Conclusion
1
The Definition of Canon
Must We Make a Sharp Distinction Between the Definitions of Canon and Scripture?
Once a distinction is made between scripture and canon, the idea of a New Testament canon does not appear applicable until the fourth century.
GEOFFREY M. HAHNEMAN
The Muratorian Fragment and the Development of the Canon
Brevard Childs once declared, “Much of the present confusion over the problem of canon turns on the failure to reach an agreement regarding the terminology.” Although Childs made this statement in 1979, it could just as easily been written in our current day. As scholars continue to probe into the origins and development of the biblical canon, debates and disagreements about canonical semantics have not abated. What exactly do we mean by the term canon? Does it refer to books that were widely used by early Christians? Does it refer to books that function as Scripture? Or does it refer only to books that are included in a final, closed list? While these discussions over the definition of canon will certainly continue, and no universal agreement appears to be forthcoming, something does seem to have changed since Childs’s original observation. The definition of canon as a final, closed list of books has begun to emerge as the more dominant one—at least in some circles. In particular, advocates of an “extrinsic” model of canon are typically committed to this particular definition and insistent that all scholars must adopt it, lest the entire field become plagued by confusion and anachronism.
Such claims are difficult to resist—after all, no one wants to plunge canonical studies into disarray. Moreover, there is certainly something attractive about having a single, unified definition of canon on which we can all agree (and build upon). Nevertheless, we must ask whether this “consensus” position, and the attitude with which it is held, is justified. Does this single definition adequately capture the complexities and nuances of the concept of canon? And are we required to adopt only this definition to the exclusion of all others?
The Exclusive Definition of Canon
The definition of canon as a fixed, final and closed list of books—what might be called the exclusive definition—was put forth originally by A. C. Sundberg in 1968. Sundberg drew a sharp distinction between the terms Scripture and canon and, on this basis, argued that we cannot speak of the idea of canon until at least the fourth century or later. Although Scripture would have existed prior to this time period, Sundberg argues that we must reserve the term canon until the end of the entire process. It would be anachronistic to use the term canon to speak of any second- or third-century historical realities. Thus, simply marshaling evidence of a book’s scriptural status in the early church—as is so often done in canonical studies—is not enough to consider it canonical. The book must be part of a list from which nothing can be added or taken away.
Sundberg’s exclusive definition of canon was initially supported by a number of key scholars such as D. H. Kelsey, James Barr and Harry Gamble, and, in more recent years, has continued to gather adherents. John Barton, while rightly recognizing that multiple definitions of canon have some validity, still seems to prefer the exclusive definition: “Much clarity could be gained if we agreed to distinguish sharply between these two concepts [of Scripture and canon].” Geoffrey Hahneman has been a vigorous advocate of the exclusive definition, declaring, “Once a distinction is made between scripture and canon, the idea of a New Testament canon does not appear applicable until the fourth century.” Lee McDonald has consistently promoted Sundberg’s definition in his many writings over the last twenty years and is no doubt one of the reasons for its recent popularity. Eugene Ulrich is quite forceful in his approach, arguing that unless scholars accept the exclusive definition, discussions will be “confusing and counterproductive.” Likewise, the recent work of Craig Allert insists on the “necessity of proper distinction between the terms ‘Scripture’ and ‘canon.’” Even this brief survey of scholars (and more could be added) suggests that David Nienhuis was correct when he observed that “Sundberg’s position has enjoyed widespread acceptance.”
But is the widespread acceptance of this position justified? We begin our analysis by noting that there are many positives to this position that ought to be acknowledged. For one, the exclusive definition of canon rightly captures the reality of the canon’s “fluid” edges prior to the fourth century. It took some time for the boundaries of the canon to solidify, and the exclusive definition accommodates this historical fact by using different terms for different stages. Moreover, this definition helps remind us of the important role played by the church in the recognition and reception of the canon. By restricting the term canon to only the final stage when the church has decisively responded, the exclusive definition keeps church and canon from being unduly divorced from one another—the two concepts go hand in hand. However, there are a number of concerns about this definition that need to be explored.
First, it is difficult to believe that the sharp Scripture-canon distinction drawn by modern advocates of the exclusive definition would have been so readily shared by their historical counterparts in the second century. Would early Christians have regarded “Scripture” as fluid and open-ended and only “canon” as limited and restricted? If they were able to say that certain books in their library were Scripture, then that implies they would have been able to say that other books in their library were not Scripture. But, if they are able to say which books are (and are not) Scripture, then how is that materially different than saying which books are in (or not in) a canon? Thus, it seems some degree of limitation and exclusion is already implied in the term Scripture. As Iain Provan observes, “The question I am asking is whether the idea of scripture does not itself imply the idea of limitation, of canon, even if it is not yet conceived that the limits have been reached. I believe that it does so imply.” If so, then the necessity of a strict demarcation between Scripture and canon largely disappears.
Second, while the exclusive definition insists the term canon cannot be used until the New Testament collection has been officially “closed,” significant ambiguity remains on what, exactly, constitutes this closing. If it is absolute uniformity of practice, across all of Christendom, then, on those terms, there was still not a canon even in the fourth century. Indeed, on those terms we still do not have a canon even today. If the closing of the canon refers to a formal, official act of the early church, then we are hard pressed to find such an act before the Council of Trent in the sixteenth century. The fact of the matter is that when we look into the history of the canon we realize that there was never a time when the boundaries of the New Testament were closed in the way the exclusive definition would require. Stephen Chapman comments on this problem: “Rather than being a minor problem, this inconsistency casts significant doubt upon the appropriateness of the entire approach. Why should scholars adopt as the correct usage of the term ‘canon’ a meaning that does not correspond fully to any historical reality?” Ironically, then, the exclusive definition is as guilty of anachronism as any of the views that it critiques.
This leads us to the third, and arguably the most foundational, problem for this definition. Inherent to the exclusive definition is an insistence that the fourth century represents such a profoundly different stage in the development of the New Testament that it warrants a decisive change in terminology. Indeed, Dungan refers to the stage of Scripture and the stage of canon as “very different.” But was the canon so very different in the fourth century? While a broader degree of consensus was no doubt achieved by this point, the core books of the New Testament—the four Gospels and the majority of Paul’s epistles—had already been recognized and received for centuries. Whatever supposedly happened in the fourth century neither altered the status of these books nor increased their authority. It is precisely at this point that the limitations of the exclusive definition become clear. The abrupt change in terminology gives the impression that these books bore some lesser status prior to this point; it communicates that Christians only had Scripture and not a canon. Or, as one scholar put it, prior to the fourth century Christians only had a “boundless, living mass of heterogenous” texts. At best this is obscurant, and at worst misleading. Moreover, it feeds the notion that the canon was somehow the result of “a great and meritorious act of the church.” And this is why this definition is a core tenet of the extrinsic model—it implies there was no (and could be no) canon until the church officially acted. Stephen Dempster highlights this problem: “Reserving the terminology ‘canon’ for only the final collection of books obscures the continuity that exists at earlier times. To accept such a limiting definition might suggest that the canon did not have a history, only to be created ex nihilo, the result of a [church] council.”
An example of this third issue can be seen clearly in the recent work of Craig Allert. The stated goal of his volume is to “emphasize the centrality of the church in the formation of the New Testament.” It is no surprise, then, that he is such a strong advocate of Sundberg’s definition of canon because, as he acknowledges, “Sundberg’s work has had the effect of pushing the decisive period, that of formal canonization, into the fourth and fifth centuries.” Such a late date for canon allows Allert to raise the profile of the church—it was there from the beginning, whereas the canon only arrives late on the scene. He declares, “The Bible was not always ‘there’ in early Christianity. Yet the church still continued to function in its absence.” While Allert is right to remind us of the important role of the church, this whole approach to the development of the canon raises some concerns. If the core books of the New Testament were functioning as authoritative Scripture by the middle of the second century, then is it really helpful to claim that early Christians did not have a “Bible”? This sort of language seems to bring more confusion than clarity. Although it may prevent one kind of misperception (that the canon was neat and tidy in the second century), it ends up promulgating what is arguably a bigger one (that early Christians had little interest in a New Testament until the fourth century).
With these concerns on the table (and more could be added), one might get the impression that this critique has been offered to challenge the overall legitimacy of the exclusive definition. However, that is not the intent here. If the above concerns are addressed, then the exclusive definition still has an important role to play. After all, the exclusive definition is correct that the boundaries of the canon were not solidified until the fourth century—and, in this sense, we did not have a “canon” until that time. The exclusive definition just needs to acknowledge that this is a general consensus and not an official act of “closing” with airtight boundaries that somehow increased the authority of these books. Thus, the main point of this critique is not to do away with the exclusive definition entirely but to challenge those advocates of the exclusive view who claim that it is the only legitimate perspective on canon. Given the limitations and weaknesses of the exclusive definition we have observed, we should be hesitant to think it completely exhausts the meaning of the term. If we are to fully appreciate the depth and complexity of canon, we must also let other definitions have a voice.
The Functional Definition of Canon
Although the exclusive definition of canon may be the dominant one at the current time (or at least the one that has enjoyed increasing popularity), it is not the only option on the table. Childs has played a central role in promoting an alternative definition, arguing, in contrast to Sundberg, that the term canon need not be restricted to a final, closed list but can “encompass the entire process by which the formation of the church’s sacred writings took place.” If a collection of books functions as a religious norm, regardless of whether that collection is open or closed, then Childs is comfortable using the term canon. Or, put differently, the term canon can be employed as soon as a book is regarded as “Scripture” by early Christian communities. Thus, Childs argues against any rigid separation between Scripture and canon, saying that they are “very closely related, indeed often identical.” For our purposes here, we shall refer to this definition as the functional definition of canon.
Of course, the pedigree of this functional definition goes back further than Childs. Barton points out how Harnack’s entire reconstruction of the origins of the New Testament canon is predicated upon this very definition. Harnack argued that a book could be considered canonical when it was expressly regarded as “Scripture”—which usually required the use of formulaic markers such as γραφή or γέγραπται. On this definition, the origins of the New Testament canon would be dated to the middle of the second century, dramatically earlier than the fourth/fifth century date advocated by Sundberg. Barton also distinguishes Harnack’s approach from that of Zahn, who was willing to regard a book as canonical apart from formulaic markers, as long as it enjoyed some degree of widespread use by early Christians (allowing for an even earlier date for canon). However, the distinction between Zahn and Harnack should not be overplayed. Their disagreement hinged upon the way to determine the church’s view of a book (formulaic markers versus widespread use), not whether the church’s view of a book was the key factor in deciding canonical status. The definitions of Harnack and Zahn are in agreement on the critical point: canon is determined by the function of a book in the church and not whether it was regarded as part of a final, closed list. In this sense, Harnack and Zahn really hold to the same general approach to canon.
The functional definition has also found support from a number of modern scholars, particularly those who have an association with Childs and/or the “canonical criticism” camp. James Sanders recognizes that the functional perspective on canon is valid because it is the predecessor of the exclusive perspective: “Canon as function antedates canon as shape.” G. T. Sheppard provides a helpful distinction between canon as a “rule, standard, ideal, norm” and canon as “fixation, standardization, enumeration.” He designates the former “Canon 1” and the latter “Canon 2,” recognizing the legitimacy of both. Chapman has persistently critiqued the exclusive definition while suggesting that the functional definition should have its place at the table, as have Provan, Meade, Riekert and Dempster.
The functional definition has many positives and provides a welcome balance to the exclusive definition. For one, it accurately captures the historical reality that early Christians did possess an authoritative corpus of books long before the fourth century, even if the edges were not entirely solidified. Thus, it does not run the risk of unduly diminishing the perceived authority of these books in pre-fourth-century context. In this sense, the functional definition would fit well with an intrinsic model of canon. In addition, this definition seems less prone to artificially inflate the role of official church declarations about the canon—as if those declarations somehow “created” or “established” the authority of these books.
That said, however, the functional definition still has its weaknesses. Two of these can be noted here. First, McDonald has pointed out that the functional definition struggles to account for books that were regarded as Scripture by some early Christian communities but never made it into the final, closed canon: for example, the Shepherd of Hermas, Apocalypse of Peter, and so on. What shall we call these books? McDonald argues that the functional definition leads to confusion because it is forced to call these books “canonical.” How can a book be canonical and then cease to be canonical? If we would only use the exclusive definition, he argues, then such confusion could be avoided. McDonald is correct to point out this issue, and it should be acknowledged that there is some imprecision in the functional definition here. However, it is not clear that the issue is as serious as McDonald suggests, nor that it mandates the sole use of the exclusive definition. For one, there does not appear to be anything particularly problematic or confusing about saying that some early Christian communities had different functional canons. There was widespread agreement about the core canonical books, but some disagreement over the peripheral books was inevitable. Some books were “canonical” in the eyes of certain communities, even though they would never become part of the church’s permanent collection. The functional definition appropriately captures this reality. The exclusive definition claims to avoid the problem of imprecision because it waits until the fourth century when the canonical boundaries were finally fixed. But, as noted above, the boundaries of the canon were not absolute even in the fourth or fifth centuries. Indeed, disagreements continue to the modern day. Thus, we are reminded that all definitions, including the exclusive definition, suffer from a level of imprecision—that is unavoidable whenever a definition seeks to capture an evolving historical situation (such as the development of the canon).
The second weakness of the functional definition is more significant and is also one that is shared by the exclusive definition. Both of these definitions fail to adequately address the ontology of canon. That is, these definitions do not incorporate what canon is in and of itself, apart from what it does in the church (functional) or how it is delineated by the church (exclusive). If we only have the functional and exclusive definitions, then we can only conclude that this thing we call canon cannot exist prior to it being used as Scripture or prior to the church reaching a final consensus. The church must act for there to be a canon. In this regard, the functional and exclusive definitions seem to confuse (or at least are prone to confuse) the church’s reception of the canon with that which makes a book canon. A book can become canonical, but on its own it is nothing. Of course, for some modern scholars, this would not be viewed as a problem. Viewing the canon as a purely community-dependent entity is central to the extrinsic model we are discussing in this volume—it is what John Webster calls the “naturalization” of canon. If the canon is nothing in and of itself, then it must be the result of contingent (and to some extent, arbitrary) human processes. Harnack is a prime example of this naturalization as he attributes the existence of the canon to the church’s “creative act” in response to Marcion. Others have argued that the canon is merely a sociocultural concept that reflects the relationship between a religious society and its texts. Still others have suggested canon is just a social phenomenon that arises when a community desires to express its identity. As Kelsey notes, canon is the church’s “self-description.” And always popular is the idea that “canon” is just a political construct, an ideological instrument, created to wield power and control.
The problem with these community-dependent views is that they do not represent the historical Christian position on the canon. Although it is out of vogue in some critical circles today, Christians have traditionally believed that the canon is a collection of books that are given by God to his corporate church. And if the canonical books are what they are by virtue of the divine purpose for which they were given, and not by virtue of their use or acceptance by the community of faith, then, in principle, they can exist as such apart from that community. After all, aren’t God’s books still God’s books—and therefore still authoritative—prior to anyone using them or recognizing them? Surely, the existence of canon and the recognition of canon are two distinguishable phenomena. Why, then, should the term canon be restricted to only the latter and not the former? Thus, our definition of canon cannot be limited to only the functional or exclusive definitions because neither of them account for this phenomenon; neither allow for the ontology of canon to play a role. Now, this doesn’t mean that all those who use the functional or exclusive definitions do not have an ontology of canon. It simply means that these definitions themselves do not allow for an ontology of canon. Unless this limitation is addressed, such definitions, whether intended or not, inevitably encourage the “naturalization” of canon.
The Ontological Definition of Canon
In order to accommodate the historical Christian approach to the canon, we need a definition that moves beyond the functional and exclusive ones. So, we shall call this the ontological definition. The ontological definition focuses on what the canon is in and of itself, namely the authoritative books that God gave his corporate church. One might say that this definition looks at canon from a divine perspective, rather than from only an ecclesiological perspective. Books do not become canonical—they are canonical because they are the books God has given as a permanent guide for his church. Thus, from this perspective, it is the existence of the canonical books that is determinative, not their function or reception. On this definition, there would be a canon even in the first century, as soon as the New Testament books were written. Of course, such a definition is inevitably retrospective in nature. The Gospel of John would have been “canon” ten minutes after it was written, but the early church would not yet have known it. It was only at a later point, when the corporate church had finally recognized which books belonged in the canon, that it could then look back and realize that there was a “canon” even in the first century. But, there is nothing illegitimate about affirming this reality. If canonicity is not merely something that happens to a book, then we can affirm a book is canonical when that book is produced. B. B. Warfield employs the ontological definition when he says, “The Canon of the New Testament was completed when the last authoritative book was given to any church by the apostles, and that was when John wrote the apocalypse, about A.D. 98.”
No doubt, those committed to a rigid historical-critical approach to the study of the canon will balk at the ontological definition as inappropriately theological. One cannot use a definition for canon that involves any theological considerations, we might be told. But why are we obligated to study the canon on purely historical-critical terms? Why should we be obligated to use the term canon in a way that prohibits the very approach to the canon that Christians have held for two millennia? Indeed, one might argue that, in this sense, the historical-critical approach is offering its own theological perspective—just in the opposite direction. More and more, scholars have recognized that theological and historical concerns are not easily separated, nor should they be. Iain Provan makes the point that, “All the great giants of biblical study in the last 200 years have worked within certain dogmatic and philosophical positions.” Francis Watson has pressed the case that, “Theological concerns should have an acknowledged place within the field of biblical scholarship.” This is especially true in the field of canonical studies. Floyd Filson has made the simple, but often overlooked, observation that, “The canon is a theological issue.” Kevin Vanhoozer concurs, “History alone cannot answer the question of what the canon finally is; theology alone can do that.”
Although the ontological definition brings a healthy balance to our definition of canon, we are not arguing here that it should be the only definition of canon. On the contrary, the ontological definition is being offered to complement (or round out) the functional and exclusive definitions. All three of these definitions make important contributions to our understanding of canon and therefore all three should be used in an integrative and multidimensional manner. The exclusive definition rightly reminds us that the canon did not fall in place overnight; it took several centuries for the edges of the canon to solidify. The functional definition reminds us that prior to the determination of the final shape of the canon there was a core collection of books that functioned with supreme authority in early Christian communities. And the ontological definition reminds us that books do not just become authoritative because of the actions of the church—they bear authority by virtue of what they are, books given by God. When all three perspectives on the canon are considered in tandem, a more balanced and more complete vision of the canon is realized. Thus, we should not be forced to choose between them.
In addition, this multidimensional approach to the definition of canon provides much-needed clarification to the ongoing debate over the “date” of canon. As Barton and others have already noted, the date assigned to the canon is, to some extent, correlative to the definition of canon one brings to the table. On the exclusive definition, we do not have a canon until about the fourth century. On the functional definition, it seems that we have a canon at least by the middle of the second century. On the ontological definition, a New Testament book would be canonical as soon as it was written—giving a first-century date for the canon (depending on when one dates specific books). When these three definitions are viewed together they nicely capture the entire flow of canonical history: (1) the canonical books are written with divine authority; →(2) the books are recognized and used as Scripture by early Christians; →(3) the church reaches a consensus around these books. The fact that these three definitions are linked together in such a natural chronological order reminds us that the story of the canon is indeed a process; and therefore it should not be artificially restricted to one moment in time. Put differently, the story of the canon is organic. It’s like a tree at different stages of its life: the young seedling just inches high, the adolescent sapling, and the full-grown adult. Even though there are changes, at each stage we can still use the same terminology, namely a “tree.” Perhaps, then, we need to rethink the whole concept of the canon’s “date.” Instead of discussing the date of canon, we might consider discussing the stage of canon. This latter term brings out the multidimensional nature of canon, whereas the former implies that canon is, and only can be, one point in time.
Once these three definitions are allowed to interface with one another, it also becomes evident that they, in some sense, imply one another. If a canonical book is a book given by God to his church (ontological definition) then we might naturally expect his church to recognize it as such and use it as an authoritative norm (functional definition). And if a canonical book is a book used as an authoritative norm (functional definition), we might naturally expect that the church would eventually reach a consensus on the boundaries around such books (exclusive definition). And if the church has reached a consensus on the boundaries around certain books (exclusive definition), then it is reasonable to think these are the books that have already been used as an authoritative norm (functional definition), and also that they are the books that God intended his church to have (ontological definition). The manner in which these definitions reinforce one another suggests that they are not contradictory as so many suppose, but instead are to be seen as complementary.
It is also worth noting that these three definitions of canon fit quite well with the established categories of modern speech-act philosophy. Speaking (which would also include divine speaking) can take three different forms: (1) locution (making coherent and meaningful sounds or, in the case of writing, letters), (2) illocution (what the words are actually doing; for example, promising, warning, commanding, declaring and so on), and (3) perlocution (the effect of these words on the listener; for example, encouraging, challenging, persuading and so on). Since any speaking act can include some or all of these attributes, it would be out of place to suggest that only one of them is the proper definition for what we call speaking. These three types of speech-acts generally correspond to the three definitions of canon outlined above. The ontological definition of canon refers to the actual production of these books and thus refers to a locutionary act. The functional definition refers to what the canonical books actually do in the life of the church and thus refers to an illocutionary act. And the exclusive definition refers to the reception and impact of these books on the church and thus refers to a perlocutionary act. Again, a multidimensional approach to the definition of canon brings out these nuances in greater richness and depth.
The manner in which speech-act philosophy uses three complementary definitions for the term speaking can provide some practical insight into how the same can be done with the term canon. Speech-act philosophy sees no need to choose just one of these definitions to the exclusion of all others, nor should we do so in regard to canon. Of course, authors may employ one particular definition of canon at any given time, but this need not be viewed as problematic. The particular definition employed may be determined simply by what an author desires to emphasize. If an author wants to emphasize the ecclesiastical dimension of canon, then the exclusive definition may be most appropriate. If an author wants to emphasize the authoritative role played by canonical books, then the functional definition is best. And if an author desires to view canon from the perspective of its divine origins, then the ontological definition is most suitable. But, even when just one of the definitions is employed, the other two definitions can still be viewed as legitimate and complementary (just as in speech-act philosophy). Moreover, it should be acknowledged that it is not always necessary for an author to choose which definition he or she is using (nor feel the need to explain to the reader which definition is being used). Sometimes the term canon, like the term speech, is used in such a general manner that all three definitions could be in view. In the end, this term can be employed with a substantial amount of flexibility, and this flexibility is a reminder of the depth and richness of this thing we call canon.
Conclusion
Brevard Childs was correct that much of the confusion over the history of the canon has to do with differences in terminology. However, that problem is not solved, as the extrinsic model suggests, by imposing a single definition of canon on modern scholars. On the contrary, insisting that only a single definition rightly captures the depth and breadth of canon may end up bringing more distortion than clarification. While the exclusive definition correctly reminds us that a general consensus on the boundaries of the canon was not achieved until the fourth century, it can give the misleading impression that there was little agreement over the core books prior to this time period. While the functional definition correctly reminds us that New Testament books served as an authoritative norm at quite an early time, it still does not address what these books are in and of themselves. While the ontological definition brings the necessary balance to both of these approaches—offering a reminder that these books do not become canonical simply by the actions of the church—it too cannot stand alone. To have only the ontological definition would lead us to wrongly conclude that these books were basically lowered from heaven as a completed canon with no development or history in the real world. Ironically, then, perhaps the debate over canon is best addressed not by choosing one definition, but by allowing for the legitimacy of multiple definitions that interface with one another. If canon is a multidimensional phenomenon, then perhaps it is best defined in a multidimensional fashion.