Today when we say “NT,” we mean a closed, 27-document collection that Christians view as authoritative. How did the Church come to accept only these documents from among other Christian writings of the first century?
Three main answers have been given, and to equip ourselves to understand them, it is necessary to begin with the definition of the word “canon” (Metzger, The Canon of the New Testament, 289–93). Canon is a standard or norm—something against which other things are measured, and as such it is used as a reference to both the OT and NT. When we hear the NT described as “canon,” it is an acknowledgment that this collection is limited and has authority for the Church.
Much of the study of the formation of the NT canon can be understood as an ever-narrowing definition of the term “canon” in reference to Christian writings. Confusion has resulted because not all canon historians understand the term in the same way. This can be seen in the three main answers to how the Church came to accept only these documents from among other first-century Christian writings.
The first answer was given by Theodore Zahn in the late 19th century when he argued that the NT arose as a spontaneous occurrence. Zahn believed that once a NT document was cited by a church father, a canon should be seen as present—citation proved canonicity. Thus, according to Zahn, there was already a NT in existence by the end of the first century that was not forced on the Church, but rather was a spontaneous creation that occurred in the life of the Church (Allert, A High View of Scripture?, 41–42).
Zahn’s position received important qualification in the early 20th century from Adolf von Harnack who developed the second answer. Harnack argued that citing a NT document as Scripture is very different from simply citing or alluding to NT documents. Harnack paid attention to the way a document was cited. Thus, if a citation was preceded by a formula referring to it as “Scripture,” this then becomes the test for canonicity because this placed the document at the same status as the OT . The effect of this qualification was to move the emergence of a NT from the first century into the mid- to late-second century because this is when citations of NT documents as Scripture emerged (Allert, A High View of Scripture?, 42–44).
The third answer to the question was offered by Albert C. Sundberg, Jr., in 1968 (Sundberg, “Towards a Revised History of the New Testament Canon,” 1216–24). Sundberg continued to narrow the definition of canon because of his reassessment of the OT canon in early Christianity. Sundberg observed that the church fathers cited more than just documents known to us as canonical as Scripture (Allert, A High View of Scripture?, 177–85). He concluded that the Church did not receive a closed OT canon, but rather, Scripture on the way to a canon. Because of this, Sundberg believed Harnack’s answer was difficult to sustain. The church fathers cited, as Scripture, documents not in our closed OT canon. Thus, one cannot claim, as Harnack did, that citation of a document as Scripture proves canonicity. If this were true, we should have a larger OT canon than we presently do. If the Church did not receive a closed OT from Judaism, but rather Scripture on the way to canonization, then the comparison of the citations of Christian literature with OT citations cannot establish canonicity for Christian writings.
Sundberg’s research has led some to agree that an essential distinction be made between the terms “Scripture” and “canon” (Gamble, The New Testament Canon; compare Seitz, The Goodly Fellowship). Sundberg thus argued that “Scripture” should be understood as writings that are held in some sense as authoritative for religion. “Canon,” on the other hand, should be understood as a defined collection that is to be held exclusively, with respect to all other documents, authoritative. The issue here is one of anachronism—we should not refer to a document as “canon” that should historically be referred to as “Scripture.” Thus, one cannot claim canonicity for a NT document that is cited with the same formula as an ot document unless one is prepared to say that the church fathers had a larger ot canon than we currently have. Based on these conclusions, Sundberg argues that a NT canon does not appear in Christianity until the latter half of the fourth century, when lists of canonical books begin to appear.
Note the increasing narrowness for the term “canon.” The way one defines canon will go a long way to determine which of the answers are accepted. Zahn and Harnack encourage us to understand “canonical” as a writing that functions authoritatively. If this is accepted, then a canon emerges quite early (later first to the end of the second century). Sundberg encourages us to view “canon” in a stricter sense—as a closed list of writings. If this is accepted, a canon emerges much later in the fourth century when these lists begin to appear.
The tendency is to view these three answers as mutually exclusive. But, as John Barton points out, when one looks at the actual argument of each position, there is much of substance in each (Barton, Holy Writings, 11–14): Zahn is correct that most NT documents did have authority in the late first and early second centuries; Harnack is correct that these books were discriminately added to in the second and third centuries; Sundberg is correct that it is only from the fourth century onward that authoritative rulings about the exact limits of the canon appear.
However, we could also say that each position is overstated: Zahn asserted that the NT books would one day form a canon, but it is an overstatement to claim that this was the intent of first-century Christians; Harnack does not give enough attention to the fact that in the second century there was still an openness to receive other books, that is, to add them to the “canon”; Sundberg states that the latter part of the fourth century is decisive because this is when strict canonical lists begin to appear, but it is probable that these lists were codifying what were already accepted lists.
It appears that how or whether a particular book was cited is not the best question to ask. Rather, perhaps a better question would be how often they were cited. As Barton explains, “The picture that emerges is surprisingly clear. From the Apostolic Fathers onwards, the Synoptic Gospels (especially Matthew), the Fourth Gospel, and the major Pauline letters are cited much more often than one would predict, if one supposed that the whole of the NT we now have was equally ‘canonical’ or important. Correspondingly, the rest of the NT (including Acts) is manifestly less important. The third category, books scarcely cited at all, contains most of those which later decisions and decrees affirm to be noncanonical; even in the earliest period none of them is cited even so often as the books of the second class” (Barton, Holy Writings, 17).
With attention to how often books are cited, we can understand the three answers given not as mutually exclusive positions, but as three phases of NT canonicity that correspond to the chronological positions given by Zahn, Harnack, and Sundberg (Barton, Holy Writings, 18–24; see also Sundberg, “Towards a Revised History,” 459–61).
In the first phase, the central core of the present NT is already beginning to be treated as the main source for Christians. The identification of this central core was completed before the end of the first century. Technically speaking, however, it was not a canon. At this stage, it is inappropriate to say that the NT canon is fixed, but it would be equally inappropriate to say there was no core collection of writings.
In the second phase, during the second and third centuries, certain other documents begin to be cited more often, which indicates addition to the core collection, albeit not necessarily explicitly acknowledged. In this phase, there is still no clear distinction between documents in the NT and documents outside it. Instead, the distinction is between documents cited often, documents cited little, and books whose use is discouraged. While the core has ceased to grow, the thought of forming a fixed collection had still not appeared. Thus, to speak of a canonical/noncanonical distinction is misguided.
In the third phase, fourth century indications about the canon become firm. This is the century in which lists of canonical documents proliferate, giving very strong indication that the Church was explicitly thinking about a closed canon. But we must realize that even here to speak of a closed canon is difficult because these lists differ in some details. That is, a few documents that appear on some of the lists are not in the present canon. Further, a few documents that are in our present canon are absent from some lists (Metzger, The Canon of the New Testament, 305–15).
Framing the history of the NT canon in these phases is properly concerned with the function of Scripture in the early church. Rather than conceiving of a closed NT in the second century to which the Church appealed for its sole source of teaching, this three-class paradigm forces one to consider how the Church judged and appropriated the very writings it included in its canon. Thus, we can conceive of an authoritative body of Christian Scripture in the first century, but we cannot claim that body of literature to have been closed even into the fifth century. This is why it is so important to distinguish between Scripture and canon.
Consistent with this focus on the function of Scripture, NT canon historians have employed a rubric called the criteria of canonicity—the criteria that documents needed to fulfill for inclusion in the canon. This is not to speak of an explicit list of criteria to which the early Church referred and through which each and every document was sifted and subsequently placed in the canon as a result of satisfying each criterion. The criteria are a retrospective scheme by which we attempt to understand why certain documents came to be valued above other documents. It is devised through an examination of the writings of the fathers and their use of these documents. We must avoid the temptation to view these criteria as hard and fast rules. Thus, it is difficult to rank them in importance because they were not invoked with great consistency or rigor. Rather, they operated interdependently or concurrently—not independently or sequentially. Further, some churches and leaders gave different weight to certain criteria, which explains why some documents took longer to gain universal acceptance in the Church.
The first criterion is apostolicity. While this could mean that a document was written by an apostle, it was not necessarily essential. Some of our NT documents were received as written by an apostle, like Paul’s letters. But other documents gained wide acceptance because of a direct link to the apostles. Some documents experienced difficulty when it came to widespread acceptance by the Church. The best-known case of this is Hebrews. Origen of Alexandria’s (circa 184–circa 253) comments are instructive: “If I gave my opinion, I should say that the thoughts are those of the apostle, but the style and composition belong to someone who remembered the apostle’s teachings and wrote down at his leisure what had been said by his teacher. Therefore, if any church holds that this Epistle is by Paul, let it be commended for this also. For it is not without reason that the men of old time have handed it down as Paul’s. But who wrote the Epistle in truth, God knows” (Origen, Homily 6.25.11–14). Even though Origen believes those who accept Hebrews as written by Paul to be mistaken, he does not dismiss it on those grounds, but rather accepts the document because it is apostolic in teaching.
The second criterion is orthodoxy. Orthodoxy means the congruity of a document with the apostolic faith. To see how this criterion functioned in the life of the Church, it is helpful to see how Serapion of Antioch dealt with a document in the early third century.
Eusebius tells us that Serapion wrote a refutation of the Gospel of Peter, the content of which had led a parish in his jurisdiction astray (Hist. eccl. 6.12.1–6). The church at Rhossus was using the Gospel of Peter in their teaching and worship. Initially, this did not trouble Serapion because he believed that they held “the true faith” and could discern this gospel’s doctrine. But he soon came to learn that this discernment did not occur after he secured a copy of the document and found that it taught Docetism, thus denying “the true faith.”
This shows the criterion of orthodoxy in action. Serapion did not appeal to a NT to see if the Gospel of Peter was included therein. Rather, his appeal was to “the true faith” to discern its teaching. Serapion’s issue was not that the Church was using a document outside of a canon, but that they did not discern the heterodoxy of the document. This was how the criterion of orthodoxy functioned. The rule of faith was used in the early Church as this standard of orthodoxy—something against which teaching and documents were measured. R.P.C Hanson calls it a “graph of the interpretation of the Bible by the Church of the second and third centuries” (Hanson, Tradition in the Early Church, 127). Even the Christian writings that were eventually included in the NT canon were subjected to this rule of faith (Allert, A High View of Scripture?, 54–56; 78–84; 121–26).
The third criterion is catholicity and traditional use. This criterion is best illustrated in Augustine (354–430): “Now, in regard to the canonical Scriptures, he must follow the judgment of the greater number of catholic churches; and among these, of course, a high place must be given to such as have been thought worthy to be the seat of an apostle and to receive epistles. Accordingly, among the canonical Scriptures he will judge according to the following standard: to prefer those that are received by all the catholic churches to those which some do not receive. Among those, again, which are not received by all, he will prefer such as have the sanction of the greater number and those of greater authority, to such as are held by the smaller number and those of less authority. If, however, he shall find that some books are held by the greater number of churches, and others by the churches of greater authority (though this is not a very likely thing to happen), I think that in such a case the authority on the two sides is to be looked upon as equal” (Augustine, Doctr. chr. 2.8.12).
We see here the dynamics at work between apostolicity, orthodoxy, catholicity, and use. For Augustine, widespread use of a document carried considerable weight for its acceptance. Some churches—like Rome, Antioch, Alexandria, and Constantinople—were given more preference concerning the documents they used than other churches. In the case of a writing that was accepted by all compared to a writing that some did not accept, preference was given to the writing accepted by all. When dealing with those documents that were not accepted by all, one was to accept those with the greater representation among the churches, with greater weight given to the more important churches. Augustine also makes reference to the improbable possibility where there may be the majority of the churches using one document while the most important churches employ another document. If this occurred, Augustine’s counsel is to accept both.
Augustine reveals an important reality in the life of the early church. For a document to be used in a church, it had to be accepted and valued as Scripture by a local church. Through gradual and more widespread recognition, that same document gained an even higher stature in the church catholic. This criterion capitalized on practices of the churches. But although the passage displays an explicit consideration of canon issues, the variety of canonical lists in the fourth and fifth centuries shows that the issue was not settled for all even then.
Some have offered another criterion of canonicity—inspiration. This is sometimes claimed as the predominant criterion. Thus, R. Laird Harris states, “… the test of canonicity is inspiration. The Early Church put into its canon, and we receive, those books which were regarded as inspired, and no others” (Harris, Inspiration and Canonicity of the Bible, 200). Harris’ argument is that the Christian documents that came to form the NT were the only documents that the early church viewed as inspired.
This argument was called into question by Everett Kalin, who asked whether the early church viewed only the documents that went into the NT canon as inspired, and those alone (Kalin, “Argument from Inspiration,” 541–49). The short answer is no. The church considered not only other documents as inspired, but also many aspects of the Church’s life, including bishops, monks, interpreters of Scripture, martyrs, councils, and a wide array of prophetic gifts.
From one perspective, it is accurate to say that inspiration was a criterion for canonicity. That is because all documents considered orthodox by the early church were, by implication, believed to be inspired. But from another perspective, it is inaccurate to say that inspiration functioned as a criterion of canonicity if we mean that inspiration was believed to have been the possession of only the documents that later became part of the nt.
Kalin concluded that the fathers did consider the books that later came to be included in the canon as inspired. But it was not inspiration that distinguished these documents from all others. This does not mean that the fathers did not regard these NT documents as unique—their elevation to canonical status clearly indicates otherwise. It does mean, however, that it was not inspiration that contributed to their uniqueness above all other Scriptures. There were other more practical reasons for the Church’s acceptance of certain documents and exclusion of others.
If the argument of writers like Harris is correct and everything in the canon was considered inspired and everything outside of canon was considered noninspired, one would expect this to be indicated somewhere by the leaders of the early church. But this is precisely what cannot be found. Kalin showed that while the fathers do apply the concept of inspiration to documents that came to be included in the NT canon, they almost never applied the concept “noninspired” to distinguish those canonical books from Scripture. In other words, canonical/noncanonical was not synonymous with inspired/noninspired. On the rare occasion when a father did declare a writing not to be inspired, he was not saying that it was not a canonical document. He was saying, rather, that the document was heretical, that it lay outside the community of faith where the Spirit was at work.
The assertion that the early church included in the NT canon only those books that were seen as inspired has one further problem—references in the fathers to noncanonical books as inspired (Allert, A High View, 60–65; 185–88). For example, Gregory of Nyssa (circa 330–circa 395) makes reference to his brother Basil’s commentary on the first six days of creation as “an inspired [theopneustos] exposition … [admired] no less than the words composed by Moses himself” (Gregory of Nyssa, Apologia in Hexaemeron). In the second century, Abercius Marcellus of Hierapolis composed an inscription that was set up over his future tomb. The Life of Abercius, which was written about this bishop in the fourth century, contains a text of this inscription and describes it as an “inspired inscription” (Life of Abercius 76). The Council of Ephesus (431) issued a synodical letter that describes the council’s condemnation of Nestorius as “their inspired decision” (Schwartz, Acta Conciliorum Oecumenicorum 1.1.2: 70).
These three examples should give one pause when claiming that the early church reserved the term “inspired” for only the canonical documents—each example describes a noncanonical document as “inspired.” This would be very high praise for these documents if inspiration was a designation for only canonical documents. Thus, inspiration did not guarantee inclusion because inspiration was not viewed as the unique possession of only the canonical documents.
It is important not to force a 21st-century perspective back onto the sources of the ancient Church. Christianity had a fairly fluid body of literature that the Church used as authoritative. Among this breadth, certain documents rose to preeminence in the life of the Church. But that rise, in some cases, was not immediate. This is not meant to deny the providence of God in the process but, rather, to say that there were very practical reasons why certain documents came to be valued (and eventually canonized) by the Church, and it is on this very practical road to canonization that God providentially lead His people by His Spirit.
Fairly early in the Church’s existence did it possess all the documents that went, eventually, to make up what we call the NT. But the early church did not necessarily and consciously collect and separate a select few (27) documents for the direct purpose of including them, and only them, in the NT. This conscious collecting did occur, but it started in the fourth century.
 Barry, J. D., Heiser, M. S., Custis, M., Mangum, D., & Whitehead, M. M. (2012). Faithlife Study Bible. Bellingham, WA: Logos Bible Software.