יום שלישי, 11 בנובמבר 2014

Computer-assisted Tools for Textual Criticism


Prof. E. Tov, The Hebrew University[i]
Prof. E. Tov
Since the last quarter of the 20th century, the study of textual criticism has been aided greatly by computer-assisted tools and research. Such tools consist of flexible, interactive databases and programs that aid the researcher in obtaining and analyzing data, while computer-assisted research presents in print non-flexible[ii] results of investigations that were compiled with the aid of machine-readable data. The latter type of research will not be referred to in this chapter. It is probably true to say that involvement in textual criticism is virtually impossible in the 21st century without the aid of electronic tools. An ever-increasing number of Bible computer modules and databases are becoming available and the possibilities for using them profitably within existing programs are expanding. This increasing availability enables several types of data retrieval, and allows scholars to access data and formulate conclusions that would not have been possible with the conventional research methods. The present chapter briefly describes the available data, but it should be remembered that due to ongoing developments some statements may be in need of updating.
a. Available Tools
The great majority of the text modules also include a morphological analysis.[iii] This analysis allows for searches of all the words, combinations of words, and grammatical categories, and also enables the creation of concordances.[iv] At this stage, the searches can be applied to the following texts:
MT (according to codices L or A, without medieval variants), allowing for searches of words or meaningful parts of them, vocalization patterns, open and closed sections, Ketib–Qere forms, accents, and the MM and MP.[v]
Sam. Pent. (edition of Tal, Shekhem 1994, without variants). Accordance enables the automatic indication of all divergences from MT, and searches of all words and grammatical categories in the Sam. Pent.
Biblical Dead Sea scrolls, or any group of them such as the pre-Samaritan texts, with indication of the divergences from MT,[vi] enabling searches of all words and grammatical categories.
LXX (editions of Rahlfs, Septuaginta, 1935; Rahlfs–Hanhart, Septuaginta,  2006; Göttingen editions).[vii] In addition, the CATSS module of LXX indicates divergences from MT, reconstructs details in the Vorlage of LXX, and marks select features in categories in translation technique.[viii] The LXX module allows for inner-translational searches, while the CATSS module also allows for searches of the LXX – MT equivalents.
T S V (usually without variants, while V in SESB includes variants), allowing for inner-translational searches.
Apparatus of BHS: fully searchable.
Since the wish to retrieve readily available information is constantly growing, it should be stressed that in 2013 the data available in machine-readable format are limited, only allowing for some types of computer-assisted research. Most research requires combinations of electronic and manual study of the data, depending on the topic. The machine-readable form of the BHS apparatus is unsatisfactory for text-critical analysis since it provides far too little information and is much too subjective. The equivalent tool for BHQ, when available in full, is more helpful, since that edition encompasses more data. However, both tools contain merely a selection of textual data, and therefore are not substitutes for other types of research, among them manual research. A machine-readable form of the HUB apparatus, if and when available, would come closer to the needs of the researcher, but that edition also provides too little information.
As a result, there is no substitute for the manual use of the data in the source languages, with the supplement and aid of machine-readable modules. In 2013, only two of these modules include variant readings (for the LXX, see n. 6), making the manual use of the editions of the versions a necessary supplement to electronic searches. Furthermore, use of the modules would be more efficient if they included a reconstruction, however tentative, of the Vorlage of the ancient versions when they supposedly differ from MT, providing the researcher with additional research possibilities. In 2013, such facilities are available (partially) only for the CATSS module of LXX.[ix]
Subjectivity and Caution. Machine-readable modules should be used cautiously, which applies as well to the use of any printed source. Since modules are compiled manually, we must remember that a machine-readable text, having been encoded by humans, may contain mistakes.[x]
The evaluation of variant readings may be based partially on data that are collected electronically, but ultimately the evaluating process of these variants remains subjective. Use of electronic data does not replace judgment.[xi]
b. Select Categories of Information
Textual criticism involves the comparison of all details in the texts and therefore necessarily deals with all types of information that may be extracted from them. This includes the areas of language, theology, geography, history, chronology, mythology, and all literary genres (wisdom, psalmody, prophecy, laments, narratives, etc.), and all other categories of information contained in texts. Variants in all these areas may be equally as relevant for these disciplines as the central text (for example, the LXX).
Not all types of information can be retrieved using electronic tools; in fact, in the long list of areas mentioned in the previous paragraph, only a very small percentage of the variants can be researched en bloc. For example, the reason that variants in the areas of history, geography, or prophecy cannot be researched is because the details have not been marked as “historical,” “geographical,” etc. in the databases. Users can indicate historical books and sections as their “search range,” but their content is not necessarily “historical.” In order to search for historical details, they first need to be indicated manually.
As a result, most, but not all, information that we wish to retrieve from the computer modules needs first to be encoded. However, also the computer programs can help us in retrieving some information that has not been expressly indicated. Thus, we can find vocabulary differences between witnesses in the same language, enabling the user to focus on the differences between MT, the Sam. Pent., and the Judean Desert scrolls, the various Greek translations, modern translations of MT and the LXX,[xii] the various Targumim, etc. Other types of information that have not been inserted are the number of letters, words, and verses, the distribution of words, and word combinations in Hebrew Scripture as a whole, in a book, or chapter. Among other things, the computer can determine the most frequent or infrequent words or Greek-Hebrew equivalents in a unit, patterns in the use of the divine names, etc.
With the exception of the types of information mentioned in the previous paragraph, the user cannot retrieve information from computer modules that has not been encoded.[xiii] An indication of any type of detail (“predetermined information”) would significantly expand the search and research facilities in the post-modern world. In my view, commentators on all biblical books are unable to obtain the necessary data in the textual witnesses in many areas, making predetermined information a desideratum for the writing of commentaries.
c. Predetermined Information
In the future, hopefully the biblical modules will include several types of predetermined information. In 2013, the major sets of analysis that have been pre-determined are limited, covering only the morphological and syntactical analysis of the main texts and some features of LXX:
a. Morphological and syntactic analysis of MT;[xiv]
b. Morphological analysis of the Sam. Pent., the Judean Desert scrolls, LXX, T, S, V;
c. Indication of some features of the translation technique of the LXX and the relation between the LXX and MT in the CATSS module.
It would be advantageous for the development of research if many additional sets of information such as those mentioned in section b were to be incorporated into the existing databases. For example, historical data may be added to extant databases such as CATSS or may be included in a new database based on MT,[xv] enabling their automatic retrieval. By the same token, it would be advantageous if all possible theological variations between textual sources were to be encoded. Undoubtedly, the very recording of such variations is subjective, but it could inspire continued research. For example, MT probably contains more examples of theological changes in Samuel than in other books, but this impression needs to be substantiated statistically.
Two areas are singled out for special attention:
 Textual analysis. In the area of textual criticism, merely a beginning has been made in computer-assisted research. Future databases should include textual variations and some categories of notes on the translation technique of the ancient versions similar to the notations in CATSS, as well as textual phenomena, such as harmonizations, pluses and minuses, exclusive inner-translational agreements, and several types of information such as described in section b.
The computer-assisted analysis of the interchanges of consonants between MT and the LXX shows some distribution patterns in the Scripture books. It also shows the letters that are most frequently interchanged.
 Linguistic analysis. The area that is best covered in the computer modules is that of morphological analysis. This analysis enables the search for all words on the basis of the headword of each text word.



[i] This introduction to the list of computerized tools updates chapter 10 of my book Textual Criticism of the Hebrew Bible (3rd ed., revised and expanded; Minneapolis: Fortress Press, 2012).
[ii] This term refers to the results of research that was performed with the aid of computers, while the computer files or computer applications are not accessible to the researcher, and as a result the user cannot work with the data themselves.
[iii] The key to the effective use of any software program of Scripture texts is the availability in the background of lemmatization and morphological analyses (grammatical tagging) of all the text words. For example, the “lemmas” or “headwords” of rmayw are w (particle, conjunction) and rmay (verb, Qal, waw consecutive, 3rd person masc. sing.). The morphological analysis is mentioned in parenthesis.
[iv] In the example listed in n. 3, each of the morphological features is searchable within the text module, such as all Qal forms of this verb or all other verbs in the Torah, all the 3rd person masc. sing. forms of all hiphil verbs in 2 Kings, etc. In this way, all the types of nouns, verbs, prepositions and conjunctions are searchable in any search range (the Bible, a certain book, or part of a book).
[v] These features are not all searchable in the same program.
[vi] This feature is available only in Accordance.
[vii] The module of the Rahlfs–Hanhart edition in SESB includes the variants of that edition (“Handausgabe”), while the Göttingen edition in the Accordance and Logos programs includes the full apparatus of the complete edition (“editio maior”).
[viii] For example, differences between the LXX and MT in number, person, addition/omission of lk. It also includes notes on transliterations and doublets in ©, interchanges of consonants between MT and the LXX, and the relation between the LXX and KetibQere.
[ix] In the world of tomorrow, we hope to see sophisticated modules of all the ancient versions and an improved version of CATSS. Further, we would like to see interconnected modules of Hebrew manuscripts and the ancient versions. However, even then the researcher would need to resort to several printed sources as supplements to the computer modules. In future research, the most efficient first step would be the use of electronic data, allowing the researcher to access more data than was previously available. After all, in the study of textual criticism we need access to a multitude of details, and while judgment remains the major component in the study of texts, we can only hope to ever access all the data with the aid of electronic databases.
[x] Furthermore, text modules involve editorial judgment in the choice of, for example, a printed edition or manuscript, recording of Ketib–Qere, and the indication of verses, etc. This subjective element comes to light especially in the recording and analysis of the Judean Desert scrolls because of their fragmentary nature. The morphological analysis of all the texts and the determining of the “headwords” are also subjective and may be incorrect. As a result, incorrect decisions limit the usefulness of searches. For example, a search for all the participles in a text module will be defective if a certain participle is parsed as a noun or vice versa. Likewise, the recording of the reconstructed parent text of © in col. b of the CATSS module is subjective, and an examination of a Hebrew word will provide incomplete results if a researcher expects to find a certain reconstruction in that column that has not been accepted by the editors of CATSS.
[xi] Evaluations are based on the raw material such as that included in concordances, dictionaries, and electronic tools. Other evaluations are based on processed data provided by electronic tools, such as reconstructed readings, which contain a greater percentage of subjective information than the readings of the ancient sources themselves.
[xii] L.L. Brenton, The Septuagint Version of the Old Testament with an English Translation (London: Samuel Bagster and Sons, 1879); NETS.
[xiii] Thus, the computer will not list, for example, the names of the kings of Israel and Judah, the number of such kings, the names of the sons of Jacob, the cities in the tribal area of Asher, a list of the prophets and false prophets, a comparison of the prophecies against the foreign nations by Isaiah, Jeremiah, and Ezekiel, etc.
[xiv] The morphological analyses are listed in the attached list; for syntactic analyses, see SESB and the Andersen–Forbes database (both available in Logos).
[xv] For example, the recording of geographical variations would involve the many variations between the sources in the tribal lists in Joshua 15–19. The recording would necessarily involve forming judgment on the data. For example, one would have to decide which of the many Greek variants should be included and excluded. One would also have to decide on categories of notes on matters of content, for example, alternative forms of the same names (Timnath-Heres in Judg 2:9 and Timnath-Serah in Josh 19:50, 24:30). One should also decide on a special notation for place names listed in one source as belonging to one tribe and in another source as belonging to a different tribe. E.g. Jerusalem is connected with both the tribe of Benjamin (Judg 1:21) and the tribe of Judah (Josh 15:63; Judg 1:8). An important type of notation pertains to variations in place names, such as 1 Sam 9:4, where “the district of Shaalim” is reflected in LXXLuc as “Gaddi of the town Segaleim (= Shaalim).”

See: Electronic Tools for the Textual Criticism of the Hebrew Bible - The List

אין תגובות:

הוסף רשומת תגובה