This article was distributed on arxiv.org and published in Journal for the Study of the Historical Jesus.
https://arxiv.org/pdf/2403.14883
History can be understood as a field of fields, where many different disciplines can be potentially relevant to the questions studied. One of these fields is statistical analysis. We have known some forays into such study already, for example in various authorship attribution studies.
What's involved here is the development of a certain valid method of anaylsis for discerning historicity. The categories involved are a certain essential historicity, which doesn't speak to the historicity of specific episodes in a narrative but which does involve a narrative that is constructed with reference to historical events, and fiction. The category of fiction doesn't mean that it has to be considered naive and uninformed fiction. It means just that it is fiction in the way fiction has been constructed, which can involve an attempt to achieve verisimilitude. Developing a method of analysis for this kind of question is difficult, as we all should be aware.
This image shows visually the data discussed in the paper:
The analysis of the paper considers the proposed alternative of a sort of historical novelization:
This involves allowing for a robust test of the alternative hypothesis of fiction, including the kind of fiction that aims for verisimilitude. That analysis involves samples of actual historicizing fiction.
The authors discuss the fact that any statistical analysis is probabilistic in nature. Any answer from such an analysis falls short of 'proof', except that it provides a 'probabilistic answer'. Fortunately, this is the same kind of texture that historical analysis has, as any sound analysis of history would show. Nothing can be proved in the sense of going beyond a 'probabilistic answer' to an apodictic one, like some kind of proof in geometry about a theorem following from given axioms.
This chart outlines the results of the statistical analysis.
The authors discuss some of the nuances of the comparison of Josephus and the Gospels and Acts with the name frequencies in Palestine:
Beyond the statistical analysis they did, they discuss other ancient Christian texts as well:
This is not a vindication of Bauckham, who went well beyond what conclusions can be justified on the basis of the evidence. Yet credit must be assigned for considering it a possible mode of analysis.
It's not often that a test of historicity is developed. This paper is bold enough to try to develop the methodological underpinnings of this analysis, updating the discussion to bring it to a point where it can be accepted as showing its probabilistic answer, as far as it is the leading edge of research on the subject.
https://arxiv.org/pdf/2403.14883
History can be understood as a field of fields, where many different disciplines can be potentially relevant to the questions studied. One of these fields is statistical analysis. We have known some forays into such study already, for example in various authorship attribution studies.
What's involved here is the development of a certain valid method of anaylsis for discerning historicity. The categories involved are a certain essential historicity, which doesn't speak to the historicity of specific episodes in a narrative but which does involve a narrative that is constructed with reference to historical events, and fiction. The category of fiction doesn't mean that it has to be considered naive and uninformed fiction. It means just that it is fiction in the way fiction has been constructed, which can involve an attempt to achieve verisimilitude. Developing a method of analysis for this kind of question is difficult, as we all should be aware.
This image shows visually the data discussed in the paper:
Below, we demonstrate that the Gospels and Acts, alongside Josephus, reflect a population of historical Palestinian Jewish names from 4 BCE–73 CE as opposed to a population of Diaspora males from the same period. Among other conclusions drawn from our analysis, we also demonstrate that name samples from the most robust historical novels available to us – novels that did indeed rely on Josephus, GA, and the Hebrew Bible for their naming practices – were not able to achieve naming statistics that significantly reflect Palestinian Jews from 4 BCE–73 CE. This latter analysis implicitly responds to a historically nuanced scenario envisioned by GB wherein the Gospel authors invented certain names but also relied on historical source material found in Josephus.19 In short, we show that name statistics in GA fit into their historical context well, and that they fit significantly better than fictitious samples.
This involves allowing for a robust test of the alternative hypothesis of fiction, including the kind of fiction that aims for verisimilitude. That analysis involves samples of actual historicizing fiction.
The authors discuss the fact that any statistical analysis is probabilistic in nature. Any answer from such an analysis falls short of 'proof', except that it provides a 'probabilistic answer'. Fortunately, this is the same kind of texture that historical analysis has, as any sound analysis of history would show. Nothing can be proved in the sense of going beyond a 'probabilistic answer' to an apodictic one, like some kind of proof in geometry about a theorem following from given axioms.
This chart outlines the results of the statistical analysis.
Ilan’s database relies on epigraphic as well as literary sources, including ossuaries from the region of Jerusalem, which show a preference for Biblical over Greek names. Conversely, Josephus’ audience, focus, and literary milieu results in a greater preference for Greek over Biblical or Semitic names of Jewish persons. This is exemplified in his generally more Hellenized orthography and formal Greek case suffixes, while GA show a preference toward less official orthography coinciding, as Ilan notes, to common pronunciation.
Beyond the statistical analysis they did, they discuss other ancient Christian texts as well:
Even if, contrary to the evidence we possess, accurate and intricate naming patterns were achieved by a fictionalizing author of antiquity, there are several additional reasons to doubt that the name statistics in GA result from a fictionalizing process. First, in second- to thirdcentury Gospels like The Sophia of Jesus Christ, The Gospel of Judas, and The Gospel of Mary, any hint of Jesus’ Palestinian environment is overshadowed by mythological concerns and persons. Second, Gospels like The Infancy Gospel of Thomas and The Infancy Gospel of James may add an unusual name like Zenon, a New Testament name like Zaccheaus, or a popular Western Diaspora name like Samuel, but the general trend reflects a disinterest in capturing any kind of realistic onomastic verisimilitude. This is seen, particularly, in the lack of qualifiers to disambiguate or distinguish persons with popular names in compositions like The Gospel of Mary and The Coptic Gospel of Philip (for more on name disambiguation, see below). Third, to assume that a later author would diligently examine a source like Josephus or other material from Palestine, take a rough inventory of name popularity, and attempt to reproduce it assumes, against available evidence, a hyper-sensitive focus on names that no critical NT reader considered meaningful for two millennia until Richard Bauckham began the research that resulted in Jesus and the Eyewitnesses. These three observations align with the published data on personal name retention, which suggests that names, being arbitrary and easily forgettable, are among the least integrated and lasting pieces of information in the recollection of stories or experiences.
This is not a vindication of Bauckham, who went well beyond what conclusions can be justified on the basis of the evidence. Yet credit must be assigned for considering it a possible mode of analysis.
It's not often that a test of historicity is developed. This paper is bold enough to try to develop the methodological underpinnings of this analysis, updating the discussion to bring it to a point where it can be accepted as showing its probabilistic answer, as far as it is the leading edge of research on the subject.
Statistics: Posted by Peter Kirby — Sat Nov 02, 2024 6:02 pm