5 — Types of Speech
Conceptions of speech
By now we are not surprised to find that before al‑Juwaynī starts to analyze the language of revelation, he feels he must define what he is analyzing. He is not analyzing language per se—the system of word definitions and syntactical rules that govern speech. Revelation is not language but speech—particular utterances composed of particular strings of words following the rules and word definitions of a particular language (which happens to be Arabic). So in describing the stuff of revelation, he is mainly describing speech (though he will mention nonverbal actions of the Prophet as well).
If he had chosen to give a technical definition of speech, al‑Juwaynī probably would have defined it as “an attribute (or meaning) subsisting in the soul (or self) and expressed by articulated sounds and ordered letters.” This reflects the Ashʿarī doctrine of “inner speech,” which understands speech (very roughly) as the idea that the speaker has in mind, while words are merely the “expression” of that mental idea. This definition of speech was used to support the Ashʿarī doctrine that God’s speech is an eternal attribute or meaning (the term maʿnā means both) subsisting in God’s essence, just like his attribute of knowledge. This doctrine had important hermeneutical implications: it meant that God’s law was a timeless reality rather than a response to particular historical circumstances, which made the context of revelation less relevant to its interpretation; and it meant that the words or revelation were not themselves God’s speech, but merely pieces of evidence that might or might not indicate the existence of a command or a prohibition subsisting in God’s essence. The Muʿtazila had a very different conception of God’s speech: it is the very sounds and letters and words that human beings actually hear when the Qurʾān is recited; hence it is created in time, and there is no gap between God’s words and his speech—the imperative “pray!” is itself God’s command, not evidence from which to infer God’s command. This Muʿtazilī definition of God’s speech kept meaning closely identified with words, whereas al‑Juwaynī’s Ashʿarī doctrine of God’s eternal speech created a large gap between words and their meanings, which could only be crossed through interpretive reasoning.
Speech is composite because it conveys information
Al‑Juwaynī, however, does not give such a technical definition of God’s speech. Instead he speaks of speech being composed of words, as though he were a Muʿtazilī. He knows better than that, but he avoids theological subtleties and jumps right to the words of revelation, which are technically just the expressions of God’s inner speech, and calls them speech just to keep things simple. Let us watch and see whether that simplification will lead him to think more like a Muʿtazilī!
The point he wants to make about speech-expressions is that they never consist of just one word, but always two or more. Classical commentaries like to illustrate the four combinations of words that he mentions. An example of speech consisting of two nouns is “Zayd standing” (zayd qāʾim), which is how one says “Zayd is standing” in Arabic. (The participle “standing” is a kind of noun meaning “a standing one,” so the sentence really has the form “Zayd [is] a standing one.” Zayd is the name Arab grammarians use for stock examples, rather like John Doe.) “Zayd stood” (qāma zayd) is an example of a noun with a verb. “[He] did not stand” (mā qāma) is an example of a verb with a particle: the particle is “not” and the verb is “he stood,” but the pronoun “he” is not expressed because the form of the verb already shows that it is in the third person masculine singular. Some Arab grammarians count the implicit pronoun “he” as a word even though it is not expressed, but al‑Juwaynī does not. Finally, “O Zayd” (yā zayd) is an example of a noun with a particle. Of course, most utterances consist of more than two words, but if their words are grouped together into noun and verb phrases, they should all be reducible to one of these four combinations. “My friend Zayd stood on the table” has the same basic structure as “Zayd stood.” Some commentaries contend that not all speech can be reduced to these four combinations, but that will not detain us, because that is not al‑Juwaynī’s point. Legal theory is not concerned with diagramming sentences; it assumes that we already know the rules of Arabic grammar, and that we can distinguish a subject from a predicate and a verb from a preposition.
Al‑Juwaynī’s point is simply that speech is always composed of more than one word, and this claim reinforces his idea that knowledge is always knowledge about something. Just as awareness of appleness does not count as knowledge, so the word “Zayd” all by itself does not count as speech. To know something, we have to know that “the apple is sweet,” and to say something we have to say that “Zayd is standing,” or, more pertinently, we have to say that “drinking wine is forbidden.” Speech, like knowledge, is about the way things are; it gives information about the properties of things, and most importantly about the legal values of actions. So speech has to have at least two components: it can’t just refer to Zayd; it has to say something about Zayd, or convey some meaning in addition to the concept of Zaydness.
The Muʿtazilī ʿAbd al‑Jabbār said that speech only had to consist of at least two letters, not two words. He, therefore, could count the lone imperative verb “stand!” as speech, whereas for al‑Juwaynī to count this as speech he would have to fudge, counting the implicit pronoun “you” as a separate word (even though he did not count “he” as a separate word in “[he] did not stand”). For al‑Juwaynī, the most basic function of speech is to express indicative statements, which normally requires either a subject and a verb (Zayd stood), or a subject and a predicate (Zayd [is] standing). We might ask how “O Zayd” conveys information, or fits al‑Juwaynī’s statement in his Burhān that speech is either subject and verb or subject and predicate. Al‑Juwaynī does not answer that question, but one commentary argues that the real meaning of “O Zayd” is “I am calling Zayd,” which is a subject / verb sentence and does convey information. The real job of speech, in al-Juwaynī’s mind, is not to bring about new situations performatively, or transform people, or conjure up images; it is to convey information. “Stand!” is just shorthand for “standing is obligatory.”
Speech is essentially indicative, not performative
This reduction of speech to the indicative mood, and of revelation to its informative dimension, is reinforced by al‑Juwaynī’s next point: speech has four basic functions, command (“stand!”), prohibition (“do not sit!”), statement (“Zayd is standing”), and question (“is Zayd standing?”). This is not about grammatical moods (indicative, imperative, subjunctive) but rather about the performative effects that speech can be intended to accomplish: it can impel someone to act or not to act, or it can convey or solicit information. Many commentators find this list incomplete, and like to add other functions of speech such as expressing a wish, making an offer, or taking an oath; in fact these three were eventually added to al‑Juwaynī’s text, and later legal theorists came up with even longer lists including such speech-functions as threat, promise, intimidation, and even creation—as when God says to something “be” and it is.
Al‑Juwaynī, however, is not trying to be comprehensive. He assumes we already know about different grammatical forms and moods, and the various goals speech can accomplish. The important thing for al‑Juwaynī is that not all of revelation comes in the form of ready-made statements about the legal values of acts; some of it is in other forms, especially commands and prohibitions, which will have to be translated into indicative statements. That will be the goal of Chapter 6 on commands and prohibitions. For legal theorists the only modes of speech that really matter are commands and prohibitions and statements; some even categorized prohibitions as negative commands, and thus ended up with only two basic speech categories.
By stating the obvious point that not all speech consists of indicative statements, al‑Juwaynī is simply legitimating the task of translating revelation into the indicative. The fact that he does not bother to mention more speech-functions than he does suggests to me that he has not really thought about how radically he is truncating Qurʾānic language.
I pointed out in Chapter 2 that God’s speech sometimes serves not to convey information but to create things, and that al‑Juwaynī could have imagined law as a set of binding interpersonal relationships of responsibility, gratitude, and submission, brought into being by the Qurʾān’s words about how people should relate to each other. If the purpose of revelation was understood this way, law, rather than consisting of statements of the form “this act by this person at this time under these circumstances has this legal value,” would take the form “this person stands in this relationship to this other person.” Those relationships would be discovered by tracing the history of human and divine speech and interaction through which the relationships were performatively brought into being.
Alternatively, we could imagine that the function of God’s speech is not to convey information but to transform people’s character. That would certainly fit the tone of the Qurʾān, and it would result in a very different kind of ethical or legal system, in which Muslims evaluate the development of their personal virtues, rather than the consequences of their actions. Legal science would then take the form “this person has achieved this degree of patience or that degree of generosity.” At present one finds such discourse not in law, but in the Sufis’ accounts of the stages of the soul’s journey to God.
Alternatively, we could think like modern existentialists, and focus on the way God’s speech challenges people, confronts them, and demands choice and response. This too fits well the tone of the Qurʾān, and it would result in an ethic that focuses not on consequences but on the authenticity of each individual’s response to revelation. In such an ethic, the same action could be evaluated quite differently depending on one’s reason for doing it: “this person’s choice to drink beer last night constituted a spineless sellout to peer pressure, whereas this person’s choice to drink was a courageous act of solidarity with downtrodden laborers….” You can see how radically different Islamic law or ethics would be if it approached God’s speech as a direct personal challenge rather than as a timeless and impersonal statement about the eternal consequences of human actions. Clearly al‑Juwaynī is interested in legal information, not existential challenge. Yet even he, at the end of the book, will hint that it is the integrity of one’s choices, rather than their correctness, that determines their validity in God’s sight.
In any case, whatever we might say the performative effect of God’s speech is, simply choosing to regard it as performative rather than merely informative would force us to look at revelation as an interpersonal event rather than as a text. That would make it just like the nonverbal revelatory actions al‑Juwaynī will discuss below. As he will say in Chapter 7, nonverbal actions are not statements expressing general legal rules; they are just individual events, from which any general law must be inferred by an argument about the relevance of that event to other situations. If even verbal revelation was viewed as such a historical event, it would not be viewed as a set of general propositions, but as a record of God’s interaction with a particular people in a particular time and place. The work of diligent inquiry would then be to investigate what we can learn about the present from God’s past interactions with people, rather than to apply the words God used in those past interactions directly to present situations.
This is exactly how some modern Muslims have tried to reform Islamic law. Fazlur Rahman (d. 1988), for example, is famous for his “double movement” theory of interpretation: first one investigates what God was trying to accomplish through his interaction with people at the time of the Prophet’s revelations, then one investigates how that same moral purpose can best be accomplished in the here and now—even if the answer is quite unlike the specific commands God gave in the seventh century. Rahman did not say he was looking for the “performative effect” of God’s speech, but he treated God’s speech as a historical event rather than as a timeless statement, which is much the same thing.
These alternative ways of understanding the language of revelation, and of imagining Islamic law or ethics, would have seemed ludicrous to al‑Juwaynī—but only because he was living at a time when Muslim jurists had already made the choice to regard revelation as indicative statements of legal values, and had come to take this for granted. It is worthwhile trying to imagine other possibilities, because we do not really understand the significance of al‑Juwaynī’s choices—of his definitions of speech and knowledge and legal science—until we imagine what it would have meant if he had chosen differently.
A flexible literalism
Al‑Juwaynī has said that speech is always composite, and that its form is not always indicative. He now adds that speech can make use of language in two different ways. Ḥaqīqa and majāz are two different ways of using the system of word definitions and grammatical rules within which speech operates. These two words are usually translated literal and figurative, and I have retained those terms in my English translation, but they express only part of what is meant. Only al‑Juwaynī’s fourth type of majāz is what we would call figurative speech: to say that a wall “wants” to collapse is figurative in the sense that it personifies the wall, when we do not really mean that the wall has a will, but only that it is about to collapse. To say “there is nothing like God’s likeness” instead of “there is nothing like God” is what we would call redundant, not figurative. To say “ask the town” might be taken as a figurative personification of the town, as though it could answer; but to al‑Juwaynī’s mind this is not personification but ellipsis (deficiency): the words “people of” are understood but have been omitted. To use the word “hollow” (a depression in the ground where one goes to screen oneself from view while relieving oneself) instead of the word “excrement” is what we could call a euphemism. All these manners of speaking deviate somehow from ordinary and straightforward speech, and thus “transgress” the normal rules of language; so when speaking of majāz in this broad sense in what follows I will call it “transgressive usage,” not in the sense that it is wrong, but in the sense that it violates normal usage (ḥaqīqa, which I will also call “normative usage”).
By giving obvious examples of transgressive usage, three of them from the Qurʾān itself, al‑Juwaynī is simply affirming that revelation does sometimes use language in transgressive ways; hence it is legitimate for an interpreter to interpret revelation figuratively or transgressively. He is not saying that revelation can be interpreted allegorically, or as a metaphor for something completely different from what it appears to be talking about: “wash yourselves before congregational prayer” cannot be transformed into a command to purify one’s conscience in order to achieve a group mystical experience. Al‑Juwaynī simply means that in day-to-day speech words are routinely used in well-recognized ways that stretch the rules of ordinary usage just a bit.
Notice, however, that none of al‑Juwaynī’s examples of transgressive usage have anything to do with law. This is no accident. Most legal theorists recognized that the Qurʾān sometimes transgresses the rules of normative usage; those few who denied this were mostly from the Ḥanbalī and the now-defunct literalist or Ẓāhirī school of law. But the concept of figurative interpretation fell into disrepute, because of the way Muʿtazilī theologians used it to explain away what the Qurʾān said about God predestining people to hell and such. Most legal theorists, therefore, preferred not to call their legal interpretations majāz. We will see how al‑Juwaynī avoids that term, and manages to make most of the Qurʾān’s legal language entirely literal. He will argue that many words have several possible normative or literal meanings, and this enables him to pick and choose among meanings, while still being able to say he is interpreting revelation literally. All of Islamic law is literalist, but that doesn’t mean it adheres rigidly to the single most basic meaning of each word. Only a few schools tried to do that: the Muʿtazila and especially the Ẓāhiriyya. Ironically, they were the most rationalist of legal theorists, and ended up with the most rigid interpretive theories. But they both died out among Sunnīs—in part because their literalism was too rigid. Al‑Juwaynī’s literalism was an expansive one, which managed to cram all kinds of interpretations under the umbrella of literal interpretation. His discussion of figurative or transgressive usage, therefore, is not all that important for law. The part of this section that matters most is what al‑Juwaynī says about literal or normative usage.
The author’s intention governs meaning
Notice first of all that al‑Juwaynī defines normative speech as speech that “uses” language in a certain way. Language is an abstract system that assigns every word and grammatical form one or more meanings; and every time someone speaks, he chooses which words and grammatical forms to use, and uses each one to express one or another of its possible meanings (or even to express some other slightly different meaning, in which case he is using it transgressively, as when he uses “the town” to mean “the people of the town”). He can also choose to use a word “directly” (ṣarīḥ), as when he uses “marriage” to refer to a marriage contract, or “indirectly” (kināya), as when he uses “marriage” to allude to sexual union; but al‑Juwaynī does not mention this distinction.
This way of picturing the relationship between language (an abstract system) and speech (a particular use of that system) assumes that the meaning a word has in a particular utterance is determined by how the speaker decides to use the word. In other words, meaning is determined by the speaker’s intent. If by “hollow” he intends to express the idea that something has a cavity inside it, that’s what “hollow” means in that sentence; if he means a depression in the ground, that’s what the word means; and if he uses it as a euphemism for excrement, then that’s what it means, regardless of how many other possible meanings are listed in the dictionary. As long as the speaker stays within the bounds of literal or recognizably figurative usage, his words mean what he intends them to mean, not what the dictionary says they can mean, or what the hearer understands them to mean.
The idea that meaning is governed by the speaker or author’s intent has been challenged by Western structuralist hermeneutics, which says that the meaning of a text is completely determined by the words and structure of the text itself, and that the author retains no control over the meaning of a text once it is written. Poststructuralist and postmodernist hermeneutics emphasize the reader’s role in determining the meaning of a text in light of his or her own context and word associations. But al‑Juwaynī, like virtually all premodern Muslim and Western thinkers, assumes that meaning is governed by the author’s will to express one idea rather than another. There are limits to what a speaker can intend a word to mean: he cannot intend a meaning that no one else thinks that word can have, and he cannot intend a word to mean two things at once, or at least not two contradictory things at once. But he can literally make white mean black (because that happens to be one possible meaning of white, bayḍ, in Arabic); or more significantly he can use an imperative when he doesn’t really want someone to act (as in “go ahead, make my day!”); or he can even say “obligatory” when he really means “recommended.”
To say that meaning is determined by the speaker’s intent places upon the interpreter the burden of discovering what God intended by each word in revelation. It means that the human construct of legal science is subject to God’s will, not human whim. But how can al‑Juwaynī ensure that this human construct is in fact dictated by God? The Muʿtazilī ʿAbd al‑Jabbār could appeal to God’s wisdom and justice to argue that when God speaks he must make his intended meaning clear; hence an interpreter can always assume that he already has available to him all the evidence he needs to figure out what God intended by a particular word or phrase. Therefore, if he finds no evidence that a word was meant figuratively, he can be certain that it was meant literally; if he finds no evidence as to which of several possible literal meanings God intended, he can be sure that God meant all of them at once (unless they are contradictory, in which case God must have meant them all as equally valid alternative interpretations). This gave ʿAbd al‑Jabbār a great deal of confidence in legal science: even though it is a human construct, we can be sure that its conclusions are correct, as long as it makes proper use of all the available evidence. Al‑Juwaynī, however, does not believe God is subject to any standard of justice, so he has no way to prove that God must reveal his intent clearly. He believes that God’s words must be interpreted in accordance with the intent and inner speech-meaning that exists in God’s mind (or, more precisely, in his essence); but that inner speech-meaning can be known only through the words that express it, and there is no way to know for sure that God has given the interpreter all the evidence needed to determine which of its various possible meanings God intends each word to express.
Thus al‑Juwaynī assumes that speaker’s intent governs meaning, but his treatise never discusses intent or how to know it;1 it only discusses the mechanisms by which meaning is constructed by human interpreters. Will al‑Juwaynī’s hermeneutic be able to guarantee that what fallible human reasoning constructs is in fact God’s intent? Or will he give the interpreter full control of the meaning of revelation, while simply claiming that its meaning is governed by God’s intent, so as to lend human interpretation the aura of divine authority?
The slipperiness of language
The next thing to notice is that al‑Juwaynī’s definition of normative usage assumes a certain view of language: words are arbitrary signs invented to denote non-linguistic ideas and things. Exactly how language was first invented is a matter of dispute among legal theorists; most ended up saying that the Arabic language (or the first language, whatever it was) was invented either by God, who taught Adam “the names of things” (see Q 2:31), or by some primordial human community that got together and assigned names to things by pointing at them. Either way, the words are arbitrary—their sounds have no natural connection to their meanings—but they denote real objects and properties that exist independently of language. To understand a word, then is to know which objects or properties it denotes in Arabic, or which ones a particular speaker intends it to refer to in a particular utterance.
This view of language as signs for things may seem self-evident, but it has been challenged by structuralist and postmodern theories of language. The linguist Ferdinand De Saussure argued that to know the meaning of “car” is not to know which objects are cars, but to know that “car” differs from “truck” in size and differs from “boat” in having wheels. In other words, signs do not refer to things, they refer to other signs. Language was not invented by pointing at things, but by differentiating one sign from another.
Postmodernists have pointed out that De Saussure’s view of language undermines the objectivity of meaning. Meaning is never objective, in that words do not denote specific objects; instead it is subjective, in that the meaning of a word is always a function of what other words the interpreter associates it with and distinguishes it from in his or her mind. Meaning is never grounded in extra-linguistic reality, but floats suspended in a web of word associations. To the extent those word associations are shared, people can communicate, but they can never be sure that the meaning they conjure up in someone else’s mind is the same set of associations they had in their own mind. You will understand what I mean by car as long as you understand what I mean by truck and boat, but those too are defined by their differences from other signs. No sign ever points directly to a thing, so it is impossible to ever finally verify whether you and I understand “car” in the same way; all we can do is keep on defining how “car” differs from “truck” and how “truck” differs from “house” and so on ad infinitum, when all the time you might be thinking about toy cars while I am thinking about full size human transportation machines.
If De Saussure is right, then interpretation does not mean figuring out what facts revelation states about things—what legal values every action actually has; it is a matter of finding a network of word-understandings that allows all of revelation to be understood in relation to all the rest of it. This is in fact a remarkably good description of what legal theory actually makes possible: we will see as we go along that the principles of legal theory do not provide clear objective interpretations; instead they provide a toolbox of interpretive maneuvers that can be used to fit together all the different and conflicting things that revelation says so that they form a coherent law. This project of reconciling texts does not yield a single authoritative interpretation objectively grounded the real meanings of words; rather it produces just one of many possible coherent set of interpretations, such that if another jurist came up with another coherent way of fitting together all the word-uses in the Qurʾān and ḥadīth, that interpretation would be just as “true” as the first—since true is defined not in terms of correlation with extra-linguistic reality, but in terms of whether signs are consistently differentiated from one another. Several interpretations could be internally consistent, but there would be no way to tell which, if any, actually assigned the same legal values to the same actions that God had in mind. If signs refer only to other signs and not to things, then there is no way to determine whether I have understood God’s or anyone else’s use of signs the way he intended.
Ludwig Wittgenstein solved this problem by saying that concrete interaction in the material world fails when two people’s sign systems are out of synch, and this permits us to be confident that as long as we continue to interact successfully on practical matters, we must be communicating, using the same sign system, and understanding each other’s intent. But recall what we said earlier: we have no direct material interaction with God, so if De Saussure’s theory of language is correct, even Wittgenstein’s solution gives us no way to know whether we are really understanding God.
Al‑Juwaynī’s view of language says that I can know I am understanding God because I know that the things a word denotes in my sign system are the same things they denote in God’s sign system. This depends on the original Arabic sign system that God revealed to Adam being handed down intact from one generation to another from the beginning. Al‑Juwaynī himself, however, casts doubt on whether this is the case. For one thing, Arabs disagree among themselves about exactly what certain words mean. Some say that the imperative mood was coined to indicate an obligation (“pray!” means prayer is obligatory), while others say it only conveys a recommendation (“pray!” means only that prayer is recommended, unless the speaker also threatens to punish those who do not pray). Some say that a general plural noun like “the believers” is coined to mean all the believers who have ever existed, and that if it is used to refer to any smaller group of believers then it is being used transgressively; but others say “the believers” is coined to indicate only that the speaker has two or more believers in mind, rather than just one. These disagreements have vast implications for legal interpretation, as we will see when al‑Juwaynī discusses commands and general expressions.
Furthermore, even if all Arabs could agree on what each word meant originally, al‑Juwaynī affirms that the Arabic language has evolved over time. This is evident in his second definition of normative usage as using words “according to the conventions of speech.” This definition recognizes that humans can come to agree to use words in new ways that they were not originally coined for. He goes on to explain this when he says that “normative usage can be linguistic, revealed, or customary.” Linguistic usage accords with the original coining of a word; for example prayer (ṣalāh) has the original meaning of “calling out” or “supplicating” (duʿāʾ). Revealed usage accords with new meanings that God specifically gave to existing words in the Qurʾān or Prophetic Sunna itself. The revealed meaning of prayer, for example, is a specific sequence of words and actions that the Prophet taught his followers to perform five times a day. Thus we know from the Prophet’s Sunna that when the Qurʾān says “pray!” it is not commanding us to call on God any way we want; it is telling us to do so in the specific manner demonstrated by the Prophet. Customary usage accords with the way people have come to use words over time: a meaning that was once figurative or technical, for instance, might in time come to be the primary literal meaning of a word, as fiqh originally meant all kinds of understanding but came to refer primarily to law.
Not all legal theorists recognized these different types of normative usage. Al‑Bāqillānī, whose work on legal theory al‑Juwaynī studied and even summarized, claimed that the notion of revealed usage was just something the Muʿtazila made up to justify reading their own preferred definitions of words like “faith” and “believer” into the Qurʾān, even though those definitions did not accord with ordinary linguistic usage. Consequently, al‑Bāqillānī denied revealed usage, and insisted that when the Qurʾān says “pray!” it means simply to call upon God, without intending any specific form of prayer. We might ask whether al‑Juwaynī, and all the other legal theorists who allowed that the Arabic language and the meanings of words could evolve over time, were perhaps doing what al‑Bāqillānī accused the Muʿtazila of doing: taking their own sets of word-associations, which had been built up through centuries of legal thought, and projecting those technical legal definitions back onto the Qurʾān and the Prophet.
When a text is lifted out of the linguistic sign system within which its author wrote it, and is transplanted into a later culture with its own slightly different set of definitions for Arabic words, the meaning of the text changes. Al‑Juwaynī’s theory of language assumes that meaning is objective and grounded in God’s original assignment of names to things; but at the same time it allows sign systems to evolve, and thus opens up the possibility that the meaning of a text may change as it is reread by people whose webs of word associations do not match those of its author or its original audience. Al‑Juwaynī calls meaning objective, but makes it dependent on the presuppositions of the interpreter.
On theories of speech, the reduction of revealed language to the indicative, literalism, and speaker’s intent, see David R. Vishanoff, The Formation of Islamic Hermeneutics: How Sunni Legal Theorists Imagined a Revealed Law (New Haven, Connecticut: American Oriental Society, 2011).
On Islamic theories of the origins and evolution of language, see Bernard George Weiss, “Language in Orthodox Muslim Thought: A Study of ‘waḍʿ al lughah’ and Its Development” (Ph.D. diss., Princeton University, 1966); and Mohamed Mohamed Yunis Ali, Medieval Islamic Pragmatics: Sunni Legal Theorists’ Models of Textual Communication (Richmond, Surrey: Curzon Press, 2000).
- The term intent (qaṣd) may appear once in the treatise, when he says that “command does not require the repetition of the act unless some evidence indicates that repetition is intended;” but the words “that repetition is intended” are not in the most reliable text, and probably were added by a copyist or commentator.