Hedging in legal writing: how certain should you sound?

Two students answer the same problem question. One concludes that "the defendant is clearly liable". The other concludes that "it may perhaps be possible that the defendant could be liable". Both sentences are grammatical, and both fail the same test from opposite ends: the certainty of the words does not match the strength of the analysis behind them.

The fixed phrases of professional caution, "it is arguable that", "on balance", "is likely to", are covered on the Legal English register page, and learning them takes an afternoon. This page is about the harder half of the skill: choosing between them. That choice is not a style decision. It is a legal judgment, and markers grade it as one.

Certainty is a claim you have to earn

When you write "a court will find" or "a court may find", you are reporting the result of your own analysis. The reader takes your certainty language as a summary of how strong the position actually is. So there are two ways to get it wrong. Words that claim more than your analysis showed read as bluff. Words that claim less read as analysis you did not finish. Calibration means making the language an honest receipt for the work.

The scale, from settled to speculative

Notice that the top and bottom rows exist. Some points of law are settled, and hedging a settled point is as much a miscalibration as overclaiming a doubtful one. "The parol evidence question is arguably governed by ..." about a rule every textbook states without qualification tells the marker you cannot tell your strong ground from your weak ground.

The modal verbs carry fixed weights inside this scale: must, will, should, may, could, in descending order of strength. The register page covers them; the point here is that moving one step on that ladder is a substantive claim about the law, so make the move deliberately.

What actually earns certainty: three checks

Your hedge should summarise three checks, not your mood at 1am.

1. The authority check. Binding authority directly on point earns the most certainty. Binding authority you are applying by analogy earns less. Persuasive authority from another jurisdiction earns less again. No authority means you are in "arguable" territory no matter how convinced you feel.
2. The fact check. Facts squarely inside the rule earn confidence. Facts near the rule's edge, or facts the other side can plausibly recharacterise, pull the conclusion down a row.
3. The counterargument check. Certainty is earned by surviving the other side's best argument, not by declining to look at it. If your application section never tested a counterargument, you have not earned "likely"; you have earned "subject to the arguments below". The discipline of meeting the counterargument belongs to the application stage of IRAC, and it is what converts a hopeful conclusion into a calibrated one.

Run the three checks, find your row in the table, and use that row's language. That is the entire method.

The four failure modes

1. Overclaiming. "Clearly", "obviously", "undoubtedly", "it is certain that". Markers read these words as a substitute for reasoning, because that is what they usually are: if the point were clear, the question would not have been set. The professional version of this habit is worse than a style problem, since advice that overclaims and turns out wrong damages the client who relied on it.
2. Fence-sitting. "It will depend on all the circumstances" is a conclusion the reader could have written without reading the facts. Committing to a position is the assessed skill; refusing to commit is not caution, it is a blank answer wearing careful clothes.
3. Hedge stacking. "It could perhaps be argued that it may be possible that ..." Each added hedge multiplies the last, and three of them together signal fear rather than judgment. One hedge per claim, matched to the row it belongs to.
4. Monotone hedging. If every issue in your answer concludes "the court is likely to find", the reader learns nothing about which of your points is strong and which is marginal. Calibration is comparative. A strong point and a weak point in the same answer should sound different, and an answer whose conclusions vary honestly in strength reads as the work of someone who can tell the difference.

If your training pulls you toward one end

Writing cultures train different defaults. An education that rewarded producing the single correct answer trains overclaiming: the instinct that an answer without certainty is not finished. An education that treated firm assertion by a student as presumptuous trains underclaiming: the instinct that committing is not your place. Both instincts transfer straight into English legal writing, and neither matches what it expects, which is a stated position with its strength honestly marked. It helps to know which pull is yours, because then the fix is a mechanical pass in a known direction rather than a vague worry.

Translation adds a second trap: certainty words do not map one to one between languages, and the everyday "maybe" of speech is far weaker than the load-bearing "may" of legal English. Treat the table above, not your first language's equivalents, as the reference.

Calibration changes with the document

• Problem answers: conclude every issue, hedge matched to the reasoning for that issue. The structure is covered on the IRAC page.
• Advice memos: the reader needs an answer they can act on plus an honest statement of risk: "there are reasonable prospects" belongs next to what would improve them. See How to write a memorandum of advice.
• Essays: a thesis may be asserted more firmly than a prediction, because the whole essay exists to defend it. "This essay argues that the doctrine rests on a mistake" is proper essay register.
• Case notes: critique of judicial reasoning stays hedged as a matter of professional courtesy: "the reasoning is open to criticism on the ground that ...".

One conclusion, three honest versions

• "The exclusion clause does not protect the supplier." Honest when binding authority is directly on point and the facts sit squarely within it.
• "The exclusion clause is unlikely to protect the supplier." Honest when the rule fits these facts well but the supplier has an argument with some force.
• "It is arguable that the exclusion clause does not protect the supplier." Honest when the point extends existing authority and a court could properly go either way.

None of the three is the correct sentence in itself. Each is correct only when the three checks land on its row, and each is a miscalibration on any other row.

A self-check before you submit

1. Find every conclusion in your draft and run the three checks behind it: authority, facts, counterargument. Move the language to the row the checks earned.
2. Search for "clearly", "obviously" and "undoubtedly". Replace each with the reasoning it was standing in for.
3. Look at your conclusions in sequence. If they all carry the same hedge, re-rank them; your strongest and weakest points should be tellable apart from the language alone.
4. Check every "may", "might" and "could" for stacking. One hedge per claim.
5. Confirm that at least one conclusion commits. An answer with no committed position anywhere has not answered.

If you would like feedback on how your own conclusions read, Legal Writing Lab is a free feedback tool for international law students writing Australian legal English. It flags clarity and register issues in your draft, explains each one, and never writes your work for you. Your words stay yours.