A European pharmaceutical company was defending its patent for a novel drug delivery mechanism in a US International Trade Commission investigation. The challenger submitted a Japanese-language journal article as prior art, along with a certified English translation. The translation described a compound that appeared, on its face, to anticipate the claimed invention. The patent holder’s survival depended on showing that the prior art did not, in fact, disclose the specific claimed configuration.
The patent holder’s litigation team hired a Japanese-English pharmaceutical linguist to review the submitted translation. The linguist identified a critical error in how a technical term had been rendered. The Japanese term described a specific physical arrangement of polymer layers. The certified translation had used a broader English term that encompassed the arrangement but also included configurations the prior art did not disclose. The prior art did not anticipate the claimed invention. The translation had made it appear that it did.
The patent survived. Not because the science changed. Because the translation was corrected.
In global patent litigation, translation is not a support function. It is evidentiary infrastructure. Prior art references, expert declarations, deposition transcripts, technical specifications, licensing agreements, and the patent claims themselves all enter the legal record through translation. When that translation is inaccurate, ambiguous, or technically imprecise, the legal record itself becomes unreliable. The consequences range from unfavorable claim construction to complete patent invalidation.
Why patent litigation translation is a distinct discipline
Patent language is not ordinary language. Patent claims are written in a specialized register that borrows from technical terminology, legal convention, and deliberate strategic ambiguity. A single claim element — “comprising,” “consisting essentially of,” “consisting of” — determines whether the patent covers a broad class of products or a narrow specific embodiment. These terms have been construed by courts across decades of case law. They are not synonyms. They are legal instruments.
When patent documents cross linguistic boundaries, the translator must preserve not only the technical meaning but the legal function of every claim element. The translator who renders “comprising” as a generic equivalent of “including” in a target language has narrowed the claim scope without the patent holder’s knowledge or consent. In jurisdictions where claim construction is outcome-determinative, this narrowing can decide the case before the merits are even reached.
The “person skilled in the art” standard — the legal fiction that courts use to interpret patent claims and assess obviousness — applies to translation as it applies to every other aspect of patent analysis. A translation of a prior art reference must be assessed not by whether it is grammatically correct but by whether a person skilled in the relevant art would understand the translated text to disclose the same technical teaching as the original. Grammar is insufficient. Domain expertise is required.
The three evidentiary surfaces where translation fails
Prior art translation. In cross-border patent disputes, prior art frequently originates in a language different from the litigation forum. Japanese technical journals, Chinese patent filings, German engineering specifications, Korean academic papers — all of these enter US, European, and other jurisdictions as translated documents. The translation must be accurate at the level the court requires: not merely conveying the general idea but preserving the specific technical disclosure, the scope of what was and was not taught, and any limitations or qualifications in the original text.
Opposing counsel will scrutinize prior art translations for over-breadth. If the translation discloses more than the original, the prior art appears more anticipatory than it actually is. If the translation discloses less, the patent holder appears to be concealing relevant art. Both errors carry consequences. The first can invalidate a valid patent. The second can result in sanctions for discovery misconduct. The translator working on prior art is not performing a courtesy. They are constructing a piece of evidence that will be tested under adversarial conditions.
Claim translation for foreign filing. When a patent is asserted in a foreign jurisdiction, or when a patent application is filed in multiple jurisdictions, the claims must be translated. This is the highest-stakes translation in the entire patent system. The translated claims define the scope of protection in that jurisdiction. A translation that broadens the claims beyond what the original language supports may be rejected by the patent office or invalidated in litigation. A translation that narrows the claims surrenders protection the patent holder is entitled to.
The European Patent Convention requires that the claims of a European patent be available in all three official languages. Discrepancies between language versions have been the subject of EPO Board of Appeal decisions. In one line of decisions, the Board addressed the question of which language version governs when the claims in different official languages appear to have different scope. The answer depends on the specific facts, but the underlying principle is clear: the translation is not a convenience for foreign readers. It is a legal document with independent legal force.
Expert declarations and deposition transcripts. When a technical expert testifies in a language other than the litigation forum’s language, the transcript enters the record as a translated document. The translation must capture not only the technical content but the precision of the expert’s qualifications. An expert who describes their work using a specific technical term in their native language and finds that term rendered as a generic equivalent in the translated transcript has been misrepresented to the court. Opposing counsel will exploit the gap between what the expert said and what the transcript records them as having said.
The translation challenge in IP litigation support across jurisdictions
Patent litigation increasingly spans multiple jurisdictions simultaneously. A single technology dispute may involve parallel proceedings in the US (ITC and district court), Europe (EPO and national courts), China (CNIPA and Beijing Intellectual Property Court), Japan (JPO and IP High Court), and Korea (KIPO and Patent Court). Each proceeding generates documents that must be translated for use in the others.
The terminology must be consistent across all proceedings. A technical term translated one way in the US filing and a different way in the European filing creates an inconsistency that opposing counsel in either jurisdiction can exploit. The term means what it means. If the US translation says the component is a “filter” and the European translation says it is a “separator,” and these terms carry different technical implications, the patent holder has given opposing counsel a gift.
This is not a hypothetical risk. In Inter Partes Review proceedings before the USPTO Patent Trial and Appeal Board, petitioners have successfully challenged patents by identifying discrepancies between the claims as translated for US filing and the claims as they exist in the original-language specification. The Board treats the original-language specification as the authoritative source. When the English translation diverges, the patent holder faces the choice of narrowing their claims to match the original or defending a translation that may not be supportable.
The complexity multiplies when the underlying technology is itself multilingual. A semiconductor patent may reference specifications originally written in Japanese, Korean, and English. A pharmaceutical patent may cite prior art in German, French, and Chinese. The litigation team needs a translation infrastructure that can handle all relevant language pairs while maintaining consistent technical terminology across every document.
What IP litigation support requires from translation
Effective international patent litigation support demands a translation methodology designed for evidentiary use, not general comprehension:
Domain-matched linguists. The translator must have demonstrated expertise in the relevant technical field. A pharmaceutical patent requires a linguist with pharmaceutical training or experience. A semiconductor patent requires a linguist who understands semiconductor fabrication terminology. Generalist translators, regardless of their linguistic skill, will produce technically imprecise translations that opposing counsel can challenge.
Literal and interpretive rendering. For evidentiary documents, the translation should include both a literal rendering and an interpretive rendering where the target language does not permit a direct equivalent. The literal rendering preserves the original text’s structure for the court’s review. The interpretive rendering conveys the technical meaning a person skilled in the art would understand. When these diverge, the divergence itself is significant and must be documented.
Terminology management across proceedings. A centralized glossary of translated technical terms, validated by the litigation team and maintained across all jurisdictions, prevents the inconsistencies that opposing counsel exploits. Every translator working on the case draws from the same validated glossary. Deviations are flagged and resolved before the translated document enters the legal record.
Certification and testimony readiness. Translations used as evidence in patent litigation must meet the certification requirements of the relevant jurisdiction. In US federal court, this may require a translator declaration under 28 U.S.C. § 1746. In EPO proceedings, it may require specific formatting and attestation. The translator must be prepared to testify about their qualifications and the accuracy of the translation if challenged. A translation that cannot withstand a Daubert challenge or equivalent scrutiny is a liability, not an asset.
The cost of getting it wrong
The pharmaceutical case described at the outset illustrates the stakes. A technical term was translated with a broader English equivalent than the Japanese original supported. The certified translation made it appear that a prior art reference anticipated the claimed invention. The patent holder’s litigation team spent significant resources identifying the error, retaining a qualified linguist, and presenting the corrected translation to the tribunal. The patent survived, but the cost of correcting the error was substantial.
Not every case ends this way. In proceedings where the patent holder does not identify the translation error in time, or where the error is in the patent holder’s own translated claims rather than in the opposing party’s prior art, the consequences can be severe. Patents have been invalidated. Claims have been narrowed to the point of commercial worthlessness. Licensing negotiations have collapsed because the translated patent claims did not match what the parties understood the patent to cover.
The investment in qualified patent translation is a fraction of the cost of patent litigation itself. A single IPR proceeding can cost several hundred thousand dollars in attorney fees alone. A district court patent trial can cost millions. The translation of the prior art, the claims, and the key evidentiary documents represents a small fraction of that budget — but it is the fraction upon which the entire evidentiary structure depends.
Artlangs Translation provides international patent litigation support across 230+ language pairs: domain-matched patent linguists for prior art analysis, claim translation for multi-jurisdiction filing, expert declaration and deposition translation, terminology management across parallel proceedings, and translation certification and testimony preparation. We serve IP litigation teams in New York, Washington DC, Munich, London, Tokyo, Seoul, Beijing, and beyond. Because in patent litigation, the translation is not background paperwork. It is the evidence. And the evidence must hold.
