Paul does use a shortened version of the phrase (ἐκ πίστεως without a genitive qualifier) no fewer than seven times in Galatians alone [3:7, 8, 9, 11, 12, 24; 5:5; add 3:23 twice, 25, 26; 5:6; 6:10] and in each instance it refers to their ‘faith’ in Christ, by which they were justified, not to the faithfulness of Christ that made such justification possible. Since ἐκ πίστεως by itself seems to be a pickup of the longer phrase, it would be especially strange for Paul to give it a different meaning from that intended in the two instances where the same phrase occurs with ‘Christ’ as the genitive qualifier (2:16; 3:22). That is, ἐκ πίστεως means that we live ‘by faith’=trust in Christ Jesus, how is it that the longer version of the phrase makes Christ the subject of the ‘faith’ rather than the object?” (Pauline Christology: An Exegetical-Theological Study [Peabody, MA: Hendrickson, 2007], 226-227).
It seems that I need to clarify my question.
The issue under dispute: The meaning of Paul’s longer phrase, ἐκ πίστεως Χριστοῦ [ek pisteos Christou], in Galatians 2:16. Does the phrase mean “by faith in Christ” or “by the faithfulness of Christ”?
Gordon Fee's reasoning:
First: The grammatically unqualified phrase, ἐκ πίστεως [ek pisteos] in Galatians 3:7, 8, 9, 11, 12, 24, means, by implication, “by faith [in Christ]” even if "in Christ" is not present in those passages.
Second: The later grammatically unqualified phrase, ἐκ πίστεως [ek pisteos], derives its meaning from the earlier longer and grammatically qualified phrase, ἐκ πίστεως Χριστοῦ [ek pisteos Christou], in Galatians 2:16.
Third: It would be quite strange for Paul to give ἐκ πίστεως [ek pisteos] a different meaning from the meaning he intended in those instances where he uses the same phrase but with the grammatical qualifier, ἐκ πίστεως Χριστοῦ [ek pisteos Christou] as in Galatians 2:16 and later in 3:22).
Therefore: The earlier and longer grammatically qualified expression (ἐκ πίστεως Χριστοῦ [ek pisteos Christou]) must mean "by faith in Christ" and must not mean "by the faithfulness of Christ" because the later grammatically unqualified expression (ἐκ πίστεως [ek pisteos] speaks of our faith in Christ.
If the meaning of A is the issue under question, and B derives its meaning from A, then how can we appeal to the meaning of B to define A without assuming the very thing that we are endeavoring to prove? This is how I understand Gordon Fee's argument. If I am correct, then, has he not begged the question? But, have I correctly understood Gordon Fee?