A week-late dvar Torah for Noach

A strange pasuk occurs at then end of Sheni in Noach, and the last pasuk in the sixth perek of Bereishit (6:22):

וַיַּעַשׂ נֹחַ כְּכֹל אֲשֶׁר צִוָּה אֹתוֹ אֱלֹהִים כֵּן עָשָׂה.
Thus did Noah; according to all that God commanded him, so did he.

There are two ways to parse this pasuk, neither of which is entirely smooth. The first is rendered above, with the semicolon after Noach. That is, Noach did. Further (or to reinforce), according to all Hashem commanded him, so did he do. In this case, וַיַּעַשׂ נֹחַ seems a phrase that ends too abruptly. What did Noach do. It reads somewhat stiltedly.

The other possibility is to but the pause after Hashem. That is,
וַיַּעַשׂ נֹחַ כְּכֹל אֲשֶׁר צִוָּה אֹתוֹ אֱלֹהִים | כֵּן עָשָׂה
Noach did in accordance with all Hashem had commanded him, indeed, so he did.
Here, the כֵּן עָשָׂה is stilted and looks like is does not belong, particularly after the completion of the previous phrase. That is, we would expect the phrase to be כְּכֹל אֲשֶׁר צִוָּה אֹתוֹ אֱלֹהִים כֵּן עָשָׂה, or rather KaAsher Tziva Oto Elokim Ken Asa.

In other words, both the first phrase וַיַּעַשׂ נֹחַ and the last phrase כֵּן עָשָׂה make powerful claims on the middle phrase, and when each claims it, the other is left in the lurch.

My intuition from seeing many psukim tells me that וַיַּעַשׂ נֹחַ should be separate, since כְּכֹל אֲשֶׁר צִוָּה אֹתוֹ אֱלֹהִים כֵּן עָשָׂה is a pattern seen in other places, and וַיַּעַשׂ נֹחַ can truly stand on its own. כֵּן עָשָׂה on its own is more problematic. Both choices are still stilted, but this is the better of the two alternatives.

The trup agrees. There is etnachta, the primary dichotomy of the verse, after וַיַּעַשׂ נֹחַ. Onkelos is no help, since he just translates word for word, and the Aramaic can be parsed in the same two ways.

Ramban seems to suggest the other way. The dibbur hamatchil cites the pasuk וַיַּעַשׂ נֹחַ כְּכֹל אֲשֶׁר צִוָּה אֹתוֹ אֱלֹהִים, leaving off כֵּן עָשָׂה, and then speaks of the כֵּן עָשָׂה echoing the וַיַּעַשׂ נֹחַ. I have to wonder if Ramban really intended that, and if so, did he have the same trup as we have (and did he look at trup as a perush?)

While Onkelos was no help, Targum Pseudo-Yonatan (for Yonatan only wrote on Nach, not Torah) might be. Targum Yonatan translates only וַיַּעַשׂ נֹחַ כְּכֹל אֲשֶׁר צִוָּה אֹתוֹ אֱלֹהִים, leaving the כֵּן עָשָׂה out. This might reflect a different girsa of Chumash that Yonatan was translating. Or, he might have been trying to argue on the trup by showing the connection. Or, the text might have simply lost the last two words of the pasuk in the targum. Most likely, though, is that the translation of a pasuk in the next perek, 5 pesukim later in 7:5, was transposed here by scribal error.
That pasuk is:
וַיַּעַשׂ נֹחַ כְּכֹל אֲשֶׁר-צִוָּהוּ ה.
And Noah did according unto all that the LORD commanded him.

Here the pasuk has YKVK, translated as "the LORD," whereas earlier the pasuk had Elokim, which was translated as "God." Here, there is no "Ken Asa" at the end of the pasuk, so no parsing problems are present.

Further, if you look at Tg Yonatan in general (at least in our Mikraot Gedolot), both Elokim and YKVK are recorded in Tg Yonatan as YKVK. And so, the chances of transposition are great, and the fact that "Ken Asa" is missing is not meaningful, and does not denote a possible change in girsa (even if it did, the fact that our Masoretic text is more difficult to read and understand would make it more authentic, just in terms of the flow of textual changes. It is easy to remove "Ken Asa," making it parallel to a later pasuk, than to add it where it would cause textual problems.)

Update: I neglected to mention that Rashi sides with the trup. See inside.