Patent claims, and therewith infringement, can get lost in translation. The Dutch first instance court limited a patent’s scope of protection based on the Dutch translation of the claims. The Court of Appeal saw it differently. A thorough review of the translation of the claims remains necessary to avoid unwanted discussions on claim interpretation.
To validate a European patent , the Dutch Patent Act (‘DPA’) requires a Dutch translation of the claims to be filed with the Dutch Patent Office within three months after publication of the grant. The quality of the translation and its conformity with the original (English, French or German) text of the European patent is not examined. The patentee has the option to file an amended/improved translation of the claims. As boring as the above information may sound, it becomes interesting when the Dutch translation does not correspond to the text of the claims as granted in English, French or German. For this situation Article 52(9) DPA prescribes that if the scope of protection of the Dutch translation of a European patent (application) is narrower than the protection conferred by that patent (application) in its original language, the translation shall be deemed to be the authentic text. This applies only to the assessment of infringement.
You could say that a Dutch translation that limits the scope of protection for the patentee, compared to the original claims, is at the patentee’s risk. It is beneficial to the legal certainty of third parties doing business in the Netherlands, who can then rely on the text filed with the Dutch Patent Office. However, in this case the patentee argued that a non-Dutch speaking party should not be entitled to rely on a Dutch translation of the claims. Non-Dutch speaking defendants could and simply take note of the authentic text of the claims which imply a broader scope of protection. Hague District Court ruled that a foreign party operating on the Dutch market can also rely on the more limited Dutch translation of the claims (par. 4.11, see the judgment here in Dutch):
“A different decision could give Dutch parties an (unjustified) competitive advantage over foreign companies operating on the Dutch market, which is all the less justifiable if it concerns companies with an establishment in another EU country…”
The District Court ruled on the scope of protection of a patent claiming a “simulated fire effect apparatus”, which, according to the original English claims, included, inter alia, “an apertured bed”. The Dutch translation of the claims translated “an apertured bed” more or less as a bed with openings (“een van openingen voorzien bed”). Although it is up for discussion whether this English translation is correct or not (as I re-translated a Dutch translation of an English claim into English), most relevant is that the Dutch translation gave the bed multiple openings (by using the plural form of the word opening). The question the court had to answer was whether the patent also covered the electric fireplaces of the alleged infringer, which had only one opening. The District Court relied on the Dutch translation of the claims. It considered that legal certainty for third parties dictates a narrow interpretation of the scope of protection of the patent, protecting only products with multiple openings. It took into account the absence of clear indications that the bed may also be provided with only one opening, whereas there were many indications that there must be multiple openings.
One month after the first instance decision, the patentee filed an amended translation of the claims, translating “an apertured bed” (again more or less) as a bed with one or more openings (“een van een of meer openingen voorzien bed”) and filed an appeal (de novo assessment). This change in translation raises interesting questions, such as whether a patentee by filing a modified translation of its claims, while an infringer has relied on the previous narrower claims, and which set of translated claims should be considered by the court after a change in translation. Unfortunately, the Dutch Court of appeal did not answer these questions and decided that it did not need to decide which set of claims should be interpreted.
The Court of Appeal considered that even the earlier Dutch translation (a bed with openings) protects a fireplace with one opening (see the judgment here in Dutch). The Court of Appeal considered that the literal text of the claim of the earlier Dutch translation indicates a bed with multiple openings. However, the Court of Appeal also took account of the inventive concept behind the invention from the standpoint of the skilled person with a mind willing to understand and the description of the relevant patent (EP 2 029 941 B1) (e.g. par. [0019], which mentions “one or more larger bodies, each of which has one or more apertures”).
While one can write another blog post on the Court of Appeal’s interpretation, there is a lesson learned: review your translations or see them go up in flames!
________________________
To make sure you do not miss out on regular updates from the Kluwer Patent Blog, please subscribe here.
Dear Jasmijn
Thank you for your report.
The importance of good translations, be it for claims or the full description is all too often forgotten? swept under the table? neglected? At the same time the fairy tale of expensive translations has been repeated so often that all too many operators have accepted it as fact. However, translation costs are minute when compared to:
1. Search fees (the quality of EPO searches being questioned regularly in this blog).
2. Maintenance fees (annuities), a competent claims translation in the Netherlands may cost € 100-150,- which would cover a fifth but not even a sixth annuity.
3. As well as filing fees at the EPO as opposed to validation fees in the Netherlands of € 25,-.
While the importance of translation has come under fire originally for a lack of transparency regarding costs, this lack of transparency is being identified by a growing number of general translation companies as a business opportunity. Artificial Intelligence provides machine translations with quality levels ranging from perfect to nightmare. We are working on combining artificial and real intelligence, where real intelligence is based human patent translation experience, but this is a litlle more complicated than some of the general translation companies may try to lead you to believe. Especially when you consider that in translation only almost perfect is acceptable in most instances.
Daniel X. Thomas commented on a recent decision of the Tech Boards of Appeal (EPO) on translation errors. The case related to an EP application filed in Chinese with a subsequent, inaccurate, translation into English. The latter extended subject-matter (Art.123(2) EPC). When noted after grant, this may lead to an inescapable trap between 123(2) and 123(3) – see also Guidelines A-VII, 7 (https://www.epo.org/en/legal/guidelines-epc/2024/a_vii_7.html) and the patent may be irrepairable and lost due to a translation error on the far beginning.
For Daniel’s blog post: see https://blog.ipappify.de/t-354-22-on-the-importance-of-a-correct-translation-from-chinese-to-english/
Also see DII of 2019 discussed in an earlier blog post bij DeltaPatents: https://eqe-d.blogspot.com/2019/02/dii-2019-shoe-soles.html
When you see the problems linked with translations at national level and at the EPO, I dare imagine what problems could arise with respect to translations if the procedure before a local division of the UPC is not conducted in the language in which the patent was granted!