Memory as Method: Some Observations on Two Recent Accounts

In: Journal for the Study of the Historical Jesus
Author: Andrew Gregory1
View More View Less
Full Access

This review essay considers the use of social memory theory in two monographs on the gospels, and the extent to which that theory aids their arguments and conclusions. In the case of Jesus’ Literacy: Scribal Culture and the Teacher from Galilee, by Chris Keith, I argue that the author uses social memory theory to provide a helpful account of what historians do, but that his conclusion could stand without explicit appeal to his theoretical understanding. In the case of Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition, by Alan Kirk, I argue that his use of social memory theory, alongside his account of individual neurobiological memory and cognitive processes, is a vital part of the argument that he presents.


This review essay considers the use of social memory theory in two monographs on the gospels, and the extent to which that theory aids their arguments and conclusions. In the case of Jesus’ Literacy: Scribal Culture and the Teacher from Galilee, by Chris Keith, I argue that the author uses social memory theory to provide a helpful account of what historians do, but that his conclusion could stand without explicit appeal to his theoretical understanding. In the case of Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition, by Alan Kirk, I argue that his use of social memory theory, alongside his account of individual neurobiological memory and cognitive processes, is a vital part of the argument that he presents.

1 Introduction and Summary

This review essay considers the use of social memory theory in two monographs on the gospels, and the extent to which that theory aids their arguments and conclusions.1 With respect to Chris Keith, whose monograph is a study of the historical Jesus,2 this means that I focus on his discussion of how tradition, the product of social memory, shapes the understanding of the authors who receive it, and the texts in which they inscribe it. I note how persuasive is the way in which he uses social memory to describe how traditions originate and evolve, but I question to what extent his argument and conclusion depends upon the theoretical framework that he employs. With respect to Alan Kirk, and his work on the synoptic problem,3 I note the importance both of social memory theory, and of an understanding of individual neurobiological memory or cognitive processes for the argument that he makes, and how each contributes to his explanation of how Matthew used two written sources, Mark and Q. According to Kirk, social memory shaped both those two written texts, as also it shaped their authors. In a similar way, it shaped Matthew the scribe who through neurobiological memory was able to internalise these two written texts, these artifacts of social memory, and use them to produce a new written text, itself the product of social memory. In Kirk’s case I find these arguments to be essential to his thesis.

2 Chris Keith, Jesus’ Literacy

So let me turn to the work of Chris Keith, and the wider questions that he askes about the nature of the historian’s task.4 For, as Chris observes in his monograph on Jesus’s literacy, ‘arguments about the Gospels and historicity are actually about the nature and development of the Jesus tradition, and thus its possible connections to the actual past.’5

As Keith has observed elsewhere, tradition is ‘a cultural artifact that inherently and necessarily negotiates the past and present simultaneously’.6 Thus tradition, as a product of cultural memory, both shapes the person who receives it, and is in turn shaped by that person as he or she passes it on to someone else. In other words, to put my own gloss on it, one of the things that social memory theory might be said to do, in comparison with form criticism, is to put more emphasis on the tradent rather than on the tradition. Certainly it recognises that traditions shape those who transmit them. But rather than put tradition before the tradent, it also emphasizes that any enduring identity-forming tradition, i.e. any social or cultural memory, has already been shaped by earlier tradents; shaped first as the product of neurobiological memory, and second as the product of social or cultural memory, before it contributes to shaping the next person who receives it and passes it on to others. Thus every stage in this chain of transmission and reception reflects not only the circumstances of the present but also the impact of the past.

To put it in other words, social memory theory, as Keith and others present it, recognises two things. The first is the contribution and impact of the redactional interests of particular authors or tradents who shape the traditions or social or cultural memories that they receive. The second is that those traditions or memories shape the tradents who receive them. Further, this approach not only recognises but perhaps emphasizes that the past helps to shape the concerns and interests of the present.7

Closely related to the realisation that the past shapes the present is the now widely accepted point that there is no tradition without interpretation. As Keith argues, in what he characterises as his criticism of the criteria approach to the study of the historical Jesus, it is not possible to disaggregate early authentic traditions about Jesus from the accretions of a later interpretative framework. All tradition is interpreted tradition, whether it concerns claims about Jesus or about anyone or anything else. Historians can and should make judgments about verifiable claims about the past; but this does not mean that they can access or reconstruct what happened in the past, except insofar as past events have been interpreted or presented or remembered by the witnesses on whose testimony historians must rely, along with any other evidence that might be available. Thus, as Keith observes, historians may hypothesize about ‘what combination of socio-historical conditions and past reality could have led to the interpretations of Jesus that we find in the Gospels.’8 But they are proposing a model to account for the evidence, not a reconstruction of what really happened. They are using their own critical skills, themselves shaped by the weight of inherited academic traditions and understandings, and asking questions of what evidence we have from the past, so their models are themselves the product of past and present, just like the evidence on which they draw.

Keith’s account of how both historians and their presents are shaped by the past is clear and persuasive, but it is hardly an account with which many if any historians would disagree. So my questions are not about the theoretical framework of studies in social and cultural memory. Nor are they about the relevance and value of this approach for the study of the gospels, whether with reference to the historical Jesus, or with the formation and development of the gospels as written texts. Rather, they concern the extent to which social or cultural memory really changes our understanding of the relationship between past and present, and about the nature of the historian’s task.

Social memory theory, as Keith presents it, conceptualises and makes explicit the continuing and changing impact of the past on the present.9 Thus it offers an illuminating theoretical framework in which historians (and others too) may reflect on the fact that all human claims must be processed and therefore reconceived, or even to varying degrees distorted or deconstructed and reconstructed by neurobiological and cognitive processes that shape the way in which all human beings perceive the world and their relationship to everyone and everything else. But I question the extent to which it need affect or change the way in which most historians already approach their task, even if they do so without the benefit of social memory theory (as some do today, and many have done in the past). As a theoretical or conceptual framework social memory theory may be helpful to the extent that it casts lights on what historians do already, and may help them to be more reflective about the constraints under which they operate, but it does not suggest a new methodological framework or approach for their work. Thus the main benefit of social memory theory is that it helps to clarify and to articulate what many historians already do, not that it provides fundamentally new insights that mean that many historians should be doing something that they are not doing already. Of course there is no tradition that is not shaped by human interpretation. And certainly the gospels ‘are not the type of ground that can be dug’, as Keith memorably puts it.10 Rather, the best way for the historian to approach questions about Jesus is on the basis of the written texts that they have, instead of ‘on hypothetical reconstructed tradition-histories they do not.’11

But do we need social memory theory to make these points? And does it matter if we refer to the artefacts as ‘traditions’, as many scholars are accustomed to do, or as ‘Jesus-memories’, to use the term that Keith prefers?12 Perhaps the term ‘Jesus-memories’ may help us to think more clearly about the relationship between past and present, and between tradition and tradent. If so, I would happily adopt it. But at one point in his monograph Keith suggests that it is possible to do history in the way that he suggests that it be done without any appeal to social memory theory.13 I think that I have to agree. Keith does not need social memory to support the claims that he makes about why the historical Jesus was very likely not scribally literate, although some people claimed that he was. But his use of the theory, whether applied to this particular historical question or to others, certainly casts light on what it is that historians do when they acknowledge that the present is shaped by the past.

3 Alan Kirk, Q in Matthew

Turning then to the work of Alan Kirk, we find a focus on how a single scribe could have written Matthew’s gospel, using his memory to draw on Mark and Q.14 Put like that, it might be easy to suppose that Kirk’s interest in memory is in purely individual and neurobiological memory as the product of an isolated mind, the way in which an individual can train his or her mind in order to increase his or her powers of recall. That would be incorrect, as anyone familiar with his earlier work would expect.15 Rather, Kirk frames his study of how scribes used memory in the wider context of ancient media conditions and practices; here, as elsewhere, he contributes to the removal of false binary oppositions between oral and written media, emphasizing the importance of the interface between them, and the interfacial way in which each related to the other. Thus while Kirk agrees with those who critique ‘conceptions of synoptic source relations as closed circuits of copying and editing’16 he rejects what he characterizes as their over oralization of synoptic source relationships. If the pendulum has swung from an unduly written to an unduly oral model, Kirk seeks to bring it back to the middle and find equilibrium in a model of synoptic source relationships; a model that ‘can accommodate rather than marginalize the powerful effects of the written media’ and can recognize that ‘oral utilization practices are a major factor in the rise of variation in synoptic parallels’.17 Thus the first step of his argument is to establish ‘the interactive or interfacial relationship between the properties of orality and writing’.18

Step two is to establish which model for ancient writing might best fit what we find in the synoptic gospels, which leads him to posit scribal literacy as the most likely context in which the gospels were written. Thus he characterizes their anonymous writers not as analogous to elite literary authors such as Livy or Plutarch, but as scribal tradents. Not as elite authors, but as scribally educated and scribally literate tradents who participate in the process of transmitting the tradition, already formed through social memory, and recorded in written structured texts such as Mark and (on his hypotheis) Q. Thus they are scribes who have a significant cultural stake in the tradition that they transmit, and who are educated in such a way that they have so internalized their tradition in their minds that they can use their memory to work through the texts that they know, and to combine them in new forms that will continue to reflect and transmit the tradition that has already shaped both them and the communities for which they write. Here, as elsewhere, I summarise detailed and sophisticated arguments, and can note only that Kirk draws on a wealth of data that demonstrates why it is not only historically plausible but also historically likely that the gospel writers were able to operate in the way that his hypothesis suggests. Scribes as tradents were ‘individuals in a community who through an act of memory bring to bear the tradition upon the present realities of the tradent community’.19

So, on Kirk’s hypothesis, what did these scribal tradents do? Using the normal capacities of trained individual human neurobiological memory capacity they gained operational control of written texts (which are themselves the product of cultural memory). This means that they could move through them again and again to pull out details that they could combine in new ways. And they did this by following cues embedded in the structures of the texts; cues that were themselves the product of those texts themselves having been carefully scribally structured as cultural artifacts, which meant that subsequent scribes could access and use them in this way. This is the hypothesis that Kirk proposes, the model that he constructs, to explain how a scribal tradent to whom we habitually refer as Matthew could have combined two written sources, a structured written Mark and a structured written Q, to produce a structured written Matthew. To test this hypothesis, he considers two sections of Matthew that are problematic for the two-document hypothesis, to see if an appeal to the use of cultural and scribal memory of the sort that he describes will work.

One is the double tradition found in Matthew 5–7, the Sermon on the Mount. Here, on the two-document hypothesis, an explanation is required as to how and why Matthew used double tradition in the way that he did in the Sermon on the Mount. For, as Kirk notes, although the Sermon on the Mount and the Sermon on the Plain share a common macro-sequence, the former contains a large amount of double tradition that appears much later in Luke. Further, there are significant differences of wording between double-tradition materials in the two discourses, even in material that appears in a common sequence.20 The other is the triple tradition found in Matthew 4.23–11.30. Here, especially in Matthew 8–11, Matthew breaks up and reorders Mark in a way that he does not do elsewhere.21 Matthew has transposed or reordered material that he took both from Mark and from Q. Thus what Matthew has done here, on the two-document hypothesis, is a counterpart to what he did in the Sermon on the Mount. His transpositions of Mark ‘are of a piece with his reordering of Q.’22 In each case it is difficult to see how one author could have drawn on two written sources if he needed visual access to both as he wrote; for how could he copy from two scrolls, or even from a scroll and a pile of wax notebooks or similar? Furthermore, it is difficult to see even how he could have drawn on one written source open before him if that written source were a scroll, given the difficulties in randomly accessing the content of a scroll as those difficulties are usually understood. Thus the two-document hypothesis struggles to explain the composition of this part of Matthew’s gospel when it makes assumptions that are based on visual access to written sources, or based on short term memory recall of a recently-read or recently-heard written text.

And so we come to the strength and explanatory power of Kirk’s hypothesis: that a scribal tradent drew on his written sources not by looking at them, but by having absorbed those written texts in such a way that he had operational control of them in his working memory; in other words, that he had the ability to access and draw on them in an order that differed from, but can be explained by, the order in which they were written. At certain points in his use of Mark and Q, Kirk suggests, Matthew addresses the ‘dauntingly technical problem’ of ‘how to combine two overlapping but independently ordered sources coherently while harnessing their authority for his own reenactment of the tradition.’23 In each case, as I understand it, the explanation goes like this (and really, I am simplifying very drastically indeed.) Mark and Q are both texts that had a recognizable literary structure at both a macro and a micro level. At the macro level we might think of them as having a certain number of chapters, and the scribe could access the text through memory at the beginning of any chapter. This is what Kirk means by the scribe working horizontally ‘across the horizontal axis of topoi sequence’24 in either Q or Mark. Then, at the micro level, we might think of each text having a certain number of sub-divisions in each chapter, and the scribe could access the sub-divisions through memory once he was in any given chapter. This is what Kirk means when he argues that the scribe is able to move down the vertical axis of the text through the constituent elements of each chapter. Ex hypothesi, his proposal for how the scribal tradent who composed Matthew actually worked has enormous explanatory power, for it offers a plausible account of how that scribe drew on written sources and combined them in the order that he did. It also fits with how scientists appear to understand the working of neurobiological memory, and how historians understand cultural memory and what at least some ancient scribes were able to do. The use of memory, as Kirk presents it, helps to resolve difficulties in the two-document hypothesis, and so provides further support for the model of synoptic relationships that it supports. To quote Kirk’s own words, the patterns that he finds in the way that Matthew uses his sources ‘attest to a memory-grounded competence … combined with parsimony of cognitive effort. In antiquity, memory played an instrumental role in utilization of written works, making it possible to overcome the constraints of the scroll.’25

Of course there are questions, even if in my view they do not challenge the fundamental achievement of Kirk’s groundbreaking work. Some may wish to question whether there is a degree of circularity in his reasoning about Q, since his argument both presupposes and concludes that it was a structured written text. But that, I think, would be to misunderstand his point. Kirk notes that other reasons have been given in support of the claim that Q was likely a structured written text that are quite independent of his proposal, and it is in any case the nature of a hypothesis that it makes a proposal and then seeks to test it.

More significant perhaps are questions about the degree of memory control that a scribe might have had of his sources, and whether that might mean that Kirk’s work could be used in favour not only of the two-document hypothesis but also of the Farrar hypothesis (i.e. Markan priority, but without Q). Kirk’s understanding of how a scribe with full access to both Mark and Q through the operation of his memory can certainly explain the transpositions in order of both double and triple tradition in Matthew, so it addresses what was a significant weakness in the two-document hypothesis. But that need not mean that Kirk’s model of how scribes used memory when composing new texts could not also be deployed in support of a rival solution to the synoptic problem. I find Kirk’s account of how Matthew used Mark and Q not only plausible but also persuasive, but it remains to be seen whether an advocate of the Farrer hypothesis could make use of Kirk’s account of how scribal tradents worked, but with reference to Luke drawing on both Mark and Matthew. Such use would involve a scribally literate Luke working forwards and backwards through his sources, but without the need to be accessing one or more scrolls. Such a scribe might work forward less consistently than Kirk’s Matthew does, but Kirk does allow his Matthew to go backwards as well as forwards, so I wonder if his case for the two-document hypothesis might be said to be better than that of the Farrar hypothesis, but not necessarily to rule it out altogether? On Kirk’s model, the two-document hypothesis seems to be simpler (though certainly not simple from a print perspective) and more economic than the Farrar hypothesis, but I wonder how decisive that is? If Kirk’s arguments for a scribally ordered Q explain how Matthew could work through both it and Mark then I wonder if similar arguments might explain how Luke worked through both Mark and Matthew, especially if he were a scribe who felt more bound to one of his sources than to the other?

Kirk’s use of memory theory, which locates trained individual neurobiological memory in the context of cultural memory and scribal tradition, and at the interface between writing and orality, offers a significant advance in our understanding of how the synoptic gospels came to be written. It is a remarkable achievement, which provides further support for the explanatory power of the two-document hypothesis, and for a structured written Q. Readers who disagree with Kirk’s conclusions about the synoptic problem and Q may nevertheless agree with his insights about how scribes composed on the basis primarily of memory, so his contribution to insights about memory may be more important than their application to a particular solution to the synoptic problem.


An earlier version of this essay was presented at a joint session of the Synoptic Gospels and Q sections at the 2018 Annual Meeting of the Society of Biblical Literature. I am grateful to Elizabeth Shively for her invitation to participate, and to fellow panelists Helen Bond, Francis Watson, Chris Keith and Alan Kirk. For a much wider ranging survey of the use of social memory theory in the study of the gospels see Chris Keith, ‘Social Memory Theory and Gospels Research: The First Decade (Part One)’, Early Christianity 6 (2015) 354–376, and Idem, ‘Social Memory Theory and Gospels Research: The First Decade (Part Two)’, Early Christianity 6 (2015) 517–542.


Chris Keith, Jesus’ Literacy: Scribal Culture and the Teacher from Galilee (LNTS 413/LHJS 8; London: Bloomsbury T&T Clark, 2011).


Alan Kirk, Q in Matthew: Ancient Media, Memory, and Early Scribal Transmission of the Jesus Tradition (LNTS 564; London: Bloomsbury T&T Clark, 2017).


The focus of this essay on methodological issues around the use of social memory theory means that I do not engage with the specific question that Keith addresses, namely whether Jesus was scribally literate. On that point let me observe only that I find his argument and approach both elegant and compelling.


Jesus’ Literacy, 29.


Here I quote from his own summary of his approach, as presented in Boston at the 2018 ­Annual Meeting of the Society of Biblical Literature.


Jesus’ Literacy, 174.


Here I quote from his own summary of his approach, as presented in Boston at the 2018 ­Annual Meeting of the Society of Biblical Literature. See further Jesus’ Literacy, 175.


In this paragraph in particular my observations are shaped by Keith’s response to an earlier version of this essay.


Jesus’ Literacy, 54.


Jesus’ Literacy, 69. Here Keith is critiquing the way in which Form Criticism has been practiced. His point is not intended to address the existence of a hypothetical source (as opposed to a hypothetical reconstructed tradition-history) such as Q, which plays an ­important part in Kirk’s arguments about Matthew’s use of memory, as discussed below.


But see Jesus’ Literacy, 124, where Keith glosses ‘Gospel traditions’ as ‘Jesus-memories’, perhaps implying or acknowledging that they are interchangeable terms.


Jesus’ Literacy, 69.


Kirk also addresses a number of wider implications of his work for historians of early Christianity, including the nature of the relationship between reified communities associated with Mark and with Q, and the question of whether Q may be used as a source for a distinctive Galilean form of early Christianity, but these aspects of his work (which I find persuasive) fall outside the scope of this review.


See, for example, Alan Kirk, ‘Social and Cultural Memory’, in Alan Kirk and Tom Thatcher, eds., Memory Text and Tradition: Uses of the Past in Early Christianity (Semeia Studies 52: Atlanta and Leiden: SBL and Brill, 2005), 1–24, now reprinted, with other essays, in Alan Kirk, Memory and the Jesus Tradition (The Reception of Jesus in the First Three Centuries; London: Bloomsbury T&T Clark, 2018).


Q in Matthew, 18.


Q in Matthew, 19.


Q in Matthew, 28.


Q in Matthew, 111.


Q in Matthew, 184.


Q in Matthew, 225–231.


Q in Matthew, 225.


Q in Matthew, 261. Here Kirk refers to Matthew’s use of Mark and Q in Matthew 3–13.


Q in Matthew, 183, where he refers to Q as a cultural and structured work, but could equally be referring to Mark.


Q in Matthew, 234. Here Kirk refers specifically to Mark, but it comes in a context where he has just noted (233–234) that the pattern of Matthew’s use of Mark is ‘virtually identical to his utilization of Q in the Sermon on the Mount: consistent forward movement from his absolute position in the source, from this moving fulcrum reaching forward in the source for additional materials, while maintaining the internal order of transposed sequences.’

Content Metrics

All Time Past Year Past 30 Days
Abstract Views 351 40 2
Full Text Views 209 8 2
PDF Views & Downloads 92 16 0