Skip to main content Accessibility help
Hostname: page-component-59b7f5684b-hd9dq Total loading time: 1.089 Render date: 2022-10-03T03:16:36.603Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "useRatesEcommerce": false, "displayNetworkTab": true, "displayNetworkMapGraph": false, "useSa": true } hasContentIssue true

6 - Canonicity As a Defining Feature of Legal and Religious Discourse

A Programmatic Essay

from B - Legal-Religious Language

David C. Flatto
Hebrew University of Jerusalem
Benjamin Porat
Hebrew University of Jerusalem


We define canonicity as a variable property of all cultural discourses: it refers to the extent to which the components of a given discourse – its terms, texts, practices, etc – tend to inhere within it over time. Law and religion may both be characterized as strongly canonical discourses, meaning that they tend not to change over time. This property stems from the particular functions that law and religion play within human culture; hence it is not dependent on the political ends to which law and religion are often used.

We follow Peter Berger’s assertion that any religion posits a particular conception of the cosmos; hence religion is irreducibly descriptive. In contradistinction to science, the descriptive elements of religion are generally not falsifiable by empirical observation. This resistance to falsification is what makes religion strongly canonical.

In contrast, legal discourse is irreducibly prescriptive. It is distinct from other prescriptive systems, such as morality, in the paramount value it places on consistency, both at any given moment and over time. This diachronic consistency is reflected in the fact that, once established, laws tend to remain in force until formally changed or annulled; this is the strong canonicity of law.

Publisher: Cambridge University Press
Print publication year: 2022

I Introduction

The theme of this volume – the affinity between law and religion – may be productively analyzed at a range of resolutions, from a microhistorical focus on specific religio-legal phenomena, to a broader perspective on how law and religion interact within particular societies or cultures, to an overarching comparison of law and religion as cultural discourses. However, the broader the scope of the analysis, the more pressing the issue of methodology becomes. Specific phenomena may be examined using any number of academic disciplines (history, sociology, cultural studies, gender or race studies, etc.), while a broader analysis demands a sure grasp of philosophy of religion, philosophy of law, or both. At its most all-encompassing, a study of the affinity between law and religion requires a methodology distinct from both philosophy of law and of religion which is able to analyze both fields using a single set of tools.

The field of semiotics is particularly well suited to this task, as it offers a technically precise definition of a discourse which can serve as a model for both religion and law. Following a brief survey of key concepts in semiotic theory, we will analyze some of the distinctive features of legal and religious discourse and then narrow our focus to an important feature that they share, namely their conservative character. Our goal is to develop a methodological framework for analyzing this common feature of law and religion. We will define conservativism in terms of the general properties of semiotic systems, and argue that this characteristic is an intrinsic feature of both religion and law, rather than a function of particular social or cultural circumstances.

II Defining Discourses

The proper subject of semiotics is the sign – any discrete, repeatable representational form. A sign may be simple, which is to say irreducible – a phoneme, a letter, a physical movement; or it may be compound, which to say composed of smaller sign-units – a statement, a text, a ritual, a painting or photographic image. One of the key principles of semiotics is that meaning is not inherent in the form of the sign, but rather the product of relationships between signs. (Umberto Eco proffers that “properly speaking there are not signs, but only sign-functions.”Footnote 1) Hence an important aspect of semiotic analysis is the way that the meaning of any sign is determined by the system of signs within which it is situated. Language is one example of a semiotic system, but any system that conveys information through the medium of discrete, repeatable forms may be treated as the object of semiotic analysis. Semiotics is relevant to law not only because it is articulated using language, but more fundamentally because legal systems consist of a range of physical forms – terms, texts, images, practices, and procedures – which generate meaning and facilitate communication. The same is true of religion, science, the arts, or other material cultures – any of the systems of meaning that constitute human society.

An essential aspect of any sign system is the fact that its structure is that of a network rather than a hierarchy. Ferdinand de Saussure posits that sign systems generate meaning through patterns of difference between the component signs of the system, and it is only these patterns which divide the amorphous mass of semiotic space is into discrete units of signification. Thus, the most basic element of a sign’s meaning is its value – its position within the semiotic network, as defined by the way its semantic range is demarcated from that of other signs in that network.Footnote 2 As Eco explains, semiotic values issue from within the system itself; “they are not defined in terms of their content … but in terms of the way in which they are opposed to other elements of the system and of the position which they occupy within it.”Footnote 3

While Saussure demonstrates the concept of value with regard to simple signs, such as phonemes and words, later semioticians such as A. J. Greimas, Julia Kristeva, and Michel Foucault explain that compound signs are also defined by their position within a network of similar signs. The difference between simple and compound signs is that a compound sign invariably has at least two layers of meaning, the first determined by the rules of a particular code regarding the combination of individual units, the second determined by the sign’s position within the network of other compound signs within which it is situated. Foucault articulates the difference between these two layers in his distinction between a sentence and what he refers to as a “statement” (énoncé):

[T]here is no statement in general, no free, neutral independent statement … : it is always part of a network of statements, in which it has a role, however minimal it may be, to play. Whereas grammatical construction needs only elements and rules in order to operate … the same cannot be said of the statement. There is no statement that does not presuppose others; there is no statement that is not surrounded by a field of coexistences, effects of series and succession, a distribution of functions and roles. If one can speak of a statement, it is because a sentence (a proposition) figures at a definite point, with a specific position, in an enunciative network that extends beyond it.Footnote 4

Analyzed as a sentence, a series of words derives its meaning from the grammatical rules of a given language; analyzed as a statement, the same series of words derives its meaning from its position within a network of other statements – what Foucault refers to as an “enunciative field” or simply as a discourse.Footnote 5 Foucault’s model thus explains how a given sentence may take on a range of meanings beyond its “literal” (i.e., grammar-determined) meaning depending on the discourse within which it is situated. At the risk of oversimplifying, we may say that semiotics as a field takes the set of tools designed to analyze language and uses them to analyze higher level discourses,Footnote 6 a discourse being nothing more than a structured network of compound signs. This is an important aspect of Kristeva’s celebrated concept of intertextuality: the meaning of a literary text is determined not only by its internal structure but also by its relationship to other texts within a literary canon.Footnote 7

Semiotics provides not only a technical definition of a discourse, but also an explanation of what it means for two discourses to be distinct from one another. Consider the relationship between law and morality, two discourses with a particularly complicated relationship. It is tempting to think of law and morality as part of an integrated normative system, in which moral precepts function as the overarching principles for legal adjudication – a position that has been articulated in various forms within modern legal theory.Footnote 8 Yet it is obvious that we frequently assign different moral and legal values to the same action: an act can be morally good but legally prohibited, or morally wrong but legally permissible. In contrast, an act cannot have two antithetical values within the same system: it cannot be simultaneously legal and illegal, or moral and immoral.Footnote 9 Hence to conflate law and morality with one another is to undermine a basic dimension of what it means for something to have legal or moral significance.

We can make the same point with regard to the relationship between religion and science. Peter Berger notes that if one defines religion (following Thomas Luckman) as reflecting the “human capacity for self-transcendence,” one could conceive of modern science as a sort of religion.Footnote 10 Yet Berger rejects this as patently unhelpful, since “[i]f one does that, one is subsequently forced to define in what way modern science is different from what has been called religion by everyone else.”Footnote 11 To wit, we instinctively categorize certain types of propositions about the world as scientific and others as religious. The discourses of religion and science constitute different sets of values (again in the semiotic sense of the term) that do not always correlate with one another.Footnote 12

It is important to state that our concise definition of a discourse does not do justice to the vast internal complexity of each of the discourses we have mentioned. Any legal, religious, moral, or scientific system is made up of many different types of signs – not only spoken and written language, but also actions (practices, procedures, rituals, etc.) and two- or three-dimensional representational forms: art, architecture, food, clothing, etc. Even if we limit our definition of “discourse” specifically to linguistic elements, we are still faced with an enormous variety of terms, statements, texts, and formulae that are part of any cultural system. That is to say, closer analysis reveals each of these discourses to be further divided and subdivided into countless smaller networks of signs.Footnote 13 A fuller analysis of such internal divisions is beyond the scope of this study. However, below we will argue for the heuristic value of a simplified schematic of each of these discourses notwithstanding their internal complexity.

We should also note that value is only one element of a sign’s meaning, reflecting the oppositional relationships between signs. Cultural scholars have often focused exclusively on the semiotic structures created by these oppositional relationships (hence the movement known as structuralism), thereby preventing a full appreciation of what semiotics has to offer to the study of human culture. After all, much of the semiotic functioning of cultural discourses – the way representational forms generate meaning and facilitate communication – comes from positive relationships between signs,Footnote 14 what Charles Sanders Peirce aptly describes as “the translation of a sign into another system of signs.”Footnote 15 We will not address this dimension of semiotic theory, since this chapter is less concerned with the way individual discourses function as systems of meaning than with the differences between distinct discourses and their positions relative to one another within the broader array of contemporary society.

III The Semiotic Concept of Canonicity

The notion that meaning is dependent on a network of individual signs affects how we analyze the evolution of languages and other discourses over time. One of Saussure’s major contributions to linguistics was his insistence that language must be studied first and foremost synchronically as a network of mutually defining signs before it can be analyzed diachronically. This is not because languages are static (they are not), but rather because “language is a system of pure values, determined by nothing else apart from the temporary state of its constituent elements.”Footnote 16 That is to say, from a semiotic perspective, any system of meaning is all structure; without a clear picture of the patterns of difference among the components of the system at a given point in time, there is no “thing” to analyze. Thus the existence of any language or higher-level discourse necessarily depends on there being an accumulated body of signs to constitute it and to ensure its stability over time. For the same reason, Saussure insisted that linguistic evolution can be analyzed only in terms of changes to individual elements, not to the system as a whole: in order to for us to able to speak of change to the system, the structure that is the system cannot be destabilized.Footnote 17

This property of semiotic systems – the fact that signs must inhere within the system in order for it to function – is an important aspect of what we refer to as canonicity. We often speak of a canon in terms of a set of exemplary units in a specific discourse: a body of outstanding works of literature or art, a collection of cases and statutes that are most often cited or studied in legal contexts,Footnote 18 a set of iconic experiments in scientific discourse.Footnote 19 In contrast, canonicity in the sense we are using it is a property of all units of a semiotic system: once a sign is accepted as an element of the system – a word is recognized as meaningful within a given language, a ruling is endorsed as legally authoritative, an experiment is recognized as scientifically significant – it becomes part of the internal structure of that system, and tends to inhere within it over time. Moshe Halbertal has referred to this as “formative canonicity,” the way that terms and texts that have already been incorporated into a discourse continue to be “taught, read, transmitted, and interpreted” over time, effectively functioning as the “shared vocabulary” of a given society or profession.Footnote 20 In effect, the property of canonicity changes the shape of a semiotic system in relation to the axis of time: even when signs enter the system at different points in its history, their relationship to one another may be analyzed synchronically so long as they are in use at the same time. As Krsiteva explains regarding the significance of a literary canon, “Diachrony is transformed into synchrony, and in light of this transformation, linear history appears as abstraction.”Footnote 21

Our use of the term canonicity to refer to the way that semiotic systems retain their constituent elements over time is different than the way the term has been defined elsewhere in contemporary scholarship. For instance, Jonathan Z. Smith defines a religious canon as a structured catalog of religious objects or texts that has undergone “closure,” which inevitably leads to its interpretation and application to a much broader range of phenomena.Footnote 22 Smith is certainly correct that the concept of canon is integral to the use of a set of signs as the basis for a far-reaching exegetical agenda. Yet his association of canonicity with closure is too simplistic: it is true that the initial process of canon-formation requires some sort of closure, roughly meaning a distinction between those forms that are “in” and those that are not, but there are many systems that we think of as canons that do not remain closed, and continue to accrue additional forms over time. For example, a canon of legal texts – those texts that are considered authoritative for legal practice and adjudication – may be expanded though various processes of recognition (executive or legislative actions, judicial rulings, etc.), some of which are invariably part of any legal system.Footnote 23 What distinguishes canons is not their closure to new elements, but rather their tendency to retain their constituent elements once they have been inducted.

While this tendency to retain sign elements is a defining characteristic of systems that we think of as paradigmatically canonical, the fact that any semiotic system depends on the existence of an accumulated body of signs indicates that all such systems must have this tendency to some extent; to wit, canonicity as we have defined it is a universal semiotic property, albeit a variable property. Strongly canonical discourses are conservative in the sense that existing elements to the system are preserved: they tend not to be displaced by new additions to the system nor do they become obsolete on their own. Weakly canonical discourses, on the other hand, may be characterized as progressive, in the sense that the component signs of those systems are frequently – and often actively – either displaced by new signs or simply discarded. However, even weakly canonical discourses are far more conservative than we normally assume them to be. To take the example that is the subject of Saussure’s own work, language is not a system of meaning that we normally think of as canonical, and indeed it tends to be progressive in the way that it jettisons words and grammatical features over time: even texts from fifty years ago frequently strike us as dated. Yet the concept of formative canonicity accurately expresses the fact that the vast majority of any given language remains extremely stable over time, and must remain stable in order for the language to function. Saussure ascribes the inflexibility of language to a number of factors, among them the size and complexity of the networks of relationships between signs, as well as the large number of members of the linguistic community who use them. But regardless, he concludes that language’s potential for change is appreciable only from a longitudinal perspective.Footnote 24

Just as the essential property of canonicity is an inevitable corollary of the mechanisms by which semiotic systems operate, so, too, we may propose that the degree of canonicity of a given semiotic system is determined by its specific semiotic function, the particular type of meaning it generates. The central claim that we will make in this chapter is that legal and religious systems are strongly canonical relative to other human discourses (language, the arts, material culture, science, and morality), and that this aspect of law and religion stems from the essential character of these discourses rather than from contingent factors, such as the political functions that these systems often serve. Obviously we do not mean to say that political authority is irrelevant to the dynamics of religion or law; on the contrary, authoritative bodies are frequently decisive in the process by which religious or legal signs become formally recognized. However, we will propose that once a sign has been integrated into a legal, religious, or any other discourse, its longevity is largely independent of that authority, and dependent instead on the specific character of that discourse and its broader, apolitical function within human society.

In and of itself, the notion that law and religion are conservative discourses is admittedly rather intuitive. The innovative elements of our analysis are our definition of canonicity as a variable property of all sign systems, and the notion that the degree of canonicity of a given discourse is determined primarily by the kind of meaning it serves to convey rather than by contingent factors such as the presence or absence of institutional authority. Our assertion about the strongly canonical character of law and religion is obviously extremely broad, and probably impossible to defend solely through an accumulation of data. Hence instead of focusing on historical or contemporary examples of legal or religious conservativism, we will take the opposite approach: examining the essential elements of both legal and religious discourse to understand why we would expect each of them to retain its constituent elements over time. Our purpose in doing so is not to prove this thesis conclusively, but rather to develop an analytical framework to serve as the basis for future research.

IV A Digression: Two Methodological Objections

However, before we proceed, we need to address two methodological objections that our analysis may raise. First, is it really possible to generalize about so many different types of religions and legal systems? Surely any system’s degree of canonicity is affected by the particular historical context in which it is situated and the specific nature of its component signs. Indeed, given our assumption that canonicity is a variable property of semiotic systems, it would seem reasonable to assume that just as different types of discourses may be strongly or weakly canonical, so, too, different discourses of the same type might not be canonical to the same degree. For example, there may be significant differences in the longevity of different modes of signs – performative (practices and rituals), material (art, architecture, food, clothing, etc.), or verbal signs (words and texts), as well as between oral and written verbal signs. Yet we will propose that these factors are secondary to the overall differences between law and religion on the one hand and other societal discourses on the other, such that religious and legal systems are conservative relative to other discourses within a particular society. If there is one set of contextual parameters that we must place on our analysis, it is that the concept of religion as we will analyze it is distinctive to the modern age, not because modern religion isn’t contiguous with premodern systems of meaning, but because the relationship between religion and other societal discourses – or rather the extent to which religion and other discourses can be considered separate from one another – has changed so drastically in the modern era.

A second, more fundamental objection may be articulated as follows: It is a prevalent view in contemporary scholarship that discourses are not anchored in any sort of context-independent foundation, be it rational, epistemic, or metaphysical. From this methodological starting point, some scholars have asserted that in the absence of another basis for discursive meaning, it is only the dynamics of power that provides a given discourse with its substance or legitimacy. The notion that meaning is a function of power has been asserted in philosophy of religion by Talal Asad,Footnote 25 in philosophy of science by Paul Feyerabend,Footnote 26 and is a prominent strain of thought among critical legal theorists.Footnote 27 According to these scholars, it is not possible to describe a set of principles or characteristics that defines these discourses apart from the political structures that utilize them. The thesis we have formulated specifically downplays the significance of political authority in determining the character of these and other discourses, hence prompting the question: in the absence of reason, metaphysics, or even power as the basis for meaning, what factor allows each of these fields to function as a discrete, coherent discourse?

As we suggested above, the key to understanding the way discourses function is the concept of difference, which Saussure posits as the most basic element of meaning. Just as the meaning of individual signs is generated most fundamentally by patterns of difference within a structured network of signs, so, too, individual discourses acquire meaning through patterns of difference between distinct discourses within a broader cultural framework. Our analysis going forward will be devoted to identifying some of the key characteristics that distinguish between different discourses in contemporary society.

The notion that something as firm as meaning is ultimately rooted in something as amorphous as patterns of difference seems deeply counterintuitive. Depending on one’s inclination, this makes it either a pivotal contribution of semiotics to critical analysis or a reason to be skeptical of semiotic theory. To readers who find themselves in the latter category, and who might still be inclined to view law, for example, simply as a veiled expression of power, we may note that more nuanced analyses of contemporary legal discourse have complicated the picture painted by critical legal scholars, showing how legal devices may also be used to challenge political authority and protect the disenfranchised.Footnote 28 This strongly suggests that the law has an independent dimension of meaning, which in turns allows it to be used for a variety of political ends.

V The Semiotic Character of Legal and Religious Discourse

Because semiotics analyzes any discourse as the amalgam of its constituent signs, understanding how discourses such as law or religion function differently from one another requires us to examine the kinds of individual statements and texts that we consider to be legal or religious in nature. Here we return to the problem of the internal complexity of each discourse: How can we characterize the system based on its constituent components if each discourse is comprised of so many different kinds of linguistic forms?

An instructive example for approaching this sort of problem is R. M. Hare’s analysis of morality in his book, The Language of Morals.Footnote 29 Hare’s approach is not explicitly semiotic, but he grapples with the same basic difficulty we have outlined here: how to articulate a unified definition of morality given the variety of statements within moral discourse. In the face of a pronounced bias among logicians toward indicative propositions (which can be measured against reality for their truth value), Hare argues that all moral judgments – even those formulated as indicative statements – necessarily entail prescriptive statements insofar as their purpose is to regulate behavior: “to guide choices or actions, a moral judgment must be such that if a person assents to it, he must assent to some imperative sentence derivable from it.”Footnote 30 To wit, the function of moral discourse can be fulfilled only if ostensibly descriptive phrases are taken as implicitly prescriptive. In effect, Hare reduces the semiotic variety of moral discourse to a single type of form.

In the course of his analysis, Hare addresses two counterarguments for why indicative statements may independently be considered moral propositions. One counterargument is the concept of moral naturalism, broadly speaking the idea that moral values are a function of the natural world, or at least of human nature, such that morality may be reduced to a set of indicative statements about what is “good” or “right.” Hare makes a compelling case that such statements are never merely descriptions of an ontological state, since the terms “good” and “right” necessarily reflect an evaluative dimension. Less convincing is Hare’s analysis of a second type of indicative statement: moral assessments which allude – descriptively – to what others consider to be moral without the speaker herself offering any evaluative judgment. Hare refers to these as “inverted comma moral judgments,” and rejects them as genuine value judgments.Footnote 31 Yet Hare’s critics point out that his definition would exclude from moral discourse any statement by an amoralist, a moral hypocrite, or even a weak-willed individual regarding moral behavior, since none of these individuals are actually committing to the moral standards they describe; to exclude such a wide range of individuals as not true participants in moral discourse seems rather artificial. It is clear that the relationship between indicative and imperative statements within moral discourse is far more complex than Hare makes it out to be.

Nonetheless, Hare’s assertion regarding the irreducibly prescriptive nature of morality is hard to dismiss entirely. Even if prescriptive statements cannot be implicated in the meaning of every individual moral proposition, we may conclude that they are essential to the semiotic constitution of moral discourse as a whole: the type of meaning we associate with any discourse called “morality” requires it to include a set of directly prescriptive statements, even as it may also include descriptive statements.Footnote 32 The relationship between these essential prescriptive statements and other, descriptive, statements is not one that can be reduced to logical deduction, or even one-to-one correspondence, but Hare is correct that a purely descriptive system cannot serve the basic function that morality does, namely to regulate human behavior. To wit, to claim that a certain statement has no prescriptive implications for anyone is to remove it from the realm of moral discourse. This formulation makes no claim about the source or basis for morality; it simply expresses how language we regard as moral operates as a system of meaning and communication.

This analysis provides a model for a simplified analysis of other types of discourses. For example, even though legal discourse, too, involves many different types of statements, we may say that law is irreducibly prescriptive in the sense that a set of directly prescriptive statements regulating human behavior – what H. L. A. Hart refers to as primary rules – is a necessary (though perhaps not sufficient) characteristic for a discourse to be called “law.”Footnote 33 Alternately, if we follow Lon Fuller’s thought experiment of an embryonic society developing social institutions de novo (he imagines a group of castaways afflicted by collective amnesia),Footnote 34 we might consider even a rudimentary regulatory system – consisting solely of directly prescriptive statements about the kind of behavior permitted, obligated, or prohibited in that society – to be identifiable as “law.” Again, this analysis does not address the source or basis for the law, only the minimum necessary conditions for legal language to be meaningful: all legal statements that are not directly prescriptive are in some sense dependent for their meaning on directly prescriptive statements.

Religion presents a more complex problem than law, given the range of systems that we might choose to include under that heading and the variety of modes of expression found within them: there is hardly an aspect of human culture that has not been touched in some way by religion. This complexity is reflected in the wide-ranging (and often amorphous) definitions that have been proposed by philosophers of religion. One may appreciate the sentiment that formulating a concise definition of religion – one that not only describes religion but also demarcates it from other related discourses – is a fool’s errand.Footnote 35 Yet few would deny that a discourse would not be considered religious without some conception of reality and of humanity’s place in it. This point is underscored by one of the most influential definitions of religion, that of Clifford Geertz: “[A] religion is (1) a system of symbols (2) which acts to establish powerful, pervasive and long-lasting moods and motivations in men (3) by formulating conceptions of a general order of existence and (4) clothing these conceptions with such an aura of factuality that (5) the moods and motivations seem uniquely realistic.”Footnote 36 Or to return to our imagined embryonic society, we might consider a rudimentary system of prescriptive statements to be a legal system, but we would not consider it to be a religion unless it also contained a set of descriptive statements about our existence.Footnote 37 Hence we may define religion as irreducibly descriptive, at least relative to law.

VI Canonicity in Religion and Science

The upshot of delineating religion in this way is that for all of the similarities between religion and law, there are essential aspects of religion that are best appreciated by contrasting it with another irreducibly descriptive discourse, namely science. Scientific discourse, too, consists of many different types of statements, yet it, too, must minimally consist of a set of descriptive statements about the natural world. If religion is strongly canonical in the sense that we defined it above, science constitutes a perfect foil of a descriptive discourse that is weakly canonical: it tends not only to change over time, but to change specifically by discarding of elements of the system. The most famous articulation of this idea is Thomas Kuhn’s analysis of scientific revolutions as paradigm shifts – the rejection of one conceptual framework in favor of another. But Kuhn’s idea merely reflects a broader trend within philosophy of science: since the middle of the last century, scholars have ceased advocating for a strictly cumulative model of scientific development. Rather, scientific progress invariably involves disposing of elements of the regnant paradigm – terms and theories, models, experimental procedures, and even observations – that do not fit with new data collected and outstanding problems solved.Footnote 38

The reason for this dynamic of scientific discourse may be attributed to one of the basic functions that science serves, to describe the workings of nature as accurately as possible; any element that does not fill this function is therefore abandoned. But regardless of the reason, the propensity of scientific discourse to discard elements of the system is reflected in the very different outcomes of scientific and religious upheaval. When fundamental elements of a given religion are called into question, it quite frequently results in the founding of a new religion, sect, or denomination alongside the old one. In contrast, while challenges to fundamental aspects of a given scientific paradigm invariably meet resistance, challenges which stand the test of time end up transforming the field rather than dividing it. Despite the many scientific debates and revolutions that have occurred over the past five hundred years, we still think of the scientific community as overwhelmingly unified (or rather, divided by field of interest rather than by ideology).

The notion that scientific progress is not strictly cumulative was articulated most forcefully by Karl Popper, who identified it as one of the defining elements of science. According to Popper, scientific inquiry is best characterized not by its use of inductive logic but rather by its use of falsification as a criterion for validity: “it must be possible for an empirical scientific system to be refuted by experience.”Footnote 39 It is this element that makes scientific systems so prone to revision and upheaval: “What compels the theorist to search for a better theory … is almost always the experimental falsification of a theory, so far accepted and corroborated.”Footnote 40 Popper’s proposal has been subject to extensive critique and refinement over the past half century. The most significant revision was offered by Imre Lakatos, who asserts that falsification results in the rejection of a theory only when a new, improved theory (one which “offers any novel, excess information compared with its predecessor”)Footnote 41 is available to replace it. (He calls this “sophisticated falsificationism.”Footnote 42) Moreover, Lakatos maintains that falsifiability is meaningful only in the context of an overall “research programme,” consisting of a “hard core” of theory surrounded by auxiliary hypotheses; it is these auxiliary elements that are most subject to rejection due to falsification.Footnote 43 Yet regardless of the weaknesses in Popper’s original thesis, it is difficult to deny that he has identified a critical aspect of scientific discourse. Whether we assess falsifiability on the level of an individual theory, a data set, or an integrated research program, a key element of what makes that unit of discourse meaningful in a scientific sense – what legitimates it as a point fit for discussion and debate within the scientific community – is the possibility that it will be rejected based on contradictory empirical evidence. Simply put, falsifiability is a property of the kind of meaning that we instinctively associate with modern science, and it is this element of science that makes it weakly canonical.

Popper proposed falsifiability not only as a positive description of scientific inquiry, but also as a criterion to demarcate science from pseudoscientific disciplines, such as Freudian psychoanalysis or Marxist historiography. In this regard, too, Popper’s thesis has proven to be insufficient; philosophers of science now doubt that there is a single attribute that we can use to verify a given discipline as a science.Footnote 44 Yet again, Popper’s thesis captures one important element of what distinguishes science from other disciplines, and particularly from modern religious discourse. A key attribute of many religious propositions – those statements we instinctively identify as religious in character – is that they are fundamentally unfalsifiable. Although this point is rarely found in formal scholarly definitions of religion, it is famously the subject of Antony Flew’s 1950 essay, “Theology and Falsification”:

Now it often seems to people who are not religious as if there was no conceivable event or series of events the occurrence of which would be admitted by sophisticated religious people to be a sufficient reason for conceding “There wasn’t a God after all” or “God does not really love us then.” Someone tells us that God loves us as a father loves his children. We are reassured. But then we see a child dying of inoperable cancer of the throat. His earthly father is driven frantic in his efforts to help, but his Heavenly Father reveals no obvious sign of concern. Some qualification is made God’s love is “not a merely human love” or it is “an inscrutable love,” perhaps – and we realize that such sufferings are quite compatible with the truth of the assertion that “God loves us as a father (but, of course …).” We are reassured again. But then perhaps we ask: what is this assurance of God’s (appropriately qualified) love worth, what is this apparent guarantee really a guarantee against? Just what would have to happen not merely (morally and wrongly) to tempt but also (logically and rightly) to entitle us to say “God does not love us” or even “God does not exist”? I therefore put to the succeeding symposiasts the simple central questions, “What would have to occur or to have occurred to constitute for you a disproof of the love of, or of the existence of God?”Footnote 45

Flew’s analysis of religion is quite obviously meant to be critical, based on the assumption that falsification is a feature of all rational discourse: “Some theological utterances seem to, and are intended to, provide explanations or express assertions. Now an assertion, to be an assertion at all, must claim that things stand thus and thus; and not otherwise. Similarly an explanation, to be an explanation at all, must explain why this particular thing occurs; and not something else.Footnote 46

That is to say, Flew considers an explanation to be substantive only if it can be tested against another possible explanation; he endorses falsificationism (as Lakatos defines it) as a standard for all meaning. Popper himself rejected this point in the same breath that he declared falsifiability central to scientific discourse: recognizing a particular statement as unfalsifiable does not render it meaningless, only meaningful in a different way.Footnote 47 The original respondents to Flew’s essay, R. M. Hare and Basil Mitchell, demonstrate that human knowledge is comprised of an enormous variety of propositions, many of which, such as expressions regarding one’s personal security (or insecurity) or faith in others (or lack thereof), are fundamentally not subject to empirical falsification.Footnote 48

If we turn Flew’s critique of religion on its head, we may offer a positive description of religions as systems designed to organize propositions that are not subject to empirical assessment.Footnote 49 And while they are hardly the only social systems to serve this function, they are among the most extensive and elaborate systems to do so. As Peter Berger observes, religion is more all-encompassing than other social nomoi in that it is nearly always cosmic in scope.Footnote 50 This is a critical point: as we mentioned above, Saussure posits that one of the factors which contributes to the invariability of a language over time – what we have called its canonicity – is the number of signs necessary to constitute the system, yielding a highly complex internal structure based on patterns of difference between the signs. Thus compared to other similar discourses (e.g., pseudosciences, superstitions), we would expect religions to be the least inclined to discard their constituent signs, since religious signs are more tightly enmeshed in the web of relationships that is a semiotic system.

As we indicated above, this division between science and religion is primarily a modern phenomenon. As our ability to make precise and uniform empirical observations has dramatically improved over the past half millennium, the distinction between falsifiable and unfalsifiable propositions has become increasingly relevant in the organization of human knowledge, and falsifiable propositions have coalesced into a distinct discourse with a different dimension of meaning.Footnote 51 Yet even if this distinction is contingent to our own historical period, it is critical to understanding the way that science and religion function in relation to one another within our time. The resistance of religion to change has become an increasingly pressing issue in the last generation with the rise of religious fundamentalism on the international stage. The perspective we have proposed explains why religion’s resistance to change is inherent to religious discourse as it exists within modern society, rather than merely reflecting external (i.e., social, cultural, or political) forces.

VII Canonicity in Law and Morality

Following our proposal that legal systems be analyzed as irreducibly prescriptive, we may posit that the strong canonicity of law stems from the paramount importance of consistency in legal behavior and adjudication.Footnote 52 Consistency is an element of law that has not been well studied within philosophy of law, in part because it is so self-evident (what is a rule if not something applied consistently?), and in part because consistency in-and-of itself does not suffice to explain the many different purposes that law may have. Consistency may reflect the sociological need for order and stability, the religious value of consonance with the divine will or other cosmic forces, or moral concerns such as justice and fairness. But it is eminently clear that all systems that we refer to as “law” share this common feature.

One consequence of legal consistency is the predominance of rigid categories within legal discourse: cases deemed to be similar for purposes of legal behavior are grouped together and assigned the same legal value (permissible/prohibited, obligatory/optional, liable/exempt, guilty/innocent, etc.). However, consistency demands not only that the legal system assign the same value to all similar cases at any one point in time, but also that any given case be treated the same as comparable cases that have been decided previously. This diachronic axis of consistency is reflected in the fact that, once established, legal rules are considered to remain in effect until formally amended or overturned (by what H. L. A. Hart refers to as the rules of change),Footnote 53 and behavior not governed by legal rules tends to remain ungoverned until formally addressed. Moreover, the way that those legal rules are applied is also expected to remain consistent over time: any legal subject may expect the outcome of her case to be the same as that of equivalent cases in the past. For the same reason, when existing laws are deemed not be relevant to the case at hand, law accords significance to an appeal to tradition, convention or the status quo: it is meaningful – though not always decisive – to claim before a legal authority, “But it has always been done this way.” In sum, the paramount value of consistency determines that conservativism is hard-wired into the nature of legal discourse, rather than merely a product of the personal intransigence of judges or other legal authorities.

We may underscore this aspect of legal discourse by contrasting it with another irreducibly prescriptive discourse, namely morality. As with religion, the definition of morality is a highly contentious point, but the moral philosopher Bernard Gert convincingly posits that one of the key features of morality is that it is informal, meaning that it has “no authoritative judges and no decision procedure that provides unique answers to all moral questions.”Footnote 54 In effect, Gert is observing that consistency in decision-making is not a central feature of moral discourse, since there is no context in which the moral values assigned to different cases need to be resolved with one another. This reflects a fundamental truth of moral reasoning: the moral value of a particular action is measured primarily by its accordance with broad moral principles and with its consequences for the well-being of those creatures that it impacts, and to a far lesser degree by its consistency within a category of comparable cases.

Gert’s assertion has numerous implications for the nature of moral discourse. For our purposes, however, the key implication of Gert’s thesis is the way that moral discourse relates to its own past. The relative unimportance of consistency within moral discourse means that, in contrast with law, arguments from tradition, convention or the status quo do not carry significant moral weight; in assessing the moral value of an action that will unambiguously benefit or harm another, coherence with past moral judgments tends not to be influential. This facet of moral discourse makes it far more responsive to shifting societal perspectives: without the need to cohere with past decisions, moral decisions can reflect what society presently considers to be right or good.

Analysis of the way that moral systems evolve over time is largely absent from major works in moral philosophy. When this aspect of morality is analyzed, it is almost always in reference to the idea of moral progress, the notion that humanity is advancing toward a more enlightened moral state. This idea is expressed in the abstract in Hegelian and Marxist historiography, and has been advocated in more concrete terms by Peter Singer.Footnote 55 Yet the notion that moral discourse is evolving in a positive direction remains highly controversial, caught up in disagreements as to whether there has been an improvement in overall human well-being,Footnote 56 as well as ideological debates over whether modern Western society’s more permissive attitudes toward practices such as abortion or assisted suicide constitute progress or regress. The argument we are making avoids these controversies by putting forth a more modest position, namely that moral discourse is constitutionally prone to evolution, but not necessarily in a positive direction. Even if one rejects the idea of moral progress specifically, it is difficult to make sense of the ways that common morality has dramatically changed over the past several hundred years – on issues as varied as religious and racial diversity, slavery, gender roles, sexual mores (including but hardly limited to homosexuality), and treatment of animals – without acknowledging that the propensity for change is a central aspect of moral discourse.

More relevant to our thesis, this evidence for moral change underscores the relatively conservative nature of law. On nearly every one of these issues, changes to the law (e.g., banning slavery or discrimination) were most often responses to shifts in popular sentiment that had destroyed the moral consensus on that issue.Footnote 57 These are examples of law fulfilling one of its standard societal functions: to provide resolution on issues of irresolvable moral debate.Footnote 58 But the legal changes in these areas have frequently been preceded by the use of law to enforce the morally controversial status quo. To wit, not only does law change more slowly than morality, it also serves to constrain the implementation of moral change.

As with religious intransigence, legal intransigence is often attributed to extralegal – especially political – forces, but this assumes that law is merely a tool for the exercise of power, devoid of any independent dimension of meaning. Our analysis of legal discourse offers a more nuanced explanation. Law is a discourse distinct from politics that is intrinsically predisposed to favor existing patterns of behavior and the attitudes and beliefs that attend them. This aspect of law makes it a useful tool for political authorities to maintain the status quo, but more generally makes it unsuited to serve as a vehicle for initiating social or cultural change.

VIII Conclusion

The semiotic property of canonicity refers to the extent to which systems of meaning such as languages and higher-level discourses retain their constituent elements over time. We have proposed that the discourses of religion and law have intrinsically strong degrees of canonicity, such that they function as predominant forms of “institutional memory” within human culture. While our analysis has mainly been a general, programmatic discussion, we have alluded to some of its specific implications, both for the internal dynamics of law and religion themselves, and for the frequent points of tension between law and religion and other cultural discourses. Although a fuller analysis is beyond the scope of this chapter, a further set of implications is the way that legal and religious terminology and concepts are frequently used within other discourses to express rigidity or durability, such as the way that sociologists analyze secular political commitments in modern society as “civil religion.” It is our hope that the methodological framework that we have developed here will prove useful for discussing this phenomenon and conducting future research.


1 Umberto Eco, A Theory of Semiotics (Bloomington: Indiana University Press, 1976), 49 (emphasis in original).

2 Course in General Linguistics, eds. Charles Bally and Albert Sechehaye, trans. Roy Harris (Chicago: Open Court, 1983), 110–20.

4 Michel Foucault, The Archaeology of Knowledge, trans. A. M. Sheridan Smith (New York: Pantheon, 1972), 99.

5 “[T]he term discourse can be defined as the group of statements that belong to a single system of formation; thus I shall be able to speak of clinical discourse, economic discourse, the discourse of natural history, psychiatric discourse;” Footnote ibid., 107–08.

6 This is not to imply that language is necessarily a paradigmatic semiotic system, only that it serves as a useful reference point due to its ubiquity. The question of whether language constitutes the primary semiotic system is in fact a topic of debate among semioticians. For opposing viewpoints, see Umberto Eco, “Peirce’s Notion of Interpretant,” MLN 91 (1976): 1457–72; and Émile Benveniste, “The Semiology of Language,” in Semiotics: An Introductory Anthology, ed. Robert E. Innis (Bloomington: Indiana University Press, 1985), 226–46.

7 Julia Kristeva, “Word, Dialogue and Novel” in Desire in Language: A Semiotic Approach to Literature and Art, ed. Leon S. Roudiez, trans. Thomas Gora, Alice Jardine & Leon S. Roudiez (New York: Columbia University Press, 1980), 6491.

8 The question whether one can define law as independent from morality was famously the subject of a debate between Lon Fuller and H. L. A. Hart. Fuller (“Reason and Fiat in Case Law,” Harvard Law Review 59 [1946]: 376–95; Positivism and Fidelity to Law: A Reply to Professor Hart,” Harvard Law Review 71 [1958]: 630–72) argued that morality is necessary to provide content to the form of law, while Hart (“Positivism and the Separation of Law and Morals,” Harvard Law Review 71 [1958]: 593629; The Concept of Law [Oxford: Oxford University Press, 1961]) countered that law need not always correlate with morality. Ronald Dworkin followed Fuller’s position on the necessary connection between law and morality, and incorporated this into a comprehensive theory of legal adjudication; “ The Model of Rules,” University of Chicago Law Review 35 (1967): 1446; Law’s Empire (Cambridge, MA: Harvard University Press, 1986).

9 This is not to say that there is never uncertainty or debate about what the legal or moral value of an action is, rather that states of uncertainty or debate are recognized as failures to assign value to that action. Below we will propose that moral discourse leaves much more room for such indecisiveness than does legal discourse.

10 The same could be said of many other definitions of religion, including the influential definition offered by Clifford Geertz that we will return to below.

11 Peter Berger, The Sacred Canopy: Elements of a Sociological Theory of Religion (Garden City, NY: Doubleday, 1969), 177.

12 We should acknowledge that the very notion of defining religion as a system of meaning has been challenged, notably by Talal Asad, who insists that “religion as practice, language, and sensibility set in social relationships rather than as systems of meaning”: “Religious language – like all language – is interwoven with life itself. To define ‘religion’ is therefore in a sense to try and grasp an ungraspable totality. …[D]efinitions of religion are embedded in dialogs, activities, relationships, and institutions …” (Craig Martin, “Genealogies of Religion, Twenty Years On: An Interview with Talal Asad,” Bulletin for the Study of Religion 43 [2014], 1314).

See also Asad’s oft-cited essay, “ The Construction of Religion as an Anthropological Category,” in Genealogies of Religion: Discipline and Reasons of Power in Christianity and Islam (Baltimore: Johns Hopkins University Press, 1993), 2754.

I suggest that Asad’s critique reflects an excessively narrow association of religious “meaning” with “belief.” The model of semiotics we have proposed suggests a far broader conception of meaning, in which every form that facilitates cognition and communication may be analyzed as a sign. Hence all the various elements that Asad mentions – including those social practices and institutions necessary for the exercise of power (which Asad sees as integral to a proper understanding of religion) – derive their religious significance from their position within a complex network of signs. (Below we will propose that any semiotic network that we consider to be religious must also include certain propositions that we refer to as “beliefs.”) Asad’s description of religion as “an ungraspable totality” reflects the fact that religion – like all sign systems – is nothing more than the amalgam of many individual signs. Semiotics is precisely the study of how such amalgams function as coherent and discrete systems notwithstanding their fragmentary nature.

13 See Yuri Lotman’s discussion in Universe of the Mind: A Semiotic Theory of Culture, trans. Ann Shukman (London: I.B. Tauris, 2001), 138.

14 On this point, see Robert A. Yelle’s critique of earlier applications of semiotics to the study of religion; Semiotics of Religion: Signs of the Sacred in History (London: Continuum, 2012), 912, 2426.

15 Collected Papers of Charles Sanders Peirce, vol. 4, eds. C. Hartshorne and P. Weiss (Cambridge, MA: Harvard University Press, 1933), §127; quoted in Roman Jakobson, “A Few Remarks on Structuralism,” MLN 91 (1976): 1537.

16 Saussure, Course in General Linguistics, 80.

17 Footnote Ibid., 79–98.

18 See Jack Balkin and Sanford Levinson, eds., Legal Canons (New York: NYU Press, 2000).

19 Thomas Kuhn explains that this is one sense in which we speak of “scientific paradigms”: specific instances of particular scientific phenomena “which, employed as models or examples, can replace explicit rules as a basis for the solution of the remaining puzzles of normal science”; The Structure of Scientific Revolutions, 3rd ed. (Chicago: University of Chicago Press, 1996), 175.

20 Moshe Halbertal, People of the Book: Canon, Meaning, and Authority (Cambridge, MA: Harvard University Press, 1997), 3.

21 Kristeva, “Word,” supra Footnote note 7, 65.

22 Jonathan Z. Smith, “Sacred Persistence: Toward a Redescription of Canon,” in Imagining Religion: From Babylon to Jonestown (Chicago: University of Chicago Press, 1982), 4452.

23 Hart, The Concept of Law, supra Footnote note 8, 97–120.

24 Saussure, Course in General Linguistics, 71–74.

25 Asad, “Construction of Religion,” supra Footnote note 12.

26 Paul Feyerabend, Against Method, 3rd ed. (London: Verso, 1993).

27 For example, see Joseph Singer, “The Player and the Cards,” Yale Law Journal 94 (1984): 170.

28 For a summary of this scholarship, see Jack Balkin, “Critical Legal Theory Today,” in On Philosophy in American Law, ed. Francis J. Mootz III (Cambridge: Cambridge University Press, 2009), 6472.

29 R. M. Hare, The Language of Morals (Oxford: Oxford University Press, 1952).

31 Footnote Ibid., 163–65.

32 James Lenman articulates this point especially well: “The making of inverted commas moral judgements is clearly parasitical on the making of genuine moral judgements but not vice versa … That is why we cannot imagine a world in which the only kind of moral judgements ever made are inverted commas moral judgments” (emphasis in original); “ Moral Naturalism,” The Stanford Encyclopedia of Philosophy (Spring 2014 Edition), ed. Edward N. Zalta,

33 Hart, The Concept of Law, supra Footnote note 8, 77–88.

34 Lon L. Fuller, “Reason and Fiat in Case Law,” Harvard Law Review 59 (1946): 377.

35 See William Arnal and Russell T. McCutcheon, The Sacred Is the Profane: The Political Nature of Religion (Oxford: Oxford University Press, 2013), 2327.

36 Clifford Geertz, “Religion as a Cultural System,” in Anthropological Approaches to the Study of Religion, ed. Michael Banton (London: Tavistock, 2004), 4; emphasis added.

37 This is not to say that beliefs necessarily occupy a privileged place within the structure of religious discourse (e.g., that other manifestations of religion should be seen as deriving from them). Rather, beliefs are especially significant insofar as they distinguish religion from other cultural discourses.

38 Kuhn’s choice to focus specifically on instances of cataclysmic change – leading him to distinguish sharply between “normal” and “revolutionary” science – has been one of the most frequently challenged aspects of his work. Kuhn’s critics charge that he underestimates the extent to which significant changes occur within the context of “normal” science – that is to say, without the drastic event of a paradigm shift. For an early articulation of this critique, see Stephen Toulmin, “Does the Distinction between Normal and Revolutionary Science Hold Water?” in Criticism and the Growth of Knowledge, eds. Imre Lakatos and Alan Musgrave (Cambridge: Cambridge University Press, 1970), 3947. The point that we are making – that science is progressive or weakly canonical – does not take a stand on the debate between Kuhn and his critics as to whether scientific change is primarily incremental or abrupt.

39 Karl Popper, The Logic of Scientific Discovery (London: Routledge, 2002), 18.

41 Imre Lakatos, “Falsification and the Methodology of Scientific Research Programmes,” in Criticism and the Growth of Knowledge: Proceedings of the International Colloquium in the Philosophy of Science, eds. Imre Lakatos and Alan Musgrave (Cambridge: Cambridge University Press, 1970), 120.

42 Footnote Ibid., 116–22.

43 Footnote Ibid., 132–38.

44 For a balanced perspective on this point, see Martin Mahner, “Science and Pseudoscience: How to Demarcate After the (Alleged) Demise of the Demarcation Problem,” in Philosophy of Pseudoscience: Reconsidering the Demarcation Problem, eds. Massimo Pigliucci and Maarten Boudry (Chicago: University of Chicago Press, 2013), 2943.

45 Reprinted (with the original respondents and Flew’s response) as “ Theology and Falsification: A Symposium,” in The Philosophy of Religion, ed. Basil Mitchell (Oxford: Oxford University Press, 1971), 1322, at 15.

46 Footnote Ibid., 21 (emphasis original).

47 Popper, Scientific Discovery, supra Footnote note 39, 18n*3.

48 “Theology and Falsification,” supra Footnote note 45, 15–20.

49 This distinction between falsifiable and unfalsifiable propositions does not lead to any particular approach to how religious signs should be understood. For example, a modern approach to reports of miracles might be to interpret them as psychological rather than epistemic (i.e., concerning knowledge of mind-independent phenomena), or even to view them as metaphors. Yet even those who assert that miracles are manifest in physical reality ascribe them to causes other than falsifiable, scientifically accepted principles.

An important consequence of the non-falsifiable nature of miracles is that the miraculous nature of an event can be established only negatively, by process of elimination. A good example of this is the Catholic Church’s procedure of confirming miracles for sake of beatification, which is done by systematically ruling out scientific explanations of the event in question.

50 Berger, Sacred Canopy, supra Footnote note 11, 25–28. Hence insofar as religion is necessarily anthropocentric, it may be described as “the audacious attempt to conceive of the entire universe as being humanly significant” (28).

51 This description follows Berger’s definition of secularization as “the process by which sectors of society and culture are removed from the domination of religious institutions and symbols;” Footnote ibid., 107.

52 One of the few scholars to explicitly discuss the importance of consistency to legal discourse is the German sociologist Niklas Luhmann, who identifies consistency in decision-making as the meaning of justice within modern legal systems; Law as a Social System, eds. Fatima Kastner, Richard Nobles, David Schiff and Rosamund Ziegert, trans. Klaus A. Ziegert (Oxford: Oxford University Press, 2004), 219. We are going further than Luhmann in proposing that this consistency is an operative principle of law in general, but a full exploration of Luhmann’s distinction between premodern and modern law is beyond the scope of this chapter.

53 Hart (The Concept of Law, supra Footnote note 8, 60–64) considers the concept of legal continuity – that a law endures until such time as it is repealed – to be one of the key features that distinguish law from mere habit.

54 Bernard Gert, Morality: Its Nature and Justification, revised ed. (Oxford: Oxford University Press, 2005), 11.

55 Peter Singer, The Expanding Circle (Princeton: Princeton University Press, 1981).

56 For a critical view on this point, see Theodor Adorno, Minima Moralia: Reflections from Damaged Life, trans. E. F. N. Jephcott (London: Verso, 2005), 5356.

57 Religious diversity is a notable exception to this point, since laws establishing tolerance toward religious minorities in the modern period have often been motivated by political or economic considerations (as they have throughout history) rather than moral concerns, such that these legal changes sometimes precede corresponding shifts in moral sentiment.

58 As Gert explains, “[w]hen it is important that [moral] disagreements be settled, societies use political or legal systems to supplement morality. These systems do not provide a moral answer to the question; rather, the question, being regarded as unresolvable, is transferred to the political or legal system”; Morality, supra Footnote note 54, 11.

You have Access Open access

Save book to Kindle

To save this book to your Kindle, first ensure is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about saving to your Kindle.

Note you can select to save to either the or variations. ‘’ emails are free but can only be saved to your device when it is connected to wi-fi. ‘’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Available formats

Save book to Dropbox

To save content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about saving content to Dropbox.

Available formats

Save book to Google Drive

To save content items to your account, please confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account. Find out more about saving content to Google Drive.

Available formats