But who are these water carriers? I would say since the beginning specified the important ones, the intent here of woodcutters and watercarriers are the less-important Israelites as well. And these might be a subtype of "stranger" but need not even be so. Rather, all those mentioned in the first verse are the "important" and those in the second are the "unimportant", and all, kulchem, are to enter into this covenant with Hashem. And this is the theme here, and so this is the peshat.
But other meforshim understand this differently. For instance, Rashi writes:
|both your woodcutters [and your water drawers]: [The mention of these people separate to the main community of Israel] teaches us that in the days of Moses, Canaanites came to convert [to Judaism], just as the Gibeonites came [to convert] in the days of Joshua. This is the meaning of the verse regarding the Gibeonites,“And they also acted cunningly…” (Josh. 9:4), [i. e., pretending that they had come from a far away country. When they were discovered, Joshua made them woodcutters and water drawers for Israel (see Josh. 9:3-27). Likewise here, the Canaanites attempted to deceive Moses,. but they did not succeed, and Moses did not accept them to be Jews. Rather,] Moses made them woodcutters and water drawers [i.e., slaves for Israel]. — [Tanchuma 2; Yev. 79a; see Rashi Gittin 23b]||מחטב עציך: מלמד שבאו כנענים להתגייר בימי משה, כדרך שבאו גבעונים בימי יהושע. וזהו האמור בגבעונים (יהושע ט, ד) ויעשו גם המה בערמה, ונתנם משה חוטבי עצים ושואבי מים:|
What I marked in blue in the Judaica Press translation of Rashi is their supercommentary on Rashi, but not actually in Rashi's words. And it is this point that I wish to contemplate in this post. Was Moshe fooled, or wasn't he? Were these people just slaves / servants, or were they genuine converts?
It seems that this midrash, taken from Tanchuma, has at least three inputs into it:
1) The pasuk transitions from וְגֵרְךָ אֲשֶׁר בְּקֶרֶב מַחֲנֶיךָ, where ger is understood to mean convert, into מֵחֹטֵב עֵצֶיךָ עַד שֹׁאֵב מֵימֶיךָ. This would imply that the water-drawer and wood-hewer are different people among the spectrum of converts.
2) Elsewhere, we have converts who fulfilled these same roles -- the Giveonim. Since this was written in the time of Moshe, and not in the time of Yehoshua, it is strange to talk of converts who fulfill these roles. Therefore, we have the urge to state that history repeated itself.
3) We have a basis to state that history did repeat itself. Even though in Biblical Hebrew, the word גם can often simply be a word with strengthens and emphasizes, the subject matter (see here for discussion and examples), rather than meaning "also", a midrash can interpret it literally in accordance with the meaning it later took on most often, such that it means also. Then, since the pasuk in sefer Yehoshua 9:4 reads:
|ד וַיַּעֲשׂוּ גַם-הֵמָּה בְּעָרְמָה, וַיֵּלְכוּ וַיִּצְטַיָּרוּ; וַיִּקְחוּ שַׂקִּים בָּלִים, לַחֲמוֹרֵיהֶם, וְנֹאדוֹת יַיִן בָּלִים, וּמְבֻקָּעִים וּמְצֹרָרִים.||4 they also did work wilily, and went and made as if they had been ambassadors, and took old sacks upon their asses, and wine skins, worn and rent and patched up;|
Even as "also", depending on where one places it in the phrase, it could mean different things. In the translation above, it is an additional action of theirs, rather than them being an addition people who did.
But we can midrashically state that they as well did this. We could have no impetus to say this if we didn't have the "difficulty" raised by points (1) and (2), but if the difficulty is there, and the word gam is here, we can exploit it.
Now, Rashi does not say here that Moshe was not fooled, unlike Yehoshua by the Giveonites. This has to be deduced from elsewhere. First, his source, which is the midrash Tanchuma:
This seems to imply that the Giveonim themselves came to Moshe earlier, but he did not accept them. That would be difficult -- did Yehoshua restore the Giveonites to water-carrier status? Or were these other Giveonites. Perhaps these should be other Canaanites, as Rashi makes it explicitly. Even if we make these other Canaanites, the Giveonites only assumed this role after they had tricked Yehoshua into a covenant that he could not undo. Why should Moshe let them assume this role, if they tried to trick him and did not succeed?מחוטב עציךאמר רבי יצחק בן טבלי:מלמד, שבאו הגבעונים אצל יהושע בן נון וקבלן, שנאמר: ויעשו גם המה בערמה וגו' ( שם ט ד).שמהו גם המה?ללמדך, שבאו אצל משה ולא קבלן
Also, how does the derasha work, if this is so? Point (1) was that it started with them being geirim, because the pasuk leads in from וְגֵרְךָ אֲשֶׁר בְּקֶרֶב מַחֲנֶיךָ into מֵחֹטֵב עֵצֶיךָ עַד שֹׁאֵב מֵימֶיךָ. If Moshe didn't accept them, why should they be called geirim, rather than, say, avadim? We would have to undo point (1), which is a very compelling point, and for what cause? Indeed, that is another good question. Every point in a midrash should have some textual basis. Here, it is difficult to find a basis for saying that Moshe was not fooled. Indeed, I can find a basis or two for saying the opposite! This is very strange. But maybe we can say (within the midrash rather than within Rashi) that since these are Giveonites and not Canaanites, had Moshe been entirely fooled, there would be no cause for them to come at some later point to convert.
Regardless, it matters not only what a source says, but also how Rashi modifies it. On more than one occasion, Rashi seems to slightly modify midrashim, or else apply a midrash from one source onto a foreign pasuk. Here, Rashi adds Canaanites, to help clarify, and Rashi does not say that Moshe did not accept them. And Rashi adds lehitgayer. Perhaps he changed this deliberately, in the interests of peshat (or midrashic peshat), because it works well to say that Moshe did accept them, and that this is what geirim means. Of perhaps he was working off a different girsa of the Tanchuma, but I wouldn't count on it.
Ramban assumes that they did not even try to trick Moshe. Rather, they came in peace to him, and this is how Moshe established them; or else (I think) that they tried to trick him but that it did not work. Thus, Ramban understands these water-carriers and woodcutters to be the erev rav. But in explaining Rashi and the midrash:
והזכיר עוד הטף והנשים והגרים וחוטבי עצים ושואבי מים אשר להם מערב רב. והביא שם הטף, להביאם בברית, כי אפילו עם הדורות העתידים יכרות ברית, או כדי לזכותם כטעם שאמרו בהקהל (חגיגה ג א). ש[sic]ורבותינו אמרו (יבמות עט א): ששבאו קצת כנענים בימי משה כדרך שבאו בימי יהושע ונתנם חוטבי עצים ושואבי מים לעדה ולמשכן ה'. שואין הכוונה לומר שרימו אותו, אבל באו אליו להשלים עמו, כי כן המשפט כאשר ביארנו (לעיל כ י יא). שוכך מצאתי במדרש תנחומא (נצבים ב):שללמדך שבאו אצל משה ולא קיבלם.כלומר שלא יכלו לרמותו לכרות להם ברית, אבל עשאם מיד חוטבי עצים ושואבי מים:
In Etz Yosef's commentary on Tanchuma, we read of a different proposed girsa, which also appears in the Yalkut Shimoni:
The same appears in Buber's version of Tanchuma, which he emends based on ktav yad romi and ketav yad rch"m.
I am not entirely convinced that this variant text is not an insertion / clarification. The word קבלן and the idea of multiple efforts drives this interpretation, I think. So that this was multiple efforts. And that they were not able to trick Moshe, as they were able to trick Yehoshua, but instead made them immediately into woodcutters and water carriers.
Etz Yosef notes that Rashi goes against this girsa (or perhaps created girsa) in the Yalkut Shimoni, because he says lihitgayer and says kenaanim, and kederech, such that these were different people, namely Canaanites. (And perhaps -- though he doesn't say so explicitly -- in Rashi's version, Moshe was indeed tricked. But maybe not, and his focus is only on Giveonitim / Kenaanim.) But he thinks that the gemara in Yevamot daf 74 (a typographical error for daf 79) suggests like the Yalkut.
How is it implied like Yalkut? Because in Yevamot 78b we read:
R. Hana b. Adda stated: David issued the decree of prohibition against the nethinim, for it is said, And the king called the Gibeonites, and said unto them-now the Gibeonites were not of the children of Israel etc.
and in the next amud, 79a, the setama di-gemara objects and reinterprets as follows:
Was it David, however, who issued the decree of prohibition against the nethinim? Moses, surely, issued that decree, for it is written, from the hewer of thy wood to the drawer of thy water! — Moses issued a decree against that generation only while David issued a decree against all generations.
But Joshua, in fact, issued the decree against them, for it is written, And Joshua made them that day hewers of wood and drawers of water for the congregation, and for the altar of the Lord! — Joshua made his decree for the period during which the Sanctuary was in existence while David made his decree for the time during which the Sanctuary was not in existence.
So according to this, we are speaking of the same Giveonim, with this quasi-status.
Judaica Press also refers us to Rashi's commentary on Gittin 23b. There, the when the gemara talks of slaves who are not part of the covenant, the gemara states:
מאי טעמא לאו משום דכתיב (אתם) (במדבר יח, ג) גם אתם מה אתם ישראל אף שלוחכם ישראל אמרי דבי ר' ינאי לא מה אתם בני ברית אף שלוחכם בני בריתand Rashi comments:
לא מה אתם בני ברית - מהולים ובאו בברית מצות:
אף שלוחכם בני ברית - לאפוקי עובד כוכבים אבל עבד בן ברית הוא דכתיב (דברים כט) מחוטב עציך וגו
Thus, Rashi on Shas points to our local pasuk to show that the chotev eitzecha was entered into the covenant, and was a ben brit. What the writers of the Judaica Press translation may be implying here is that once we combine this with Rashi's comments on chumash, we see that these were slaves, and these were people who tried to trick Moshe into accepting them as converts. This then should be evidence that Rashi maintains that they did not succeed. And that Moshe did not accept them except as woodcutters and water carriers, that is, they say, slaves.
I am not convinced that we should combine Rashi on Shas with Rashi on chumash. Locally, all the words Rashi does use would convince me that these were other accepted converts, who used a similar strategy. That we could look inside sources, or remote Rashis, to get a picture entirely other than what Rashi seems to want to tell us may well be irrelevant. A slave is not the same as a geir, and the ger aspect seems a peshat concern and a midrashic concern. That Rashi gives a different interpretation in explaining the halachic position of one opinion in a dispute should not necessarily concern us. Indeed, I think it should not.
I would also point out that if we look in Daat Zekenim miBaalei HaTosafot (and here) the explanation of חוטב עציך are the male slaves which שואב מימיך are the female slaves. And he does not bring in any suggestion that there were Canaanites who tried to convert. So we can establish this, if we want, as entirely separate idea, such that they do not need to be slaves.
Conclusion: At the end of the day, what do I think? I don't know. I am leaning towards saying that of course Moshe did not accept them. How so? That Canaanites came, just as the Giveonites came in the days of Yehoshua, and tried to become full Israelites. If Moshe had accepted them fully, they would have no need to serve in these particular roles, of water-carriers and woodcutters. He did let them in with some status. It is unclear if this is as converts or as slaves. The gemara which says that Moshe put this decree in force only for his generation need not be speaking of Giveonites in particular but whichever people came to join klal Yisrael. And these Canaanites were perhaps accepted as geirim, and thus can be termed gercha. Or else they are gerim in the sense of strangers, though I think this is slightly less likely. And so these people were in a special caste, just for that generation, and so he did not fully accept them. Meanwhile, Yehoshua, at least initially, was entirely fooled and accepted them as full converts and new members of Israel. Only later, after their trickery was discovered, were they unaccepted and became the water carriers and wood cutters. See in sefer Yehoshua to see how this would work. I am not certain I would deem this role to necessarily be one of slaves, however. But I am still a bit uncertain.