The Queen of Sheba Comes to Visit Solomon

Transcription

1 The Queen of Sheba Comes to Visit Solomon Ian C. McKay, 20 April 2011 I recently examined and compared four ancient versions of the story of the census of Israel and Judah ordered by King David, with a view to clarifying the stories mutual relationships and tracing their probable route of descent from hypothetical earlier stories. The approach that seemed most natural to me, being more of a biologist than a Bible student, was to construct a family tree showing hypothetical ancestral versions of the documents, and inferring the most probably features of these earlier documents. I provisionally assumed that the four versions had arisen from a single original version by progressively branching into variants, which each evolved independently from the others, undergoing numerous mutations as they were repeatedly copied and/or edited. I also assumed that each change or mutation probably originated only once, because it is unlikely that exactly the same copying mistake or editorial change would be made independently by two different scribes or editors. I found that there was one and only one family tree that was compatible with these assumptions and with the variable features of the story, and from this tree some unambiguous inferences could be made about the contents of earlier documents. Now I have tried the same approach with another Bible story, the story of the visit of the Queen of Sheba to Solomon, and the results were different. There was no simple, purely branching family tree compatible with these two assumptions and with all the variable features of the story. So I discarded the first assumption and kept the second. In other words, I allowed for the possibility that variants that had diverged might partly recombine again, perhaps because an editor of one version of the story copied bits from an alternative version so as to make the two versions agree more closely. When this possibility is allowed, there can never be a unique solution to the quest for an evolutionary tree, because the end result can always be reached in many different ways. However, out of the many possible evolutionary trees that fit the facts, it is possible to find the simplest, and this has some merits as a working hypothesis. Other things being equal, the simplest model is usually preferred. This is a principle usually attributed to William of Occam or Ockham. In practice, I defined the simplest model to be the one with the smallest possible number of recombination events, and with each mutation originating only once. The Documents The four main surviving versions of the Queen of Sheba story are in the (MT) of 1 Kings 10 and, and the corresponding Septuagint (LXX) versions of these chapters. I used the NASB translation of the MT and Brenton s translation of the LXX. queen_of_sheba_analysis.doc 20 April 2011 Page 1 of 8

2 The Differences In these four documents there are many variations that might have arisen from ambiguities, but I ignored these and identified 16 points of variation that seemed more likely to have arisen from transcription errors or editorial changes. These are the important variants for my purposes, because they provide good evidence of divergence from a common ancestor, and the number of shared variants provides a measure of how closely related any two documents are. Errors arising from ambiguity, such as the Hebrew word that can mean either stairway or whole burnt offering, are not so useful for this purpose, because it is quite likely that two translators could make exactly the same mistake independently, so documents that share such errors do not necessarily have any recent common ancestor. The sixteen points of variation that I have used are shown in the following table. Verse no. in Chron. Verse no. in Kings Details that vary 1 Kings 10 MT 1 Kings 10 LXX 2 Chron. 9 MT 2 Chron. 9 LXX 7 8 men or wives? men wives men men 8 9 establishing them/him no yes yes yes with her servants? yes yes yes no brought gold and silver to no no yes yes Solomon (repeated) shields or spears? Shields (O) Spears (I) Shields (O) Shields (O) shekels or 3 3 pounds/ 3 pounds/ 300 pounds/minas minas (S) minas (S) shekels (L) and the lavers were golden no yes no no 300 shekels (L) ivory yes no yes yes apes yes no yes yes peacocks yes no yes no stones no yes no no After After inserted passage * no yes no no v.21 v silver yes no yes yes myrrh/stacte no yes no yes how many horses/mares/chariots? 1,400 (S) 4,000 (L) 4,000 (L) 4,000 (L) gold no yes no yes * The inserted passage is This was the arrangement of the provision which king Solomon fetched to build the house of the Lord, and the house of the king, and the wall of Jerusalem, and the citadel; to fortify the city of David, and Assur, and Magdal, and Gazer, and Baethoron the upper, and Jethermath, and all the cities of the chariots, and all the cities of the horsemen, and the fortification of Solomon which he purposed to build in Jerusalem and in all the land, so that none of the people should rule over him that was left of the Chettite and the Amorite, and the Pherezite, and the Chananite, and the Evite, and the Jebusite, and the Gergesite, who were not of the children of Israel, their queen_of_sheba_analysis.doc 20 April 2011 Page 2 of 8

3 descendants who had been left with him in the land, whom the children of Israel could not utterly destroy; and Solomon made them tributaries until this day. But of the children of Israel Solomon made nothing; for they were the warriors, and his servants and rulers, and captains of the third order, and the captains of his chariots, and his horsemen. To make the raw data compact enough to fit into diagrams, I coded the characteristics of each document into a string of 16 letters, representing the 16 characteristics shown in the above table, in the same order as listed in the table. Most of the letters are the initial letters of words like yes, no, men, wives, etc, except for the codes O, I, L and S, which respectively represent shields, spears, large and small. Not knowing what words a professional textual critic would use, I shall use a biological metaphor and refer to these 16-letter strings as genotypes and the 16 components as genes. Thus the details of the above table become condensed as follows. Document Genotype MT MNYNOSNYYYNNYNSN LXX WYYNISYNNNYYNYLY MT MYYYOLNYYYNNYNLN LXX MYNYOLNYYNNNYYLY It would be satisfying if we could make no initial assumptions about which texts come from a close common ancestor. It would be better to derive the conclusions objectively from the data shown. So let us count the number of difference between all possible pairs of the documents, and show the counts in a table. The largest possible count is 16. Counts of Differences MT LXX MT LXX MT LXX MT LXX A Surprise At this point the first surprise raises its bizarre head. If I had been a professional textual critic this would not have been a surprise to me. The LXX version of is away out on a limb. It differs from all the other documents more than they differ from one another. Most surprising of all, its biggest disagreement (13 points out of 16) is with the MT version of, the very document that we would expect it to have the greatest affinity with. This phenomenon is found in other parts of 1 Kings, not only in chapter 10. Possible reasons for it are discussed by Emanuel Tov at Methods However, we can now begin to construct a family tree by joining the most closely matched pair of documents by a forked line, and inferring the most probable details of their hypothetical common ancestor. In this example we have to choose between two equally close matches: MT queen_of_sheba_analysis.doc 20 April 2011 Page 3 of 8

4 matches LXX quite closely, as we might expect, but it matches MT equally closely. To help to choose between these two possibilities, I considered the historical fact that Kings and Chronicles diverged from each other before the LXX versions of them were produced, so the divergence between the Hebrew and Greek versions of 2 Chronicles must have been more recent than the divergence of Kings and Chronicles. In constructing the tree we start from the present day and work back. Accordingly, the first step in constructing it is as shown below. Since we are assuming each mutation originated only once, we can infer most of the variants in the hypothetical common ancestor, as shown. Ancient MY?YOLNYY?NNY?L? LXX version MYNYOLNYYNNNYYLY MYYYOLNYYYNNYNLN We now have three pieces of jigsaw puzzle to fit together instead of four. We have the family tree shown above plus two versions of. The next step is to find which pair of these three are most closely related and join them along with their hypothetical common ancestor. Sometimes such a choice depends on the precise measure of relatedness that we use, but in this case the choice is clear. The most closely related pair of these three is the small family tree above and the Masoretic 1 Kings 10. Joining these together gives us a slightly bigger family tree as shown below. Intermediate M???O?NYYYNNYNLN Ancient MY?YOLNYY?NNY?L? LXX version MYNYOLNYYNNNYYLY MYYYOLNYYYNNYNLN MNYNOSNYYYNNYNSN Finally we add the least related surviving document, and postulate an ancestor common to all four, as shown below. Ancient???????????????? Intermediate M???O?NYY?NNY??? Ancient MY?YOLNYY?NNY?L? LXX Version WYYNISYNNNYYNYLY LXX version MYNYOLNYYNNNYYLY MYYYOLNYYYNNYNLN MNYNOSNYYYNNYNSN queen_of_sheba_analysis.doc 20 April 2011 Page 4 of 8

5 So far, we have identified some genes of hypothetical ancestors purely by considering the genes of their immediate descendants, but we can do better than this. The next step is to fill in more of the missing details in the ancestors by considering also their own ancestors. A different layout The logic involves a lot of details, and the data are confusing to the eye, so instead of drawing a family tree in the above style, a different way of setting out the systematic working is shown below, which makes it easier to align and compare the various genotypes (16-letter codes.) Let us recapitulate what we have done already, but put it in a different format. At first we infer the most probable genotype of each ancestor purely by looking at its descendants. If both its descendants share a particular gene, then the parent probably shared it too. LXX MYNYOLNYYNNNYYLY MT MYYYOLNYYYNNYNLN Ancestor of the 2 above (Ancient Chronicles) MY?YOLNYY?NNY?L? MT MNYNOSNYYYNNYNSN Ancestor of all the above (Intermediate Kings) M???O?NYY?NNY??? LXX WYYNISYNNNYYNYLY Ancestor of all the above (Ancient Kings)???????????????? Then we go back and fill in some more of the missing details for the hypothetical ancestors by looking also at their own ancestors. At this stage we try to minimize the total number of mutations, in accordance with the Principle of Parsimony and Occam s razor. For example, the third gene of Ancient Chronicles, intermediate Kings and Ancient Kings are all probably Y, because this hypothesis requires only a single mutation for that gene, which occurs in going from Ancient Chronicles to its LXX version. Any other hypotheses would require more than one mutation. For the same reason, the 4 th gene in Intermediate Kings and Ancient Kings is probably N. The one mutation then occurs in going from Intermediate Kings to Ancient Chronicles. Likewise, the 2 nd gene in Intermediate Kings and Ancient Kings is probably Y. The one mutation then occurs in going from Intermediate Kings to Kings MT. And the 5 th gene in Ancient Kings and Intermediate Kings is probably S, requiring only one mutation, which occurs in going from Intermediate Kings to Kings MT. The transmission of the second last gene can have occurred in only one way if we assume only one mutation. It must be L in both Ancient Kings and Intermediate Kings, with the only mutation occurring in going from Intermediate Kings to MT. Most of the genes in Ancient Kings remain indeterminate. Those in positions 1, 4, 6, 7, 8, 9, 11, 12, 13 could equally well be either of two variants, since either option would be compatible with a mutation originating only once for each gene. So we arrive at the following model. queen_of_sheba_analysis.doc 20 April 2011 Page 5 of 8

6 LXX MYNYOLNYYNNNYYLY MT MYYYOLNYYYNNYNLN Ancestor of the 2 above (Ancient Chronicles) MYYYOLNYY?NNY?L? MT MNYNOSNYYYNNYNSN Ancestor of all the above (Intermediate Kings) MYYNOSNYY?NNY?L? LXX WYYNISYNNNYYNYLY Ancestor of all the above (Ancient Kings)?YYN?S????????L? One Assumption Must Go The most difficult remaining problem concerns genes 10, 14 and 16. To fill in the missing details of these we have to allow either more than one mutation per gene, or else what biologists would call horizontal transfer of genes, i.e. transfer of information between two contemporaneous but basically different documents. Such horizontal transfer could occur, for example, if an editor or translator had access not only to the main story he was editing or translating, but also to another version of the same story. He might incorporate elements of one version into another in order to harmonize them. This seems not unlikely. We know of many occasions when such editing has taken place. The alternative explanation, that exactly the same mutation has occurred independently in two different documents, seems less likely. So to find an explanation of the descent of these three genes, let us look at the three by themselves LXX NYY MT YNN Ancestor of the 2 above (Ancient Chronicles)??? MT YNN Ancestor of all the above (Intermediate Kings)??? LXX NYY Ancestor of all the above (Ancient Kings)??? It is noticeable that the three problem genes in LXX are identical to those in LXX. It is equally noticeable that the three problem genes in MT are identical to those in MT. This suggests several possibilities. First, someone may have edited LXX to make it agree with LXX. If its genotype had originally been YNN then we could fill in the table as follows, with the need for only one mutation per gene. LXX NYY MT YNN Ancestor of the 2 above (Ancient Chronicles) YNN MT YNN Ancestor of all the above (Intermediate Kings) YNN LXX before harmonizing with Chronicles LXX YNN Ancestor of all the above (Ancient Kings) YNN queen_of_sheba_analysis.doc 20 April 2011 Page 6 of 8

7 Secondly, someone may have edited LXX to make it agree with LXX. If its genotype had originally been YNN then we could fill in the table as follows, with the need for only one mutation per gene. LXX before harmonization YNN MT YNN Ancestor of the 2 above (Ancient Chronicles) YNN MT YNN Ancestor of all the above (Intermediate Kings) YNN LXX NYY Ancestor of all the above (Ancient Kings)??? Thirdly, someone may have edited MT to make it agree with MT. If its genotype had originally been NYY then unfortunately there is no way of filling in the table with only one mutation per gene so this hypothesis is incapable of explaining the lineage of all three problem genes. Fourthly, someone may have edited MT to make it agree with MT. If its genotype had originally been NYY then we could fill in the table as follows, with the need for only one mutation per gene. LXX NYY MT YNN Ancestor of the 2 above (Ancient Chronicles) NYY MT before harmonization NYY Ancestor of all the above (Intermediate Kings) NYY LXX NYY Ancestor of all the above (Ancient Kings) NYY So there are three possible directions of horizontal gene transfer. Only one of them is needed, but more than one is possible. Let us now draw our hypothetical evolutionary tree and include arrows to show possible horizontal transfer of information from one document to another of a different lineage. The data we have looked at and the methods we have used do not tell us which of these horizontal transfers took place, but the observed similarities and differences between the surviving documents can best be explained if at least one of these horizontal transfers occurred. Results Ancient?YYN?S????????L? Intermediate MYYNOSNYY?NNY?L? Ancient MYYYOLNYY?NNY?L? LXX Version WYYNISYNNNYYNYLY LXX version MYNYOLNYYNNNYYLY MYYYOLNYYYNNYNLN MNYNOSNYYYNNYNSN queen_of_sheba_analysis.doc 20 April 2011 Page 7 of 8

8 So there we have it. It is not the only possible evolutionary tree, but it is probably the simplest one that is compatible with the facts and assumptions that we started with. It is entirely possible that the four final texts continued to evolve for some time after the hypothetical horizontal transfer, but in the interests of simplicity I have not shown this possibility in the diagram. The facts we have considered also tell us nothing about the long and obscure transmission of text from Ancient to the LXX version of. We don t know if most of the considerable changes were made before or after translation into Greek. Linguistic studies might answer this question. If the peculiarities of the LXX version contain many Hebraisms, this would suggest mainly pre-translation changes, but if the peculiarities contain more of the kinds of things a Greek writer would say, this would suggest mainly post-translation changes. This study and my previous one have each been limited to only one Bible story. I conducted these studies mainly to see whether the methods used in evolutionary biology could usefully be applied to the transmission of ancient texts. It would be interesting to conduct similar analyses of other stories from the same books. If these gave similar conclusions, it would suggest that the family tree shown above describes the history of the entire books, not just the history of a particular story. queen_of_sheba_analysis.doc 20 April 2011 Page 8 of 8