Daf 18b: If one of them is found to be a relative or an invalid witness

  • Harav Mosheh Lichtenstein
Translated by R. Shalom Mashbaum
גיטין יח: "איתמר אמר לעשרה וכו'. רש"י ד"ה נמצא.
מכות ה:, המשנה השנייה. רי"ף שם (ב: מדפי האלפס) "וחזינן לגאון... שלא מדעת חברו". רמב"ם הל' עדות פ"ה ה"ה-ה"ו.
בבא בתרא קסב: "דתניא הרחיק את העדים... פוסל בארבעה". תוד"ה נמצא (עד סוף העמוד).
[חידושי רבנו חיים הלוי על הרמב"ם, הל' עדות פ"ה ה"ו, ד"ה ובעיקר]
Questions for preparation
1. Is it obvious that the law of “if one of them [the witnesses] is found to be a relative or an invalid witness, their testimony is cancelled] applies to shtarot [legal documents] as well? Try to answer this question while relating to the various possibilities regarding the nature of a shtar, and the testimony in it.
2. What is the impression we get from within the gemara in Bava Batra 162: in regards to this question? What is the opinion of the Rishonim on this question?
3. Is it possible to explain the disqualification of “if one of them is found to be a relative or disqualified” in the sugya in Gittin 18b in a way different from that of Rashi?
The sugya our shiur deals with (18b) discusses a case in which the husband requests that additional witnesses sign the get, in addition to the two who are required mi-ikkar hadin [as the essential law mandates].
“It was said: One who said to ten men ‘Write a get for my wife”, R. Yochanan said: “two are witnesses, and the rest are to fulfill the condition.” And Reish Lakish said: “All of them are witnesses.” … On what do they differ? They differ as to the law if two signed the get on that day, and the others in the course of ten days. The one who says [the additional people] sign to fulfill the condition, in this case the get is valid, and the one who says [the additional people] sign as witnesses, the get is pasul [invalid]. If at the outset one of the witnesses is a relative or an invalid witness, some say the get is valid, and some say the get is invalid. Some say it is valid – the additional people are signing to fulfill the condition [not as witnesses]. The one who says the get is invalid, because this case can be confused with the case of shtarot [legal documents] in general.
In R. Yochanan’s opinion, all the people who sign beyond the two who are absolutely necessary are just signing ‘to fulfill the condition,’ and only two witnesses are considered to be testifying regarding the get. Therefore, we don’t care if the others are invalid witnesses. In contrast, Reish Lakish is of the opinion that the get is invalid [if one of the signers is an invalid witness], because all those who signed the get are witnesses.
On the face of things, R. Yochanan’s opinion is straightforward and understandable – for a person can make any conditions he wants, including a condition that invalid witnesses sign the get – while Reish Lakish’s opinion requires explanation. Why should a get that relatives signed, in addition to the valid witnesses be invalid, for in fact there are two witnesses to validate the get;
what changed by the fact that invalid witnesses were added?
a. Rashi’s approach
The answer to our question is simple, and we can get to it immediately via the commentary of Rashi on our sugya. Relying on the famous mishna in Makkot 5b (which for some reason he cites in particular from a sugya in Sanhedrin), Rashi explains that a set of witnesses composed of more than two witnesses is disqualified if among its members is a relative of an invalid witness. The very existence of invalid witnesses in the framework of the set of witnesses damages the integrity of the set of witnesses, despite the fact that it has within it the number of valid witnesses required to make the testimony of the set valid.
The law Rashi cites sees the requirement of two witnesses as arising from their superior legal significance, and not as a result of the greater trustworthiness of two witnesses over one witness. Similarly, it is based on regarding the set of witnesses as one organic unit, and not merely a chance combination of individuals who happened to see something. Therefore, if the unit includes invalid witnesses, this disqualifies the entire unit.
This law is halakhically derived by R. Akiva in the cited mishna, and constituted on of the basic drashot of the laws of witnesses. Therefore, Rashi’s explanation of the invalidation of a shtar which was signed by relatives or invalid witnesses is an intuitively convincing explanation.
The nature of testimony in legal documents
However, things are not so simple. Indeed, the invalidity of a unit of witnesses which includes invalid witnesses is straightforward and clear regarding verbal testimony, but our sugya is dealing with testimony in a legal document, and not verbal testimony delivered in court, and it is not at all clear that testimony in a document is similar to verbal testimony. If so, we must examine the applicability of the rule of ‘if one of the witnesses is found to be a relative or an invalid witness [their testimony is invalidated] in regard to shtarot.
The root of the above dilemma is in understanding the mature and essence of a shtar; is it fundamentally testimony – indeed delivered in writing, and not verbally by the witnesses in court, but nevertheless testimony, or perhaps a shtar is not testimony but a separate legal institution, which acts not as testimony, but [independently] as a shtar?
We can say that there are two main schools of thought regarding this question, which already arose in the shiur on dating gittin. The first one, which expresses itself mainly in the viewpoint of the Baalei Tosfot[1], sees a shtar as testimony in all regards, and the processes connected to it as testimonial processes. The second school sees a shtar as a document given by one of the parties of a case to the other; fundamentally, it consists of an obligation the giver assumes in regard to the receiver.
Indeed, even according to this opinion a shtar is considered trustworthy evidence similar to that of testimony, but in its essence it does not belong to the world of testimony, but to the relationship between the two parties in the case. This perception is expressed in the concise definition of the Baal haMaor (Yevamot 31b), who ascribes a shtar to the writer of the shtar, and not the witnesses. An expression similar to this perception can be seen also in the words of the Ramban (Bava Batra 77a, at the end of s.v. chozer, regarding a shtar that “we need a book [i.e. a document] that effects a transaction, and the witnesses act as the agents of the concerned party.”
The nature of testimony in our sugya
In our matter this point is extremely essential. If a shtar is fundamentally testimony, then the laws of testimony will apply to it[2], including the rule of ‘if one of the witnesses was found to be a relative or invalid, their testimony is cancelled’; while if a shtar is not testimony, but rather an obligation of one party to another, it seems that there is less to apply the rule of ‘if one of the witnesses was found to be a relative or invalid...,’ since this rule is embedded in the creation of a set of witnesses with the authority of testimony. Therefore, we must examine Rashi’s explanation, which makes Reish Lakish’s rule dependent on R. Akiva’s derivation which invalidates relatives and invalid witnesses as part of an extended set of witnesses, in light of the opinions which exist regarding the nature of a shtar, and to accept or reject his explanation accordingly.
However, despite the fact that this seems logical, we must still examine the picture which arises from the sugyot. On the face of things Rashi’s opinion is inescapable, for our sugya says that if the people sign as witnesses, then the get is invalid if one of them is found to be a relative or an invalid witness, if so it is explicit that the psul [disqualification] of ‘if one of them is found to be a relative or an invalid witness’ applies to shtarot as well. In actuality, this is not merely Rashi’s opinion, but rather a straightforward and clear explanation of a rule stated explicitly in the sugya.
However, we are not compelled at all to explain the sugya thus, and we can easily explain it without recourse to the sugya in Makkot, for there is an essential difference between the two cases; in Makkot we are speaking of witnesses who come to testify in court, to whom additional people who are invalid as witnesses are joined, while in our case the husband explicitly demanded that ten people sign. The case in which invalid witnesses were added to the valid ones, when it is absolutely clear that the basic requirement of the biblical text is only for two valid witnesses, in which case we can invalidate the testimony only because the additional witnesses disqualify the entire set of witnesses, is not comparable to the case in which the demand was defined in advance as a demand for more witnesses.
In other words, the invalidation in our sugya is not necessarily because of the damage to the set as a set of witnesses but rather the simple fact that the number of witnesses that the husband demanded as a condition for writing the get is lacking. The husband demanded a get that ten witnesses appear upon, but in reality there are only nine valid signatures. Here we are not dealing with the principle “additions detract,” but rather here the required number of signatures is deficient. Just as a get with one witness is not valid – for a required witness is missing – so too a get with nine witnesses {when the husband demanded ten] is invalid, in Reish Lakish’s opinion, for it does not have the required ten witnesses.
Indeed, I did not find this explanation in the Rishonim in regards to our sugya, but it is possible that it was noted by one of the Rishonim in the course of his discussion of rabbinic (d’rabbanan) concerns. See Sefer HaIttur (ot kuf, kabbalat eidut, page 40 in the Rm”i edition), who assumes as something obvious that the rule “if one was found to be a relative or invalid” applies to a shtar, and raises an objection regarding the details of the decrees the Rabbanan should make in this context, and answers his objection with the claim “alternately, when we are dealing with a case when the husband says “and all of you sign the get” that “all of you” are considered like two.
If this explanation is correct, then only in the case of a condition, in which the husband demanded many witnesses, would the get be invalidated by the absence of the required number of witnesses. In a regular shtar, in regarding to which no condition was made, it is definitely possible that it would be valid if two valid witnesses sign it, even if relatives or invalid witnesses were added to the signatories. [Similarly, there is a practical difference [nafka mina] between this explanation and that of Rashi if one of the ten witnesses is not qualified to be a witness, but is not defined formally as an “invalid witness” but as “the absence of testimony.” The invalidation of “if one of them be found a relative or invalid” would not apply in this case, but still the signature of a valid witness would be lacking.
In fact, if we examine other sugyot, we will arrive at the sugya in the beginning of the chapter 10 “Get Pashut” in Bava Batra, which discusses the case where the signatures of the witnesses are written a great distance from the body of the shtar. This is what is said there:
We learned, if the [signatures of the] witnesses were distanced two lines from the writing [body of the shtar], it is invalid. Less than that, it is valid. Four or five witnesses signed the shtar, and one of them was found to be a relative or an invalid witness, the testimony is authenticated [titkayem] by the others. This supports the opinion of Chizkiya, for Chizkiya said: if it [the gap] was filled by relatives, the shtar is valid. And do not be puzzled by this, for air [a gap] [in the sekhakh] of a sukka invalidates it if it is three [t’fachim wide], and invalid sekhach invalidates the sukka only if it is four [t’fachim wide]. (Bava Batra 162b).
On the face of it, it is explicit here that there is no problem with a shtar which was signed by relatives or invalid witnesses, in addition to the valid ones, and on the contrary, [if their signatures filled in a gap] this shtar is better than one without their signatures. In other words, our claim that the rule of a relative or invalid witness [invalidating the testimony] is corroborated and confirmed here. In summary: while in regards to the sugya in Gittin we are not compelled to explain that there is a problem of ‘if it is found that one of them is a relative or an invalid witness,’ in the sugya in Bava Batra the clear impression is received that there definitely is no such problem.
However, the Rishonim refused to go on this path. Beginning with Rashi in our sugya [in Gittin] and the Rif in Makkot, until the discussions of Tosfot and the Sefardic sages, the Rishonim accept the premise of ‘if one of them is found to be a relative or an invalid witness’ applies to a shtar. This assumption is used to explain the sugya in Gittin, as we saw above, while the sugya in Bava Batra is explained in light of the internal principles of ‘if one of them is found to be a relative or an invalid witness’[3].  Even the Baal HaIttur – upon whose comment we relied in order to escape the necessity to assume that the sugya in Gittin invalidated ‘if one of them is found to be a relative or an invalid witness’ in regards to a shtar – thinks like the other Rishonim regarding this question. As so did the Rambam rule regarding this subject (Edut 5:6).
However, that which the Rishonim assumed to be something straightforward and clear, was found to leave room for doubt by the Achronim, who came back and were compelled to discuss this sugya on the basis of their understanding of the essence of a shtar. In the following lines we will endeavor, in the path of R. Chaim (Edut 5:6) and those who followed in his path, to clarify the principles which arise from their discussions.
In practice, every shtar includes within it two elements: a) The very narrative of the action the shtar presents to us. b) The fact that this narrative constitutes a legal document with binding legal authority, and is not merely a story. As Tosfot commented in Ketuvot[4], the witnesses role is crucial in transforming the shtar from a mere narrative to a legal document. In other words, the witnesses are needed to “confirm the shtar,” apart from the fact that they are vital to the act of narration that is included in the shtar. Therefore, if we come to discuss the relationship to the rule ‘if one of them is found to be a relative or an invalid witness’ in regards to a shtar, we must divide the discussion into two different components, in light of the two different roles the witnesses play.
If so, it seems that we must distinguish between the narrative and the process of creating the shtar [as a binding legal document]. In regards to the narrative, there is no room to apply the rule ‘if one of them is found to be a relative or an invalid witness,’ for here there is not testimony but the intention of the one who obligates himself, which comes from the admission of the party concerned, and the invalid witnesses have no connection to what one party writes and delivered to the other party. However, the process of creating the shar [as a binding legal document] is an act of testimony in all respects, in regard to which there is room to invalidate it because of the rule ‘if one of them is found to be a relative or an invalid witness,’ and therefore the Rambam and other Rishonim ruled that the rule ‘if one of them is found to be a relative or an invalid witness’ applies to shtarot.
The rule ‘if one of them is found to be a relative or an invalid witness,’ regarding a shtar is not explicit in the gemara, and the Amoraim do not relate to it explicitly. Logically, resolving this question depends on the nature of shtar: is it, fundamentally, written testimony, to which the rules of testimony apply, of perhaps it is an obligation of one party to another, which is not connected to the system of the laws of testimony.
There are two relevant sources we must perforce examine, one in Gittin 18b and the other in Bava Batra 162b. We can understand the gemara in Gittin as assuming this pasul [invalidation] and explain the gemara in Bava Batra according to the specific details of the rule ‘if one of them is found to be a relative or an invalid witness,’ but we could adopt the straightforward sense of the sugya in the chapter Get Pashut, according to which the rule ‘if one of them is found to be a relative or an invalid witness’ does not apply to shtarot, and to understand the problem in the sugya in Gittin as grounded in the condition which appears in it. Resolving this dilemma must be done on the basis of an understanding of the nature of shtarot.
However, the Rishonim as a group are incline to adopt the invalidating approach, despite their differing opinions on the subject of shtarot, and lack of a compelling conclusion in the sugyot. The Achronim returned and analyzed this question in light of their understanding of the essence of a shtar, and suggested distinguishing between two processes of testimony which take place in the shtar, by which they limit the rule of ‘if one of them is found to be a relative or an invalid witness’ to the ‘internal testimony’ regarding the validity of the shtar as a binding legal document, while the ‘external testimony’ of the narrative is not [really] testimony, and the things which invalidate testimony are not relevant to it.
Sources for the next shiur:
1. דף יט. "איתמר המעביר דיו... את הקרעים דיו", ירושלמי "תמן תנינן... אשת איש".
2. דף כ. "אמר רב חסדא... אבל הכא לא", שבת דף קד: "כתב על גבי... מן המובחר".
3. תוספות בסוגיתנו ד"ה דיו, רמב"ן ד"ה דיו.
4. תוספות דף ט: ד"ה מקרעין, תוספות רי"ד דף יט. "ופירש המורה... ר"ח זה לשונו".
5. רמב"ם הלכות שבת פי"א הט"ז, הלכות גירושין פ"א הכ"ג.
[6. ספר התרומה הלכות תפילין סימן ר"ה "והיכא שנפלה... לכולי עלמא אינו מועיל".]
Questions for preparation
1. What problem arises from bringing our sugya together with the ones on daf 20a and Shabbat 104b?
2. What are to various solutions of the Rishonim to this problem?
3. The Ramban distinguishes at some stage between the area of the melachot of Shabbat and the other areas. What does this distinction mean?
4. How do explain the answer of R. Yochanan to Reish Lakish’s objection according to Rashi? Does his explanation conform with the words of the Yerushalmi? Can you suggest a different explanation?

[1] See Tosfot Ketuvot 20a-b s.v. VeRabbi Yochanan, and Bava Batra 77a s.v. Chozer. Also see
Netivit haMishpat 46 note 9.
[2] See Chidushei R.Chaim HaLevi, Edut 3:4, who excludes from this general principle the laws related to the process of delivering testimony in court.
[3] Which establishes that anyone who saw the event in question without intention of testifying is not considered to be part of the set of witnesses, and thus cannot blemish it. Thus we can consider a shtar signed by relatives to be valid, on the basis of the assumption that their signatures do not indicate their intention to join the set of witnesses. See the Tosfot in Gittin.
[4] 18b s.v. amrei, and Bava Batra 162b s.v. nimtza, Chiddushei HaRamban Gittin 18b s.v. michtam and Bava Batra 162b s.v. miluhu, and the commentary of the other Rishonim to the above sugyot. And see further the Shulchan Aruch Choshen Mishpat 45:12, and the Shach there note 23.