Shemot 2:5 reads:
“The daughter of Pharaoh went down to bathe in the Nile, and her maidens were walking on the side of the Nile, and she saw the basket within the marsh, and she sent forth et-amata and took it.”
The word amata is ambiguous. Pharoah’s daughter might have stretched forth her hand (amah), or she might have sent forth her maidservant (amah). Her maidservants were mentioned earlier in the verse, and so this would be one of those many maidservants. But, stretching forth a hand to thereby take something also works well as a phrase.
Rashi discusses this ambiguity, and pits Chazal against dikduk:
her maidservant: Heb. אֲמָתָהּ, her maidservant. Our Sages (Sotah 12b), however, interpreted it as an expression meaning a hand. [The joint from the elbow to the tip of the middle finger is known as אַמָּה, hence the cubit measure bearing the name, אַמָּה, which is the length of the arm from the elbow to the tip of the middle finger.] Following [the rules of] Hebrew grammar, however, it should have been vowelized אַמָּתָהּ, with a dagesh in the mem. They, however, interpreted אֶת אֲמָתָהּ to mean her hand, [that she stretched out her hand,] and her arm grew many cubits (אַמוֹת) [so that she could reach the basket]. [From Sotah 12b, Exod. Rabbah 1:23]
את אמתה: את שפחתה. ורבותינו דרשו לשון יד. אבל לפי דקדוק לשון הקודש היה לו להנקד אמתה מ"ם דגושה. והם דרשו את אמתה את ידה, שנשתרבבה אמתה אמות הרבה:
(This is a wonderful derash in which we get all three meanings -- maidservant, hand, or cubit. When translating this, I like to preserve the ambiguity by translating amata as “her hand-maiden.”)
Some would point to this Rashi and say -- perhaps simplistically and crudely -- that peshat in the pasuk must accord to dikduk. And so, the word means maiden. And Chazal are making things up which are not correct, due to a desire to fabricate, embellish, or even for lack of grammatical knowledge. And they cannot both be true, because either Pharaoh’s daughter fetched it or her maidservant did. Others might say -- perhaps in a more nuanced fashion -- that the peshat is indeed maiden, and the derash which invokes hand and cubit means to convey a deeper message, where that deep message might be a life-lesson (“Try, and you too can obtain things out of reach!”) or an insight into a theme or undercurrent in the Biblical narrative. (E.g. While the events seem a natural progression or happenstance, this supernatural stretching of her hands shows that divine guidance, which is surely a theme in the story.)
However, I would question whether we must really assign “maid” to the peshat and “hand” to the derash.
ב,ה וַתֵּרֶד בַּת-פַּרְעֹה לִרְחֹץ עַל-הַיְאֹר, וְנַעֲרֹתֶיהָ הֹלְכֹת עַל-יַד הַיְאֹר; וַתֵּרֶא אֶת-הַתֵּבָה בְּתוֹךְ הַסּוּף, וַתִּשְׁלַח אֶת-אֲמָתָהּ וַתִּקָּחֶהָ.
וּנְחַתַת בַּת פַּרְעֹה לְמִסְחֵי עַל נַהְרָא, וְעוּלֵימְתַהָא מְהַלְּכָן עַל כֵּיף נַהְרָא; וַחֲזָת יָת תֵּיבְתָא בְּגוֹ יַעְרָא, וְאוֹשֵׁיטַת יָת אַמְּתַהּ וּנְסֵיבְתַּהּ.
The word וְאוֹשֵׁיטַת means that she stretched something forth. And note how, in the Aramaic, they place a dagesh in the mem, just as Rashi asserted would need to stand in the Hebrew for it to mean “hand”. I suppose you could say that Onkelos is deviating from peshat in order to follow the midrash of Chazal, but maybe you could say that he either (a) disagreed from a dikduk perspective or (b) had a different nikkud.
Saadia Gaon, as well, translates this as hand in the Tafsir, his translation of Tanach:
Note the word yadha, “her hand”. He is not one to follow derash over peshat.
Perhaps we could look as well to the gemara in Sotah 12b:
ותשלח את אמתה ותקחה ר' יהודה ור' נחמיה חד אמר ידה וחד אמר שפחתה מ"ד ידה דכתיב אמתה ומ"ד שפחתה מדלא כתיב ידה ולמ"ד שפחתה הא אמרת בא גבריאל וחבטן בקרקע דשייר לה חדא דלאו אורחא דבת מלכא למיקם לחודה ולמאן דאמר ידה ליכתוב ידה הא קמ"ל דאישתרבב אישתרבובי דאמר מר וכן אתה מוצא באמתה של בת פרעה וכן אתה מוצא בשיני רשעים דכתיב (תהלים ג, ח) שני רשעים שברת ואמר ריש לקיש אל תיקרי שברת אלא שריבבתה
Or, in English:
And sent her handmaid to fetch it2 — R. Judah and R. Nehemiah [differ in their interpretation]; one said that the word means 'her hand' and the other said that it means 'her handmaid'. He who said that it means 'her hand' did so because it is written ammathah;7 he who said that it means 'her handmaid' did so because the text has not yadah [her hand]. But according to him who said that it means 'her handmaid', it has just been stated that Gabriel came and beat them to the ground!8 — He left her one, because it is not customary for a king's daughter to be unattended. But according to him who said that it means 'her hand', the text should have been yadah! — It teaches us that [her arm] became lengthened; for a master has said: You find it so9 with the arm of Pharaoh's daughter and similarly with the teeth of the wicked, as it is written: Thou hast broken [shibbarta] the teeth of the wicked,10 and Resh Lakish said: Read not shibbarta but shirbabta [thou has lengthened].11
(Shemot Rabba 1:23 is more or less the same.)
Before we get into the back-and-forth analysis by the setama degemara and the harmonization with other midrashim, we see that Rabbi Yehudah and Rabbi Nechemiah had this dispute as to the meaning of the word. They don’t discuss the nikkud of the word -- not even the gemara discusses the implication of the nikkud or dikduk arising from that nikkud. It is, rather, the implication of the word choice.
There is a possibility of arguing with the nikkud. Shadal, in his Vikuach al Chochmat Hakabbalah, points to the Rashi on this pasuk as determining peshat based on nikkud:
“For instance, Rashi za”l on the pasuk vatishlach et amata pushes off the Midrash of our Sages because of nikkud…”
And this is a cause for determining the age of trup and nekudot. Shadal agues that Chazal lacked the written signs (orthography) for nekudot, and in some places were therefore not fixed and were ambiguous. He offers various proofs to this end. For instance, no gemara talks about the written signs, and there are several places (he describes) where one would very much expect them to come up. If the trup and nekodot are post-Chazal, rather than e.g. halacha leMoshe miSinai, then one can (safely, theologically speaking) argue with them. After all, they just represent the opinion of the author of the trup / nekudot, rather than the definitive meaning as given by Hashem to Moshe.
The other day, I saw something surprising in Michlal Yofi (authored by Rabbi Shlomo Ibn Melech, Fez, Constantinople, 1548):
“וַתִּשְׁלַח -- a language of sending [an agent].
אֲמָתָהּ -- her maidservant. And some say that this word is divided in its pronunciation, for there are some who read it with a dagesh, as אַמָּתָהּ, from the [language of] אַמָּה אָרְכּוֹ [which has a dagesh]. And so does Rabbenu Saadia explain it in his [Judeo-]Arabic commentary, ומרת דרעהא [which is more or less what was about, with a “fa” in פמרת rather than a “va” in ומרת and דרעהא as “arm” rather than ידהא as “hand”] -- which would mean “and she sent forth her arm”. And according to this explanation, וַתִּשְׁלַח would be a language of extending.”
According to this, then, there is an actual dispute as to how the word is to be read, and the peshat would go along with it.
Having seen this, I looked again at Minchat Shai and saw that he said the same thing about Saadia Gaon having a different girsa in the reading of the pasuk:
He cites the midrash in Sotah and in Shemot Rabba, that it either means hand or maidservant. And then writes:
“And according to the one who darshens it as a language of hand, the aleph needs to be with a [full] patach and the mem with a dagesh [אַמָּתָהּ], like the girsa of Rabbenu Saadia, as well as in the Targum. And according to the one who darshens is as a language of maidservant, the aleph is with a chataf patach and the mem is weak [without a dagesh chazak], as it is in our sefarim. And so wrote Rashi, Ibn Ezra, and Chizkuni. And see Mizrachi and Shorashim.”
Thus, Minchat Shai sees this as a dispute in the Masorah, as to what nekudot should appear in the Masoretic text. We have our tradition and apparently Saadia Gaon and Onkelos had a different tradition.
Are these two pronunciations really that distinct? Well, some who lain are quite medakdek in pronouncing and putting particular stress to those degeishim. For instance, listen to Rabbi Jeremy Weider, at the 1 minute 38 second mark, lain this pasuk:
וַתֵּרֶד בַּת פַּרְעֹה לִרְחֹץ עַל הַיְאֹר וְנַעֲרֹתֶיהָ הֹלְכֹת עַל יַד הַיְאֹר וַתֵּרֶא אֶת הַתֵּבָה בְּתוֹךְ הַסּוּף וַתִּשְׁלַח אֶת אֲמָתָהּ וַתִּקָּחֶהָ:
vaTTeired bas-Par’oh lirchotz al-hay’or, vena’aroseha holechos al yad hay’or, vaTTeire es haTTeiva besoch haSSuf, vaTTishlach es-amasah vaTTiKacheha [should be vaTTiKKacheha].
This overstress and over-enunciation strikes me as a bit unnatural and ridiculous, even though many baalei kriyah do this.
Yes, it is true that the dagesh chazak represents the gemination -- that is, doubling -- of the letter. But this gemination is a natural feature of spoken language, while the deliberate overstressing of the dagesh feels anything but.
To digress for a bit, consider the word וַתֵּרֶד. The gemination of the letter Tav there means that the Tav works both to close the preceding syllable and open the next syllable. In Hebrew, short vowels (such as segol, chirik chaser and, in this case, patach) in unstressed syllables need to be closed. CVC, consonant vowel consonant. So, VAT is a closed syllable. Long vowels can be in open syllables. CV, that is, consonant vowel. So, TEI is an open syllable. Since we have geminated (doubled) the Tav, we have VaT-Tei.
The same analysis goes for the word ammatah. If it meant “hand”, then there would be a patach (a short vowel) and the syllable would need to be closed. By geminating the Mem, we get (recalling that aleph is a consonant) `aM-Ma-tah. If it meant “maidservant”, then the chataf patach under the aleph (which is a reduced kamatz) is a sheva na, and not something that needs to be part of a closed syllable. So with a chataf patach, we can have `a-Ma-tah, with no gemination of the Mem.
But this gemination might be something natural and slight, and something we all do in English. Think of the word “attack” or “attach”. Why it the letter t doubled? There are phonological rules at play, in which certain short vowels appear to cause a doubling of the following, where the long vowel will not. Compare platter vs. plater, matter vs. mater, happen, sadder, gladder, adder. And this geminated consonant might well be closing the first syllable and opening the second one, even though the distinction is quite slight. For instance, consider these words with short vowels, saying each a few times: happen [hap-pen] and rabbit [rab-bit] vs. rabid [ra-pid]. You might expend a very slight amount of time and energy on the doubled letter. But it is nothing at all light the gemination you hear baalei keria (like Rabbi Weider above) perform. In modern Hebrew as well, according to the rules of Hebrew grammer (and phonology and morphology), those degeishim are present. But I’ve never heard anyone speak like this, in normal, correct speech. I have doubts somewhat doubt that, throughout the generations, when people spoke Hebrew, they regularly doubled their consonants in a manner that feels so unnatural.
However, it could well be that this impression is due to my own deficiency and English-language bias. There are many languages out there which have gemination. See this Wikipedia article:
Consonant length is distinctive in some languages, for instance Arabic, Berber, Catalan, Danish, Estonian, Finnish, Classical Hebrew, Hungarian, Italian, Japanese, Latin, and Tamil. Most languages (including English) do not have distinctive long consonants. Vowel length is distinctive in more languages than consonant length, although several languages feature both independently (as in Arabic, Japanese, Finnish, and Estonian), or have interdependent vowel and consonant length (as in Norwegian and Swedish).
So, since I am used to English, and lack familiarity with those languages which have distinctive consonant length, it sounds weird to me.
If the gemination of consonants is really something very slight, then the difference between amatah and ammatah would similarly be very slight. And, absent written signs for this gemination (via dagesh) and patach vs. chataf patach (invented post-Chazal), one could see how this was up in the air, and how Chazal could see both as legitimate possibilities. And the Masoretes wrote down what they heard, when they heard it, in one of those two ways.