Vayechi (5765)

From the book Mida Tova: Articles on the Hermeneutical Principles by Rabbi Michael Avraham. Translated from Hebrew using gpt-5.4 (reasoning_effort=high, batch API).

---

With God’s help — Midah Tovah — Sabbath eve of Parashat Vayechi, 5766

Questions

1. Are the hermeneutical principles technical and mathematical procedures, or are the interpreter’s own lines of reasoning also involved in the interpretations?
2. What is the relation between analogy and induction?
3. Two planes in the activation of the hermeneutical principles.
4. The place of reasoning within the picture of those two planes.
5. Two kinds of generalization and two kinds of comparison.
6. On inferential principles and local principles.

The Principles

gezerah shavah (verbal analogy), heqesh (juxtaposition-based analogy), binyan av (derivation of a general category), kelal u-ferat u-khelal (general-particular-general), davar she-hayah bi-khlal (a case originally included in a general rule).

A. Summary of Last Year’s Essay

Rabbi Yohanan said: Three keys are in the hand of the Holy One, blessed be He, and they were not entrusted to any messenger. They are: the key of rain, the key of childbirth, and the key of the resurrection of the dead. The key of rain, as it is written: “The Lord will open for you His good treasure, the heavens, to give the rain of your land in its season.” From where do we know the key of childbirth? As it is written: “God remembered Rachel, God heeded her, and He opened her womb.” From where do we know the key of the resurrection of the dead? As it is written: “You shall know that I am the Lord when I open your graves.”

— Babylonian Talmud, Taanit 2a; Sanhedrin 113a; Bereshit Rabbah 73

“The angel who redeems me from all evil,” and so forth. Rabbi Yose bar Halafta said: Sustenance is twice as difficult as childbirth. Regarding childbirth it is written, “In pain you shall bear children,” whereas regarding sustenance it is written, “In toil you shall eat of it all the days of your life.” Rabbi Elazar and Rabbi Shmuel bar Nahman differed. Rabbi Elazar said: Scripture linked redemption to sustenance, and sustenance to redemption. Just as redemption is double, so sustenance is double; just as sustenance is every day, so redemption is every day. Rabbi Shmuel bar Nahman said: Sustenance is even greater than redemption, for redemption comes through an angel, whereas sustenance comes directly from the Holy One, blessed be He. Redemption through an angel: “The angel who redeems me.” Sustenance through the Holy One, blessed be He: “You open Your hand and satisfy every living thing.” Rabbi Yehoshua ben Levi said: A person’s sustenance is as difficult as the splitting of the Sea of Reeds, as it is said, “To Him who divided the Sea of Reeds into parts,” and elsewhere it is written, “Who gives bread to all flesh.”

— Bereshit Rabbah 97, on “The angel who redeems me”; see also Bereshit Rabbah 20, on “In toil you shall eat of it”

The midrashim above deal with the relations among childbirth, redemption, and sustenance. We pointed to links and contrasts among these concepts. We saw that redemption, which is usually perceived as mystical and beyond nature, is presented here specifically as natural. Sustenance, by contrast, which is the most everyday thing imaginable, is presented as a kind of miracle. At the end of the essay we noted that both of these perspectives are correct. In every event in the world there are two dimensions: the natural and the miraculous. There is a tendency to see a contradiction between them, and that tendency forces us to choose one explanation and reject the other. Yet, as we have seen many times, both explanations can be adopted together.

In that context we dealt with heqesh, and to a lesser extent with gezerah shavah. In the sheet for Parashat Vayera, 5765, at the end of part II, we discussed the dispute between Rashi and Rashbam concerning the rule that a gezerah shavah may not be applied “by halves.” According to Rashi, in Babylonian Talmud, Keritot 22b, this rule does not establish symmetry between the two directions of the gezerah shavah; it only establishes that in the chosen direction of derivation one must compare all relevant features. According to Rashbam, in Babylonian Talmud, Bava Batra 120b, this rule also requires the gezerah shavah itself to be symmetrical, that is, bidirectional. As we saw in the sheet for Parashat Vayigash, 5765, many have compared heqesh to gezerah shavah, and it seems that the meaning of the rule that rejects a partial derivation is similar in both contexts.¹ Afterward we discussed the significance of Rabbi Elazar’s two-sided heqesh above.

We noted there that the symmetry of a gezerah shavah or a heqesh indicates that they are types of analogy, and not merely a technical-formal indicator for deriving a specific conclusion.

We then discussed the structure of an ordinary heqesh. Before determining its premises, we must decide on the axis of the comparison: which parameter it concerns. This is a decision made by the interpreter himself, either from reflection on the two sides of the equation—in our case, sustenance and redemption—or on the basis of a tradition handed down to him. In our example, the axes chosen were everydayness versus miraculousness. At the next stage, the interpreter determines his premises: redemption is supernatural and not everyday, whereas sustenance is natural and everyday. After that, he must decide the direction of the heqesh: from redemption to sustenance, or vice versa.

At that point the following questions arise: when we make a heqesh, part of the picture given in the original data changes. Redemption turns out to be everyday, and sustenance turns out to be miraculous. How do we choose along which axis, or in which direction, to perform the heqesh? Apparently there are four different possibilities, each of which would yield a different result:²

1. To learn from redemption that sustenance is miraculous.
2. To learn from redemption that sustenance is not everyday.
3. To learn from sustenance that redemption is everyday.
4. To learn from sustenance that redemption is natural.

All of this strongly suggests that two of those four data points are not absolute. It is not known in advance that sustenance is natural; it is only usually perceived that way. Likewise, it is not known that redemption is not everyday; it is only usually perceived that way. The other two data points—the supernatural character of redemption and the everyday character of sustenance—are apparently regarded by the interpreter as certain, and therefore only they can serve as possible starting points for the heqesh. Each direction of the heqesh modifies our perception of one of the other two data points, the ones that are not certain. Thus we are left with two absolute data points from which one can proceed. What remains is to determine the direction of the heqesh. It is therefore clear that the interpreter’s prior assumptions participate in the performance of the heqesh.

We noted there that such a phenomenon characterizes every analogical inference. There are several criteria for evaluating the quality of an analogy:³ the number of parameters, that is, the axes of the analogy, in which the similarity has been observed; the diversity of those parameters; the degree of speculation in the conclusion; and so on. But the most important parameter for evaluating analogies is the relevance of the observed features, those that form the basis of the analogy, to the feature inferred from it.

However, as we saw there, relevance cannot be defined a priori. So long as we have not investigated the subjects of the analogy, we cannot know which of their properties are essential and which are accidental. How, then, can such a sense of relevance be grounded a priori? That ought to be the result of the inquiry, not the assumption on which it rests.² From these and other considerations it emerges clearly that analogy always rests, at least in part, on a priori intuitions that exist within us even before we carry out the inference, or the investigation, itself.

In the end, we returned to apply these points in the interpretive context. In that context, the decision about similarity is not the result of a generalization from observing some collection of parameters of resemblance, but of a textual hint. Of course, as we have seen more than once, some of those textual hints themselves rest on the intuitions of the interpreters. This distinction blurs the importance of most of the evaluative criteria mentioned above, but the question of relevance certainly remains. We would not accept an inference that sustenance will come through the Messiah son of David, like redemption, nor that in order to bring redemption one must receive a salary and shop at the grocery, like sustenance. This is a more general expression of the question of how one chooses the axes of the heqesh, with which we dealt above.

B. The Place of the Hermeneutical Principles in the Midrashic Process

Introduction

We have already noted several times that the interpreter’s reasoning is involved in the midrashic, that is, rabbinic interpretive, process. We have seen that the interpreter chooses the axes of the heqesh and its conclusions. Rabbi Shmuel Ariel discussed this at length in his article “On Human Involvement in Derashot,” in Galut 6, Yeshivat Otniel. We will now briefly present several of the main conclusions reached there.

The involvement of the interpreter’s reasoning in choosing the axis of the derashah

Rabbi Ariel brings as an example the heqesh found in Babylonian Talmud, Kiddushin 35a, from the verse: “And it shall be for you as a sign upon your hand and as a memorial between your eyes, so that the Torah of the Lord may be in your mouth” (Exodus 13:9). The Gemara learns from there that the whole Torah is compared to tefillin, and from this it infers that just as tefillin is a positive commandment dependent on time and women are exempt from it, so too women are exempt from all positive commandments dependent on time.³

Let us assume that the language of the verse clearly indicates that the whole Torah is indeed compared to tefillin. Even then, it is still unclear what we are supposed to infer from that comparison. Even if women are exempt from every commandment similar to tefillin, the question of the parameter of similarity remains open. Perhaps women are exempt from all commandments connected with writing Torah passages on parchment—mezuzah, the writing of a Torah scroll, and the like—or from all commandments connected with animal hide. More fundamentally, who told us that this heqesh teaches the exemption of women from any commandments at all? Perhaps it teaches that just as tefillin require preparation with proper intention, so too all commandments must be performed with proper intention.

It is fairly clear that the decision on both of these questions was made by the interpreter. He decided that the heqesh teaches us about the exemption of women from commandments similar to tefillin, and he also decided that the basis of the similarity is dependence on time.

It should be noted that women’s exemption from positive time-bound commandments is therefore grounded in reasoning. Were there no such reasoning, the interpreter would have had no basis for preferring this possibility over the other possible ways of reading the heqesh. At the same time, however, it is clear that the hermeneutical principle itself is also required, and without it the reasoning that exempts women would apparently not suffice to yield a legal conclusion. Once there is a heqesh, we review all the possible ways of activating it, and the most plausible among them is the derivation concerning time-dependent commandments.

This move resembles what we saw in the Meiri’s approach in Sanhedrin regarding the concept of a “scriptural decree.” We saw there that even laws derived from a “scriptural decree” are not devoid of rationale, and their rationale may be clear to us as well. The only conclusion is that reasoning by itself would not have sufficed to teach us the legal conclusion, and therefore a “scriptural decree” is also required. In our case, that “scriptural decree” takes the form of a derashah derived through one of the hermeneutical principles.

An even more extreme example appears in Babylonian Talmud, Shabbat 64b. There the early sages derive from the verse “and she that is sick in her menstruation” that a woman may not paint her eyes or adorn herself during the days of her menstruation. Then Rabbi Akiva comes and rejects that tradition on the basis of a value judgment: “You thereby make her repulsive to her husband.” He therefore offers a different derashah on that verse, and by doing so permits a menstruating woman to adorn herself; that is indeed the accepted law. Rabbi Akiva is not merely drawing conclusions from a derashah by means of value-based reasoning; he is even setting aside an earlier tradition that interpreted the verse differently, by force of that same reasoning.

Additional ways in which the interpreter’s reasoning is involved

The interpreter’s reasoning can be involved in the midrashic process in other ways as well. Sometimes it precedes the derashah. That is, there is an existing line of reasoning, and the interpreter then finds a midrashic anchor for it. Here too one must ask: if he already has a line of reasoning, why look for a midrashic anchor for it? Clearly, the reasoning alone is not sufficient to derive the legal conclusion. Yet once we find a midrashic anchor for it, it becomes clear that this reasoning is indeed solid enough to support the conclusion we wished to derive from it.

Rabbi Ariel suggests another possibility as well. The interpreter may be troubled by a certain question, and that very concern leads him to determine the parameter along which the midrashic inference will proceed. For example, in Babylonian Talmud, Keritot 9a, there is a derashah that teaches that conversion requires circumcision, immersion, and the presentation of sacrificial blood. This rule is based on a heqesh that is made explicitly in the verse, Numbers 15:15: “As you are, so shall the convert be before the Lord.” The Gemara learns from this that the convert is like our ancestors, who entered the covenant through such a process; therefore the convert, too, enters the covenant through such a process.

At first glance, there is no hint in the verse that the comparison concerns the mode of conversion, for the plain meaning of Scripture indicates that the convert is to be compared to us in every respect. Rabbi Ariel suggests that the interpreter was troubled by the question of how the conversion process is supposed to work, since the Torah clearly teaches that conversion is possible—it deals with converts—yet it does not tell us how the process is carried out. When that question nags at us and we read this verse, the possibility immediately suggests itself that one may extract from this heqesh the manner of conversion.

General conclusions regarding the place of the principles in derashot

We have already seen several times that the great majority of the hermeneutical principles are based on the methods of analogy and induction. In fact, in the essays on Parashat Vayigash and Parashat Vayakhel, 5765, we noted that the term heqesh simply denotes analogy, and as such it is not the name of a specific hermeneutical principle, but a Hebrew rendering of the term “analogy.”

However, the principles of binyan av, whether from one verse or from two verses, are also principles of analogy; and so too are qal va-homer (a fortiori inference), gezerah shavah, and the rule of a case singled out from a general category, and many others. If so, what is the difference among all these principles? At first glance, they are all nothing more than different expressions denoting the same kind of reasoning: analogy, or heqesh.

It appears that the distinction among the principles does not lie in the question of what kind of inference is being performed on the logical plane. The distinction lies on another plane, one that precedes the inference itself. Gezerah shavah is a textual indication that an analogy is to be performed. The fact that two similar words appear in two different contexts hints to us that a comparative heqesh—that is, an analogy—can be performed here. Once we have noticed this textual hint, we determine, through reasoning, what the axis of similarity is and what can be learned from it between the two contexts under discussion.

In light of this, we can understand that qal va-homer is likewise nothing but an indication for performing a comparative inference. A hierarchical relation of greater and lesser stringency tells us that there is room to compare the two contexts.

By contrast, the principle of a case singled out from a general category instructs us to learn from the particular about the entire general category. Here too there is a textual indication, since the Torah takes the trouble to remove the particular from the general category and to write a specific case separately in addition to the general commandment. From this we learn that we are to perform a generalization from the particular to the whole. See the fuller discussion in the sheet for Parashat Vayakhel, 5765, at the end of part I. Likewise, the principle of kelal u-ferat u-khelal is a textual indicator for performing an inductive generalization: “You may infer only what is of the same kind as the particular.”

Thus, the principles are nothing but a collection of different indicators for performing an analogical or inductive inference. The difference among them lies only in the question of what triggers the analogy. But once we have noticed that trigger, we perform the same logical operation: a comparative inference, along the axis chosen through reasoning.

The principle of binyan av

The difficulty that arises in light of this distinction concerns the principle of binyan av. At first glance, this principle seems to be nothing but analogy pure and simple. In binyan av we compare contexts that resemble one another in content, and not because of any textual hint. If so, this is a pure analogy, without any prior indicator.

This gives rise to two questions:

1. Why do the sages use the term heqesh to denote analogical inference, rather than the existing term binyan av?
2. In what sense is binyan av a hermeneutical principle at all? At first glance, it is merely an analogical inference.⁴

In all the other principles, a rule given at Sinai is required in order to teach us the hints that direct us to perform the analogical inference. But binyan av seems to be nothing more than a comparison made by reasoning between two similar contexts. Why are the tools of straightforward interpretation not sufficient for that? Had the principle of binyan av not been given to us, would we not have performed analogies at all?

In the sheets for Parashat Shemot and Parashat Vayakhel, 5765, we cited a dispute among medieval authorities regarding the meaning of binyan av from one verse: is this a derivation from two sources in the structure of a “common denominator,” with both sources appearing in one context, or is it a derivation from a single source only? The difficulty exists only according to the second view. According to the first view, binyan av from one verse is also a complex analogical inference, with the logical structure of a common denominator. And indeed, from the tannaitic teaching that lists the examples, it appears that even the example brought for binyan av from one verse is built in the structure of a common denominator.

According to the view that understands binyan av from one verse as an analogy based on a single source, it would seem that we must conclude that had this principle not been given at Sinai, we would not perform an analogy without additional support, for fear of error. This principle would then teach us that one may perform an analogy without concern unless a refutation is found. But this is not correct, for as we saw in the sheet for Parashat Shemot, 5765, binyan av is a generalization, not an analogy. It is a move from the son or sons to the father that includes them. If so, this is not a principle of comparison but of generalization, and that is why the term the sages use for a midrashic analogy is specifically heqesh.

Summary: Two planes in derashot

We thus find that, generally speaking, derashot are composed of two planes:

1. The indicator, whether textual or substantive, that tells us to perform the derashah.
2. The midrashic inference itself, namely analogy or induction.

The principles differ from one another only with respect to the first plane. Different principles provide us with different indicators—some substantive and some textual—for performing a derashah. The second plane is determined by reasoning, or by tradition in the case of supportive derashot. All the rules that accompany the use of the principles—for example, that one does not expound a kelal u-ferat when the two are textually far apart; see the sheet for Parashat Ki Tissa, 5765—concern only the first plane.

The conclusion that emerges is that on the first plane there are thirteen principles, that is, different methods of interpretation. But once we have found a basis for performing the derashah by means of one of them, only two possible modes of operation remain open to us on the second plane: analogy, that is, comparison, or induction, that is, generalization. Those two modes will occupy us in the next chapter.

C. Two Types of Rational Consideration: The Relation Between Analogy and Induction

Introduction

The picture sketched in the previous chapter describes a situation that is highly complex on the first plane, the plane of the various indications that license the derashah. But on the second plane the picture appears, at least at first glance, much simpler: either one performs an analogy or one performs an induction.

In logic, from ancient Greece until the present day, discussion has focused almost exclusively on deductive inferences. They are classified into their various types, the relations among them are examined, and so forth. By contrast, analogy and induction seem not to lend themselves to such classifications, because they are modes of thought whose character is not logical-mathematical but intuitive. Even so, there is a strong sense that our intuitions themselves belong to different types. The fact that it is very difficult to classify them does not necessarily show that there are no such types. We will therefore examine two principal ways of carrying out a generalization or a comparison.

Two ways of generalizing

As we have already noted, the principle of davar she-hayah bi-khlal is one of the principles that serves as an indicator for generalization, not for comparison. From the fact that the Torah takes the trouble to specify the law of a particular case in addition to the general statement regarding the entire group, we learn that we are to take features of the particular and attribute them to the group as a whole. This description fits mainly one of the two manifestations of the principle of davar she-hayah bi-khlal—see the sheet for Parashat Vayakhel, 5765—namely, the one that infers from the particular to the general category, and not the one that narrows and characterizes the general category.

For example, the prohibition of kindling fire is written explicitly in the Torah, in addition to the general prohibition against all Sabbath labors. In the Talmud, tannaim disagreed as to whether kindling was singled out to teach that it is merely a negative prohibition, and not punishable by stoning and karet like the other Sabbath labors, or whether it was singled out in order to divide the liabilities. According to the view that it was singled out in order to divide the liabilities, we learn from kindling that just as for kindling itself one is stoned when the act is done intentionally, and when it is done unintentionally one must bring a sin offering, so too the same is true of each of the other labors.

It should be noted that here the group with respect to which the generalization is performed is explicitly given to us in the verse. What we have to do is only choose the feature with respect to which the generalization will be made. We must decide which feature of the particular is the one that should characterize the whole group. Thus, the generalization in the principle of davar she-hayah bi-khlal does not determine the group to which the law will apply; it determines only the law itself.

By contrast, in the principle of kelal u-ferat u-khelal, which is also one of the principles that serves as an indicator for generalization rather than comparison, we are given only the particular, together with the law that applies to it. The hermeneutical principle tells us that in such a case the law applies to everything that is “of the same kind as the particular.” Thus, the group to which the generalization will apply is not known at the initial stage of the derashah. We must perform a generalization through which we can define the group to which the given law will apply.

For example, as discussed in the sheet for Parashat Va’era, 5765: “Whatever your soul desires” is a general term; “cattle, sheep, wine, and strong drink” is a particular term; and “whatever your soul asks of you” is a renewed general term. The law around which the derashah revolves is the redemption of the second tithe. The particular to which the law applies is cattle, sheep, wine, and strong drink. From this particular we learn and define a group to which the given law will apply: anything produced from the produce of the earth and growing from the soil, excluding mushrooms and fungi.

Thus, although both of these principles belong to the class of indicators of generalization, they are diametrically opposed to one another. In one, the law is given explicitly in Scripture, and the task of the midrashic inference is to define the group to which the law applies. In the other, the group is given, and our task as interpreters is to determine which of the laws of the particular will apply to it. We have therefore already found two different types of generalization.

Application to analogy

The same distinction can be made with respect to analogy. Binyan av from one verse—according to the view that takes it to be an analogy from a single source—is an analogy based on a given law and a given particular. The purpose of the analogy is to define an additional particular to which that law will also apply.

By contrast, in the ordinary principle of heqesh, that is, the comparison of two particulars that appear side by side in a verse, the particular that is the target of the derashah is already known, and we must decide, by reasoning, which law will be transferred from one particular to the other. This is exactly the reverse of the case of binyan av from one verse. The same is true of gezerah shavah. There too, the target of the midrashic inference is a known particular context, and we must decide which parameters will be transferred in that inference.

Ambiguous principles

The picture that divides the principles into indicators of analogy and indicators of induction is too superficial. For example, in the principle of kelal u-ferat, the rules of derashah instruct us to derive from it “only what is in the particular,” not “what is of the same kind as the particular,” as in kelal u-ferat u-khelal. One might therefore have understood this as an indicator for performing an analogy, that is, a comparison with those particulars that are completely identical to the given particular, rather than a generalization to everything merely “of the same kind as the particular,” as in kelal u-ferat u-khelal. Yet, as part of the family of kelal u-ferat principles, it seems more plausible to view this principle as an indicator of generalization rather than comparison—albeit a generalization to a narrower group than in kelal u-ferat u-khelal, namely, only to contexts that are fully similar to the given particular, “in three respects”; see the sheet for Parashat Vayikra, 5765.

The principle of binyan av from one verse raises the same problem. Sometimes it is an analogy from one given particular to another, or from one category to a parallel category, and sometimes it is a generalization from a given particular to the general category—the “father”; see the sheet for Parashat Shemot, 5765—that resembles it.

The relation between analogy and induction

This phenomenon raises the question of the relation between induction and analogy. We have already noted in the past, see the sheet for Parashat Va’era, 5765, that there is a complex relation between these two modes of inference. On the one hand, every analogy is in fact a hidden induction. For example, if we wish to perform an analogy between one frog and another, and infer from this that if the first is green then the second is green as well, then in the background there stands an induction concluding that all frogs are green, since we assumed nothing at all about the frog currently under discussion. On the other hand, every induction presupposes at its base a whole collection of hidden analogies. For example, if we wish to infer from the fact that frog A is green that all frogs are green, it would seem that we are relying on a collection of analogies. With respect to each frog, one can infer by analogy that its color is green, and the sum of all those analogies is the inductive generalization. The question, then, is which of these is the more basic inference, and which is derivative.

In light of the division we made above, perhaps this fog can be dispelled somewhat. When the law is given explicitly in the verse and we are trying to define the group to which the law will apply, this is a generalization built from a collection of analogies. We build a group each of whose members resembles our particular, and we apply the given law to each one separately. But when the target group is given, and we are merely seeking which law, from among those that characterize the particular, should be applied to that group, then it seems that here we are performing a direct generalization. The inference is made with respect to the whole group at once, and we seek which of the features of the particular can suit that whole.

With respect to the two corresponding kinds of analogy, one can see a similar division. In the analogy where the target particular is given and we are merely looking for the law to apply to it, we are performing a direct analogy. But in the opposite case, where the law is given and we are looking for the particular to which it is to be applied, then it seems that we are relying on induction. The inference is directed toward all particulars for which the given law is relevant, and from among them we seek the specific particular that will be the object of the midrashic inference under discussion.

Summary

We have presented here, in broad outline, a picture of the system of hermeneutical principles. We classified them into two main groups: indicators for performing analogy, and indicators for performing induction. Each of those two groups was then divided again into two subgroups, according to whether the analogy or induction is seeking a rule or a field of application. We saw that there are principles that stand between these sectors and can be assigned to more than one of them.

It is of course clear that there are hermeneutical principles that do not appear on this map. For example, the principle of “two scriptural verses that contradict one another” performs neither analogy nor induction, but rather attempts to resolve a contradiction between two biblical sources. This principle is not an inferential principle that expands the given context; rather, it is a principle that helps us solve local interpretive problems within that very context itself.

Footnotes