Word order

Why does word order carry the meaning in some languages and barely matter in others?

"The dog bit the man" and "The man bit the dog" use identical words and mean opposite things. An English speaker takes that as simply how language works. But a Latin speaker could shuffle the same three words into any order and the meaning would not budge — the endings would keep saying who bit whom. So one language treats position as sacred and another treats it as free. Is one of them doing something the other cannot? Or are they solving the same problem with different equipment?

Start with the problem itself, before either solution. A sentence names some participants and an action, and the listener has to work out which participant did the acting and which was acted upon. That information has to be carried somehow. It cannot be left out, or the sentence is a list of ingredients rather than an event.

What are the ways to carry it? Ask what a signal needs: it must distinguish the roles, and it must survive to the listener. Two obvious candidates fall out. You can mark the participants — attach a flag to each noun saying "this one is the actor", "this one is the undergoer". Or you can fix the positions — agree in advance that the actor comes first and the undergoer comes after the verb, so a noun's role is read off where it sits.

Now notice something. If you have the flags, position is free for other work. And this is exactly what free-order languages do with it: Russian and Latin move constituents not randomly but to signal what is old news and what is the point of the sentence, which English mostly has to do with stress, clefting, or the passive. Free order is not disorder. The slots got repurposed.

And if you have rigid position, the flags are redundant — and redundant machinery tends to erode. This is where English becomes a genuine demonstration rather than a thought experiment. Old English had case endings on nouns, adjectives and determiners, and correspondingly freer arrangement. Those endings sat in unstressed final syllables, which are phonologically the most vulnerable part of a word; they reduced and merged, a process probably accelerated by intense contact with Norse speakers whose stems were similar and endings were not. As the marking blurred, position had to be relied on, and the order hardened into the subject-verb-object pattern we now find unnegotiable. Modern English keeps case only where it is cheapest to keep: on pronouns, which is why he/him still varies when dog does not.

Which way does the causal arrow run, though? Honestly, this is argued. It is tempting to say case loss forced order fixing, but you can tell the story in reverse — that order was already stiffening, which made endings uninformative and so unprotected against erosion — and most historical linguists treat the two as having reinforced each other rather than one cleanly driving the other.

One more observation, and it is the one that resists tidy explanation. Across the world's languages, two arrangements dominate: subject-object-verb and subject-verb-object. Verb-first orders are a distinct minority, and orders that put the object before the subject are vanishingly rare. Why the subject should so overwhelmingly come first, and why those particular two should split most of the world between them, has proposed answers — processing efficiency, the tendency to mention topics early, historical descent from a small number of ancestors — but no settled one. It is worth naming as an open question rather than dressing it up.

That last point is the correction worth making, because the tidy version of this story overstates itself. The relationship between case marking and word-order freedom is a strong tendency, not a law: German has a working case system and yet places its verb by rigid rule; Mandarin has essentially no case marking and its order, while firmly subject-verb-object, is looser in places than English. Role information is also carried by things this account has left out — verb agreement, adpositions, and plain plausibility, which is why "the man bit the dog" is understood as a joke rather than as nonsense.

The deeper point survives the mess. A language cannot simply omit who-did-what-to-whom, so the question is never whether it is marked but where the marking lives. When it lives in endings, position is free to carry emphasis; when the endings go, position is conscripted. English did not become simpler by losing its cases. It moved the same information somewhere else and gave up the freedom it had been buying with it.

• How languages with free order actually use it — topic and focus positions, and what English does instead.
• Why the two dominant orders in the world's languages are the ones they are, and what evidence could settle it.

Every term the collection defines is gathered in the glossary.