1) A SECOND SET OF WITNESSES WHO SAY THAT "THE OPPOSITE OCCURRED"
QUESTION: The Gemara attempts to prove from the Mishnah (92a) that a single witness is believed to exempt a woman from Yibum. The Mishnah states that if testimony was given that her husband died before her son died (and thus she is exempt from Yibum), and then she remarried, and afterwards two other witnesses state that the opposite occurred (her son died before her husband died, and thus she is obligated to do Yibum), she must leave both husbands.
The Gemara asks, who gave the initial testimony that the woman may remarry? It cannot be that two witnesses testified, because when the second pair of witnesses testify they should not be able to override the first pair (rather, they create a situation of "Trei u'Trei"). It must be that the initial testimony was given by a single witness, and when two witnesses come they are able to override his testimony.
The Gemara rejects this proof and asserts that the first testimony indeed was given by two witnesses. The reason why the second pair is able to override the testimony of the first pair is that the second pair of witnesses are Edei Hazamah who are believed to refute the testimony of (or, more accurately, disqualify) the first pair of witnesses, and thus the situation is not one of "Trei u'Trei."
The Gemara's assertion that the second pair of witnesses are Edei Hazamah is problematic. The Mishnah says that the second pair of witnesses testified that "the opposite occurred" (the son died before the father died). This is not the testimony of Edei Hazamah. Edei Hazamah disqualify the first pair of witnesses by testifying that the first pair could not have seen the event about which they testify because they were not where they claimed to have been when they saw the event in question. The second witnesses do not testify that "the opposite occurred" but merely that the first witnesses could not have seen the event.
The Mishnah cannot mean that the second pair of witnesses said two things -- that the first pair was with them at the time the events occurred, and that the opposite occurred. Such testimony would not be able to make the other pair of witnesses Edim Zomemim. (See Makos 5a; RAMBAM, Hilchos Edus 18:2, and Lechem Mishneh.) (ARUCH LA'NER)
The Mishnah also cannot mean that there are two sets of witnesses which counter the testimony of the first pair, one which testifies that the first witnesses are Edim Zomemim, and one which testifies that the opposite occurred. The words of Rashi and the Rishonim clearly imply that in the Mishnah's case only one other set of witnesses came.
ANSWER: Perhaps when the Mishnah says that the second pair of witnesses testify that "the opposite occurred," it does not mean that those witnesses testify that they saw the opposite actually occur, but rather they say that the opposite might have occurred (because the first witnesses were with them in a different place and could not have the seen the event).
Why, though, should the Mishnah emphasize that the "opposite" might have occurred (the son might have died before the father)? If the second pair of witnesses testified that the first pair are Edim Zomemim, there is a far greater concern -- the husband is still alive!
The answer is that the Mishnah means that the second pair of witnesses made the first into Edim Zomemim only with regard to the first pair's testimony about the timing of the death of the son. The husband certainly is dead, as the first pair testified, and the only point of doubt that remains is whether the son died before or after the father died. The Mishnah uses this case to demonstrate that if the son indeed died first, the children which the woman subsequently bears from another husband are Mamzerim, in accordance with the view of Rebbi Akiva who says that the children of a Yevamah la'Shuk are Mamzerim. (M. Kornfeld)
2) MARRYING ONE'S MOTHER-IN-LAW AFTER THE DEATH OF ONE'S WIFE
OPINIONS: The Gemara cites a dispute between Rebbi Yishmael and Rebbi Akiva with regard to whether a man is punished for living with his wife's mother (Chamoso) after his wife has died. (This is according to Rava's understanding of the dispute.) The Gemara concludes that Rebbi Akiva, who maintains that one is punished with Sereifah only when he lives with his mother-in-law while his wife is alive, agrees that it nevertheless is forbidden even after the wife's death.
What prohibition remains after the wife's death? The verse in Parshas Kedoshim (Vayikra 20:14) which imposes the death penalty for one who marries a woman and her mother applies only when the wife is alive (according to Rebbi Akiva). What is the source to prohibit it even after the wife's death?
(a) TOSFOS (DH me'Isura, and 98b, DH Kalash), the RAMBAM (Hilchos Isurei Bi'ah 2:5), and the RA'AVAD (quoted by the ME'IRI in Sanhedrin 76b) write that the prohibition is an Isur of Kares and a Lo Ta'aseh, as derived from the verse in Parshas Acharei Mos (Vayikra 18:17) which imposes the punishment of Kares and a Lav for the act and which does not limit the Isur of Kares and Lav to when both women are alive.
(b) RASHI in Sanhedrin (76b, DH Rebbi Akiva), however, writes that according to Rebbi Akiva no Isur Kares or Lo Ta'aseh prohibits one from marrying his wife's mother after the death of his wife. His wife's mother is only "Asur b'Arur" to him -- the Torah places a curse on a man who marries his mother-in-law (see Devarim 27:23); no explicit Torah prohibition forbids it.
Support for the view of Rashi may be derived from the wording of the Gemara in Sanhedrin (as the MAHARSHAL here points out), which says that one who marries his mother-in-law after the death of his wife transgresses an "Isura b'Alma" -- only a "minor prohibition." Additional support may be adduced from the Gemara later in Yevamos (98b), as TOSFOS there (DH Kalash) writes.
Rashi's opinion is shared by the RASHBA and RAMBAN (98b) who write that one who marries his mother-in-law after the death of his wife transgresses only an Isur Arur.
1. The ARUCH LA'NER here asks that there actually are six relatives who are forbidden by the Isur of "Ishah u'Vitah": one's mother-in-law, her mother, and her mother-in-law; one's wife's daughter, her son's daughter, and her daughter's daughter. The verse of "Arur," however, refers to only one of those relatives -- one's mother-in-law. Consequently, there should be no prohibition or Arur at all against having relations with one's wife's daughter or granddaughter, or with one's wife's mother's mother or mother-in-law, after the death of one's wife!
If it is true that one's wife's mother's mother is permitted to him l'Chatchilah after the death of his wife, then the question of the Gemara returns: why does Rebbi Akiva not say that another Ervah needs a Get? In the case of a man who married a woman (with Kidushin) and then heard that she died, and he went and married her mother's mother (who was permitted to him l'Chatchilah since his wife was presumed to be dead), if his wife then returned alive he should be required to give a Get to her mother's mother. The Gemara has not answered its question that there is an additional Ervah which Rebbi Akiva could have included in his list.
The RAMBAN and RASHBA (98b) answer that the verse which teaches the Isur of Arur includes not only one's mother-in-law, but also her mother and her mother-in-law. (See Aruch la'Ner who suggests a different answer.)
2. However, this answer addresses only why Rebbi Akiva does not include one's wife's grandmother in his list of Arayos who need a Get if one marries them by accident under the presumption that his wife is dead. However, why does he not mention the Ervah of one's wife's daughter or granddaughter? When one presumes that his wife is dead and he marries his wife's daughter or granddaughter, he should be required to give her a Get since those relatives are permitted to him after his wife's death and the verse of Arur does not apply to them.
The RAMBAN and RASHBA (ibid.) explain that Rebbi Akiva's opinion that the prohibition of "Ishah u'Vitah" is not in force after the death of one's wife applies only in the case of one's wife's mother or grandmother (the relatives discussed in the verse, "Oso v'Es'hen," Vayikra 20:14.) One's wife's daughter or granddaughter, in contrast, are not permitted after the death of one's wife.
However, the RAMBAM (Hilchos Isurei Bi'ah 2:7-8) understands that Rebbi Akiva refers to any of the Arayos in the two generations above or below one's wife, and not just to those in the two generations above his wife (i.e. his wife's mother and grandmother). According to the Rambam, in all of these cases Rebbi Akiva maintains that no Isur Sereifah applies.
Rashi also appears to agree with the Rambam in this regard. Rashi earlier (2a and 9b; see Insights to Yevamos 9:3) implies that one's wife's daughter or granddaughter becomes permitted to him after his wife's death. According to Rashi, since the Torah does not write an Arur in the case of one's wife's daughter or granddaughter, he should be permitted to marry her l'Chatchilah!
Apparently, Rashi indeed understands that the Arur that applies to one who lives with his wife's close relatives applies to all of the Arayos included in "Es Ishah v'Es Imah." Just as Rashi understands that the words "Es Ishah v'Es Imah" (from which Rebbi Akiva derives that one's mother-in-law is permitted after the death of his wife) refer either to the wife and her mother or grandmother, or to the wife and her daughter or granddaughter, so, too, when the Torah writes "Chosanto" (his mother-in-law) in the verse of the Arur it includes all of those included in "Es Ishah v'Es Imah."
(Although Rashi in Sanhedrin writes that the "Isur" mentioned by Rebbi Akiva here means only an "Arur," the words of Rashi as quoted by the TOSFOS YESHANIM here and the RITVA (98b) state that after the death of one's wife, his mother-in-law is prohibited with Kares and a Lav, but not with Sereifah. They apparently infer this from the words of Rashi later (98b, DH Kalash), where Rashi writes that after the death of his wife the man is "not punished with Sereifah" for living with his mother-in-law. This implies that an Isur Kares and a Lav do remain. However, the words of Rashi in our texts, both in Sanhedrin and in Yevamos, state otherwise and prohibit one's mother-in-law and one's wife's daughter after the death of his wife with only an Isur of Arur, but not with an Isur of Kares and a Lav.)