Monday, 23 January 2017

Authorship and Date

As I've worked through the dates of the New Testament texts, the one thing I've wanted to avoid is making this a study of authorship. The more I work on the material though, the more I realize that the relationship between authorship and date is unavoidable. Take the example of the undisputed Pauline letters. I would argue that the ultimate reason that we suppose that these cluster to the 50s of the first century is because we attribute them to Paul. Attributing them to Paul, and knowing that he likely died no later than 68 (the end of Nero's reign), we reasonably suppose that any letter that he wrote must predate 68. We have other data, primarily given by Acts, that leads us towards the 50s as the primary time frame for these writings. In the most recent major study of Pauline chronology, Douglas Campbell recognizes the close relationship between authorship and date in the study of the Pauline epistles, making the question of authenticity the first issue that he discusses in each case; although there is much in Campbell's study to which I object, I am increasingly seeing the necessity in chronological studies of first querying the traditional authorship, not only of the Pauline texts but in fact of any text with a traditional attribution.

So, this has me thinking: how does one go about deciding authorship? When one looks at the scholarship, it's a morass. Paul is just assumed to have written the undisputed Pauline epistles, without question. If Acts' presentation of Paul are thought to differ from how Paul is presented in these epistles, Acts is judged without question to be in error. If the presentation of Paul in the Pastoral Pauline epistles is thought to differ from that in Acts, then the Pastoral Epistles are judged to be in error without question. If the presentation of James or Peter in Acts or the undisputed Pauline epistles differs from that found in the letters of James or 1 or 2 Peter, James or 1 or 2 Peter are judged to be in error without question. There is an unspoken hierarchy: data from undisputed Pauline epistles are supposed without question; data from Acts is accepted where it does not conflict with data from undisputed epistles; data from other epistles, Pauline or Catholic, are accepted where it does not conflict with data from undisputed Pauline epistles or Acts. No attempt to justify this hierarchy of data is advanced; it is just tacitly assumed.

So, then, the question of how to establish authorship really becomes: how do we know that Paul wrote the undisputed epistles? Who can answer that? Very few, I suspect. But how we answer that, it must become the paradigm for answering the question of authorship of any other text. I would suggest that the following procedure is at work. It's very simple. We see that it is attributed to Paul, and find no compelling reason to think that to be in error. This judgment is not offered without data: the crucial datum in each case is the attribution. In most cases this will be a necessary condition, but due to the phenomenon of pseudepigraphy in no case will it probably be sufficient. Thus to this positive test we must add the negative test of compelling reason to judge against the attribution.

The question thus becomes: what reason do we have to say that Paul did not write this? James? Peter? Mark? Matthew? Etc. In each case, the nature of the case will be different, and must be argued from the particularity of the data. Of course, in many cases the particularity of the data will be similar, and some data even will recur from letter to letter (such as with the details of Paul's life, which will recur throughout deliberations on the thirteen canonical Pauline epistles as well as extra-canonical texts attributed to Paul). There will be cases in which the answer is very obvious, in either direction; there will be cases in which that is less the case. Chronology will be related to this in a complex fashion. Sometimes it might exclude the traditional authorship. If a letter clearly references an event that occurred after the author's known time of death, then the traditional authorship must be in error. Sometimes the question of authorship will aid in answering chronological questions, by establishing that the text was likely written during the traditional author's lifetime.  The putative authorship establishes an initial time frame in which to look. If we judge that James, brother of Jesus, wrote the Epistle of James, then we can hardly date the letter later than 62, the year in which (on the basis of Josephus) we know that he died; if Peter wrote 1 Peter, the letter cannot date likely date any later than 68; etc. In most cases, the lower end of possibilities is circumscribed by references to Jesus' life, ministry, or death, making a date earlier than c. 30 impossible.

Authorship and date turn out to be indissolubly linked.

Monday, 16 January 2017

James and Programmatic Skepticism

In his 2013 International Critical Commentary on the canonical Epistle of James, Dale Allison notes rightly that there were numerous pseudepigraphical works written in James’s name. Yet, he also notes rightly, no one argues that these were in fact written by James, brother of Jesus, whereas there are arguments made to support the notion that James wrote the canonical epistle. Thus he asks rhetorically, “Might there not be a canonical or theological bias at work here?” (p. 13). Of course, the answer to the question as asked needs to be “Yes.” There might well be a canonical or theological bias at work. Indeed, it seems hard to deny that such is not at work among those who hold explicitly to any variant of the doctrine of inerrancy. Yes, does it follow that anyone who argues for the authenticity of the Epistle of James must be guided by such a bias?

An immediate observation: the hypothesis is self-condemning. This is obvious if extended beyond the Epistle of James. No critical scholar of whom I am aware argues or indeed has argued that Paul wrote 3 Corinthians, although it is attributed to him. Yet these same scholars have no problem affirming that Paul wrote at least seven of the letters attributed to him and which are found in the New Testament canon. Is there a canonical or theological bias at work? By Allison’s implicit reasoning, there should be, as the situation is precisely parallel: the same scholars who do not seriously entertain the possibility of extant Pauline literature outside the NT canon seriously consider the possibility of extant Pauline literature inside the NT canon. The only difference is one of degree, not kind: with only one rather than thirteen texts attributed to James in the NT, one either affirms 0% or 100% of those attributed to him therein; the possibility of affirming some but not all texts attributed to James in the NT is non-existent. If those who affirm the traditional authorship of our one canonical Jacobean text but not of any non-canonical Jacobean texts are ipso facto guilty of canonical bias, then so too must be those who affirm the traditional authorship of some of our canonical Pauline texts but not of any non-canonical Pauline texts. If we follow this line of reasoning, then a canonical bias is operative in the judgment that Paul wrote at least these seven texts; yet, we continue to affirm that judgment as true; if the affirmation reflects reality, then the bias cannot be thought to constitute a barrier to genuine insight, and why then should we suppose that it would in the case of the canonical James? The question of canonical bias thus seems a non-starter (and this doesn't even mention that if used to vet possible hypotheses, the possibility of bias is really nothing but an instance of the genetic fallacy).

More interesting to me though is the empirical significance of the canon. The reality is that the canon is a historical datum. It exists, in time and space. The observation that the Epistle of James is the only text attributed to James that made it into the canon cannot be programmatically assumed to be a datum of irrelevance for establishing authorship. We know that the early Christians were concerned to admit into the canon only those texts that they had good reason to suppose were written by members of the first Christian generation. It is not a given that this was simply rhetoric or myth-making meant to justify other, truer, but often unspoken interests. That might be the case, or it might not; it needs to demonstrated, not supposed on a programmatic basis. Too often however, it is precisely on a programmatic basis that it is supposed; it is become disciplinary common sense in many circles, but common sense has a tendency to dissolve under closer examination. We cannot dismiss programmatically the hypothesis that the Epistle of James made it into the canon for precisely the reason that our ancient informants claim: that there was good reason to think that it was written by James.

What we’re getting at here is really what it means to be a critical scholar. There is an unfortunate tendency in many circles to suppose that critical scholarship consists of pronouncing negative judgments on early Christians’ own self-understanding of their origins. I would suggest that this is a misunderstanding of what it means to be a critical historian. The critical historian is one who formulates a question, attends to the data relevant to answering that question, weighs possible answers, and then affirms that answer which handles the relevant data best. Sometimes that will much resemble early Christians’ self-understanding of their own origins; sometimes it will be remarkably at variance therewith. The skeptic supposes programmatically that the best answer will be at variance with traditional narratives. That is bias, the bias known as skepticism, which takes as its sinister twin the bias known as credulity: the programmatic supposition that the best answer will be fully congruent with traditional narratives. Both arbitrarily close off possible answers before the investigation even begins. As such, the spirit of critical thought is programmatically opposed to both.

Wednesday, 28 December 2016

The Ascension of Isaiah and the Gospel of Thomas

John A.T. Robinson's Redating the New Testament is one of those books whose footnotes are often mini-essays in their own right. Over the holidays I've had the chance to pore over his notes, and in so doing been well-rewarded. I want to consider two notes: one in which he addresses the Ascension of Isaiah, the other in which he addressed the Gospel of Thomas. Perhaps not surprisingly, given his general tendency, Robinson argues that the Ascension of Isaiah (which is frequently dated to the 80s through 110s) might well date to the late 60s. His reasoning is expressed briefly, and he does not reach as firm a conclusion as he does with the texts that he considers in the main body of the monograph, but his argumentation is quintessentially "Robinsonian." Most notably, he argues that the descriptions of persecution within the book fit well the details that we know about the Neronian persecution. Certainly, the description of Beliar-in-the-flesh in Ascension 4 sounds a lot like contemporary descriptions of Nero, especially the emphasis upon matricide in 4.2; and it is stated in 4.3 that this incarnation of evil would kill one of the Twelve, a quite plausible reference to Peter's death under Nero. This, and a few other details in 4, incline me to think that the identification of Nero as the king in this chapter is probably strong. Robinson suggests also that 4.13 refers to the flight of the Jerusalem church into Pella as a contemporary event, something that I'll grant as possible but a bit on the speculative side. All considered, Robinson makes a convincing argument that a late-60s date for the Ascension of Isaiah is at the very least plausible. I would be inclined to say the following: if one dates Revelation to the 60s, then one will have a hard time arguing that Ascension must date much later.

A second footnote of interest is his off-hand statement that he is open to considering the possibility that the Gospel of Thomas is earlier than often supposed (late-first through mid-second century). He gives no explanation as to why he is thus open: are their particular details in the text that lead him to suspect an earlier date, or he is just in principle open to such possibilities? This has led to me rereading the Gospel of Thomas, with an eye to the sort of thing that might have caught Robinson's intention and made him suspect an earlier date. In so doing, I am struck by logion 12, wherein Jesus states that after he dies, James the Just will take his place in leading his disciples. As it stands it does seem to be concerned with post-dominical succession; of course, one could come up with all sorts of other rhetorical purposes for this logion, but one suspects that they might be closer to eisegesis than exegesis. Now, it is highly unlikely that Jesus uttered anything like these words, as James appears not to have been one of his followers during his lifetime. Indeed, it is not clear that James took on a leadership role before the early 40s. Yet, if the author of Thomas, or his sources, felt sufficient freedom to place upon Jesus's mouth instructions regarding the succession of leadership after his passing, then one is struck by the absence of comparable instructions regarding James's death. I find it difficult to imagine that the logion, in its present form, could have been written after 62.

Now, of course, it has been argued that Thomas is the product of a long period of development, extending certainly into the second century; indeed, this notion of development is usually used to justify a first-century date, by arguing that there was effectively a proto-Thomas in the first century that developed into the Thomas that we now have. Thus, it could be argued, logion 12 could represent an early stage in the development of Thomas, while much of the balance of the text dates much later. I am uncertain if this resolves the chronological problem posed by logion 12 however, for two reasons. One, although such hypothetical developmental theories are often carefully argued, they tend towards the speculative, especially given the relative dearth of textual material that we have for Thomas. We do not even have a complete copy in the original language, which leaves me with less-than-complete confidence in our ability to clearly define the text's literary development. Two, and I think more crucial, such a developmental theory would in fact serve to make more acute the question of why logion 12 either has not dropped out or has not been elaborated to address post-Jacobean succession. The more that one emphasizes Thomasine fluidity, and the more one emphasizes that the text was a perpetual work-in-progress that was responsive to external conditions potentially well into the second century, the more conspicuous the fact that it gives instructions that could be followed no later than 62. And further indications of a date for Thomas in, say, the 50s might be adduced. For instance, the text is concerned with the necessity of circumcision, and our extant evidence would seem to situate intra-Christian debates over the necessity of circumcision primarily in the the late-40s through 50s. That's where Acts would seem to place such debates, and the texts otherwise most directly concerned with the matter and readily dated,, i.e. the core Pauline epistles, all emerge from this decade or so. If Robinson had been so inclined, I think that he could probably have made a solid argument for a Gospel of Thomas in the 50s.

A possible objection: does not Thomas include doctrinal material that could not reasonably preexist the second century? I'm not altogether convinced that this is the case, and I doubt that Robinson would have been also. He tended to be very rigourous when it came to such arguments from development, seeing them as generally circular in nature: development theories are used to date the texts, and then the sequence of texts is used to confirm the development theories. And in any case, there seems to be a growing consensus that Thomas' gnostic character has been much exaggerated. Certainly, it does not contain much in the way of hints of the more elaborated systems of gnostic speculation that emerge in the second century. And already in the Corinthian literature Paul is addressing what has often been described as a sort of "proto-gnosticism." I don't see anything in Thomas's theology that requires second century date, or even one later than c. 60.

That having been said, let it be emphasized, I am not arguing that we should date Thomas prior to 62. Rather, I am considering the sort of argument that Robinson might have advanced in favour of an earlier dating for the text, given his general approach to such questions and had he been so inclined to present such an argument.

Thursday, 22 December 2016

Why Robinson Still Matters

Robinson's Redating the New Testament is of crucial significance because, to the best of my knowledge, it was the 20th-century's only monograph-length critical study devoted wholly to the dates of the New Testament texts. Indeed, Robinson, writing in 1976, had to look back to Harnack's Chronologie, published in 1896-7, to find something comparable to his own work. It was controversial because it argued for what I describe heuristically as a "lower" chronology for the New Testament texts. In my own work on the dates of the New Testament texts, I distinguish between lower, middle, and higher chronologies. While all three chronologies agree that the undisputed Pauline epistles cluster in the 50s, lower chronologies argue that the balance of the NT texts for the most part date between c. 40 and c. 70, middle chronologies that the balance dates largely between c. 70 and c. 100, and later chronologies push one or more larger works of the NT (these days, usually Luke and Acts) into the second century. Obviously this distinction is fuzzy, as are most heuristic schemes, but it's a good starting point.

Middle chronologies can also be described as the "consensus position," as by far the majority of NT scholars would subscribe to such a view. But the middle chronology has never been given a full, synthetic, defense, comparable to what Robinson did with the lower chronology, and nor has the higher chronology. The closest for the middle chronology was Harnack's aforementioned 1896-7 work, but--not surprisingly--what Harnack wrote 120 years ago does not quite correspond with the general contours of the modern middle chronology. Moreover, turning to Harnack for a synthetic defense of the middle chronology is to open a door to the lower chronology, as Harnack subsequently revised certain key dates downwards (notably those of the Gospels of Mark and Luke, and also Acts), such by the end of his career his chronological scheme straddled the lower and middle chronologies. The closest to a synthetic defense of the higher chronology is the work of F.C. Baur in the mid-19th century, but as the middle chronology is largely a response to that, and the early to the middle, an updated synthetic defense would be desirable. Indeed, Baur's chronology is hardly viable today, as even the most die-hard advocates of a higher chronology would tend to recognize. For synthetic treatments of the matter of the dates of the New Testament text, we are left with Robinson. He is virtually the only player on the field. Yet he only represents one of the three major options that are out there. This leaves us in a position where we tend to evince greater confidence in the dates of the New Testament than the current state of the research permits.

The way to begin correcting this situation is for three monographs to be produced: one giving an up-to-date defense of a lower chronology, one giving a defense of a middle chronology, and one giving a defense of a higher. I myself am presently working to produce a defense of the lower. I have no intention of writing defenses of the middle or the higher chronologies: the ideal situation is for each to be written by an enthusiastic advocate of the respective chronology, and I find that I can only be such with regard to the lower chronology. I would, quite simply, not be the right person to write the other two, necessary monographs. But I do believe they are necessary, and hope that they might emerge.

Wednesday, 21 December 2016

1 Clement--"A Post-Apostolic Postscript," Pt. 4

With this blog post, I discuss Robinson's treatment of 1 Clement, the final text that he considers in Redating the New Testament. I will post in the next few days on other matters that come out of my reading of Robinson, especially on how it relates to my own current research programme, but this completes my "review" of Robinson's chronology proper.

1 Clement is generally dated to the mid-90s, and this because it is supposed (largely following Lightfoot) that the text became associated with Clement of Rome (who is never actually mentioned in the body of the text) because it was sent from Rome to Corinth at a time when he was the bishop of the former church. Robinson, not surprisingly opts for a date a couple decades earlier, in early 70. Some of his argumentation here is related to that which he advanced for the Shepherd of Hermas: that Clement was likely associated with this text not because he was bishop of Rome when it was written, but rather because he was the Roman church's "foreign secretary." In terms of generating an exact date, Robinson argues that the "sudden and repeated misfortunes and reverses that have happened to us" of 1.1 (following Holmes's translation) refer, not as Lightfoot and others have argued, to the events of Domitian's reign, but rather to the Neronian persecution and the chaos of the Year of the Four Emperors. He argues also that the discussion in chapter 41 of the Jerusalem temple supposes that it is yet standing. Thus, he argues, the text must have been written in the latter part of 69 or the earlier part of 70.

I find these arguments to be of varying weight. Robinson is wholly correct to note that the association with Clement needn't indicate that the letter was written when he was bishop, and thus that a date as early as the 70s or even into the late-60s is theoretically possible. He is of course correct in judging that it cannot date much earlier than that, as it talks about the deaths of Peter and Paul as past events (cf. chapter 5), and they appear both to have died in the mid- to late- 60s. I'm less convinced that 1.1 refers to the Neronian persecution. Indeed, it's not even clear to me that 1.1 references persecution at all; "misfortunes" and "reversals" could entail a great number of things. On the upper end of things, I am unconvinced that chapter 41 necessarily supposes that the temple yet stands; unlike Hebrews 10:2, the argument in this passage seems quite conceivable after the destruction of the temple. I think frankly that Robinson puts too much weight on this datum. 100 however seems like a reasonable terminus ante quem, as Clement seems to have died around that time. Thus, a range from c. 70 up to c. 100, around which time Clement seems to have died, seems by far the most viable range.

A potential objection to the earlier end of this range is that 47.6 refers to the Corinthian church as "ancient" (ἀρχαῖος). Robinson addresses this quite well, and is surely correct in arguing that this hardly makes the 90s more likely than c. 70. All Clement seems to be saying here is that the church has been around for some time. C. 70 the church in Corinth would be about twenty years old, and by c. 90 it would be about there something about forty that would make it ancient when twenty is not? When we remember moreover that ἀρχαῖος can simply be translated as "old," it doesn't seem unreasonable to think that in a movement barely forty years old (as of 70) a church that is twenty years old might well have been one of the older communities then in existence. Another objection is the fact that Clement probably cites Hebrews, which is part of why 1 Clement is often dated to the 90s: if Hebrews is dated to the 80s or early 90s, then 1 Clement cannot date any earlier. But Robinson has already given very good reason to date Hebrews before 70, thus anticipating and obviating this objection.

That all being said, at this point, I am unconvinced that we can be as precise in that range as either Lightfoot or Robinson desired. The reality is that we don't know that much about the church in these latter decades of the first century, at least not relative to our detailed knowledge of the 30s through 60s. In particular, we lack precisely that form of data most needful for chronology, namely a solid narrative account. We have that for the earlier decades in Acts, but that breaks off c. 62. The one argument that might incline me towards an earlier date in this range is Robinson's observation that 1 Clement does not seem to suppose a sole bishop in the church at Rome, whereas by the time Ignatius is writing during Trajan's reign he can suppose precisely that. But even there, Ignatius could be writing as late as c. 117, thus allowing for the better part of twenty years to intervene between the latest possible date for 1 Clement. As such, I am generally inclined to think that, as with the Shepherd of Hermas, we have to conclude that it likely dates to the last three decades of the first century, but that on the basis of the extant data greater precision is excluded.

A final note, not raised by Robinson: such a conclusion would allow for the possibility that the Clement of Phil. 4:3 is the Clement of 1 Clement. If Clement was one of the most prominent--perhaps even the most prominent--leaders in the Roman church of the 90s, then it's not at all improbable that he was active in Christian ministry in the late-50s. That said, as I concluded in a comparable discussion with regard to the Shepherd of Hermas, while not impossible or even improbable, I do not think that the data permits us to make such an identification with confidence. We cannot even be certain from Philippians that Clement is associated with the Roman church at all (although if Philippians is written from Rome, that would be quite probable), and moreover "Clement" appears to have been a very common name in Rome. This seems to be a matter upon which we might need to remain agnostic.

Monday, 19 December 2016

The Didache--"A Post-Apostolic Postscript," Pt. 3

We turn now to the third text that Robinson treats in Chapter X of Redating the New Testament, namely the Didache. Generally speaking, the Didache is one of the more difficult early Christian texts to work with, as our only Greek copy wasn't found until 1873, dates from the 11th century, and appears to be incomplete (incidentally, this copy is part of the Codex Hierosolymitanus, which also contains our only complete Greek copy of 1 Clement). Still, there is a general consensus that the Didache dates to the first century or so after Jesus's life. There is probably a general tendency to place it around 100, give or take a decade or two in either direction, and not surprisingly Robinson argues that this is on the high side. He instead argues that the Didache, more or less as we find it in the Codex Hierosolymitanus, was completed by c. 60.

The word "completed" is crucial here, as virtually all scholars agree that the text was composed over time, and might have originally constituted a multitude of disparate texts. Robinson agrees with this assessment, but notes correctly that "over time" is a relative term. He follows J.P. Audet, who argues in his 1958 work La Didachè: Instructions des apôtres (a classic work whose influence has been vitiated by the lack of an English translation) that the Didache was produced over a twenty-year period, spanning from 50 to 70. Robinson's only modification to Audet is to propose that the span was closer to 40 to 60. Why does Robinson place the Didache in this period? Numerous reasons. He suggests for instance that the ecclesiology of the Didache is most similar to that found in the authentic Pauline letters and paralleled in Acts: itinerant ministers, with a movement towards establishing permanent bodies of elders in the various communities. He suggests that the eschatology of the final chapter, 16 (which seems to breaks off, thus suggesting incompletion) finds its readiest parallels in 1 and 2 Thessalonians. In short, it seems to fit better with what we know about Christianity c. 50, give or take a decade, than the time around c. 100. He further suggests that the use of Synoptic materials looks it comes more from the period that the gospels were being formed than from the use of fully-formed gospels. Admittedly, these are all among the weaker arguments for establishing a text's date, but given the state of the data it might all that we can do. They do seem stronger than those used to defend the almost-entirely speculative hypothesis advanced in the late-19th century and still held by many today, namely that they represent an otherwise-unknown form of early-second-century Syrian (or, less commonly, Egyptian) Christianity.

One thing that I find interesting about this narrative is how well it coheres with another datum, namely the longer title of the Didache, which reads in English "The Teaching of the Lord to the Gentiles by the Twelve Apostles." It has long struck me as a remarkable coincidence that we have the Didache, the Epistle of James, the Gospel of Matthew, and the Gospel of Mark all being attributed to figures associated with the first-generation Jerusalem church and all evincing a close literary relationship. If there is no connection between this body of densely-related literature and Jerusalem than we have virtually to posit among the ancients an intentional effort to deceive on this point. It’s far from clear to me that such a conspiracy is better defensible on the data than the hypothesis that a religious movement that began in Judea happened to produce a considerable among of its earliest texts in Judea. And it also happens that, as best we can tell, during the 40s and the 50s the Jerusalem church expended considerable energy on thinking about the place of the Gentiles in the nascent movement. Given the state of the text-critical evidence we probably should not dismiss out of hand the possibility that the longer title provides data for considering the initial conditions under which the Didache was produced, and we should seriously entertain any hypothesis that can make good sense of the text overall as well as the title(s).

Sunday, 18 December 2016

The Shepherd of Hermas--"A Post-Apostolic Postscript," Pt. 2

While I find Robinson on shaky ground in his dating of the Epistle of Barnabas, I find him on pretty solid ground with the dating of the Shepherd of Hermas, at c. 85 (compared to the mid-2nd century of broader consensus). Before we discuss his arguments however, we need to consider a question that he does not raise, namely the unity of the Shepherd. The text is remarkably complex, consisting of three separate parts (often cited separately, a practice I will follow here as it was followed by Robinson): the Visions, the Mandates, and the Similitudes, abbreviated typically and respectively as Vis., Mand., and Sim. Indeed, the text is sufficiently complex that the Society of Biblical Literature Style Guide, 2nd ed., dedicates an entire appendix to the discussion of how best to cite the text. No other text is signaled out for such special treatment. Historically, this complexity was seen by critical scholars as evidence of a composite work, written by multiple hands over an extended period of time. In contrast, Robinson simply assumes the unity of the text, without arguing for it. This is something of a weakness in his treatment, although coming as the discussion does in a "postscript" it perhaps can be forgiven: this is more an appendix than it is the main body of the work. It should also be noted that more recent scholarship has moved somewhat in the direction of seeing the text as a unity. In her 1999 Hermeneia commentary on the text, for instance, Carolyn Osiek argues that it was written by a single author, over an extended period of time. My own inclination is to think Robinson and Osiek more likely to be on the right track than not. In general, I think that the tendency to disintegrate texts that come to us as an unity into hypothetical composite parts is a hold-over of certain habits of speculation introduced into the discipline in the 19th-century: habits that tend to suffer from a dearth of evidence and a fuzziness of concept that vitiate their arguments.

Robinson's date for the text turns on on two matters: the irrelevance of the Muratorian Canon for the discussion, and the crucial relevance of Vis. 2.4.3. As Robinson notes, the argument for a second-century rests almost entirely upon the Muratorian Canon, which states that the text was written "recently," by a man named "Pastor," the brother of Pius, bishop of Rome from 140-154. Now, no question, the Shepherd is situated in Rome: that's clear from the text itself. But, as Robinson notes, the Canon is not generally considered to be a reliable source, and more recent scholarship seems inclined to push it from a late-second century date into the third and even fourth centuries. As Robinson notes, for no other text would chronological judgments rest almost entirely upon the canon's evidence, and thus it seems questionable to do so with the Shepherd. Moreover, Robinson is able to produce a compelling explanation for the text's likely error: the Latin title of the Shepherd is Liber Pastoris (Book of the Shepherd), and suggests that the author of the canon has simply made the mistake of assuming that this was the name of the author, i.e. Pastor.

More compelling is the data given by Vis. 2.4.3. Here, Hermas is having a vision of an elderly lady, who in Holmes' translation says to him
Therefore you will write two little books, and you will send one to Clement and one to Grapte. Then Clement will send it to the cities abroad, because that is his job. But Grapte will instruct the widows and the elders. But you yourself will read it to this city [i.e. Rome], along with the elders who preside over the church.
There is a longstanding and likely correct supposition that the Clement referenced here (even if fictionally) is likely Clement of Rome, the putative author also of 1 Clement, and according to tradition the third bishop of Rome after Peter. It is sometimes supposed that this verse indicates (even if fictionally) that Clement was bishop at the time that Hermas wrote these words. Clement's episcopacy is typically dated to the 90s, although arguments have been made that would date its advent to the mid-80s. Robinson argues that this description of Clement's role suggests that Clement was more of a "foreign secretary" (his term) than a bishop proper, and thus that the text should predate his episcopacy. Given that the exact role of the "bishop" of Rome in these early decades remains unclear, I find myself generally unable to adjudicate between these two positions with confidence, and would be more inclined to say more generally that the Visions date to the last two or so decades of the first century. Other evidence adduced by Robinson points also in this direction. For instance, the reference to the elders who preside over the city (in Vis. 2.4.3, cited above) does seem to suggest an ecclesiology more like that which we know from the first rather than the second century.

We are left with the problem, identified by Osiek, that the text overall, while likely written by a single author, might well have been written over an extended period of time. I am perfectly happy to grant such a scenario, not least of all because of the sheer size of the work. But would this require us to move the completion of the text significantly into the second century, if at all? I'm not convinced, and thus am inclined to situate the Shepherd overall in the late first-century. Note that this would permit one to identify Hermas the author with the Hermas mentioned by Paul in Romans 16:14. Robinson however does not make this identification, and for good reason. Both are Roman Christians, in the first century (if the above hypothesis is deemed correct), named Hermas, but I'm not sure that suffices for an identification. This differs from the identification of the Clement of Vis. 2 with Clement of Rome, wherein the description of his role in the Shepherd matches closely the description of his role that we find in relation to 1 Clement. Absent such data, I am reluctant to think the Hermas who wrote the Shepherd is likely the Hermas mentioned by Paul. Certainly, the recurrent efforts to use Rom. 16:14 as a datum for dating the Shepherd need to be greeted with reserve.