Mode	Text / Speech	Sentence #			Subordinations
					English 1	English 2
Subtitle translation	Sir Ken Robinson	50			1	1
Target language	Reordering Σi=1n-1 Σj=i+1n     I(xj<xi)	± Nestings   { }                  {{ }}                {{{ }}}			Semantic changes Δ
Turkish 1	—	—	—	—	—
Turkish 2	—	—	—	—	—

Difference in analysis: None

Comment on parsing: The status of [2], [3] and [4] as functionally independent or subordinate is slippery. For a modifier like a relative clause, the distinction between defining and non-defining is useful, since a non-defining modifier is generally seen as making an independent assertion. But that distinction can be blurred when what’s described is an indefinite entity. That’s the case in the English version of the sentence above, where [2] describes “a conversation” in [1], and [3] and [4] provide information about “a wonderful woman” in [2]. In the English version, [2] is seen as a modifier defining the conversation, while [3] and [4] are seen as functionally independent comments on the woman. But the distinction is far from clear-cut. In the Turkish version of the sentence, [2], [3] and [4] describe definite rather than indefinite entities – “the conversation” in [1] and “the wonderful woman” in [2]. Also, [2], [3] and [4] in Turkish are all grammatically deranked for person and tense. Both these features strengthen the feeling that [2], [3] and [4] in Turkish are all defining modifiers and should be seen as functionally subordinate to their respective parents. Despite this difference in feeling, [2] has been marked as functionally subordinate, and [3] and [4] as functionally independent, in both English and Turkish, and indeed in all language versions of the sentence. That choice is in keeping with the principle followed in this study of parsing all language versions of a given sentence the same way for ease of comparison, unless there are felt to be major differences in structure between them.