By Greg Corcoran, Brenda Holmes, Takahiro Koshibu and Toby Mak.
Systematic limitations on the efforts for global patent harmonization exist. As patent professionals, we encounter the risks of such linguistically derived differences on a regular basis. Awareness of such differences between the linguistic characteristics of different languages can help mitigate or even minimize their impact.
This article reviews the differing linguistic approaches to some concepts as basic as expressions of numerical limitations, including the singular and plural. The reasons for the occurrence of these linguistic differences are reviewed to help practitioners understand why they occur and, more importantly, when they may occur and how they can be avoided.[1] The article then notes that the degree of consistency of language in a specification can fundamentally impact claim scope if addressed inappropriately on translation. The insights provided by this article suggest how drafting in a source language (e.g. English) can be adjusted to improve the clarity of resulting claim and specification in the target language (e.g. Chinese (i.e. Mandarin) or Japanese as a language of prosecution[2]) to more closely correspond with the initial, intended scope on drafting (i.e. source text); and vice-versa.
We, the authors, found that effective communication between the parties from the client and to local prosecuting attorney is key to reduce the chance of unintentionally introducing linguistic limitations during translation. The best practical translation is purposeful and holds true to the original concept expressed in the source text, and thus the commercial application of the claimed invention as originally defined and intended by the patent applicant.
‘Alice felt dreadfully puzzled. The Hatter’s remarks seemed to her to have no sort of meaning in it, and it was certainly English. “I don’t quite understand you”, she said as politely as she could.’[3]
As patent attorneys we work in an international commercial context. Applicants often want certainty that their inventions should benefit from the same patent protection worldwide. As patent law is generally harmonized, applicants may well feel justified in their expectation of achieving a uniform worldwide patent claim scope through the equivalent (or corresponding) patent filings and consequential granted patents. Yet in the reality of actual practice, it is no surprise to patent attorneys that this is an idealized objective – uniform claim scope of equivalent granted patents is rare.
The corresponding independent claims drafted and prosecuted in corresponding patent applications in a patent family filed in any country in the world are intended to cover just one, common inventive concept. However, there can be as many different interpretations of a claim in a patent application as there are countries in which the corresponding applications are filed, indeed due to the different practices of the respective patent offices of those countries. The reasons that the resulting granted claims of the corresponding patents would have different scope from each other are various. The different patent offices may have differences in approach to interpretation of the chosen claim language, the requirements of patentability such in assessing inventive step and the scope of allowable amendments.
There are thus different reasons for the inconsistent claim scope between the equivalent patents. One underlying cause of inconsistency is how the original claim wording of the source text can be translated into an applicable official language of the jurisdiction in which an equivalent patent application is filed: the target text. Like Alice in Alice in Wonderland, sometimes not everything is what it seems to be. There can be fundamental differences between languages in how to express concepts. Sometimes such a concept is basic, such as the use of indefinite articles and the use of numerals. Therefore, even the simplest and clearest choice of words for an elementary expression in the language of the source text could result in unclear language or text that even means something very different from what was intended when drafted. There is a consequence when the source text of a draft is literally translated into the language of the equivalent application of a different jurisdiction, i.e. the target text, the meaning and thus scope of the translated target text ‘unexpectedly’ differs from the source text.
The simplest way of avoiding such problems is through effective communication with external foreign counsel about the causes of mistranslation that create a difference in scope. The occurrence and frequency of such mistranslations and consequential impact of claim scope may be reduced and even eliminated. Some concepts, which can be fundamental, do not easily mutually translate between English, and Chinese or Japanese.[4] Therefore, in translating between languages it is helpful to try and understand how a certain choice of words in the source language would be understood after translation in the context of the target-language. Often translation is ‘literal’ in that individual words are selected for their literal meaning, rather than in context of the surrounding text, and the intended purpose of the text, , i.e. context, such as claim language. Good translation for patent applications is thus more than understanding the best literal translation of a word or expression – it is purposeful.
In this article, the authors hope to share some of the insights they have had with such unintended mis-translations. They hope that this article helps readers avoid similar experiences of mis-translation in their own practice. Indeed, the authors each learnt something when preparing the original seminar on which this article is based. No amount of experience is sufficient. There is always something new to learn. The exemplary topics considered in this article include how to use indefinite and definite articles, singulars and plurals; and the use of common claim terminology. General considerations are reviewed for language use in patent specifications that reduce the impact of differing linguistics between source and target texts.
Note: the material of this article was presented in a seminar in 2012. It may well be expected that the standard and quality of translation between different languages has improved over the years, including with assistance of machine translation, especially artificial intelligence (‘AI’) enabled translations. The need for the present article would then seem to a reasonable and legitimate question. The authors agree, however, that, based on their experiences, the topics mentioned in this article remain relevant today as in 2012 for the drafting and prosecuting patent attorney despite the advancements in computer aids such as AI translation. For example, AI translation tools still have issues with basic concepts such as the use of numbers and use of indefinite articles when translating between English and the languages selected for this article. The continued relevance of the general theme addressed in this article is one reason why a similar session related to drafting of a specification for multiple jurisdictions was included in the programme for CIPA Congress 2024.
The Chinese and Japanese languages conceptually treat numbers in different ways from English, including:
A mistranslation though any of these points, however trivial, could result in claim scope very different from that intended, which can be to detrimental to the applicant expectations.
The Japanese and Chinese languages both distinguish between singular and plural words differently from English. This creates a risk that, for example, ‘a component’ and ‘components’ are translated into the same word ‘component’ without any distinction between singularity and plurality. (Notice ‘component’ translated into Japanese and Chinese neither has an indefinite article – ‘a’ nor an indication of a plural, ‘–s’.) If the context surrounding the word ‘components’ clearly and without doubt indicates that the component needs to be plural in defining the invention, then ‘components’ would be translated into ‘a plurality of components’ into the target language. However, making such a judgement could often be difficult at the time of preparing a translation because the techniques and practice of drafting claims in China and Japan can be substantially different from those in the UK and the US.
To avoid such ambiguity and make the translation properly reflect the claim drafter’s true intention that the plural is indeed intended, it is suggested that the expression ‘a plurality of components’ should be consistently used throughout the claims as well as in the corresponding locations in the description.
When just one component is intended, use of an expression such as ‘one component’ should be considered. However, when ‘one component’ is used, care should be taken in view of a possible limitation to just and only one component. The word ‘one’ in this context in Japanese and Chinese is narrowly construed in Japan and China, respectively. If the source text is intended to cover a single component, consider alternatives to e.g. ‘one component’ that do not risk a narrow translation and consequential interpretation in the target language. To ensure such an alternative wording in the source language results in a better choice of language in the target language, draw the potential issue to prosecuting counsel’s attention when sending instructions for filing the equivalent application. Prosecuting counsel could verify that the wording can be translated as intended in the source text to refer to more than one of the components, e.g. ‘at least one’, or that the available translation in the target language of the identified expression is inevitably at risk of being limited to read ‘just one’. It is of course better to mitigate such a risk at drafting.
Another issue as noted in the previous section is the translation of the indefinite article ‘a’. If the number of components is not a concern for defining the scope of the invention and the target language is Japanese, the ‘a’ in the source language is not supposed to be, and will generally not be, translated into Japanese. Thus, for example, ‘a component’ in English is usually translated in Japanese as ‘component’.
The situation is different in Chinese. The indefinite article ‘a’ could be literally translated as ‘one’ (“一”). So the text ‘including a component’, for example, could be translated to “包括一组件”, which means ‘including one component’. Such a translation into Chinese risks imposing undesirably limitations to the claim scope because this claim language is considered under Chinese practice to be ‘close-ended’. Even though the expression in Chinese ‘including one component’ uses the word ‘including’, which is usually considered to be ‘open-ended’, it can be considered to have the same ‘close-ended’ effect on the scope of the claim as the expression ‘consisting of’. The scope of such a claim can thus be effectively limited to just a single component. Therefore, it is suggested that the indefinite article ‘a’ in an expression such as ‘including a resistor’ should generally not be translated, resulting in the translation into Chinese as ‘including resistor’.
To ensure a broad claim scope that both includes a single feature and an indefinite multiple number of the feature in a Chinese specification, the expression ‘at least one component’ should be used in the text of the source application in English.
This phrasing may be objected to during examination for a lack of support especially if the description does not clearly specify the upper limit of the number of components or a limit cannot generally be understood. Chinese examiners generally object to an expression in a claim that does not have an express upper limit for lacking support by the description, unless the description clearly indicates that there is no upper limit.
If indeed an upper limit is not known, then it would be useful to include in the description a statement for the benefit of prosecution in China that the number of components is to be determined by, and would be understood by, a person skilled in the art according to the operational environment. As always, it depends on the particular expression used in the language of the application as filed.
In the US and Europe if a claim recites a specific number of components, then the claim is generally interpreted to cover a device with the recited number of components, as well as a device with more than the recited number of components, for example a claim to a table with three legs includes a table with four or more legs,[5] but not a table with two legs or one leg. Therefore, English claims tend to recite the fewest number of components needed to define the invention. This is a clear distinction from Chinese practice.
In Japan, if the number is intended to be a lower limit, then the use of ‘at least’ should be considered. The specification or the drawings should also be drafted to support the intended range by explaining embodiments or examples using more than the claimed number in connection with the intended effects. For example, if the claim recites ‘a chair with at least three legs’ then the specification should explain, preferably with reference to drawings, that the chair with four or more legs will have the same or similar effects as those achieved by the chair with three legs. However, if the granted claim ends up with ‘a chair with three legs’ by dropping the term ‘at least’ in view of for example a prior art document disclosing a chair with four legs, then the number just covers three legs because of the file wrapper estoppel which is applicable in Japan. The situation in China is the same.
From a patentability perspective, in most jurisdictions, the number of claimed components is not usually helpful in distinguishing over prior art cited for inventive step or obviousness unless changing the number of components provides a technical advantage. If, for example in the US, the prior art shows a device with one component and the claim requires two or more of the components, most US examiners would find that increasing the number of components is an obvious design choice or variation, unless this difference achieves a technical benefit. The practice in Europe, China and Japan is similar.
In English it is common to recite a ‘plurality of components’ (i.e. a group) and then to further define one or a subset of the components (i.e. one or more subgroups of the group). Although the ‘plurality of components’ may be considered to be a group of components, often the claim language and even the wording in the description avoids the specific word ‘group’ and thus ‘sub-group’. In a more specific example, a claim in English may recite ‘a first one of the components’ (where later in the claim there is reference to a second one of the components) or ‘a selected one of the components’ in addition to a ‘plurality of components’ in order to refer to a feature of one or some of the components of the ‘plurality of components’, and not necessarily a feature of each and every one of all the components of the ‘plurality of components’.
However, there is no precise corresponding equivalence in Chinese because Chinese does not distinguish between singular and plural words in the same way as English. These sorts of expressions therefore become unclear and convoluted, for example using the solutions mentioned earlier in this article. Even if the source text is expressed carefully, the text translated into Chinese may inevitably be unclear.
One way in Chinese of referring to additional features of one of the components, but not the others, is for a claim to recite a group including a first component, a second component and a third component, for example. Yet, although such claim language attempts to clarify the claim scope, it may present additional translation problems in China.
Consider the example ‘a group including a first component…’. It is unclear in Chinese whether each of the words ‘group’ and ‘component’ are in the singular or plural. Unless specified explicitly, the claim could be mistranslated into Chinese to refer to only one group of the components. Alternatively, the language of the claim can be translated to refer to multiple groups (i.e. more than one group of the components).
Consider the example claim language written in English of ‘a table with legs, each leg having three sectional lengths’ (e.g. an upper, a middle and lower length) and how this text could be translated into Chinese.
To avoid ambiguity and as broad claim scope as possible, the English expression should be drafted to ensure that the conceptual elements can be accurately translated. In context of the example, such a claim could be written in a manner to highlight the different groups of feature present in the claim as:
‘a table with one or more legs (groups) including one or more first sectional lengths (first components), one or more second sectional lengths (second components) and one or more third sectional lengths (third components)’.
This language can be simplified for an actual claim for example, in which it is not required to cover a table with a single leg (see section “Dealing with numbers” above):
‘a table with a plurality of legs, each leg having an upper sectional length, a middle sectional length, and a lower sectional length’
where sensitive translation of ‘a’ would be applied so it is not translated as ‘one’ in the target language.
As shown in this specific example physical characteristics of the components within a group, which here are the positional relationships of the components of the leg, i.e. upper, middle and lower, would help to reduce the risk of unintended mistranslation.
If such a claim were used as an independent claim in Japan, the US or Europe, it may be considered to be unnecessarily narrow relative to the initial proposal in this example of a claim featuring a group of components and selecting one or more of the components from the group i.e. as a subgroup. Suitable drafting and prosecuting strategies could be adopted to address this sort of issue so that a broader claim is prosecuted in Japan, the US and Europe and a claim with wording addressing the translation issues is available in China. For example, the details of the narrow claim for Chinese prosecution, such as of physical characteristics, could be included in the description and dependent claims, or alternatively an extra independent claim may be provided for prosecution in China. Further the description and dependent claims may also include enumerated features of the components, such as sectional lengths, despite the described clarity risks in China, to provide fall back positions in other jurisdictions. Such wording may provide broader scope than the use of physical features suggested for use in China.
In Japan, there should not be a translation issue when referring in the claims to a ‘a plurality of components’ and then defining one or a subset of the components, for example using language such as ‘a first one of the components’ (so as long as there is a second one of the components in the claim) or ‘a selected one of the components’. However, if the wording ‘a group including a first component, a second component and a third component’ is used in the claims instead of or in addition to ‘a plurality of components’ then it is likely that ‘a group’ is translated differently from ‘a plurality’ in Japanese. The translated term ‘group’ then may have a different scope from ‘plurality’. The scope of ‘group’ can as a result be narrower than ‘a plurality’. In any event, if the claim language raises a question of its meaning or scope, the specification and drawings, as well as statements on the file wrapper, will be considered in the claim construction.
There are certain English words that are generally understood to have a number or amount associated with them and they should be carefully used. English often builds words from syllables having specific meaning, such as ‘bi’ meaning two. Japanese and Chinese have pictorial elements, or pictograms, which are used in combination to correspond to an English word, but are not in the form of constituent syllables as in English. For example, a bicycle is usually understood to have two wheels in English because of the syllable ‘bi’; yet the Japanese and Chinese words for bicycle do not have this limitation. None of the characters of either the Japanese word for bicycle, “自転車” or the Chinese word for bicycle “自行车” are a representation of two wheels, (unlike the syllable ‘bi’ in English which means ‘two’). Consider an invention for a bicycle which could have commercial relevance for a vehicle propelled by pedals, with three wheels or more, such as a chain set or handlebars for the bicycle. If a claim for the invention is intended to have scope which could cover the embodiment with three or more wheels, it is better to avoid the English word ‘bicycle’ in the original version (or source version) of the claim. Instead, the expression which uses explicit scope when translated between the languages should be used, for example:
‘a vehicle configured to be propelled by pedals having at least two wheels’.
US practitioners use open-ended claim language, such as ‘comprising’, instead of closed-ended language, such as ‘consisting of’. The preferred choice of wording ensures that the claim will cover a device with more than the recited number of features rather than being limited to the number of recited features. Yet arguments made during prosecution (i.e. the doctrine of file wrapper estoppel) may result in limitations to the scope of a claim. Usually, a claim can be expected to be amended in view of the presented arguments. However, on occasion nuanced arguments have been made in a response for which the examiner did not require a clarifying amendment to the claims. Such arguments in the file wrapper may modify to the scope of such a claim.
Yet there are limitations to the benefit of relying on such argumentation. If a specification discloses only a few components, the specification has disclosure that may provide support for no more than a few components. If an applicant argues during prosecution (e.g. in view of prior art) for a claim interpretation that covers hundreds of components, the patent application may be at risk of receipt of a rejection for the ground of lack of written description or for lack of enablement. If the application grants without such a rejection, the patentee’s arguments made during prosecution provide a risk of invalidation of the patent for the same grounds. The position is similar in other jurisdictions such as China and Japan.
Note that different arguments on file can be used in different jurisdictions for example because different prior art is cited in the different jurisdictions. Thus, the respective prosecution record of similarly worded claims of patents in, say, Japan and the US can have different scope.
Another example of relatively open claim language at least in the US and Europe, is the use of the word ‘for’ as in ‘suitable for’ in an apparatus claim which may be used to express an intended functional aspect, purpose or use of the field of the invention or of a feature of the claim. In European case law, such language may imply certain limitations for novelty without which the apparatus could not be used for the intended function or purpose and is treated as being limiting only to the extent that the article had to be suitable for that use.[6] Such limitations are likely to be obvious and thus not bestow patentability in themselves. However, in Japan such language, when literally translated from English, may be considered to have a limiting effect such that claimed invention is limited to the claimed functional aspect, purpose or use. There are many scenarios which may be dependent on the technical field.
These examples show that even the most elementary numerical aspects of claim drafting in English can lead to unclear claims in the non-English target language or claims with unintended scope. It is only to the applicants benefit that carefully considered measures are taken at the time of translating, if not at the time of drafting, to ensure the claims in the target language have the scope and adequate support as intended by the drafting attorney. Relying on machine-aided translation, such as AI translation, is likely to be an incomplete solution. Often the effective measures are only improved through open and frequent communication between the applicant, or their local patent attorney, and prosecuting patent attorneys in other jurisdictions.
Avoiding or even minimizing the effect of translation can be achieved by informed selection of terminology when preparing a draft and in anticipation of amendment.
Some jurisdictions, such as Japan and China, prefer consistent terminology throughout a specification and its claims. In other jurisdictions such as the US, it may be preferable to use synonyms to broaden the legal scope of claim language. These approaches can appear consistent: which approach is preferable?
There may be differences in the degree of similarity between the wording used in the claims and the wording used in the description in the different jurisdictions. The practice in the US is generally more lenient than in Europe, China or Japan. In the US, the differences between the language in the claims and in the specification may result from the US style of drafting or from amendments made during prosecution.
It is relatively common for a US application to include specific examples in the description using specific language, but use more generic language in the claims. Under established US practice claim terms may be defined in the description without a formal glossary. In some instances a claim term is defined by listing specific examples that are intended to be covered by the term, and to indicate that the examples are non-limiting. This type of open-ended definition may not be consistent with the practice in other jurisdictions.
In Europe, consistency of terminology of the claim and through the description is recommended. Where inconsistent terms are used (either in different parts of the description or between the claims and the description), it is advisable to include in the description an appropriate definition of the terms, or a statement explaining how the terms relate to each other. Without such measures there is a risk of adding matter during prosecution. Such a finding of added matter may be in consequence of amending a claim to introduce text from a passage of the description that has different wording but similar meaning. The addition of the new text to the claim may cause the claim to have scope that was not present in, and thus is added matter relative to, the text of the application as filed.
In China, there is an additional reason for using consistent terms, as in Europe. Because of pictorial nature of the Chinese language, it is strongly recommended that terms should be used consistently in the specification and claims. There is a risk that different pictograms could be used for different terms, which could have different meanings from the original terms in the source language. Where more than one term is used in the specification, the relationship between the terms should be clearly set out to avoid ambiguity, which is a ground of invalidation in China.
In Japan, the claims can use generic language or the language expressing the broader concept provided such language is well supported with specific examples in the specification and drawings. However, if the claims are intended to cover the same level of narrower concepts as the specific examples in the specification, then the language used in the claims and the language used in the specification should match, otherwise the claims might face a clarity or support rejection.
In the US, a claim may be amended to include a term or phrase that did not appear in the original claims. Although the term or phrase should be supported by the description, it does not need to appear verbatim in the description. Support can even be present in the Figures. In certain circumstances, an added term may describe an aspect or feature of the invention shown in a figure, but the added term may not appear at all in the text of the description.
This is another area where US practice differs from other jurisdictions, because such an amendment is very likely to constitute added matter in China and Europe. In these jurisdictions an amendment should be directly and unambiguously derivable from the specification as filed.
Yet in Japan, the new matter requirement (added matter) for an amendment has lessened over time. A number of years ago, the requirement was very strict in the sense that the amended matter must be directly and unambiguously derivable from the specification. In practice it was quite difficult to rely on the drawings for amendments. The strictness of the requirement, however, has been eased and there is now more opportunity for finding support for an amendment in the drawings.
Compared to the US, the requirements of Japan, China and Europe require a different approach to the choice of terminology which may permit less flexibility than US drafting counsel may be accustomed. In view of linguistic limitations on translation, it can only be beneficial to ensure that terminology is consistently used in an application. By taking appropriate measures, prosecution in foreign jurisdictions may be facilitated and claims better suited to achieve the leverage sought by the applicant can be achieved at grant. Such patents may have claims which are more robust when challenged.
Achieving the objectives of international harmonization of patents is implicitly limited by the use of different languages essential for developing an international patent portfolio, and how different jurisdictions can differently interpret the same language (e.g. even English claims in the US and the UK) – sometimes with results that are contrary to initial expectations. When similar claims in different countries are interpreted and applied under their respective local laws by their own courts, there will be an impact of the different languages and their linguistics: in the priority filing, and for the prosecuted application. A difference in the claim scope is generally inevitable – a final hurdle to harmonization.
This article looked at how some of the key issues can be avoided, and is intended to be a pointer to better outcomes in prosecution. It has discussed some approaches that can be adopted, with examples, to help reduce the impact of different linguistics in the source language of drafting and in the target language of prosecution and grant.
The key solution appears to be continuous improvement in communicating between the applicant and counsel to learn about the common linguistic issues. Such communication for better mutual understanding can result in improved outcomes – in reducing if not removing the hurdle from achieving similar scope for equivalent patents. For example: the language and choice of words and phraseology of the original claims and specification of the draft (i.e. source language) can be prepared with a view of future filings and translations in other jurisdictions the languages of those jurisdictions (i.e. target language). Such an applicant would ensure that the translation of the source text into the respective target texts is purposeful, with an applicant’s commercial objectives in mind. Such a careful approach may be used to improve the quality of translation prepared by an applicant traditionally or using machine aids such as AI translation. The applicant can thus expect to obtain granted patents that are more likely to be fit for purpose as the commercial tools that the applicant requires, such as enabling the business objectives the applicant sought when filing the applications for the patents.
[1] This article is based on a session of CIPA Congress 2012, and builds on the authors’ experiences of working together in preparing the session
[2] The Chinese and Japanese languages were chosen for this article because they are very different to English but are commonly encountered by most attorneys in their daily practice.
[3] Alice’s in Adventures in Wonderland, Lewis Caroll
[4] This article considers translation between English and Japanese and between English and Chinese. The principles may apply to translations between English and other Asian languages. The general concept of inaccurate translation may apply between different language combinations, but they are not considered in this article, as such.
[5] This construction is taken with a view to the specification according to the local laws of construction.
[6] Case Law of the Boards of Appeal Tenth Edition, July 2022 (selected sections updated June 2024) I.C.8.1.5 Novelty criteria for product claims with purpose characteristics, https://www.epo.org/en/legal/case-law/2022/clr_i_c_8_1_5.html
