Does Onkelos translate וַיִּקַּח as וּדְבַר or ונסב?

Summary: The answer is that it is the former, despite R' Meir Wolf's variant nusach. Shadal and Lechem Abirim explain.

Post: In Onkelos, we encounter the following translation of Vayikach:

א,יז וַיִּקַּח מֹשֶׁה, וְאַהֲרֹן, אֵת הָאֲנָשִׁים הָאֵלֶּה, אֲשֶׁר נִקְּבוּ בְּשֵׁמֹת.

וּדְבַר מֹשֶׁה וְאַהֲרוֹן, יָת גֻּבְרַיָּא הָאִלֵּין, דְּאִתְפָּרַשׁוּ, בִּשְׁמָהָן.

This makes sense, since the general pattern is that taking of objects is NSB is Aramaic ('took') while taking of people is DBR (I would say: 'directed').

Shadal, for some reason, endorses this translation, in Ohev Ger:
ויקח משה  ואהרן ,  ודבר  ( קע"א  וג"ת  = )  לא  ונסיב .

He rejects ונסיב and endorses ודבר, based on two manuscripts. One, קע"א, is Codex 1411, which is defined here. Also, I would guess that it is the same one mentioned here:

The other, ג"ת stands for gimel targumim, and is so called because it has Targum Onkelos, Targum Pseudo-Yonatan, and Targum Yerushalmi. This Chumash also included Rashi, and was printed in Amsterdam. I don't know when it was printed.

In Lechem Abirim, a commentary on Onkelos by R' Bentzion Berkowitz, writes:

"ודבר משה, so it is in all the various texts before us. But the מעמר (R' Meir Wolf) brought the variant text of ונסיב, and said that the Meturgeman (=R' Eliyahu Bachur) misled the copyists to correct to ודבר, when he said in the root of דבר and in the root of נסב that all language of 'taking' by man is translated with דבר, end quote. And he (=Meamer) did not mention that his =(R' Bachur's) words were the words of Rashi za'l on the pasuk {Bereishit 43:15} וּמִשְׁנֶה כֶּסֶף לָקְחוּ בְיָדָם וְאֶת בִּנְיָמִן, and see what I wrote in the Simlah (on Bereishit 11:31)."

The pasuk and Rashi in question are in Miketz:

15. So the men took this gift, and they took double the money in their hand[s] and Benjamin, and they got up and went down to Egypt and stood before Joseph.		טו. וַיִּקְחוּ הָאֲנָשִׁים אֶת הַמִּנְחָה הַזֹּאת וּמִשְׁנֶה כֶּסֶף לָקְחוּ בְיָדָם וְאֶת בִּנְיָמִן וַיָּקֻמוּ וַיֵּרְדוּ מִצְרַיִם וַיַּעַמְדוּ לִפְנֵי יוֹסֵף:
		ואת בנימין: מתרגמינן ודברו ית בנימין, לפי שאין לקיחת הכסף ולקיחת האדם שוה בלשון ארמי, בדבר הנקח ביד מתורגם ונסיב, ודבר הנקח בהנהגת דברים מתרגמינן ודבר:

That is, the Targum on this pasuk is:

מג,טו וַיִּקְחוּ הָאֲנָשִׁים אֶת-הַמִּנְחָה הַזֹּאת, וּמִשְׁנֶה-כֶּסֶף לָקְחוּ בְיָדָם וְאֶת-בִּנְיָמִן; וַיָּקֻמוּ וַיֵּרְדוּ מִצְרַיִם, וַיַּעַמְדוּ לִפְנֵי יוֹסֵף.

וּנְסִיבוּ גֻּבְרַיָּא יָת תִּקְרֻבְתָּא הָדָא, וְעַל חַד תְּרֵין כַּסְפָּא נְסִיבוּ בְּיַדְהוֹן וּדְבַרוּ יָת בִּנְיָמִין; וְקָמוּ וּנְחַתוּ לְמִצְרַיִם, וְקָמוּ קֳדָם יוֹסֵף.

In the Biblical Hebrew, there is but one verb, lakechu, which distributes across both items. This is acceptable because Biblical Hebrew will use that verb for both these items. But the Targum, written in Aramaic, does not have that option, since NSB and DBR are verbs which apply to different sorts of objects. Therefore, Onkelos offers both, and associates each with the appropriate object.

R' Eliyahu Levita Bachur, in Meturgeman, writes as follows on DBR:

and similarly on NSB.

Yes, he may have gotten this from Rashi, or maybe not. This sort of thing is evident from examination of the patterns of translation.

According to מעמר, people saw R' Bachur's work and were misled into correcting their text of Onkelos. It seems quite possible to me that they would see this major work, a dictionary on the Targum, and consult that, more than noticing a random Rashi somewhere. I am actually uncertain what Lechem Abirim's objection is. Maybe that מעמר might be comfortable challenging R' Eliyahu Bachur, but not Rashi.

I would note, though, that we might be able to disprove מעמר's theory. If קע"א, Codex 1411, is so dubbed because it was published in 1411, then it was published before R' Eliyahu Bachur was born, in 1469! Perhaps the Gimel Targumim text from Amsterdam is similar in its early date, and that is why Shadal is citing these manuscripts specifically in his discussion in Ohev Ger.

Lechem Abirim referred to his discussion in Simlat Ger, earlier in parashat Noach, Bereshit 11:32. These are two works of the author, Rabbi Bentzion Berkowitz, running side by side. To explain the relationship, I will quote myself:

Lechem veSimlah, containing Lechem Abirim, which is based on R' Eliyahu Bachur dictionary of Targum called Meturgeman and Shadal's Ohev Ger, discussing differentnuschaot of Onkelos; and Simlas Ger, discussing the meanings of Onkelos's choice of translation, as he received from his Tannaitic teachers.

So while here in Lechem Abirim he was focused on whether ונסיב could be correct, over there in Noach, he wrote in Simlas Ger, he is focused on the choice of translation for לקח, in different contexts. This, as we will now see, will reveal some of מעמר's novel theory.

In parshat Noach, Rabbi Berkowitz wrote:

"ויקח תרח את אברם - it is translated ודבר. See Rashi on the pasuk {in Miketz} וּמִשְׁנֶה כֶּסֶף לָקְחוּ בְיָדָם וְאֶת בִּנְיָמִן. That he wrote that something which is taken from hand to hand, we translate it as ונסיב, while something taken via leading with words,we translate ודבר. And so in the Meturgeman {the dictionary by R' Eliyahu Bachur}. And I already commented on this in Oteh Or {another sefer by R' Berkowitz, in essay form} that even on taking women he translates with נסיבה, for she comes to his hand in one of the methods of acquisition, with kesef, shetar, etcetera, and so was it presumably before the giving of the Torah -- the woman which a man chose to establish children via her, who is called via his name, he would make with her some kind of connection and establishing of a covenant, just as it is until today for each and every nation and tongue. (And see what I wrote in parashat Lech Lecha) upon ותקל האשה בית פרעה, which is translated as ואתדברת, and not like all other languages of taking a woman, such as the aforementioned ויקחו להם נשים which is translated as ונסיבו; and so too ויקח אברם ונחור as ונסיב, and many like them. (And I have seen in the מעמר parashat Bemidbar 1:17 that he distinguishes in a different way between the language of דבר and נסב in the Targum; and the chooser shall choose.) And is it a wonder that he did not mention from the words of Rashi, zal, and brings from the Meturgeman and wrote that which he wrote, see there."

Alas, I don't think I will be able to check out the מעמר inside, as per his recommendation, since there is none freely previewable online. From what I can reconstruct, the מעמר has a different theory distinguishing these languages. And as evidence, or one data point, towards his theory, he noted that this instance in במדבר was originally ונסיב, as we see in some early manuscript. (Though I think I disproved the derivation from confusion from R' Elijah Levita Bachur.)

I am not convinced that R' Berkowitz's elaboration would synch with the first Rambam in Hilchos Ishus:

א  קֹדֶם מַתַּן תּוֹרָה, הָיָה אָדָם פּוֹגֵעַ אִשָּׁה בַּשּׁוּק--אִם רָצָה הוּא וְהִיא לִשָּׂא אוֹתָהּ--מַכְנִיסָהּ לְבֵיתוֹ וּבוֹעֲלָהּ בֵּינוֹ לְבֵין עַצְמוֹ, וְתִהְיֶה לוֹ לְאִשָּׁה.  כֵּיוָן שֶׁנִּתְּנָה תּוֹרָה, נִצְטַוּוּ יִשְׂרָאֵל שְׁאִם יִרְצֶה הָאִישׁ לִשָּׂא אִשָּׁה--יִקְנֶה אוֹתָהּ תְּחִלָּה בִּפְנֵי עֵדִים, וְאַחַר כָּךְ תִּהְיֶה לוֹ לְאִשָּׁה:  שֶׁנֶּאֱמָר "כִּי-יִקַּח אִישׁ, אִשָּׁה; וּבָא אֵלֶיהָ" ש

I don't think an actual kinyan of some sort needs to be established. Rather, I would associate דבר as the kal, devar, with the more intense pael, dabbar (indeed listed next in Meturgeman). The former would mean to convey, the latter to direct. This is appropriate only when taking someone some place, or I suppose gathering them to some purpose. Meanwhile, if a man marries a wife, this is a different form of לקיחה. There need not be kesef, shetar, or biah as a formal method of acquisition, which we then must establish as occurring pre-Matan Torah. Perhaps נסב is the more general. Or perhaps taking into one's household is considered gaining something to one's personage, and thus covered under נסב. Indeed, I wonder how closely my explanation hews to that of the מעמר.