Persian has no grammatical gender, no noun cases, one plural ending that covers almost everything, and adjectives that never change shape for anything. All of that is true, and none of it is why the first month is hard. The first month is hard because of a vowel that is not written down.
It is called the ezafe, it is a short e, and it does the job English does with of and with word order: it binds a noun to whatever is describing it. Ketab-e man is my book. Ketab-e bozorg-e baradar-e man is my brother's big book, and every join in that phrase is a sound that does not appear on the page at all.
Three facts that decide how Persian is learned
- no gender, no articles, no cases. One pronoun covers he and she, nouns do not decline, and there is nothing to memorise alongside a word. For a learner arriving from German, Russian or Arabic this removes most of what made those hard.
- the ezafe joins a noun to what follows, an unwritten short e linking a noun to its adjective or its owner: ketab-e man, my book, ketab-e khub, a good book. It is the single most frequent piece of Persian grammar and it is invisible on the page.
- the written and spoken forms differ systematically. This is not slang against formality; it is a regular set of sound changes that turns nearly every sentence into a different shape. A learner who studies only the written form will read a newspaper and not follow a taxi driver.
A noun phrase is a chain with invisible joins
The consequence is immediate and it catches every beginner. To read a Persian phrase aloud you have to know where the ezafe falls, and to know where it falls you have to know already which words are describing which. So reading is not decoding followed by understanding; the understanding has to come first, at least in outline, or the phrase comes out as a list of unconnected nouns.
Nothing about that is difficult once it is named, and being named is most of the fix: hearing the joins in speech, where they are perfectly audible, is what teaches your eye to put them back on the page. Thirty two letters, joined, with the vowels left for you to supply is where the mechanics sit, and the short vowels are the second thing missing from the same line of text.
The alphabet is borrowed, and it fits the way borrowed clothes fit
The script came from Arabic for a language that is not related to Arabic. Four letters were added for sounds Arabic does not have, the ones written pe, che, zhe and gaf, and the fit is good in that direction. It is worse in the other one. Arabic distinguishes several consonants that Persian pronounces identically, and the spelling kept all of them, so there are four ways to write the sound z and three to write s, and which one a word takes is a fact about that word.
There is no rule available, because the rule lives in a language you are not learning. This is one of the few places where Persian asks for straight memorisation, and it is worth knowing early that the difficulty is inherited rather than designed. Stress is the third unwritten thing, and it is the kindest of them: it sits on the last syllable of the word almost always, and the exceptions are systematic rather than scattered. Stress that sits at the end, and what the spelling hides takes both.
The verb system is genuinely as tidy as advertised
Every Persian verb has two stems, a present one and a past one, and once you have those two everything else is regular endings applied without exception. The irregularity that exists is in which past stem belongs to which present stem, and that is a list of a few hundred rather than a system.
The number is small for a reason that is worth understanding early: most of what English does with separate verbs, Persian does with a noun plus one of a handful of light verbs, chiefly the ones meaning to do, to become and to give. To telephone is telephone plus to do. So the vocabulary you need for actions is mostly nouns you were learning anyway. No gender, no cases, and one of the tidiest verb systems anywhere is the full account, and it is the part of this subject where the reputation is accurate.
Written into spoken
Four rules turn the language in the book into the language in the room
The gap between written Persian and Tehrani speech is a set of regular conversions rather than a second vocabulary, which is the useful news almost no course delivers early.
- Long a becomes uBefore m and n. nan is nun, and Tehran is Tehrun
- To be attachesThe verb contracts onto the end of the word in front of it
- Verbs contractmiravam becomes miram, and the pattern holds across the common verbs
- ra shortensTo ro after a vowel, and to a bare o after a consonant
An afternoon to learn, and months of listening before they stop being computed
Learning three of the four does not get you three quarters of a conversation. Ordinary speech applies all of them at once and announces none of them.
The book and the street are two registers, and the conversion is regular
Written Persian and Tehrani speech differ systematically, and a learner taught only the written form can read a newspaper and miss half of a conversation about lunch. What makes this bearable, and what almost no course says early enough, is that the differences are rules rather than a separate vocabulary.
Long a becomes u before m and n, so nan is nun and Tehran is Tehrun. The verb to be contracts and attaches to the word in front of it. Common verbs contract their own middles, so miravam becomes miram. The object marker ra shortens to ro, or to a bare o after a consonant. Four rules, an afternoon to learn, and then months of listening for them to become automatic rather than computed.
This is the stall in Persian, and it is not a vocabulary stall. It is the gap between a learner who has done everything the book asked and a conversation that was never written in that register. Two registers, and why the book does not sound like the street is the systematic version, and Tehrani speech, and the vowels it swallows is the practice that turns the rules into hearing.
A dictionary with the vowels in, and a machine that supplies them
The single most useful new tool here arrived with the models. Paste a paragraph of Persian and ask where the ezafe falls and what the short vowels are, and the answer comes back as text you can read aloud. It replaced asking a teacher, which is the only thing that previously worked, and it is wrong often enough that you should use it on material you can check rather than on the thing you are about to say to someone.
Wiktionary is the free dictionary that solves the smaller half of the same problem: its Persian entries carry a romanisation with the short vowels written in, which the Persian spelling does not, so a word looked up there arrives pronounceable. It replaced a bilingual dictionary that gave you the meaning and left the sound to chance.
Chai and Conversation has been publishing Persian lessons for English speakers since 2011, in the spoken register rather than the written one, which is unusual and is the reason to use it. The early episodes are free and the structured course is paid. The Arabic layer, and the words English quietly took is worth an hour alongside it, if only for the pleasure of discovering how many ordinary English words came this way.
Whether the script or the language is what you want next
Persian sits at a junction and the two roads out of it lead to different kinds of work. The script, the sounds, and the distance from standard to street shares the alphabet and a large layer of vocabulary, and shares almost nothing of the grammar, so the transfer is real and narrower than it looks. Devanagari, and the Urdu it is mutually intelligible with is the other direction: centuries of Persian vocabulary went east and stayed, so the words are familiar and the script is not. Both are inside every tongue this atlas speaks. The question worth asking before either is whether the four conversion rules above have become automatic yet, because until they have, the language you can hold a conversation in is still only the written one, and the method, which is the same whichever language you picked is where that gets turned into a schedule.