Persian and Farsi Speakers in English: Ezafe-Influenced Phrasing and Detection

You wrote the whole thing yourself, in careful English, and a detector still called it AI. If you grew up speaking Persian, this stings in a particular way, because the habits that make your English feel correct are often the same ones a tool reads as machine-made. When it comes to Persian Farsi ESL AI detection, much of the trouble traces back to one small piece of grammar you stopped noticing years ago. It’s called the ezafe, and it quietly shapes how you build noun phrases in every language you write.

Persian ezafe trains you to build noun phrases head-first and to chain modifiers with an “of”-like link, so your English fills up with long, uniform “of” phrases and stacked noun strings, and that low-variation phrasing is exactly what AI detectors tend to score as predictable, even when every word is your own.

(Persian and Farsi are two names for the same language, so this applies whether you say *fârsi* at home or studied “Persian” in class.)

Key takeaways

• Ezafe is a Persian linking particle that puts the head noun first and trails its modifiers behind it, which maps neatly onto English “of” phrases and stacked noun chains.
• Leaning on that one structure produces long, even noun phrases that read as low-perplexity and low-burstiness, the statistical signature detectors associate with machine text.
• English lets you branch two ways. Native writers mix “of” phrases with possessives and compounds; ezafe habits push you toward “of” almost every time.
• The fix isn’t new vocabulary. It’s variety: convert some “of” links to compounds or possessives, turn heavy noun phrases back into verbs, and vary sentence length.
• A detector reports a probability, not proof. Keep your drafts and notes, and revise for real clarity rather than chasing a score.

What ezafe actually is

Ezafe (Persian *ezâfe*, “addition”) is a short, unstressed vowel, an /-e/ sound, that links a noun to whatever describes it: an adjective, a possessor, or another noun. English speakers meet it most easily as a rough stand-in for the word “of.”

Two facts matter for your writing. First, it usually isn’t written. The Persian script leaves short vowels off the page, so the ezafe lives in how you say a phrase, not how you spell it. You’ve been producing it your whole life without a symbol to point at. Second, the head noun comes first, and everything that modifies it trails behind, each piece clipped on with another ezafe.

So in Persian you’d say the equivalent of *ketâb-e bozorg-e Ali*, literally “book big Ali,” to mean “Ali’s big book.” The adjective follows the noun; the owner follows both. And you can keep chaining: the results of the survey of the students of the university slots together as one smooth, head-first string. It only branches one direction.

Where Persian Farsi ESL AI detection trouble starts

Here’s the collision. English gives you two directions to build a noun phrase, and ezafe only trained you for one of them.

You can put modifiers in *front* of the head noun, the left-branching way: “Ali’s big book,” “the student survey results.” Or you can trail them *behind* with “of,” the right-branching way: “the big book of Ali,” “the results of the student survey.” Both are correct English. Native writers switch between them without thinking, and that switching is part of what makes their prose uneven.

Persian branches one way only, head first. So the “of” version feels native and the front-loaded version doesn’t. Reach for a noun phrase under time pressure, and “of” is the default that comes out. One “of” is fine; the problem is the chain. Carry the ezafe habit into English and you get sentences like this:

> *The purpose of this study is to examine the effect of the level of anxiety of the students of the program on the quality of the outcomes of the assessment.*

Every word is correct. But count the “of”s: six, all doing the same job at the same interval, so your reader (human or machine) can predict each next word before it arrives.

Why long, uniform noun phrases flatten your writing

AI detectors mostly measure two things, and ezafe-style phrasing hits both. You can read the full breakdown in our piece on perplexity and burstiness, but the short version fits here.

Perplexity is how surprising each next word is; a repeated “of X of Y of Z” frame is about as unsurprising as English gets. Burstiness is how much your sentence length varies, and noun-chain sentences tend to run to the same medium length, one after another, which keeps it low.

Low perplexity plus low burstiness is the signature language models leave, which is why your handwritten paragraph can land in the same statistical zone as generated text. The detector isn’t reading your effort; it’s reading the evenness of your phrasing. This isn’t a fringe worry: a 2023 study by Liang and colleagues found detectors flagged non-native English writing far more often than native writing, partly because it draws on a narrower, more predictable band of structures. The tool guesses, and it guesses against you.

How to revise ezafe-heavy English

You don’t need bigger words. You need more than one way to attach a noun. Try these in order.

Break the “of” chains

Scan for any noun phrase with three or more “of” links in a row. Keep one and convert the others: turn a link into a possessive with ‘s, fold two nouns into a compound, or split the phrase across two shorter sentences. That anxiety example becomes *We looked at how student anxiety affected assessment scores* — same meaning, the “of”s gone, and a verb now carrying the load.

Turn heavy nouns back into verbs

Persian and formal English both love nominalization, packing the action into a noun so it can take an ezafe or an “of.” “The examination of the data” hides a perfectly good verb. Free it: *we examined the data.* De-nominalizing adds verbs, shortens phrases, and naturally varies your rhythm, all at once.

Mix your branching on purpose

Once you notice the two directions, alternate them. Let one phrase front-load (“the student survey results”) and the next trail with “of” (“the limits of that survey”). No single structure should dominate a paragraph. That variety is what raises your burstiness without touching your argument.

Vary sentence length

This is the highest-leverage habit for any ESL writer. Break one long sentence in two. Let another run long with a subordinate clause. Drop in a short, blunt one to land a point. That deliberate unevenness is hard to fake with a formula, and it reads as more confident English.

The false-positive caveat, and why to keep your drafts

Be clear-eyed about what these tools are. A detector outputs a probability, not a verdict, and as the Liang study showed, that probability skews against non-native writers. A high score on your own writing is a false positive, not evidence you did anything wrong.

So protect yourself the boring, reliable way: keep your outlines, rough drafts, notes, and version history. A document with an edit trail is your proof of authorship, and it beats any score a detector can print. Revise your phrasing because clearer, more varied English is better writing, not to win a number. When you want a second read before you submit, you can check a draft and humanize it to see which sentences still look predictable, then fix only those.

Frequently Asked Questions

What is ezafe, and why would it affect my English writing? Ezafe is a Persian linking particle, an unwritten vowel that connects a noun to whatever modifies it: an adjective, a possessor, or another noun. It roughly corresponds to English ‘of,’ but Persian puts the head noun first and chains the modifiers after it. When you write English, that head-first habit follows you, so you reach for long ‘of’ phrases and stacked noun strings instead of mixing in short compounds and possessives. The result is even, predictable noun phrases, and evenness is what detectors tend to flag.

Does speaking Persian actually make my English get flagged as AI? Not directly. A detector doesn’t know your first language. It reads statistics: how predictable your words are and how much your rhythm varies. Ezafe-influenced English leans on long, uniform noun phrases with the word ‘of’ repeated, which flattens that variation. A 2023 study by Liang and colleagues found detectors misclassified non-native English far more often than native English. So the bias is real, but it keys on the phrasing pattern, not on Persian itself.

How do I break up long ‘of’ chains without losing my meaning? Find any noun phrase with three or more ‘of’ links in a row. Turn one link into a possessive with ‘s or into a compound, and split the phrase across two shorter sentences. ‘The analysis of the results of the survey of the students’ can become ‘We analyzed the student survey results,’ which says the same thing with fewer moving parts and a verb doing the work.

Is it cheating to revise my writing so it stops getting flagged? No. Editing your own English to read more naturally is ordinary revision, the thing every writer does. The real harm here is false positives against non-native speakers, not the students fixing them. The one line you shouldn’t cross is passing off text a machine generated as your own original work.

What is the fastest single change a Farsi speaker can make? Vary how you build noun phrases. Instead of stacking ‘of’ after ‘of,’ mix in possessives, compounds, and the occasional short sentence that carries just one idea. That single habit breaks the even rhythm detectors respond to, and it reads as more fluent English at the same time.

The takeaway

Ezafe gave you a fast, elegant way to stack meaning onto a noun, and that habit doesn’t switch off when you write in English. It shows up as one long “of” chain after another, the flat rhythm a detector rates as predictable. Give your nouns more than one way to connect and your sentences more than one length. That’s not gaming a tool; it’s writing English that varies the way English is supposed to. To keep sharpening the habit, browse more guides on writing and detection, or see which plan fits if you’ll be checking work often.