This thought occurred to me after having read a closed question here. I drew five language trees for an identical sentence rendered in five different languages, and the result was quite interesting.
Sentences were structured a little bit differently according to a type, and so were the language trees.
I took English, Russian, Finnish, Estonian and Swedish as sample languages. Then I wrote and structured it in Dutch, Spanish, Polish, Bulgarian and French.
The languages, however different in their typologies, have some similar features (e.g. English is more like Swedish, and Finnish is much like Estonian). There were also differences in language trees depending on language analytism, flections and possible word order. The differences may be, however, described by submodalities of personal ideolect as well. I have not yet compared the languages thorough an atlas with wide parameter scope.
Unfortunately, my knowledge of other languages (e.g. that of Arabic, Yoruba or Chinese) was not sufficient to present more data on languages of other types, so I am presenting this question to the Community.
If different types are indeed described by different structures, then what is the main typological factor for the language tree discrepancy?