct division of tasks under Article 267 TFEU. If what must be established is that there is no reasonable doubt as to the correct application of EU law in a specific case in order for an exception to the duty to refer to be triggered, where then, at least approximately, does the boundary between the tasks of this Court and that of the national courts lie?

99.

There is no need to rekindle the debate on whether the CILFIT criteria, in particular those relating to there being no reasonable doubt as to the correct interpretation and application of EU law in a specific case (the acte clair exception), amount to a checklist or are simply a tool kit. The fact is that they do not work either way. If they are a checklist, then it is impossible to complete that checklist. If they are a tool kit, then the problem inherently becomes one of selectivity as to which tool should be used in an individual case. The latter problem then becomes evidently serious at the stage of potential enforcement of the duty to refer: if there are no clear criteria, how could that duty ever be enforced without that enforcement being arbitrary?

100.

A number of Advocates General have commented on the lack of feasibility of the criteria in the past. Moreover, what is also intriguing is the absence of application of those criteria by the national courts and the European Court of Human Rights (‘the ECtHR’), including the courts that are in fact enforcing the duty to make a reference.

101.

First, starting famously with Advocates General Jacobs’ hesitations with regard to certain women’s garments and their classification as pyjamas for customs purposes, a number of my learned predecessors have criticised the practical difficulties associated with the acte clair exception. In Wiener, Advocate General Jacobs rejected the view that ‘the [CILFIT] judgment should be regarded as requiring the national courts to examine any Community measure in every one of the official Community languages … That would involve in many cases a disproportionate effort on the part of the national courts’, noting that it would be rather strange to require national courts to have recourse to a method ‘which appears rarely to be applied by the Court of Justice itself’. ( 75 ) He also pointed out that ‘if the [CILFIT] judgment were applied strictly, then every question of Community law … would have to be referred by all courts of last instance’. ( 76 )

102.

In her Opinion in Intermodal Transports, Advocate General Stix-Hackl held that requiring national courts to examine a provision of the then Community law in every one of the official Community languages ‘would place a practically intolerable burden on the national courts’. ( 77 ) In her view, the CILFIT criteria cannot be used as ‘a type of instruction manual … which is to be rigidly adhered to’. ( 78 ) In his Opinion in Gaston Schul, Advocate General Ruiz-Jarabo stressed the need to revisit CILFIT in its entirety, noting that ‘the proposed test was unviable at the time it was formulated, but, in the reality of 2005, it seems preposterous’. ( 79 ) In his Opinion in X and van Dijk, Advocate General Wahl sought to emphasise the importance of the inherent discretion national courts of last instance must have in assessing whether the duty to refer has arisen, suggesting that the CILFIT criteria should be seen as a tool kit and concluding that ‘if a national court of last instance is sure enough of its own interpretation to take upon itself the responsibility (and possibly the blame) for resolving a point of EU law without the aid of the Court of Justice, it ought to be legally entitled to do so’. ( 80 )

103.

I do not think that there is any need to repeat the arguments already so eloquently made by the Advocates General quoted above, or to demonstrate why and how the CILFIT criteria, if taken at face value one by one, are utterly unfeasible. Indeed, as pertinently and wittily put by Advocate General Wahl, ‘coming across a “true”acte clair situation would, at best, seem just as likely as encountering a unicorn’. ( 81 )

104.

Simply put, the CILFIT criteria relating to the identification of the acte clair exception are inescapably plagued by the conceptual problem already highlighted above. ( 82 ) On the one hand, there is a healthy portion of non-ascertainable and thus non-reviewable subjectivism: the national courts must be ‘convinced’ that the matter is not only ‘equally obvious to the courts of other Member States and to the Court of Justice’ but also ‘necessary for deciding a case’, and must have subjective ‘reasonable doubt’. On the other hand, those elements that are stated in objective terms are simply unattainable, at least for mortal national judges not possessing the qualities, time, and resources of Dworkin’s Judge Hercules (comparing (all) language version; interpreting each provision of EU law in the light of EU law as a whole, while having a perfect knowledge of its state of evolution at the date on which that provision is interpreted).

105.

Second, and somewhat understandably in view of the previous point, the CILFIT criteria for ‘acte clair’ are hardly applied in a consistent and systematic manner by national courts of last instance within the examination of the third exception to the duty to refer. ( 83 ) There are certainly instances in which CILFIT is referred to by national courts of last instance. However, that does not mean that the criteria set out in that judgment, in particular the specific requirements relating to the existence of acte clair, would in fact be applied as such. ( 84 ) There are instances in which national courts replace the CILFIT criteria with their own criteria and standards. For example, the French supreme courts frequently consider that the duty to refer is triggered where there is a ‘serious difficulty’ in interpreting EU law, thereby making the duty to refer much looser than that which follows from the letter and spirit of CILFIT. ( 85 ) Similarly, other supreme courts have also focused on the nature of the question raised, regardless of the existence of interpretative doubts, as the basis of the duty to refer. ( 86 )

106.

Third, there are some Member States in which compliance with the duty to refer is checked by constitutional courts by means of an individual constitutional complaint, invoking the right to a lawful judge or fair trial. ( 87 ) However, in those systems, the overall yardstick adopted is, in fact, far lighter than the CILFIT one, revolving around the concepts of an obviously untenable or arbitrary interpretation of EU law, often coupled with the duty to properly justify the reason why a reference to the Court of Justice was not necessary. ( 88 )

107.

It could be suggested that national constitutional courts review the obligation to make a reference to the Court of Justice under the third paragraph of Article 267 TFEU in line with their national standards and notions of constitutionality and fundamental rights protection. That is certainly true with regard to whether the identification of the relevant national right per se (the right to a lawful judge, or a fair trial, or due process of law, or whatever else) will be one guaranteed in the national legal system. It is a different story for the internal yardstick, that is to say the specific conditions under which a request for a preliminary ruling ought to be made. On the latter issue, it is fair to assume that national constitutional rules are silent in that regard. Yet, when it comes to the effective enforcement of the duty, none of the national constitutional courts appear, in fact, to have embraced CILFIT.

108.

Fourth, the ECtHR invokes the CILFIT criteria when reviewing the failure of national courts of last instance to seek a preliminary ruling from the Court under the heading of the right to a fair trial under Article 6(1) of the European Convention on Human Rights (ECHR). The ECtHR found, for instance, a violation of Article 6(1) ECHR in the case of Dhahbi v. Italy due to the lack of reasons given by the domestic court for its refusal to refer a question to the Court. It noted that the Italian Court of Cassation had made no reference to the applicant’s request for a preliminary ruling, nor explained the reasons why it had considered that the question raised did not warrant a referral to the Court, nor made any reference to its case-law. ( 89 ) However, elsewhere, the ECtHR considered that a summary reasoning to refuse a request for a preliminary ruling was sufficient where the national court had already concluded in another part of its judgment that such a request was redundant. ( 90 ) The ECtHR has also held that, where a national court of last instance refuses or fails to refer a question for a preliminary ruling, that court is required to give reasons for such refusal in the light of the exceptions provided for by the case-law of the Court. It must in particular indicate the reasons why it considers the question to be irrelevant, whether the EU law provision in question has already been interpreted by the Court, or whether the correct application of EU law is so obvious as to leave no scope for any reasonable doubt. ( 91 )

109.

In general, while the review carried out by the ECtHR refers to the CILFIT decision, its focus remains on the reasoning behind a national decision not to make a reference. Thus, without dwelling much on the substance, and certainly not in terms of a detailed examination of the ‘no reasonable doubt’ factor which is in fact required by CILFIT, the ECtHR reviews whether national courts of last instance have duly explained why they consider that the CILFIT criteria have been fulfilled without the ECtHR examining the merits as to whether that is indeed the case.

110.

In summary, the lack of any reasonable guidance as to the logic or application of the CILFIT criteria is not only reflected in the (surprisingly consistent) criticism given by the past Advocates General over the years. The same point also emerges from the simple fact that, among those tasked with applying that obligation, and in particular those who are actually enforcing it, no one appears to be following the guidance of the Court. This is, in my view, not because they seek, in one way or another, to disregard the Court. It is rather a natural self-preservation mechanism. Impossibilium nulla est obligatio.

111.

There is another argument as to why it is necessary to revisit CILFIT: the systemic, horizontal coherence of EU law remedies. In a nutshell, the CILFIT criteria are even oddly disconnected from EU law’s own means of enforcing the obligation to make a reference under the third paragraph of Article 267 TFEU.

112.

Certainly, at present, there is no specific EU law remedy available to the parties to which they could have recourse if they believe that their right to have a matter submitted to the Court under the third paragraph of Article 267 TFEU has been infringed. ( 92 ) That is a logical consequence of the Court’s insistence that the parties to the main proceedings do not have an automatic right to have a request for a preliminary ruling made in as much as Article 267 TFEU does not constitute a means of redress available to the parties to a case pending before a national court. ( 93 ) In view of the (now) robust national constitutional case-law, ( 94 ) as well as the case-law of the ECtHR, ( 95 ) according to which, if the (objective) criteria for the existence of an obligation to refer are met, the parties to those proceedings have a (subjective) right to have their case referred to the Court which is inherent in their right to a fair trial, one may nevertheless consider that that is perhaps not the only possible approach. ( 96 )

113.

Since there is no ‘direct’ remedy, the possible enforcement of the duty to refer under the third paragraph of Article 267 TFEU could, possibly, be a matter of either state liability or an infringement action. However, that is exactly where the issues get complicated.

114.

On the one hand, since Köbler, there is the possibility of obtaining redress in the national courts for the damage caused by the infringement of individual rights owing to a decision of a court adjudicating at last instance. ( 97 ) In order for such an action to be successful, the legal rule infringed must be intended to confer rights on individuals; the breach must be sufficiently serious; and there must be a direct causal link between the breach of the obligation incumbent on the State and the loss or damage sustained by the injured parties. State liability for loss or damage caused by a decision of a national court adjudicating at last instance is governed by the same conditions. ( 98 )

115.

There are, however, two problems. First, since the third paragraph of Article 267 TFEU is not a rule ‘intended to confer rights on individuals’, the non-compliance with the obligation to refer cannot, on its own, trigger state liability. Second, irrespective of that fact, the CILFIT criteria play no role in the assessment of whether or not there has been an infringement of other rules of EU law, ( 99 ) at least objectively speaking. The standard in such situations is manifest infringement of the applicable law that may lead to a sufficiently serious breach. ( 100 )

116.

On the other hand, there is the infringement proceedings pursuant to Article 258 TFEU. What, for a number of years was possible only in theory, ( 101 ) was fully implemented in 2018. In its judgment in Commission v France, ( 102 ) the Court found, in the operative part of the judgment, that a Member State was in breach of EU law specifically for the failure of a last-instance court to make a (single) reference to the Court of Justice in order to fulfil their obligation to refer, under the third paragraph of Article 267 TFEU, in a situation where the interpretation of the substantive provisions of EU law in question was not so obvious as to leave no scope for doubt.

117.

In reaching that conclusion, the Court relied on CILFIT, ( 103 ) or at least on its general requirement of there not being any reasonable doubt. The Court first noted that the Conseil d’État (Council of State) had chosen to depart from a previous judgment of the Court concerning the United Kingdom legislation on the ground that the British scheme in question was different from the French scheme ‘while it could not be certain that its reasoning would be equally obvious to the Court’. ( 104 ) The Court thus held that the absence of request for a preliminary ruling on the part of the Conseil d’État (Council of State) in two of its judgments ‘led that court to adopt, in those judgments, a solution based on an interpretation of the provisions of Articles 49 and 63 TFEU which is at variance with that of the present judgment, which implies that the existence of reasonable doubt concerning that interpretation could not be ruled out when the Conseil d’État (Council of State) delivered its ruling’. ( 105 )

118.

There are three elements worth noting that relate to the latter judgment. First, in those infringement proceedings, it was undoubtedly for the Court itself to apply CILFIT. However, in that judgment, the Court contented itself with merely stating that the general criterion was laid down in CILFIT, without reviewing any of specific criteria or applying them. There is a conspicuous absence of discussion not only of any potentially contradictory judgments on the matter emerging from the courts of last instance of other Member States, or even of other French courts, but also of previous case-law of the Court itself on the matter, except for one precedent.

119.

Second, such effective affirmation of ‘stricter CILFIT’, at least in terms of its spirit, sits somewhat uneasily with the Court’s latest case-law on the same matter, as embodied, in particular, in the judgments in X and van Dijk and Ferreira Da Silva e Brito discussed above. ( 106 ) Whereas a light touch appeared to be the new rule in X and van Dijk in particular, that light touch suddenly turned into a tight grip once more.

120.

Third, all that is likely to leave certain actors with a somewhat bitter aftertaste in the form of a hardly defendable selectivity as to what is in fact being applied and enforced, as well as why, and how that application and enforcement takes place. This in no way suggests that the Commission would not have, as the indeed established case-law has it, complete discretion in deciding whether to bring an action under Article 258 TFEU. ( 107 ) It is also not intended to challenge the outcome of Commission v France: certainly, if CILFIT is now to be taken seriously, then there indeed was reasonable doubt as to the correct application of EU law in Commission v France. ( 108 )

121.

The point is a different one: the overall lack of (horizontal) consistency in penalising the duty to refer under the third paragraph of Article 267 TFEU as a matter of EU law. The Court’s own case-law on the scope of that duty, certainly the more recent case-law, does not appear to be in line with the recently (re)discovered enforcement of that duty under Article 258 TFEU, and wholly disconnected from state liability. However, should these not be the facets of one and the same obligation to make a request for a preliminary ruling under the third paragraph of Article 267 TFEU?

122.

I do not think that it is necessary to dwell in any great detail on the obvious: just how much the European Union has changed over the past 40 years. With one recent exception, membership has grown continuously. So too have the number of official languages and the number of courts able to seise the Court of Justice. At present, the scope and breadth of EU law itself is simply unparalleled. Following the completion of the internal market and no fewer than five successive Treaty amendments, it has become difficult to find an area for which EU legislation is not available or in which the interpretative help of the Court is not needed. Those factors lead to a staggering number of new requests for a preliminary ruling, while the judicial resources of the Court of Justice are not limitless.

123.

Nevertheless, in the midst of this vastly altered legal landscape stands the motionless Titan of a long bygone era, CILFIT, insisting that references be made by courts of last instance in every case where any form of reasonable doubt exists. I do not think that it is necessary to go through all the systemic changes, one by one, in order to demonstrate how they alter the input configuration for the logic of CILFIT. ( 109 ) Instead, in conclusion, I prefer to point to a different, deeper matter.

124.

As the environment has changed and the system has matured, the nature of the preliminary rulings procedure has evolved as well. A procedure originally conceived as being one of partnership and judicial cooperation amongst equals has gradually and rather inevitably developed into one which places greater emphasis on precedent building for the purpose of systemic uniformity. Certainly, the language of ‘assistance’ and ‘partnership’ still remains, but the more long-standing and careful observers of the system have noticed a number of vertical elements gradually being introduced.

125.

This places greater emphasis on the macro (or public) purpose of the adjudication of the Court in the form of ensuring uniform interpretation and further development of the law. Certainly, there always is, and always will be, the individual case and the more micro (or private) dimension of a litigation. However, increasingly, a fortiori in a procedure such as the preliminary ruling procedure, where the facts and the individual case are for the referring court to settle, the focus shifts beyond the specific case file.

126.

To make another systemic parallel: the recently introduced Article 58a of the Statute of the Court of Justice of the European Union established a filtering mechanism for appeal in certain areas from the General Court to the Court of Justice. Any such appeal shall not proceed to an examination on the merits unless the Court of Justice decides that it should be allowed to do so because it raises ‘an issue that is significant with respect to the unity, consistency or development of Union law’. ( 110 ) If such considerations are valid with regard to direct actions, in which only the Court of Justice of the European Union being the exclusive judge, and hence the litigation before both EU Courts is primarily a matter for settling the individual case, should the same logic not be applicable a fortiori to preliminary rulings?

127.

Finally, maturation of a judicial system also implies a maturing of its constituent parts. Today, national courts are much more familiar with EU law in general and with the preliminary rulings procedure in particular. There are, and there will always be, individual exceptions. However, a potential (reluctant) tree should not be allowed to overshadow the (compliant) forest. National courts of last instance, especially those structurally tasked with ensuring unity and uniform application of the law within their respective jurisdictions, have been privileged partners of the Court in identifying cases with structural importance for the EU legal order. Should (the often rather asserted than actual) mutual trust not also be of relevance vertically?

128.

The fact that the national courts of last instance are able to handle the preliminary rulings procedure is, at present, in my view, evidenced in a rather unorthodox way: namely that they are not following CILFIT. Heretical as it may sound, by exercising self-restraint and discretion incompatible with the criteria set out by the Court in CILFIT, national courts of last instance are in fact demonstrating a very good understanding of the true nature of the system. One may only imagine the reverse scenario, in which (some of) the national courts of last instance were to embrace the approach of the Good Soldier Švejk, ( 111 ) and were indeed to apply CILFIT to the letter to all cases before them. The annual docket of the Court of Justice would suddenly have several more zeros attached at the end and the system would collapse within a short period.

129.

Not only has CILFIT been problematic in terms of its feasibility, but above all (or before that) it was mistakenly conceptualised. The nature of the CILFIT exceptions does not match the nature of the Hoffmann-Laroche duty to refer that it was supposed to enforce. A duty which was established in order to ensure the uniform interpretation of the case-law across the Union cannot be made subject to the absence of any subjective doubts as to the correct application of EU law in the specific case.

130.

All the other issues outlined in this section are partly the consequence of that conceptual mismatch, while adding problems of their own. The CILFIT criteria are thus not being applied either by the Court of Justice itself, or by the national courts, including those national or international jurisdictions that are actually enforcing them. The CILFIT criteria also remain disconnected from other EU law methods of enforcing the obligation to refer that is incumbent on courts of last instance. Such considerations are, in a way, hardly surprising: since CILFIT (as it stands) cannot be reasonably enforced, other yardsticks must be introduced.

131.

In my view, the Court should, in the first step, re-affirm the purpose and the scope of the duty to refer under the third paragraph of Article 267 TFEU as stated already in Hoffmann-Laroche. In the second step, however, CILFIT must be revisited in order for the possible exceptions to match that duty.

132.

Hoffmann-Laroche stated that the aim of the obligation to submit a reference is ‘to prevent a body of national case-law that is not in accordance with the rules of EU law from being established in any of the Member States’. ( 112 ) Following that logic, three points can be singled out. First, what shall be achieved is uniform interpretation, not correct application. Second, the focus shall be on ‘a body of case-law’, not on the correctness of the outcome in every single case. Third, it concerns case-law divergences within any of the Member States and of course, a fortiori, across the European Union. Logically, both types of divergence must be avoided: from the point of view of systemic coherence of EU law, if one Member State or parts of it, or even a system of jurisdiction within that Member State, were to operate according to rules other than those applied elsewhere in the European Union, then there would be no uniform interpretation within the Union.

133.

In other words, the focus of the duty to refer must change, from there being no subjective reasonable doubt as to the correct application of EU law in respect of the outcome of the specific case, to there being, an objective divergence detected in the case-law at the national level thereby threating the uniform interpretation of EU law within the European Union. In this way, the focus also moves from establishing the correct answer in the case before the national court to the identification of the right kind of questions.

134.

Following that logic, I would suggest that, under the third paragraph of Article 267 TFEU, a court or a tribunal of a Member State against whose decisions there is no judicial remedy under national law is to refer the case to the Court of Justice provided that it raises (i) a general issue of interpretation of EU law (as opposed to its application); (ii) to which there is objectively more than one reasonably possible interpretation; (iii) for which the answer cannot be inferred from the existing case-law of the Court of Justice (or with regard to which the referring court wishes to depart from that case-law).

135.

Stated in this way, the duty to refer already contains within itself its own exceptions. The potential exception(s) to the duty to refer will be triggered only if one of the three cumulative conditions for there being an issue of interpretation of EU law that is subject to the duty to refer is not met. However, should a national court of last instance be of the view that, even if faced with an issue of interpretation of EU law in the main proceedings, one of the three conditions is not met, that court is obliged to identify clearly which one of the three conditions is not met and state the reasons why it believes that to be the case.

136.

Before exploring these three conditions in further detail, I wish to point out two important elements.

137.

First, the fact that there is no duty to make a reference in a given case under the third paragraph of Article 267 TFEU certainly does not preclude a court of last instance from seeking assistance from the Court under the second paragraph of Article 267 TFEU, should it deem that to be necessary in order to settle the individual case before it. The absence of an obligation to do something does not preclude the possibility to do that same thing. A case that may not fall within the scope of the obligation of the third paragraph of Article 267 TFEU may still be subsumed under the second paragraph: ‘a court or tribunal of a Member State against whose decisions there is no judicial remedy’ is in any event and remains ‘a court or tribunal of a Member State’. The former is a logical subset of the latter.

138.

Second, it might again be repeated that everything which has been stated, and all that follows, concern exclusively preliminary questions on interpretation. The obligation to make a reference concerning the validity of an EU act remains subject to a strict duty to refer, without exceptions. ( 113 )

139.

This condition is obvious at first sight. Indeed, it might even be suggested that this is what the preliminary ruling procedure has always been about. The reality is nonetheless somewhat more diverse. The Court’s occasional insistence on there being no reasonable doubt as to the correct application of EU law in the case before the court of last instance has led a number of courts to also submit factual and rather specific issues to the Court. Three examples may illustrate this phenomenon.

140.

First, the Court previously provided an interpretation of the concept of ‘extraordinary circumstances’ within the meaning of Article 5(3) of Regulation (EC) No 261/2004 ( 114 ) at an early stage, starting with Wallentin-Hermann, where the Court stated that the concept of ‘extraordinary circumstances’ refers to ‘an event which … is not inherent in the normal exercise of the activity of the air carrier concerned and is beyond the actual control of that carrier’. ( 115 ) Certainly, the likelihood is that there will be a number of additional cases which confirm and make clear the exact scope of such a definition. However, are all new factual scenarios, each merely asking whether or not a particular set of facts (premissa minor) may once again be subsumed under the interpretative definition already provided for in EU law (premissa maior), really instances of interpretation of EU law? In a number of cases thereafter, the Court was invited to categorise as ‘extraordinary circumstances’: a collision between an aircraft and a bird; ( 116 ) the closure of part of European airspace as a result of the eruption of an Icelandic volcano; ( 117 ) the unruly behaviour of a passenger which justified the pilot diverting the flight ‘unless the operating air carrier contributed to the occurrence of that behaviour or failed to take appropriate measures in view of the warning signs of such behaviour’; ( 118 ) or damage to an aircraft tyre caused by a foreign object, such as loose debris, lying on an airport runway. ( 119 )

141.

Second, in a similar vein, in matters of insurance against civil liability in respect of the use of motor vehicles, the Court has issued several judgments on the concept of ‘use of vehicles’ within the meaning of Article 3 of Directive 2009/103/EC. ( 120 ) According to the Court, that concept covers any use of a vehicle that is consistent with the normal function of that vehicle, ( 121 ) that is to say any use of that vehicle as a means of transport. ( 122 ) However, subsequently, the Court has been invited to confirm whether the following factual situations also amount to the ‘use of vehicles’: ‘the manoeuvre of a tractor in the courtyard of a farm in order to bring the trailer attached to that tractor into a barn’; ( 123 )‘the situation in which an agricultural tractor has been involved in an accident when its principal function, at the time of that accident, was not to serve as a means of transport but to generate, as a machine for carrying out work, the motive power necessary to drive the pump of a herbicide sprayer’; ( 124 )‘a situation in which the passenger of a vehicle parked in a car park, in opening the door of that vehicle, scraped against and damaged the vehicle parked next to it’; ( 125 ) or a ‘situation … in which a vehicle parked in a private garage of a building, used in accordance with its function as a means of transport, has caught fire, giving rise to a fire which originated in the electrical circuit of that vehicle and caused damage to that building, even though that vehicle has not been moved for more than 24 hours before the fire occurred’. ( 126 )

142.

Third, a final illustration might be drawn from the concept of ‘working time’ under Directive 2003/88/EC. ( 127 ) According to established case-law, the key requirement for the classification as ‘working time’ within the meaning of Directive 2003/88 is that the worker be physically present at the place determined by the employer and be available to the employer in order to be able to provide the appropriate services immediately in case of need. Those obligations, which make it impossible for the workers concerned to choose the place where they stay during stand-by periods, must be regarded as coming within the ambit of the performance of their duties. ( 128 ) However, following that general definition, the Court has again been asked to effectively subsume under that definition factually detailed situations concerning various forms of on-call or stand-by duties of medical and other emergency personnel. The Court thus held that time spent on call constitutes working time if the employee is required to be at the health centre; ( 129 ) even where the person concerned is permitted to rest at his place of work during the periods when his services are not required; ( 130 ) or that the concept of ‘working time’ should apply to ‘a situation in which a [firefighter] is obliged to spend stand-by time at his home, to be available there to his employer and to be able to reach his place of work within 8 minutes’. ( 131 )

143.

Certainly, in all these areas, the Court has no doubt provided a useful answer to the national courts. However, I cannot but agree again with Advocate General Jacobs that ‘detailed answers to very specific questions will not always promote … uniform application. Such answers may merely provoke further questions’. ( 132 )

144.

In my recent Opinion in Van Ameyde España, another case concerning motor vehicle civil liability insurance, submitted by the Tribunal Supremo (Supreme Court, Spain), I sought to suggest, in more concrete terms, where the (by definition, rather elusive) line between the interpretation and application of EU law could lie. I cannot but refer to that Opinion as an actual case study. ( 133 ) For the purpose of the first condition of where the obligation to make a reference ought properly to lie, I would add the following.

145.

The duty to refer ought to be triggered whenever a national court of last instance is confronted with an issue of interpretation of EU law, formulated at a reasonable and appropriate level of abstraction. That level of abstraction is logically defined by the scope and purpose of the legal provision at issue. In the particular context of (indeterminate) legal concepts of EU law, the task of the Court is to provide an interpretation of that concept. Its application, including the subsumption of specific facts under that definition, is a matter of application of EU law.

146.

Certainly, a mere application may quickly turn into interpretation, provided, for instance, that the referring court were to invite the Court to narrow, broaden, qualify, or depart from the definition already provided. However, provided that that is indeed what is being requested, the referring court ought to state this element clearly, explaining specifically why the case being referred is more than merely another confirmation (and in this sense application) of the previously stated premissa maior.

147.

Moreover, in this sense and in that dimension, the interpretation sought ought to be one of general or generalisable impact. The question of interpretation ought to concern a general, potentially recurring question of interpretation of EU law. I wish to emphasise that this is not a criterion relating to the legal significance or importance of the question asked. It is another, much simpler question for the national court: is the question I am now facing likely to arise again, either before me, or before my colleagues in other Member States? Should I seek guidance from the Court in the interest of uniform interpretation of EU law?

148.

Such a mental exercise not only helps with the formulation of the question to be referred at an appropriate level of abstraction, because it forces one to identify general, transversal problems, but it also helps with the weeding out of very narrow, singular, or one-off cases which, even if potentially raising an issue of interpretation of EU law, are simply of no general, structural impact. In any case, it is safe to assume that the majority of the national courts of last instance are very well versed in such thinking, although perhaps primarily within the confines of their own systems. A similar type of thinking and reflection ought simply to be applied to the broader, larger level of the EU legal order.

149.

In summary, I am certainly not suggesting that we get rid of one unicorn in order for it to be replaced immediately by another: ( 134 ) the idea that there is any clear border between where interpretation ends and applications starts and vice versa. However, the aim of the duty to refer is to ensure the uniform interpretation of EU law, not the correct application of that law. Thus, the uniformity sought is not and has never been at the level of the single outcome of each individual case, but at the level of the legal rules to be applied. This means that, logically, while there is a reasonable degree of uniformity of the legal rules, there may be diversity in terms of specific outcomes.

150.

Is it objectively possible to interpret a specific point of EU law in different ways? A necessary part of any of the conditions of the duty to refer will always be a study of the alternatives. That being said, in contrast to the inherently internal and subjective uncertainty in the form of any reasonable doubt as to the outcome of an individual case, it is the existence of plausible alternatives that place similar considerations on a more external and objective footing. Hence the suggestion to refocus attention from a mere ‘I do not know’ to ‘these are the alternatives I have to choose from’.

151.

I stress the importance of the adjective ‘possible’ or ‘plausible’. It is certainly not suggested that a referring court would be obliged to set out the alternatives fully, to reason through them, or even to explain which one it favours. ( 135 ) However, there ought to be more than a mere subjective doubt, or even absence of knowledge, to justify a duty to refer.

152.

However, where there are two or more potential interpretations proposed before the national court of last instance, the duty to refer becomes strict. After all, it is precisely for this reason that the duty was originally imposed: it is to apply when there is a choice between a number of options ‘on the table’ available. The following scenarios are particularly pertinent in providing an illustration as to when such a situation may arise.

153.

First, there are examples where different interpretations of the same rule have been adopted in final decisions of national courts. Any other court of last instance being seised of the dispute involving the same element of EU law and having detected a divergence in interpretation of the same rule is bound to make a reference, in order to clarify which of the strands of case-law is in fact correct. It does not matter whether such a detected divergence is within the same Member State or across several Member States. Indeed, I also fail to see the logic under which divergence in the strands of national case-law only would not be sufficient for there objectively to be divergent case-law on the same matter within the Union ( 136 ) unless it were to be a singular variant of reverse discrimination. After all, the logic underlying Hoffmann-Laroche is that there be deviation from the common approach ‘in any of the Member States’, and for a good reason: there can be no uniform interpretation if parts of a Member State simply go their own way?

154.

However, I would place emphasis on the final existing decisions issued at a national level. What is sought is an articulated difference in interpretation at the horizontal level, be it within the same Member State (different courts of appeal, different chambers or formations of the highest jurisdictions) or across the Member States. The same might not be necessarily true of one single set of proceedings which is still being decided. In such a situation, in which, for instance, the interpretation embraced by the first-instance court differs from that adopted by the court of appeal, with the case now pending before a supreme court, might also involve two or more approaches to the same rule. However, in contrast to a situation where there are divergent interpretations across individual final decisions of various courts, this does not automatically mean more reasonable and plausible ways of interpreting the same EU law rule. Within one and the same procedure, it indeed cannot be excluded that one court simply made a mistake. By contrast, divergence in case-law across different proceedings is no longer a mistake, but a structural problem for EU law and national law alike.

155.

Second, all other situations, even including that of different interpretations adopted within one and the same set of proceedings, are then simply to be assessed in the light of each individual case. In the proceedings before the court of last instance, were there indeed plausible alternatives as to the interpretation of the same rules, regardless of where they came from? They could have been in the submissions of the parties; they could have come from the various actors involved in the national proceedings; or indeed could have stemmed from the various previous judgments within the same proceedings, where the difference was not a mistake but rather the manifestation of plausible alternatives.

156.

In addition, such doubts as to the choice between possible interpretations of one rule may naturally always come from the national court itself. However, in view of all that has been stated above in this section, I wish to emphasise one rather important point: reality and realism. One can hardly expect the national courts of last instance to suddenly turn into comparative EU law research centres, whereby they themselves, so to say ex officio, carry out searches into the case-law of other national courts in other Member States, or start looking proactively into finding interpretative problems.

157.

That being said, what certainly can be expected of national courts of last instance is that they should recognise that there is objectively a divergence in legal interpretation if that divergence has been brought expressly to their attention by any of the actors in the proceedings before them, in particular by the parties themselves. Provided that there is indeed a divergence in possible legal interpretation, demonstrated by possible alternatives, then, using the CILFIT terminology, (reasonable) doubt can be considered to have been objectively and externally established in the dispute before them and the duty to refer in the interest of ensuring uniform interpretation of EU law cannot then be ignored.

158.

There is likely to be no shortage of (heated) discussion about what exactly, in a specific scenario, is ‘established case-law’, and whether everybody understands exactly what has been ‘established’ in the same way. However, at a conceptual level, it is indeed Da Costa, ( 137 ) as retained and expanded by CILFIT, ( 138 ) which requires the least revisiting. A national court of last instance is obliged to make a reference if it faces a new question of interpretation of EU law, or a question which cannot fully be settled on the basis of existing case-law of the Court, or where it is appropriate to invite the Court to clarify or to revisit some of its previous decisions.

159.

Put simply, the second condition and the present, third condition are facets of the same concern – the first condition concerns the uniform interpretation of EU law – in different types of directions and with different actors. The second condition concerns horizontal coherence and national case-law, while the third, focuses on the decisions of the Court of Justice and their effects.

160.

A national court of last instance is therefore not obliged to refer a question of interpretation of EU law if the same provision has already been interpreted by the Court. The same applies in the situation where previous decisions of the Court, handed down in whatever type of proceedings, have already provided sufficient interpretative guidance allowing the national court to settle the issue before it with confidence on the basis of existing case-law.

161.

I would just conclude with three additional clarifications in this regard.

162.

First, it might be worth pointing out again the logical connection between the third condition and the first one: what is to be established and should clearly be inferable from the case-law is the EU law rule to be applied, not the outcome of the specific case. Thus again, for example, the EU law interpretation of what constitutes ‘extraordinary circumstance’ was established the moment the definition of that event was stated and confirmed by the Court. Unless a court of last instance were to face a situation in which it wished to revisit, refine, or to depart from that definition, it is simply to apply it without being obliged to seek guidance on whether, for example, in addition to all the factual scenarios already covered by the Court, ( 139 ) extraordinary circumstances would also apply in the event of a deer or a naked man running on the runway.

163.

Second, it may be fair to acknowledge that, while the language of the Court uses terminology such as ‘established case-law’, or ‘previous decisions’, this may in fact sometimes amount to a single precedent. Certainly, that all depends on the content, context and clarity of the EU legal rule that was supposed to be established in the past decision. ( 140 ) However, a clearly articulated legal position, even if stated only once (and thus hardly amounting to genuine ‘established case-law’ in true civil law tradition, in which it would have been repeated many times before becoming effectively binding) may, from the outset, exempt a national court of last instance from making a reference.

164.

Third, a national court, and in particular a national court of last instance, is always allowed to invite the Court to adapt, refine, clarify, or even depart from its previous decisions. However, if a national court of last instance wishes to depart from the interpretation of EU law previously adopted by the Court, that national court is under an obligation to make a reference, explaining to the Court the reasons for its disagreement and, ideally, setting out what ought to be, in the view of the referring court, the proper approach. ( 141 )

165.

For the sake of completeness, it might be added that such ‘clarification’ scenario may not only include situations in which a national court indeed wishes the Court to modify its case-law, ( 142 ) but also instances of divergence within the case-law of the Court detected by a national court called upon to apply that guidance at the national level. In such (hopefully rare) situations, there is a duty to refer, precisely in the interest of uniform interpretation of EU law across the European Union, in order to prevent divergences between national courts with some of them relying on one strand of the case-law of the Court, while others effectively apply a different one.

166.

In the proposal made in this Opinion, the scope of the duty to refer already contains the exceptions. They are two sides of the same coin. For there to be a duty to refer, all three conditions outlined in this section must be met. In order for the duty not to arise (or, depending on the construction, for an exception to be triggered), the presence of one of the three conditions must be negated: either there is no issue of interpretation of EU law; there is only one reasonably possible interpretation of the EU law at issue; or an answer can be found in the existing case-law of the Court.

167.

That being said, there is, in any case, a transversal duty, or even a fourth condition: regardless of which of the three conditions is to be invoked by the national court of last instance, that court is under an obligation to provide adequate reasons for its conclusion that the case before it does not fall under the obligation to make a reference pursuant to the third paragraph of Article 267 TFEU.

168.

Naturally, there is no universal yardstick as to what is an adequate and thus sufficient statement of reasons. It all depends on the nature of the case, its complexity, and above all the arguments brought before the deciding court and those contained in the case file. However, and in any event, provided that a relevant point of EU law has indeed been raised before a national court of last instance, that court is under an obligation to state clearly and specifically which of the three conditions (exceptions) is supposed to apply to that case and to provide at least a summary explanation as to why that is the case.

169.

I believe that it is important to emphasise this obligation clearly. A general, vague, and largely unsubstantiated reference to ‘acte clair’ or CILFIT, without any real and case-specific reasoning being provided as to why exactly there is no duty to refer on the merits of the case in question, does not meet that minimal requirement. ( 143 ) By contrast, in the admittedly rather objective and external approach to the duty to refer in the interest of the systemic uniformity in interpretation advocated in this Opinion, the reasoning to be provided by the national court of last instance is a reaction to the elements brought to the attention of that court, either by the parties, or because they clearly emanate from the proceedings and case file itself. The duty to state reasons thus naturally correlates with the duty of a judge to react to all the relevant circumstances and arguments brought before him or her.

170.

Put simply, a judge is not obliged to go out and search for interpretative problems which may possibly arise with respect to a particular EU law provision. However, the interpretative problems brought before him or her, in particular those invoked by the parties, cannot be ‘swept under the carpet’ without proper reasoning, by simply stating in one sentence that ‘all is clear and beyond reasonable doubt’.

171.

Finally, in my view, the duty to state adequate reasons, although it is likely to already flow from relevant national rules, is also an obligation imposed by EU law under Article 47 of the Charter. It is rather logical that if a point of EU law is validly raised during national judicial proceedings, then that case is likely to be, in so far as concerns the applicability of other elements of EU law, within the scope of EU law. There is, moreover, Article 267 TFEU itself, which would be applicable in such a case. Therefore, such a case, and any national decision taken in relation to it, are clearly an instance of implementation of EU law under Article 51(1) of the Charter.

172.

I shall deliberately leave it there. Issues relating to the enforcement of the duty to refer are perhaps a matter for future cases. However, before such matters are capable of being discussed in the future, the scope and the nature of the duty to refer must first be considerably revisited. Only if and when that first step is undertaken, it may become relevant to discuss the subsequent matter of remedies.

173.

I conclude with three general remarks.

174.

First, why is it necessary to revisit the CILFIT file now? Indeed, having read the compelling Opinions of all my learned predecessors quoted throughout this Opinion, it is clear that the Court is not likely to reopen that file lightly. Moreover, with tongue in cheek, it could be suggested that, since CILFIT has not worked for 40 years, a few more years or decades will make no difference. There is also a certain beautiful simplicity and wisdom in inertia, particularly once the system, as a whole, finds some sort of equilibrium. Indeed, returning to the metaphor used in the introduction to this Opinion, it may be wiser to let sleeping dogs lie. If they should wake, no one knows whom they may bite.

175.

Appealing as such thinking may be, it has its clear limits. It is not healthy for the institutional authority and legitimacy of any court to be considered irrelevant, as the guidance coming from the centre is being disregarded, and for good reason. In addition, if such a lack of relevance touches upon one of the crucial parameters of the entire judicial system, for the proper working of which and for at least some of its enforcement, that system must rely on others, a rather unhealthy rule scepticism may come into being and eventually spread to other areas and other issues. Finally, tensions are likely to emerge if such a careful equilibrium is thrown off balance by a sudden burst of selective enforcement of such rules, in which those subject to such an enforcement might correctly ask: why us? In this way, with the dogs now awake, it becomes imperative to revisit the rules so that they may be enforced equally against everyone.

176.

Second, it may be suggested that by focusing on the macro or public function of the duty to refer, the proposal made in this Opinion neglects the individual litigants and their subjective rights. Moreover, when separated from subjective doubts in the individual case, the definition of the scope of the duty to refer and the exceptions to that duty become vague in their articulation, relying on abstract concepts such as divergence in interpretation.

177.

In my view, it was precisely the lack of clarity in conceptually distinguishing between the macro/micro (public/private) function and objective/subjective conditions of the duty to refer that account for the problems of CILFIT. Moreover, the choice made in CILFIT was unique in yet another way: the duty to refer was made subject primarily to individual and subjective conditions of the specific case pending before a court of last instance without, however, the individuals being given any rights whatsoever to enforce that duty.

178.

If, as proposed here, the predominantly systemic and structural nature of the duty to refer were acknowledged, which would then be based on more objective considerations of the systemic needs arising in the individual case, that could possibly all provide individual litigants with much more on which to base arguments in an individual case than subjective judicial doubts. Moreover, while in part detached from the circumstances of an individual case, the conditions suggested here are, as to their nature and articulation, in fact much more precise than those present in CILFIT. In addition, in contrast to the ongoing and endless discussions of whether the exceptions in CILFIT do in fact amount to a checklist or simply to a tool kit, the conditions proposed here clearly form a checklist, with confirmation of the correlating duty to specifically and adequately state reasons.

179.

Third and final, would such a ‘loosening’ of CILFIT, no longer focusing on the correct application of EU law in every case before a last-instance court, not amount to the abdication of the key responsibility of the Court in terms of ensuring unity and uniformity of EU law?

180.

I have taken quite some space in this Opinion to try to explain why I believe that the CILFIT uniformity with regard to the correct application of EU law in each individual case is a myth. In view of the decentralised and diffused nature of the EU judicial system, the best that can ever be achieved is a reasonable uniformity in the interpretation of EU law, with that type of uniformity already being a rather tall order. As to the uniformity in application and outcomes, the answer is rather simple: ‘no man can lose what he never had’. ( 144 )

181.

I recommend that the Court answer the first question referred to it by the Consiglio di Stato (Council of State, Italy) as follows:

Under the third paragraph of Article 267 TFEU, a court or a tribunal of a Member State against whose decisions there is no judicial remedy under national law is to refer the case to the Court of Justice, provided that, first, that case raises a general issue of interpretation of EU law, which may, second, be reasonably interpreted in more than one possible way and, third, the way in which the EU law at issue is to be interpreted cannot be inferred from the existing case-law of the Court of Justice. Should such a national court or tribunal, before which an issue of interpretation of EU law has been raised, decide not to submit a request for a preliminary ruling pursuant to that provision, it is obliged to state adequate reasons to explain which of the three conditions is not met and why.