Ορισμένες πρώτες σκέψεις σχετικά με την απόφαση 221/2026 του Αρείου Πάγου.

Η υπόθεση αφορούσε ένα μακροχρόνιο έργο αιολικής ενέργειας, στο οποίο συμμετείχαν μια ελληνική εταιρεία ανάπτυξης έργων και ένας Ισπανός επενδυτής. Παρότι το έργο επρόκειτο να κατασκευαστεί στην Ελλάδα, η μεταξύ τους σχέση διεπόταν από συμβάσεις συνταγμένες στην αγγλική γλώσσα.

Όταν η σχέση αυτή οδηγήθηκε τελικά σε ρήξη, ο Ισπανός επενδυτής επέδωσε εξώδικη δήλωση με την οποία γνωστοποιούσε ότι είχε αποφασίσει να «terminate the … Contracts and withdraw from the Project».

Η έννοια και οι έννομες συνέπειες της δήλωσης αυτής αποτέλεσαν αντικείμενο της δίκης. Τα δικαστήρια που εξέτασαν την υπόθεση ασχολήθηκαν με το συμβατικό λεκτικό και τις ελληνικές μεταφράσεις.

Για περισσότερα, διαβάστε το άρθρο. ⬇️

Hellenic Supreme Court Judgment No. 221/2026 [1] contains an unusually useful example of translation becoming part of the legal problem in a cross-border dispute.

The underlying case concerned a long- running wind-energy project involving a Greek developer and a Spanish investor. Their relationship was governed by English-language contracts. When that relationship eventually broke down, the Spanish investor served an extrajudicial notice stating that it had decided to “terminate the … Contracts and withdraw from the Project”. The meaning and legal effect of that declaration became part of the litigation. The courts had to consider the contractual language, the Greek translations relied upon by the parties and the legal mechanism that the notice was attempting to exercise

The way the documents had been translated had become relevant to the court’s legal analysis of the dispute.

When terminology becomes part of the argument

The central translation issue concerned the English terms “withdrawal” and “withdraw”. The parties had produced different Greek renderings of the same contractual language. In the Framework Agreement, the Greek developer translated the heading “Withdrawal” as «Υπαναχώρηση». The Spanish investor’s translation used «Απόσυρση». Within the operative provisions of the Spanish investor’s own translation, however, “withdraw from their obligations” was rendered using «υπαναχώρηση», and the same term appeared again in the provisions describing the consequences of exercising that right. The Court of Appeal expressly identified the inconsistency.

That finding has an immediate practical significance for lawyers. A translation commissioned during the early stages of a matter may still be sitting in the file several years later when the meaning of the underlying provision becomes disputed. The terminology adopted at that earlier stage can then be compared with later translations of notices, correspondence, pleadings and exhibits. If the same contractual mechanism has been rendered differently across the record, counsel may have to explain why. In Judgment No. 221/2026, the inconsistency was visible to the court and formed part of its assessment of the competing terminology.

The legal significance of the words also depended on context. The Court of Appeal considered «υπαναχώρηση», «απόσυρση» and «αποχώρηση» in light of the contractual mechanism created by the relevant provisions and the legal consequences attached to it. It concluded that, in this particular contractual setting, “withdrawal” and “withdraw” were to be understood from the viewpoint of Greek law as «υπαναχώρηση» and «υπαναχωρώ». That reasoning is useful precisely because it is contextual. An English verb that appears straightforward at first sight may acquire a much narrower legal significance once it is placed inside a clause dealing with the exercise of a particular contractual right.

For counsel, this creates a clear risk-management issue. A translation can contain language that is perfectly intelligible and still create difficulty because the chosen terminology does not sit coherently with the legal mechanism expressed elsewhere in the contract. The problem becomes more serious where several translations have been commissioned at different times and the terminology has drifted. Once the dispute reaches court, those choices become part of the documentary history of the case.

Terminology drift

Cross-border disputes rarely arrive fully formed. They develop over time[2]. An agreement may be translated when advice is first sought. Correspondence may be translated months later. An extrajudicial notice may be prepared during an escalation of the dispute. Pleadings, exhibits and witness material may follow. Appeals and cassation proceedings can generate further translation requirements. The same underlying contractual concepts may therefore reappear in documents produced years apart.

That makes translation consistency a case-management issue. The important question is whether recurring concepts have been treated coherently across the file. Counsel should know how a key term was translated in the agreement, whether the same terminology was used in subsequent correspondence, and whether a later translation introduces a different Greek expression for the same contractual mechanism. If a change is deliberate, it should be capable of explanation. If it has happened accidentally because a later document was translated without access to the earlier material, the file has acquired an avoidable vulnerability[3].

Judgment No. 221/2026 gives this point real substance. The court did not examine one translated sentence in isolation. It moved across the Framework Agreement, the Project Agreement, the parties’ respective translations and the December 2015 notice. Earlier terminology informed the analysis of the later declaration. The documents formed a chain. For lawyers managing a bilingual dispute, the translation process needs to reflect that same chain.

This has implications for the way instructions are given. A translator dealing with an operative notice may need the underlying agreement and its existing translation. A translator working on pleadings may need to see the terminology already adopted in the evidence. Where a concept is genuinely ambiguous, counsel may need to decide how that ambiguity should be handled before the translation is finalised. Those steps reduce the chance that terminology will develop independently across documents that will ultimately be read together.

The legal effect of a few words

The December 2015 notice brought these issues into focus. Its English text stated that the Spanish investor had decided to “terminate the … Contracts and withdraw from the Project”. The Court of Appeal considered that the wording raised a question about the juridical character of the declaration, because the two expressions potentially pointed towards different legal mechanisms.

The court then examined the consequences contemplated by the notice. Among other things, the Spanish investor proposed the transfer of its shares in the project company for a nominal price. That consequence corresponded with the contractual mechanism provided for in Article 5.2 of the Framework Agreement in connection with withdrawal. The court therefore characterised the declaration as an exercise of contractual υπαναχώρηση.

For a lawyer reviewing a translation, this is an important point of method. Operative language has to be read through the legal function of the document. A notice may be exercising a right, triggering a contractual consequence or attempting to bring a legal relationship to an end. A translation choice affecting the characterisation of that act deserves a different level of scrutiny from a purely descriptive passage in the factual record.

That distinction is easy to miss where the disputed English word appears ordinary and is not a term of art. “Withdraw” is common language. In this case, it was connected with a contractual right carrying defined conditions and consequences. A translator who treats the word as an isolated lexical problem may produce a fluent version while missing the legal architecture around it. Counsel then inherits that decision when the document is relied upon in proceedings.

The judgment provides another example in the word “feasibility”. Article 3.5 of the Project Agreement referred to the project’s “technical and financial feasibility”. One translation rendered the term as «βιωσιμότητα». The Court of Appeal considered «σκοπιμότητα» the appropriate rendering in the context of the provision.

The point is useful because “feasibility” does not immediately look like a difficult legal term. Within Article 3.5, however, it formed part of the condition governing whether the investor could exercise its right to withdraw from the project. Its importance came from the work it was doing in the clause.

Commercial contracts contain many expressions of this kind. “Material”, “reasonable”, “completion”, “default”, “acceptance”, “notice”, “cause” and similar words can acquire considerable significance through their contractual function. Their translation requires attention to the surrounding rights, obligations and consequences. The presence of familiar vocabulary should not encourage a lighter review where the clause itself performs important legal work.

For counsel, the practical implication is straightforward. Review effort should follow legal significance. A term embedded in a provision governing termination, withdrawal, liability, payment, default or another important contractual mechanism deserves close attention even if the English itself appears routine.

The risk often lies in the relationship between the word and the clause, not in the word viewed on its own.

AI does not remove the need for file-level verification

The same issue arises when machine translation or generative AI is incorporated into the workflow. Glossaries and terminology instructions can improve consistency across large volumes of material. They are useful tools, particularly where a case contains recurring technical or contractual language. Their effectiveness still needs to be verified against the documents that matter.

Generative systems produce language probabilistically. A preferred term may be supplied in a glossary and still appear differently in another passage. The system may preserve the chosen word while misunderstanding the legal relationship expressed by the clause. Large-scale use therefore creates a practical review question for counsel: have the legally significant occurrences actually been checked against the agreement and against the terminology already used elsewhere in the file?

That question becomes particularly important where translation work is divided across several stages of a dispute. An AI-assisted translation of a new notice may be linguistically polished and internally coherent while departing from terminology established years earlier in the contract. Without a case-level review, the divergence may remain invisible until the documents are placed side by side in proceedings.

Translation risk

Judgment No. 221/2026 does not support the claim that the Spanish investor lost because of a bad translation. After the December notice had been characterised as an exercise of υπαναχώρηση, the courts still had to determine whether the contractual conditions for exercising that right had been met. Article 3.5 allowed withdrawal where the project’s technical or financial feasibility had proved insufficient under the relevant conditions. The Court of Appeal found that the notice did not specify whether the alleged inadequacy was technical, economic or both, and did not identify the supporting data. The attempted exercise of the right therefore failed on substantive grounds. The Supreme Court rejected the cassation challenge to that conclusion.

The translation issue nevertheless mattered because it formed part of the route by which the legal act was characterised. That is a realistic model of translation risk in contentious work. The danger is often cumulative. An inconsistent term, an unexamined synonym, a later document translated without reference to the underlying agreement, or an ordinary commercial word carrying an important contractual threshold can create an additional point requiring submissions and judicial analysis.

Cross-border litigation already contains enough genuine uncertainty. Translation should be managed so that it does not introduce further avoidable ambiguity into the record. Judgment No. 221/2026 shows how easily terminology can become part of the forensic material once a dispute turns on the legal effect of the documents themselves. For Greek lawyers working with English-language contracts, the case offers a practical reason to treat translated terminology as part of the case strategy from the outset, with the same attention to consistency, context and legal consequence that is applied to the original documents.

#GreekLaw #LegalTranslation #ContractualDisputes

[1] https://shorturl.at/f0m4s. The judgment is in Greek.

[2] Scott, J., O’Shea, J. How Legal Documents Translated Outside Institutions Affect Lives, Businesses and the Economy. Int J Semiot Law 34, 1331–1373 (2021). https://doi.org/10.1007/s11196-020-09815-5

[3] Scott, J. & O’Shea, J., ‘Legal Systems Exposed: Translation and Vulnerabilities’ in Anne Wagner and Aleksandra Matulewska (eds), Research Handbook on Jurilinguistics (Edward Elgar Publishing 2023) 300 https://doi.org/10.4337/9781802207248.00029