THE EQUAL AUTHENTICITY OF OFFICIAL LANGUAGE VERSIONS OF EUROPEAN LEGISLATION IN LIGHT OF THEIR CONSIDERATION BY THE COURT OF JUSTICE OF THE EUROPEAN UNION

The aim of the paper is to analyse the judgments of the Court of Justice of the European Union in order to determine whether all official language versions of EU legislation are considered in the course of interpretation. The Court itself has stated in many of its judgments1 that all official versions should be taken into account in the interpretation process as they all have an equally authentic character. Moreover, the comparison of all official language versions is a common feature for all methods of reconciliation of differences between various language versions that had been distinguished in literature. The author has examined 80 judgments in order to determine whether the comparison of all language versions really forms an important part of the interpretation conducted by the Court, or if it should be regarded as an unfulfilled wish expressed by the scholars and the Court itself.


Introduction
This study examines judgments of the Court of Justice of the European Union and opinions of the Advocates-General in cases where there have been problems concerning differences of wording (and of meaning) between versions of a legal act formulated in different official EU languages.The author will briefly describe the established methods of reconciliation of differences between various official versions of EU legislation, search for their core features, and analyse how the theoretical ideas are applied in the judicial reality of the Court of Justice of the European Union (hereinafter referred to as the Court or the CJEU).A quantitative comparison of selected judgments will be presented and analysed in order to verify whether the Court examines all official language versions of EU legislation that is claimed to ensure the realisation of the principle of equal authenticity of them.
The article concludes several parts of the research undertaken by the author within the project "The certainty of law and multilingual EU legislation", being conducted at the University of Silesia as a part of her PhD studies.3

INTERPRETATION OF MULTILINGUAL LEGISLATION BY THE CJEU
The concept of legal multilingualism and the unique approach towards it taken by the CJEU has been analysed by many scholars (inter alios Šarčević 2000 and 2002, Derlén  2009, Doczekalska 2006 and 2009, Fryźlewicz 2008).All agree that the European Union's diversity, proclaimed in its official motto "Unity in Diversity"4, has a linguistic dimension, expressed through multilingualism 5 .Doczekalska (2009, 344) underlines that it: "not only guarantees the citizens of the European Union access to European law, but is also essential for a uniform application of EU law in Member States".This uniform application can be achieved only when European legal acts are understood in the same way in each Member State, regardless of the plurality of languages in which EU legislation is expressed.
The multilingualism of EU law makes its interpretation different from that of legislation drafted in only one language.The official versions all have the same legal value, therefore, according to the most popular approach 6 , each one should be considered in the course of interpretation.The factual consideration of all versions by the CJEU has been given much attention in the conducted study and will therefore be presented separately.
The legal basis for multilingualism in the European Union has been established in its legislation, in the Treaties (art.55 TUE), as well as in secondary law (Regulation 1/587 ).Further principles concerning the equal value of all official languages and the need for consideration and comparison of all official versions of a disputed provision of EU law have been developed through legal science and case law.One of the most important judgments is CILFIT8 , where the Court stated that: "Community legislation is drafted in several languages and that the different language versions are all equally authentic; an interpretation of a provision of Community law thus involves a comparison of the different language versions"9 .
It is important to underline that the Court does not regard the chronological order in which different versions were created as important in the course of its interpretation.Meanings represented in one or some versions are not accepted on the grounds that that they were the first to be drafted.Such an approach can be explained through the theory of original texts (as well as by the principle of equal authenticity of all official versions10 ).
Another important consideration for this approach is in the method of drafting official language versions in the EU.The proposal of a legal act is usually presented in French or English (sometimes in German).The proposal is then immediately translated into other official languages.All the following preparatory work and discussion can be held in any official language and the act that is finally passed may have been drafted simultaneously in more than one language.It is not possible to separate the process of drafting and translation (particularly with regards to the working languages of the Commission, namely French, English and German 11 ).This process is called the codrafting of EU law in many languages 12 .For these reasons, no "pure" originals nor "pure" translations exist.The CJEU and the Advocates-General go even further by taking into account language versions that did not exist at the time of enactment of a particular piece of legislation.A. Doczekalska (2006, 19) presented some very interesting examples of this phenomenon 13.For the aforementioned reasons, differences between particular versions cannot be solved in favour of the original.No single language version would prevail over the others just because it had served as a source for the other versions in the process of legal drafting.
National courts of Member States may submit an inquiry for a preliminary ruling to the CJEU when they have doubts concerning the interpretation of Community Law14 .The CJEU will then provide a uniform interpretation of EU legislation to enable it to be understood in the same way in all Member States.It can therefore be seen that the judgments of the CJEU serve as important material in understanding the interpretation of multilingual law.
Linguists and legal scholars have made attempts to clarify and classify the methods applied by the CJEU in the process of reconciliation of differences observed by comparison of official versions.Detailed descriptions of various approaches can be found in Derlén (2009, 36-43).The reconciliation process has been analysed from a purely linguistic perspective (Loehr, as referred to by Derlén 2009, 42), with the result that all differences are to be reconciled in the process of interpretation only by linguistic means.According to Derlén (2009, 50) the reconciliation process requires the application of another, non-linguistic methods of interpretation (e.g teleological interpretation.) Elements of quantitative analyses are sometimes used by the Court in the process of reconciliation of different language versions (the Court applies the provision interpreted in harmony with the majority of official versions)15 .As it had been noted by Jacobs (2004, 304), numerical considerations are taken into account by the Court to "justify the conclusion that the version or versions in minority contained a drafting error16, or that the ambiguity inherent in those versions must be resolved in a particular way".According to Advocate-General Sharpston, the Court gives preference to the meaning represented in the majority of official versions when it corresponds with the objectives of the regulation17 .
Other proposals for reconciliation of different language versions refer to the preference for the most clear meaning, or the meaning that is the most liberal for an individual.
The most important non-linguistic method of interpretation, stated to be useful for reconciliation of differences between authentic versions, is the teleological method.Various authors distinguish different types (for example, Derlén -2009, 43-48, writes about radical teleological method, classical reconciliation and examination of the purpose), but the core issue of its application is the choice of the interpretation that is the most accurate to the scheme and the aim of the provision in question.This choice usually occurs after the examination and comparison of the official text of a legal act.Therefore, the first step to be taken is the consideration of wording of the provision in different languages.The Court's aim is to achieve uniform interpretation and resolve any problems with the application of the disputed provision in all Member States.

EXAMINATION OF ALL OFFICIAL VERSIONS AS THE CORE FEATURE OF ALL RECONCILIATION METHODS AND A SINE QUA NON CONDITION FOR THE EQUAL AUTHENTICITY OF OFFICIAL VERSIONS
Regardless of the classification and character of a certain method of reconciliation, all methods are based on a presumption that the Court's analysis includes all official versions of the disputed provision.If this was not the case, teleological reconciliation methods (aiming at the choice of meaning that is most accurate to the general aim and scheme of the analysed provision) could not be strictly applied because the choice would be limited to the examined language versions.Potentially, versions not examined could have proved to realise the aim and scheme of the provision better than the chosen ones.Consequently, when the examination is limited, the comparison is not complete and the choice not fully justifiedthere is a potentially more acceptable solution, not discovered by the Court.
The presumption of examination of all official language versions is the sine qua non condition for quantitative methods, where the interpretation in accordance with the majority of official language versions is considered to be justified.
The question of a right to rely on only one version versus the obligation of the courts (CJEU and national courts) to analyse and compare all official versions had been undertaken by several authors.Some observations contain a distinction between position of a national courts and CJEU. 18The consideration of all official versions of a disputed provision is treated as an obligation of CJEU because it has the necessary resources (i.e.multi-national staff, and specialist services such as lawyer linguists and translators).The same analysis is not required from national courts as they do not have these resources at their disposal 19 .National courts should request the CJEU for a preliminary ruling in cases where there is an interpretative doubt concerning differences between language versions of the provision in question.
The dominant approach by authors 20, confirmed by numerous CJEU judgments 21 and opinions of Advocates-General 22 , is that the principle of equal authenticity of all language versions requires them all to be taken into account in the process of interpretation.Some authors claim that the Court tries to make comparisons of all official versions. 23However, some authors state that the examination of all the versions is simply impossible, first and foremost for national courts, which is why they should ask the CJEU for a preliminary ruling (Jacobs, 2004,303).The most clear example of an explicit doubt in the factual comparison of official language versions by the CJEU itself is the statement of Advocate-General Jacobs: "I do not think that the CIĹFIT judgment should be regarded as requiring the national courts to examine any Community measure in every one of the official Community languages (now numbering eleven-or twelve, if the Treaties and certain other basic texts are in issue).That would involve in many cases a disproportionate effort on the part of the national courts; moreover reference to all the language versions of Community provisions is a method which appears rarely to be applied by the Court of Justice itself, although it is far better placed to do so than the national courts."[emphasis added] 24 However, in the judgment C-498/03, the Court confirmed that all language versions should be considered in course of interpretation of EU legislation 25.In practice, the reconciliation choices made in course of adjudication already limit the material equal authenticity of certain language versions.All versions can never be equal because only some of them will be approved in the course of the decision-making process.Such "inequality" is acceptable, because particular language versions often represent opposite or contradictory meanings and the Court is required to make a decision by choosing one of them.With respect to the equal authenticity principle, the choice of a certain meaning requires a justification, usually based on other criteria, such as the expression of the aim and scheme of the disputed provision in case of teleological reconciliation methods.
The impossibility of realising the equal authenticity principle in its material meaning does not prevent the Court from realising its formal meaning.Formal equal authenticity enables the Court to choose the meaning that remains in accordance with any of the official versions.In author's opinion, this formal equal authenticity requires the consideration of all official versions to ensure that they all have a potential influence on the final decision.Even if certain versions do not prove to represent the correct meaning, they will at least be taken into account.The author further considers this to be a sine qua non condition for the realisation of equal authenticity principle.Without it, there is no reason to claim equal authenticity of official language versions.The conducted study determines whether and how far this principle is fulfilled by the CJEU.

Methodology
Before the presenting the study results a few words need to be said about the selection of the analysed material and interpretative choices made to classify the judgments.
One of the biggest problems in selecting judgments for the study was the number of cases where the Court did not openly state all the details its interpretative processes and decision-making choices.In cases where the analysed versions had not been mentioned by name, but indicated as "all the versions", "all other languages", etc., the author counted the number of official EU languages at the time of adjudication as being the number examined.Where the multilingual comparison had been described with vague expressions like "other languages" or "some versions", or where one or several versions had been specified clearly as examples of wording that is relevant to an unspecified number of further versions, the cases are classified in the "no data" category.Versions that appeared in the Advocate-General's opinion or a statement of a party, are counted as as "examined" (i.e. the wording of examined versions is known by the Court regardless of whether the CJEU conducted the analysis itself).Overall, any indication of a language version by a party, Advocate-General, or the Court is counted as an "examined version".

The Choice of analysed material
The author found 96 judgments using the search engine at the official Internet site of the CJEU using the keywords "wersja językowa" ("language version" in Polish ) Only CJEU judgments (no cases decided by General Court or Civil Service Tribunal) delivered before 30.06.2013 have been taken into consideration. 26Interestingly, the results have revealed that many judgments frequently cited in the literature, especially those passed at the very beginning of EU, still have not been translated into Polish.To complete the picture, nine judgments widely commented on in in the literature and quoted in this article 27 were added to the study material.A total of 105 judgments were pre-selected for further analysis.
Of these, some of judgments from the Internet search turned out not to deal with multilingualism of legislation or interpretation of multilingual law at all 28 , or concerned other language issues (such as the language of a TV programme 29 , the language of a contract 30 , or the the lack of publication of apiece of EU legislation in an official language 31 ).Consequently, 25 out of 105 selected judgments proved to be useless for the study.The remaining 80 were chosen for further examination.
The fact that the number of official EU languages has not been constant in course of the Union's history had an important impact on the study.New languages are usually added at the time of accession of new Member States (but sometimes the status of official language is granted later, as with Irish and Maltese).Due to the changing number of official versions, the results of the analysis are displayed in relation to the total number of relevant cases (%) based on the time of adjudication rather than the time of enactment of a the particular legislation (this follows the CJEU's own approach towards considering versions created after a piece of legislation has come into force).
Presentation of proportional values enables comparisons of the number of language versions and the frequency of the examination of certain language versions.The first part of the study aims establishing how many language versions had been examined by the Court in particular cases.In order to estimate how many language versions existed at the time of the adjudication, the judgments are divided into groups representing the consecutive extensions to the number of official EU languages.

 The examined language versions in analysed judgments
Table 1 presents the number of examined language versions in analysed cases.The study reveals that in the most cases only a few language versions were examined.Nearly 70% of the judgments examined fewer than half the official languages, most far fewer.The last line contains 28 cases where the analysed versions cannot be estimated due to the wording of the judgment ("no data").The Court has used expressions that do not allow to determine which versions had been analysed nor the number of examined versions. 32However, although the exact number of examined versions was not given, it was possible to count the minimum number of languages examined (using mentions by the Court or Advocate-General as examples for an unspecified number of versions).It can be stated that in only 2 out of 28 "no data" judgments the Court or Advocate-General mentioned 50% of official language versions or more.For reasons of vagueness by the Court explained above, it is difficult to be precise, but overall, it is estimated that the Court examined half or more of the official language versions in only 25 (23 +2 "no data" judgments) out of 80 examined cases.
Only in 19 judgments did the Court declared to have studied all language versions.Very often there was neither any indication of the differences between various versions nor any detailed comparison of wording.Of these, only 4 judgments (all been passed when the number of official EU languages equalled 11 33 ) contain an analysis of the provision in question in all the official languages.In a further 4 judgments the Advocates-General conducted comparison of all official versions 34 .However, in the remaining 11 cases the Court or Advocates-General only declared to have examined all versions, using expressions like "all versions" and, "none of versions".Croatian * 2013 -------------Table 2 shows the official languages and the number of cases in the study relevant to each language group.Each group contains languages granted official EU language status in the same year.The total number of cases relevant to one language group differs from the total of other groups.For the oldest EU languages (group I), the number of cases equals 80 (all examined cases).For Finnish and Swedish (group V), the total number equals 78, because 2 cases had been adjudicated before those languages became official in 1995.For the youngest official EU languages (group VII), the number of relevant cases equals only 49 (all cases adjudicated before the recognition of those languages in 2007 had to be put aside).Croatian only became an official EU language in 2013, after the Court's adjudication of analysed cases, and therefore could not be taken into account.The total number of languages for the study was therefore 23.

Certain language versions examination and their influence on final decisions
The variation of the number of relevant cases for each group makes a direct comparison more difficult.To overcome this, the results for each official language are presented proportionally to the number of cases relevant to their particular language groups.The general observed tendency is that the examination rate corresponds with the period of time in which a certain language has served as an official EU language.The "older" the official language, the higher the examination rate.Irish, one of the youngest official languages, was the least frequently examined, appearing in just under 15% of cases.Results for Bulgarian and Romanian also introduced in 2007 were significantly higher (20% and 25%), in line with the group VI -languages that became official in 2004).
The results for the groups I-V are more differential.The working languages of the Commission, French and German, the oldest of the EU official languages, as well as English were the most frequently examined (94 % for French, 93% for German and 89 % for English).Interestingly, Italian and Dutch, also amongst the original EU languages, have been examined much less frequently (71 % and 65%), comparable to Spanish and Portuguese, that became official only in 1986.The examination of Danish and Greek, which also have also long tradition as official EU languages, showed results similar to the more recently added Swedish and Finnish languages.

Conclusion
The examination of all official language versions of a certain provision is stated to be the core feature of all distinguished methods of interpretation of multilingual EU law.It is very important for the justification of choices made by the CJEU in the process of reconciliation of differences between the wording of particular versions where a provision is disputed.
In the author's opinion, the most finding of her work is that the examination of all official language versions of a legal act is hardly ever employed by the CJEU.Therefore, it should be regarded rather as a wish expressed by the Court itself and by the scholars, than as a real judicial practice.
The author agrees that the examination of all official versions would grant equal authenticity to all official versions of a legal act (by assuring them an influence on the final decision), but its practical use is impossible for private entities, or national courts (due to a large and constantly growing number of authentic languages).As this study has showed, even the CJEU, with al the necessary resources to conduct such a comparison, does not use this possibility very often.

Table 1 .
The examined versions in analysed cases.

Table 2 .
The frequency of examination of particular language versions.