Criteria in "Lower" Criticism

So I've been doing a lot of my own text-critical work lately, and it behooves me to go back and check my method. It has, after all, changed a bit over the course of making decisions—even over just three chapters of Hebrews, not to mention the time I've been playing with scriptural texts and manuscripts, period.

So today I'm going to meditate a bit on the text-critical criteria and rules of Bruce Metzger and Kurt and Barbara Aland. Two great and very important books, both entitled "The Text of the New Testament," and both deeply relevant to the state of contemporary text-critical scholarship in the New Testament. If you've used the Nestle-Aland or UBS Greek texts, you owe your textual basis to these folks and their colleagues—and their critical judgments.

(The Hebrew scriptures are another mess entirely, with no longstanding tradition of eclectic critical texts, because that canon has no such broad and varied basis of manuscript evidence available. The BHQ is not seriously changing this status quo, though it enlightens the variant situation considerably since BHS. But then, the Masoretic Text is not analogous to the text of the Byzantine majority, either.)

It's important to understand the principles with which one makes decisions between textual variants among the manuscripts. You'll have principles, even if you don't set them out and understand them solidly, and naive principles, once they solidify into bad habits, can be an intractable mess to sort out. These, at least, are time-tested good ideas, and they have a good bit of humility and self-criticism built in, because they understand the nature of the object.

Metzger's Criteria

Bruce Metzger sets out two categories of evidence: external and internal. External evidence is external to the text. It is about the means of transmission. And Metzger categorizes external evidence as dealing with three things:
  • The date of the text (not the manuscript), considered especially in terms of its genealogy, and not just its material date as a copy
  • The geographical distribution of concurring witnesses to a text, considered rather like catholicity in terms of the preference for non-genealogical agreement across regions
  • The value of a text with its genealogy, under the rubric ponderantur testes, non numerantur (i.e. we do not count texts like votes in an election; we weigh them like evidence in a trial)
That last brings us very close to internal evidence, as we consider the value of a reading, its "weight" in the tradition. While the number of votes cast for the Byzantine reading is high because the production runs were high, other considerations devalue it, especially its late date and its evident disagreement with earlier manuscripts into which it was inserted editorially as "corrections." But that reminds us that value is often considered as a kind of product of date and region concerns. The "text type" labels—Alexandrian, Western, Ceasarean, and Byzantine—represent a kind of total familial judgment, and the modern verdict comes out overwhelmingly against the Byzantine and for the Alexandrian. (Which is lately being modified as the pendulum swings, given the severity of that judgment and the idea that there is some historical value to the Koine traditions, and not pure novelty.) But farther than that we really can't say without summing over the internal considerations.

Metzger lists a larger range of internal evidence, lumped into two categories: scribal habits ("transcriptional probabilities") and authorial habits ("intrinsic probabilities"). It's important to know which classic rules come under which heading, too. Under scribal habits, we have three general preferences:
  • Difficult readings, under the rubric lectio difficilior potior (i.e. since harmonization and "correction" of "errors" seem more likely as editing than making a text harder to understand, we try to understand the more difficult reading as authorial—we prefer rough truth to smooth sense)
  • Shorter readings, under the rubric lectio brevior praeferenda (i.e., since it seems more likely that a scribe will harmonize something to something else by addition of material, the reading without it ought to be closer to the author)
  • "Verbally dissident" readings, meaning that passages in the author's own words are to be preferred to passages that line up with external sources like the LXX or another gospel—or even a different Pauline epistle
So: beware of variants that a) tie up problems too neatly, b) have extra words in them, or c) match other sources too closely. Editors do things like that intentionally more often than they do the opposite, and unintentional mistakes are usually simpler and more obviously on the whole. These are good rules, as long as you realize that individual scribes and individual manuscripts all have particularities with their own characters, and identifiable tendencies in both authors and editors trump general rules.

Under authorial habits, Metzger gives us several general concerns to watch for. A few are gospel-specific, but I'm going to generalize them because I believe they apply just as well outside:
  • Style and vocabulary—is a variant consistent with the language the author uses elsewhere in this writing?
  • Immediate context—does it make sense in the flow of this piece of text?
  • Overall usage—does it sound like how the author uses this language elsewhere, or like someone else is using the same language differently?
  • Sense in character—does the variant fit the cultural context of the author or original speaker? (Metzger cites the Aramaic background of Jesus' teaching, but we could just as well invoke the Hellenistic Judean context of Paul's teaching under this rubric.)
  • Source priority—does a variant look like authorial usage of an earlier text (e.g. Mark in Matthew, or LXX in Hebrews), rather than editorial harmonization to it? (We expect flexible usage from authors, and literal corrections from editors.)
  • The role of tradition—as Metzger says, "the influence of the Christian community upon the formulation and transmission of the passage in question" (This seems to be its own intermediate category of concern, for how non-scribal preservation history influenced the reading of a text.)
There's a basic holistic sense here, something that you can only gain by familiarity not only with a broad base of Greek literature across periods and regions, but also with a deep selection of an author's work. You have to know the language, in its grammar, syntax, vocabulary, and stylistic range, well enough to pick up on the specific verbal character of this author. This is, of course, easier to do with authors from which we have a corpus of authentic works, rather than the singletons in the collection. And things like Synoptic usage are sort of an intermediate category, because you still have to know Matthew, Mark, and Luke as individuals.

The Alands' Rules

Kurt and Barbara Aland set out twelve basic rules, rather than make such a systematic taxonomy of manuscript, editorial, and authorial qualifications. They lean very clearly on a similar distinction between exterior and interior criteria to Metzger's taxonomy, but this set of rules is additive to such an ideology of the text. Think of these as guidelines to the guidelines. (I've simplified these somewhat; you can find the originals on pp. 275-6 of the book.)
  1. However many variants there may be, only one reading can be original.
  2. Only the reading that best satisfies the requirements of both external and internal criteria can be original.
  3. Text criticism must always begin from external criteria and only afterward turn to internal criteria.
  4. External criteria can never be trumped by internal criteria, nor can internal criteria be the sole basis for the critical decision. (An original may be good, but goodness cannot make a reading original; a late text is late.)
  5. The Greek manuscript tradition is the primary authority; versions in other languages (Latin, Coptic, Syriac, etc.) and the Patristic witnesses can only supplement and corroborate readings, especially where the Greek text is unclear. (They cannot be considered as variants in their own rights.)
  6. Manuscripts are weighed, not counted, and every passage and text should be considered individually from the available witnesses ("the local principle"). Consider the peculiar traits of each manuscript. No single manuscript or group of manuscripts, whether papyri, uncials, or minuscules, no matter how important or early, can be considered reliable in any simple sense, even though we may trust certain combinations more than others.
  7. It is only hypothetically possible that the original reading may be found standing alone against the rest of the evidence. In practice, following minority readings as a general rule would lead to the worst sort of eclecticism, based on personal bias rather than the search for an original text.
  8. Genealogical analysis of the available readings in terms of possible descent from an original is very important, because it lets us understand which reading may best explain how the tradition produced the other variants. That reading is more likely original.
  9. Never treat variants in isolation from one another. Context is essential if we are to avoid deciding in favor of an eclectic text of a passage, from the best of each individual variant, that appears as a whole in no actual manuscripts.
  10. Do not choose difficult readings as though their difficulty made them original.
  11. Do not choose shorter readings as though their brevity made them original.
  12. "A constantly maintained familiarity with New Testament manuscripts themselves is the best training for textual criticism. In textual criticism the pure theoretician has often done more harm than good."
These could be boiled down more than they are, though I don't expect that from a German, especially one as detail-oriented as Kurt Aland. Essentially, what we have here is concern for the original reading of the text. Examination of the whole collection of manuscript evidence is not about finding the best-looking text, or of taking the best turn at every point, but rather about finding the most original text. The original is a real thing, and it may be sloppy, and it may be nice, but what matters is that it is the text of the author. "Original" cannot be classified in any other way. Every other rule must have enough flexibility in it to accommodate the reality that the original text of any writing has no other qualities to commend it.

To that end, the Alands tell us that we need to understand the texts at a range of scales, from the very local context of the individual variant, to the local context of the passage in which it appears (possibly with other variants), to the context of the whole writing in which it appears, to the context of the author's oeuvre, to the context of the whole New Testament as a collection of Greek literature. And it is textual familiarity and discernment across these scales that will tell us the most—attention to actual texts pays off far more often in the search for a real original text, than attention to theory. Theory is suited more often to pursuit of an ideal, and there is no ideal text of the New Testament. There are only the many texts we have, and the originals we know were there before them.

What Does This Mean?

In some ways, the result is an idea unpalatable to the postmodern mind: we wind up looking for the Author, the dead man, as the source of authority. But we are looking for no ideal rational Author, nothing that conforms to universal principles. We are looking, instead, for a very specific, peculiar thing for each text in the canon: the original text, without any later meaning-making inscribed upon it. And we can expect such a thing, in each case, because all of the related things we have descend from it. We expect to be able to at least point more closely to it by knowing about the likelihoods involved in transmission history, and by learning the character of each author and each text as best as it may be possibly known. We are looking for a text with its own character, written by an author with its own character, to an audience with its own character. And we are looking along the way for editors and copyists with their own characters in manuscripts with their own characters, which have been transmitted through communities with their own characters. That's a lot of particularities to keep track of, in a very human process from top to bottom.

We don't posit the unreality of the manuscript traditions at any point. They are, in fact, all we have. The real text of the New Testament is not any one of these manuscripts—it is the thing witnessed by all of them taken together. (Awfully like the Word of God and its relation to the texts of scripture.) Ideally, the whole manuscript tradition can be explained by a set of originals and a set of paths taken through history. And this text does not presuppose its interpretation. The idea in "lower" criticism is to find the solid ground from which we can most reliably make meaning in "higher" criticism—but to keep that search for meaning as separate from the search for the text as possible.


I find myself saying this about many things lately: the text is a sort of seminarium, and the "Bible" a cold-frame or a greenhouse built to house this soil in which faith grows. It can support a vast range of forms and contents of trust in God. The question of "getting it right," when it comes to scripture, is decidedly not about aligning the text with our sense. This is one of the reasons why, though it has been edited for content across its history, we (at our best) refuse to edit the texts of the Bible for content today. We edit translations freely, and much comes in and slips out of our English—and well it should! There is nothing of lasting value about the KJV, nothing of lasting value about the NRSV, the NIV, or any of the others. The English is not the text of scripture. Translation is always in some sense a betrayal of the witness. It is interpretation, and there is nothing of lasting evidentiary value in its secondary witness to the witnesses, that can stand itself up next to the witnesses themselves. Interpretation is where the game is played—and it is a serious game—for the understanding of the text. And so we do not edit the text, except to secure the best objective basis for all later interpretive work. And at that, we do the best we can in every time, subject to the limitations of our knowledge.

This is among the major reasons why my views on the authority of scripture will never line up with the various text-authoritarian positions that use words like "inerrant," "verbally inspired," etc. It's not that I'm disillusioned, by any means; scripture is a profound basis for faith, and I trust in God because I have seen this God in scripture before knowing this God in life and prayer. But scripture is that basis as what it is—and it is perfectly secure as what it is, in ways we cannot improve upon by preferring some particular version, or any of the other ways we try to make the text more secure for ourselves. All we can do by those means is attempt to shore up our interpretations, which is a worthless task. Interpretations are the leaves of a tree, and they are meant to change and fall away. You will grow new leaves, as the Spirit gives you new life every day.

The witness of the manuscript traditions is to something not wholly preserved in any one of them, just as the witness of the texts is to something not wholly preserved in any one of them. And our witness must be to something not wholly preserved in any one of us, any more than we could ever contain or protect or preserve God. It is God who upholds and supports and preserves us, as we go about our life and work.


  1. Incidentally, stuff like this is also why I don't place much stock in any of the archaeological freakouts, like Karen King's very recent and widely engrossing advertisement of what is turning out to be a fraudulent text on a papyrus scrap. They take advantage of the pop imagination in which Christianity has one source, which has one voice and one vote on its truth.

    And the scripture-principles of fundamentalisms exacerbate this problem by asserting exactly that, and attempting to make it the strongest and only sovereign voice, the only one that matters. And especially if it has one voice that contradicts itself, as is frequently asserted, any external disconfirmation pushes against the authority of scripture and therefore of the religious system that upholds it. Enter every archaeological freakout, no matter how legitimate the discovery or the discoverer, that might bear on ancient history in the region—because it might disprove the authority of Christendom.

    None of which has any effect on the reality of what scripture is, in its plurivocal witness to God in history, and to the relative univocity of its preservation. The variants add no uncertainty to the reality they all testify. The differences among the writings that we accept as canon add no uncertainty to the reality they all testify. Not as long as we remember and respect what scripture is, and what it does.

    You may stack up all the real writings that do indeed differ in opinion against scripture and not disprove it. And you may of course disagree with scripture and its collective witness. But nothing found yet, no matter the popular exaggerations of its meaning, can do better than that. And none of it can therefore serve as any real threat to the witness to God. Threaten the systems all you like, and topple them if you like, as long as you don't harm the people; you cannot damage the deity with history—not even real history.


Post a Comment