Saturday, October 29, 2011

How Well Does the Jesus Seminar Follow Its Own Rules? Part II

In a previous post concerning whether the Jesus Seminar follows its own rules, I noted that its voting members appear to favor the Gospels of Thomas and Q over the canonical gospels (i.e., Matthew, Mark, Luke, and John), but I had yet to take the dating of the various sources into consideration. I do so in this post.

The dating of sources is no easy task, however; it is more art than science. In the table below, I summarize two dating schemes: one that I refer to as the conventional dating scheme because it reflects the dates that most critical biblical scholars assign to the various sources of Jesus's sayings. It is based largely on Bart Ehrman's work, primarily because he has explicitly documented the dates of more sources than have most other biblical scholars. Where Ehrman is "silent" on the date, I use the "Early Christian Writings" website for dates. The other dating scheme I refer to as the Jesus Seminar because it appears to reflect the dates that most members of the Jesus Seminar ascribe to. It is based primarily on John Dominic Crossan's work and is supplemented by his Jesus Seminar colleague Robert Miller's book, "The Complete Gospels" where Crossan does not provide a date.

Looking at the table, what is striking is the difference between the two schemes. Note that Crossan dates all of the canonical gospels no earlier than 75 AD, while dating Q and Thomas to the mid-50s. These dates are at odds to the dates that most biblical scholars assign to these gospels. Q is typically dated to the mid-60s and Thomas to the late 1st-Century or early 2nd-Century.

Table 1: Conventional and Jesus Seminar Dating of Sources


Conventional
(Ehrman)
Jesus Seminar
(Crossan)
Canonical



Matthew
80-85
90

Mark
65-70
75-80

Luke
80-85
95

John
90-95
105-125

Acts
80-85
120

1 Corinthians
53
53

1 Thessalonians
49
50

1 Timothy
80-110
125

2 Peter
80-110
135

2 Timothy
80-110
125

James
110
100

Revelation
95-100
95-100
Noncanonical



Thomas
110-120
55-70

Q
65-70
55

1 Clement
95
98-100

2 Clement
150
150

Apocryphon of James
100-150
125

Epistle of Barnabas
130
95-100

Dialogue of the Savior
120-180
60-80

Didache
100
100-150

Egerton Gospel
100
55

Papyrus Vien 2325 - Fayyum
125
55

Gospel of Peter
110-120
150

Gospel of the Ebionites
110
150

Gospel of the Hebrews
110
55

Gospel of the Nazareans
125
150

Shepherd of Hermas
120-140
100

Letter of Ignatius - Ephesians
125
110

Letter of Ignatius - Philad
110
110

Letter of Ignatius - Polycarp
110
110

Letter of Ignatius - Smyrna
110
110

Letter of Ignatius - Trallians
110
110

Justin Martyr - Trypho
150
150

Papyrus Oxyrhynchus 1 (Greek Thomas)
110-120
55-70

Papyrus Oxyrhynchus 654 (Greek Thomas)
110-120
55-70

Papyrus Oxyrhynchus 655 (Greek Thomas)
110-120
55-70

Papyrus Oxyrhynchus 840
110-160
85

Papyrus Oxyrhynchus 1224
50-140
55

Letter of Polycarp - Philip
110
140

This early dating takes on even greater importance, at least when considerng Crossan's own work on the historical Jesus and the sayings he includes in his inventory for determining whom Jesus was. After assigning dates to the various sayings, he then groups them into what he calls "complexes" (what most people would probably call "themes") and sorts them into four different strata: 30-60 AD, 70-80 AD, 90-120 AD, and 120-150 AD. Then he declares that for his reconstruction of the historical Jesus, he will only consider sayings that appear in the first strata.

This methodological move on his part has profound consequences. By dating the first stratum from 30-60, Crossan is able to include all of Jesus' sayings in Thomas and Q and eliminate all sayings in the canonical gospels that are not traceable back to Q. In other words, sayings found in Mark as well as those that are unique to Matthew ("M") and Luke ("L") are not considered by Crossan in reconstructing the historical Jesus.
Crossan also builds upon John Kloppenborg's work in order to argue that the non-apocalyptic portions of Q (Q1) predate the apocalyptic portions (Q2) although he believes that all of Q came together in the 50's. As Bart Ehrman notes, however, claiming to distinguish layers within Q and then drawing conclusions on such layers is probably more than what is warranted given the fact that we don't even have the document.
According to this line, the original edition of Q did not have the apocalyptic traditions about Jesus. These were only added later, when the document was edited by Christians who were a bit obsessed with the imminent end of the age. Thus, according to this theory, Q as we have it (well, even though we don't have it), may be an apocalyptic document. But in fact it provides evidence of a non-apocalyptic Jesus" (Bart Ehrman, "The New Testament: A Historical Introduction to the Early Christian Writings," 3rd ed., p. 257).
All we know about Q is what Luke and Matthew included in their gospels. If there were portions of Q that neither decided to include, they are most likely lost to history. And, of course, some of the material that is only found in Matthew ("M") or in Luke ("L") could also have come from Q:
"Despite the exuberant claims of some scholars, we cannot fully know what Q contained because the document has been lost. We have access to it only through the materials that Matthew and Luke both decided to include in their accounts, and it would be foolish to think that one or both of them included the entire document. Indeed, if only one of them included a passage from Q, then we would have no solid grounds for knowing that it came from Q rather than, say, M or L. It is entirely possible, for example, that Q had a Passion narrative, and that neither Matthew nor Luke chose to use it., or that only one of them chose not to do so (so that some of the verses of Matthew's or Luke's Passion narrative not found in Mark actually derive from Q). At the same time, it is equally possible that Q was almost entirely sayings, without a Passion narrative (or nearly any other narrative). Regrettably, we will never know, unless, of course, Q itself should serendipitously turn up" (Bart Ehrman, "The New Testament: A Historical Introduction to the Early Christian Writings," p. 88)!
So, what does an analysis of the voting yield. In separate multivariate regressions (i.e., an OLS and an ordered logit -- Table 2, which is quite large but summarizes the OLS results, appears at the bottom of the post) that included the conventional dating scheme as one of the independent variables, the two most significant predictors of a higher score were whether a saying could be traced back to Q or Thomas, while the number of indpendent sources had a negative effect on a saying's score. Just as interesting is the finding that a saying's date had virtually no impact on whether the saying was voted red, pink, gray, or black. There was a slight negative effect, such that the older the source, the lower the saying's score. However, the effect was minimal and of little consequence.

The exception to this occurs if one includes Crossan's strata in place of the estimated dates of the gospels. When you do that, sayings located in the first stratum receive higher scores than those that are not. Recall, however, that only sayings that can be traced back to Q or Thomas are included in this strata (well, sayings found in Paul's letters are also included, but they are few and far between), so this lends empirical support to the charges of critics who argue that for Crossan and the Jesus Seminar, non-canonical sources are superior to canonical gospels. Bart Ehrman summarizes the concerns of many scholars quite nicely:
Crossan engages in a detailed analysis to argue that other sources not found in the New Testament are earlier than the sources that are. These others include such documents as the 'Egerton Gospel,' a fragmentary text from the second century that contains four stories about Jesus; the Gospel of the Hebrews, which... no longer survives, but is quoted a bit by some church fathers in the late second to the early fifth centuries; and parts of the Gospel of Peter, which survives again only as fragments. Such sources, Crossan claims, provide more reliable access to Jesus than the New Testament Gospels, which everyone, including Crossan, dates to the first century.
But this strikes most scholars as a case of special pleading. Most recognize clear and certain reasons for dating the New Testament Gospels to the first century. But giving yet earlier dates to noncanonical Gospels that are, in most cases, not quoted or even mentioned by early Christian writers until many, many decades later seems to be overly speculative and driven by an ultimate objective of claiming that Jesus was not an apocalypticist even though our earliest sources indicate that he was (Ehrman, "The New Testament," p. 258).
What then are we to conclude? The modern tools of social science and literary theory to not appear to prevent modern scholars from creating Jesus in their own image, a critique that Albert Schweitzer made against 18th and 19th century biblical scholars. As the table above illustrates, the dating of sources is more art than science, and while multiple, independent sources are in theory useful for separating the wheat from the chaff, their presence (or absence) does not seem to get in the way of scholars rejecting (or accepting) the historicity of a particular saying or deed.  In short, in spite of scholars' best efforts to objectively uncover what Jesus said and did fraught with difficulties. The temptation to create Jesus in our own image is so powerful that it is difficult to transcend in our quest for objectivity.

Table 2: OLS Regression

Model
1
2
3
4
5
6
Intercept
.102*
.117*
.126**
.062**
.137
.062**
Gospel







Matthew
.061**
.070***
.060**
.061**
.070***
.071***

Mark
.076**
.079***
.076**
.085***
.081***
.090***

Luke
.082***
.088***
.080***
.075***
.086***
.082***

John
-.010
-.004
.005
.007
.014
.012

Thomas
.033
.044
.019
.044
.022*
.049*
Source







Matthean
.077*
.060
.083**
.090**
.069*
.077**

Markan
.106
.090**
.109***
.122***
.098***
.112***

Lukan
.105**
.090**
.115***
.125***
.104**
.114***

Johanine
.023
.007
.024
.017
.011
.005

Thomas
.187***
.173
.186***
.181***
.176***
.173***

Quelle
.212***

.205***
.177***



Q1

.226***


.221***
.195***

Q2

.151***


.147***
.123***

Q3

-.086


-.092
-.121

QU

.125**


.120**
.095**

# of Sources
-.066*
-.056*
-.066**
-.061**
-.060**
-.056
Dating







Date (Ehrman)
-.000
-.000





Date (Crossan)


-.001

-.001


1st Stratum



.062**

.065**
Adjusted R2
.270
.286
.271
.275
.286
.291