Thursday, June 30, 2011

Kuzari: A Reply to Dovid Kornreich on Evidence and Hypotheses


I have yet to fulfill my promise to Dovid Kornreich: I agreed to explain how I think the Sinai story originated and developed.

Today I want to take yet another step toward directly formulating this explanation, but let me first review earlier steps:
  • In "Kuzari: Belief and Evidence (and Bias, Oh My!)," I bracketed the task--i.e., my speculative explanation on the Sinai story--to give what I hope is proper perspective on its value. The best I can hope for is a fair approach to and accounting for the observed evidence. This means we cannot simply grant that the story might be true as it appears in the Torah because that smuggles in the assumption (among others) that the God of Moses existed. Anyone who wants to claim that the story is true as reported in today's Torah must show both evidence and argument for the existence of that God and his involvement in the event in question. Incidentally, that anyone might also want to show both evidence and argument for Moses, as the existence of Moses is considered unlikely.
  • In "Kuzari: Deuteronomy Doesn't Validate the Sinai Revelation" I examined Deuteronomy 4:9-40 and concluded that it presented a later account of the Sinai event and interpretation of it. The passages did not, I said, provide us with a report of Sinai as it was happening. My reading was based in part on understanding the context established in Deuteronomy 1. I concluded that we could not use the Deuteronomy 4 passages to validate the Sinai event itself, but that we could use them to discuss the understanding of the Sinai story.
  • I presented the Sinai stories from the J, E and P sources in "Kuzari: Three Sinai Stories." They are quite different and remarkable accounts. J is about the coming of God to Sinai and the establishment of Moses and Aaron as the official go-betweens of God and Israel. The account is more personal in E. There remains a distance between God and the people, but Moses functions as more a translator in E, whereas I see him as a representative in J. God is a black box in P, and Moses alone enters. All knowledge and authority rest with Moses.
  • Most recently, I posted "Kuzari: Why Aren't There More Sinai-Like Stories?" to address the question in the title. My answer is that we have three Sinai-like stories: J, E, and P. We also have stories with one or more elements such as we find in the Sinai story. What we do not have is another story from another tradition or culture that is exactly like Sinai. But we don't need carbon copies of Sinai, and a demand for them is unreasonable.
Before I can get to a direct formulation, I have to address one more topic, which comes from comments made by Dovid Kornreich:
Please specify (in future posts, perhaps) 1) the observed evidence and 2) tested hypotheses which reconstruct the textual history of Deuteronomy--which do not commit logical fallacies. Namely: of assuming the conclusion at the outset. Meaning they do not initially view the evidence through the prism of the conclusion.

I have yet to come across such fallacy-free evidence and hypothesis testing in Biblical scholarship.
This is a great comment deserving serious consideration, and the topic it raises concerns the nature of evidence. What is the evidence? What does it mean for something to be taken as evidence? What is the relationship between evidence and hypothesis?

These are huge questions that I think can be usefully approached by first establishing the big picture. For us, the big-big picture is essentially a model of the world and how it works. There are several ways to specify the model, but let's try this:
  1. The natural world operates according to physical laws.
  2. Events in the natural world have physical components.
  3. Events can cause other, subsequent events.
  4. Some events can literally be more effective than other events.
  5. Some events are more likely to have regular causes than other events.
I assume we all can agree with the general outline of this model, although some may question or object to specific elements. It is hopefully beyond question that the natural world, per specifications 1 and 2, has enough regularity and predictability to allow us to develop a stable picture of it. Now I'm aware that our human ability to apprehend and describe the world breaks down at some extremes, and this is fine. Ultimately, however, with physical laws we know we are talking about the measurable behavior of matter and the transfer of information.

If someone wants to modify the model by saying--for example--"The natural world operates according to physical and spiritual laws," then I need to know what we are talking about when we use the term spiritual laws. I need to know what spirit is, what it does, and how we build knowledge of it.

But I want to return to the model I sketched out before because we do not need to commit ourselves to it. We can test the model and ask questions about it:
  • (a) Given the model I described, how do we assess the likelihood of specific observations?
  • (b) Given a set of observations and the model above, how do we choose the cause-effect chain that best explains the observations?
  • (c) Given a set of observations and different models or different variations within the model above, how do we find the best model or model variation that best explains the observations?
Using all the above, let's consider a statement such as we find in Deuteronomy 16b:
there were thunder claps and lightning flashes, and a thick cloud was upon the mountain, and a very powerful blast of a shofar
In our model, how likely are thunder claps, lightening, and clouds by mountains? Likely within normal ranges. In other words, weather events fit right into our model. So far as I know, there is nothing related to climate or geography that would make a weather event practically impossible at Sinai.

Now, what happens if we use our model and the second bullet (b) above? Well, we can establish different configurations of natural causes and events that would lead up to thunder, lightening and clouds on Sinai at a particular time. Our proposed configuration may or may not be close to the truth, but they will be complete. We have no need to invoke anything beyond the model to develop a minimally viable hypothesis.

How about using our model and the third bullet (c) above? This could be the time to ask whether a model that included "and spiritual" might perform better than the original model for the given observation. It won't, unless we have a way to identify what specifically is spiritual in the observed event. In other words, unless and until we can agree on what information does and does not fall within the category "spiritual," then the concept is superfluous for our purposes.

We can now return to Dovid Kornreich's question to me and talk about the observed evidence. I'll cite biblical scholar Richard Elliott Friedman (from The Bible with Sources Revealed) and list the evidence of the multiple source hypothesis as follows:
  1. Linguistic evidence (p. 7): The different sources reflect the Hebrew language of several distinct periods. The change in language is attested through Hebrew texts outside the Torah.
  2. Terminology (p. 8): Certain words and phrases appear disproportionately and even entirely in some sources but not in others.
  3. Consistent content (p. 10): This is the "different sources use different names for God" line of evidence. More correctly, the sources differ on when the name of God was first revealed to humans. A second line of evidence in this same category concerns sacred objects such as the Tabernacle, the Ark of the Covenant, Urim and Tummim, and so on. Some sources dwell excessively on one or more of these objects while other sources make no mention at all.  A third line of evidence involves the priestly leadership. In the P source, the line of Aaron has exclusive access to the divine. The arguments for this line are more substantial than I can relate here and now, so do read Friedman and others on this. Finally, P is unique among the sources in its concern over ages, dates, measurements, numbers, order, and precise instructions.
  4. Continuity of texts (p. 13): When the sources are separated from one another, each makes a flowing, sensible text. In discussing this line of evidence Friedman addresses an objection I already know is coming, as it is expressed in Kornreich's question. The objection is that the multiple source hypothesis came first, and then the Torah was divided to produce this result. Friedman anticipates this type of objection:
    So much of the text flows smoothly flows smoothly...that it is not possible that any scholar could have constructed it to do so while keeping all the evidence consistently within sources. The scholar would still have to keep all the sources' similar versions of common stories (known as "doublets") separated. The scholar would still have to keep all of the characteristic terminology of each source within the passages attributed to that particular source. The scholar would still have to keep all of the linguistic  evidence for the stages of Hebrew intact, all the occurrences of the divine name consistent within sources, and all the other lines of evidence intact--all of this while producing stories that flow smoothly.
  5. Connections with other parts of the Bible (p. 14): I'll let Friedman's words make the case here.
    When distinguished from one another, the individual sources each have specific affinities with particular portions of the Bible. D has well-known parallels of wording with the book of Jeremiah. P has such parallels with with the book of Ezekiel. J and E are particularly connected with the book Hosea. This is not simply a matter of a coincidence of subject matter in these parallel texts. It is a proper connection of language and views between particular sources and particular prophetic works.
  6. Relationships among the sources to each other and to history (p. 18): We see that each source has connections to specific circumstances in history and to other sources. J appears connected to the kingdom of Judah in the south of Israel. E has connections with northern Israel. Our time frame here is between 922 and 722 BCE. P is connection to the time of Hezekiah, king of Judah from 715-687 BCE. D is associated, as we have previously discussed, with the reign of Josiah, king of Judah from 640-609 BCE. Finally, the P source has a consistent relationship with the prior sources J and E. Its content and order of episodes show it to be an alternative composition to JE.
  7. Convergence (p. 27): I'll once again let Friedman state the case:
    Above all, the strongest evidence establishing the Documentary Hypothesis is that several different lines of evidence converge....The most compelling argument for the hypothesis is that this hypothesis best accounts for the fact that all this evidence of so many kinds comes together so consistently.
We now have the set of seven observations listed above and a hypothesis purporting to describe the causal chain accounting for the observations. This puts us in range of the second bullet item (b) above. With a set of observations and a viable hypothesis, we can propose a model of how the Torah was constructed and then test it through bullet points (a) to (c). The model must recognize all of the seven observations as possible outcomes. The model must also be generally compatible with our big-picture model from before.

I will apologetically avoid sketching out a personal, provisional model of how the Torah was constructed. One reason for this is that Kornreich's question to me can be fully addressed now without such a model. The second reason is that I may need to provide it in the next post, which I expect will be my promised explanation of how the Sinai story originated and developed.

To answer Kornreich's question, then:
  • The observed evidence is such that is enumerated above. We observe, for example, words in the Bible from different periods in the history of the Hebrew language. One explanation for this observation is that preserved content from earlier times was later combined with other content and the whole thing became one composite text.
  • The tested hypotheses are not only the species of the Documentary Hypothesis but species of what I'll call the Divine Inspiration Hypothesis. The latter set ranges from taking the Torah as the word of God transmitted through Moses to taking it as assembled (a la the DH) by divinely inspired redactors. The tests include incorporating new observations and data points and reconciling lines of evidence with each other. In other words, we are not looking simply for an explanation to the language history observation, we are looking also for an explanation that is compatible and consistent with other lines of evidence.
  • Do modern biblical scholars such as Friedman presume the truth of the Documentary Hypothesis at the outset? Friedman's statements in the "Continuity of texts" line of evidence (#4) argue against circularity. These statements also suggest how circularity could be exposed and the DH challenged. Now, we do need to bring some assumptions to the table beforehand. For example, if we assume that there are no contradictions at all in the Bible, we can come up with all sorts of ingenious ways to explain apparent contradictions to make them "go away." The real question is how do we choose between the assumption that Torah contains no contradictions and the assumption that it may contains contradictory accounts and statements? I don't think that we can answer this question without thinking long and hard about our big-picture model.
To conclude, we arrive at the heart of the disagreement in my perspective and Kornreich's. If I understand his position correctly, he will argue that at least some of the seven observations do not constitute "problems" at all; that is, these are not things that need to be explained.

But this is where I really should invite Dovid himself to respond. And so...my questions to Dovid:
  1. Are the seven observations valid? Which ones are not, and why not?
  2. Of the observations that are valid, how do you explain what we see in the text? 
  3. How do your explanations better account for the observations than explanations under the Documentary Hypothesis?
  4. How would you modify or alter the big-picture model I developed earlier in this post?
  5. Assuming you subscribe to a version of the Divine Inspiration Hypothesis, how do you personally avoid assuming its truth when you are reasoning about what you observe in the Bible and in the sacred works of other religions?
Dovid, I'll look forward to your answers. I'm willing to give them a full post here or to link to your blog if you like. Next up for me: My explanation of how the Sinai story originated and developed.

    6 comments:

    1. "...reconstruct the textual history of Deuteronomy--which do not commit logical fallacies. Namely: of assuming the conclusion at the outset."

      "Do modern biblical scholars such as Friedman presume the truth of the Documentary Hypothesis at the outset? I think that Friedman's statements in the "Continuity of texts" line of evidence (#4) argues against circularity."


      Indisputably not every element of the conclusion is assumed at the outset. Hence, the appearance of circularity is easy to avoid. However, one wouldn't want to assume anything false, and so must examine assumptions in the big picture model.

      The secular approach clearly does not at all assume the DH. However, secular scholars (at least appear to) assume the falsity of supernatural hypotheses. We should examine their reasons for doing so and replace that (apparent) assumption with reasoned conclusions.


      1. Are the seven observations valid? Which ones are not, and why not?

      It's unfortunate that most of the post elaborates on these when they are not very controversial. A computer program recently provided evidence from outside biblical scholarship that it parses the texts coherently and along its natural joints.


      2. Of the observations that are valid, how do you explain what we see in the text?

      Explanation after the fact is easy, as through every set of data there is a perfectly fitting explanatory function. Once we know that we are dealing with a text, the explanation that always works is the intent of the author, which we don't have to stipulate. To demand it would be to make an argument from ignorance.

      3. How do your explanations better account for the observations than explanations under the Documentary Hypothesis?
      4. How would you modify or alter the big-picture model I developed earlier in this post?


      This is where the action is.


      Assuming [one] subscribe[s] to a version of the Divine Inspiration Hypothesis, how [would one] avoid assuming its truth when...reasoning about what [one] observe[s] in the Bible and in the sacred works of other religions?

      Regarding other religions, regardless of one's criteria it will be possible to rate religions by that set of criteria and observe that they have different likelihoods. This is true even for reasoning that does not explicitly assume the truth of one religion.

      The same basically applies when comparing religious to non-religious explanations, except less obviously.

      Note that one can observe which starting assumptions ought to lead to which conclusions, and subjectively select to espouse the neutral criteria that supports one's favored belief, even if the reasons for this selection were assumptions explicitly and illegitimately declaring the truth of a system.

      E.g., if my society demands I believe in Australia, I may notice that modern science compels and justifies belief in Australia, and become an advocate of science. Had my society demanded belief in Atlantis, I would have noticed (consciously or subconsciously) that science is in tension with that belief and believed in something else.

      It is thereby not surprising to find facially neutral premises invoked by all people believing things. It is a special trait of Christianity to say faith is sufficient justification. Not an endearing one.

      ReplyDelete
    2. Are the seven lines of evidence uncontroversial? I wouldn't have thought so, but I don't get around much. I have seen ""different sources use different names for God" paraded around, but this is not actually what DH claims. What I want is responses to the DH as I understand it through Friedman and others.

      "Once we know that we are dealing with a text, the explanation that always works is the intent of the author, which we don't have to stipulate."

      Not sure I read you right, but some say that to posit authorial intent is always a fallacy (i.e., the intentional fallacy). Even the authors themselves are not able to settle intent. Dennett has a nice article on this.

      ReplyDelete
    3. "Not sure I read you right, but some say that to posit authorial intent is always a fallacy (i.e., the intentional fallacy)."

      Regardless of whether or not authorial intent is always important or necessary, one can't assume about any given thing that its cause was not intent, and furthermore one can't demand that a third party tell you a plausible intent. One has to be open to the idea that there was an intent regardless of whether or not one can think of a plausible one.

      "How do you explain that the text looks exactly like one made by combining multiple human sources?"

      "God did it."

      "Why would he do such a thing?!"

      "I can't imagine, but are you implying that were I to present you with a good reason one might have to fake the evidence just so (or why a totally different process would result in the same appearance), you would no longer be troubled by the fact that the text 'looks man-made? And if not, why did you ask?"

      "Yes."

      "So you're saying our inability to imagine a reason is conclusive reason to say it doesn't exist? I'm flattered. Most arguments from ignorance rely merely on the speaker's assumption of having all relevant knowledge, but here you merely assume that between us we have all relevant knowledge."

      ReplyDelete
    4. Ah, but hold on. Intent to communicate is different than intended message.

      If we agree that such-and-such is a text, then we may also agree that the author(s) had an intent to communicate something. Yet we may never agree as to the what and why of the author(s).

      Intent to communicate is, by itself, not especially workable.

      We agree that the Torah is a text or a composite text. Whatever. We therefore agree that there are one or more intents to communicate.

      I am not asking DK to give me an answer of intent for the seven lines of evidence. I am asking him two questions: (1) Do you agree that the observations are valid? and (2) What caused the text to be the way it appears to us in these observations?

      Our job as scholars is to identify features, identify possible causes of these features, and try to evaluate which causes are most likely. We are not making, nor are we demanding, arguments from ignorance. On the contrary, we are trying to make the strongest, most reliable inferences we can based on the little that we actually do know.

      ReplyDelete
    5. "The different sources reflect the Hebrew language of several distinct periods...Relationships among the sources to each other and to history (p. 18): We see that each source has connections to specific circumstances in history and to other sources. J appears connected to the kingdom of Judah in the south of Israel...D is associated, as we have previously discussed, with the reign of Josiah, king of Judah from 640-609 BCE. Finally, the P source has a consistent relationship with the prior sources J and E."

      "(1) Do you agree that the observations are valid?"


      Depending on how they are phrased some of the observations might be presuming the conclusion while others do not. "I saw a homeless man today" presumes one didn't see a CEO pretending to be homeless.

      Does the word "reflect" in the following sentence capture what would be going on if I extrapolated the development of language, wrote a text in today's English, wrote a text in the language of 100 years from now, and jumbled the texts together: "the combined text reflects the English language of distinct periods." I'm not sure. What if I were smart enough to jumble them in my head, as I was writing them, such that there were never two separate texts? Would there be "different sources"?

      "(2) What caused the text to be the way it appears to us in these observations?"

      The text

      I. has all of the signs of being a mishmash of consistent, regular human works of coherent authorship or
      II. does not?

      If I., is it because:

      a) it was created by the crude combination of consistent, regular, human works of coherent authorship;
      b) of the intent of a single author to make it look exactly as if it were a);
      c) unfathomable coincidence?

      If I. b), then

      i. the single author was human or
      ii. the single author was non-human?

      If I. b) ii., the non-human author
      1. is restricted by laws of physics or
      2. unrestricted by laws of physics?

      Establishing I. ever more firmly in the face of II. does nothing to address I. b) and its problems.

      ReplyDelete
    6. Brian,

      Your first point is, of course, valid. I hope for the indulgence of my readers in understanding that for the sake of brevity and clarity I am talking about putative sources. To use your example, I am always saying "I saw [what appeared to be but which I cannot verify with absolute certainty at this time] a homeless man today."

      So, given the assumption of different sources--that is with multiple sources as our provisional hypothesis--the appearance of archaic language forms, the commonality of expressions and concerns across extra-Torah books, etc. can be explained in an intellectually satisfying way. The pattern of the languages can be used to suggest multiple sources. The distribution of content with apparent relationship to history and to other content in the Bible also suggests multiple sources. And so on. What's interesting is that the different lines of evidence suggest roughly the same number and type of sources.

      In your next point, I'm not sure how you would be able to extrapolate the future of a particular language. Words can change meanings and acquire very new ones. They can change part of speech. The syntax can change. The sounds, morphology, and pronunciation can change. Contact with other cultures and technologies can affect vocabulary and register. I wouldn't think it possible for a person to make consistent use of either archaic or future language forms. I would love to see an experiment where someone attempted to write about 10K-50K words in the style of colonial New England elite. Could such a text be written well enough to fool scholars and casual observers?

      Finally, I don't know that DK will agree that the text "has all of the signs of being a mishmash of consistent, regular human works of coherent authorship." DK may argue that the text has two levels, both of which must be taken into account equally. T1 is the plain meaning of the statements, insofar as we can agree on what's plain and what the boundaries of statements are. T2 is the moral and practical teaching of the text. Apparent contradictions at the T1 level can be resolved at the T2 level. But I want to hear DK's actual answer.

      As to what I think is your larger argument: We cannot fully rule out that the Torah is the product of a single mind who crafted the text to appear as a composite work assembled over centuries. But it's highly impossible, perhaps practically so, in our as-yet-unformulated model of textual production.

      In this model, individual authors have have personal styles of derived from the compositional situation, the present state of the language, the technologies of textual production and dissemination, prevailing ideologies of the self and community, and the social institutions that authorize textual production.

      You can't just imagine a single author. You also have to imagine an early Israel that receives his text and accepts it. You have to imagine how a person, long before the concept of "authorship" is ever formulated, finds the time and space to compose in solitude as an author, and then the time to present the full-borne text to the political elite or the people broadly.

      What you need, then, is a new model of textual production and a new narrative of the history behind it.

      But if we agree on the model that I've hinted at, wherein texts encode various elements from their historical moment(s) of composition, then I. b) is not significant enough to carry forward after careful consideration.

      ReplyDelete

    Feel free to comment if you have something substantial and substantiated to say.