On Commandments and Parts of Commandments – An Analysis of the Nature of Concepts in Jewish Law and in General

Akdamot – 2008

Introduction

The principles concerning the commandments that Maimonides formulated and defined in Sefer HaMitzvot present a very rare and unique text, in which a central halakhic sage offers a reflective picture of his methodological and meta-halakhic assumptions. In each of Maimonides’ principles one can discern a philosophical principle, and in almost all of them this is not merely a meta-halakhic principle, but one that touches on issues in general philosophy as well.[1] In this article I will try to propose a philosophical foundation for understanding the dispute between Maimonides and Nachmanides in the eleventh principle, which concerns the parts of commandments, and I will seek to derive from their halakhic (or meta-halakhic) positions philosophical conclusions regarding the nature of halakhic concepts in particular, and the nature of concepts in general. More specifically, we will consider whether the halakhic concept is an ‘entity,’ or an Idea that possesses some sort of existence, or whether a concept is nothing more than a purely linguistic convention.

Preface: A Brief Look at Maimonides’ Principles

The point of departure for what follows is the dispute between Maimonides and Nachmanides in Principle 11 of Sefer HaMitzvot, and therefore I will preface the discussion with a few words about Maimonides’ fourteen principles. Although these constitute a very rare and unique text, in which a central halakhic sage reflects on his own working method,[2] to the best of my knowledge relatively little systematic work has been done in an attempt to understand these principles in depth.[3] Studying them exposes a number of philosophical questions, both with respect to the philosophy of Jewish law and with respect to philosophy in general. In this article I will present one of those questions.

Maimonides’ halakhic work was carried out in an almost deductive manner. In order to write the Mishneh Torah, the only comprehensive halakhic code ever written,[4] he needed a framework that would help him classify and organize this vast and complex material, and ensure that he forgot or omitted nothing. Maimonides chose as that framework the enumeration of the commandments. He combed through the Written Torah and all the literature of the Sages and the Geonim that was available to him, in order to extract from them the 613 commandments.[5] He classified and arranged these commandments by topic, and these served as the basis of his monumental Mishneh Torah.

However, as every student knows, extracting and defining the 613 commandments from the full corpus of halakhic literature is an extremely difficult task, involving numerous methodological assumptions. Until Maimonides’ time, the enumeration of the commandments was dominated by the count of the author of Halakhot Gedolot (hereafter: Behag), and all those who enumerated the commandments (at least those known to Maimonides) followed him. In his introduction to Sefer HaMitzvot, Maimonides refers to this very critically, and even mocks those who followed Behag as though reason had come to a halt at this man’s statement (as though reason had come to a halt with this man’s statement).[6] Despite the mockery, Maimonides had to substantiate his revolutionary claims regarding the enumeration of the commandments, since they stood in opposition to the dominant approach that had prevailed until his time.

For this reason, Maimonides found it necessary to write fourteen principles, in which he included the criteria by which he decided how to classify and categorize the commandments. The principles fall into several types: there are principles dealing with the halakhic status of laws as a basis for their inclusion or non-inclusion in the count (for example, the first and second principles); there are principles dealing with duplications (for example, the sixth and ninth principles); there are principles dealing with the question of what constitutes a command (for example, the fifth principle); and there are principles dealing with parts of commandments that should not be counted separately, but only as a whole.

The principle with which we are concerned here (Principle 11), as well as the one that follows it, belongs to the principles that deal with parts of commandments. The conclusion emerging from both is that parts of commandments should not be counted independently, but rather all the parts should be gathered together and included in the enumeration as a single commandment.

A. Maimonides’ and Nachmanides’ Approaches in Principle 11[7]

In Principle 11, Maimonides establishes the following rule:

One must not count the parts of a commandment separately, each part on its own, when their combination constitutes a single commandment. (One must not count the parts of a commandment separately, each part on its own, when their combination constitutes a single commandment.)

The example Maimonides himself gives is the four species. We are commanded to take all four of them, and if one is missing we have not fulfilled our obligation in the commandment. Therefore, Maimonides argues that one should not count separately the commandments to take the lulav, the willow, the myrtle, and the citron, but rather count one inclusive commandment of taking the four species, within which the four species appear as particulars (see positive commandment 169 in Sefer HaMitzvot).

In his discussion, Maimonides examines the implications of the halakhic question whether the parts of a commandment are mutually indispensable for one another, in relation to the rule under discussion here. At first glance, these are two sides of the same coin: if parts of a commandment are mutually indispensable, we ought to count them as one commandment; if there is no such indispensability, then there are two commandments here. And indeed, the four species of the lulav are mutually indispensable, as are the showbread and the frankincense that accompanies it. The same is true of the purification of the metzora and of his removal. In all these cases, Maimonides states, there is no doubt that these are parts of a single commandment, since they are mutually indispensable.

Maimonides’ explanation is that the desired end is not realized unless all the parts are performed. Therefore it is clear that all these parts form one whole. But what of cases in which the parts are not mutually indispensable? Maimonides writes that there is a subtle element at the root of this principle: the criterion for determining whether we are dealing with one commandment or two distinct commandments is one-directional. If the parts are mutually indispensable, then we are dealing with different parts of one commandment. But if they are not mutually indispensable, we must not necessarily infer that these are two commandments. In such cases, sometimes the parts will be counted as one commandment, and sometimes as separate commandments.

The two main examples around which the medieval authorities debate are brought in the Mishnah in Menahot (38b):

The blue thread does not preclude the white, and the white does not preclude the blue. The arm-tefillin do not preclude the head-tefillin, and the head-tefillin do not preclude the arm-tefillin. (The blue thread does not preclude the white, and the white does not preclude the blue. The arm-tefillin do not preclude the head-tefillin, and the head-tefillin do not preclude the arm-tefillin.)

Now according to Maimonides, the commandment of tzitzit (ritual fringes), with its two components, is counted as one commandment (see positive commandment 14 in Sefer HaMitzvot), even though the white component does not preclude the blue one and the blue one does not preclude the white. Tefillin, by contrast, he counts as two separate commandments (positive commandments 12-13).

Maimonides brings proof for his position from the Mekhilta of Rabbi Ishmael:

One might have thought that these are two commandments, the commandment of the blue thread and the commandment of the white thread. Scripture therefore says, 'and it shall be a fringe for you'—it is one commandment and not two. (One might have thought that these are two commandments, the commandment of the blue thread and the commandment of the white thread. Scripture therefore says, and it shall be a fringe for you—it is one commandment and not two.)

At the end of the principle Maimonides explains the matter as follows:

You now see that even parts that do not preclude one another may sometimes constitute one commandment, when the matter is one. For the purpose of tzitzit is remembrance. Therefore the whole thing that produces remembrance is counted as one commandment. Thus, in counting the commandments, what ultimately remains for us to examine is not whether they are mutually indispensable or not, but only whether this is one matter or many matters. (You now see that even parts that do not preclude one another may sometimes constitute one commandment, when the matter is one. For the purpose of tzitzit is remembrance. Therefore the whole thing that produces remembrance is counted as one commandment. Thus, in counting the commandments, what ultimately remains for us to examine is not whether they are mutually indispensable or not, but only whether this is one matter or many matters.)

Maimonides explains here that the criterion is whether there is ‘one matter’ here, that is, one purpose and one end. According to Maimonides, the Sages’ midrash in the Mekhilta teaches us that these two parts serve one idea, and therefore, even though they are not mutually indispensable, they are parts of one whole.[8] This shared idea is apparently the remembrance of the commandments: so that you may remember and perform all My commandments and be holy to your God (so that you may remember and perform all My commandments and be holy to your God) (this is also how Nachmanides explains it in his glosses). Apparently, according to Maimonides, this is the Torah’s own intention when it says and it shall be a fringe for you. The blue and white components join into one whole called tzitzit.

As noted, Nachmanides, in his glosses to this principle, agrees with the basic principle in Maimonides’ formulation. He also accepts that mutual indispensability is not the criterion; rather, the question is whether the parts have a single subject matter. However, he challenges Maimonides on the basis of his count regarding tefillin, since the arm-tefillin do not in fact preclude the head-tefillin, and vice versa, yet it is nevertheless clear that their subject matter is one:

And if we look at their content, tefillin should be considered all the more one matter, for everything written about this is written about that, and their content is one: "so that the Torah of the Lord may be in our mouth"—they are placed opposite the heart and the brain, the dwelling places of thought. (And if we look at their content, tefillin should be considered all the more one matter, for everything written about this is written about that, and their content is one: so that the Torah of the Lord may be in our mouth—they are placed opposite the heart and the brain, the dwelling places of thought.)

Thus Nachmanides himself agrees in principle with Maimonides that even parts that are not mutually indispensable, if their subject matter is one, are counted as one commandment.

From Maimonides’ words it emerges that apparently there is no shared idea joining the arm-tefillin and the head-tefillin, and therefore he counts them as two commandments. Here it is worth sharpening the issue of the shared idea of parts of a commandment. Even if two actions are intended to achieve a similar goal, that does not thereby make them one commandment. Dwelling in a sukkah and eating matzah are both commemorations of the Exodus from Egypt, and yet it is clear that these are two commandments. Parts of a commandment are counted as one commandment only when the goal toward which they are directed is achieved through the combination of them all. And this is what Maimonides himself writes:

Likewise, wherever it is clear to you that the sought end cannot be attained by one of those parts alone, it is clear that their combination is what is counted… So too, his purification—the metzora’s purification—is not attained except through everything mentioned: the birds, cedar wood, crimson thread, and shaving; only then is his purification attained. (Likewise, wherever it is clear to you that the sought end cannot be attained by one of those parts alone, it is clear that their combination is what is counted… So too, his purification—the metzora’s purification—is not attained except through everything mentioned: the birds, cedar wood, crimson thread, and shaving; only then is his purification attained.)

Maimonides’ approach to tefillin can likewise be explained as follows: the arm-tefillin correspond to the heart (inclination and emotion), while the head-tefillin correspond to the brain (thought and belief). Even if these are two similar goals, Maimonides apparently holds that because neither requires the other, the commandment of tefillin is counted as two commandments. By contrast, in tzitzit the remembrance of the commandments is achieved optimally only through the combination of blue and white.[9] Therefore tzitzit is counted as one commandment.

Be that as it may, the definition of these matters on the level of general logic still requires clarification: what is the difference between parts of a commandment that are not mutually indispensable but do have a shared idea (as in tzitzit), and two parts that are mutually indispensable (as in the four species)? If in both cases we are dealing with one commandment, then why are there cases in which the parts are mutually indispensable and cases in which they are not? And as for the third case, in which there are two parts of a commandment that are not mutually indispensable and also have no shared idea (as tefillin according to Maimonides), why does the possibility even arise that such two parts belong to one commandment?

B. Three Types of Relation Between the Parts of a Commandment and the Whole: Three Approaches Among the Medieval Authorities

Between the whole of the commandment and its parts, three possible types of relation may obtain:

1. Each part possesses existence of independent value, and the parts do not combine at all into a complete whole.
2. The fulfillment of each part has value, but only the joining of all the parts together constitutes the full whole of the commandment.
3. No individual part has value in itself, and the value of the commandment is created only when all are fulfilled together.

In case 1, it is clear that one part does not preclude the other, nor does it preclude fulfillment of the commandment as a whole (for there is no inclusive commandment that gathers them together). At first glance, these are not parts of one commandment at all, but two different commandments. In case 3, it is clear that the absence of one part precludes the entire commandment, and therefore there is no point in performing the second part alone (for its entire purpose is the fulfillment of the inclusive commandment).

In case 2, it can be said that one part precludes fulfillment of the commandment as a whole, but does not preclude the second part. In such a case there is reason to perform the second part by itself, even if we do not attain the benefit that arises from the commandment as a whole. However, this case too can appear in two forms:

2a. When one performs one part, there is fulfillment of that part itself, but not of the whole.

2b. There is here a deficient fulfillment of the commandment as a whole.[10]

Parts of a commandment that are mutually indispensable are certainly parts of one inclusive commandment. That is case 3. Parts of a commandment that are not mutually indispensable, according to Maimonides’ position, may sometimes be counted as one commandment and sometimes as two. The difference depends on whether the fact that one can perform the second part without the first is rooted in the independent value of that performance (case 2a)—in which case they are counted as two commandments—or whether it is a deficient fulfillment of the whole (case 2b)—in which case there is one commandment here.

This is the difference between tzitzit and tefillin: in both cases the parts are not mutually indispensable, that is, they certainly do not belong to the third type of relation. However, according to Maimonides, in tzitzit the white and the blue combine into a whole, and therefore when one fulfills the commandment only with white, or only with blue, there is a partial fulfillment of the entire commandment of tzitzit. Therefore the commandment of tzitzit is one inclusive commandment, even though the parts do not preclude one another. Regarding tefillin, Maimonides proved from the wording of the above midrash that the two parts do not combine into one whole, and therefore these are two distinct commandments.[11]

It would seem that one can see this in the verses themselves (Numbers 15:38-40):

Speak to the children of Israel and say to them that they shall make for themselves fringes on the corners of their garments throughout their generations, and that they shall put on the fringe of each corner a cord of blue. (Speak to the children of Israel and say to them that they shall make for themselves fringes on the corners of their garments throughout their generations, and that they shall put on the fringe of each corner a cord of blue.)

The object is called tzitzit even before blue is placed upon it. Upon the white tzitzit one also places a blue cord, that is, this is an addition. The whole is also called ‘tzitzit’ (the next verse: and it shall be a fringe for you). The conclusion is that the white alone is considered tzitzit (and therefore its fulfillment is a deficient fulfillment of the overall commandment of tzitzit, and not merely of a commandment of the white thread), while tzitzit is also the whole. No parallel structure appears in the verses dealing with tefillin (indeed the term tefillin is not a biblical term at all).

It thus emerges that there are three main approaches among the medieval authorities regarding the relation between the parts of the commandment and the whole: Maimonides holds that both the question of ‘indispensability’ and the question whether the parts join into a whole (the distinction between 2a and 2b) are relevant to the enumeration of the commandments. Nachmanides holds that only ‘indispensability’ is relevant (he does not distinguish between 2a and 2b). Behag, by contrast, holds that even the blue and white in tzitzit are counted as one commandment, as are the arm-tefillin and the head-tefillin. Nachmanides explains that, according to Behag, ‘indispensability’ is apparently not a relevant parameter for the enumeration of the commandments, and only the question whether the parts are integrated into one matter (the difference between 2a and 2b) is relevant from his perspective.[12]

C. A Note: Why Parts of a Commandment That Are Neither Mutually Indispensable Nor Integrated Are Not Two Commandments

Thus far we have almost ignored the existence of category 1, since that category describes two different commandments, not parts of one commandment. Is there any difference between it and category 2a? In other words: why does category 2a belong to the discussion of parts of a commandment? Why do we not simply say here that these are just two different commandments? More concretely: are tefillin two parts of a commandment that, for technical reasons, are counted separately, or are they two different commandments with no connection between them?

From the wording of Rabbi Abraham in his above-mentioned responsum to Rabbi Daniel the Babylonian, it clearly follows that we are dealing with parts of one commandment. The same also appears from Maimonides’ own wording. It is only for this reason that Maimonides was troubled at all by the question why they should not be counted as parts of a single commandment.

It is possible that the explanation for why the two tefillin are parts of one commandment (even though they are counted as two commandments and are not mutually indispensable) lies in Nachmanides’ argument cited above. The two tefillin indeed operate independently of one another (one to internalize the commandments in the heart, the other to internalize them in the mind), yet both are intended to achieve similar goals that join together. Each part is not dependent on the other, but both together still possess the significance of a whole. For Nachmanides, that suffices to define them as one commandment. For Maimonides, they are indeed counted as two commandments, but he still holds that these are two independent parts of a commandment, and not two entirely separate commandments. That is why he also addresses the commandment of tefillin in his discussion of this principle.

A further possible ramification of viewing tefillin as independent parts of a commandment (2a), rather than as two separate commandments (1), may concern the dispute among the medieval authorities over whether tefillin do not preclude one another even when both are available to him and yet he puts on only one of them.[13]

D. Conventionalism and Essentialism with Respect to Concepts[14]

We have identified several types of relation between the parts of a commandment and its overallity, and we have seen how this relation is expressed in the enumeration of the commandments according to the various medieval approaches. We must now move forward and clarify the logical foundations of this picture. Our suggestion is that the basis of the matter lies in the logical attitude toward concepts and theoretical ‘entities.’ We will therefore preface the discussion with a brief philosophical excursus, and afterward return to the disputes among the medieval authorities presented above.

The pair of terms ‘substance’ and ‘accident’ has two meanings, and the two are interrelated; both are relevant to our subject: a. ‘Substance’ is the thing itself, the entity. ‘Accident’ is what happens to it, or what is attached to it. These are its characteristics or properties. b. The distinction can also be made among the various properties or characteristics of the substance itself: some properties belong to it essentially, that is, they are part of its essence; others are attached to it only accidentally, and in that sense they are ‘accidental’ for it.

When one speaks of ‘substance’ and ‘accident’ with regard to concepts (rather than substances), there is an intuition that only the second meaning can be relevant. This intuition rests on the assumption that concepts have no being. They are not existing things. According to this assumption, concepts are fictive constructs, produced through agreement among human beings. This is the conventionalist view of concepts. According to this view, the concept is a cluster of characteristics to which the speakers in a community have, by agreement, attached a common linguistic term that describes the whole.[15]

For example, let us take the concept ‘a democratic state.’ According to conventionalism, this concept has no entity, and it is created by human agreement. There are indeed democratic states in the world, but their democracy is not a real entity. It is a property of theirs, but is not itself an entity. Of course, this concept has characteristics (properties), some essential and some accidental. For example, the feature that the state’s name begins with the letter yod is accidental, whereas the features of its system of government—such as the election of governing institutions by all citizens—appear essential. A state whose name begins with another letter would still be democratic; by contrast, if it has no elections for its governing bodies, it would not be considered such.

A definition of a concept will contain only its essential properties, not its accidental ones. These are the necessary conditions for constituting the concept. There is no point in introducing accidental components into the definition, for even in their absence we are still dealing with the very same concept.

According to the conventionalist account, it would seem at first glance that one cannot change the definition of a concept, since such a change in an essential property means turning the original concept into a different one. A different definition points to a different concept. Likewise, it would seem that conventionalism also does not allow disagreement concerning the meanings of concepts. Anyone who defines the relevant concept differently is simply dealing with a different concept. The dispute thus becomes purely semantic.

Yet there is a strong intuition that definitions of concepts can indeed be changed, and that one can argue about them. Over time, concepts can take on a different character from the one they had in the past. Likewise, it is entirely possible that in different places the very same concepts will be understood in different ways.

For example, ‘morality’ requires different—and at times even opposite—actions in different societies and at different times. Some will say that morality requires aborting a sick fetus, while others will strongly oppose this and claim that it is murder. The same is true with respect to homosexual relations, and so forth.

Is there here a disagreement about the meaning and characteristics of the concept ‘morality,’ or are we simply dealing with two different concepts? It seems that if the concept ‘morality’ were merely a bundle of characteristics and nothing more, then there would be no room for any dispute. It would be clear to both sides that they are dealing with different concepts, and the confusion would stem only from the fact that they happen to use the same linguistic term to describe them. But, as noted, there is a strong intuition that this dispute is real, and that it concerns the meaning of one and the same concept. How can we explain this? What is there in a concept beyond its characteristics and properties?

To explain this, we must assume that the concept ‘morality’ has selfhood. There is an entity (or Idea) called ‘morality,’ and the dispute concerns its characteristics: from one point of view it appears in one way, and from another angle it is perceived differently. But both sides are arguing about the same concept itself, and therefore the dispute is real and not merely semantic.

This is essentially a Platonic approach. According to this view, there exists a world of Ideas, in which abstract Ideas exist, such as horse-ness, redness, kindness, triangularity, and so on. The concrete horse that exists in our world is the expression of the Idea of horse-ness that exists in the world of Ideas. The Idea is by its nature perfect, but the concrete horse can sometimes appear in a deficient way. When the horse is lame, or deformed, or simply imperfect, we will still continue to refer to it as a horse. The reason is that the Idea of horse-ness is indeed perfect, but its appearance in concrete beings can sometimes be deficient. In such cases we will say that some of the characteristics of the Idea do not appear fully in the concrete object.

Thus this intuition conceals behind it a position that is not conventionalist but essentialist. The conventionalist holds that the set of characteristics is what constitutes the concept. The essentialist, by contrast, maintains that a concept also has an essence, some ontic background, which the set of characteristics merely attempts to describe but does not constitute. The concept exists prior to its characteristics, and beyond them (and perhaps even without them). In other words: the concept is that which is described by those characteristics, not the collection of characteristics itself.[16]

We have seen that conventionalism does not allow us to view concepts with different characteristics as different expressions of the same concept (for the concept is the totality of its characteristics). Essentialism, by contrast, leads us to the conclusion that even the absence of a property—and even an essential property—can still describe that very same concept under different circumstances. The identity of the new concept with its predecessor can be preserved despite the change in characteristics (if its ‘substance,’ or ‘matter,’ is preserved), since the characteristics are not the constitutive basis of the concept, but only something attached to it.

E. Essentialism and Conventionalism in Jewish Law

We now return to the meta-halakhic discussion conducted above, and examine it in light of what has just been said. We saw that the initial inclination of all the medieval authorities is to say that when two features of a commandment are not mutually indispensable, they are presumably two different commandments (case 1). The reason is that we tend to think that every part of a commandment is essential, and therefore if it is missing, this is no longer the same commandment. From this it follows that if some part of a commandment does not preclude its fellow, then necessarily these are separate commandments.

The logical basis for this is apparently a conception of halakhic conventionalism. The commandment is not a real entity, but a theoretical concept that is nothing more than the sum of its parts; therefore, when we remove some part of it, the commandment in question does not exist at all. For example, tzitzit is composed of blue and white. According to the conventionalist conception, there are indeed concrete objects called ‘tzitzit,’ but the concept (or Idea) ‘tzitzit’ itself has no substantial reality. It is nothing more than a whole constituted by a certain combination of tying blue and white threads in the manner the Torah defined in the section of tzitzit. If we place only blue or only white on the corner of the garment, then this is not tzitzit in its halakhic sense, for its full definition is not realized. And if Jewish law nonetheless determines that placing only white is possible, then the conclusion is that the white component is a commandment separate from the blue one.

But as we have already seen in the verses themselves, placing only white, without blue, on the garment entitles it to the designation ‘tzitzit.’ There is not merely a determination that there is value to the commandment of the white thread without the blue; there is also a determination that the white alone is itself tzitzit. This necessarily pushes us toward an essentialist conception, according to which the concept ‘tzitzit’ indicates some entity (a Platonic Idea). We can now discuss whether white alone also counts as an appearance (albeit a deficient one) of that concept, or whether, because the appearance is so different, it is clear that it is no longer the same concept. Thus, at least with respect to the commandment of tzitzit, the conventionalist picture seems incorrect. For this reason, all three medieval authorities with whom we dealt count the white and the blue together as one commandment, even though they are not mutually indispensable.

By contrast, the concept ‘tefillin,’ as noted, does not appear in Scripture. Nachmanides and Behag also see tefillin as one inclusive entity (and they too rely on a biblical hint, according to which both serve the same purpose: so that the Torah of the Lord may be in your mouth, ‘so that the Torah of the Lord may be in your mouth’), and therefore they count them as one commandment. According to Maimonides, by contrast, they are counted as two commandments.[17]

F. The Relation to the Nominalism-Realism Question

Here it is worth commenting on the relation to another issue that has already been discussed in the meta-halakhic literature.[18] Many have already distinguished between nominalism and realism with respect to Jewish law. According to the nominalist approach, Jewish law does not refer to reality or reflect it (according to this approach, the eruv, for example, is not an expression of the existence of any real partition, but a fictive halakhic construction). The realist approach, by contrast, sees in eruv, or in ritual impurity, kinds of (spiritual) realities that produce the halakhic phenomena (repulsion or displacement in the case of impurity, and partition in the case of eruv).

Many link the nominalism-realism question to the issue of the reasons for the commandments. If the conception is indeed realist, then the real plane behind Jewish law is the reason for the law itself. For example, nominalism views the laws of impurity as rules constituted by the Torah, with no significance beyond the commands and laws of impurity and purity themselves. The realists, by contrast, hold that behind the concepts of impurity there stands a defective reality from which one ought to distance oneself, and that the laws of impurity were created for that reason.

Thus the distinction between nominalism and realism lies outside the halakhic sphere, since it deals with meta-halakhah and the reasons for the commandments. By contrast, the distinction we have raised here lies wholly within the halakhic world. We are dealing with the meaning of halakhic concepts as such, not with their relation to external reality. Our question is whether a halakhic concept has existence as some Platonic Idea, or whether it is nothing more than a combination of properties and definitions.

Let us sharpen this distinction by examining the commandment of tzitzit. After the command regarding the white and the blue, the verse returns and states: and it shall be a fringe for you. At first glance, this is a rather strange formulation. It suggests that the term ‘tzitzit’ is already familiar to us, and that the Torah is merely commanding us to perform the commandment in a certain way—this way and not another—as if to say: this shall be tzitzit for you, and not other forms you might have imagined. If the content of these verses itself were the definition of the concept ‘tzitzit,’ then this verse would be superfluous, for the features (four threads, white and blue, tied to the four corners of the garment) would constitute the concept tzitzit and not merely direct it. What is the point of saying that this combination shall count as tzitzit for us, if it is only a definition of purely semantic significance?

The required conclusion is that the Torah here is not inventing and creating the concept ‘tzitzit,’ but commanding us regarding a particular mode of performing a familiar act, or creating a familiar object. Thus, for example, according to Ibn Ezra on this verse, ‘tzitzit’ means a distinguishing sign, something like a banner or emblem, as in the forelock of the head (the forelock of the head). According to Ibn Ezra, then, the Torah commands us that the combination defined in these verses—and that combination specifically—is what shall count as tzitzit, that is, as a symbol, and not any other combination.

At first glance, we seem here to be returning to the plane of the rationale of the verse. Yet precisely in the context of tzitzit it is easy to see that this is not precise. As is well known, the commandment of tzitzit is not a positive obligatory commandment but a conditional commandment,[19] that is, only when one wears a four-cornered garment must one place tzitzit on it. Without that garment, there is no obligation to symbolize and distinguish ourselves. Indeed, strictly speaking there is no obligation at all to wear a four-cornered garment. Thus the Torah does not impose on us an obligation to symbolize and mark ourselves by means of tzitzit. It merely defines that this collection of components is the symbol (the tzitzit), and if one wears a four-cornered garment, then one must mark oneself specifically in this way and not otherwise.

In this sense, the question we have raised here is not connected to the rationale of the verse; rather, it is an internal halakhic question (one that does not concern the relation between Jewish law and the external world). We are defining the halakhic concept ‘tzitzit,’ not its relation to an idea or fact found outside Jewish law (that is, to the everyday concept ‘tzitzit’). The claim that tzitzit is not the aggregate of its components but has something beyond them is directed not toward the rationale of the verse, but toward the definition of Jewish law from within itself.

In other words: and it shall be a fringe for you establishes a relation between the commandment of tzitzit and the everyday concept of tzitzit (symbol). Our inference from the fact that the white alone is also called ‘tzitzit’ concerns another level: the definition of the halakhic concept ‘tzitzit.’

In analytic philosophy, a distinction is commonly drawn between CONSTITUTIVE systems of rules and REGULATIVE systems of rules.[20] A constitutive system of rules is a set of rules that defines the field in question; hence one who does not act in accordance with those rules is simply not acting within that field. For example, the rules of chess define the game of chess. One who plays according to other rules is simply not playing chess, but another game. By contrast, a regulative system of rules is a set of rules that directs activity within an already existing field (but does not constitute it). For example, traffic laws do not constitute the field of driving, or of walking in the public domain, but merely regulate it. Someone who does not act in accordance with traffic laws is indeed breaking the law, but it is still correct to say that he is engaged in driving.

Inspired by this philosophical terminology, we can say that the halakhic definition of tzitzit does not constitute the concept ‘tzitzit,’ but regulates it. Someone who does not make tzitzit in the way the Torah defines will still be engaged in producing tzitzit or using it in its everyday sense (for he is still expressing himself by means of some emblem or banner), but he will not be doing so correctly. These claims all belong to the nominalist-realist axis, and not to the axis of the reasons for the commandments.

On the internal halakhic plane with which we are concerned here (the one relating to the question of conventionalism versus essentialism), we would say that there are two partial levels of ‘incorrect’ tzitzit:

• Wearing white without blue. This is a deficient fulfillment, which is still ‘tzitzit’ in the halakhic sense.
• Wearing something else (without white and without blue, or a tassel with fewer threads, or without a knot). This is ‘tzitzit’ in the everyday sense, but not ‘the commandment of tzitzit’ in the halakhic sense.

Despite the distinction we have presented up to this point, it may perhaps be possible to see a connection between these two planes: if a given commandment is regulative and not constitutive, then it has a parallel in the world of everyday concepts. Alongside halakhic ‘tzitzit’ there is everyday ‘tzitzit.’ In such a state of affairs, it is more plausible that the halakhic concept is not the aggregate of its components, but rather an expression of the everyday concept on the halakhic plane. The everyday concept ‘symbol’ finds halakhic expression in tzitzit. That is, from here it is easier to understand the claim that the halakhic concept ‘tzitzit’ exists beyond the sum of its components (its halakhic characteristics). It is the halakhic symbol, and its halakhic characteristics are only its properties. Of course, the converse is not true: even commandments that have no parallel in the everyday world can possess an essentialist character. In such a case the halakhic concept would have no everyday counterpart.

Many commandments are intuitively perceived by us as constitutive, and closer examination reveals that they are nevertheless regulative. In the next section we will briefly see another example that sharpens the relation between the questions and the conclusion regarding halakhic essentialism.

G. Another Example: Kiddushin and Divorce

There are cases in which a constitutive approach to the command seems called for, as for example in the laws of marriage. It appears that the Torah concept of marriage is a constitutive concept. One who has not effected it in accordance with Jewish law is simply not married.[21] But all this is true only in the halakhic sense. The fact that the halakhic rules are indispensable does not mean that they are constitutive in the everyday sense. It may be that a person who did not betroth a woman has in fact married her, except that this does not exist in the halakhic sense and on the halakhic plane. He is married to her on the plane of ordinary reality, but not in the way directed by the Torah. This may be what Maimonides means at the beginning of the laws of marriage, when he describes marriage as it functioned before the giving of the Torah and treats halakhic marriage as a second tier built upon it.

As noted above, we may infer from this that marriage too is not the aggregate of the components that characterize it (two witnesses, a declaration, an object worth a perutah, huppah, and the like), but that there is a concept ‘marriage’ of which all these are merely characteristics. The consequence is that sometimes an act will still have significance even if it is performed in a deficient way.

The concept of divorce also appears, quite plainly, to be a constitutive concept, due to the fact that, on the face of it, it is not a commandment at all, but entirely a halakhic procedure. One who wishes to divorce his wife finds himself bound to a defined and precise halakhic procedure. One who does not act in accordance with the procedure set out in the Torah’s commands (written and oral) is not a transgressor. He simply has not divorced his wife. Seemingly, this is precisely the characteristic of a constitutive rather than regulative rule, as we saw above.

Despite the above description, this picture is not so simple. In Sefer HaChinukh, commandment 579, ‘That one who wishes to divorce his wife should divorce her by means of a document,’ the conclusion reads as follows:

One who transgresses this and divorces his wife without writing her a bill of divorce in accordance with the Torah’s command and in the manner explained by our Sages of blessed memory has failed to fulfill this positive commandment, and his punishment is very great, because in law she is a married woman while he treats her as divorced; and the punishment concerning a married woman is well known, for it is among the gravest transgressions in the Torah. (One who transgresses this and divorces his wife without writing her a bill of divorce in accordance with the Torah’s command and in the manner explained by our Sages of blessed memory has failed to fulfill this positive commandment, and his punishment is very great, because in law she is a married woman while he treats her as divorced; and the punishment concerning a married woman is well known, for it is among the gravest transgressions in the Torah.)

There is a very surprising claim in these words. One who divorces his wife in such a way that the divorce is ineffective commits a transgression. We would have expected such a person not to be classified as a transgressor; rather, Jewish law would determine that the divorce is ineffective in this case, and that the woman is still his wife. The author of Sefer HaChinukh writes that not only does the divorce fail to take effect, but this person is also a transgressor.

One might have thought that the problem lies in the state that is created, in which a woman is presumed to be unmarried even though in truth she is not. Such a state can lead to transgressions. But it seems clear that this is not the intention of Sefer HaChinukh, for that would not be a transgression inherent in performing an invalid divorce, but merely a problematic situation bound up with the state that resulted. Put differently: according to that explanation, such an act may indeed lead to a problem, but cannot itself be considered a transgression. Most clearly of all, a transgression of this type cannot be called a ‘failure to fulfill a positive commandment,’ as the author of Sefer HaChinukh puts it, and certainly not the failure to fulfill the positive commandment of divorce.

It therefore seems that the intention of Sefer HaChinukh is to say that divorce is not a concept constituted by Jewish law. The Torah merely regulates divorce. What this means is that to divorce a woman does not mean to hand her a bill of divorce, but rather—as in ordinary language, and as was customary before the giving of the Torah (see the beginning of Maimonides’ laws of marriage)—to send her out of his house. When a person divorces a woman unlawfully, it is not correct to say that no divorce at all has taken place here. A divorce has taken place—after all, she has left his house—but that divorce has not been effected in the manner directed by the Torah.

This is why the author of Sefer HaChinukh writes that one who divorces his wife unlawfully has failed to fulfill the positive commandment of divorce. The conclusion is that the commandment of divorce is indeed a commandment, and not merely a procedure. The commandment is that when a man wishes to divorce his wife, he must release her completely and permit her to remarry by means of a get. Thus the Torah regulates divorce rather than constituting it.

We can now go on to say, by way of example of what we noted above, that if there is indeed an everyday concept parallel to the halakhic concept, it is reasonable to continue and argue that, from a halakhic standpoint as well, divorce is an Idea and not merely the sum of its characteristics. The implication is that there are deficient manifestations of the Idea of divorce—this time on the halakhic level.

And indeed, we certainly do find such manifestations, and we will refer to them briefly. It can be shown that when a man divorces his wife in the everyday sense—that is, he sends her out of his house—even if he has not placed the bill of divorce in her hand, this too has halakhic consequences. One cannot say that nothing at all has happened here; rather, there is a state of ‘deficient’ (partial) divorce.

Once a man has set his mind on divorcing his wife, he is again forbidden to have relations with her (see Gittin 90a and Nedarim 20b). Likewise, he has no right to the produce of her melog property (see Gittin 17b–18a). The Ran (on the Rif in Gittin 8a in the Rif’s pagination), as well as Nachmanides (there), raised the possibility that he is also no longer obligated to redeem her if she is taken captive. These three are rabbinic consequences, but Rashbam (Bava Batra 146b, s.v. ‘Naflah aleha,’ and the author of Meshekh Chokhmah, Beha’alotekha) went further and wrote that from that moment he is also no longer entitled to inherit her. Rashash, in the sugya in Bava Batra there, added that he is also no longer obligated to bury her (for she is no longer his flesh, his flesh/kin). In Gittin 35b it is explained that a priest who married a divorcee is disqualified from Temple service, but from the moment he has set his mind on divorcing her (though there he must vow that he will divorce her) he regains his fitness.[22]

All these are examples of divorce carried out without most of the conditions halakhically required, and yet the woman is divorced in many respects (some of them even on the Torah level). Thus we discover that the concept of divorce, too, is not merely the sum of its characteristics. There are ‘deficient’ manifestations of the Idea of divorce. If a man divorced a woman with an invalid get, or sent her out of his house with no get at all, she is partially divorced.

This is another example of the connection between the two planes on which we have focused above: from the realism of the concept of divorce we arrive at its essentialism.

Summary

We began by analyzing the words of Maimonides and the medieval authorities on Principle 11. At the outset, it seemed that two parts of a commandment that are not mutually indispensable should be counted as two different commandments. Yet, surprisingly, we saw that all the medieval authorities are willing to accept a state in which there are different parts of a commandment that are not mutually indispensable, and nevertheless these are parts of one commandment. We explained this by saying that the fulfillment of each part involves a deficient fulfillment of the commandment as a whole, as opposed to other situations of parts of a commandment that are not mutually indispensable, in which the fulfillment of each part has independent value (a value not identical with the value of the inclusive commandment).

We saw that the initial intuition contained within it a conceptual conventionalism, whereas the position of the medieval authorities apparently conceals behind it an essentialist assumption regarding halakhic concepts, and perhaps regarding concepts in general. As we saw, the view according to which a halakhic concept reflects some abstract Idea, while the characteristics merely describe it and do not constitute it, makes it possible to understand a deficient, or partial, appearance of that Idea.

At the end of our discussion, we considered the connection between the question with which we dealt (essentialism versus conventionalism) and the question of realism versus nominalism. The second question concerns mainly the relation between Jewish law and reality, whereas the first is an internal halakhic question. And yet we saw a possible connection between these two questions, and even illustrated it through the laws of divorce. If there is indeed an everyday concept parallel to the halakhic concept (realism), it is plausible that the halakhic concept is not merely the sum of its components, but an Idea with existence of its own. Such an Idea has existence beyond its components as well, and therefore it can appear—albeit partially or deficiently—even when some of them are absent.

[1] Another example can be seen in my two articles in Tzohar: ‘The Logical Status of the Methods of Derash,’ Tzohar 12, Tishrei 5763, pp. 9-22; and ‘Induction and Analogy in Jewish Law,’ Tzohar 15, Summer 5763, pp. 23-34.

[2] See the remark on this in Hanina Ben-Menahem’s article, ‘The Individuation of Laws and Maimonides’ Sefer HaMitzvot,’ Shnaton HaMishpat HaIvri, 14-15 (5748-5749), p. 95, at the beginning of chapter 7.

[3] With the exception of a few isolated cases (such as the article mentioned in the previous note, and likewise my ‘The Secret of Yesod Mora and the Secret of the Torah,’ Dinei Yisrael, 22 [5763], p. 177), the overwhelming majority of the research and traditional literature dealing with the principles touches on the first and second principles, because they have direct halakhic implications. For a survey of the literature on this subject, see my article in the jubilee volume for Rabbi Prof. Nahum Rabinovitch, which is supposed, God willing, to appear in the coming year.

[4] Perhaps with the exception of Arukh HaShulchan, which is the only other halakhic decisor to have organized the entire range of Jewish law in a format similar to a code.

[5] The source of this number is a statement of Rabbi Simlai, Makkot 23b. Some medieval authorities did in fact question it (see a survey in Rabbi Baruch Brenner’s article, ‘The Ralbag’s Attitude toward Maimonides’ Method in the Enumeration of the Commandments,’ Ma’aliyot, 20; it also appears on the ‘Daat’ website).

[6] It should be noted that, apart from one place (in Principle 14, and incidentally also in Principle 7), Behag is not mentioned by name in the principles, but only as there is one among our predecessors who erred in this (‘there is one among our predecessors who erred in this’) and the like. As is well known, opinions are divided concerning the identity of Behag, but this is not the place to elaborate.

[7] Nachmanides’ glosses to this principle were printed in the Frankel edition, while in the standard editions they appear (with omissions and corruptions) as a continuation of the glosses to Principle 9.

[8] Rabbi Abraham ben Maimonides (hereafter: Rabbi Abraham), in responsum 5 in Ma’aseh Nissim (printed at the end of the Frankel edition of Maimonides’ Sefer HaMitzvot), sharply qualifies this criterion, for if we push it too far we may arrive at a very small number of commandments with different ends.

[9] To be sure, it can also be achieved, though less optimally, with either one of them, and therefore neither precludes the other.

[10] The second formulation seems more plausible, since in the first formulation we come very close to returning to case 1.

[11] Rabbi Abraham, in the above-mentioned responsum, writes: However, when they said regarding the lulav that there are four commandments, it is known that these are parts of one commandment. The difference between this and their statement concerning tefillin—that there are two commandments—is that the four species of the lulav are mutually indispensable; if, for example, one found three of them and was unable to find the fourth, he has fulfilled no commandment at all with what he found. Therefore the blessing can be recited only over all of them together. Tefillin are not like this… And consider their statement regarding tefillin, 'If one lacks two commandments, should he not perform one commandment?'—if one lacks two commandments, should he not perform one commandment? Nothing like this would be said regarding the lulav; indeed the law there is the opposite… (However, when they said regarding the lulav that there are four commandments, it is known that these are parts of one commandment. The difference between this and their statement concerning tefillin—that there are two commandments—is that the four species of the lulav are mutually indispensable; if, for example, one found three of them and was unable to find the fourth, he has fulfilled no commandment at all with what he found. Therefore the blessing can be recited only over all of them together. Tefillin are not like this… And consider their statement regarding tefillin, if one lacks two commandments, should he not perform one commandment? Nothing like this would be said regarding the lulav; indeed the law there is the opposite…).

According to his words, the split within the commandment of tefillin does not stem only from the wording of the midrash that calls them ‘two commandments.’ Maimonides derives it from the entire wording of the Talmudic passage he cites there: ‘If one lacks two commandments, should he not perform one commandment?’ If this were a deficient fulfillment of the whole, there would be no room for the Talmud’s astonishment at the possibility that one should not perform one part in the absence of the other. From here Maimonides proves that the Talmud understood the commandment of tefillin such that the two tefillin do not preclude one another not in the sense of 2b, but precisely in the sense of 2a (and perhaps even 1).

[12] To be sure, were it not for Nachmanides’ comments, it would certainly have been possible to learn that Behag agrees with Maimonides, except that unlike Maimonides he thinks tefillin too are integrated into one matter. See also Abraham Fיינטוך, Pikudei Yesharim: Explanations on Maimonides’ Sefer HaMitzvot, Jerusalem 5752, on Principle 11.

[13] See Beit Yosef, end of siman 26, and Or Zaru’a, siman 578, in the name of Rabbi Simhah, and Rabbi Yeroham Fishel Perla’s commentary on Saadia Gaon’s Sefer HaMitzvot, positive commandments 5-6. Compare Maimonides’ wording regarding tefillin (chapter 4, law 4), where it is clear that according to him they are not mutually indispensable in any case, with his wording regarding tzitzit (chapter 1, law 4), where it appears that they are indispensable if he in fact has both in hand.

[14] For more on this distinction, see my book Shtei Agalot VeKadur Pore’ah, Bet El 5762, second section.

[15] The analytic philosopher Saul Kripke, in his book Names and Necessity (translated from English by Avishai Revah, Tel Aviv, 1994), refers to this metaphorically as the ‘baptism’ of a concept.

[16] See my article ‘What Is Legal Effectiveness?,’ Tzohar 2, Winter 5760, pp. 71-88. There, as in the above-mentioned book, one may find examples.

[17] Chapter 3 above is devoted mainly to showing that this still does not prevent us from applying an essentialist conception to tefillin as well, and from saying that they are nonetheless parts of the same commandment, though for purposes of the enumeration they are counted as two.

[18] The credit for first applying the basic distinction between these approaches with respect to Jewish law belongs to Yohanan Silman, in several articles. For sources and further discussion, see also Rabbi Yoel Bin-Nun’s article, ‘The Search for Truth versus Halakhic Formalism,’ in Amihai Breholtz (ed.), Derekh Eretz, Religion and State, Jerusalem 5762, p. 195 (especially note 2). See also David Henshke’s article, ‘On Legal Reality in Maimonides’ Thought,’ Sinai, 92 (5743), p. 228 (he also connects this discussion to the dispute between Maimonides and Nachmanides in the second principle).

[19] A conditional commandment differs from a facultative commandment (as opposed to an obligatory one). A facultative commandment is one whose fulfillment is voluntary, whereas a conditional commandment is obligatory, except that the obligation arises only under certain circumstances that need not necessarily occur. The commandment of Torah study, beyond one chapter in the morning and one in the evening, is for many authorities a facultative commandment. So too the commandment of charity beyond one-third of a shekel per year (see Maimonides, Mishneh Torah, Laws of Gifts to the Poor, 7:5). The commandment of tzitzit, by contrast, is obligatory if the condition is met that one is wearing a four-cornered garment.

[20] See more on this in my book Enosh KaChatzir, Kfar Hasidim 5768, p. 374 and onward.

[21] For example, one who betroths a woman in the presence of only one witness, contrary to the halakhic requirement.

[22] Generally speaking, one may say that this is in fact a return to the state of betrothal (these are precisely the halakhic consequences that distinguish betrothal from marriage). Just as in the creation of marriage the process consists of two stages—kiddushin (betrothal) and marriage—so too the process of dissolution proceeds in two stages: removing her from his house (a return to her father’s house, as in the state of betrothal), and only afterward her full release. Usually this is done all at once by means of the get, but here we see a case in which the stages are separate. To do this only partially—that is the transgression of which Sefer HaChinukh speaks.