When a Lawyer Realises Their English Needs a Second Pair of Eyes

The case for native English proofreading

in legal documents written by non-native lawyers

There is a moment that comes for most European lawyers who work in English. The document is finished. The argument holds together. The clauses do what they need to do. Everything has been checked. Then the lawyer reads it back one more time, slowly, and a small doubt arrives. Does this read the way an English lawyer would write it? Not whether it is correct. Whether it sounds right.

That doubt is well-founded. Spotting the difference between correct English and natural English is the hardest part of writing in a second language at a professional level, and it is the part that automated tools and bilingual colleagues are least able to help with. By the time a lawyer notices the gap, they have usually been staring at the document for too long to fix it themselves.

Why the problem is invisible to the person who wrote it

A Polish, German or French lawyer drafting in English has spent years reaching a standard of fluency most native speakers never achieve in any second language. Their grammar is sound. Their vocabulary is technical and precise. Their reading comprehension is often better than that of younger native lawyers. However, none of this prevents the small errors that mark a document as non-native.

The errors that matter are not the kind a textbook covers. They are calques from the writer's first language, idiomatic phrasings that translate cleanly but read awkwardly, prepositions that follow the wrong noun, articles missing where a native would put one and inserted where a native would not. They are the comma before "that" which is correct in Polish and wrong in English. They are false friends like "actual" used where "current" is meant, or "eventually" where "possibly" is meant. None of these errors look wrong to the person writing them, because nothing about the writer's training in English has flagged them as wrong.

The further difficulty is that most automated tools do not catch them. Grammarly and similar services have improved dramatically over the past five years, but they still flag mainly textbook errors. A sentence that is grammatically correct but reads strangely will pass through unchanged. A document full of such sentences will pass through unchanged. The writer reaches the end of their checking process and sees no problems flagged, but this is not the same as no problems existing.

What goes wrong when the English is not quite right

The consequences of imperfect English in a legal document run on a spectrum. At the mild end, the document simply reads slightly oddly to a native speaker. The reader might not be able to articulate what is wrong, but they will sense that something is off. In a cross-border transaction where the other side is choosing whether to trust your client and your firm, that subliminal impression counts for something.

A counterparty's lawyer reading a document drafted in awkward English makes a quiet assessment about the firm that produced it. Sometimes that assessment is unfair. Sometimes it is not. Either way, the lawyer on the other side now has a small piece of leverage they did not have before, and a small reason to push back on terms they might otherwise have accepted.

Further along the spectrum, ambiguity starts to appear. A modal verb used inconsistently, a defined term used in a slightly different form on its third appearance, a sentence whose subject becomes unclear as the sentence grows to be a bit too long. These are the openings that the other side's legal team will look for if a dispute arises later. The document does not need to be wrong on its face. It needs only to admit two readings, one of which favours the counterparty.

There is also the client to think about, and this is the part lawyers most often underestimate. Much of what a lawyer writes goes not to opposing counsel or to a tribunal but to the client: opinions, advice memos, status updates, explanations of what a contract actually says. If the client is a native English speaker, awkward English creates the same quiet impression of unprofessionalism that it does in any other professional service relationship. The client may not raise it, but they notice it, and over time it erodes confidence in the firm's competence on matters that have nothing to do with English at all.

The bigger problem is when the client is not a native English speaker either. There is a persistent assumption among lawyers in continental Europe that English written for a French, Italian, German or Spanish client does not need to be particularly good, because the client will not be able to tell the difference. The truth is closer to the opposite. When two non-native speakers communicate in English, each is operating in their second language, and ambiguity that a native reader would resolve instinctively can become a real source of misunderstanding. A native English speaker reading a slightly off-key sentence in a legal opinion can usually work out what was meant. A French or Italian client reading the same sentence may not, or worse, may work out something other than what was meant. The English in those documents does not need to be more relaxed because the audience is non-native. It needs to be more precise, more direct and stripped of the nuance and indirection that a native speaker might see through but a non-native reader will simply misread.

The cost of a client who quietly misunderstands a piece of advice is not visible until it surfaces, and by the time it surfaces, it is often as a complaint, a fee dispute or a lost relationship.

At the serious end, badly worded clauses cause real legal problems. A contractual penalty that uses language a court or client interprets differently from what was intended. A notice provision that does not actually require notice in the form the parties assumed. A termination right that turns out to be narrower than the drafter believed. These are not common outcomes, but they are not rare enough to ignore.

There is also a different category of error that has nothing to do with language at all. The date on the cover page that says 2025 when every other reference in the document says 2024. The cross-reference to Section 7.3 when the relevant clause has been renumbered to 7.4 during drafting. The party name spelled two different ways across the same agreement. The schedule that refers to an annex that does not exist. The defined term used in three clauses and never actually defined. The figure in the summary that does not match the figure in the calculation. A document the lawyer has worked on for weeks is exactly the document in which the lawyer can no longer see these errors. An experienced proofreader catches them precisely because they are reading the document fresh.

What a good piece of writing looks like to a native reader

When the English is right, the document disappears. The reader does not notice the language; they focus on the substance. That is the goal of good legal drafting in any language, and it is the highest compliment a legal document can receive.

A native English reader picking up a contract drafted by a non-native lawyer who has had the document properly proofread will not think "this is well-written English". They will think nothing about the English at all. They will read the clauses, follow the argument and form an impression of the firm based on the substance of what is in front of them. That is the impression any European law firm should want to make on a foreign client, whoever and wherever that client is.

There is also a quieter benefit. Lawyers who routinely work with a good proofreader gradually internalise the corrections. The first document comes back with thirty changes, the tenth with ten, the fiftieth with three. The lawyer's own English improves through exposure to consistent, expert correction in a way that no language course can match, because the corrections are specific to the kind of writing the lawyer actually does.

How to find a proofreader who is actually right for legal work

Not every native English speaker is qualified to proofread any kind of document, let alone a legal one, and not every proofreader is qualified to work on the kind of documents lawyers produce. The market is full of general editing services that will happily take a legal contract and treat it as if it were a piece of marketing copy. The result is usually worse than no proofreading at all, because the proofreader will "improve" defined terms, soften precise language and introduce variations the lawyer then has to undo.

A proofreader worth using will have specific experience with legal documents. This does not mean they need to be a lawyer themselves. In fact, most of the best legal proofreaders are not. What they need is sustained exposure to the kind of writing lawyers produce: contracts, opinions, submissions, due diligence reports. Ask how long they have been doing this work and which kinds of documents they have handled. The answer will tell you most of what you need to know.

It is also worth understanding what a good proofreader actually does, because the word itself is misleadingly narrow. A genuine professional proofreader is not only a language check. They are a fact-checker, a logic-checker, a name-checker and a consistency-checker. They notice that the buyer is "Smith Holdings" in clause 2 and "Smith Holding" in clause 14. They notice that a clause refers to "the date specified in Schedule 3" when Schedule 3 contains no dates. They notice that paragraph 11 logically contradicts paragraph 4. They notice that a sum given as €4.5 million in the recitals appears as €4.6 million in the operative provisions. They notice that the lawyer has referred to a counterparty as "she" in one place and "they" in another.

None of this is what the lawyer hired the proofreader for, strictly speaking. All of it is what the lawyer needs. A proofreader who only checks language is doing a fraction of the job. A proofreader who reads with the kind of attention that catches small inconsistencies of fact, logic and reference is doing the job properly. Lawyers who have worked with a good proofreader over time will recognise this as the single biggest difference between a competent one and an excellent one.

They should also be a native English speaker, ideally British or American depending on the audience the document is aimed at. A native British proofreader is the right choice for documents intended for UK or European audiences, and generally a safe default for international arbitration where British English remains the dominant register. A native American proofreader is the right choice for documents going to US clients or US courts. The distinction is not trivial. Spelling conventions, punctuation and certain word choices differ enough between the two registers that mixing them in a single document looks careless. A good proofreader should be knowledgeable and experienced enough to be able to handle both consistently upon request.

Practical things to expect from a professional proofreader. They should be willing to sign an NDA without being asked twice. They should work in track changes so the lawyer can review every edit. They should not silently rewrite passages; any substantive changes or doubts should be flagged in a comment and explained. They should turn work around within an agreed timeframe and meet deadlines. They should price the work transparently, either by word, by page or by hour, and they should be able to estimate the cost before starting.

They should also know what they do not know. A good legal proofreader will not pretend to be a lawyer. If they spot something that looks like a substantive legal issue, they will flag it as a query rather than correcting it. The decision about whether to change a substantive point belongs to the lawyer, not the proofreader.

What the work is worth

Native English proofreading is not the cheapest line item in a transaction, but among the costs a law firm bears in producing an English-language document, it is one of the smallest. A 200-page due diligence report might require a few days of proofreading work at a rate that represents a tiny fraction of the legal fees being charged for the underlying matter. A 30-page contract may take a day. A 10-page legal opinion may take a couple of hours.

The calculation is straightforward. If the document is going to be read by a native English speaker, used in proceedings, form part of a transaction or represent the firm to an international client, the cost of proofreading is small in absolute terms and trivial as a proportion of what is at stake. If the document is internal correspondence between Polish-speaking colleagues, proofreading may well not be worth it. Almost everything in between is a clear case for getting a second pair of eyes on the document before it goes anywhere.

It is also worth emphasising again that the value of a good proofreader extends well beyond catching language errors. The same person who notices the awkward sentence will also notice the duplicated paragraph, the wrong cross-reference, the inconsistent party name and the date that has not been updated since the last round of negotiations. None of these are language errors. All of them are errors a fresh, careful reader catches and the document's author no longer can. The lawyer is paying for language check and receiving, as part of the same fee, a structural and factual sanity check on a document they have been too close to for too long. Any good law firm will have a “four-eye” policy where all documents are read by at least two people before being sent out. An experienced legal proofreader can certainly be a useful pair of eyes.

What is genuinely expensive is the alternative. A clause that turns out to mean something other than what the drafter intended. A submission that an arbitral tribunal reads more sceptically than it would have done because the English is awkward. A client relationship that ends quietly because the work product never quite hit the standard the client expected. These costs do not appear on any invoice, but they are real, and they are larger than the cost of preventing them.

There is also the time saved. Lawyers who try to perfect their own English writing spend hours on a task that a specialist can do in a fraction of the time, and produce a worse result while billing at lawyer's rates. The hours saved by handing the proofreading task to someone else are hours that can be spent on substantive legal work that only the lawyer can do.

The harder question

The hardest part of using a proofreader for the first time is psychological, not practical. It involves admitting that the English the lawyer has worked for years to perfect is not quite at the level required for the document in front of them. For a senior lawyer used to being the most authoritative person in any room, that admission can sit uncomfortably.

It should not. A native English barrister sending a French-language pleading to a French court would obtain a native French speaker's review of it without hesitation, and would think nothing of doing so. The standard required for legal writing is high enough that even highly competent non-native speakers benefit from a native check, and pretending otherwise serves only the lawyer's pride at the expense of the client's interest.

The lawyers who use proofreaders consistently tend not to be the ones with the weakest English. They tend to be the ones who have understood that legal writing in a second language is hard enough to require help, and that the help is cheap, available and works. These are also, not coincidentally, often the lawyers whose international practices grow fastest.

A word on the obvious alternative

It would be dishonest not to address the question every lawyer is asking themselves about now. Why pay for a human proofreader when ChatGPT, Copilot, Harvey, Claude or any of half a dozen other tools will check the English for free? The honest answer is that AI is genuinely useful for some of this work, and genuinely unreliable for the rest. A language model will catch many of the textbook errors a spellchecker misses. It will smooth awkward sentences into something that reads more naturally. For low-stakes documents read by people who will not look closely, that may be enough.

For legal documents, it is not. AI tools confidently rewrite legal language in ways that change its meaning, replace defined terms with synonyms because they read better, and silently introduce errors of fact while polishing the prose around them. They do not flag inconsistencies between clause 3 and clause 14, because they are not reading the document as a single argument. They have no idea whether the figure in the recitals matches the figure in the schedule. They will happily turn a precisely drafted obligation into something that reads beautifully and means something subtly different. Worst of all, they do all of this with complete confidence, so the lawyer reviewing the output has no way to tell which changes are safe and which are dangerous without checking every single one against the original, which defeats the purpose of using the tool in the first place.

The lawyer who runs a contract through an AI tool and sends it on without further checking has not saved money. They have transferred risk from a known cost (a proofreader's fee) to an unknown one (whatever the AI got wrong that the lawyer did not catch). When the consequence appears, it appears as a counterparty exploiting an unintended ambiguity, a court reading a clause differently from the way it was meant, or a client asking why the advice they relied on said something other than what their lawyer thought it said. None of those costs is small. All of them are larger than the proofreading fee that would have prevented them.

There is a place for AI in legal writing. It is useful for drafting first attempts, summarising long documents and surfacing questions worth asking. It is not a substitute for a human reader who knows what they are doing, has read the document properly, and is willing to put their professional reputation behind the result.