How could the Chachamim be lenient and permit her to marry based on the
testimony of a single witness, when, according to the Torah, she is an Eshes
Ish? How can the Chachamim override the Isur d'Oraisa of Eshes Ish?
(Although the Gemara later (89a-90b) says that there are times when the
Rabanan do override Torah law, we do not find that they have the power to do
so "b'Kum va'Aseh," actively, but rather "b'Shev v'Al Ta'aseh," passively.
In the case under discussion, they allowed the woman to go and marry another
person, in what would appear to be an active violation of Torah law.)
(a) TOSFOS (DH Mitoch) explains that the Chachamim are authorized to
override an Isur in the Torah, but they may not *uproot* an Isur in the
Torah (unless it is done b'Shev v'Al Ta'aseh, as the Gemara says later on
90b).
The difference between "overriding" and "uprooting" an Isur d'Oraisa is that
to "uproot" an Isur means to do something that everyone can see is
contradictory to the Torah. In contrast, when the Chachamim are doing
something that seems to have some logical (if not scriptural) basis and does
not *appear* to be contradicting something in the Torah, then it is
considered to be "overriding" an Isur and not uprooting it. In the case of
our Gemara, there is good reason ("Ketzas Ta'am u'Semach," as Tosfos writes)
to assume that the woman's husband is dead, since a woman always
investigates the matter before remarrying. Thus, it is not considered
"uprooting" an Isur in the Torah, but rather "overriding" it.
It seems that according to Tosfos, the Gemara accepts that this is one of
the powers with which the Torah invests the Chachamim; the Torah gives the
Chachamim the authority to do as they deem appropriate, as long as they do
not "uproot" something written in the Torah. When they are relying on sound
judgment and reason, even though the Torah does not give Halachic
recognition to such reason, it is not called "uprooting." Since the Torah
does not openly say *not* to accept such reasoning (and thus, issuing such a
ruling will not inadvertently cause people to disrespect the Torah or the
Rabanan), Chazal have the leeway to institute what they see as necessary.
(b) RASHI in Shabbos (145b, cited by REBBI AKIVA EIGER in Gilyon ha'Shas)
writes that in this situation, the Chachamim uprooted the Kidushin
retroactively (a principle the Gemara later, 90b, calls "Afke'inhu") in
order to permit the woman to remarry. Thus, she was retroactively never
married to her original husband.
The RITVA and Rishonim ask that if she was never married to the first man,
then her marriage to the second man is a full-fledged marriage mid'Oraisa.
However, from the Gemara it is clear that the *first* marriage is considered
the real one (91a), and not the second!
The Rishonim (ME'IRI, TESHUVOS HA'RASHBA 1:1162) answer that according to
Rashi, the Chachamim uprooted the first Kidushin *on condition* that the
husband does not come back. If he eventually returns, then they never
uprooted the original Kidushin. (The reason the woman may remarry meanwhile,
even though there exists the possibility that her first husband will return
alive, is because there is a Chazakah that since right now he is not here,
he will not return; see Tosfos Gitin 33a DH Afke'inhu.)
(c) The RITVA cites the RE'AH who says that the testimony of the single
witness is not accepted as "testimony," but rather as a "Giluy Milsa b'Alma"
that she is no longer married. That is, the Torah does not always require
the formal testimony of two witnesses to prove facts. Rather, any clear
circumstantial proof ("Anan Sehadi" -- "*we* bear testimony") that she is
not married suffices. The Torah left it to the Chachamim to decide what
proof is considered clear enough to permit her to remarry.
In the case of our Mishnah, the fact that a single witness testified to the
death of the husband is considered to be very strong proof that the husband
is dead, because a person would not lie in court about an indisputable fact
when his lie will eventually become known to all (through the return of the
supposedly dead husband -- the Gemara introduces this logic later in the
Sugya, 93b). Additional proof is afforded by the fact that a woman would not
rely on the testimony of a single witness to get married unless she herself
has investigated the matter to her satisfaction. Since the Rabanan decided
that there is sufficient circumstantial proof of the death of her husband,
the woman may get married mid'Oraisa and not just mid'Rabanan. (See also
TOSFOS YESHANIM.)
This might be the intention of the RAMBAM as well. In several places (see
Hilchos Edus 5:2; Hilchos Yibum 3:11) the Rambam writes that the
trustworthiness of a single witness is "mi'Divreihem" (mid'Rabanan).
However, in the end of Hilchos Gerushin (13:29) the Rambam comments that we
should not wonder why the Chachamim permitted an Ervah to marry based on the
testimony of a single witness (or a woman), because the Torah only requires
formal testimony for an event that can *only* be known through witnesses
(for example, that a person transgressed a specific prohibition, or that one
person lent another person money). However, if a fact is in question, but it
will eventually become known on its own (in our case, such as the husband's
return), the Torah does not require formal testimony, since it is very
uncommon for a single witness to lie about such a matter. The Rambam
concludes, "Therefore, the Chachamim were lenient and believed a single
witness... in order that women should not remain Agunos."
The Rambam's words at the end of Hilchos Gerushin are unclear. He begins and
ends his comments by saying that the single witness is believed
mid'Rabanan -- like he seems to write elsewhere -- while in between he says
that the witness is believed mid'Oraisa!
The SHEV SHEMAITSA (7:11) explains that even though the witness is believed
mid'Oraisa as the Rambam says, the Chachamim are usually very stringent with
Isurim such as that of Eshes Ish and they require two witnesses. Here, also,
they should have been stringent and decreed that an Eshes Ish may not
remarry on such questionable testimony, and they should have required two
witnesses, mid'Rabanan. The reason they did not require this here was in
order to prevent women from becoming Agunos (like the Ritva wrote).
This is what the Rambam means when he says that the Chachamim were lenient,
while at the same time the witness is believed mid'Oraisa. This is also why
the Rambam writes in Hilchos Yibum that a single witness is believed only
"Mishum Iguna." When the Rambam (in Hilchos Edus) says that a single witness
is believed "mi'Divreihem," he means that it is not explicitly written in
the Torah that he is believed in this case, but rather the Chachamim derived
it from the verses. (The Rambam consistently refers to such Halchahos
mid'Oraisa as "mi'Divreihem.") The witness, through, is trusted mid'Oraisa.
(The Rambam, unlike the Ritva, does not mention the logic that a woman
thoroughly investigates the testimony of a witness, as our Gemara says,
apparently because it would seem from our Gemara that this logic is entirely
mid'Rabanan and does not affect the Halachah mid'Oraisa. The Ra'avad, ibid.,
notes -- like the Ritva -- that this logic also serves to lend credence to
the testimony of the witness.)
(d) The RAMACH in his comments on the R ambam (end of Hilchos Gerushin)
takes a different approach and differentiates between the trustworthiness of
a valid single witness (i.e. a Jewish male), and that of a single witness
who would not be valid as one of a pair of witnesses. In the former case,
the testimony is accepted to allow the woman to remarry mid'Oraisa (like the
Ritva said), while in the latter, the witness is trusted only mid'Rabanan
(like Tosfos).