I'm interested in what is known about the structure of languages and how much they might differ. In Indo-European languages (and Hebrew as well), the basic sentence structure is (not necessarily in this order and with variations) Subject, Verb, Object.

I wonder if there are known languages, ancient or modern, in which this is not the case. I'd be especially interested to know if languages of the descendants of the first arrivals in the Americas, around 15 thousand or so years ago) are not like this, say Navajo or Hopi or ....

Even non-verbal languages would give me some hints if they differ.

Underlying this is the deeper question whether that sentence structure is a basic, fundamental, human thought pattern that, perhaps, precedes language.

Edited to add that I'm interested in something deeper than word (element) order. As I said above, "not necessarily in this order". Is there anything with a radically different structure, say a language without verbs or the equivalent? Or with no concept of object? So, not the order of the three common elements, but a language with different elements altogether?

Note that linguistics isn't my field, so any help is appreciated.

  • 1
    Yes, S-V-O sentence order is very common world wide; it's the second-most common type.
    – jlawler
    Commented Jun 8, 2022 at 19:53
  • 1
    @jlawler, yes, but I'm interested in whether it is universal. And thanks for the link. Just learning here.
    – Buffy
    Commented Jun 8, 2022 at 19:55
  • 1
    It's not. Read the article. For starts, not all languages have subjects and objects.
    – jlawler
    Commented Jun 8, 2022 at 19:57
  • 2
    @Buffy in ergative-absolutive languages (like Basque, most Australian Aboriginal languages, and in the Americas, the Eskimo-Aleut languages & Mayan languages amongst others) the patients of transitive verbs behave like the single argument of intransitive verbs (i.e. in the sentences "the horse ate the apple" & "the child laughed", the word "apple" is the one that behaves like the word "child", rather than the word "horse"). In these languages, it's not clear how to meaningfully define subject and object
    – Tristan
    Commented Jun 8, 2022 at 20:11
  • 1
    In Classic Maya of the hieroglyphic texts the word order was VOS (Verb - Object - Subject), although Wikipedia wrongly states it was VSO. Besides, since Classic Maya was an ergative-absolutive language, the right way to say it would be the word order was Verb - Patiens (in absolutive case) - Agens (in ergative case).
    – Yellow Sky
    Commented Jun 8, 2022 at 20:44

3 Answers 3


I'm interested in what is known about the structure of languages and how much they might differ. In Indo-European languages (and Hebrew as well), the basic sentence structure is (not necessarily in this order and with variations) Subject, Verb, Object.

Underlying this is the deeper question whether that sentence structure is a basic, fundamental, human thought pattern that, perhaps, precedes language.

There are linguists who strive to prove that languages are based on a universal grammar and that this grammar includes certain linguistic universals. Many linguists increasingly believe that hypothesis is incorrect or at least not a good approach to understanding language.

At a certain level, all humans share certain fundamental biology and a drive to survive, reproduce, and find happiness. Also, at a certain levels, all languages have expressions and ways to refer to things, convey information, ask questions, and make demands. All languages have ways of distinguishing who is doing what to whom or to express aspect. The difficulties start when we try to move beyond these abstract levels and analyze details of structure that accomplish these goals.

At the abstract semantic level, you could say that all languages have "subjects" and objects, but if you try to use this term to describe the actual structure of many languages, you run into difficulties.

In ergative-absolutive languages, like Tibetan, Hindi (an Indo-European language), Sumerian, and Basque, there are expressions of the following type:

  1. The city defended (There was a defense of the city)

  2. The army approached (There was an approach of the army)

  3. The army attacked the city (There was an attack of the city by the army)

In these languages, the word "city" patterns the same way in sentences 1 and 3, regardless of whether it is the subject or object of what is shown in the parentheses. This is called an "absolutive" construction. The word "army," however, patterns differently in sentences 2 and 3 depending on whether it is an intransitive verb, as in sentence 2 (an "absolutive construction"), or is the subject of a transitive verb, as in sentence 3 (an "ergative construction"), even though it is the "subject" in both sentences from an English point of view. I have tried to show the logic of this patterning by using "of" or "by" in the versions of the sentences in the parentheses.

Given that the patterning has nothing parallel to the English concepts of "subject" and "object," we use the terms "ergative" and "absolutive" to differentiate these functions in these languages to avoid confusion.

In many languages, "objects" are often incorporated into verbs in a special form so that the expressions are more like "he rock-threw," rather than "he threw rocks" so that it is awkward to talk about objects other than in a general semantic sense, any more than you would talk about the prefixes in "proceed," "exceed," "precede," and "accede" as really just adverbs incorporated into the verb "ceed"/"cede." In such languages, overt "subjects" and "objects," when they exist, often merely coindex things already within the verb and are not treated as arguments of the verb. An example is Mohawk.

In Mandarin, it is more efficient to talk about topics and verb complements than "subjects" and "objects," since verbs have no agreement rules, nouns and verbs do not inflect, and word order has many variations that are arguably more determined by pragmatic and semantic criteria rather than by syntax. Furthermore, in Mandarin and Japanese, sentences like "elephants have long noses" generally could mark both "elephants" and "noses" equally as "subjects," which is something inconsistent with the usual conception of subjects, which requires only one per clause.

It turns out that languages include very few universals at the structural level, even if we can subject them to some forms of universal analysis. Let's look at a few cases of things you might expect to be structured in the same way.

  1. Number of objects: Some languages have no grammatical structure to indicate it, like Japanese. Some differentiate singular and plural, but apply different rules to animate and inanimate things. Some add a mandatory dual for anything numbering exactly two. English has one way to say "trees." Arabic has at least four, depending on the number and what you mean. Some languages force you to indicate the number of objects on nouns; some, on verbs. Number is a universal physical reality, and yet languages use different structures or no structures to convey it.

  2. Pronouns: When Japanese do not want to repeat a noun, the usual practice is simply to omit it. There are a myriad ways to refer to participants in a conversation and to third parties with a convenient short reference, but no real systematic way that amounts to a system of pronouns. Many times its system of honorific and in- and out-group distinctions clarifies who is doing what to whom in a way completely different from what we are used to in English.

  3. Relative directions: A few languages have no expression for "behind," "left" or "in front of"; instead they use cardinal directions to locate objects in space relative to each other. In these languages, you say "step to the north," rather than "step to the left," in the appropriate situation.

  4. Tense: Mandarin verbs have no person, tense, or mood; and yet Mandarin speakers have no difficulty at all expressing the semantics of these things by other means. Time of event relative to time of speech is a physical reality, but this reality is not built into a single grammatical structure of Mandarin, but is accomplished through a variety of different means.

  5. "Yes" and "no": All languages have ways to respond to "yes-or-no questions," but most languages have no exact equivalents of "yes" and "no," and others have no such words at all. French and German use three words. Japanese has two words that equate better with "right" and "wrong" than "yes" and "no." Mandarin, Irish, and Latin really have no such universal words that can invariantly be used by themselves to affirm or deny propositions. In these languages, you typically use a short version of the question to affirm or deny something.

I'd be especially interested to know if languages of the descendants of the first arrivals in the Americas, around 15 thousand or so years ago) are not like this, say Navajo or Hopi or ....

The indigenous languages of the Americas belong to many different language groups and do not really have any unifying language features, nor do they generally have features that clearly mark them off as a group from other languages in the world.

It is true, however, that there are many languages in the Americas that have structures extremely different from English. For examples, read up on verbs in Navajo or on obviation in nouns and verb inflection in Munsee or on (evidentiality) in Quechua. However, I could also point out "unusual things" in German (word order) Ancient Greek (verb conjugations), Mandarin (aspects), Arabic (derivation from roots), Swahili (noun classes), American Sign Language (double question words), and Coptic (second tenses). These languages have structures that are extremely different from English and extremely different from each other with implications throughout their grammar.

A simple way to make sense of all this is a hierarchy that separates what humans linguistically have in common and what they don't. If you imagine seeing a man who is hungry. You are witnessing a state that is understandable in all human societies. At this level 1, languages do not differ.

If you are motivated to communicate what you see, you are conceiving a generic communication at level 2 that is again common to all languages. All languages have some way of doing this.

At level 3, different languages have a different range of tools to express this communication. These tools allow you to express certain things but also force you to express certain things because of how they work. At this level, all languages differ, but some have various tools in their tool boxes that are similar to some others.

At level 4, the speaker composes a specific sentence. At this level, each language has diversity, since different sentences can be used for the same communicative purpose. In English, I might say: "he is getting hungry," "he's hungry," "he's gotten hungry," or "that guy looks like he is starving." This menu of options is unique to English, but other languages may have some of these tools or something similar. Other languages may have a completely different set, and there is little you can say about any universal structure in how different languages would express this general thought.

  • Thanks. This gives me the background I need and helps clear up my misconceptions.
    – Buffy
    Commented Jun 21, 2022 at 14:50

It seems likely that “thought” (whatever that is) is different from “expression”, and that language is a system for encoding thoughts into some system of expression. Linguists analyze that system of expression – grammar – in terms of various analytic classes, for example “subject”, “object”, “indirect object”, “preposition”, “noun”, “verb” and so on. In light of research into how you might diagnose or define these things across languages, it has become increasingly difficult to maintain that there are many hard universal analytic entities that every language has. For example, many languages do not have a class of words “postposition” (English is an example), and many languages do not have a class of words “preposition” (Japanese is an example) – postposition vs. preposition has neven been a candidate for “universally required category”, it has always been simply “a possible distinction”. Certain Austronesian and Salishan languages are proposed as examples of languages that do not distinguish nouns versus verbs. Even in English which seems to be a paradigm case of a language with nouns versus verbs, many words can be both nouns and verbs (boil, plow etc) and some inflected forms are both nouns and verbs at the same time (cooking, boiling, plowing).

There is a universally-available structural fact found in all languages, namely “constituency”, which is the property that some two words can be “tightly knit” to the point that we say that they act as a single unit. In English, in the sentence “I think that the old goat chased the young cat”, we can minimally say that “the old goat”, “the young cat” and “the old goat chased the young cat” are units of some type (“noun phrase” and “sentence”), and that one can substitute one noun phrase for another and still get a well-formed utterance. Even then, we cannot actually agree on what things are constituents. In particular, one view of constituent holds that the string of words forming a constituent must be all and only the words of an utterance from i to j, but there are also theories that say that “what” and “eat” are a constituent in “What did the child eat”, because it makes the semantic interpretation work out better. Parentheticals are a particularly clear illustration of the problem of discontinuous constituents, for example McCawley’s 1982 example “Tom may be, and everyone is sure Mary is, a genius” – the single NP a genius is “in” two structural slots in the sentence. This is not fatal for a strict view of constituent that they have to be “all and only the words from i to j”, because many alternative structures have been proposed. We may simply need to relax the claims about what a “constituent” is, maybe a constituent is a “group of words that worm a unit”, not necessarily words in a row. But then, what is the point of “constituent”?

Well, it still remains a fundamental concept for forming propositions. Perhaps (I think I heard this idea about a half century ago) the “thought” aspect that gets encoded via grammar into a physical manifestation (an actual sentence) is the human-invariant thing, and the invariant part of grammar is extremely small.

  • words that "worm a unit"? Commented Jun 9, 2022 at 16:36

All possible permutations of Subject, Verb, and Object are attested as the basic word order in at least one language.

There are different estimates of the proportions but per the Wikipedia page on Word Order a 2013 update to Dryer's 2005 study got the following proportions:

  • Subject Object Verb: 41.0%
  • Subject Verb Object: 35.4%
  • Verb Subject Object: 6.9%
  • Verb Object Subject: 1.8%
  • Object Verb Subject: 0.8%
  • Object Subject Verb: 0.3%
  • Unfixed: 13.7%

The precise percentages vary quite a bit between estimates, but the order seems fairly consistent, with SOV slightly more common than SVO, a big jump down to VSO, then VOS, OVS, and only a very small number of OSV languages.

Even languages with free word order tend to have one particular order which is typical, with alternative orders being used to emphasise different parts of the sentence.

A far from complete set of examples of languages with each order as their default unmarked word order are given below:

  • Subject Object Verb: Ancient Greek, Bengali, Hindi, Hungarian, Japanese, Kannada, Korean, Latin, Malayalam, Persian, Sanskrit, Urdu
  • Subject Verb Object: Chinese, Dutch, English, French, German, Hausa, Italian, Malay, Russian, Spanish, Thai, Vietnamese
  • Verb Subject Object: Amazigh, Biblical Hebrew, Classical Arabic, Irish, Māori, Tagalog, Welsh
  • Verb Object Subject: Malagasy, Baure, Car
  • Object Verb Subject: Apalaí, Hixkaryana
  • Object Subject Verb: Warao

Of the example languages listed here, Baure, Apalaí, Hxkaryana, and Warao are all indigenous to the Americas. Navajo and Hopi both use subject-object-verb as its basic word order.

Trends observed here, and also within languages that permit multiple word orders (i.e. have "free" word order) are that elements nearer the start of the sentence are generally perceived to be more pertinent. This likely explains why in the vast majority of languages, subjects appear before objects except when the object is being emphasised.

In some models of syntax, the verb and object are taken to constitute a single predicate phrase at the underlying level, although one or more part may be moved somewhere else in the surface realisation of the sentence (meaning the words may not actually be said one after the other). This may explain why word orders with the verb and object separated by the subject are rarer than those with them together (assuming the same ordering of subject and object).

  • See the edits to the question. I'm looking for something different than just word order. Sorry that it's hard to be clear in a field not my own.
    – Buffy
    Commented Jun 8, 2022 at 20:28
  • (A small quibble -- I'm not sure Ancient Greek's unmarked order is SOV, or that it has a single default unmarked order at all; SVO order is also very common and doesn't seem to be in any way more marked than SOV.)
    – TKR
    Commented Jun 8, 2022 at 20:35
  • the examples given here are taken from Tomlin, Russell S. (1986). "Basic Word Order: Functional Principles" via wikipedia, with some adjustment to the names of a few languages by me
    – Tristan
    Commented Jun 9, 2022 at 7:31
  • But the question specifically states "not necessarily in this order and with variations" in other words the question already accepts that there are different permutations of subject "verb" and object. Commented Jun 9, 2022 at 14:13
  • 1
    @Araucaria-him that's true, I did miss that bit. And the subsequent clarifying edit makes it extra clear that this answer does not address the intended question
    – Tristan
    Commented Jun 9, 2022 at 14:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.