Defining And Non Defining Relative Clauses

Relative clauses in English: what actually matters

Most people learning English get confused between defining and non defining relative clauses because textbooks present them as if there is some magical rule you can memorize. There isn't. It is about information structure, not grammar rituals. Let me explain how I actually handle this with students and why the usual explanations fall apart in real usage. I worked on an edit last year for a legal translation project where a client kept insisting that commas in relative clauses were "stylistic." They had sent back a document with every restrictive clause unnecessarily punctuated with commas, which changed the meaning in at least three passages. In one case, "the employees who resigned last month" (only some of them) became "the employees, who resigned last month" (all of them). That distinction is not cosmetic. It is the difference between accurate and legally ambiguous text.

Understanding the difference between defining and non defining relative clauses

A defining relative clause narrows down which noun you are talking about. It is essential to identifying the referent. Remove it and the sentence either falls apart or means something completely different. A non defining relative clause adds extra information about a noun that is already identified. Remove it and the core meaning stays intact. That is the technical distinction, and it is correct. But knowing that will not stop you from making mistakes, because native speakers do things that break this rule regularly. Here is the thing nobody tells you: spoken English routinely drops commas on non defining clauses, especially in informal contexts. I spent two years transcribing business meetings and watched experienced native speakers say things like "My brother lives in London he is a lawyer" without any pause, no written comma, nothing. The grammar is technically wrong by the prescriptive standard, but it is incredibly common. Meanwhile, some writers will deliberately add a comma to a restrictive clause for a rhetorical effect, treating it as if the information were parenthetical when it is actually essential. The map and the territory don't always align.

Let me give you practical examples before we get into the edge cases. The man who called yesterday is here. No commas. We need to know which man. There could be multiple men. The clause "who called yesterday" defines which one. If you say "The man, who called yesterday, is here," you are implying there is only one man in the relevant context and the fact that he called yesterday is just additional detail. That changes the pragmatics entirely.

My sister, who lives in Manchester, is coming over. Commas required. You likely have one sister. The fact that she lives in Manchester is supplementary. If you remove that clause, "My sister is coming over" still identifies the person. The information is non defining by nature.

How to decide which type you actually need

The most reliable test I use is the substitution test. Can you replace the noun phrase with a proper name or a demonstrative and the sentence still work without losing critical identification? If yes, the clause is likely non defining. If no, it is defining. Take "The book that I borrowed from you is fascinating." Replace "the book" with "that book" and you get "That book is fascinating." You have now specified which book, so the relative clause is no longer doing the identifying work. But in the original sentence, "the book" is vague without the clause. The clause is doing essential work. Therefore it is defining.

Now try "Paris, which is the capital of France, attracts millions of tourists." Replace "Paris" with "this city" and you get "This city attracts millions of tourists." The identification is already complete. The relative clause is just adding a fact. Non defining. This test is not perfect. I found a case recently where a client wrote "The new software update, which fixes the security bug, is available for download." On paper this is clearly non defining because "the new software update" points to a specific known entity. But in their internal communication context, there were actually three updates pending and they meant only one of them fixed the bug. The comma was wrong. The clause was functioning as a defining clause in practice even though the grammar suggested otherwise. Context overrides the surface structure every time.

Common problems and what to do about them

The biggest issue I see is over-comma-ing. Learners, especially those influenced by Romance languages where relative clauses are often introduced with a preposition plus relative pronoun regardless of whether they are defining or non defining, tend to default to commas everywhere. In Portuguese you might say "O homem, que ligou ontem, está aqui" even when you mean a specific man among several, and that error carries over into English writing. The workaround is straightforward but requires discipline. Before you write any relative clause, ask yourself: is the noun phrase already sufficiently identified without this clause? If the noun is a proper name, a unique descriptor like "my mother" or "the president," or a demonstrative like "this" or "that" used deictically, the clause is almost certainly non defining and needs commas. If the noun is a general identifier like "the book," "the man," "the company," "the solution," then the clause is likely defining and should not have commas.

Another issue is the use of "that" versus "which." In American English, "that" is generally reserved for defining clauses and "which" for non defining ones. In British English the distinction is softer, and "which" can appear in defining clauses, particularly in formal writing. I learned this the hard way when a British colleague corrected my manuscript for using "that" in a non defining context and I had no idea what the fuss was about. The fix was simply to adopt the American convention for clarity unless you are writing for a British audience that expects otherwise. Here is a nuance that almost nobody mentions: possessive relatives. "Whose" works in both defining and non defining clauses, but it sounds increasingly formal and stiff in defining clauses when applied to inanimate objects. "The company whose stock collapsed" is grammatically fine but "the company that collapsed" is far more natural. I recommend using "that" or rephrasing entirely when the antecedent is not a person, even if "whose" is technically defensible.

When the rules stop working

There are constructions where the defining versus non defining distinction becomes genuinely murky. Relative clauses with "as" and "which" after a whole main clause are a good example. "He failed the exam, as expected" and "He failed the exam, which was unexpected" both have commas and are technically non defining, but they function differently. The "as" version is almost a fixed expression meaning "in the way that." The "which" version is genuinely supplementary. Both take commas, but the "as" construction has no defining counterpart that sounds natural. You wouldn't say "He failed the exam as expected" without the comma in formal writing, and even then it is ambiguous. Another problematic area is reduced relative clauses. "The woman standing by the window is my aunt." This is a defining clause with no relative pronoun and no finite verb. Add commas and it becomes "The woman, standing by the window, is my aunt," which implies there is only one relevant woman and her position by the window is incidental. The presence or absence of commas in reduced clauses is just as meaningful as in full clauses, but the signal is weaker because there is no overt relative pronoun to anchor the analysis.

I ran into this exact problem when coaching a student who wrote "The data collected from the survey shows a trend" and I marked it as ambiguous. Without context, we don't know if there was only one survey (non defining, with implied commas around "collected from the survey") or multiple surveys and we are specifying which data (defining, no commas). The fix was to rewrite it as "The data that were collected from the survey" or to restructure entirely depending on the intended meaning. Sometimes the grammar itself cannot resolve the ambiguity and you have to change the sentence.

A practical reference for defining and non defining relative clauses

I keep a simple checklist for my own editing work and for the students I coach. It is not exhaustive but it covers the cases that come up most often. First, identify the antecedent noun phrase. Second, determine whether that antecedent is uniquely identifiable without the relative clause. Third, if it is uniquely identifiable, set the clause off with commas. If it is not, do not use commas. Fourth, check whether the relative pronoun is "that" or "which" and apply the relevant convention for your target variety of English. Fifth, read the sentence aloud and listen for the natural pause that a comma would represent. If you feel a pause, the clause is likely non defining. If the sentence runs straight through without a break, it is likely defining. The listening test is imperfect but useful as a secondary check. Native speakers naturally pause before and after non defining clauses. That is not a stylistic choice, it is a prosodic feature of the grammar. If you can hear the pause in your head while reading, the commas belong there. If the sentence flows without interruption, no commas needed.

I will leave it at that. The topic is straightforward in principle and messier in practice than most guides admit. If you are writing formally, stick to the comma convention. If you are speaking or writing informally, you will hear and see the rules broken all the time, and that is normal. The important thing is to know what the rule is so you can decide when breaking it is acceptable and when it changes the meaning.