The Tricky Prepositions: je, po and the Accusative of Direction
The prepositions survey makes the general case: each Esperanto preposition has one fixed meaning, none governs a case, and about twenty of them are consequently the easiest thing in the language. This article is the depth piece on the corners that bite - je, the preposition with no meaning at all; po, whose arithmetic everyone gets wrong once; the scope question hiding inside anstataŭ and krom; and the -n of direction, which turns any preposition of place into a preposition of arrival. ( had its own reckoning in quantity; it is not re-tried here.)
je: the wildcard, on purpose
Rule 14 of the Fundamento makes two promises. Every preposition has a definite, constant meaning - and when no meaning fits, you use , which has none. Most language designers would have quietly stretched an existing preposition and hoped nobody noticed; Zamenhof instead labelled the overflow pipe and made it official. It is the single most honest decision in the grammar, and it exists because prepositions are exactly where a planned language would otherwise start to rot - every natural language jams its leftover relations into some overworked little word (at three o'clock, good at chess, ill at ease), and the jamming is arbitrary. Esperanto's arbitrary bucket at least announces itself.
A century of use has settled onto a short, learnable list:
tria- at three o'clock. Clock time is's biggest job.
- taller by a head. The measured difference.
- I believe in God. Kredi takesfor belief-in.
- he held me by the hand. Grabbing, holding, pulling by a part.
sano!- to your health. The fixed phrases.
Two pieces of discipline keep respectable. First, it is a last resort, not a comfort blanket: before writing it, check whether , per, or would say something sharper, because one of them usually would, and seasoned readers clock lazy the way you would clock someone who answered every question with "it depends". Second, do not invent new uses; the wildcard is only understood where convention has already filled it in. The same rule 14 offers an escape in the other direction: wherever would go, a bare accusative is equally correct - trian means exactly tria , and the time expressions article leans on that swap throughout.
po: distribute the number
shares an amount out. It stands immediately before a number, and it says that the number applies to each member of a group, not to the group in total:
.- Three children received two apples. Two apples in total, and presumably an argument.
.- Three children received two apples each. Six apples, no argument.
The rule that prevents every classic mistake: modifies the number, and nothing else. English each floats around the sentence (they each got two, they got two each) and learners try to float the same way - onto the recipients ( , wrong), onto the verb, anywhere but where it belongs. It is glued to the numeral. If you cannot see a number (or a numberish quantity like , a bit each), has no business in the sentence. The same logic runs the rate readings: is ten kilometres per hour - the distributing the ten over the hours.
Then the famous dispute, which deserves the honest version. Does the noun after take -n when the phrase is an object? Zamenhof wrote in the Fundamento's own exercises - as a true preposition, blocking -n like any other. Most modern writers reason instead that belongs to the numeral rather than governing the noun, so the phrase as a whole is still the object of the verb and marks itself normally: . PMEG, surveying a century of both, declines to condemn either - and so do we, though the house style takes the side the survey article already took: write the -n, because the apples are what got received. What you should not do is treat either form as an error when you meet it. This is one of Esperanto's very few genuinely unsettled points, and pretending otherwise would be marketing.
anstataŭ and krom: whose role?
(instead of) and (except / besides) look like ordinary prepositions and mostly behave like them. The depth question they hide is one of scope: when the sentence has both a subject and an object, which one does the replacement or exception apply to?
.- He wrote the letter instead of me: in my place. I was meant to do the writing.
.- She invited him instead of me. I am the one who went uninvited.
The plain pronoun reads as a stand-in for the subject; the -n moves the replacement to the object role. This is not the ordinary no-accusative-after-prepositions rule being broken - it is a deliberate disambiguation device, sanctioned by PMEG precisely because alone cannot tell you who would have done what. works identically: - I saw everyone except him; he went unseen, so he wears the object's -n. Where no object exists there is nothing to disambiguate, and the plain form is all you need: .
carries a second ambiguity all of its own, flagged in the survey and worth pinning down properly here: it means both except and besides, which are close to opposites.
.- Everyone except him came. (He did not.)
, .- Besides him, she came too. (He did, and so did she.)
Context resolves most cases, and two habits resolve all of them. If you mean in addition, put ankaŭ in the main clause - it announces the besides reading before any doubt forms. If you mean except, a negative or a totality word (, ) nearby does the same job in the other direction. And when the stakes justify a longer word, the language keeps two unambiguous specialists: escepte only ever means except, aldone only ever means besides. Reaching for them is not defeat; it is what they are for.
The accusative of direction: en la domon
Prepositions of place tell you where something is. Add -n to the noun and they tell you where the motion ends up:
.- I am in the house.
.- I ran into the house.
.- The cat jumped onto the table.
.- The mouse darted under the bed (and is now there).
.- The bird flew in among the trees.
This is the accusative's second job, seen from the preposition's side. English burns a separate word on each contrast (in/into, on/onto) and then runs out - there is no under-to or between-to - so it leaves the mouse ran under the bed genuinely ambiguous: was it running about under there, or did it dart under? Esperanto marks the difference every time, with one ending: is where the running happened, is where it finished.
The boundary that keeps the rule tidy: prepositions that already mean direction never take it. Al is towards, el is out of, ĝis is up to - direction is their entire content, so -n would add nothing, and is plainly wrong. The candidates for the direction -n are the position words: , sur, sub, inter, apud, , , and friends. (Trans takes it too, when the crossing lands you on the far side: .) The same -n builds the direction adverbs - , homeward; , whither - and that thread is picked up where it belongs, in the adverbs article's direction section.
What to take away
is the official wildcard: clock times, measured differences,, holding by a part, the fixed phrases - and nothing new of your own invention. A bare accusative can always stand in for it.distributes a number across a group and is glued to the numeral:is two apiece. Never.- The
-nafteris a real, unresolved dispute: Zamenhof wrote, modern majority practice writes, PMEG accepts both. We write the-n; mark nobody wrong. - After
and, a plain pronoun stands in for the subject,-nfor the object:(my job),(I got dropped). means except and besides.flags besides, a negative orflags except, andescepte/aldoneare unmistakable.- Place prepositions plus
-nmean arrival:,,.,andnever take it - they are direction already.
See also
- Prepositions - the survey this article deepens: the full system, da versus de, and the tio-ke workaround.
- The accusative - the -n ending itself, whose second job the direction rule is.
- Adverbs - the -en direction adverbs (hejmen, kien) built from the same -n.
- Quantity - da, the other preposition with a whole article to itself.
- Time expressions - je la tria in the wild, and the bare-accusative alternative.
- The -um wildcard - je's exact counterpart among the suffixes: the official piece with no meaning of its own.