Proclisis and Enclisis in Detail

Two Greek words name the two positions. Πρόκλιση (proclisis) is a clitic leaning forward onto the word after it — σε βλέπω. Έγκλιση (enclisis) is a clitic leaning back onto the word before it — βλέποντάς σε. The terms matter because the two are not stylistic variants: each one belongs to particular verb forms, and using the wrong one produces a sentence that is not merely marked but ungrammatical.

The basic rule is on where clitics go. This page is the full system — the complete inventory of both positions, the reason the split falls exactly where it does, the living variety that does it the other way round, and what happens to the accent when the clitic moves. Competing resources routinely get this wrong in one specific way: they present enclisis as an option available "for emphasis" or "in older style". It is not an option anywhere in standard Modern Greek. It is obligatory with two verb forms and impossible with every other one.

What a clitic actually attaches to

A clitic is not a small word — it is a piece of a word looking for a host. It has no accent of its own, it cannot be said in isolation, and it cannot begin an utterance. Whatever it attaches to, host and clitic are pronounced as one phonological word with one main stress. That single fact predicts almost everything else on this page: it is why nothing can be inserted between them, why the direction of leaning has to be fixed rather than free, and why the written accent sometimes has to be repaired when the unit gets too long.

Proclisis: the default, with no conditions attached

In standard Modern Greek, a clitic goes in front of every finite verb — every form that carries a person ending. There is no context in which a finite verb loses proclisis; there is only the separate matter of two non-finite forms that never had it.

Older descriptions call δεν, θα, να, ας, the conjunctions and the wh-words triggers of proclisis, as though the clitic would otherwise drift backwards. In the modern standard language that is a historical fossil rather than a synchronic rule — remove every one of those words and the clitic still stands in front, as Τον ξέρω and Μου τηλεφώνησε show. The trigger list is genuinely useful, but as a list of the sentences where an English speaker is most tempted to get it wrong, not as a list of licensors.

Enclisis: exactly two hosts

Two verb forms take their clitics behind them, and they are the two that stand outside the tense system — neither can take δεν or θα in front of it: the bare positive imperative and the gerund in -οντας.

The positive imperative. Full treatment on clitics with the imperative.

The gerund. The -οντας form is invariable and non-finite, and it always takes its clitics behind it — even when μην is standing in front of it (Μην ξέροντάς το, υπέγραψε), because the position is fixed by the verb form itself and not by what precedes it. Its formation is on the gerund in -οντας.

That is the entire inventory for verbs. (A possessive clitic leans backwards too, but on a noun — το βιβλίο μου, ο δάσκαλός μου — a different host, handled on the second accent.) One further construction looks like a third case and is not: the pointing word νά and the question word πού take a tiny set of subject forms behind them — Νά τος! Πού 'ναι τος; Those are a different, closed set of pronouns, not the clitics on this page, and they are covered on subject pronouns.

The negative command is not an exception

Πες μου has the clitic behind; Μη μου πεις has it in front. That looks like an exception to the imperative rule, and it is not — because the second sentence contains no imperative. Greek has no negative imperative at all: a prohibition is μη(ν) + subjunctive, an ordinary finite clause, which takes proclisis like every other finite clause. See negative commands.

Once you see that the negative "command" is a different construction wearing the same English clothes, the two-host rule stays intact: there is no imperative in the sentence, so there is no enclisis in it either.

Where the system came from, and where it survives

The modern split is the residue of an older, tidier arrangement. In Medieval Greek the clitic was essentially a second-position element: it followed the first stressed word of the clause, which meant enclisis after a clause-initial verb (είδα τον) and proclisis whenever something else — a negator, a complementizer, a question word, a fronted phrase — occupied that first slot. The standard modern language generalised the proclitic pattern to all finite verbs and left enclisis stranded on the two forms that can never have anything in front of them.

Cypriot Greek kept the medieval system. In everyday Cypriot speech a clitic follows a plain indicative verb, and it moves in front exactly where the medieval rule said it should — after negation, after a complementizer, after a wh-word, after a focused phrase. Both of the sentences below are correct in their own variety; neither is a mistake. The Cypriot one is written with standard-Greek words so that only the clitic's position differs — a real Cypriot speaker would also change the negator (μεν) and the verb ending.

Notice that the second half of the sentence is identical in both: να μην ανησυχεί has proclisis in Cypriot too, because να and μην are exactly the environments the old rule reserved for it. More on the variety on Cypriot Greek. What this means practically is that if you learn Greek from a Cypriot friend and then write a formal letter to an Athenian office, this is the feature that will show — and that είπα του is a dialect feature in a Cypriot mouth and an error in a standard one — same string, and the verdict depends entirely on which variety you are speaking.

What the switch does to the accent

Because host and clitic form one phonological word, moving the clitic to the back lengthens that word — and Greek allows a stress only within the last three syllables. When enclisis pushes the original stress outside that window, the spelling has to repair it with a second accent.

Nothing comparable happens under proclisis: με άκουσε adds a syllable to the front of the unit, which the three-syllable window does not care about. The exact conditions, including the cases where no second accent is needed at all, are on accent on clitic clusters.

Common Mistakes

1. Leaving the clitic behind the verb under θα. The future particle does not change the position; it just adds a word in front of it.

2. Treating a να-clause as if it were a command. The verb after να looks a lot like an imperative in some persons, but it is finite and it takes proclisis.

3. Putting the clitic in front of a gerund. The gerund is one of the two forms that require enclisis, so proclisis is impossible there.

4. Enclisis in a wh-question. A question word does not license a following clitic; the verb is finite, so the clitic is in front.

5. Treating ας like an imperative. Ας πούμε translates as "let's say", and the English feels like a command — but ας takes a finite subjunctive.

Key Takeaways

• Proclisis (clitic in front) is the default with every finite verb in the standard language, with or without δεν, θα, να, ας, a conjunction or a question word.
• Enclisis (clitic behind) belongs to exactly two verb forms: the bare positive imperative and the gerund in -οντας. No other verb form allows it — a possessive clitic after a noun (ο δάσκαλός μου) is enclisis on a different kind of host.
• The old "trigger" lists describe a Medieval system, not the modern one — in Modern Greek nothing has to license proclisis.
• The negative command is not an exception: μη(ν) + subjunctive contains no imperative, so it behaves like any finite clause.
• Cypriot Greek preserves the medieval pattern (Είπα του) and is a variety, not an error — but it is not the standard.
• Enclisis lengthens the phonological word and can force a second written accent (Άκουσέ με, Φώναξέ τον); proclisis never does.