A.  Introduction

1  The autonomy of European Union law is a constitutional trait of the legal system of the European Union (‘EU’) (a ‘notion structurante’ according to Marti (2023) 55) and is applied as a general principle of EU law by the Court of Justice of the European Union (‘CJEU’; European Union, Court of Justice and General Court). Autonomy of EU law refers to EU law constituting a self-referential legal system without normative gaps, a legal order with its own norms, notions, methods, and concepts separate from both the national legal orders of member states (‘internal autonomy’) and international law (‘external autonomy’), although EU law has benefited and borrowed from both. As an autonomous legal system, EU law determines by and for itself not only how a norm of its legal order is interpreted, but also whether a norm is valid in its legal order and to what extent a norm stemming from a different legal order can enter EU law, for example when it violates fundamental human rights provided for under EU law. EU law thus occupies its own constitutional space, separate from though not unrelated to and indeed intertwined with the national law of the EU’s 27 member states on the one hand and international law on the other (Lenaerts, Gutiérrez-Fons, and Adam (2021) 48).

2  The notion of ‘autonomy of EU law’ shows some commonalities to two notions of international law, even though it differs in key aspects. First of all, it is related to the ‘certain autonomy’ that international organizations possess as subjects of international law (Legality of the Use by a State of Nuclear Weapons in Armed Conflict (Advisory Opinion) (1996) para. 19; see also Nuclear Weapons Advisory Opinions; International Organizations or Institutions, General Aspects), arguably an attribute that is the equivalent of a state’s sovereignty. It is part of the definition of international organizations that they possess a degree of internal autonomy from their member states and make their own decisions (International Organizations or Institutions, Decision-Making Process). Similarly, they benefit from external autonomy from the body of general international law (Konstantinidis (2024) 96, 99; General International Law (Principles, Rules, and Standards)). The second international law notion of which the autonomy of EU law is reminiscent is that of self-contained regimes (Self-Contained Regime). But where the two international law notions are relative, allowing for a larger or smaller degree of autonomy/self-containment, autonomy of EU law makes a sweeping, absolute claim. Autonomy of EU law establishes EU law as its own, separate legal order stemming from an independent source distinct from international and national law. That claim does not deny that EU law was linked at birth to the international legal order and that of member states, but the umbilical cord linking EU law to them has been cut (Eckes (2020)). Today’s notion of autonomy of EU law is more closely linked to other concepts that emphasize the sui generis character of the EU legal order, such as supranational law. Despite the absolutist theoretical claim underlying the notion of autonomy of EU law, however, the scope of the notion in practice is the subject of constant recalibration and is, by necessity, relative.

3  The objective of the principle of autonomy of EU law is, at its core, to protect the distinctive character and independent decision-making of the EU. This objective can be broken down into a number of more detailed functions of the autonomy of EU law. Normatively, it protects the core values and principles of the legal system of the EU and allows an independent interpretation of the rules of that legal system, ensuring that these rules are applied uniformly across all member states despite their different legal systems and thereby furthering the rule of law within the EU. Institutionally, autonomy safeguards the institutional framework of the Union including mutual trust between member states as well as the division of competences. The protection of the role of the Union’s judiciary stands out in this regard (on the typology see Vajda (2019)). While the different functions can be clearly identified in theory, practice in this regard is somewhat more confusing. First of all, the different functions cannot always be neatly separated. For example, the role of the EU’s judiciary is vital for the unity and uniformity of EU law across the Union. Furthermore, the application of the autonomy of EU law in different contexts serves the various functions of autonomy to different extents. This helps explain the complexity and at times apparent incoherence of the notion of autonomy. Thus, for example, the internal autonomy of EU law—its independence from member states’ laws—protects the uniformity of EU law. External autonomy—its independence from international law, which turns the Court into a gatekeeper deciding under what conditions international law enters the EU system (Odermatt (2016) 18)—only does so in some constellations, namely where member states have signed differing treaties, as is the case with investment treaties (Investments, Bilateral Treaties; Investments, International Protection). In many other instances external autonomy primarily pursues the protection of core values of the EU, namely where it prevents an external rule from being applied in the EU system because of human rights.

4  The autonomy of EU law is not explicitly mentioned in the EU treaties. The CJEU has developed the notion in its case law, drawing on a variety of provisions of the EU treaties, which support different aspects of the autonomy of EU law to a greater or lesser extent. The case law first focused on the internal autonomy of EU law, i.e. the autonomy vis-à-vis member states’ legal systems, featuring such landmark cases as Van Gend en Loos v Administratie der Belastingen Case 26/62 (1963) and Costa v ENEL Case 6/64 (1964) and establishing Community (and later EU) law as an autonomous order applying uniformly, according to the interpretation the Community legal order itself provides, in the whole Community with direct effect and primacy (Van Gend & Loos Case; Costa v ENEL Case). Later cases tackled the external autonomy of the EU legal system, i.e. the relationship between EU law and international law (European Union Law and International Law). In decisions such as Kadi and Al Barakaat International Foundation v Council and Commission Joined Cases C-402/05 P and C-415/05 P (2008) (Kadi Case), Creation of a Unified Patent Litigation System Opinion 1/09 (2011), Accession of the European Union to the European Convention for the Protection of Human Rights and Fundamental Freedoms (Opinion 2/13) (2014), and Achmea Case C-284/16 (2018) (Achmea Case), the Court held that the EU’s and member states’ treaties may not affect the autonomy of EU law, with significant consequences for the reception of international law within the EU legal system and the EU’s capacity to enter into treaties, particularly where those establish a system of judicial dispute settlement (European Union, Party to International Agreements). In its case law on external autonomy, the Court of Justice has increasingly applied the notion of autonomy as a general principle with its own normative content. With regard to the interpretation of EU law, the Court has defended its role as an authoritative interpreter of EU law both internally vis-à-vis national courts (e.g. Foto-Frost v Hauptzollamt Lübeck-Ost Case 314/85 (1987)) and externally vis-à-vis international tribunals (e.g. European Economic Area I Opinion 1/91 (1991)).

5  The case law on internal autonomy is largely accepted, though it is still constantly tested by member states’ political systems and courts (European Union Law and Domestic (Municipal) Law). In contrast, the Court of Justice’s case law on external autonomy, particularly on accession to the European Convention for the Protection of Human Rights and Fundamental Freedoms (1950) (‘ECHR’) and on the relationship between EU law and investment law, has been subjected to much criticism. The Court is accused of exaggerating the importance of autonomy, erroneously trying to insulate EU law from international law (Eeckhout (2015) 992) causing rather than resolving problems for the rule of law (Kochenov and Lavranos (2020)), and of protecting above all its own constitutional role in the guise of autonomy (Spaventa (2015) 56). Some scholars also criticize the application of autonomy as a general principle, preferring to limit the notion to that of a descriptive umbrella term (Moreno-Lax and Ziegler (2022)). So far, the Court has continued to pursue its path undeterred. As will be shown later, at least in two regards the criticism is hard to refute: while external autonomy is supposed to protect only essential characteristics of EU law, the Court at times seems to embrace a limitless notion of what these characteristics are, possibly because of an unwarranted parallelism between internal and external autonomy. Secondly, the protection of the Court’s interpretative authority also suffers from a loss of focus, risking a similar unwarranted extension (Moreno-Lax and Ziegler (2022)).

6  This entry will first retrace the development of autonomy in the case law of the Court of Justice (see Section B), before tackling in turn the internal (Section C) and external aspects of autonomy (Section D). The distinction between these two aspects is well-established and reflects the history of the case law (Pernice (2013)). The entry will then cover interpretative autonomy (Section E), which has both internal and external aspects (Marti (2023) 63–67), before evaluating the notion and offering an outlook (Section F).

B.  The Development of Autonomy in the Case Law

7  The Court of Justice started to develop its case law on autonomy with regard to the notion’s internal aspect, i.e. the relationship between EU law and member states’ laws, developing in parallel the notion of autonomy and two of the most fundamental constitutional characteristics of EU law, direct effect and primacy, which are intertwined with the internal notion of autonomy. In later cases, the Court also expounded on external autonomy (concerning the relationship between EU law and international law), analysing international agreements the EU enters into for compliance with the autonomy of the EU legal order and deducing legal requirements from the need to prevent adverse effects on that autonomy.

8  The first explicit reference to the notion of autonomy with regard to the Community legal order dates back to 1956, when Advocate General Maurice Lagrange referred to the ‘caractère autonome du droit du Traité’ to argue against applying national time limits in a case where the treaties failed to provide for such limits. The term was translated as ‘independent’ in the English translation of his opinion (Opinion of AG Lagrange in Mirossevich v High Authority Case 10/55 (1956)).

9  It was not until the later cases of Van Gend en Loos and Costa v ENEL, however, that the Court fleshed out the concept of internal autonomy, hand in hand with establishing the direct effect and primacy of Community law. In Van Gend en Loos the Court famously held that unlike in international law, where states decide for themselves whether and to what extent treaties produce direct effect, in the Community legal order this decision was a matter of Community Law. The Court decided that the standstill clause for customs duties (Art. 12 Treaty Establishing the European Economic Community; ‘EEC Treaty’) produces direct effect so that individuals can invoke the provision before national courts. The Court reached this conclusion in substance by appealing to the autonomy of the Community legal order, even though it would not use the term itself until a year later in Costa v ENEL. ‘[T]he Community constitutes’, wrote the Court,

10  This quote has become the locus classicus for the description of EU law as a new legal order that is separate and distinct from municipal law, i.e. an autonomous legal order. The modifier ‘of international law’ shows that at the time the Court’s focus was on internal autonomy and it arguably still viewed Community law as part of international law. The Court dropped the modifier in later cases, illustrating that EU law is autonomous from both municipal and international law. The Court arrived at its conclusion by analysing the spirit and general scheme of the EEC Treaty, describing it as ‘more’ than a normal treaty creating mutual obligations between states, as it establishes a common market of direct concern to interested parties in the Community, refers to the peoples and not just governments in its preamble, establishes institutions with sovereign rights, the exercise of which affects citizens, includes nationals through the European Parliament and the Economic and Social Committee, and provides for the uniform interpretation of Community law through the Court of Justice via the preliminary reference procedure.

11  The Court used the term ‘autonomous’ for the first time in Costa v ENEL when discussing the particular nature of Community law, though it was again translated as ‘independent’ in the English version of the judgment. In the case, the Court established the primacy of Community law over national law, arguing with the special nature of the EEC Treaty, which establishes its own legal system ‘issue d’une source autonome’—‘stemming from … an independent source of law’. Without primacy, the force of Community law would vary from member state to member state, resulting in discrimination and preventing the attainment of the objectives of the Treaty. Obligations of the Treaty would become contingent rather than unconditional. Notably, the Treaties’ explicit provision that regulations are directly applicable in member states could be nullified by member state law (Costa v ENEL 594).

12  Through these groundbreaking cases the Court of Justice emancipated EU law from a status in which all of its legitimacy is based on delegation of sovereignty by member states within the bounds of their constitutional orders and established EU law as its own, self-standing legal order. This new order is based on the founding treaties but goes beyond a normal treaty-based order, limiting the sovereign rights of states and interacting directly with the people of the Union.

13  In later cases, the Court began to rely on the autonomy of the EU legal system as a principle governing international treaties concluded by the EU and its member states. A first inkling of this approach (though it did not mention autonomy by name) can be seen in Draft Agreement establishing a European Laying-up Fund for Inland Waterway Vessels Opinion 1/76 (1977), where the Court held that the draft Agreement on the establishment of a European Laying-up Fund for Inland Waterway Vessels was incompatible with the EEC Treaty. The first full analysis of whether an agreement undermines the autonomy of the Community legal order is contained in Opinion 1/91 with regard to the proposed system of courts of the Agreement creating the European Economic Area (EEA).

14  The Court refined its approach in later cases, the most significant of which are Agreement on the Establishment of a European Common Aviation Area Opinion 1/00 (2002), the Kadi Case (2008), which tackled the application of restrictive measures imposed by the UN Security Council in the Community legal system, Opinion 1/09 (2011) on the Unified patent litigation system, and Opinion 2/13 (2014) on the Accession of the EU to the ECHR, as well as the line of cases on intra-EU bilateral investment treaties starting with the Achmea Case (2018) through to CETA Opinion 1/17 (2019) on the provisions on investor-state dispute settlement in the Comprehensive Economic and Trade Agreement between Canada, the EU and its Member States (‘CETA’).

15  In addition to these leading cases on the autonomy of EU law, numerous other decisions of the Court of Justice have helped clarify the relationship of EU law with international and member state law, whether or not they refer to the concept of autonomy, such as Commission of the EEC v Luxembourg and Belgium Joined Cases 90/63 and 91/63 (1964), in which the Court held that the public international law notions of countermeasures or the defence of a non-fulfilled reciprocal treaty (exceptio non adimpleti contractus; Exception of Non-Performance; Reciprocity) do not apply in EU law, or Portugal v Council Case C-149/96 (1999), where the Court held WTO law in principle not to be directly applicable within the EU legal order (World Trade Organization (WTO)).

16  As the doctrine stands today, the Court explicitly recognizes that the EU legal order is autonomous with regard to both the law of member states and international law. It bases the autonomy on the unique constitutional framework and characteristics of the EU. The precise description of this framework varies from case to case, but its essential elements are the independent source of EU law in the Treaties, its primacy and direct effect, the allocation of powers, its constitutional structure and institutional framework, and the EU’s founding values set out in Article 2 Treaty on European Union (‘TEU’). The Court has, at times, gone beyond these essential elements, stating that the constitutional framework also encompasses ‘the general principles of EU law, the provisions of the Charter, and the provisions of the EU and FEU Treaties’ (Opinion 1/17 para. 110). The protection of the autonomy and the constitutional characteristics of the EU is ensured by the judicial system established by the Treaties that safeguard the full application of EU law in all member states as well as its consistency and uniformity of interpretation, namely through the national courts and the Court of Justice of the European Union. In this regard, the preliminary ruling procedure under Article 267 Treaty on the Functioning of the European Union (‘TFEU’) plays a particularly important role (Opinion 1/17 para. 111; Preliminary Ruling: European Court of Justice (ECJ)).

C.  Autonomy of EU Law from National Law

17  As described above, the Court formulated the basic principle of autonomy of the EU legal order vis-à-vis national law in Van Gend en Loos and Costa v ENEL, establishing EU law as its own legal system, separate from those of member states, even though it interacts with those in a number of ways. The concept of the autonomy of the EU legal system both enables the direct effect and primacy of EU law and is in turn reinforced by those two doctrines, which have become part and parcel of the autonomy of EU law. A corollary of the primacy of EU law over member states’ law including constitutional law is that only EU law itself can serve as a yardstick for the validity of EU measures (Internationale Handelsgesellschaft mbH v Einfuhr- und Vorratsstelle für Getreide und Futtermittel Case 11/70 (1970) para. 3; Internationale Handelsgesellschaft Case).

18  Autonomy from national law thus understood enables a uniform interpretation and application of EU law across all member states—independent from member states’ law, ensuring equality of both member states and Union citizens before the law (Art. 9 TEU, Art. 20 Charter of Fundamental Rights of the European Union; Charter of Fundamental Rights of the European Union (2000)) as well as enabling mutual trust between member states of the EU, which is based on the presumption of compliance by other member states with EU law, including the fundamental values of EU law set out in Article 2 TEU.

19  The internal autonomy of the EU legal order is in practice largely accepted. Significant disputes remain with regard to both the theoretical foundation of the primacy of EU law and the recent trend towards direct defiance of the primacy of EU law by some national apex courts. The dispute as to the theoretical conceptualization of primacy stems from different mandates of national apex courts and the CJEU: While the latter is tasked with interpreting the autonomous legal system of the EU, national apex courts view the law through the lens of national (often constitutional) law and thus through the lens of a different, equally autonomous system, for which the legal order of the EU is derived by way of delegation of sovereign rights under national constitutional orders. Member states thus remain the masters of the treaties. Where CJEU and national courts interact and question each other in a mutually respectful manner, the difference in outlook can be productive and improve the legal system through a dialogue of courts, as illustrated by the development of fundamental rights within the EU system. Where the different outlook leads to a wholesale refusal to accept the primacy of EU law, it puts the EU system at risk (Hestermeyer (2019) 459).

D.  Autonomy of EU Law from International Law

20  When the Court changed its description of the Community legal order from a ‘new legal order of international law’ in Van Gend en Loos to ‘its own legal system’ in Costa v ENEL, it had not only cut the umbilical cord between EU law and the legal systems of member states, but also between EU law and public international law: Today, EU law is autonomous both with regard to domestic and international law. External autonomy to some extent puts the EU in a similar position vis-à-vis international law as states. Autonomy from international law does not mean that EU law does not interact with international law or stopped being open to the application of norms of international law, nor does it liberate the EU from the obligations under international law it has as a subject of international law in its own right (Art. 47 TEU; see also Art. 3(5) TEU).

21  The autonomy of EU law from international law protects the essential characteristics of the EU and its law against interference from the international legal system. This protection encompasses both the institutional framework of the Union and—much as with internal autonomy—the interpretation of its norms, ensuring their uniform application across the Union and preventing actors from being exposed to (and bound by) varying, inconsistent interpretations of the same provisions.

22  Questions of external autonomy are raised in two situations. The first of these arises when the EU wants to sign an international agreement and the Court analyses whether that agreement violates the principle of autonomy, as it first did explicitly in Opinion 1/91. As a subject of international law, the EU can sign international agreements, which, according to the case law of the Court, once concluded ‘form an integral part of [EU] law’ (Haegemann v Belgian State Case 181/73 (1974)). However, such agreements may not have an adverse effect on the autonomy of EU law. Second, the principle of autonomy can come into play when the question is raised whether a rule stemming from international law can apply in the EU legal order. It cannot do so if it contravenes the principle of autonomy.

23  Three aspects of the essential characteristics of EU law are particularly relevant in the analysis of a possible violation of the external autonomy of EU law: substantive principles of the EU constitutional framework, the allocation of power to and within the EU constitutional framework, and the judicial system set up by the Treaties. These three aspects are closely linked and cannot be neatly separated. Thus, the allocation of power and the judicial system of the EU are part of its constitutional framework and closely connected to its substantive principles. The protection of the judicial system is instrumental for the whole framework of autonomy and is one aspect of institutional autonomy, which in turn plays an instrumental role for the substantive constitutional framework.

24  The first way in which a treaty or a provision of international law can affect autonomy relates to the substantive principles of the EU constitutional framework, most importantly to the values the Union is founded upon according to Article 2 TEU, in particular human rights. The Court of Justice had to tackle the issue in the Kadi Case, where it was asked to rule on the legality of restrictive measures imposed by the EU on individuals under a UN Charter Chapter VII sanctions regime (United Nations Charter; Sanctions). The Court acknowledged the binding nature of the sanctions under international law but argued that

The Court accordingly reviewed the lawfulness of the EU measure implementing the sanctions and provided for the protection of fundamental rights despite the primacy that obligations under the UN Charter enjoy in the field of international law according to Article 103 UN Charter. The Court thus made it clear that international law obligations can only become part of the EU legal order where they comply with fundamental rights, an approach that was confirmed by the Court’s analysis of the agreement between Canada and the EU on the transfer and processing of Passenger Name Record data in EU–Canada PNR Agreement Opinion 1/15 (2017).

25  The second relevant aspect of autonomy relates to the allocation of powers and functions to and within the EU and its institutions. According to the case law of the Court, ‘the essential character of the powers of the [Union] and its institutions as conceived in the Treat[ies must] remain unaltered’ by treaties the Union and/or its member states sign, including the division of powers between the EU and the member states (Opinion 1/00 para. 12). Similarly, agreements cannot set up institutions, in particular tribunals, that may hand down awards preventing EU institutions from operating in accordance with the EU constitutional framework (Opinion 1/17 para. 119).

26  This issue is of decisive importance for the design of dispute settlement mechanisms for international agreements to which both the EU and its member states are parties, i.e. so-called ‘mixed’ agreements (European Union, Mixed Agreements). The logical consequence of mixity is that a dispute with a non-EU party arising in the area of EU competences is a dispute with the EU, whereas a dispute relating to member state competences is one with member states. The determination of the correct party to the dispute boils down to the determination to whom EU law allocates a competence. As the Court of Justice ruled with regard to both the EEA Court and the co-respondent mechanism designed for EU accession to the ECHR, it would violate the autonomy of the EU legal order to leave this determination to a dispute settlement system outside of the EU (Opinion 1/91 paras 33 et seq.; Opinion 2/13).

27  In the context of investor-state dispute settlement in CETA, the Court of Justice has also expressed its concerns regarding the possible impact of a tribunal’s powers on the democratically legitimized legislature. It indicated it would not tolerate endowing a tribunal with jurisdiction to issue awards that double-guess the democratically set level of protection of public interests in the EU, forcing the legislature to amend or withdraw legislation (Opinion 1/17 paras 149 et seq.).

28  The third relevant aspect of autonomy protects the judicial system set up by the Treaties, which serves to provide for a uniform and consistent interpretation, application, and enforcement of EU law across the Union and ensures effective judicial protection, thereby in turn preserving and protecting the constitutional characteristics and framework of the EU. According to Article 19 TEU the judicial system of the Union consists of both member states’ courts and the CJEU. An essential element of the judicial structure of the EU linking the two and effectively protecting and enabling the uniform interpretation and application of EU law is the ability of member states’ courts to request a preliminary ruling from the CJEU under Article 267 TFEU on the interpretation of EU law or the validity of EU acts (Opinion 2/13 paras 174 et seq.).

29  This third aspect of external autonomy is of particular relevance for the creation of dispute settlement mechanisms in international treaties. According to the case law of the Court of Justice, an international agreement signed by member states and/or the EU can, in principle, establish ‘a court responsible for the interpretation of its provisions and whose decisions are binding on the EU institutions, including the Court of Justice of the European Union’, even though a treaty concluded by the EU forms an integral part of EU law (République de Moldavie v Komstroy LLC Case C-741/19 (2021) para. 61; Opinion 2/13 para. 182). However, such an international agreement cannot affect the essential allocation of responsibilities to the CJEU (just as it cannot do so with other institutions). This includes, first of all, the exclusive jurisdiction of the Court under Article 344 TFEU, under which ‘Member States undertake not to submit a dispute concerning the interpretation or application of the EC Treaty to any method of settlement other than those provided for therein’ (Commission v Ireland Case C-459/03 (2006) para. 123; MOX Plant Arbitration and Cases). Second, a court set up by an international treaty may not deprive member state courts of the power to request preliminary rulings under Article 267 TFEU and thus interfere with the system of cooperation between the CJEU and member state courts enabling the latter to be an essential element in the correct and uniform application of EU law. This is true even where the new international court is endowed with the power to make preliminary references unless the new court is a court common to member states (only) and an integrated element of the judicial system of the EU (Opinion 1/09 paras 65–89). Third, a treaty may not create tribunals with the power of ‘binding the EU and its institutions, in the exercise of their internal powers, to a particular interpretation of the rules of EU law’ (Opinion 2/13 para. 184). Where a tribunal is empowered to interpret EU law in such a manner, it must be integrated into the EU system so that through member state courts and preliminary references the uniform interpretation of EU law is safeguarded. The Court thus protects the exclusive authoritative interpretation of EU law through the EU judicial system.

30  The Court has found fault with a number of envisaged (or existing) international tribunals with regard to the autonomy of the EU legal system, amongst them the EEA Court in Opinion 1/91 (leading, ultimately, to the creation of the EFTA Court and European Economic Area II Opinion 1/92 (1992); Court of Justice of the European Free Trade Association (EFTA Court)), the European and Community Patents Court in Opinion 1/09, the accession to the ECHR in Opinion 2/13, and intra-EU investor-state dispute settlement whether provided for in a bilateral investment treaty, ad hoc arbitration, or under the Energy Charter Treaty in Achmea, République de Moldavie v Komstroy, and PL Holdings Case C-109/20 (2021). The approach of the Court has been much criticized, arguing amongst others that under the guise of autonomy the CJEU protects its own judicial supremacy and that it creates a system that lacks openness towards international law. Criticism has been particularly severe with regard to the EU’s accession to the ECHR and the Court’s approach to international investment arbitration (de Witte (2014), Spaventa (2015), Tomuschat (2015), Kochenov and Lavranos (2020)).

31  In the case of the ECHR the Court rejected the draft accession agreement for the accession of the EU to the ECHR despite an explicit obligation for the Union to accede in Article 6(2) TEU. Besides the mentioned flaws of the co-respondent mechanism the Court found fault, amongst others, (i) with the possibility that the European Court of Human Rights (ECtHR) could call into question the Court of Justice’s finding in relation to the scope ratione materiae of fundamental rights of the EU with regard to Article 53 Charter of Fundamental Rights and its coordination with Article 53 ECHR, (ii) the requirement imposed on member states to check that other member states have observed fundamental rights despite the EU law principle of mutual trust, (iii) the failure to clarify the relationship between the advisory opinion mechanism under Protocol No 16 to the ECHR and the preliminary ruling procedure provided for in Article 267 TFEU entailing a risk that the preliminary ruling procedure can be circumvented (Advisory Opinion: European Court of Human Rights (ECtHR)), and (iv) the failure to exclude ECtHR jurisdiction under Article 33 ECHR over disputes between member states or between member states and the EU in contravention of Article 344 TFEU (Opinion 2/13 paras 179–214). Some critics regarded the Court as simply unwilling to submit to the ECtHR. For them, the Court’s approach to autonomy is building a state-like system that, unlike most modern European states, regards external control of human rights as undue interference in its sovereignty (Eeckhout (2015); Tomuschat (2015)). More charitable observers viewed the opinion as an attempt to adapt the accession process to the non-state character of the EU, requiring further work and creating a task list for negotiators to resolve (Halberstam (2015)). Fears that the opinion had closed the door to accession seem overstated, as negotiations for accession resumed in 2020, dividing the issues raised by Opinion 2/13 into four baskets. In 2023 negotiators reached a provisional agreement on the issues in three of these baskets, with the EU intending to resolve the remaining ‘basket 4’ issue internally (COE ‘18th Meeting of the CDDH Ad Hoc Negotiation Group (‘46+1’) on the Accession of the European Union to the European Convention on Human Rights’ Meeting Report (2023); Council of Europe (COE)).

32  With regard to investor-state dispute settlement, the Court of Justice’s landmark judgment in Achmea, which for the first time held an investor-state dispute settlement clause in an intra-EU bilateral investment treaty to violate the autonomy of EU law, proceeded in three steps: It determined that the investment tribunal was empowered to interpret and apply EU law as part of the applicable law, that the tribunal is not situated within the judicial system of the EU and hence not empowered to make a reference to the Court of Justice for a preliminary ruling, and finally that the limited possibility of review of an arbitral award by national courts within the EU does not suffice to ensure the proper application of EU law. Criticism ranged from the technical detail, like the fact that the Court focused on the theoretical possibility of (rather than an actual) application of EU law, to the general—arguing that Achmea contributed to the deficient judicial protection in some member states (Kochenov and Lavranos (2020)). The most problematic aspect of the Court’s approach in Achmea is the focus of its autonomy analysis on the possibility of an interpretation of EU law by an investment tribunal rather than the specific characteristics of intra-EU investor-state dispute settlement (Hestermeyer (2020) 88–89).

33  The case law on external autonomy continues to be much contested. Overall, two weaknesses of the case law of the Court of Justice are particularly prominent. The first one is that the Court repeatedly emphasizes that autonomy relates to the essential characteristics of the EU. In theory, this establishes autonomy as a firewall, blocking international law only where particularly important elements of EU law are put at risk. However, the Court then dilutes the notion of ‘essential characteristics’ by adding an ever more expansive list explaining where autonomy resides, which reads as though at least all primary EU law is included. The reason for this expansion might be a parallelism between internal and external autonomy, as the former unproblematically includes the primacy of all EU law over member states’ law. However, external autonomy should operate in a more limited manner.

34  The second weakness relates to the proper definition of the ultimate interpretative authority over EU law, which will also be discussed in the next section. The interpretation of EU law by an external tribunal only interferes with autonomy where the tribunal makes internally binding pronouncements on EU law, for example, settling disputes on the application or interpretation of EU law between member states and thus violating Article 344 TFEU, but not where the external tribunal merely considers EU law (Moreno-Lax and Ziegler (2022)).

E.  Autonomous Interpretation of EU Law

35  The third aspect of autonomy, the autonomous interpretation of EU law, cuts across both the internal autonomy of EU law (ensuring an interpretation autonomous from member states’ legal systems) and its external autonomy (arriving at an interpretation autonomous from international courts and tribunals).

36  EU law concepts have, in principle, an autonomous meaning, independent from concepts of national law and international law, even where those are identically worded. The interpretation of EU law follows its own, EU-law–specific methodology, with the ultimate interpretative authority vested in the CJEU (see CILFIT v Ministero della Sanità Case 283/81 (1982); Foto-Frost; Opinion 1/91). The following paragraphs describe the purpose, content, and limits of this aspect of autonomy as well as the institutional mechanism ensuring the autonomous interpretation of EU law.

37  The purpose of the autonomous interpretation of EU law is to ensure a uniform interpretation and application of EU law across all member states. Such a uniform application ultimately serves the rule of law and its essential component, equality before the law. It also enables the creation of an area of freedom, security, and justice without internal frontiers as well as an internal market (Art. 3(2), (3) TEU), subject to the same rules in all member states. As the Court of Justice stated in Ekro Case 327/82 (1984):

38  As to the content of the principle, autonomous interpretation means that EU law has its own method of interpreting concepts within its legal order, which arrive at EU-law–specific results that may not align with the interpretation of the same terms in other legal orders. The methods EU law employs find their origin in the traditions of both member states and the international legal order. Much like in many national systems and international law, the wording of the provision (though in a multilingual setting), its context, and its object and purpose as well as the travaux préparatoires are the decisive elements of interpretation (Interpretation in International Law; Contextual Interpretation). However, the importance attached to each of these is specific to the EU legal order (Lenaerts and Gutiérrez-Fons (2013)). Where EU concepts are identically worded to international law concepts and yet given a different meaning, this is often due to the context and object and purpose of EU law. For example, EU law provisions resembling (and historically based on) GATT provisions have developed very differently, following the integrational logic of EU law (General Agreement on Tariffs and Trade (1947 and 1994)). Similarly, the Court of Justice has emphasized the different interpretation of EU law provisions and identically worded provisions of the EEA in Opinion 1/91 (paras 14 et seq.).

39  Autonomous interpretation is not unlimited, however. It does not imply that the meaning of concepts in different legal orders does not affect their meaning in EU law. As to international law, where international agreements of the EU are interpreted as such, even though they are an integral part of the EU legal order they are interpreted according to the methodologies of international law (European Union Law and International Law). Given the rank accorded to international law in the EU system, secondary EU law has to be interpreted as far as possible in a manner consistent with the obligations of the EU under international law (HK Danmark Joined Cases C-335/11 and C-337/11 ((2013) para. 29). As to national law, Article 19 TEU grants the CJEU the authority to engage in a comparative study of the laws of member states and appeal to general principles common to the legal systems of the member states (Lenaerts and Gutiérrez-Fons (2013) 35). Indeed, in some cases, the EU Treaties themselves accord specific relevance to national or international law, creating instances of permeability.

40  Institutionally, the autonomous interpretation of EU law is safeguarded by the judicial machinery under Article 19 TEU and the principles of direct effect and primacy, which imply that member states’ courts apply and enforce EU law, the uniform interpretation of which is safeguarded in the final instance by the power of any court or tribunal of any member state to request a preliminary ruling on the interpretation of EU law or the validity of acts of the EU by the CJEU.

F.  Evaluation and Outlook

41  The autonomy of EU law is a core trait of EU law, asserting its distinctive character and distancing it from both national and international law, while retaining a certain amount of permeability with regard to both orders.

42  Internal autonomy is, at least in principle, widely accepted. Nevertheless, the tug of war between the autonomous orders of member states and the autonomous EU legal order over who has the final say in the composite EU/member states multilevel system—member states or the Union—will continue. It has gained in explosiveness with the ongoing rule of law crises. In the face of the risks of a destabilization of the system, it is worthwhile to distinguish actual attacks on the primacy of EU law from the mere theoretical posturing about who has the final say. The latter debate does not deserve the space it is sometimes accorded. The theoretical tenet that there has to be one final voice determining interpretation in a legal system does not reflect the reality of any legal system, whether a federal one, a multilevel one, or a unified national state one.

43  When it comes to external autonomy the concept presents a complex conundrum (Odermatt (2016)). On the one hand, external autonomy is a necessary corollary of a solidifying system increasingly becoming an actor in the international order. On the other, it limits the capacity of the EU to open its system to the broader world of international law, even though the EU itself is arguably the result of such an opening up of member states’ systems. The solution to the conundrum lies in a clearer definition of autonomy.