It wouldnвЂ™t be as relevant to count the total number of words displaced, since this would penalise languages with larger amounts of short words more than others. My solution of counting the total length of clusters is not a perfect indicator of the amount of word order change, but at least it doesnвЂ™t decrease when word order change is at a maximum. The reason why I found it was misleading to count only the number of clusters (and not their length) was that once the scrambled-ness of word order passes a certain point, the number of clusters in a passage actually decreases as several clusters are absorbed into each other.

It is a rough method, but at least it can be applied with relative consistency. Whereas the cluster below would count as a 3:ДЄnd this other cluster would also count as a 3: To measure how much word order has changed, I counted the total number of clusters (that is, how many groups of crossed lines there were) and the total cluster length (the combined total number of lines in all clusters).ДЇor example, the cluster below would count for a length of 2: King James Version, 1611, with modernised spelling Clusters of crossed lines indicate a change in word order. I placed the original Greek text just above corresponding sections of the translation, and drew a line from word to word (I ignored little words like вЂњthe,вЂќ which functioned differently in different languages or were absent altogether).

Where possible, I tried to find translations which were more formal and вЂњword for wordвЂќ rather than вЂњthought for thoughtвЂќ, although in practice all readable translations are to some extent вЂњthought for thoughtвЂќ since truly mechanistic translations (interlinear translations) canвЂ™t really convey the sense of the passage for anyone who does not already know the rules of Greek. The most convenient choice was, naturally, a passage from the Greek New Testament, Matthew 16:24-28.

I found a Koine Greek text which has been translated very carefully many times throughout history into many different languages. Although the sample text is short, at least this exercise can give a taste of how similar and distant the various languages are to Ancient Greek in this particular respect. ItвЂ™s a crude, chunky experiment looking at how blocks of data move positions in different languages to express more or less the same thing according to different syntax rules. This is not a subtle study about ineffable shifts in semantics. What I like about studying word order is that itвЂ™s a fairly obvious part of the language and you can see it move around. In this study, IвЂ™ll explore how much word order change occurs in several published translations of the same Koine Greek sample text, analysing eleven translations spread out over six languages: Latin, English, German, French, Mandarin Chinese, and Modern Greek. For instance, how closely can a Latin translation mirror Greek word order? Are Romance languages any closer to Greek word order than Germanic languages? Was medieval English more similar to Koine Greek syntax than contemporary English? And roughly how close is Modern Greek to its ancestor? What I want to investigate is how much the word order changes when translating a passage from Ancient Greek (specifically Koine) into various other languages. Art of Conversation II (LвЂ™Art de la conversation II), Rene Magritte, 1950

Ancient numbers in different languages