2) TWO WITNESSES WHO SAY THAT THEY DO NOT KNOW
QUESTIONS: The Mishnah (40a) teaches that if one witness responds to a question of the Chakiros by saying, "I do not know," the testimony is invalid. In contrast, if one witness responds to a question of the Bedikos by saying, "I do not know," and even if two say, "We do not know," the testimony is accepted.

The Gemara asks what the Mishnah means to teach by saying that "even if two say, 'We do not know.'" What difference does it make if one of the two witnesses does not know or if both of the witnesses do not know? In either case, Beis Din does not have complete testimony, and yet their testimony is valid.

Rav Sheshes proposes that the Mishnah means that even when there are three witnesses and two of them say that they do know in response to a question of the Chakiros and the third witness says that he does not know, the entire testimony is not valid.

Rava rejects this proposal, because the Mishnah says that the testimony is valid when two witnesses say that they do not know. How can we change the words of the Mishnah such that it says the opposite?

Rava therefore explains the words of the Mishnah to mean that when there are three witnesses and two of them say that they do know and the third witness says that he does not know, with regard to the Chakiros the entire testimony is valid.

The only difference between the answers of Rava and Rav Sheshes is that Rava keeps the world "valid" ("Kayemes") and does not change it to "invalid" ("Beteilah"). However, Rava himself changes the words of the Mishnah, as Rav Sheshes does! The Mishnah says that two witnesses say, "We do not know," and Rava changes those words to read, "We do know." Why does the Gemara not challenge Rava with the same question that Rava asks on Rav Sheshes? How can Rava change the Mishnah to say "we do know" when the Mishnah says that they say "we do not know"? (CHIDUSHEI HA'RAN)

ANSWERS:
(a) The CHIDUSHEI HA'RAN and CHAMRA V'CHAYEI explain that it is more likely that the correct text of the Mishnah is that the testimony is valid, as is written, than that the statement of the two witnesses was that they do not know. The reason for this is that the Mishnah uses one word ("Kayemes") to teach the Halachah for two cases -- the case in which a witness says "I do not know" with regard to the Bedikos, and the case in which two witnesses say that they do not know (or that they do know) with regard to the Chakiros. In the former case, the testimony is valid. Therefore, the Mishnah's ruling must have been that the testimony is also valid when the two witnesses say that they do (or do not) know with regard to the Chakiros.

(b) The ARUCH LA'NER cites RASHI earlier (10b, DH v'Ha Ibur) who writes that it is common for a Tana to inadvertently omit a word that he forgot, but the Tana does not change one word into an entirely different word. Similarly, the Tana here might have added the word "Ein Anu Yod'in" inadvertently, but the Tana would not confuse the word "Kayemes" with the word "Beteilah."

(c) Another answer might be as follows. How can Rav Sheshes and Rava suggest that the words, "even if both witnesses say, 'We do [not] know'," are referring to the witnesses' response to a question of the Chakiros? The Mishnah at this point is clearly discussing the Bedikos!

Perhaps Rava actually means to suggest that the Mishnah should be read as it appears in our texts, and that these words indeed refer to the witnesses' reply to the Bedikos and not to the Chakiros. The Halachah that Rava teaches is the inference from the words of the Mishnah and is not explicitly stated in the Mishnah.

According to Rava's reading, the Mishnah refers to a single case in which three witnesses come to testify and are asked the questions of the Bedikos. The Mishnah teaches that if one of the three says, "I do not know," then even if the other two also say that they do not know, the testimony is valid. (That is, the word "va'Afilu" in the Mishnah should be read as "Afilu," without the Vav.) This phraseology of the Mishnah implies that with regard to the Chakiros, the Halachah is different: if the third witness says -- in response to a question of the Chakiros -- that he does not know, the testimony will be valid only if the other two witnesses say that they do know. This is Rava's intention.

According to this explanation, Rava's emendation of the Mishnah is insignificant (it merely omits the Vav of "va'Afilu"). This also seems to be the intention of the KOS HA'YESHU'OS.

(d) Consistent with the third answer above, one may suggest further that Rava is not changing even a single letter in the Mishnah. Rather, Rava's intention is to explain why the Mishnah considers it a Chidush to say that when two witnesses say that they do not know, the testimony is valid. Rava answers that the Mishnah is discussing a situation in which there are three witnesses. Accordingly, the Mishnah is teaching an additional Chidush by saying that two witnesses say they do not know. When a single witness says "I do not know," Beis Din still has the valid testimony of the other two witnesses. The Mishnah adds that even when two of the three witnesses say "we do not know," leaving only one witness who does know, the testimony remains valid.

However, this explanation is acceptable only if the Mishnah does not consider all of the witnesses to be giving a single testimony. If all of the witnesses must be valid in order to accept the testimony, then once the Mishnah teaches that the testimony is valid when one says "I do not know," we know already that the testimony is valid when two witnesses say that they do not know. Consequently, from the words "even if two say, 'We do not know'," it can be derived that when more than two witnesses testify together, it is necessary only to validate the testimony of two of them (and not all of them). This is what Rava means when he says that these words teach that when two witnesses answer the Chakiros and the third does not, the testimony is valid.