14 — Reasoning by Analogy
Growing branches from other branches
Analogy is not accepted by Shīʿī legal theorists, but among Sunnīs it is usually listed as the fourth source of law, after the Qurʾān, the Sunna, and consensus. Al-Juwaynī presents it as a tool for extending the reach of revelation. The way he imagines jurists using this tool can be seen most clearly from his comment that “the root must be established by evidence that is accepted by both parties.” He is imagining a debate, in which each party is arguing for its own school’s position on a point of law (the branch), and attempting to prove or defend its position by appealing to an analogy. For one’s argument to be convincing, one must start from a premise (the root) that one’s opponent accepts. The point of analogy is to defend law in argument, not, apparently, to construct law from scratch in the privacy of one’s own study.
Notice that al‑Juwaynī is using the terms root and branch in a new way. “Branch” still means a specific legal ruling: a statement of the form “this act by this person at this time under these circumstances has this legal value.” “Root,” however, does not mean the revealed source from which the branch is derived, or the foundation upon which the branch is built; now “root” refers to another specific legal ruling of the same form as the branch. When we speak of analogy, the revealed sources of law do not come into the picture; if we had a source text that gave evidence of a legal rule we would not need to appeal to analogy at all. In the absence of a suitable revealed text, we must take an established legal rule, and use it as the source or evidence for the legal rule we want to defend. The established rule thus comes to function as the source (the root), from which we infer our new legal rule (the branch).
For example, there is in the Qurʾān and the reports from the Prophet no text that specifically addresses the permissibility of smoking marijuana, for the simple reason that it was not a phenomenon that the Prophet or the early Muslims encountered. Nevertheless, according to al‑Shāfiʿī and the mainstream jurists who followed him, the law is absolutely comprehensive: God did not leave any aspect of human life unregulated—though he did grant humans a great deal of freedom by specifically declaring many things “permissible.” There must be, therefore, a divinely legislated legal value for smoking marijuana. One could perhaps appeal to a modern consensus on the matter, since it would be hard to find a traditionally trained jurist who permitted it, except perhaps in case of medical necessity. But even a consensus would have to be based on some evidence, according to most legal theorists. That evidence would probably be an analogy based on the well-established prohibition against drinking grape wine (khamr). Marijuana shares with wine the common characteristic of being intoxicating (yukhāmiru al‑ʿaql, it veils one’s reason), and since that common characteristic fits the condition given by al‑Juwaynī—it always entails prohibition, as far as revelation indicates, wherever it is found—then things with that common characteristic must share the legal value of grape wine, which is “prohibited.” In this analogy, “drinking grape wine is prohibited” is the root, and “smoking marijuana is prohibited” is the branch.
Gradations of obviousness
The analogy between grape wine and other intoxicants is a classic example, but it is somewhat exceptional in that the Prophet actually stated on one occasion that the reason grape wine is forbidden is that it intoxicates. This might seem intuitively obvious to us, given the stigmas against intoxication in our society, and it might have seemed intuitively obvious to some of the Prophet’s contemporaries as well. But there were disputes among early Muslims about whether other intoxicants were in fact forbidden. Date wine (nabīdh) was a more common example at the time. Some jurists declared it licit, at least in small quantities, on the grounds that the Prophet had never explicitly forbidden it. After all, there is no necessary connection between grape wine and date wine, and even less of a connection to marijuana. What if grape wine was forbidden because it was made from grapes? Or because of its color or taste? According to the logic of reasoning by analogy, Welch’s grape juice and red Coolade share common characteristics with grape wine, and what is to say that intoxication is more important than color, taste, or what something is made from? Well, in this case there was a ḥadīth report available in which the Prophet specifically stated that God was concerned with intoxication, not color or taste. For most jurists that settled the matter, and as long as one accepted the method of reasoning by analogy, then unless one could find some other stronger evidence in favor of date wine or marijuana, one had to admit that they were forbidden.
More tricky (and much more numerous) are the cases in which jurists are left to debate among themselves which characteristic, among all the qualities of an act, is the one that “brings about” its legal value, and can therefore be used to determine the legal values of other acts. In some cases the relevant characteristic seems so obvious that one can say it obviously entails the legal value. The classic example is the Qurʾānic command (Q 17:23) not to scornfully say uff to one’s parents. Obviously the point is not to do them the least harm, even so little as insulting them; hence any other act that does them greater harm, such as beating them, is a fortiori prohibited. You may recall (from Chapter 3) that some legal theorists said this was so obvious that not hitting one’s parents was actually an implicit part of the linguistic meaning of “not saying uff.” They called this an example of positive implication, not reasoning by analogy. Al‑Shāfiʿī, however, said this was not linguistic meaning but actually a form of inference, albeit a pretty obvious one. The fact that al-Juwaynī never mentions implied meaning anywhere in this short treatise suggests that he was content to treat obvious inferences such as “beating your parents is prohibited” as instances of reasoning by analogy. As we noted earlier, he was not embarrassed by the involvement of human reasoning in the construction of law. Many other legal theorists preferred to ascribe as many of their interpretations as possible to the linguistic meaning of revelation, because that gave their interpretations the authority of revelation itself, instead of leaving open to debate.
More often, the reason for a prohibition, permission, or obligation is not so obvious. In such cases, one cannot say that the common characteristic one identifies as the basis for a legal value actually entails the legal value, but only that it indicates the legal value. Even the fact that a drink intoxicates is not obviously evil, in and of itself; indeed if God had not forbidden it, drinking at least small quantities of wine might have been considered one of the beneficial and pleasurable things that God made and permitted for his creatures. The intoxicating nature of wine is therefore taken merely as a sign—a marker by which God enables us to identify a group of actions to which he desires to attach the same legal value. The sign indicates the legal value, but it does not entail it; we can easily imagine that God might have grouped actions differently, according to different characteristics, and assigned legal values on the basis of taste, color, or origin.
Another example would be the question of whether it is obligatory to pay the alms tax (zakāh) on the wealth of a child who is not yet legally responsible. One could argue that no tax is due because the child cannot be under any legal obligation at all. This was the position of Abū Ḥanīfa. Others argued that the child’s ownership of his wealth is just like an adult’s ownership of his wealth, and that this common characteristic indicates that the obligation that exists in the case of the adult also exists in the case of the child—though the obligation would presumably fall on an adult guardian, who would pay the alms tax out of the child’s wealth. This is the most common kind of analogy. It is always, by definition, disputable.
A third kind is even more open to dispute, since it is defined as a case in which two different analogies are possible, starting from two different established legal rules (roots). For example, if someone kills a slave, what compensation is due to his owner? Since a slave is a human being, one could start from the root that establishes the compensation for murder, and say that the same compensation is due to the owner of the slave. However, it is also indisputably true, in the logic of classical law, that a slave is a kind of property, just like livestock; hence one could argue that compensation should be based on the market value of the slave. How does one decide between these two roots? Is the slave like the first by virtue of the shared characteristic of being human, or like the second by virtue of the shared characteristic of being property? Some jurists decided on the latter, not because they doubted the slave’s humanity, but because there were so many other legal rules that treat a slave as property: he can be sold, he can be inherited, etc. Overall, in the minds of the jurists, all these rules made him more “similar,” in legal terms, to a piece of property than to a human being; hence compensation should be calculated as with other kinds of property.
Restraints on analogy
It is not hard to see how much this kind of interpretation is shaped by the prejudices, the imagination, the moral values, and the social presuppositions of the interpreter. Who in our society would say that a slave is more like a piece of property than a human being? Our commitment to the latter is so strong as to rule out the very possibility of slavery, at least in its most overt forms.
Al‑Juwaynī and his peers, however, do not want to make reasoning by analogy a free-for-all in which a jurist can reach whatever legal rules he wants. Al‑Juwaynī lays down several conditions that an analogy must fulfill; that is, he gives interpreters several ways to check their own imaginations, and he gives participants in a debate several ways to criticize each others’ arguments.
First of all, the branch and root must “correspond” to each other, or be “fitting.” This gives jurists some vocabulary for criticizing one another’s analogies, by declaring that they are “not fitting.” This “fit” between root and branch, however, is precisely what jurists are evaluating and debating when they assert that “case X is analogous to case Y,” so the requirement that they must “correspond” or “fit” does not provide any objective measure of the analogy’s validity. Ibn al-Firkāḥ, in his commentary on this part of the Waraqāt, says that this requirement is redundant because it simply means that the root and branch must share a common characteristic, which is true by the very definition of analogy. This first requirement, therefore, does not place any substantial restriction on the use of analogy.
Second, the parties must agree on the evidence undergirding the root rule. This sounds obvious: one’s argument would not be persuasive if one’s opponent did not share one’s premise. But it is actually much stronger than mere persuasion requires. For one thing, one must agree not only on the root rule itself, but on the evidence that supports that rule. This suggests that al‑Juwaynī thinks of analogy not simply as an extension of legal science itself, in which jurists extend their own human construct by their own reasoning; it means that al‑Juwaynī thinks of analogy as an extension of the revealed evidence on which that construct is built. This was what al‑Shāfiʿī had argued almost three hundred years earlier: analogy is simply the means God prescribed for figuring out the meaning of his speech. Furthermore, one must actually agree with one’s opponent; one cannot employ the classic tactical debating move of adopting a position one’s opponent holds, and arguing from his premises to force him to accept one’s own view, regardless of whether one believes in those premises oneself. Again, analogy is not simply an exercise in human argument; it is a method for building something true, upon the foundation of revelation itself. One must not build on a foundation one does believe to be God’s revealed will.
Third, the common characteristic on which one bases one’s analogy must always correlate with the presence of the legal value one is deducing from it. One cannot say that intoxicatingness is the common factor indicating the prohibited nature of marijuana if there exists some intoxicating substance (let’s say cocaine) that is not prohibited. And one cannot wiggle around such an objection by saying that the term intoxicating is not strictly speaking the right word for the effects of cocaine; it’s the idea that counts, not just the words. This requirement is tricky, because Islamic law is filled with exceptions, so the correlation between a characteristic and a legal value is seldom completely consistent, and it will be easy to find inconsistencies in just about any analogy. Drinking wine, for instance, is permissible if one does so under duress. Thus there is always a way to argue against an analogy, and its proponent will always have to defend himself by explaining away such inconsistencies.
Fourth, the legal value one deduces must exist precisely where the common characteristic exists. This requirement is equivalent to the prior one, and is subject to all the same argumentation. Al‑Juwaynī’s second through fourth rules thus place some pretty significant limitations on analogy, which is also to say that they all allow wide scope for disagreement and debate.
The law’s coherence
Two things strike me as noteworthy about this brief discussion of analogy: it assumes a certain coherence to the law, and like the previous devices we have studied, it provides a way to either maintain the status quo or reform the law.
The idea that God’s law is coherent and predictable—that is makes sense—is natural for scholars like the Muʿtazila, who consider it to be based on the natural qualities of acts. The reason God forbids intoxicating things, even in small quantities, is that consuming even small quantities can lead to consuming larger quantities, which can lead to harmful consequences. We know that God’s law is designed for our benefit; it can hardly be designed for God’s benefit, since he neither needs anything nor can be augmented by anything. Hence we know that as we interpret revelation, we will find stable patterns in God’s commands, corresponding to human nature and the natural consequences of each action.
Why, however, would an Ashʿarī like al-Juwaynī look for patterns in God’s laws, if he believes that God simply commands what he wishes to command, without having to follow any higher or natural standard of justice? Sometimes, in fact, God’s laws do not seem to follow any kind of discernible pattern. A number of early Muslim scholars had pointed to what they thought were utterly arbitrary and incomprehensible rules, and concluded that God’s law could not be expected to make any sense; it could only be followed. These scholars included some Muʿtazilī scripturalists, as well as the “literalist” Ẓāhiriyya who followed their lead; they argued that one should simply follow the plain meaning of God’s words, without seeking to identify what rationales or moral patterns lay behind them. Consequently, they rejected analogy entirely, and simply refused to regulate actions that were not specifically addressed in revelation. They might regulate marijuana, because the Prophet did once say that “all intoxicants are forbidden;” but if he had only explicitly forbidden grape wine, then other intoxicants would be perfectly permissible. Similarly, since the Prophet forbade the practice of charging interest (ribā, unequal amounts and delay) in trading six specific kinds of goods (gold, silver, wheat, barley, dates, and salt), they allowed ribā in other kinds of transactions. This put them quite at odds with the mainstream, who said that interest was likewise prohibited in trading other commodities, on the basis of analogy to the original six. The Ẓāhiriyya’s literalism was most unsatisfying to the mainstream jurists, and it went against the grain of the very systematic sets of rules they had worked out by means of analogy. The Ẓāhiriyya did not survive as a school of law.
Ashʿarī scholars, including al‑Juwaynī, had to find a way to imagine God’s law as coherent and predictable, without saying that God was subject to some natural or higher standard of justice. They did so by imagining that the “common characteristics” that indicate patterns in the law are not actually God’s “reasons” for legislating as he did; God requires what he wills, without having to follow any kind of rationale. Rather, the common characteristics are just signs or indicators—qualities of things that God has arbitrarily chosen to associate with certain legal values, so that we could extend our knowledge of his law without having to hear him explicitly mention every conceivable human action. Analogy does not indicate a moral coherence per se, it is just an epistemological coherence that God gave the law so that we could come to know it more easily through human reasoning.
Notice, however, that in his closing sentence al‑Juwaynī says that “the common characteristic brings about the legal value.” This sounds like he thinks the common characteristic is not just a sign, but also a cause of an act’s legal value. Perhaps he does imagine God’s law to have a natural coherence to it. Or perhaps, since he is talking about human legal science rather than God’s law per se, he just means that the common characteristic “brings about” our knowledge that an act has a certain legal value.
A tool for creating, justifying, preserving, or reforming the law?
What is the purpose of this interpretive device called analogy? On the surface, it is a means of creating law where no revealed texts exist. It extends the law, and extends the meaning of revealed texts, beyond the topics they explicitly address to topics about which they are silent.
Notice, however, that al‑Juwaynī’s rules about analogy do not support this understanding of analogy. Take for example his classification of analogy into three types. If analogy were a formal means of constructing law from scratch, we could look at any analogy and put it into one of these three categories, and then we would know which of several possible analogies should trump which other ones. As al‑Juwaynī will say below, clear analogy is given priority over obscure analogy. This classification tells us which are clearer and which are more obscure: analogy by common characteristic is the most obvious, analogy by indication comes next, and analogy by resemblance is the weakest.
But how is one to classify a given analogy? To classify it we have to determine whether it is obvious, debatable, or merely one of several good possibilities. In other words, to find out how clear or obscure an analogy is, we have to decide how clear or obscure it is. This is the same problem we encountered with al‑Juwaynī’s classifications of the clarity and obscurity of language. What we concluded then, we may also conclude here: the classification does not tell us which interpretation is better; it merely gives us a way of justifying our claim that one is better than another.
After all, unless the Prophet has explicitly stated the relevant common characteristic that lies behind a ruling, it will always be possible to suggest several possible common characteristics relating a new branch to any of several possible roots. All analogies are what al‑Juwaynī calls “analogy by resemblance:” there are always several options, and which one seems more obvious, or a better “fit,” will depend on our presuppositions about how one should relate to one’s parents, the purpose of the alms tax, the humanity of slaves, or whatever other question we are dealing with.
We have already made the same point about al‑Juwaynī’s rules for analogy: rather than providing objective criteria for evaluating analogies, they provide ways of disputing one’s opponent’s analogies. His whole discussion of analogy, therefore, seems designed more to give interpreters tools or arguing for or against rules that they or their opponents have already decided upon, than for creating new laws from scratch by simply observing the characteristics of things. His reference to the parties in the debate makes this doubly clear.
Analogy, therefore, is conceived of by al-Juwaynī as a device for extending the meaning of revelation, but is designed to function as a tool for justifying and debating interpretations.
Like many other devices we have encountered, this one has potential first of all as a means of reinforcing the legal status quo. It gives jurists great flexibility in constructing and critiquing arguments, and that flexibility can most easily serve a conservative legal vision. The treatment of slaves is a case in point: because the jurists already had other rules that treated slaves as property, they argued that in calculating compensation for their murder the law must also treat them as property rather than as human beings.
Those jurists who approved of analogy—and they were almost all the Sunnī jurists—were typically very concerned to maintain the systematic consistency of the law, and this resulted in a self-reinforcing and conservative law. Whenever new questions arose, they were resolved so as to fit into the law of prior generations.
Kecia Ali, who has written extensively on Islamic laws regarding women, has identified the classical jurists’ conception of marriage as a case in point. In her book Sexual Ethics & Islam, she points out that the law consistently treats the marriage contract as a contract of sale, and treats the marriage relationship as one of sexual ownership parallel to the ownership of slaves (in which a male owner had legitimate sexual access to his unmarried female slaves). Jurists, of course, were quite aware that marriage was not simply a sale, and wives not simply slaves. In arguing about certain details of marriage law, they tried several kinds of analogies, sometimes saying that marriage is like an act of worship (and therefore that the required marriage payment has a divinely ordained minimum value that is established by a ḥadīth), and sometimes saying that marriage is like a sale (and therefore that the marriage payment can be any amount the parties agree upon, just like the price of a commodity). The jurists recognized the conflict between these analogies, and very consciously chose to side, in many instances, with the analogy that makes marriage like sale and slavery. As Ali has forcefully argued, this logic of male ownership is so entrenched in legal thinking that to reform the law in line with feminist or egalitarian values would require a complete rethinking of the whole notion of marriage. Analogy reinforces and replicates established ways of thinking by transferring them onto new questions, and makes the law an interdependent whole that cannot be modified piecemeal without destroying its systematic nature.
On the other hand, the tremendous potential of analogy as a means of adapting the law to modern societies has also been highlighted by numerous modern Muslim thinkers. Analogy is perhaps the single aspect of legal theory that has received the most attention in the modern period, as other questions like the analysis of language have been relatively neglected. This is surely because analogy constitutes a frank recognition that legal science is a human construct, and because it allows a great deal of room for the interpreter’s own values to play a role in identifying the rationales that he or she feels must motivate the law. For an Ashʿarī like al‑Juwaynī, analogy is simply a way of analyzing signs God has used to indicate his arbitrary will; but even al‑Juwaynī gave several hints that he cannot help thinking of analogy as based on some real moral coherence behind the law. Modern thinkers, enamored (like Western ethics) with general moral principles, can find in analogy an opportunity to identify overarching rationales or moral goals that lie behind the words of the law, and extend those moral principles to new cases. In fact, those principles can even be extended to old cases, since a new analogy, based on a new understanding of the rationale behind the law, can always be used to challenge an old analogy, or can function to modify the meaning of a revealed text. Once again, it is the gradual evolution of Muslim social and cultural norms that will govern how much change can be introduced by means of analogy. The texts of revelation are too few, their meaning is too malleable, their authenticity is too debatable, consensus is too ephemeral, and analogy is too open-ended, for the law to remain bound to its past. It is bound only by the imagination of Muslims, who continually reconstruct it generation after generation.
For an overview of the technicalities of reasoning by analogy, see Mohammad Hashim Kamali, Principles of Islamic Jurisprudence, rev. ed. (Cambridge: Islamic Texts Society, 1991), ch. 9.
For an alternative explanation of how the Ashʿariyya understood the coherence of the law, see Anver M. Emon, Islamic Natural Law Theories (Oxford: Oxford University Press, 2010).
Kecia Ali provides a fascinating example of how profoundly the logic of analogy shaped Islamic marriage law in Sexual Ethics & Islam: Feminist Reflections on Qur’an, Hadith, and Jurisprudence (Oxford: Oneworld, 2006).