When a U.S. Examiner Says Your Claims Read Like a Translation: What Went Wrong?

A U.S. examiner does not reject a claim merely because it came from a Chinese application. But the USPTO has a form paragraph for claims that appear to be a literal translation from a foreign document and are so narrative, grammatical, or idiomatically defective that the claim becomes indefinite. A CrossVision IP Chinese article examined a mechanical application in which all nine claims were rejected and the examiner expressly complained that the claims read like a translation. The useful lesson is not “write prettier English.” It is to fix the claim logic before translation artifacts become §112 problems.

What the Examiner Was Actually Complaining About

The Chinese source article described a mechanical application with nine claims. All nine were rejected. The striking part of the Office Action was not that a translated application had been recognized as foreign; that happens every day. The examiner's concern was that the claims contained extensive grammatical and usage problems, read as though they had been translated directly from another language, and were difficult to understand well enough to determine their scope.

That distinction matters. U.S. patent examination is not an English-composition contest. The MPEP itself reminds examiners that examination is not conducted for the purpose of securing grammatical perfection. Minor errors can be objected to and corrected. A §112(b) rejection requires something more important: uncertainty about what the claim actually covers.

The USPTO's current form paragraphs reflect the same line. Form paragraph 7.34.07 addresses claims that are generally narrative and indefinite, appear to be literal translations from a foreign document, and contain grammatical or idiomatic errors. It is used with the substantive §112(b) indefiniteness rejection—not as a stand-alone rule against foreign drafting.

Problem One: The Claim Introduces Elements in the Wrong Order

One issue highlighted in the source article was antecedent basis. Chinese-to-English translation can preserve a sentence order that sounds understandable in Chinese but creates a U.S. claim such as “the locking member engages the groove” before any locking member has been introduced. The drafter knows what “the locking member” means because the figure is open on the screen. The claim, however, is expected to define its own relationships with reasonable clarity.

In ordinary U.S. claim drafting, an element is commonly introduced with “a” or “an,” and later references use “the” or “said.” This is not a mechanical grammar statute, and lack of explicit antecedent basis is not automatically fatal if the reference is reasonably clear. But when the reader cannot tell which earlier structure is being referenced—or there is no earlier structure at all—the defect can make the metes and bounds indeterminate.

Problem Two: One Chinese Structure Becomes Several English Structures

The source article also identified inconsistent terminology. The same Chinese component appeared to have been rendered with different English names in different parts of the claims. To the translator, the words may have looked like natural synonyms. To the examiner, three different nouns can indicate three different structures.

This is a recurring patent-translation trap. “Part,” “member,” “portion,” “unit,” “module,” “assembly,” “body,” and “component” are not interchangeable simply because they are all broad nouns. If the source intends one component, the English patent record should normally give it one principal name. If the source intends three structures, the translation should preserve that distinction deliberately.

Patent consistency can be more important than stylistic variety

Ordinary English rewards avoiding repetition. Patent claims often reward the opposite: repeat the same term when you mean the same thing, and change the term only when you mean a different thing.

Problem Three: Literal Syntax Hides the Technical Relationship

A word-for-word translation can also preserve all the nouns while losing the legal-technical relationship among them. Chinese patent sentences often carry several positional, functional, and causal relationships in one long sequence. If those relationships are moved into English without restructuring, the reader may not know which modifier belongs to which component, which action occurs first, or which element performs a recited function.

This is not merely stylistic. The claim may be cited against prior art limitation by limitation. An unclear modifier can determine whether the examiner thinks a reference discloses the limitation. The same ambiguity can later affect infringement analysis. U.S. adaptation therefore needs to preserve the source meaning while reorganizing syntax so each structural and functional relationship is explicit.

Problem Four: The Claim, Description, and Drawings Stop Agreeing

The Chinese source article recommended checking claim terminology against the figures before filing. That is particularly important when a Chinese source uses several labels for the same structure or when the translation team works on text separately from the drawing labels. A claim may say “support member,” the detailed description may say “supporting element,” and the figure may label the same part as “bracket.”

No single one of those words is necessarily wrong. The combined record is the problem. The examiner has to decide whether the documents describe one part or three. A pre-filing terminology map can eliminate that uncertainty at almost no cost compared with fixing a §112 rejection later.

This Is Why “Native English” Alone Is Not the Solution

A fluent editor can make sentences smoother and still damage a patent. If an editor replaces repeated terms with synonyms, compresses a relationship, or “improves” a technical phrase without understanding claim scope, the result may read better while becoming less precise.

Conversely, a patent professional can preserve every Chinese word and still produce a poor U.S. claim if the structure is never adapted. The required skill is legal-technical bilingualism: understand what the Chinese disclosure says, understand what the U.S. claim needs to communicate, and change the form without silently changing the invention.

A Pre-U.S.-Filing Claim Review

Check every claim element at first introduction and every later “the” or “said” reference.
Create a terminology map linking Chinese terms, English claim terms, description terms, and figure labels.
Use one English principal term for one structure unless a real distinction is intended.
Break long translated syntax into clear element-to-element and step-to-step relationships without adding new technical content.
Confirm singular/plural consistency and which actor performs each function.
Read the claim without the drawings open. If the scope is unclear without guessing from the figure, revise the words.
Then read the claim with the drawings and description to confirm the entire application tells the same technical story.

Takeaway

The examiner's phrase “looks like a literal translation” is not really an insult to translation. It is a warning about the patent record. The U.S. problem begins when translation artifacts make the claim boundaries unclear, not when the examiner can tell that the application originated in another language.

For a Chinese application entering the United States, the safest workflow is therefore not “translate first, fix the Office Action later.” It is to conduct a claim-level adaptation review before filing: antecedent basis, terminology consistency, technical relationships, and figure correspondence. Those are small edits when the file is still on the drafting desk and expensive edits once they have become a §112 record.