Talmud - Pesahim 12a פסחים י״ב א׳

and one of them errs by an hour and a bit.

Rav Huna, son of Rav Yehuda, went and said the halakha of Abaye before Rava. Rava said to him that Abaye’s explanation must be rejected: And were we to closely examine the statements of these witnesses and ask them when precisely the incident occurred, and find that the one who says at three hours means that it occurred at the beginning of the third hour, and the one who says at five hours means that it occurred at the end of the fifth hour, it would be contradictory testimony and we would not kill the accused on the basis of this testimony; and will we arise and kill based on uncertainty? Although their testimony could be valid, it could also be void. Can the court execute the accused based on that uncertainty? And consider that the Merciful One says in the Torah: “And the congregation shall judge…and the congregation shall deliver” (Numbers 35:24–25), from which it is derived that judges must do everything in their power to save an accused from the death penalty.

Rather, Rava said: In fact, according to the statement of Rabbi Meir, a person errs by two hours less a bit; according to the statement of Rabbi Yehuda, a person errs by three hours less a bit. The Gemara elaborates: According to the statement of Rabbi Meir, a person errs by two hours less a bit. How so? When the incident occurred, it was either at the beginning of the second hour or at the end of the third hour and one of the witnesses erred by two hours less a bit, as that is a reasonable discrepancy.

According to the statement of Rabbi Yehuda, a person errs by three hours less a bit. How so? When the incident occurred, it was either in the beginning of the third hour or at the end of the fifth hour, and one of them errs by three hours less a bit. According to Rabbi Yehuda, that is a reasonable discrepancy.

We learned in a mishna: The judges of the court would examine witnesses with seven forms of interrogation: In which seven-year Sabbatical cycle of the Jubilee did the incident occur, in which year of the seven-year Sabbatical cycle, in which month of that year, on which date of that month, on which day of the week, at which hour of the day, at which place did the incident occur? And it was taught in a baraita: What is the difference between interrogations and examinations? With regard to each of the forms of interrogation, if one of the witnesses said: I do not know, their testimony is void. However, with regard to examinations, even if they both say: We do not know, their testimony is valid.

The Gemara continues. And we discussed this baraita: In what way are interrogations different and in what way are examinations different? Why is the testimony of a person who cannot answer one of the interrogations void, but his testimony is valid if he cannot answer one of the examinations?

And we said in response: With regard to interrogations, if one of them says: I do not know, their testimony is void, as this is testimony that you cannot render as false, conspiring testimony. Witnesses are determined to be false, conspiring witnesses when other witnesses arrive in court and testify that the first pair of witnesses could not possibly have witnessed the incident in question, as they were at a different place at that time together with them. If a witness says he does not know when the incident occurred or where it occurred, it would be impossible to testify that he was elsewhere and render him a false, conspiring witness. The Torah disqualifies the testimony of a witness who cannot be rendered a false, conspiring witness.

However, with regard to examinations, even if the witness cannot remember every detail, it remains testimony that you can render as false, conspiring testimony. Even without that detail, the second pair of witnesses can still testify that the first pair of witnesses was with them in a different place at that time.

Based on the above, the Gemara returns to discuss the matter of discrepancies in the testimony of the witnesses with regard to the hour that the incident occurred. And if you say that a person errs by that much, e.g., several hours, with regard to the interrogations of what time of day the incident occurred, this is also testimony that you cannot render as false, conspiring testimony, for if other witnesses claim that they were elsewhere at that time, the witnesses can say: We are mistaken, and the incident did not actually occur at that time, but a few hours earlier or later. Consequently, it is impossible to prove that they are false, conspiring witnesses.

The Gemara explains that we give leeway to the first pair of witnesses for the entire duration of the potential margin of their error. In order to render them false, conspiring witnesses, the second pair of witnesses must testify that the first witnesses were with them for the entire period during which their testimony would be accepted, despite the discrepancy with regard to the hour in which it occurred. According to Rabbi Meir, we give the witnesses who testified that the incident occurred at two or three hours of the day, from the beginning of the first hour until the end of the fifth hour. If they testified that the incident occurred in the second hour or the third hour, the court must take into account a mistake of two hours in either direction.

And by right we should give them even more leeway at the beginning. Since one errs by two hours less a bit, if the witnesses testified that the incident occurred at approximately two hours of the day, they might have erred by roughly two hours and the incident actually occurred before sunrise. However, since between day and night people do not err, and as the witnesses testify that the incident occurred during the second or third hour of the day, it could not have occurred before sunrise.

And according to Rabbi Yehuda, we give the witnesses a margin of error from the beginning of the first hour until the end of the sixth hour. And by right we should give them even more leeway at the beginning, as according to his opinion one errs by three hours less a bit, and therefore the margin for error should be greater.

Traduction : The William Davidson digital edition of the Koren Noé Talmud, with commentary by Rabbi Adin Even-Israel Steinsaltz.

Pesachim 12a explores how much witness testimony discrepancy is acceptable in capital cases, weighing Rabbi Meir and Rabbi Yehuda's views.