was successfully added to your cart.


Textual Evidence

Learning the Language, Part 3: Canons of Criticism

By | Biblical Criticism | No Comments

In the first two parts (see part 1 & 2) of this article, I discussed and defined many of the key terms and concepts involved in Textual Criticism. Now it remains to examine something of how these work in practice.

To begin, it is worth taking a look at the origins of Textual Criticism:

In the early Church, we have very little evidence of this, although clearly manuscripts of the New Testament were continuously copied and collected—some evaluation of the accuracy of a manuscript must have been involved in establishing an “exemplar” for copying. Early Church Fathers occasionally mention variant readings of the NT in their writings. But it is only with Origen that we see anything like a critical concern for the text of Scripture. Alexandria (the site of Origen’s early career), in particular, had a strong tradition of scholarship and scribal practice, going back into the Hellenistic period. However, even Origen does not address variants and other textual issues as carefully or consistently as one might expect. The only other Church Father (we know of) who came close to Origen’s level of technical scholarship was Jerome.

In the Middle Ages, serious Criticism of any sort was largely absent in the Church. The situation was a little better in the Greek (Byzantine) East; while Jewish and Islamic scholars, much further ahead at the time, helped to preserve a certain level of scholarship for future generations in Western Europe. The Rabbinic scribes, in particular (the Masoretes), had established a traditional practice of evaluating and addressing variants and textual corruption in the Hebrew manuscripts of the Old Testament. However, Textual criticism, in something like the modern sense, only began following the Renaissance. After centuries of relying upon translations, scholars such as Lorenzo Valla, Jacques Lefevre d’Etaples, and Desiderius Erasmus began looking afresh at the text, publishing critical studies and editions. The hunt for manuscripts was also on, though at the time of Erasmus’ critical edition of the Greek NT, relatively few MSS were available for study.

The modern science (or art) of Textual Criticism begins in earnest in the 18th and 19th centuries, following many decades of adopting Critical methods and principles in the study of Scripture. The earlier Critical editions of the 16th and 17th centuries (the so-called “Textus Receptus”) no longer sufficed. Men such as Gerhard von Mastricht, Richard Bentley, Johann Albrecht Bengel, Johann Jakob Wettstein, Johann Jakob Griesbach, Karl Lachman, Samuel Tregelles, and Constantin von Tischendorf, all edited and published Critical editions of the Greek NT, sometimes prefacing these with detailed prolegomena or critical studies that included principles for establishing the text. Griesbach, in his second edition (1796-1806), was perhaps the first to enunciate clearly a set of rules, to be adopted and modified by future scholars. (See below on “Canons of Criticism”).

In 1881-2 came the publication by Brooke Foss Westcott and Fenton J. A. Hort of their Critical edition (The New Testament in the Original Greek), and here, for the first time, was a decisive break with the Textus Receptus tradition. This proved to be momentous (and the subject of no small controversy!) for all subsequent Textual Criticism. In devaluing the Koine/Byzantine Text, they gave clear priority to the ‘earlier’ Text-types—particularly the Alexandrian Text, as represented by Codex Vaticanus (B), which they referred to as their “Neutral” Text. While future scholars were to abandon some of their premises and terminology, Westcott and Hort’s evaluation of Vaticanus and the Alexandrian Text has generally been confirmed—”a very pure line of very ancient text”.

CRITICAL EDITION: It is worth defining this here: Any Edition (see) which has been produced according to the methods and principles of Textual Criticism. That is to say, rather than simply publishing a single manuscript, multiple manuscripts are collated and analyzed, usually with the purpose of establishing (the most likely) form of the original text. Today the UBS (United Bible Societies) and Nestle-Aland critical editions are the most popular and widely used; the two editions are largely identical.

However, the question must be asked: How exactly does one determine the most likely original text?


A “canon” (Greek kanw/n, kanœ¡n) is a rule or standard, a regulating formulation or principle. So, the “Canons of Criticism” are simply the rules or guiding principles used in (Textual) Criticism. Sometimes these are referred to as “Canons of Authenticity” or “Criteria for Authenticity”—that is, for attempting to determine the authentic (original) text. There are many different rules or principles that have been established or adopted over the years, but the most important of these, I think, can be reduced down to the following set of fundamental rules—an even half-dozen (to which, of course, there are exceptions!):

  1. (Preliminary:) Negligible or insignificant variants can be discounted. This includes all nonsense readings, obvious scribal errors, and true singular readings. Only substantive/significant variants are to be considered (See Part 2 on these)
  2. The earliest and “best” manuscripts are to be preferred. Readings which occur in the earliest manuscripts—i.e., the 2nd-4th century Papyri, and the 4th-6th century uncials (such as Codex Vaticanus [B] and Sinaiticus [a])—are to be given priority. Determining the “best” manuscripts is a bit more difficult: to be considered are both (1) individual “purity”, the accuracy of the copy (number of scribal errors); and (2) relative “purity”, the relation to other accurate/reliable manuscripts. Most scholars today tend to give priority to the “best” representatives of the so-called Alexandrian text (especially Vaticanus and the related early Papyri).
  3. Readings in early and (geographically) diverse witnesses are to be preferred. For example, if a reading occurs in early Alexandrian and “Western” manuscripts as well as (relatively) early versional (i.e., Old Latin, Syriac, Coptic, etc.) manuscripts, it is likely to be original.
  4. The more difficult reading is to be preferred. This was most famously stated by J. J. Griesbach by the Latin proclivi scriptioni praestat ardua, but usually formulated as the maxim difficilior lectio potior. It can be broken down under several categories:
    * The more obscure reading vs. the more “intelligible”
    * The rougher (harsher or ungrammatical) reading vs. the “smoother”
    * The more unusual expression or vocabulary vs. the more customary
    * The reading which more conforms to ecclesiastical or doctrinal norms vs. the more challenging
    The tacit assumption being that scribes are more likely, in the process of copying (which involves some degree of interpretation), to smooth over rough, difficult or challenging passages. Such changes, usually intentional, were typically made simply for the purpose of explaining or clarifying the text (not for mischievous or malicious reasons!)
  5. (Corollary:) The shorter reading is typically to be preferred. This is the maxim brevior lectio potior. Again, the assumption is that scribes were more likely to expand the text, than to shorten it. Some have questioned this assumption, but analysis of the manuscripts and ancient scribal practice seems to bear it out. This is one of the more difficult areas for evaluation, especially when manuscript evidence is evenly divided between a longer and shorter reading—which is original? Is the longer reading an expansion/interpolation, or is the shorter reading the result of a scribal mistake (parablepsis, homoioteleuton)?
  6. The reading is to be preferred which best explains the rise of all others. Griesbach worded it as the reading that “…lies midway between the others”. A most useful rule (often decisive) when the manuscript evidence is evenly balanced; unfortunately, it can be most difficult to determine which is, in fact, the “midway” reading. Here theories regarding scribal practice and transcriptional probability (see below) need to be examined most closely.

* TRANSCRIPTIONAL PROBABILITY: This refers to the probable (or likely) process by which variant readings arose in the copying (transcription) of the text. Arguments and theories in this area can be complicated, and subject to considerable dispute among scholars. However, a careful analysis of transcriptional probability is often necessary in evaluating variants where the manuscript evidence is evenly divided.


The citing of Textual Witnesses is central to the science of Textual Criticism, but it can be a complicated and laborious process. Fortunately, professional scholars have already done most of the legwork in collating manuscripts, documenting and evaluating variant readings. The student is likely to encounter this information at first in two places:

1) In Textual Commentaries or Commentaries which contain specific textual notes.

2) In the “apparatus” (footnotes and appendices) to critical editions, such as the UBS or Nestle-Aland Greek New Testament (or the BHS Hebrew Old Testament).

Textual witnesses are cited according to fixed conventions, which nearly all scholars have adopted. To avoid cumbersome citation of manuscripts, etc., a technical shorthand was developed, using a standard set of abbreviations, referenced in the following order (for the New Testament):

  1. The early Papyri: Indicated by capital P, usually in script (Ë), followed by the number (according to a standard list),  typically in superscript—for example, Ë45, Ë 46, Ë 47 (Chesty Beatty Papyri); Ë66, Ë75 (Bodmer Papyri)
  2. Greek Majusucules (Uncials): These are numbered (standard list), prefixed with “0” (to distinguish them from the Minuscules)—i.e., 01, 02, 03, etc. However, up to manuscript 045, it is more common to use a letter to indicate the MS. Hebrew aleph (a), followed by English A-Z, followed by select Greek letters up through 045. Here are some of the most commonly cited manuscripts in the list:
    a [01] — Codex Sinaiticus
    A [02] — Codex Alexandrinus
    B [03] — Codex Vaticanus
    C [04] — Codex Ephraem Syri Rescripti
    D [05] — Codex Bezae (Codex Claromontanus [06] also indicated by D)
    W [032] — Codex Freerianus
    Q [038] — Codex Coridenthianus
    Y [044] — Codex Athous Lavrensis
    (The name of the codex is, by convention, a latinized version related to its provenance—where it was discovered or where it is being housed)
  3. Greek Minuscules: These are simply referred to by number (1, 2, 3, etc.) according to a standard list. f1 (“Family 1”) and f13 (“Family 13”) are commonly cited groups of related minuscules.
  4. The broader Greek MS evidence: Following the citation of specific MSS, the following abbreviations are sometimes used to indicate further support of a reading:
    Byz — the majority of (Byzantine) manuscripts (generally synonymous with the Majority text)
    Koine — also synonymous with the majority (Byzantine) reading, commonly indicated by K in decorated script (Š)
    ª — M in script, to indicate the Majority text; this has somewhat replaced Koine/Š.
    pc — (pauci), a few manuscripts
    al — (alii), some manuscripts
    pm — (permulti), a large number of manuscripts
    rell — (reliqui), the rest of the manuscript tradition (includes the Majority text)
  5. Versional Manuscripts: that is, the important early translations.
    Latin Versions
    it — (Itala), the Old Latin versions (sometimes also indicated by “OL”);
    individual manuscripts are represented by lower case letters and abbreviations (a, b, d, ff, gig, etc.)
    vg — the Latin Vulgate
    lat — part of the wider Latin MS tradition (including the vulgate)
    latt — the entire Latin MS tradition (possibly with exceptions)
    Syriac Versions — indicated by “sy” or “syr”; individual Syriac versions are indicated by superscript, or with separate abbreviations:
    s (or c) = Old Syriac [Sinai or Curetonian] p = Peshitta; ph = Philoxenian; h = Harklean
    Coptic Versions — indicated by “co” or “copt”; individual (dialect) versions are indicated by superscript, or, more often, with separate abbreviations, such as:
    ac (or ach) = Akhmimic
    bo (or boh) = Bohairic
    sa (or sah) = Sahidic
    Other Versions — aeth (or eth) = Ethiopic; arm = Armenian; geo = Georgian; arab = Arabic; goth = Gothic; slav = Old Church Slavonic.
    Typically Latin witnesses are cited first, then Syriac, Coptic, and so forth, generally as indicated above.
  6. Church Fathers (as textual witness): These leading figures and authors of the early and early-medieval Church, who cite/quote Scripture in their writings, are indicated (along with a few anonymous works), by a set of abbreviations. For example, Or = Origen, Aug = Augustine, Chr (or Chrysost) = Chrysostom, Cyr (or CyrAl) = Cyril of Alexandria, and so forth.
  7. Lectionaries: Usually indicated by a cursive “l” (Û) followed by a number, again from a standard list. Lectionaries are only rarely cited.

In addition, there are some superscript notations which may be applied (primarily with the Papyri and Uncials):

* — an asterisk indicates the original hand of a manuscript (when it has been corrected)
c — when the manuscript has been corrected
1, 2, 3, — a first, second, third, etc., hand which has corrected a manuscript
mg — textual evidence in the margin of a manuscript
v.l. — (varia lectio) when a variant reading has been recorded as such in a manuscript
txt — the text of the manuscript when a variant reading (above) has also been recorded
vid — (ut videtur) = apparent support for a reading when the condition of the MS makes it impossible to be certain how it reads
supp (or s) — when a portion of the manuscript is missing, and the text has been supplied by a later hand

I will now provide a couple of specific examples, to demonstrate how these textual witnesses are cited in practice:

Example 1: John 1:18monogenh$ qeo$ (monogen¢s theos) vs. monogenh$ ui(o$ (monogen¢s huios)
The textual evidence is fairly evenly balanced, and the question remains disputed among scholars, though the majority would accept the reading qeo$ (God) rather than ui(o$ (Son). The diagrams below show the evidence as it is cited in the critical apparatus of the UBS and Nestle-Aland (27th ed.) Greek texts (with some modification and simplification); first the evidence for qeo$, then that for ui(o$:


Example 2: Luke 23:34a — addition (interpolation) vs. omission
Here again the evidence is divided, between witnesses which include Jesus’ famous prayer (“Father, forgive them, for they know not what they do”) and those which do not. The diagram first indicates the witnesses which either do not have this portion (or have it marked with asterisks); then those which include it:



The information in this three-part article relates specifically to New Testament Textual Criticism; however, many of the terms and concepts apply to analysis of the text of the Old Testament as well. By comparison to the NT, Old Testament Textual Criticism is still, one might say, in its infancy, having really begun in earnest only in recent decades (with publication of all the Dead Sea Scrolls). In a future post, I hope to explore some of the issues relating to the Text of the Old Testament.

Learning the Language, Part 2: The Maze of Textual Variants

By | Biblical Criticism | No Comments

Establishing the Text (Continued):


In the prior article, the terms “Textual Variant” (or “Variant Reading”) and “Variation Unit” were defined. However, it is especially important to distinguish between different sorts of variants. The New Testament MSS contain thousands of variants—how are these to be analyzed? Are some more important than others? To begin with, there are three primary types of variants:
1. Add/omit: when a word or phrase is either added (or deleted) in the process of copying.
2. Substitution: when a word (or specific form of the word) is ‘substituted’ for another during copying. Instances of substitution involving an entire phrase or sentence are extremely rare.
3. Changes involving word order. To which I might add a fourth type of variant:
4. Conflation: when two different variants are combined together or side by side in the same MS.
These definitions, of course, assume that there was an original text, and that (all) variants are the result of textual corruption (see part one for a definition and explanation of this term). Whether, or to what extent, this original text is recoverable today, is a matter of some debate — see below under “Methods of Textual Criticism”

In order to sift through the enormous number of variants, it is common practice (indeed, a practical necessity) to distinguish between:

1) Substantive (or significant) variants, and
2) Negligible (or insignificant) variants

I prefer the term “substantive” to “significant”, as variants may have “significance”, but be essentially irrelevant for the purposes of establishing the text. It is easier to determine substantive variants by simply eliminating all negligible ones first. Negligible variants include:
a) Nonsense readings
b) Obvious scribal mistakes
c) Singular readings
The first two sorts are accidental, the third may or may not be. Scribal errors will be discussed separately below. “Nonsense readings” would seem to be self-explanatory, almost always the result of simple carelessness. However, the third needs to be defined further:
• Singular reading: Any reading which occurs in a single (one) Manuscript; the assumption being, that a reading which occurs in but a single MS would almost certainly not be original. Some critics would limit this definition to a single Greek Manuscript for the NT; this seems to me unnecessarily restrictive, although, in general, the proposition is valid. Occasionally a reading may be found in just one Greek MS, but have significant versional support: this occurs especially in the “Western” text (i.e., Codex D + Old Latin/Syriac versions); in such instances, I would still regard these as substantive variants (though rarely, if ever, representing the original text). The following terms are also used at times:
• Subsingular reading: This generally refers to a variant found in just two (or, perhaps several) otherwise unrelated MSS.
• “Nearly singular” reading: Similar to subsingular: any variant reading which occurs in just a few manuscripts.

Once negligible readings are discounted and/or eliminated from consideration, only substantive readings remain. It is these that are the basis of Textual Criticism—both in terms of defining manuscript relationships and of establishing the (original) text. However, it is at least worth outlining one further distinction, namely between variants which are (or may be):
a) Accidental, or
b) Intentional/Purposeful
I distinguish between “intentional” and “purposeful”, as a variant reading may be purposeful, but not necessarily the result of (conscious) intent by the scribe. Certain harmonizing additions, conflations, pious substitutions, and the like may often be done unconsciously. Here I define and provide some examples:

 Accidental Variants (= “Scribal Errors”):
These are the more common; for an exhaustive list, consult the standard textbooks.
Dittography: repeating a letter, word or phrase
Haplography: skipping over of a letter, word or phrase, resulting in an omission. Sometimes this is called “Parablepsis”.
Homoioarchton: skipping over a portion of text, from one word to another which has a similar beginning.
Homoioteleuton: skipping over a portion of text, from one word to another which has a similar ending.
Transposition: reversing two letters or words. Occasionally more than two words may be involved.
Substitution: accidentally substituting one letter or word for another, usually a mistake due to a trick of the eye.

 Purposeful Variants (possibly accidental):
Substitution and/or conflation: specifically, in the New Testament, with regard to the (sacred) names and titles, i.e., “God”, “Lord”, “Jesus”, “Christ”, “Son”, et al. It is often hard to tell when, or if, these are intentional, especially since it was visually easy to confuse them (almost always presented as two-letter abbreviations); on the other hand, they were often infused with tremendous theological import. I will discuss these nomina sacra (“sacred names”) in a later article.
Harmonization: that is, modifying the text to match another similar (and familiar) passage elsewhere in Scripture. In the NT, these are of two main kinds:
a) Gospel Harmonization: The manuscripts are rife with these sorts of changes, especially in the Synoptic Gospels (Matthew-Mark-Luke); with so many similar narrative passages, sayings of Jesus, and the like, these could easily be blended together. It is often hard to tell for certain when they are intentional. A large part of textual criticism in the Gospels involves these sorts of variants.
b) Harmonization to the LXX: Quotations of the Old Testament in the Greek NT, are quite often modified to match the Septuagint (or Old Greek) version. More rarely they may be harmonized to match the underlying Hebrew, or another version.
One might also refer to harmonization to the lectionary and/or liturgical practice. Consider, for example, the so-called doxology to the Lord’s Prayer, which is likely not original to the text—was its insertion accidental or intentional? So familiar it must have been to early Christians, scribes might easily have inserted it unconsciously during copying.

 Other Purposeful (Intentional) Variants:
These are more difficult to determine and define, but the following areas should be noted:
Attempts to clarify the text: primarily by means of
a) Substitution: replacing a word or form (with one that makes more sense), or
b) Addition: adding in words (sometimes referred to as “gap-filling”), especially within a narrative context. Commonly, for example, a name may be inserted to clarify the subject—who is speaking, etc. (e.g., “he said…” might be changed to “Jesus said…”)
Changes to safeguard doctrine: more often than not, these are Christological in nature, though sometimes it can be difficult to know for certain that they were intentional. For example, every time Joseph is referred to as Jesus’ “father” (or Joseph and Mary as his “parents”), the text was altered in a number of manuscripts, presumably to protect the doctrine of the virgin birth.
Interpolations: these represent typically larger additions—a phrase, sentence, or entire section. They are usually intentional, but occasionally may have been inserted accidentally (from a marginal comment, etc.). They are also among the most hotly disputed variants, affecting as they do a larger portion of the text. The “pericope of the Adulteress” (John 7:53-8:11) is probably the most famous major interpolation—critics and commentators are nearly unanimous in the opinion that the passage was not part of the original text.
Harmonizations: For these, see above. Many instances of harmonization are intentional, but it can be difficult to know for certain.


There are many techniques and principles used in textual criticism. In the third part of this article, I will discuss the so-called “Canons of Criticism”. I refer here to three fundamental methodological approaches:

• The Majority Method: The reading found in the majority of manuscripts is adopted, except in rare cases where the evidence is more or less evenly divided. Very few scholars today follow this approach in any meaningful sense—it is largely limited to extremely Traditional-Conservative circles. However, for most scholars, when the majority of manuscripts include many of the “earliest and best” witnesses, this does usually prove to be decisive.
MAJORITY TEXT: The text of the majority of manuscripts, often called “Byzantine”, as it stems from the time and region of the Byzantine Empire (prim. 9th-15th centuries), and is more or less synonymous with the so-called Byzantine Text (or Recension). The Textus Receptus (“Received Text”) of the Renaissance period is an off-shoot of the Majority Text.

• The Genealogical Method: So called because the genealogy of manuscripts is traced, by the grouping together of related manuscripts into a stemma (pl. stemmata)—that is, into clusters, families, and broader types. By tracing back the ancestors of these groups, one would, theoretically, end up with an approximation of the original text. Again, few scholars today will strictly follow such an approach; however, relationships between manuscripts are still studied intensively, and textual groups (or “types”) remain helpful in analyzing the text. The “Genealogical” method has been modified in recent decades; the most important to mention are:
1. The Local-Genealogical Method: As the term was meant when coined by Kurt Aland, the history of the text can be traced back (or forward) and established genealogically, but only at the level of each variation-unit.
2. Documentary-Historical Method: As practiced by many scholars today, this involves examining and grouping together manuscripts—in particular the early Papyri and Uncials—into related clusters. By comparing these, one attempts to determine the most reliable form of the (original) text. However, this can only be done from book to book (or for portions of a book), not for the NT as a whole.

• The Eclectic Method: This involves an objective analysis of each individual variant, examining both external and internal evidence (see below), and determining the most likely original reading on a case-by-case basis. This category is usually subdivided into:
1. Moderate or “Reasoned” Eclecticism: Here external and internal evidence is more or less evenly weighed and balanced, often with a priority given to external considerations (i.e., the age and quality of the manuscripts involved).
2. Strict or “Rigorous” (or “Consistent”) Eclecticism: Here consideration is given primarily (sometimes almost exclusively) to internal evidence. Rather few critics today follow this method.

Scholars generally utilize some combination of these approaches—most adopt a “moderate” eclectic method, but with a strong “documentary” emphasis on external evidence (the “earliest and best” manuscripts). As indicated above, when the best manuscript evidence coincides with the “majority” reading, this is almost always regarded as representing the original text.


In the first part of this article, I discussed the different kinds of “Textual Witnesses”. These witnesses make up the “Textual Evidence”—that is, all the evidence scholars and students must examine when attempting to evaluate variant readings and establish the original text. This evidence is further grouped into two classes:

• External Evidence: That is to say, all of the material evidence of the textual witnesses: the Greek MSS, versional MSS, writings of the Church Fathers, and the lectionaries. When examining this external evidence, one must look at each particular textual witness that contains a variant reading, in terms of: the age, quality, and reliability of the witness; also if the reading is supported by other reliable witnesses.

• Internal Evidence: This involves everything intrinsic to the text itself: style of writing; grammar, vocabulary, and theology of the author; scribal tendencies during copying; and so forth. Internal evidence is admittedly of a more subjective nature than the external: questions of authorial style, and the like, can be very much open to debate. More reliable are points related to scribal tendencies and practice, which are a bit easier to determine.

Note: This article will conclude with the third part: “Canons of Criticism”


Part 1  <> Part 3 coming soon