A New Numbering System for Greek New Testament Lexemes James Tauber jtauber@jtauber.com Ulrik Sandborg-Petersen ulrikp@scripturesys.com Abstract Numbering systems (such as Strong’s) are a popular way to reference the lex- emes of the Greek New Testament cor- pus but a straight enumeration is not without problems, particularly when there is disagreement about whether two forms are the same lexeme or not. We present a way of referencing lexemes that allows competing view- points to be represented simultane- ously. Existing numbering systems can be mapped into this new system with- out any loss of granularity and new analyses can be expressed without vi- olating the integrity of existing refer- ences into the system. 1 Introduction In the late 19th century, a concordance of the King James Bible was produced under the di- rection of James Strong (Strong, 1890). This concordance provided a comprehensive cross- reference of every word of the King James text back to the corresponding words in the origi- nal texts. A dictionary was then included that provided a glossary for the various lemmas in Hebrew, Aramaic and Greek. The cross- referencing was achieved by assigning each lemma a number. While there has been considerable criticism made of the dictionary itself over the last cen- tury, there is no doubt that the numbering sys- tem he devised for that dictionary has proved useful independent of the definitions them- selves. The mapping of numbers to lexemes pro- vides a means of referencing lexemes in a way that is unambiguous with regard to homo- graphs (distinct words of identical spelling). It also keeps users of the system isolated from different choices of lemma—for example, how to handle deponency (Taylor, 2004) or whether to use the first person present active indicative of a verb or, say, the aorist infinitive. 1 Deci- sions like this can be made without affecting the integrity of analyses that refer to lexemes by number. Limitations of Strong’s numbers, however, have long been recognized and there have been both attempts to improve Strong’s num- bering and to develop alternative numbering systems (Goodrick and Kohlenberger, 1990). These limitations (and hence the nature of their correction) generally amount to errors of omission or disagreement over where to draw the line between certain lexemes (see Section 2 for examples). However, in all cases, the improvements or alternatives remain simple enumerations of a set of lexemes. Any decision as to whether to distinguish two lexemes or conflate them into one is locked into the numbering system itself. A disagreement in how to slice the lex- icon up results in competing numbering sys- tems which prevents external databases and analyses based on one to be integrated with those based on another. What is needed is a manner by which al- ternative viewpoints can be represented at the same time in a single data structure that can additionally be integrated with existing num- 1 While the traditional approach has been to identify verb lexemes by the former, there have been strong arguments made in favour of the latter (Buth, 2004; Taylor, 2004).