What Generates the Obligation of a Shevu'a

  • Rav Joshua Amaru


This week’s shiurim are dedicated in commemoration of the yarhzeit of
Rabbi Lipman Z. Rabinowitz, by his family




1. גמרא ג: "אלא אתיא מעד אחד ... תורת הזמה לא פריך" (ד.). רש"י וריטב"א.

2. רמב"ן ד"ה 'הכא גרסינן', רשב"א ד"ה 'הכי גריס רש"י'.

3. תוספות ד"ה 'הצד השוה', תוספות רבינו פרץ ד"ה 'שעל ידי טענה וכפירה וכו''.

4. שבועות מ. "אמר רב נחמן בר יצחק אמר שמואל ... מחייבו שבועה", חידושי הרמב"ן על אתר "איכא דדייקי...".


1. Pay attention to the various stages in the attempt to understand how Rav Chiyya's derived his halakha. On what basis does the Gemara reject the possibilities of learning the din from single witness and gilgul shevu'a?

2. The Gemara concludes that Rav Chiyya bases his halakha on the common element found in two sources tzad hashaveh. The Spanish Rishonim had a different reading than Rashi regarding the final position that the Gemara proposes for Rav Chiyya. What are these two sources according to Rashi and according to the Spanish Rishonim? How does this difference affect the final understanding of the derasha?

3. What may be learned from the Tosafot regarding the factor that obligates an oath? Does the Ramban agree?

4. The Ramban cites a disagreement between the Rif (Shevu'ot 28a in the Rif's pagination) and Rabbenu Efrayim (and Ri Migash), on the issue whether the testimony of a single witness obligates an oath when the plaintiff is himself is uncertain that anything is owed to him. How is this disagreement reflected in the various understandings regarding the criteria for taking an oath?


I.  The Derivation of R. Chiya's Ruling

            In a previous shiur, we addressed R. Chiya's assertion that the defendant is obligated to take an oath in a situation where two witnesses testify that he owes part of a claim which he completely denies.  According to R. Chiya, this conclusion arises out of the principle "that his (the defendant's) admission should not be greater than the testimony of witnesses."  The gemara follows with various attempts at establishing the exact content of this kal va-chomer which prefers witnesses to admission.  After these attempts fail, we take up our shiur at the point where the gemara changes strategy and tries to deduce R. Chiya's ruling from a comparison of the testimony of a single witness to that of two witnesses.

            Ultimately, the gemara deduces R. Chiya's ruling from the tzad ha-shaveh, or 'common denominator' of both previously suggested sources, i.e., the admission of the defendant ('hoda'at piv') and the testimony of a single witness ('ed echad').  Both a partial admission and the testimony of a single witness create a situation wherein we require an oath from the defendant.  The common ground between these two halakhot is that both of them arise out of a situation of a 'claim and a denial' (ta'ana u-kefira) and we can therefore assume that this situation of 'claim and denial' is that which generates the obligation to take an oath in general.  Two witnesses also arise out of a situation of 'claim and denial' and therefore they too can obligate an oath.

            It would seem that beyond furnishing a justification for R. Chiya's ruling, the gemara here is telling us something fundamental about the requirement of oath-taking.  This requirement (both in the case of modeh be-miktzat and in the case of ed echad) arises from a common source - a situation of 'claim and denial.'  In the rest of the shiur, we will explore what the gemara could possibly mean by 'a situation of claim and denial,' and the repercussions of this concept on our general understanding of what factor generates the obligation to take an oath (in other words, the 'mechayev' of shevu'a).

            'A situation of claim and denial' is the tzad ha-shaveh of two cases, hoda'at piv and ed echad.  In order to understand the set described by the tzad ha-shaveh, we need to understand its two (or really three, since two witnesses are included in this category,) members.  Hoda'at piv has already been discussed in previous shiurim and we will not go into it in detail.  Hoda'at piv being referred to is that of a case of modeh be-miktzat which obligates the defendant to take an oath on the remainder.  But what is the case of ed echad?  On this question there is a disagreement between the Rishonim which also involves a textual question.  The gemara, before arriving at the tzad ha-shaveh as its source, tried to deduce R. Chiya's ruling from ed echad alone.  This was rejected because a single witness only obligates an oath about that which he testifies while we are looking for a source for an oath on the part regarding which there is no testimony.  As the Ritva points out, the problem is not that ed echad is too strong a source such that there is no kal va-chomer to the case of two edim, but rather that ed echad is too WEAK a source - since the defendant is permitted to take an oath to contradict the testimony of a single witness.  Therefore, we cannot deduce anything about the oath he would have to take in order to avoid paying the rest of the plaintiff's claim regarding which there is no testimony.  In other words, an ed echad is only able to create obligations about his own testimony and the fact that he can require an oath about his own testimony does not tell us anything about the remainder of the claim.

            Following the rejection of ed echad as the sole source, the gemara tries to deduce R. Chiya's ruling from 'gilgul shevu'a' of ed echad; i.e., from that fact that when a defendant was forced to take an oath to contradict the testimony of a single witness, the plaintiff can "roll" onto him additional oaths that under normal circumstances he would not be obligated to take.  Here we seem to have a case in which an ed echad, through the mechanism of gilgul shevu'a, can obligate an oath even about the remainder of the claim.  The gemara rejects this suggestion, because the oath in a case of gilgul shevu'a can be viewed as an extension of the primary oath - 'an oath drags along another oath' - and therefore cannot be ascribed to the ed echad.  Hence we cannot derive R. Chiya's ruling from ed echad.  [This gemara has important repercussions for the understanding of gilgul shevu'a which will be discussed in an upcoming shiur.]

            At this point the gemara combines ed echad and 'piv' which gives rise to the tzad ha-shaveh.  However, it is not completely clear which case of ed echad is being combined with piv.  Is it our original suggestion of the oath generated by the testimony of the ed echad or is it the case of gilgul shevu'a attached to an ed echad?

II. The Ramban's Text of the Gemara - Gilgul Shevua of an Ed Echad as a Source

            Chakhmei Sefarad, or the Spanish School, led by the Ramban, gloss the gemara text differently than the text printed in our gemara.  Following a Gaonic text, the Ramban glosses "ma le-ed echad she-ken shevu'a goreret shevu'a."  In other words, the Ramban learns the tzad ha-shaveh not from the standard case of ed echad but from the case of gilgul shevu'a generated by an ed echad.  For the Ramban, the gemara's rejection of gilgul shevu'a as a source for R. Chiya's ruling does not imply that the ed echad is in no way connected to the second oath.  Rather, we understand the case of gilgul shevu'a as a mechanism by which the ed echad can obligate an oath even on the part of the claim not included in his testimony.  We can deduce from the obligation of an oath, both in gilgul shevu'a of an ed echad and in modeh be-miktzat, that two witnesses can also obligate an oath on the remaining claim.  We can make this deduction, the gemara claims, because the oath obligation in the case of both gilgul shevu'a of an ed echad and of modeh be-miktzat is a function of the same principle - they both arise out of a claim and denial situation - and this principle is also applicable to R. Chiya's case. 

            What is a 'claim and denial situation'?  The Ramban s.v. Hakhi garsinan, explains that both the claim and the denial refer to the defendant.  In both gilgul shevu'a of an ed echad and modeh be-miktzat, the defendant completely denies a portion of the claim while admitting to the remainder.  (In the case of ed echad, of course, he does not admit anything, but the Ramban claims that the single witness's testimony is "like an admission" insofar as it raises a serious suspicion as to the honesty of the defendant.)  Therefore, in both cases the defendant is required to take an oath about the portion he 'denies,' and from this we can deduce an obligation for an oath in R. Chiya's case.  Through the tzad ha-shaveh, we can broaden the concept of modeh be-miktzat.  Without the combination with gilgul shevu'a of ed echad, we would have thought that the shevu'a mandated by an modeh be-miktzat is limited to cases of actual admission.  After the tzad ha-shaveh, we conclude that the oath arises out of a 'claim and a denial situation,' which includes cases where the 'admission' is provided by some form of testimony.  The tzad ha-shaveh does not, however, inform us in any way about other cases where an oath is mandated, such as the regular case of ed echad.

III. Rashi and Tosafot's Text of the Gemara - Ed Echad as a Source

            Rashi's text, which is that of our edition of the gemara, glosses "ma le-ed echad she-ken al ma she-me'id hu nishba" which implies that shevu'a generated by the ed echad's testimony without gilgul shevu'a.  Clearly, Rashi understood the gemara's rejection of the gilgul shevu'a as an element in the kal va-chomer to be conclusive - the second oath is generated by the primary oath and cannot be ascribed to the ed echad.  Therefore, the gemara reverted to the original suggestion of ed echad and combined it with 'hoda'at piv' to establish a tzad ha-shaveh.  The implication of deriving R. Chiya's ruling in this way is radical - we conclude that at least two of the three cases of "shevu'a de-oraita," of an oath mandated by the Torah (ed echad and modeh be-miktzat), are not independent of one another but are applications of a meta-principle based upon 'claim and denial.'  Only Rashi's text of the gemara (as opposed to that of the Ramban), leads us to such a meta-principle since only in Rashi's text are the components of the tzad ha-shaveh two basic categories of the oath obligation. 

            What is the meta-principle which unites both categories of oaths?  Obviously, its content will depend upon our interpretation of the tzad ha-shaveh - a claim and denial situation.  Rashi, s.v. Ta'ana u-kefira explains that a 'claim and a denial situation' is simply one in which there is a plaintiff and a defendant and they come to court.  The implications of this interpretation are very far-reaching in terms of our understanding of what obligates a shevu'a.  According to Rashi, a shevu'a is mandated in any situation where a plaintiff makes a claim that is denied by the defendant in court!  Given this approach to the shevu'a's place in a claims court, Tosafot, s.v. Ha-tzad ha-shaveh ask an obvious question - why is it that in the case of a complete denial by the defendant (kofer ha-kol) there is no shevu'a de-oraita?  Tosafot answer that not every claim can obligate a defendant to take a shevu'a - only an 'important claim' (ta'ana chashuva) obligates a shevu'a.  An important claim is one similar to that of the plaintiff in the cases of ed echad or modeh be-miktzat, i.e., cases where we have reason to believe that the defendant is lying.  In other words, according to Tosafot, the obligation of a shevu'a on the part of the defendant arises out of a situation where his denial is weak relative to the claim of the plaintiff.  A strong claim may be based upon the testimony of a single witness or partial admission to the claim, or as in R. Chiya's case where two witnesses testify that he owes part of the claim.  In a situation where the defendant's denial is of equal weight to the plaintiff's claim, there is no requirement of an oath.   So what is the meta-principle according to the Tosafot?  In my opinion, Tosafot can be interpreted in two ways:

1) If we emphasize the end of the Tosafot, which explains the 'ta'ana chashuva' as a function of the fact that we have reason to doubt the defendant, we could say that we require a shevu'a in any case where we have reason to believe that the defendant is lying.  Ed echad and modeh be-miktzat are simply specifications of situations where such suspicions exist.  This explanation is very similar to the explanation offered by many Rishonim that the oath of modeh be-miktzat is a function of 'raglayim la-davar' (grounds for belief) that the defendant is lying.  Our meta-principle is therefore that a shevu'a is obligated whenever we suspect the defendant of lying in his denial of the plaintiff's claim.

2) In my opinion, Tosafot did not use the term 'ta'ana chashuva' frivolously and our explanation must refer to this factor - the ta'ana, or claim.  In order to appreciate Tosafot's understanding of what obligates an oath, we must direct our attention not only to the defendant but to the plaintiff.  It appears that the obligation of the defendant to take an oath arises out of the necessity to respond to the plaintiff's claim.  When the plaintiff introduces a claim which cannot be swept aside and ignored, the defendant must respond with an oath.  This is true in all the cases in question, since the claim of the plaintiff has objective support, while the denial of the defendant is unsubstantiated and appears suspicious.  According to this interpretation of the Tosafot, the meta-principle which generates an oath-obligation is that in a situation where the plaintiff's claim has objective backing, the defendant must use an oath to deflect this claim.  Such is the situation in the cases of ed echad, modeh be-miktzat, and that described by R. Chiya, where the claim is supported either by some form of testimony of by a partial admission. 

            Tosafot R. Peretz s.v. She-al yedei ta'ana offers a different answer to the Tosafot's question.  According to R. Peretz, even a situation where the defendant completely denies the plaintiff's claim is included in the category of a 'claim and a denial situation.'  The only reason there is no obligatory oath in a case of complete denial is because of a 'gezeirat ha-katuv' (the Torah's decree).  This interpretation implies a radical understanding of the oath-obligation: According to R. Peretz, every case of a claim in principle requires an oath from the defendant.  One could perhaps explain this position as being based upon the chazaka that a person does not make a claim unless he is owed (Shevu'ot 40b) and thus the claim of the plaintiff always has objective basis.  In any case, according to R. Peretz, no special circumstances are required in order to mandate a shevu'a - a shevu'a is the normal response of a defendant to a claim made against him.  In order to require an oath in R. Chiya's case, all we need to show is that there is a claim being made that the defendant is denying.

            To sum up, following Rashi's text of the gemara, we have arrived at two possible meta-principles for the obligation of shevu'a: 1) Our first interpretation of Tosafot, which requires an oath of the defendant when we have grounds to believe that he is lying.  2) Our second interpretation of Tosafot, which requires an oath of the defendant in order to bolster his denial in the face of a stronger claim.  Tosafot R. Peretz goes even further, demanding an oath of the defendant in order to respond to any claim on the part of the plaintiff.  At this point, we can turn to some general issues of oath obligations and try to apply what we have learned.

IV. The Obligation of Shevu'ot In General

            The basic difference between the opinions discussed above lies in the question of whether we find the source of the oath-obligation in the plaintiff's claim against the defendant or whether we find it in the suspicious behavior of the defendant.  The Ramban (at least regarding modeh be-miktzat) and our first interpretation of the Tosafot (that which saw the source of the oath-obligation in the fact that we suspect the defendant of lying) clearly take the latter view.  Tosafot R. Peretz and our second interpretation of the Tosafot (based upon ta'ana chashuva) are rooted in the former.  What are the repercussions of the difference between these two approaches?  In the following we will list two, though there certainly are others as well.

1.  What about the third shevu'a de-oraita, the oath of a "shomer?"  Obviously, we will find no answer to this question according to the Ramban, since even if he does think that all cases of the oath-obligation arise from a single principle, that principle is not to be found in our sugya.  But how would the Tosafot and R. Peretz relate to the oath-obligation of a shomer?  If we interpret the Tosafot to mean that the oath-obligation resides in our suspicion that the defendant is lying, it seems difficult to apply this principle to a shomer.  What grounds are there to suspect the shomer of being a liar?  At most we could suggest that Rami Bar Chama, who holds that the shomrim oath requires a partial admission (Bava Metzia 5b), believes that the shomrim oath is also founded on suspicion of lying.  However, the mainstream opinion, that requires an oath of a shomer anytime he makes a claim that would absolve him of reimbursing the owner, must view the shomrim oath as founded on an independent principle.

            On the other hand, it seems possible that according to R. Peretz, the shomrim oath is an application of the same principle as found in the other cases of oaths.  If the oath is simply the response the Torah requires to a claim, then the claim of the owner on the shomer, obligates him to take an oath.  It is still possible that R. Peretz views the shomrim oath as independent only because in the case of shomrim, there is no definitive claim on the part of the owner, since in most cases he has no knowledge of what occurred to the object in question.  On the other hand, there is a certain strength in the owner's assertion that may be sufficient to obligate the oath - after all, the shomer does not in any way deny the owner's claim that the object was placed in the shomer's possession.  It thus remains an open question whether the general principle of oath-obligation that we derived from ed echad and modeh be-miktzat can be extended to the shomrim.

2.  The second instance through which we can investigate the various understandings of the oath-obligation is already mentioned in the Rishonim (see Ramban et al. Shevu'ot 40a).  There is a disagreement between the Rif (Shevu'ot 28b in the Rif) and Rabbenu Efraim (quoted in Ramban above) regarding the oath-obligation of an ed echad in a case where the plaintiff has an uncertain claim (ta'anat shema).  For example, Reuven makes the following claim to Shimon:  An ed echad told me that you owe me $100.  If the ed echad comes and testifies, is Shimon obliged to take an oath to contradict the ed?  According to the Rif, he is obliged, though Reuven himself made no independent claim.  According to Rabbenu Efraim, he is not obliged.  Rabbenu Efraim brings support to his position from our gemara that in order to be obligated in a shevu'a one needs to be in a claim and denial situation which is not the case if his claim is uncertain. 

            According to the Ramban, our sugya does not have any impact on the issue of an uncertain claim supported by an ed echad, since our gemara refers exclusively to the special case of gilgul shevu'a by ed echad (the Ramban makes this point explicitly in Shevu'ot 40b).  According to the Tosafot, if we understand the shevu'a to be based on our suspicion that the defendant is lying, there does not seem to be any reason to distinguish between an uncertain and a certain claim (and so we should hold like the Rif).  We are not really interested in what the plaintiff says but in the believability of the defendant's response.  If, however, we understand the oath as a response to the plaintiff's claim, then the nature of that claim becomes important.  If we require a 'ta'ana chashuva,' a relatively strong claim on the part of the plaintiff, then the disagreement between the Rif and Rabbenu Efraim would turn on the question of whether we regard an uncertain claim on the part of the plaintiff (albeit supported by a witness) as a ta'ana chashuva.  It is certainly not unreasonable to adopt Rabbenu Efraim's position on the basis of the view that the uncertain claim is not considered a ta'ana chashuva.  If we do not require a ta'ana chashuva, but rather any claim is sufficient to generate an oath-obligation (R. Peretz's position), the disagreement between the Rif and Rabbenu Efraim would turn on the question of whether we regard an uncertain claim as a real claim.  It is possible that there arises an oath-obligation only in situations where the plaintiff is putting the full pressure of a claim on the defendant and without this pressure a simple denial will do.

V.  Summary

            In conclusion, we have discussed in this shiur the different interpretations of the tzad ha-shaveh, 'a claim and denial situation,' from which the gemara deduces R. Chiya's ruling and the implications of these interpretations for a general understanding of the oath-obligation.  We saw two interpretations of the tzad ha-shaveh: The Ramban regards the claim and denial situation to be an extension of the concept of modeh be-miktzat where the oath-obligation is based upon our suspicion that the defendant is not telling the truth.  Rashi understands the 'claim and denial situation' more broadly and finds here a meta-principle which is the source for the oath-obligation in both ed echad and modeh be-miktzat.  We investigated different candidates for such a meta-principle:  a) our first interpretation of the Tosafot that the oath in both cases is a response to our suspicion that that the defendant is lying.  b) our second interpretation of the Tosafot that the oath is necessary to respond to the plaintiff's claim when that claim carries a greater relative weight than the defendant's denial.  c) The Tosafot R. Peretz that an oath is necessary to respond to any normal claim on the part of the plaintiff.

Sources and questions for next week’s shiur:


  1. גמ' דף ד'. "אלא אי אי תמר..." עד ד: "...וש"מ דהילך פטור"
  2. ח' הר"ן ד"ה והילך
  3. רמב"ם הל טוען פ"א ה"ג
  4. בעל המאור
  5. ר"ן ד"ה והא


  1. Compare the different definitions of “heilakh” found in Rashi, the Rambam, and the Ran. Define “heilakh” for each.
  2. What is the basis for the gemara’s comparison of “heilakh” to our mishna? See Rashi, Tosafot (technical answers), the Baal HaMaor and the Ran (logical answers).