Here’s an overview of the complexities of word order in Attic Greek. It’s a textbook excerpt, but it’s pretty approachable.
I don’t know German, but I gather from this thread and the linked text that word order is considerably more flexible in Greek than in German.
Old English (Anglo-Saxon) was a highly inflected language with a flexible word order, but it became steadily simpler over the centuries until it turned into modern English.
From the wiki on Old English grammar:
Word order was flexible:
That was one of the fun things about studying Irish - it’s VSO. Well, I thought it was fun, for a lot of people it’s part of what makes the Celtic languages hard. As Celtic languages (Irish, Gaelic, Welsh, Manx, etc.) used to be a lot more common VSO word order also used to be more common before being displace by various invasions and conquering by the likes of Julius Ceaser, Angles/Saxons/Jutes, and the Vikings.
Uh, yeah - languages are not “less mature” or “more mature”.
Nor do I see anything inherently better in one particular word order over another.
This might be off-topic. English (maybe other languages, I don’t know) has conventions that go deeper than O/S/V - the order of adjectives is also quite important; as quotably noted by author mark Forsyth:
Sometimes authors and poets deliberately break this convention in order to create impact.
And there’s this:
Johanna, you say that the frequency of the orderings of S, V, and O is as follows:
> Of these, almost all the world’s languages are either SOV, SVO, or OSV. The other
> three are rare or extremely rare. The most common word order among the world’s
> languages is SOV. Our familiar SVO comes in second place.
That doesn’t match sources I’ve checked, which say:
> In percentages, SOV order is found in 47.5% of languages, SVO order — in 41%,
> VSO order in 8%, VOS order in 2.1%, OVS in 0.9% and OSV in 0.3% of languages.
Other sources have a separate category for languages without a fixed order, so they say:
> SOV 41.0%
> SVO 35.4%
> VSO 6.9%
> VOS 1.8%
> OVS 0.8%
> OSV 0.3%
> UNFIXED 13.7%
Did you make a copying error in your claim, or do you have a better source than me?
Also, we should note that not all languages are correctly described using the distinction between subjects and objects. Some split up the arguments of a verb into ergative and absolutive instead. This means that the six orders for the S, V, and O are not useful in classifying those languages:
But you can have green great dragons!
It’s a great dragon (as in, species “great dragon”) which is green.
That seems a little unrealistic, though. Even a flexible language like Russian tends to have a preferred ordering. It kind of has to, since going out of order is typically tool for emphasis. It is like saying that Asian languages are specifically tonal while a language like English is not: while technically true, IME, a large amount of information in spoken English is carried in tone (as well as cadence), and often an indistinct utterance can be resolved through its melody (here, the rigidity of order can be immensely helpful).
That does match what I said.
Um, no, you give SOV, SVO, and OSV as being the most common, while it’s actually SOV, SVO, and VSO.
This is a fantastic podcast, especially for answering all of those “how come the English language does this weird thing?” questions.
What about a language like Arabic? If spoken Arabic is VSO, mightn’t we say written Arabic is OSV?
[runs for cover! 
]
Klingon is OVS, because … alien. There are a number of linguistic in-jokes like that built into the language, where it has some feature that is rare or absent in natural human languages.
I’m not sure how rare it is, but OVS is common in Romanian. Here’s an example: il voi transmite eu. It means, I will transmit it, where eu=I, voi transmite=will transmit, and il=it. OV is even more common than OVS because S is actually included in the verb form, which can make it hard for us (Romanians) to learn certain languages.
Linguists generally consider German to be SOV in its underlying structure. This appears to be contradicted by simple declarative sentences such as “Er hat ein Buch”. But “Er hat ein Buch gekauft” has the auxiliary send but the main verb last. Also “Er will ein Buch kaufen”. And the SOV order is always used in subordinate clauses. English does none of these things.
Another interesting point. SVO languages tend to have adjectives follow their nouns. French does this but English doesn’t and linguists consider this somewhat inconsistent. SOV languages have adjectives preposed. I remember a sentence “Der vor dem Hotel sitzende Mann” (the in front of the hotel sitting man).
Please forgive if I have mucked up the German; I studied it 60 years ago (including the poor man sitting in front of the hotel).
Oh, whoops, I did reverse the order of the letters in VSO, sorry. Meant I totally “VSO” and thinking was I of Arabic and Irish. But to process the letters in the right order was hard for my brain somehow. At least with LGBT, GLBT, TBGL, etc. it still means the same thing if you inadvertently scramble the letters.
I don’t know much about Romanian, but Wikipedia says it is classified as a SVO language. In most languages there are exceptions to the standard word order. Even in English there are common (often fossilized) phrases with unusual word order like “I thee wed” (SOV) or “You I like!” (OSV) or “Believe you me” (VSO).
My first reaction was to think Wikipedia may be wrong in this case because “let’s be honest” no one really cares about Romanian. But then I thought the article was based on information from people who study languages for a living, so there must be a mistake I was making.
I translated several other short English statements into Romanian and every time I came up with OVS-type structures. “Okay,” I said to myself. “Let me think like a linguist. How would I proceed with this if I wanted to establish whether Romanian was an SVO or OVS language?” Well, a linguist would not start from a statement like “I will transmit it,” but rather “The military will transmit the message,” where both the subject and object are nouns. And I have to admit that the Romanian translation shows an SVO-type structure: “Militarii vor transmite mesajul.” However, as soon as the speaker uses a pronoun for the object instead of a noun, the OVS-type structure becomes the norm. “The military will transmit it” -> “Il vor transmite militarii.” But because word order is less rigid in Romanian, it is possible to use an SOV-type structure as well.
I want to point out that statistically speaking OVS-type statements are by far more numerous than SOV-type statements. These are not exceptions or examples of ossified structures - they’re everyday statements that speakers build or adjust to communicate basic stuff. This is one of the reasons why I think foreigners must find Romanian impossible to study. And yet Arab people seem to learn it amazingly fast.
Like everything else in existence (particularly in linguistics), neat categorization doesn’t consistently work. Even in the cases where you can more or less say that a particular language is SOV, SVO, VSO, VOS, OVS, or OSV, there will be occasional exceptions. Sometimes the number of exceptions is so large that it’s hard to categorize the language as being consistently one of those six categories at all. Sometimes it will be better just to say that there’s no fixed order at all. Sometimes the distinction between subjects and objects used in those six categories doesn’t really correctly explain the different sorts of arguments (i.e., nouns and pronouns) of verbs in the best way, as in ergative-absolutive languages. The same is true about much of how languages work.
If you expect all languages to be explainable in rigid structures, you’re going to be disappointed. Languages don’t work that way. I should know. I spent three years as a graduate student in linguistics after getting a bachelor’s degree in mathematics. At that time (1974-1977), during the height of generative grammar’s delusional control of the field, people thought that it was possible to explain how language worked with a relatively small set of principles. It isn’t. I went back into mathematics after discovering that there was no way to consistently apply any of the theories then battling for control of the field to real languages. Languages are huge messy structures for which no single neat set of rules will work in all cases.