I was reminded today about an area that I used to think about quite a bit: textual criticism. My main interest is in seeing how various readings might impact the discourse, i.e. “What if it read like this?” The discussion that spurred this post was about Galatians, so I did some digging around in the SESB NA 27 text to find an interesting example. I landed on Gal 3:7. It is a fairly complex context, but a great illustration of how discourse considerations can inform text-critical judgments. If nothing else, it provides a minimal pair to illustrate the difference of stating it one way versus another concocting unnatural examples.  Here is the Nestle-Aland text:

5 ὁ οὖν ἐπιχορηγῶν ὑμῖν τὸ πνεῦμα καὶ ἐνεργῶν δυνάμεις ἐν ὑμῖν, ἐξ ἔργων νόμου ἢ ἐξ ἀκοῆς πίστεως; 6 Καθὼς Ἀβραὰμ ἐπίστευσεν τῷ θεῷ, καὶ ἐλογίσθη αὐτῷ εἰς δικαιοσύνην· 7 γινώσκετε ἄρα ὅτι οἱ ἐκ πίστεως, οὗτοι υἱοί εἰσιν Ἀβραάμ.

Verse 5 asks a rhetorical question in the form of a verbless clause. The topic of the clause is a thematically-loaded reference to God which characterizes Him in a particular way based on Paul’s rhetorical purposes in the context. He is referred to as “the”One who gives you the Spirit and works miracles.” This effectively makes God the agent, it is He who does these things and not the believer. The key question concerns the basis for these things being done, whether it is by works of the Law or by hearing by faith? The answer is withheld until the end of verse 7, giving the reader some time to ponder it.

Verse 6 is the beginning of a complex clause, as analyzed in the in the LDGNT:

Here is how I break it down. The καθώς clause of verse 6 establishes a comparative frame of reference for the main clause that follows in v. 7. It is fronted w.r.t. the main clause not because it is the most important part, but to establish a basis of comparison for something that follows: “Even as/just as Abraham believed God and it was credited to him as righteousness…” So just as Abraham’s believing and crediting happened,1 something else happened. This something else provides the answer to the rhetorical question of v. 5.

Verse 7 begins with what I have referred to as a “meta-comment”, or what is traditionally called a “disclosure formula.” No matter the name, γινώσκετε in this context serves to attract attention to the clause that follows. This is accomplished in the first place by not being semantically required in the context. The call to “know” something is rhetorical, and the verb could be excluded from the context without significantly harming the proposition: “The ones believing are the sons of Abraham.” Second, it draws attention to what follows by delaying the introduction of the main clause that follows, introduced by ὅτι. The meta-comment is not the only delay tactic used to build up to the climactic answer.

The proposition following the meta-comment in v. 7 contains a left-dislocation [LD], otherwise known as a pendens construction, hanging nominative, and a few other things. What all of these terms refer to is the reference to something at the beginning of (technically outside of) the clause that is reiterated again within the clause, typically using a pronoun. I claim elsewhere that there are two primary motivations for using this construction. The first is to introduce a complex entity (e.g. a hypothetical person or a proposition). This can be illustrated from Mt 12:32, where two complex hypothetical entities are introduced and juxtaposed:

ὃς ἐὰν is commonly used to introduce some hypothetical figure. If the entity is too complex to be introduced as the subject, using a left-dislocation streamlines the introduction by avoiding the use of a second “There once was an X, and X did Y.” Once it is introduced, a simple pronoun can reiterate it  in the main clause.

A second motivation for using this construction is rhetorical in nature. Introducing something and reiterating it adds prominence to the entity or proposition, especially when the reiteration is also placed in a position of prominence. This prominence motivation is particularly evident when the entity introduced is not too complex to have been introduces without a dislocation. Consider the case of Mt 5:19, where the first clause introduces a complex entity using a simple topical frame of reference, whereas the second clause uses a dislocation.

Once the first entity has been introduced, the positive counterpart should be even easier to process, meaning there is no need for the dislocation. The use in 19b is rhetorically motivated, giving the reader pause to think about “the one who would do and teach” before reading what is asserted about them: They will be called greatest in the kingdom of heaven. The pronoun is able to capture the whole concept or proposition in a single word.

Now we have the context needed to consider Gal 3:7. Note that the entity introduced is not very complex compared to the other examples we have looked at. A simple topical frame of reference would have sufficed, as with the introduction of Abraham in 3:6. This implies that the dislocation plays a rhetorical role, not a semantic-processing one. Here is where the text-critical issue arises. Here are the two readings. The underlining marks a frame of reference, the bolding indicates something placed in marked focus for emphasis sake.

NA27: γινώσκετε ἄρα ὅτι οἱ ἐκ πίστεως, οὗτοι υἱοί εἰσιν Ἀβραάμ.

Alternate: γινώσκετε ἄρα ὅτι οἱ ἐκ πίστεως, οὗτοι εἰσιν υἱοί Ἀβραάμ.

One of the challenges of information structure is accounting for all of the data. One cannot selectively apply the principles. If I were to claim that οὗτοι in NA27 is fronted for emphasis, then it leaves the placement of υἱοί before the verb unexplained. It is either fulfilling some pragmatic function (e.g. marked focus or a topical frame) or it should follow the verb as in the alternate reading. About the only way to understand the NA27 reading is how I analyzed it in the LDGNT. If the element that is fronted is complex, sometimes only the most salient element will be fronted. This appears to be the case with υἱοί being separated from Ἀβραάμ. Mike Aubrey discussed this is a series of posts on Devine and Stephen’s book Discontinuous Syntax: Hyperbaton in Greek. The NP υἱοί Ἀβραάμ is not particularly complex, leading one to think that the context presupposes that there there is some relation to Abraham, it is just a matter of what kind. Fronting the most salient element of the NP highlights that it is as sons.

Aside from the little matter of the earlier and harder reading criteria, the alternate reading provides a more satisfactory option based on the discourse context. Recall from 3:5 the rhetorical question regarding the basis for God’s activity, whether it was from works of the Law or from hearing by faith. Abraham is then held up as a model of faith, with his belief being the basis of the righteousness being credited to him. In the alternate reading, prominence is not assigned to the relationship the ones believing have to Abraham. Instead emphasis is placed on who is related to Abraham as his sons. It is those who believe by faith as opposed to some other group. Recall that the question of v. 5 focused on whether works of the Law or hearing by faith was the key. The alternate reading provides a much tighter connection to this question. One’s relationship to Abraham has not been an issue in the context, nor is it in the following context. Instead the focus is on belief as opposed to works.

So does this resolve the text-critical issue? Hardly. If anything, it provides more support for the NA27 reading since it is the dis-preferred reading, the one more likely to have been corrected by a later hand (as appears to be the case from Sinaiticus). However it illustrates what is at stake in such decisions. It is not just a matter of one little word inconsequentially being swapped with another. Such changes can radically alter our understanding of where Paul was pounding the pulpit.

  1. on the basis of faith, it just so happens. [↩]