CHAPTER 3: the Method of Identifying High-Interest Parallels
3.2 The Steps Involved in the Method
3.2.5 Linking and Scoring
The Linking and Scoring step is performed on every segment of words from the target text in order to identify the parallels (or ‘links’ between the source and target) in each segment. Within the new method, this process has two components, as follows:
1. Instantiate the Rules
Make a copy of the appropriate ‘Master file’ and replace the words in the file (represented as ‘AAA’, ‘BBB’, ‘CCC’ and ‘DDD’, and ‘111’, ‘222’, ‘333’ and ‘444’ etc.) with all the lexical alternatives of the words in the current segment and the following segment.
2. Interpret the Rules
Copy the instantiated rules into Accordance in order to find ‘valid’ or ‘true’ rules. That is, rules that return a set of source texts that have ‘singular potential.’
As an example, the first segment of 1 Timothy contains four search words: Παῦλος (‘Paul’), ἀπόστολος (‘apostle’), Χριστός (‘Christ’) and Ἰησοῦς (‘Jesus’). The
311 Nugues, Language Processing with Perl and Prolog: Theories, Implementation,
following segment has six search words,312 so the Master4x6 file was chosen as a template. This file is listed in Appendix A.
The file contains six pages of rules of which over half are for structural parallels. There are only eleven rules that relate to non-verbatim clauses/phrases in a four-word segment (see Section 3.2.3). These rules are repeated below. The numbering on the left is for the subsequent explanation on the linking process.
1. (AAA) <within 5 words> (BBB)
2. (AAA) <within 5 words> (BBB) <within 5 words> (CCC)
3. (AAA) <within 5 words> (BBB) <within 5 words> (CCC) <within 5 words> (DDD) 4. (AAA) <within 5 words> (BBB) <within 5 words> (DDD)
5. (AAA) <within 5 words> (CCC)
6. (AAA) <within 5 words> (CCC) <within 5 words> (DDD) 7. (AAA) <within 5 words> (DDD)
8. (BBB) <within 5 words> (CCC)
9. (BBB) <within 5 words> (CCC) <within 5 words> (DDD) 10. (BBB) <within 5 words> (DDD)
11. (CCC) <within 5 words> (DDD)
Each of the target words in this template was then replaced with all the lexical alternatives of the actual words in the segment. This was done using Microsoft
Word’s ‘Advanced Search/Replace’ function.313 The result of this instantiation for the eleven rules is shown below. The result for the entire file is listed in Appendix B. 1. (Παῦλος) <within 5 words> (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή)
2. (Παῦλος) <within 5 words> (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Χριστός) 3. (Παῦλος) <within 5 words> (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Χριστός) <within 5 words> (Ἰησοῦς)
4. (Παῦλος) <within 5 words> (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Ἰησοῦς) 5. (Παῦλος) <within 5 words> (Χριστός)
6. (Παῦλος) <within 5 words> (Χριστός) <within 5 words> (Ἰησοῦς) 7. (Παῦλος) <within 5 words> (Ἰησοῦς)
8. (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Χριστός)
9. (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Χριστός) <within 5 words> (Ἰησοῦς) 10. (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Ἰησοῦς)
11. (Χριστός) <within 5 words> (Ἰησοῦς)
If Accordance had a web service interface, a computer program could send these searches to the appropriate function of the web service. However, since this service
312 A ‘search word’ is a word that is used in the Linking step of the method (see Section 3.2.3). The second segment of 1 Timothy consists of twelve words in total, as follows: κατ᾿ ἐπιταγὴν θεοῦ σωτῆρος ἡµῶν καὶ Χριστοῦ Ἰησοῦ τῆς ἐλπίδος ἡµῶν. In this study, prepositions, pronouns, conjunctions, and articles are not used in the searches, meaning that the second segment of 1 Timothy has six search words: ἐπιταγή, θεός, σωτήρ, Χριστός, Ἰησοῦς and ἐλπίς.
was not available, the searches were simply manually copied to the command line of the Accordance program.
Every source text returned by these eleven searches could potentially represent a textual reference, so an exhaustive method would perform detailed analysis on every such text. However, given that this analysis is labor-intensive, this is not practical for large numbers of parallels. One approach to solving this problem is to limit the
analysis to parallels that have the greatest number of matching words. If, for example, this limit was set to 3 matching words, this would be equivalent to only performing analysis on the parallels returned from the searches on lines 2, 3, 4, 6 and 9. This would be similar to setting the ‘number of ignored words’ parameter to 1 when using the INFER search command with a four-word segment (i.e. [INFER 4, 1 …]. See Section 3.3.3, below). However, based on the theory of allusions, this study selects parallels to analyze based on the criterion of potential singularity. This allows the method to efficiently identify parallels with low verbal similarity, like those that occur in the three baseline lists for the Pastoral Epistles (i.e. UBS5, NA28 and Evans).
To illustrate how this is done, the rules above have been indented, as shown below: 1. (Παῦλος) <within 5 words> (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή)
2. (Παῦλος) <within 5 words> (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Χριστός) 3. (Παῦλος) <within 5 words> (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Χριστός) <within 5 words> (Ἰησοῦς)
4. (Παῦλος) <within 5 words> (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Ἰησοῦς) 5. (Παῦλος) <within 5 words> (Χριστός)
6. (Παῦλος) <within 5 words> (Χριστός) <within 5 words> (Ἰησοῦς) 7. (Παῦλος) <within 5 words> (Ἰησοῦς)
8. (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Χριστός)
9. (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Χριστός) <within 5 words> (Ἰησοῦς) 10. (ἀπόστολος, ἀποστέλλω, πέµπω, ἀποστολή) <within 5 words> (Ἰησοῦς)
11. (Χριστός) <within 5 words> (Ἰησοῦς)
The method would begin with the first pair of words in the segment: Παῦλος and ἀπόστολος. If the first search on line 1 returns a set of parallels with singular potential (i.e. 9 or less source texts), then it sends those that are thematically coherent for analysis. Because the method has already found all the rare combinations of Παῦλος and ἀπόστολος, it does not need to execute the searches on lines 2-4 (which are indented in from line 1 in order to illustrate this). So, the method jumps down to the
second pair of words, Παῦλος and Χριστός, on line 5 (which is at the same indentation level as line 1 in order to illustrate this).314
However, if the result of the search on line 1 does not have singular potential (i.e. there are more than 9 source texts) then the method discounts these parallels and moves to the second search on line 2 that contains an additional word. If this search returns a set of parallels that has singular potential, then it sends those that are
thematically coherent for analysis and jumps down to line 4; otherwise it moves on to line 3, followed by lines 4 and 5.
Similarly, if the search on line 5 returns a set of texts that has singular potential, then it sends those that are thematically coherent for analysis and then jumps over line 6 to the third set of words on line 7. However, if the search on line 5 does not have
singular potential, the method discounts them and moves to line 6. This process is then continued throughout the whole set of rules, with only parallels that have singular potential and are also thematically coherent being passed to the labor- intensive ‘post-processing’ step that performs detailed analysis of these parallels.