(Mishnah): The pasuk states (Devarim 17:6), "according to the words of two witnesses or three witnesses the dead man will die" - if the testimony is valid with two, why did the pasuk mention three?

To compare three with two - just as three can prove two to be false, so two can prove three to be false.

From where do we know that two could even prove a group of 100 witnesses to be false? The pasuk says '(or three) witnesses' to teach that even 100 witnesses are treated like two.

R. Shimon: Just as two (false witnesses) are only killed if both are proved to be false, so too three are only killed when they are all proved to be false.

And from where do we know that R. Shimon's ruling would apply even to a group of 100 false witnesses? The pasuk says 'witnesses'.

R. Akiva: The pasuk mentioning a third witness comes as a stringency to include him in the punishment of the first two witnesses if they are all proved to be false.

If Hash-m punished a person who was secondary to the sin (since the testimony of the two would have been accepted even without him), like the sinner himself; all the more so, Hash-m will reward a person who was secondary to a mitzvah as if he did the mitzvah himself.

The Torah compares three witnesses to two - just as if there were two and one was found to be a relative of the accused or of the other witness or disqualified from testimony (for himself being a sinner), the entire testimony is therefore cancelled; so too where there are three witnesses and one was found to be a relative or disqualified, that the entire testimony is cancelled.

From where do we know that this applies even if 100 witnesses testified together? The pasuk says, 'witnesses'.

(R. Yosi): When is this true? In capital cases, but in monetary cases, the testimony continues with the others.

Rebbi: This applies both in capital and in monetary cases, when all of the witnesses warned the offender, but if not (meaning that of the entire group, only two valid witnesses gave the warning, showing that only they intended to be able to testify, but the others merely happened to see the offence), what can two brothers do if they happened to also witness a person committing murder?!

Question: (About both capital and monetary cases, the pesukim mentioned two and three witnesses) - let the Torah only say it about monetary cases and capital cases could be learned from it?

Answer: As I would say that if in monetary cases, which are lenient, three witnesses can disprove two witnesses, but two cannot disprove three.

Question: From where do we know that two can even disprove 100 witnesses? The pasuk says "witnesses" - meaning even 100.

(And why couldn't the Torah only say that two can disprove three about capital cases but not about monetary cases?) If so, I would have said that in capital cases, which are more stringent, two can disprove three, in order to prevent any possibility of false testimony existing; therefore, the pasuk was needed to teach this even for monetary cases.

Question: And from where do we know that two can even disprove 100 witnesses? The pasuk says "witnesses" - meaning even 100.

(Mishnah): If two witnesses saw him from one window and two from another window and one in between warned him (and it was heard by the witnesses)...

If at least one of each pair saw each other, it is considered to be one testimony; if not, it is two testimonies.

Therefore, if one pair is found to be false, the accused murderer and the false witnesses are killed and the other pair (that were never proved false) are exempt.

(Gemara - R. Yirmiyah): We can see from the wording of the Mishnah that it applies when there are two witnesses in each window, and one witness from one of the pairs saw the other witnesses; however, if there were three witnesses in the window and one of them saw the witnesses in the other window, since there is valid testimony in the first window even without him, they are not drawn after them to become part of the testimony with the pair in the other window.

R. Yosi: There is an explicit Baraisa that supports this - that if there are three witnesses in each window and some of them see each other, they do not join with the witnesses of the other window (since they themselves are an independent testimony).

The Mishnah taught that when none of the witnesses saw each other at all, they are considered two testimonies. This teaches something that was not taught anywhere in Maseches Sanhedrin, that even if Ploni (the accused) did actually murder (like the testimony of the second pair), since the first pair were proven to be false, the accused and the first pair are both put to death, whereas the second pair are acquitted.

(Mishnah - R. Yosi bei R. Yehuda): (Disagreeing with the previous Mishnah) Really, the accused is not killed unless the accusing witnesses themselves gave the warning (rather than a single unrelated witness), as the pasuk states (Devarim 17:6), "through the mouth of two witnesses".

Alternatively, "through the mouth of two witnesses" teaches that the Sanhedrin may not receive the testimony through a translator.

(Gemara - R. Hoshiya - Beraisa): This pasuk refers to a case of two pairs of witnesses (Devarim 19: 16), "When a false witness will arise against a man" - that the original witnesses testified about a man (that he did a certain sin)... "to give perverted testimony against him" - this refers a second pair disproving the first pair "against him", meaning against the witnesses themselves rather than against their testimony. (This teaches that the witnesses can only become false if the second pair said that they were with them elsewhere at the time of the crime; but to say that the accused was with them does not disprove the first witnesses.)

Question: Do the false witnesses require a warning in order to convict them?

Answer (R. Yitzchak bar Tavlai citing R. Elazar): No, they don't.

Question (R. Abahu): Why not?

Answer (R. Yaakov bar Dasi): There are plenty of degenerate people who see an innocent person being taken out to be killed and they know that they gave false testimony about him, but they do not say anything. Do you think that they would be affected by warning them before they testify?!

If a person wrote over all of his property to two people on the same gift document, if there were witnesses who were valid for one recipient but invalid for the other recipient (such as if they were related to him)...

(R. Hila): There is a dispute between R. Yochanan and R. Shimon ben Lakish about whether their testimony is nevertheless valid for the first one - one said that once they are invalid for one, they are invalid for the other; the other opinion said that even if they are invalid for one, they are still valid for the other.

R. Mana did not explain this dispute;

R. Avun explained (detailing who says what...) that R. Yochanan said that once there is testimony written on the document that is invalid for one litigant, even if those witnesses are valid for the other litigant, the document is invalid for everyone. R. Shimon ben Lakish said that the document is valid for the litigant to whom those witnesses are valid.

(R. Elazar): There is a support for R. Yochanan in the following Baraisa - (The pasuk in Devarim 17:6 juxtaposes two witnesses with three witnesses) Just as in the case of two witnesses, if one of them was found to be a relative of the litigant or found to be disqualified as a witness, the entire testimony is invalid; so too for three witnesses, even if only one witness was disqualified, the entire testimony is invalid (even though a pair of valid witnesses remain).

Question: From where do we know that this would apply even to a group of one hundred witnesses? The pasuk says, "witnesses" - which implies even multiple.

(R. Yaakov bar Acha): There is a dispute between R. Chanina, friend of the Rabbis, and the Rabbis - one said that R. Elazar spoke correctly (in giving his proof to R. Yochanan) and the other said that he did not.

The one who said that R. Elazar is correct is himself correct, as R. Elazar said that it is like one testimony against one man, and anytime a testimony is partially cancelled, it is completely cancelled.

The one who said that R. Elazar is not correct is because the proof is unlike the Beraisa - even in a case of two gifts on one document, it becomes like two separate sets of witnesses who are valid for this one and invalid for that one.