Underdetermination in Late Postclassical Ḥanafī Legal Theories

In: Oriens
Author: Asad Q. Ahmed1
View More View Less
  • 1 NES, University of California
Full Access


This article argues that Ḥanafī uṣūlīs of the later phases of the postclassical period understood uṣūl to be universal propositions that were underdetermined with respect to their evidentiary bases. Though the purpose of such propositions was to confer actionable certainty to particular legal effects, the later tradition imagined the main charge of uṣūl al-fiqh on a meta-theoretic level, i.e., to determine how such propositions could themselves be suitably grounded. In casting the discourse within the framework of naturalized technical methods and distinctions from fields of logic and philosophy, the tradition generally granted the relational and systemic validity of each proposition in terms of the grounding it received from another underdetermined proposition. This second-order perspective of the tradition reveals that uṣūl were systemically and relationally valid, but individually underdetermined. Thus the application of the attribute of relational validity to them and to the effects for which they are serviceable is apt.


This article argues that Ḥanafī uṣūlīs of the later phases of the postclassical period understood uṣūl to be universal propositions that were underdetermined with respect to their evidentiary bases. Though the purpose of such propositions was to confer actionable certainty to particular legal effects, the later tradition imagined the main charge of uṣūl al-fiqh on a meta-theoretic level, i.e., to determine how such propositions could themselves be suitably grounded. In casting the discourse within the framework of naturalized technical methods and distinctions from fields of logic and philosophy, the tradition generally granted the relational and systemic validity of each proposition in terms of the grounding it received from another underdetermined proposition. This second-order perspective of the tradition reveals that uṣūl were systemically and relationally valid, but individually underdetermined. Thus the application of the attribute of relational validity to them and to the effects for which they are serviceable is apt.

1 Introduction

The purpose of this article is to demonstrate that the choice of any legal theory was taken to be underdetermined by certain influential late postclassical1 Ḥanafī uṣūlīs. In this context, by theory, I mean to refer to those general and foundational principles, articulated as universal propositions, that would justify or facilitate the extraction of the body of legal effects; the specific indicants for the legal effects would fold into the universal propositions and gain their hermeneutic force via them.2 In other words, legal theories of the sort I will discuss below allowed for the transition from a particular scriptural, linguistic command to the evaluation of the legal and moral value it communicates. Yet neither my concern nor that of the parts of texts I examine is to enumerate such theories or to show how, on their basis, Muslim jurists argued that obligations may be posited on the weight of probable epistemological states. Rather, the discussion pertains to the uṣūlīs’ higher-order reflections on the very nature of these theories.3

In simple terms, a theory is underdetermined when it and its contraries equally explain the evidence at hand; a stronger version would claim that the contraries imply the evidence. The case is analogous when the evidentiary body is a series of theories: a given theory is underdetermined when both it and its contraries within the system of which these theories are a part equally explain another theory. Conversely, in such cases, any given theory that is an explanans is also an equal explanandum of another theory and its contraries. Thus at any given layer of the system, an underdetermined theory no more explains its underlying evidentiary theories than an alternative one; nor is it explained by another theory exclusively.

The following example may be illustrative. Let us assume that John states to his nine year-old daughter, who is never fond of bedtime, “I wish you were in bed at this time!” This statement confuses the girl. For she is not quite sure whether John is obliquely issuing a threat that she would be grounded if she does not comply or whether he is genuinely leaving the decision up to her. Two propositions are, therefore, heuristically deployed by the girl: (1) “Whenever my father uses the subjunctive mood, he must be obeyed” or (2) “Whenever my father uses the subjunctive mood, he is relinquishing decisions up to me.” (1) and (2) both equally explain the evidentiary statement, “I wish you were in bed at this time!” (1) and (2) are, therefore, both theories that are underdetermined in relation to the evidence.

The choice of one theory over the other requires further theories. For (1), an enthymeme occurs to the girl’s mind: (A) “Whenever my father uses the subjunctive mood, he is upset;” “whenever my father is upset, he should be obeyed” (in an enthymeme, the second proposition, which is the major premise, would normally be suppressed); therefore, (1) “Whenever my father uses the subjunctive mood, he should be obeyed.” Yet (1) may be equally grounded in (B) “Whenever my father uses the subjunctive mood, he is being polite;” “whenever my father is being polite, he should be obeyed;” therefore, (1) “Whenever my father uses the subjunctive mood, he should be obeyed.” In this case, (1) is the evidentiary proposition, which is equally explained by (A) and (B), each of which is thus also underdetermined. The regress need not require further comment.

That the body of theories in a system should be underdetermined in the aforementioned manner does not mean that the system as a whole is necessarily impugned as probabilistic or, worse, as false. Nor indeed, as we see in the hypothetical case above, would the explanandum have to be rejected, provided the regress is sufficiently deep. Each theory, as the evidentiary explanandum of its higher-level theories is valid on the condition of their determined status. And each of the latter is valid on the condition of the determined status of their theories. And so on. Thus one may argue that, as a whole, the system is determined, and that each of its theories is conditionally determined. Understood in this fashion, no universal proposition or theory in a system is true or false in isolation; it is simply valid or invalid in relation to the whole.

My claim in this article is that this was precisely the framework within which certain late Ḥanafī legal theorists imagined the project of uṣūl al-fiqh. The discipline was meant to posit universal propositions of legal significance; yet it was also recognized that each such proposition was underdetermined in relation to its immediate evidentiary body. Determinacy—and therefore, validity—was possible for any given theory insofar as it was conditioned by theories of an upper layer; the latter would be similarly conditioned as a means to overcome their own underdetermination. And so on. Thus, as a whole series of theories, the system could in fact be issued as a closed and determined type. The system may or may not have been true or certain or probable; but it was valid. Consequently, in virtue of this broader mode of determinacy, the whole could serve as a criterion for verifying the validity—not independent truth—of any of its theoretical parts. Again, the functional validity of theories within the system did not necessarily imply their independent truth.4

2 An Earlier Thesis

The thesis I am advancing in this article ought to be clearly disambiguated from another one with which it overlaps in some respects. In his groundbreaking and now classic dissertation, The Economy of Certainty, Aron Zysow states the following:

From a very early period, for reasons upon which we are free to speculate, Muslims came to treat the question of legitimacy [i.e., of legal orders] along strictly epistemological lines. Certainty and probability were the fundamental categories with which they approached every question of law … Its constant preoccupation with epistemology makes Islamic law much more self-conscious than either of these two systems [i.e., Anglo-American law and Jewish law], and for those of a philosophical bent, much more sophisticated.5

These are general observations that hold also for the thesis of this article: the late postclassical legal-theoretical tradition maintains, first and foremost, an obsessive concern with epistemology as the foundational backdrop to its discourse; and by the late postclassical phase, what was once a philosophical inclination had intensified into a method due to the long process of the absorption of falsafa and manṭiq distinctions into the apparatus of uṣūl.6 Yet there is a fundamental difference between what Zysow unfolds in his work and that which I wish to bring to the fore. Zysow shows that the major function of legal theory was to generate functional certainty with respect to legal acts, even in cases where such certainty was not epistemologically available. For the fulfillment of this project, those who admitted probabilistic evidence deployed a number of arguments and devices, such as the notion of benefit (maṣlaḥa) and the theory of consensus (ijmāʿ), with relatively regular and consistent effectiveness. These methods either elevated the epistemological probability of the evidence to certainty or rendered the former sufficiently neutralized, so as not to impugn the obligatory nature of certain legal acts.7

Here, let me take up an example outlined by Zysow. In discussions of how unit reports (khabar al-wāḥid) of a grade of epistemological probability may generate a legal norm, the sources argue that such reports were accepted by the Prophet and the Companions. However, such a claim about their attitude toward unit reports is itself conveyed by unit reports. Thus one runs into a debilitating circularity in the proof. One way out of this circularity was by way of appeal to the method of upgrading unit reports to concurrency (tawātur). The basic argument of the sources is that reports that convey the same meaning have concurrency by virtue of their shared meaning (tawātur bi-l-maʿnā), even though no unit report itself is transmitted concurrently. This move would allow unit reports to generate norms, thus conferring instrumental certainty to probably-true reports.8

Yet late postclassical Ḥanafī sources would not pause with this step. What is of interest to them is to see on what grounds—i.e., on the basis of which universal principle (aṣl)—the rule that allows for the shift from the khabar al-wāḥid to tawātur bi-l-maʿnā could itself be endorsed. Thus the investigation would focus on a principle such as “If a whole range of unit traditions sharing a meaning exists, it confers concurrency by virtue of the meaning.” In other words, it is this universal theory—a second-order grounding proposition under which the move from probability to certainty is subsumed—that is itself suspect and that would itself receive sustained attention. And it is such principles that were considered to be the subject matter of late postclassical Ḥanafī legal theory.9

Now it should be pointed out that Zysow brings the significance of such principles to our attention. For example, in showing how the tradition used the probity of a reporter of a unit report as an argument for establishing a legal norm, he writes,

Two elements, say the Muʿtazilites, must be present for us to act on a unit-tradition. First, the unit-tradition itself and, secondly, the master-rule that informs us of the obligation to conform to the norm contained in the tradition. The unit consisting of the master-rule along with the relevant tradition now bears the load of certainty. This master-rule does not differ in any significant respect from an ordinary command of God.10

Zysow then quotes from the Kitāb al-muʿtamad of Abū al-Ḥusayn al-Baṣrī: “There is no difference between God saying to us, ‘When the veracity of the reporter seems probable to you, then act according to his report,’ and His saying, ‘Do so and so.’ ”11

Master-rules are precisely the principles/hypotheses that are the subject matter of late postclassical uṣūl; it is such principles that allow the legal status of acts to be stipulated.12 In the aforementioned case, the focus of Zysow’s argument, however, is not the principles themselves; rather, he aims to show that such principles, when compounded with the evidence, upgrade epistemological probability to certainty with respect to action, i.e., to a legal obligation. This appears to me to be a correct assessment of one of the ways in which legal theory aimed to produce certainty. Where the later tradition would diverge, however, is with regard to Zysow’s claim that the positing of the master-rule produced certainty by virtue of itself;13 rather, as I mentioned above, the master-rule itself was taken to be underdetermined; it was generated as system-specifically valid (not certain) and it grounded valid (not certain) legal norms.

The difference between the two theses is perspectival. In terms of their function and operation, legal theories were indeed concerned with generating actionable certainty out of probabilistic epistemologies by positing various general instruments. This would be an internal perspective on the avowed purpose of the discipline. On the other hand, these instruments, which were the grounds for the legal effects and were considered to be the subject matter of the discipline of legal theory, were themselves in need of grounding. Thus insofar as the stated purpose of the discipline of legal theory was to inquire into the nature of such instruments—not into how they deliver legal effects, but into how they are grounded and on what grounds they deliver legal effects—legal theories were only conditionally valid, not certain. This would be a systemic and higher-order theoretical perspective in contrast to the internal one.

In other words, any legal action was valid in virtue of the master-rule; the master-rule was valid in virtue of another rule; the latter was valid in virtue of another rule; and so on. This, in turn, meant that the validity of the legal rule was granted by virtue of the congruent complexity of the whole system. The principles/master-rules grounded legal norms and they were uncertain in virtue of themselves. Yet they were valid insofar as they constituted a whole propositional system. As I will show below, the later Ḥanafī tradition was conscious that this was the nature of the legal-theoretical system on a meta-level of inquiry, i.e., as an inquiry into the nature of the very master-rules and principles.

3 Logic, Fiqh, and Uṣūl in the Musallam al-thubūt of Muḥibballāh al-Bihārī14

Completed in 1109/1698, the Musallam al-thubūt is widely considered the most advanced text of Ḥanafi legal theory produced in India. Embracing a lode of philosophical and logical distinctions and technical arsenal, this work advances key Ḥanafī claims against the backdrop of the Shāfiʿī tradition of legal hermeneutics. Two extended commentaries were written on this work, the first by ʿAbd al-ʿAlī Baḥr al-ʿulūm15 and the second by ʿAbd al-Haqq al-Khayrābādī;16 a third modern commentary, covering the introduction and the first two maqālas (on the theological postulates and on aḥkām) was authored by Mawlawī Hidāyatallāh. All three will be deployed in the analysis in this section.

In his introductory section, al-Bihārī sets up several productive prompts that guide the discourse of his future commentators. “The introduction [of this book],” he writes, “concerns the definition (ḥadd) of uṣūl al-fiqh, its subject matter (mawḍūʿ), and its final cause (ghaya),” i.e., its objective. The first two of these two notions—the definition and the subject matter—pave a path of philosophical disquisition within the commentarial tradition on the Musallam. Thus Baḥr al-ʿulūm comments that one may take definition in its real sense, i.e., as a genus and specific difference, and that, by subject matter, the author intends “that whose essential accidents that come to inhere for a thing by virtue of its essence or due to something equivalent are investigated.”17 In other words, the topic of the book is an investigation of the essential accidents of uṣūl insofar as the latter is defined in a philosophically-robust manner.18

As for the definition of uṣūl al-fiqh, al-Bihārī explains that, when it is taken as comprising two elements that are in an iḍāfa construction, then aṣl may be understood in one of two senses: as (1) a simple lexical item, it may refer to something upon which another thing is based (mā yabtanī ʿalayhi ghayruhu);19 or (2) as a technical term, it may refer to four meanings—(A) that which has preponderance over something else (rājiḥ), (B) that whose continuity is presumed (mustaṣḥab), (C) a rule (qāʿida), and (D) an indicant (dalīl).20

In the technical sense, therefore, the expression aṣl is homonymous with respect to four meanings. Its precise and specific definition is conveyed by the specific difference that circumscribes it, namely, the fact that it stands in a governed relation to a body of knowledge. This body of knowledge is fiqh, to which I will return presently. The aṣl of a body of knowledge is explained as follows:

It is conveyed in the Sharḥ al-Mukhtaṣar21 that when aṣl is posited in a governed relation to knowledge, then what is intended is its indicant (dalīl). There is no doubt about this. However, this is not because the expression aṣl is used in the sense of indicant [as a technically-fixed meaning]. How could this be? For if it were so, the transfer [in meaning] would be entailed twice. Rather, the expression aṣl is used in its lexical meaning and, when it is posited in a governed relation with knowledge, the sense becomes ‘that upon which knowledge is based’. Its basis [i.e., the basis of knowledge] is nothing other than an indicant. It is in this sense that dalīl is intended, not in the first sense [of a technically-fixed item].22

Thus the meaning of aṣl as indicant is not produced by means of a double transfer of meaning: it is not the case that aṣl is first transferred from its lexical meaning (base) to its technical meaning (root) and then further transferred to a subcategory of the technical meaning (indicant). Rather, the meaning as indicant is derived from the linguistic context—its governed relation to knowledge—that suggests its meaning as indicant. In other words, the lexical meaning of aṣl, as noted above, is “that upon which something is based;” in this case, the aṣl is given as the base of fiqhī knowledge; and such a base is an indicant (dalīl).23 The significance of this argument, as we will see, is that it divests the notion of dalīl from its tight technical usage in legal theory and allows it to be analogized to a broader theory of derivation. Yet it ought to be noted that the possibility of analogy to a broader theory does not mean that the tradition of the Musallam conceded in any way that dalīl in legal theory was identical to any broader theory. I will elaborate on this point below.

Further, al-Bihārī explains that “one24 who takes uṣūl in the sense of rule (qāʿida) has forgotten the principle that the rules of a body of knowledge (ʿilm) are its problemata (masāʾil), not its foundational premises (mabādī).”25 As Baḥr al-ʿulūm explains, “If in the present case, aṣl meant rule, the meaning of [uṣūl al-fiqh] would be the problemata of fiqh. And this is absurd. Know that there is no doubt that taking aṣl as rule is unlikely. However, there is a kind of soundness [in this] if one were to manipulate the governing relation slightly, i.e., [if one were to take aṣl as] ‘problemata that have some connection with fiqh’ (masāʾil lahā taʿalluq bi-l-fiqh).”26

We now have a further definitional component of uṣūl, namely, that they are the foundational premises of fiqh, not its problemata. As was the case with the usage of subject terms in the context of the definition of uṣūl, so here also the argument is grounded in certain distinctions from the field of logic. A representative quotation may be supplied from Avicenna’s Najāt:

The things that constitute demonstrations are three: [1] the subject matter under investigation (mawḍūʿāt), [2] problems [related to the subject matter] (maṣāʾil), and [3] premises that are principles [and starting points] (mabādī). One demonstrates about subject matters, demonstrates [the solution to] problems [related to the subject matter], and demonstrates by means of premises.27

And so uṣūl are the indicants upon which fiqh is based; they are its foundational premises and starting points, not problemata proved within fiqh. Yet they are problemata of their specific discipline and are demonstrated within it. Now we might recall from above that what is investigated within a discipline are the essential accidents that pertain to its subject matter. Thus the discipline of geometry would investigate measurable dimensions—the subject matter—insofar as range and boundary are their essential accidents; arithmetic would investigate number insofar as it is odd and even, etc. And the problemata of the former would be something such as “A straight line through the center of every measureable dimension with a circular boundary has an approximate ratio of 7 to 22 in relation to that boundary;” the latter would include problemata, such as “Every number that is odd cannot be divided by two so as to yield a whole number.” Similarly, one subject matter of the discipline of uṣūl is the divine command and a problema that is investigated in it is “Every divine command implies obligation on the part of the person commanded.” In other words, the purpose of uṣūl is to investigate and demonstrate the validity of such propositions that stand as its problemata.

Next, fiqh is defined as “wisdom that is derivative and textually-grounded” (ḥikma farʿiyya sharʿiyya).28 Baḥr al-ʿulūm explains this expression to mean that fiqh is a knowledge of actual, not supposed, matters and that it presupposes belief in certain creedal matters and is established by means of textually-grounded indicants.29

Further, fiqh is defined more precisely as “knowledge of textually-grounded statements [derived] from their expressed indicants” (al-ʿilm bi-l-aḥkām al-sharʿiyya ʿan adillatihā al-tafṣīliyya). In this context, the expression aḥkām, normally translated as legal norms or legal effects, is parsed by Baḥr al-ʿulūm as “the occurrence of a relation or its non-occurrence”, i.e., the relation of a subject and predicate in a proposition. With this definitional component, he explains, the author precludes simple conceptions (taṣawwurāt sādhija),30 i.e., those where no relation between two things obtains, from the class of knowables that constitute fiqh. In other words, fiqh is knowledge that is propositional in nature. Yet propositions such as “fire burns” and “the heavens are spherical” are not fiqhī because they are not established on textual grounds (sharʿī).31

In sum, fiqh is the knowledge of propositions that are both textually grounded and are derived on the basis of specific and expressed textual indicants. If the knowledge of the propositions is not proof-dependent, such as the knowledge of a muqallid or the knowledge possessed by the angel Gabriel, it would not constitute fiqh.32 The same would be true of knowledge that is not textually grounded, even if it has legal content. The grounds for the derivation are the uṣūl, which, as problemata, are investigated and demonstrated in uṣūl al-fiqh.

So what exactly are the uṣūl and in what relation do they stand with the body of knowledge of which they are the grounds? Al-Bihārī writes, “Next, this body of knowledge/discipline (ʿilm) is the compressed indicants33 of fiqh which are needed when the expressed indicants are made congruent with their judgments (thumma hādha l-ʿilm adilla ijmāliyya li-l-fiqh yuḥtāju ilayhā ʿinda taṭbīq al-addila al-tafṣīliyya ʿalā aḥkāmihā).”34 Three items must be isolated in order to deliver clarity to this definition: the compressed indicants, the expressed indicants, and the judgments. The first of these are the grounds whereby the second is made operative, so as to confer validity to the third. As I noted above, the individual elements of the definition are rather well known; but the manner in which they are deployed is novel and is reminiscent of the treatment of indicants in the falsafa tradition. As an example that will clarify the motives of the matn and the interpretive angles of the commentaries that followed, one might again quote from the Najāt of Avicenna:

[Dalīl,] in this instance, is an enthymeme, whose middle term, when it exists for the minor [term,] the existence of something else for the minor [term] is always entailed, however it may do so. If its two premises were made explicit, it would occur in accordance with the first figure. An example is your statement, ‘This woman is lactating; so she has given birth.’ Sometimes this syllogism itself is called a dalīl and sometimes the middle term is called by [this name].35

An enthymeme is a syllogism wherein the major premise is suppressed.36 In the example above, the minor premise is “This woman is lactating” and the conclusion is “This woman has given birth.” The suppressed major premise is plain: “Whoever is lactating has given birth.” Now a dalīl, as explained by Avicenna, is either the lactating subject (the middle term) or the syllogism, i.e., the statement composed of the two premises, one explicit and the other suppressed (i.e., an enthymeme).37 In principle, the term dalīl may in fact be applied both to the middle term and to the syllogism composed of the two premises. This is so, because the middle term contains within it a universal property that would transfer to the minor term when the latter is taken to stand as the subject of the middle term. This is precisely the enthymeme that comes to be expressed in a first figure syllogism when both the premises are made explicit.38

These considerations must have been contained within al-Bihārī’s matn.39 For without much preliminary exposition, Baḥr al-ʿulūm states the following:

If the [compressed] indicant is rendered/opened up (idhā ḥurrira al-dalīl) according to the ordering of the first figure, its major [premise] would be taken from/extracted from among the uṣūl (maʾkhūdha min al-uṣūl). [This would be so] whether [the major premise] is a specific/exact uṣūlī problema or is incorporated in [the problema] or is taken from a number of [problemata] (sawāʾun kānat ʿayna masʾala uṣūliyya muʿayyana aw mundamija fīhā aw maʾkhūda min ʿiddat masāʾil). When the indicant is rendered/opened up according to the arrangement of the exceptive syllogism, the implication (mulāzama) would be taken from/extracted from among them [i.e., the problemata of uṣūl].40

The commentator’s interpretation does not require extensive elaboration in view of the foregoing presentation of dalīl that is found within the framework of logic. Uṣūl are compressed indicants (adilla), i.e, middle terms or enthymemes, which may be made manifest in the first figure or exceptive syllogism. In the first case, they open up as the major premise of a syllogism and, in the latter, they supply the general implication rule. In other words, the major premise and general implication rule are extracted from the compressed indicants in a universal mode, so that, as we will observe below, they may be serviceable in generating the conclusions, i.e., the aḥkām. The compressed indicants themselves are the problemata of the discipline and the extent to which they may be used to draw out the universal major premise or the implication rule, they are the subject of investigation in the discipline of uṣūl al-fiqh. Now it is the process of grounding the major premise and implication rule—the problemata of uṣūl—that is the central aspect of the investigation of the discipline; and it is fraught with the problem of underdetermination.

4 Underdetermination

The underdetermination of uṣūl may be demonstrated on the basis of the examples offered by Baḥr al-ʿulūm:

Giving alms is obligatory because of the statement of the Elevated, ‘Give alms!’ If we wanted to make [this expressed indicant] congruent with its judgment, we would say [A.] Alms are commanded by God the Elevated and [B.] Whatever is commanded by the Elevated is obligatory. [This is so] because a command is [issued] for an obligation. So this major [premise] is taken from/extracted from an uṣūlī problema. Next there must be restrictions for the soundness of the universality of this major [premise]. And these are [C.] Whatever is commanded by a non-abrogated command … is obligatory; and [D. Whatever is commanded by a command] that is not contradicted by something that is preponderant [is obligatory; and E. Whatever is commanded by a command that is not contradicted by] something equal [is obligatory; and F. Whatever is commanded by a command that is not contradicted by] something that has a reductive interpretation [is obligatory]. So the complete realization of this proposition (qaḍiyya) requires knowledge of the problemata of abrogation, contradiction, and reductive interpretation. Thus this major [premise] is taken from/extracted from a number of these problemata.41

As we observed above, a first figure syllogism would be, for example, of the form “A is B; Every B is C” and would yield the conclusion “A is C”. With terms, an example of the figure would be “Socrates is a man; every man is an animal” and this would yield the conclusion that Socrates is an animal. In the conclusion, the application of the predicate—animal—of the subject—Socrates—depends on the validity of the universal major premise that every man is an animal. For the assertion of Socrates’ being an animal is congruent with the assertion that he is a man in virtue of the major premise that allows man to be subsumed into animal.

The case of textually-grounded judgments (aḥkām) is identical. The conclusion “Giving alms is obligatory” is congruent with the minor premise “Giving alms is commanded by God” by virtue of the major premise “Whatever is commanded by God is obligatory.” This major premise is a general theory that explains the evidence of the particular claim of the conclusion insofar as it relates to the expressed indicant, the minor premise that is supplied by the textual source. Command, therefore, is among the subject matters of uṣūl; and one of its problemata concerns the theory of the universal nature of its essential accident of obligation. This is where the problem of underdetermination emerges in full view. The claim of the universality of obligation in relation to commands—a theoretical problema of uṣūl—is itself grounded in a number of other problemata.

At first blush, there is nothing remarkable about this situation, since the problemata of various disciplines are indeed evidentiary in relation to theoretical problemata explored in distinct and higher disciplines; the latter are principles (mabādī) of the former. The case at hand, however, is different. For here the problema of a discipline is investigated by means of a number of other problemata of the same discipline. Furthermore, insofar as the entire ambit of the discipline of uṣūl al-fiqh is textually-grounded, no ultimate highest-order problema can be posited that can serve as a theoretical ground and principle for other problemata within this same discipline.

That the entire ordered sequence of theoretical and evidentiary problemata is internal to the discipline is pointed out by al-Khayrābādī in the following terms: “This major [premise] is a compressed indicant from among the problemata of uṣūl al-fiqh … and just as this major [premise] is among the problemata of uṣūl al-fiqh, so that upon which its complete realization depends is also from among the problemata of uṣūl al-fiqh.”42 Indeed the assessment of the modern commentator, Hidāyatallāh, forces the problem into an even more intractable space.43 He explains that the compressed universal premise, “Whatever is commanded by God is obligatory” has the expressed indicant “Give alms!” as one of its instances.44 This claim implies that a series of such expressed indicants, the minor premises of the syllogism, yields the compressed indicant, i.e., the major premise. Yet a grounding theory that is independent of the ordered sequence of uṣūlī theories is needed both to justify and to undergird this move.45 Hidāyatallāh details his argument in the following fashion:

The author, [i.e., al-Bihārī] mentions as an objection [to Amānallāh al-Banārasī] that the [expressed] indicants of fiqh—in terms of their form and matter (ṣūratan wa-māddatan)—are among the instances of the subject matters of the problemata of uṣūl.46 For example, “Give alms!” is an expressed, i.e., fiqhī, indicant. Now its matter is “the fact of giving” and its form is the morphology that a command [has in Arabic]. With respect to both [its matter and form] “Give alms!” is an instance of the subject matter of a problema of uṣūl, namely, “The command is for obligation.” This is so because “Give!” with respect to the [fiqhī] problema of alms, and “Perform!” with respect to the [fiqhī] problema of prayer, are among the totality of commands.47

In Baḥr al-ʿulūm’s disquisition, we had witnessed the claim that uṣūlī problemata that serve as major premises depend for their realization on one or more other uṣūlī problemata; in each case, the former stood as general theories in relation to the evidentiary base contained in the latter. And this position was explicitly endorsed by al-Khayrābādī. In line with the basic framework of this argument, Hidāyatallāh turns to the very first layer uṣūlī problemata and explores their relation to their evidentiary base.48 According to him, the compressed indicants that are the problemata of uṣūl have the expressed indicants of fiqh as their instances. What this means is that the problema “Commands confer obligation” that allows for the expressed indicant “Give alms!” to yield the conclusion that alms are obligatory is itself dependent for its universal explicatory capacity on the expressed indicant both in terms of its matter and form. To put it differently, the interpretive explicans of a fiqhī indicant—which is a problema of a higher discipline, i.e., uṣūl al-fiqh—is the explicandum of its own explicandum. Now this outcome would lose its troublesome edge if one could posit some independent grounds whereby the move from the fiqhī indicants to the uṣūlī ones could be justified.

Indeed Hidāyatallāh’s foregoing comments were precipitated by a desire to vitiate precisely such a possibility. The inspiration came from the commentaries of Baḥr al-ʿulūm and al-Khayrābādī that build upon only a seemingly out-of-place prompt of the matn:

The relation of [uṣūl] to fiqh is not the [same as] the relation of logic to philosophy, as it is falsely imagined. For the expressed indicants, both in terms of their matters and forms, are the instances of the subject matter of the problemata of uṣūl. This is not the case with logic, which investigates second intentions.49

This false analogy between uṣūl and logic is attributed to Amānallāh al-Banārasī and, were it taken to be valid, it would have granted license to non-textual theories to be the ultimate determinants and grounds of uṣūlī problemata. Take, for example, the following comparative cases. The first one is from logic.

Rule T = If p, then q

Now this rule may also be expressed as:

L = If A is B, then A is C

And as an example of this formal rule, one might offer

M = If the heavens are simple, then the heavens are spherical

The problem with M is that it brings into relief the fact that the derivation of the consequent from the antecedent is lacking any grounds, though of course it occurs on the pattern of T and L. Similarly, the rule L, where “A is B” appears as an expressed form of the antecedent, p, of the logical rule T, does not demonstrate any implication, though of course L is true on the assertion of T. The validity of T, therefore, depends on the recognition that p is an enthymeme, i.e., a syllogism where the major premise is suppressed. This suppressed premise is:

K = Whatever is B is C

As such, p expresses “A is B and whatever is B is C”, so that L, properly speaking, should be “If A is B and whatever is B is C, then A is C”. From the discussion presented earlier in the article, it is clear that K is the same as a compressed indicant or a universal major premise that confers validity to the conclusion, i.e., the consequent. It can be presented as a conditional proposition “If something is B, then it is C”. And this is nothing other than L above, granted in virtue of the interpretive move called the dictum de omni:

N = “A premise containing a dictum de omni [states that] there is nothing of which the subject is said of which the predicate is not also said.”50

So if something is B, then it is C. And this is the final step in the series that is needed to validate L. The summary may be presented as follows. The logic rule T may be expressed as L. This articulation, along with the example M, makes it plain that the antecedent p in L must be an enthymeme, the major premise of which is K, which is identifiable with the dalīl mujmal. It is K that confers validity to the derivation of the consequent. Expressed as a conditional,51 K reduces to L and becomes formally operative via the deployment of the interpretive move that is allowed by N, the dictum de omni.

Now the keener philosophical problems with the foregoing argument aside, in principle, K is said to be made operative via the imposition of the independent semantics of the dictum de omni. It is in virtue of this interpretive principle that the circle of proof may be said to be complete, so that L may have formal validity and determinacy. Thereafter, within the range of L, any set of particular and material cases (such as M) may be evaluated. In the case of uṣūl al-fiqh, there appears to be no independent formal interpretive semantics that allows for an aṣl to have determinacy. Take the following simple case from uṣūl.

L2 = If A is commanded by God, then A is obligatory

M2 = If alms are commanded by God, then alms are obligatory

Now on the model of the previous argument, it becomes obvious that the antecedent in L2 is an enthymeme, where the major and suppressed premise is

K2 = Whatever is commanded by God is obligatory

But here the proof must stop, as there is no formal, materially-divested, and independently-posited interpretive semantics via which K2 itself can be made operative as L2. The evidentiary base of K2 is a materially-invested set—the specific problemata of fiqhī indicants, such as “Give alms!” “Perform prayer!” etc. So K2 is specifically sharʿī, i.e., textually bound, both in terms of its evidentiary base and the theoretical semantics that makes it operative as a rule, such as L2. Its formal validity is inseparable from its material validity. To put it differently, the conditional rule/major premise of the enthemyme is underdetermined with respect to the evidentiary base—“Give alms!” “Perform prayer!” etc.—because there are no independent grounds for its theoretical determination. The evidentiary base is itself the very grounds.52

The foregoing comparison was a summary analysis of Baḥr al-ʿulūm’s statements. He writes:

It has already become clear that the discipline of uṣūl al-fiqh has a particularity in relation to fiqh (li-ʿilm al-uṣūl khuṣūṣiyya bi-l-fiqh); it is a particularity that it does not have with something other than [fiqh]. As for logic, its relation to philosophy, uṣūl, and fiqh is one relation. [Logic] is needed only in order for one to have knowledge of the nature of deriving a conclusion. Indeed there is no [major] premise [of any discipline] whose proof/indicant is from among the problemata of logic. Yet [this last claim] may be problematized with a view to the investigations in qiyās. For these [investigations] are needed only in [determining] the nature of the derivation of its conclusion. How could it [be more?!] For qiyās supplies a judgment (mufīd li-l- ḥukm) with respect to itself without the addition of any other thing along with it.53

Thus uṣūl al-fiqh is like logic in the sense that it supplies rules of derivation. Yet it must be distinguished from it in that, whereas logic is broadly applicable to all disciplines equally, uṣūl al-fiqh is bound particularly to fiqh. The point is further driven home in the assertion that there is no major premise of any discipline that is derived from logic; the clear implication is that, on the other hand, uṣūl al-fiqh does supply such major premises for fiqh. This much of course should be no surprise, given the foregoing observations. Next a challenge is posed: is it not the case that qiyās in legal theory is as formal as logic, since it is not concerned with any specific material base? If this is so, then insofar as qiyās can serve as a compressed indicant for fiqh, at least some part of the discipline of uṣūl al-fiqh would be materially divested and independent of fiqh. In turn and in principle, one might be able to make a case for theoretical determinacy for at least some problemata of uṣūl. Baḥr al-ʿulūm lays this possibility to rest in the following statements:

Yet you must not stumble. For qiyās does not supply a textually-grounded judgment/legal effect (ḥukman sharʿiyyan) except with a view to the consideration that the originator of the text (al-shāriʿ) took overwhelming belief that obtains from [qiyās] to be reliable (iʿtabara ghalabat al-ẓann al-ḥāṣil bihi). And so no textually-grounded judgment/legal effect is established except [by means of the following argument]: ‘Qiyās leads to this judgment/legal effect; anything to which qiyās leads is established by God.’ The second proposition is taken from among the uṣūl. As for the materially-divested qiyās [that exists] without this proposition (al-qiyās al-mujarrad bi-dūn hādhihi al-qaḍiyya), well it does not supply [the conclusion] that this judgment/legal effect is from God, so that acting in accordance with it should turn out to be an obligation. The soundness of this [second] proposition requires qualifications. So one must know whether qiyās has been abrogated or not and so on. On the basis of what we have mentioned, an objection that [only] appears to be correct has been refuted, namely, that some problemata of uṣūl are not suitable to be major premises.54

In terms of their formal aspect, uṣūl can comprise at least some theories that are indistinguishable from theories of logic. Yet as such, they are not serviceable for fiqh. For example, perfectly sound theories of analogical and syllogistic reasoning may be posited as major premises for the purposes of deriving a conclusion on the concession of a particular, minor premise; on formal grounds, such conclusions would be valid. At the same time, such conclusions would not be relevant to fiqh, which requires the major premises also to be materially-invested in particular fiqhī indicants or in other uṣūl that are ultimately grounded in them.55 In the example that Baḥr al-ʿulūm offers, an underlying mode of sound formal argumentation yields a conclusion, which is not yet a legal effect of fiqh. In order for it to be so, the formal argument must include a materially-invested major premise, a major premise which is itself formally and materially grounded in specific cases. As Baḥr al-ʿulūm explains, every problema of uṣūl—even one that is manifestly concerned with the question of formal validity—is of this sort.56

5 Conclusions

This article does not contravene the idea that the aim of Islamic legal theory was to facilitate the derivation of actionable certainty out of a probabilistic base; nor does it challenge the position that such certainty was argued either by means of upgrading the modality of indicants or by the deployment of functional master-rules. Instead it observes that, with the penetration and naturalization of a formal logical and philosophical apparatus within the late Ḥanafī uṣūl tradition, the principles that are identified as uṣūl and that served as the theoretical base of legal derivation, came to be understood as underdetermined and, therefore, uncertain in themselves.

Each theoretical problema of uṣūl is both materially and formally determined by its evidentiary base. In turn, each possibility of grounding the problemata of one layer by problemata of another is similarly conditioned. Thus there appears to be no independently-established and materially-divested first problema of uṣūl—not even qiyās—that may be deployed to undergird the entire system. This situation is presented by the sources as a contrast to the problemata of logic in relation to their material and evidentiary base.

What is of note in this development is that the late tradition is conscious of this fact and engages in a higher-order theoretical disquisition about its own enterprise. In other words, there is no doubt that the problemata/compressed indicants of uṣūl al-fiqh are taken to confer legal validity to judgments by making expressed indicants congruent with them. In functional terms, therefore, the task of uṣūl remains generally constant between the classical and postclassical periods. However, the higher-order reflection of the tradition on the nature of the problemata yielded something new.

The upshot of all this should be re-emphasized. The truth and certainty of each item of legal theory is underdetermined; and no such item can be independently grounded in formal logic, because of the material investment that logic lacks. However, each system of legal theory, as a whole, is legally valid simply by dint of the fact that each of its items is placed within it as congruent with its evidentiary base in virtue of modes of derivation that are also congruent with their similarly-conditioned evidentiary bases. The system is thus valid in a specific and self-referential manner, i.e., it is legally valid. Insofar as any problema of uṣūl can be shown to stand in a relation of legal congruence with another problema, it is also relevantly valid. Again, this does not mean that the problema is true or certain or that it can confer certainty to a legal effect by virtue of itself. It can only confer the system-specific relational validity to which it is also limited. Nevertheless, it does not appear to be the case that uṣūlīs considered that such a mode of underdetermination was a weakness in their system or that the latter was inferior to the more capacious field of logic. It is simply that uṣūl were not susceptible to truth conditions; rather, they operated within the sphere of validity.

The consequence of this higher-order perspective invites reflection. For it means that the possibility of change in Islamic law is both null and maximal: underdetermined theories, taken in isolated relation to their evidentiary bases, are arbitrary and may be jettisoned; at the same time, insofar as they stand in a complex relation of systematic congruence, they are unshakeable. But this is how any economy of validity would function.


I would like to thank Robert Gleave and the reviewer for their insightful comments on this article. The errors that remain are mine.


Periodization is always a problematic enterprise. I use the expression “postclassical” to refer to the continuity of certain modes of reading and writing (e.g., the commentary), of certain pedagogical methods within institutions (e.g., the madrasa), of certain systems of patronage (e.g., the princely court, the waqf), of certain diachronic networks (e.g., the Khayrābādīs), etc. As such, certain regions would be manifestly postclassical at times when others are not. Generally, however, one might be able to argue that the postclassical period stretches from the early thirteenth century to the early twentieth century CE. I use the expression “late” simply to indicate that, as far as I can tell, the observations of this article do not pertain to Ḥanafī texts from the thirteenth-sixteenth centuries CE (i.e., to an earlier phase of the postclassical period).


Examples of such theoretical propositions would include “The obligation imposed for the performance of an act includes the prohibition of its contrary,” “The occasioning cause of an obligation that is valid over a temporal range is deferred to the last moment of that range,” “If an obligation is valid over a temporal range, then the totality of that range is valid for its performance,” and so on. These are all universal statements whose validity would be required for the validity of the legal effects that are their specific cases.


The question of how legal theory allowed one to upgrade probable epistemological states or to derive laws from probable indicants or to deploy methods below the grade of certainty to extract legal effects or to formulate preferred rules out of a body of available options, etc. has been discussed in a number of compelling works. See, for example, Aron Zysow, The Economy of Certainty: An Introduction to the Typology of Islamic Legal Theory (Ph.D. Dissertation, Harvard University, 1984); Wael Hallaq, “On Inductive Corroboration, Probability and Certainty in Sunnī Legal Thought,” in Islamic Law and Jurisprudence: Studies in Honor of Farhat J. Ziadeh, ed. by Nicolas Heer (Seattle: University of Washington Press, 1990), 3–31; Talal al-Azem, Rule-Formulation and Binding Precedent in the Madhhab-Law Tradition: Ibn Quṭlūbughā’s Commentary on the Compendium of Qudūrī (Leiden: Brill, 2017). See also the forthcoming article by Feriel Bouhafa on Averroes’ probabilistic epistemology, especially as it pertains to the unit-tradition and its casting within the discipline of rhetoric: Feriel Bouhafa, “Averroes’ Corrective Philosophy of Law,” in Interpreting Averroes, ed. by Peter Adamson and Matteo di Giovanni (Cambridge: Cambridge University Press, forthcoming).


As we will see below, the fact of the underdetermination of any legal theory within the system is brought into sharp relief due to the higher-order reflections that are facilitated by the deployment of a logical and philosophical apparatus. It is worth considering whether a similar perspective of earlier uṣūlīs has simply remained hidden from view.


Zysow, Economy, 1.


Indeed the influence of falsafa is already rather significant and easy to detect no later than the sixth/twelfth century. See, for example, Fakhr al-Dīn al-Rāzī, al-Maḥṣūl fī ʿilm uṣūl al-fiqh, ed. by Ṭ. Fayyāḍ al-ʿAlwānī (Beirut: Muʾassasat al-Risāla, 1992): on semantic theory (Nora Kalbarczyk, Sprachphilosophie in der islamischen Rechtstheorie. Die avicennische Klassifikation der Bezeichnung bei Faḫr ad-dīn ar-Rāzī, forthcoming (Leiden: Brill, 2018)); on the discussion of the quiddity of the invitation to an action (māhiyyat al-ṭalab), where not only the language of essentialism, but also the much-discussed principle of logic that absurdities imply absurdities is deployed (a-Rāzī, al-Maḥṣūl, II: 18–19); on the discussion of the ontological modalities of things commanded (al-Rāzī, al-Maḥṣūl, II: 271 ff.). The case is similar with the deployment of various distinctions that appear in falsafa and especially logic. On legal dialectics that display a heavy dependence on the discipline of logic (manṭiq) and were enshrined in the ādāb al-baḥth genre, see now Walter Edward Young, The Dialectical Forge: Juridical Disputation and the Evolution of Islamic Law (Cham: Springer International Publishing, 2016). For further on the absorption and impact of logic on legal theory, see Wael Hallaq, “Logic, Formal Arguments and Formalization of Arguments in Sunni Legal Theory,” Arabica, 37.3 (1990): 315–58.


This would be the position of those legal theorists and traditions that Zysow calls formalist. Materialists, such as the Ẓāhirīs, for example, would simply not admit probability into the legal system. As such, they also generally eschew legal mechanisms that are susceptible to uncertainty, such as analogical reasoning. Zysow, Economy, 5.


Zysow, Economy, 34.


ʿAbd al-ʿAlī Baḥr al-ʿUlūm’s commentary and Muḥibballāh al-Bihārī’s base text present the argument in the following fashion (throughout this article, italics represents the base text and roman represents the commentary): “The multiplicity of unit reports that share meaning—though this may be by means of [signification] by implication—i.e., although the meaning may be by implication—necessitates knowledge with respect to the extent [of meaning] that is shared among those unit reports. [It is stated by the commentator at this stage that this is something that is known intuitively and needs no proof, after which claim the argument proceeds as follows.] I say that there is a doubt here that depends on a premise, namely, that if it is possible for each of the instances of a universal to be non-existent—singly or together—then it would be possible for this universal also to be nullified; otherwise … the possibility of the Platonic Form would be allowed. These are quiddities that exist divested of individuation. [Further,] the reason for the implication [that if it is possible for each of the instances of a universal to be non-existent—singly or together—then it is possible for the universal to nullified] is that if the existence of the universal along with the nullification of the totality of [its] instances, each taken separately or as together, is possible, then the existence of the universal without individuation would be possible …” (ʿAbd al-ʿAlī Muḥammad b. Niẓām al-Dīn (Baḥr al-ʿulūm), Fawātiḥ al-raḥamūt bi-sharḥ Musallam al-thubūt li-l-Imām al-Qāḍī Muḥibballāh b. ʿAbd al-Shukūr al-Bihārī, ed. by ʿAbdallāh Maḥmūd Muḥammad ʿUmar (Beirut: Dār al-Kutub al-ʿIlmiyya, 2002), 2: 146). Thus one begins with the general principle/hypothesis (1) that “If unit reports share meaning, then they generate knowledge to the extent this meaning is shared.” Knowledge, in this context, amounts to certainty. This principle is grounded in another principle, the contrary of which must be disproved, namely, (2) that “It is not the case that if it is possible for each of the instances of a universal to be non-existent, then it is possible for the universal to be non-existent.” This principle requires disproving another principle, namely, (3) that “if the existence of the universal, along with the nullification of the totality of [its] instances, is possible, then it is possible for the universal to exist without individuation.” In this example, each principle is underdetermined in relation to the evidentiary theory of the lower level. The resolution provided in the text that ultimately determines the system is rather typical: “The most that can be said in response is that it—i.e., the extent [of meaning] that is shared among the transmitted unit reports—is known [with certainty] not because it is true in an absolute sense, on the basis of reason, lest what you stated be used in refutation. Rather, it is only known by means of habit/custom. For divine habit/custom generates knowledge after experience and repetition.” Put differently, the series of sequential hypotheses/principles is determined as a whole on the basis of a final argument that steps outside the series and the whole. The impact of the apparatus of logic and falsafa is obvious and does not seem to require comment.


Zysow, Economy, 39.


Zysow, Economy, 39.


This principle above is as underdetermined by the evidence as the one noted by Ghazālī: “If a bird flies past you and you think that it is a raven, then I have made so-and-so an obligation for you and have made your presumption a sign of the obligation to act.” This principle itself needs another principle to ground it. Zysow mentions this statement by Ghazālī at Zysow, Economy, 39. The position is a cognate of—though not identical to—the one adopted by certain Muʿtazilites, who argue that there may be a benefit in the act when it is performed in a state of presumption. It is such benefit that confers an obligatory status to the act.


At the end of his quotation from Ghazālī, Zysow, Economy, 40, writes, “It is because our own state of being is a condition for the obligation of action that the uncertain authenticity of the tradition is of no importance. All that matters is our knowledge of the master-rule, and that is certain.” Again, the later tradition would grant that such a principle would produce actionable certainty; and it would also assert that such a principle is certain. But this certainty would be taken to be conditional on hypotheses of an upper layer and underdetermined by virtue of its evidence. In such cases, certainty is better understood as constrained, underdetermined, and conditional validity, as we will see below.


Muḥibballāh b. ʿAbd al-Shakūr/al-Shukūr al-Bihārī was a Ḥanafī scholar of the eleventh/seventeenth century north India. He was born and raised in Karā in Bihār and gained scholarly prominence in the second half of the reign of Awrangzīb (r. 1068/1658–1118/1707). Prior to the Musallam, he had penned a demanding and celebrated logic text, the Sullam al-ʿulūm. Very likely starting in the 1090s/1680s, he secured successive and enviable appointments at the Mughal administration and court. Towards the very end of his life, he was given the title Fāḍil Khān by the emperor Shāh ʿĀlam (r. 1119/1707–1124/1712) soon after he had installed him in the central ministry. He died in 1119/1707. See Asad Q. Ahmed, “The Sullam al-ʿulūm of Muḥibb Allāh al-Bihārī,” in The Oxford Handbook of Islamic Philosophy, ed. by Khaled El-Rouayheb and Sabine Schmidtke (New York: Oxford University Press, 2016), 488–508.


ʿAbd al-ʿAlī b. Niẓām al-Dīn Baḥr al-ʿUlūm was a twelfth/eighteenth century scholar of fiqh, uṣūl al-fiqh, manṭiq, falsafa, and kalām. He was considered a mujtahid in matters of legal interpretation among the Ḥanafī scholars of his period and was also a recognized murshid (spiritual guide) of the Qādirī order. Baḥr al-ʿulūm is known for a considerable number of writings in some of the aforementioned fields and enjoyed the patronage of a number of princely states and the British East India Company throughout his career. He died in 1225/1810. Asad Q. Ahmed, “Baḥr al-ʿUlūm, ʿAbd al-ʿAlī,” in Encyclopedia of Islam III, ed. by Kate Fleet, Gudrun Krämer, Denis Matringe, John Nawas, and Everett Rowson. Consulted online on 11 October 2017


ʿAbd al-Ḥaqq b. Faḍl-i Ḥaqq al-Khayrābādī was a thirteenth/nineteenth century scholar of South Asia. Born in 1244/1828, he received the bulk of his training from his equally-celebrated father and wrote several commentaries and glosses on well-known texts in the rationalist disciplines. These included a gloss on the commentary of Mubārak on the Sullam al-ʿulūm of al-Bihārī, a gloss on the commentary of Ḥamdallāh b. Shukrallāh on this same text, and a gloss on al-Harawī’s gloss on al-Jurjānī’s commentary on al-Ijī’s Mawāqif. He died in 1316/1899. See Usayd al-Ḥaqq al-Qādirī Badāyūnī, Khayrābādiyyāt (Budaun: Tajul Fuhool Academy, 2011), 24–25.


The text of the Musallam al-thubūt, along with Baḥr al-ʿulūm’s commentary, al-Fawātiḥ, was published with the Mustaṣfā of al-Ghazālī. For ease of reference I mention the book only with reference to the work of al-Bihārī. Muḥibballāh b. ʿAbd al-Shukūr al-Bihārī, Musallam al-thubūt (Egypt: Būlāq, 1322 AH), I: 8.


The following is an instructive and relevant passage from Avicenna’s Najāt on the subject matters of disciplines: “As for subject matters, well, in the sciences, they are those posited things whose essential accidents are sought [by the investigator]. [Their examples are] measureable [dimensions] for geometry, number for arithmetic, body, insofar as it moves and is at rest, for physics, and the Existent and the One for theology. Each one of these [subject matters] has essential accidents that are particular to it, such as range, boundary, and form for measureable [dimensions], even and odd for number, change, growth, wilting, etc. for the natural body, and potentiality, actuality, perfection, deficiency, generation, and eternity and similar things for the Existent. The subject matter [of a science] may be one, such as the natural body, or it may be many things, homogeneous or commensurate, such as line and surface for geometry.” See Avicenna, Avicenna’s Deliverance: Logic, translated by Asad Q. Ahmed, with an Introduction by Tony Street (Karachi: Oxford University Press, 2011), 102. (“sought” above should properly be “investigated”.)


The lexical definitions of uṣūl, as discussed by various legal scholars, is conveniently summarized by Badr al-Dīn al-Zarkashī, al-Baḥr al-muḥīṭ fī uṣūl al-fiqh, ed. by ʿAbd al-Qādir al-ʿĀnī (Kuwait: Wizārat al-Awqāf, 1992), I: 15 ff. Most non-technical definitions share the family feature that an aṣl is that upon which something else depends.


al-Bihārī, Musallam, I: 8. Baḥr al-ʿulūm offers the following explanations and examples. (1) Figurative language has literal language as its aṣl; (2A) Scripture has preponderance over analogy; (2B) the purity of water is that the continuity of which is presumed in the absence of proof to the contrary; (2C) that the fāʿil is in the nominative is a rule of grammar; and (2D) “Pray!” is an indicant of the obligation of prayer. These well-known definitions are also mentioned, for example, in the enumeration of Zarkashī, al-Baḥr, I: 16 ff., on the authority of al-Qarāfī.


This is the Mukhtaṣar of Ibn Ḥabīb on the Manār of al-Nasafī. Both were popular legal theory texts in South Asian Ḥanafism. The Nūr al-anwār of Mullā Jīwan was perhaps the most widely read South Asian commentary on the Manār.


al-Bihārī, Musallam, I: 8.


Musallam, I: 8; Khayrābādī, 7.


The modern partial Urdu commentary on the Musallam identifies this scholar as Amānallāh al-Banārasī (d. 1133/1721), who appears to be a frequent target in the Musallam. See Mawlawī Hidāyatallāh, Baḥr al-hidāyat: urdu sharḥ musallam al-thubūt (no further publication information available), 22–3.


al-Bihārī, Musallam, I: 8.


al-Bihārī, Musallam, I: 8–9.


Avicenna, Deliverance, 98.


Thus knowledge that is a priori, for example, even if it pertains to legal matters, would not be considered fiqh. In addition, derived knowledge that is based purely on reason would also be excluded from this class. On the other hand, the application of reason in the derivation of knowledge, insofar as this knowledge is grounded in a textual source base (i.e., a scriptural base) would still be considered fiqh. It is due to these distinctions that I have chosen to translate sharʿī as “textual” and not as religious or legal. For the former would also include certain branches of philosophical theology and the latter would not exclude certain legal norms such as the obligation to believe. For further on the classification of fiqh among the ʿulūm sharʿiyya, as opposed to the ʿulūm ḥikmiyya, see Ṣiddīq Ḥasan Khān, Abjad al-ʿulūm, ed. by ʿAbd al-Jabbār Zakkār (Damascus: Wizārat al-Thaqāfa, 1978), I: 58 ff. There were of course several classification schemes elaborated in the tradition.


al-Bihārī, Musallam, I: 10: wa-l-fiqh ḥikma ay amrun wāqiʿiyyun … mutafarriʿa ʿalā l-īmān bi-dh-dhāt wa-ṣ-ṣifāt … thābita bi-adilla sharʿiyya. ʿAbd al-Ḥaqq al-Khayrābādī, Sharḥ Musallam al-thubūt maʿa sharḥ jumlat al-minhiyyāt/munhiyyāt (Quetta: Maktaba Islāmiyya Mīzān Markit, nd), 10, points out a couple of problems associated with the definition of fiqh as ḥikma, the first of which is related to the observations of Zysow. The latter term, he explains, is used only to refer to knowledge that is certain, whereas analogy (qiyās), for example, only generates overwhelming opinion (ghalabat al-ẓann) in fiqhī matters. Given this, if fiqh is ḥikma, then a broad range of its problemata would be excluded from its purview. The solution, which brings us to the crux of the matter that will be discussed below, is upgrading by the following move: “such and such a legal norm is something to which analogy and the unit-report lead; everything to which analogy and the unit-report lead is a legal norm from God.” Thus the legal norm is valid and the state of overwhelming opinion is only a means to it. As we will see below, the universal major premise is precisely the subject matter of the discipline and it is taken to be underdetermined.


See al-Bihārī’s logic work, Sullam al-ʿulūm, which is the obvious backdrop and impetus to Baḥr al-ʿulūm’s commentarial interventions at this juncture: “If [knowledge] is a belief in a predication relation [between a subject and a predicate], it is an assent and judgment (ḥukm). Otherwise, it is a simple conception (taṣawwur sādhij).” In other words, a ḥukm is the knowledge that there exists a relation of affirmation or negation between two items—the subject and the predicate—in a proposition. The translation is from Asad Q. Ahmed, Palimpsests of Themselves: Philosophical Commentaries in Postclassical Islam (forthcoming), Part III.


al-Bihārī, Musallam, I: 10.


al-Bihārī, Musallam, I: 11.


These are indicants that are general and not specific to any particular case. I use the expression “compressed” to stress that it is in virtue of being embraced by them that expressed and particular indicants yield conclusions and become congruent with them. In other words, by virtue of their generality, such compressed indicants potentially embed all the relevant particular ones. See also Hidāyatallāh, Baḥr, 23–4.


al-Bihārī, Musallam, I: 9. The formulation, which is not entirely novel, comprises key features that prompt commentarial interventions (on which see below). A similar introduction to uṣūl al-fiqh, for example, can be found in Jalāl al-Dīn al-Suyūṭī’s Niqāya: “The discipline of uṣūl al-fiqh [comprises] the compressed indicants [of fiqh] (ʿilm uṣūl al-fiqh adillatuhu al-ijmāliyya).” His self-commentary elaborates that the compressed indicants are “non-specific (ghayr muʿayyana), such as commanding and forbidding simpliciter, the acts of the Prophet … as opposed to the expressed indicants [of fiqh], such as ‘Pray!’ and ‘Do not approach fornication! …’ ” The work is embedded in the margins of Abū Yaʿqūb al-Sakkākī, Kitāb Miftāḥ al-ʿulūm wa-maʿahu Kitāb Itmām al-dirāya li-qurrāʾ al-Niqāya (Egypt: al-Maṭbaʿa al-Maymaniyya, nd), 76. Common—but not identical—features of the definition are also mentioned by al-Zarkashī, al-Baḥr, I: 16 ff.: “An aṣl is that whereby the status (ḥukm) of something else is known and a farʿ is that which is known on the basis of the status of something else.” Interestingly, in terms of signification, al-Zarkashī mentions a report on the authority of al-Māwardī’s al-Hāwī that “an aṣl is that which something else signifies (dalla ʿalayhi) and the farʿ is that which signifies something else.” For example, the derivative element (smoke) signifies the existence of the cause (fire).


Avicenna, Deliverance, 85.


Avicenna, Deliverance, 84.


Avicenna, Deliverance, 42: “A syllogism is a statement composed of [other] statements. When they are posited, a statement other than them follows from them. [The new statement is generated] through these [statements] themselves, not by accident, but by necessity.”


A first figure syllogism has the form: “A is B; B is C” and it entails that “A is C”.


I have argued elsewhere that base texts served as deliberate prompts and hints for future commentarial openings and disputes (See Asad Q. Ahmed, “Postclassical Philosophical Commentaries: Innovation in the Margins,” Oriens, 41.3–4 (2013), 317–48). In the case at hand, we noted al-Bihārī’s emphasis on the derivative nature of textually-grounded judgments (aḥkām) for the latter to count as fiqhī propositions. Now though the argument for the derivative nature of the judgments is not novel within uṣūl, the strong pivot to the logical framework certainly is unprecedented. This kind of emphasis can be located, for example, in Avicenna’s explanation of the conclusions that follow from syllogisms: “The meaning of ‘follows’ is that assent [to the truth of the new statement] is granted and that [this new statement] must be inferred due to the assent granted to the premises and their form. If [the truth of this derived statement] is apparent in itself and a syllogism [composed] out of premises of a similar apparent status is applied [to the task of generating it], this would not be a true syllogism” (Avicenna, Deliverance, 42). In other words, the conclusions of syllogisms must depend for their validity on the validity of the premises. If the premises and the conclusion are of the same epistemological grade, such that assent to the premises does not in itself confer validity to the conclusion, then such a construct would not be counted as a syllogism. As we noted above, the aḥkām of fiqh are treated in the same fashion in al-Bihārī. The following explanation is given in the logic text, the Sullam al-ʿulūm, of al-Bihārī; it may well have served as a direct commentarial inspiration. The work was composed before the Musallam and also received commentaries from Baḥr al-ʿulūm and ʿAbd al-Haqq al-Khayrābādī. “[Dalīl] is limited to three [types]. Its underlying foundation is the syllogism, which is a statement composed of propositions, from which, due to their [very selves,] another statement is entailed.” Asad Q. Ahmed, Palimpsests, Part III, whence the translation is extracted.


al-Bihārī, Musallam, I: 9.


al-Bihārī, Musallam, I: 9.


Al-Khayrābādī, Sharḥ, 8. In the quotation from Baḥr al-ʿulūm above, this would mean that problema B would stand as evidentiary in relation to problemata C, D, E, and F. The latter, in turn, would stand in a similar relation to other problemata.


The commentator builds his argument in the course of an exploration of the disagreement between al-Bihārī and Amānallāh al-Banārasī on whether the relation of uṣūl to fiqh is analogous to the relation of logic to philosophy (more on this below). The inspiration may well go back also to Baḥr al-ʿulūm’s pithy statement that the major premise is taken from a problema of uṣūl and that the question of its universality is taken from a range of uṣūlī problemata. In other words, there are two issues under consideration: how does the major premise relate to its evidence and how does the universality of the major premise relate to its evidence. See Hidāyatallāh, Baḥr, 24–5.


Hidāyatallāh, Baḥr, 25.


Kit Fine’s theories of grounding offer a very helpful theoretical perspective for articulating the issues at hand. Take the following case as an example: The fact that the particle is accelerating obtains in virtue of the fact that it is being acted upon by some net positive force. Now this claim may be rendered with a modal connection between the explanandum and explanans as: (i) Necessarily, if the particle is acted upon by some positive force, then it is accelerating. But the modal connection is not explanatory and determinative with respect to the antecedent and consequent. It asserts that a fact (consequent/explanandum) necessarily obtains in virtue of another fact (antecedent/explanans). (This is similar to the assertion that if something is commanded (antecedent/explanans) then it is obligatory (explanandum/consequent).) A stricter account is needed that would fill the gap between the antecedent and consequent. For example, if one considers that the consequent of (i) strictly holds in virtue of the increasing velocity of the object, then perhaps one may have filled the gap. Thus a second determinative claim is needed to ground the first one. In principle, unless a strict metaphysical necessity is at hand, the process can devolve ad infinitum. No such metaphysical necessity seems to ground the sequence of uṣūl problemata. The question of whether the mabādī kalāmiyya (kalām postulates/principles) and the mabādī lughawiyya (linguistic postulates) do so and are also taken to do so, on theoretical grounds, by the uṣūlīs would be fruitful to pursue. The problem would be that the former are not textually grounded, so that they cannot generate sharʿī judgments; and the latter are textually-grounded, so that, though they can generate sharʿī judgments, they lack independent grounds for their own validity. See Kit Fine, “Guide to Ground,” in Metaphysical Grounding: Understanding the Structure of Reality, ed. by Fabrice Correia and Benjamin Schnieder (Cambridge: Cambridge University Press, 2012), 37–80. I would like to thank Mohammad Saleh Zarepour for bringing this work to my attention.


Reading uṣūl for fiqh for sense. The reading I offer emerges two lines below in Hidāyatallāh’s elaboration. Hidāyatallāh, Baḥr, 25.


Hidāyatallāh, Baḥr, 25.


A briefer version of his argument is presented in al-Khayrābādī, Sharḥ, 8–9.


al-Bihārī, Musallam, I: 10.


Avicenna, Deliverance, 32.


I follow the al-Bihārī, Musallam, I: 9, which asserts the equivalence of the major premise with the conditional rule: “So would be the case when it is rendered (wa-kadhā in ḥurrira) by means of an exceptive syllogism: ‘If alms are commanded, then they would be obligatory; well, the antecedent is true; so alms are obligatory.’ The implication rule (mulāzama) is taken from our statement ‘Command confers obligation’ ”. The implication rule is nothing other than the major premise “Whatever is commanded by God is obligatory,” as discussed above.


Arguments against this differentiation between uṣūl and logic are presented in al-Khayrābādī, Sharḥ, 8 ff. The gist is that the position regarding the difference between the two disciplines lies in the claim that expressed indicants of philosophy are substrates of second intentions, the subject matter of logic, only with respect to their form, not with respect to their matter. This is so because second intentions exist only mentally, whereas the expressed indicants of philosophy may have extramental existence. Thus logic delivers only a formal congruence between the expressed indicant of philosophy (or any discipline) and its conclusions. The response, presumably from Amānallāh al-Banārasī, aims to undercut the claim upon which the argument of distinction is based. According to him, a problema of uṣūl such as “If something is commanded, it is for an obligation” is actually a statement only about the form of commands (not about the matter) and it is only as such that the expressed indicant “Give alms!” is its instance. Thus there is indeed a parallel between logic and uṣūl. Both confer formal validity. The argument is not accepted that, as an instance of the subject matter “commands” of uṣūl, “Give alms!” which is formally and materially extramental, confers extramental status to the subject matter. The reason is that the fact that an instance is extramental does not necessitate that the subject matter should also be so. For example, man is an extramental instance of species; yet the latter is a second intention. This position of Amānallāh is rejected.


al-Bihārī, Musallam, I: 9. This is a standard definition of the syllogism that, in the formal sense, is being deployed to explain qiyās. I have, therefore, not translated it as (legal-) analogical reasoning.


al-Bihārī, Musallam, I: 9.


Here al-Khayrābādī, Sharḥ, 9 f. quotes the instructive self-gloss of al-Bihārī: “[You may] say this, i.e., that the need for logic with respect to the three indicants other than qiyās [—the Qurʾān, Sunna, and Consensus—] is evident. [But] as for with respect to qiyās, well there is no [need for it], because the investigation in each of these two disciplines, i.e, uṣūl and logic, pertains only to the nature of qiyās. So there is no need for logic in [the discipline of uṣūl], given that, necessarily, each problema from among the problemata of analogy (tamthīl) that is needed [in the investigation] of the qiyās fiqhī is mentioned in that investigation [i.e., in uṣūl]. [In response,] I say that let us grant that the [case] on the surface is so; however, the verification is [as follows]. A problema of uṣūl is: Whatever is established by means of qiyās is a legal norm/judgment from God. Thus the qiyās obtains as follows: This is a judgment of God because [1] it is established by a qiyās; and [2] whatever is established by a qiyās is a judgment of God. Thus logic is needed [for determining] the quality that comes to inhere in the two premises that are brought together.” The point is that, though logic is needed for conferring formal validity to the argument, the validity that is operative for fiqh is found only in uṣūl. This is the case even when the two disciplines appear to be engaging in the same investigation. Al-Khayrābādī proceeds to refine the argument, claiming that, though there is a kind of similarity between the relation of logic to philosophy and of uṣūl to fiqh, there is no identity.


The issue may be placed in loose analogy with Kit Fine’s distinction between a metaphysical necessity and an explanatory relation. For example, the claim that a certain act is either right or wrong is grounded in the fact that it is right. For if it is right, then, by irreducible necessity, it is either right or wrong. Now in virtue of what is the act right? Well, one might posit that it is right in virtue of the fact that it delivers benefit; this would be its normative explanation. Thus, by metaphysical necessity, if an act delivers benefit, then it is right or wrong. However, the fact of it delivering a benefit is not that in virtue of which it is right or wrong. In similar fashion, one might say that perhaps a formal qiyās ultimately grounds the relationship between an act and its legal status in a metaphysically irreducible manner, but this does not mean that it has explanatory status in a sharʿī manner. One would say, for example, “if qiyās is a valid tool of derivation, then such and such a legal status is predicated of such and such an act; and this is metaphysically true.” However, this does not mean that the legal status of the act is true in virtue of qiyās, i.e., the formal qiyās does not have the relevant fiqhī explanatory value. Such would be the case only if the qiyās is itself first shown to stand in the relevant fiqhī relation to its conclusion by means of a materially-invested argument that grounds the claim of this relation. Of course the latter argument’s grounding claim would itself need similar grounding. And so on. See Kit Fine, “Grounding,” 40–45. Hallaq, “Inductive Corroboration,” 24 ff., presents the case of al-Shāṭibī, who aimed to ground the certainty of general theoretical legal principles in a theory of induction. This effort comes close to what is described in this article. However, the crucial difference lies in the fact that the sources surveyed here would argue that the inductive principle may lend formal certainty, but not legally-relevant material validity to the principle. Thus whereas in al-Shāṭibī’s case one might be able to argue for an independent grounding theory, in the cases at hand, any grounding theory itself remains ungrounded. However, as I have argued in this article, all theories of this sort would be systemically and relationally valid.


  • Ahmed, Asad Q. “Postclassical Philosophical Commentaries: Innovation in the Margins,” Oriens 41.3–4 (2013): 317–48.

  • Ahmed, Asad Q. “The Sullam al-ʿulūm of Muḥibb Allāh al-Bihārī.” In The Oxford Handbook of Islamic Philosophy, ed. by Khaled El-Rouayheb and Sabine Schmidtke. New York: Oxford University Press, 2016, 488–508.

  • Ahmed, Asad Q. “Baḥr al-ʿUlūm, ʿAbd al-ʿAlī.” In Encyclopedia of Islam III, ed. by Kate Fleet, Gudrun Krämer, Denis Matringe, John Nawas, and Everett Rowson. Consulted online on 11 October 2017

  • Avicenna. Avicenna’s Deliverance: Logic. Translated by Asad Q. Ahmed, with an Introduction by Tony Street. Karachi: Oxford University Press, 2011.

  • Al-Azem, Talal. Rule-Formulation and Binding Precedent in the Madhhab-Law Tradition: Ibn Quṭlūbughā’s Commentary on the Compendium of Qudūrī. Leiden: Brill, 2017.

  • Badāyūnī, Usayd al-Ḥaqq al-Qādirī. Khayrābādiyyāt. Budaun: Tajul Fuhool Academy, 2011.

  • Baḥr al-ʿulūm, ʿAbd al-ʿAlī Muḥammad b. Niẓām al-Dīn. Fawātiḥ al-raḥamūt bi-sharḥ Musallam al-thubūt li-l-Imām al-Qāḍī Muḥibballāh b. ʿAbd al-Shukūr al-Bihārī, ed. by ʿAbdallāh Maḥmūd Muḥammad ʿUmar. Beirut: Dār al-Kutub al-ʿIlmiyya, 2002.

  • Al-Bihārī, Muḥibballāh b. ʿAbd al-Shukūr, Musallam al-thubūt. Egypt: Būlāq, 1322 AH.

  • Bouhafa, Feriel. “Averroes’ Corrective Philosophy of Law.” In Interpreting Averroes, ed. by Peter Adamson and Matteo di Giovanni. Cambridge: Cambridge University Press (forthcoming).

  • Fine, Kit. “Guide to Ground.” In Metaphysical Grounding: Understanding the Structure of Reality, ed. by Fabrice Correia and Benjamin Schnieder. Cambridge: Cambridge University Press, 2012, 37–80.

  • Hallaq, Wael. “On Inductive Corroboration, Probability and Certainty in Sunnī Legal Thought.” In Islamic Law and Jurisprudence: Studies in Honor of Farhat J. Ziadeh, ed. by Nicolas Heer. Seattle: University of Washington Press, 1990, 3–31.

  • Hallaq, Wael. “Logic, Formal Arguments and Formalization of Arguments in Sunni Legal Theory.” Arabica 37.3 (1990): 315–58.

  • Hidāyatallāh. Baḥr al-hidāyat: urdu sharḥ musallam al-thubūt. (no further publication information available).

  • Kalbarczyk, Nora. Sprachphilosophie in der islamischen Rechtstheorie. Die avicennische Klassifikation der Bezeichnung bei Faḫr ad-dīn ar-Rāzī. Leiden: Brill, forthcoming, 2018.

  • Khān, Ṣiddīq Ḥasan. Abjad al-ʿulūm, ed. by ʿAbd al-Jabbār Zakkār. Damascus: Wizārat al-Thaqāfa, 1978.

  • Al-Khayrābādī, ʿAbd al-Ḥaqq. Sharḥ Musallam al-thubūt maʿa sharḥ jumlat al-minhiyyāt/muhiyyāt. Quetta: Maktaba Islāmiyya Mīzān Markit, nd.

  • Al-Rāzī, Fakhr al-Dīn. al-Maḥṣūl fī ʿilm uṣūl uṣūl al-fiqh, ed. by Ṭ. Fayyāḍ al-ʿAlwānī. Beirut: Muʾassasat al-Risāla, 1992.

  • Al-Sakkākī, Abū Yaʿqūb. Kitāb Miftāḥ al-ʿulūm wa-maʿahu Kitāb Itmām al-dirāya li-qurrāʾ al-Niqāya. Egypt: al-Maṭbaʿa al-Maymaniyya, nd.

  • Young, Walter Edward. The Dialectical Forge: Juridical Disputation and the Evolution of Islamic Law. Cham: Springer International Publishing, 2016.

  • Al-Zarkashī Badr al-Dīn. al-Baḥr al-muḥīṭ fī uṣūl al-fiqh, ed. by ʿAbd al-Qādir al-ʿĀnī. Kuwait: Wizārat al-Awqāf, 1992.

  • Zysow, Aron. The Economy of Certainty: An Introduction to the Typology of Islamic Legal Theory. Ph.D. Dissertation, Harvard University, 1984.

Content Metrics

All Time Past Year Past 30 Days
Abstract Views 255 45 0
Full Text Views 280 40 6
PDF Views & Downloads 83 40 8