Because at times the word “criticism” can mean “finding fault with,” it is important to note that when it is used here it means “evaluation,” the analysis of something with the intent of determining its value. The wording of the manuscripts of the Hebrew Bible and of the Greek New Testament varies here and there to a greater or lesser degree. It is necessary, therefore, to employ the criteria of textual criticism in order to evaluate the various readings so as to determine, if possible, the original author's text prior to the modifications that appear in extant manuscripts—for the original autographs were lost long ago. There are three classes of sources that scholars use in textual criticism of biblical texts: the Hebrew or Greek manuscripts; ancient translations in other languages; and quotations made by rabbis and church fathers.
The first step in the determination of the original text involves a scrupulous comparison of all the witnesses (or, at least, the important witnesses) in these three classes of texts, and then producing a compilation of the differing readings. Such a compilation is known as a critical apparatus. At times this process sheds light on how and why a scribe introduced a textual variation. The majority of differing readings occurred because of unintentional error; in other instances, the text may have been intentionally altered. Accidental variations can result from one letter being mistaken for another; from the reversal of the sequence of two letters (metathesis); from exchanging letters and words that sound similar; from confusing two successive lines that begin with the same letters or words (homoeoarchton) or that end with the same letters or word (homoeoteleuton), by allowing the eye to skip from the first to the second line (parablepsis), thus omitting the intermediate text (haplography); and from the eye accidently processing the same word or groups of words twice so that the scribe writes for a second time a text that was meant to be read only once (dittography).
Oddly enough, scribes who thought about the text were more likely to make emendations than those who simply wanted to produce an accurate copy. Deliberate changes include correcting spelling and grammar; conforming a reading to a parallel passage; expanding or polishing the text by adding a familiar word or phrase where one seemed to be called for; combining similar phrases; clarifying historical and geographical problems; substituting synonymous words or expressions; and modifying or deleting expressions considered objectionable by the scribe.
The textual critic's fundamental considerations when assessing variant readings involve both external and internal evidence. External evidence relates to the date of the witnesses, the geographical distribution of the witnesses that agree, and the family relationship (if determinable) of manuscripts and groups of witnesses. Internal evidence is concerned with transcriptional probabilities, which require analysis of paleographical details and the scribe's habits, and intrinsic probabilities, which necessitate examination of the author's style and vocabulary throughout the book.
The differing conclusions of textual critics can usually be traced to one's judgment as to which criteria are deemed most significant. For example, for the Hebrew Bible most scholars use the Masoretic text as a point of departure for textual criticism because it is a complete, established text that was scrupulously transcribed. In some cases, however, readings in the Qumran Hebrew manuscripts are considered superior to the Masoretic text by virtue of their agreement with ancient translations. The Qumran manuscripts, however, are not complete, and some were negligently copied.
As a general rule, the more difficult reading is usually to be preferred, as is also the less smooth or unassimilated reading—since in both instances scribes resisted the urge to produce a more polished, harmonious text. The shorter reading is also favored by the majority of textual scholars (unless specific omissions can be traced to homoeoteleuton, or unless the shorter reading does not conform to the character, style, or scope of the author), since scribes tended to supplement the text with explanations or material from parallel passages rather than to abridge it. Simply stated, the reading that best explains the origin of the other readings should be preferred as the original.
Bruce M. Metzger