Earl S. Johnson, «Mark 15,39 and the So-Called Confession of the Roman Centurion», Vol. 81 (2000) 406-413
Continuing examination of the grammatical, literary and historical evidence indicates that the centurion's remarks about Jesus in Mark 15,39 cannot be understood as a full Christian confession of Jesus' divine sonship, and cannot be taken as a direct challenge to any Roman emperor in particular. Jesus' identity in the gospel is not revealed by the centurion, the demons, the disciples or in the introduction to the gospel. It is made clear by God's declaration that he truly is the Son (1,11; 9,7), and in the faith of the readers as they search for Jesus' presence in their own community.
both verses should be taken in the same light as Christian and confessional. If the grammatical basis for taking the confession in 15,39 as definite is shaky and uncertain since Marks usage of anarthrous nouns does not demonstrate that ui(o_j qeou= must be taken definitely7, the supposed corroboration of son of God in 1,1 renders the interpretation even more untenable. In the UBSGNT, son of God is only given a C rating, indicating that the enclosed words in the text are those whose presence or position [...] is regarded as disputed. In a thorough study, P.M. Head8 carefully reviews the reasons for doubt about the traditional reading in 1,1. As he demonstrates, text-critical evidence indicates that the original text of Mark probably did not include the disputed words. Based on the external evidence in wide ranging manuscripts and patristic citations, as well as internal evidence (the shorter text, for example, is obviously the more difficult reading and can easily be explained as an addition to buttress the churchs Son of God theology), Heads conclusions must be taken into account: The original, shorter form of 1,1 was supplemented by many MSS with Son of God in two different forms, probably around 100 AD. The scenario is more plausible than any other, it accounts for the other variants, and fits what we know of scribal habits and the tendency of gospel traditions9. Without the support of the anarthrous ui(o_j qeou= in 1,1 the case for taking 15,39 as definite is weakened considerably. The appealing symmetry of declaring divine sonship at the very beginning, the middle and the end of the gospel is lost. Jesus is declared the Son by the bat qôl at the baptism (1,11) and at the Transfiguration (9,7) but in both cases the title is clearly defined by the definite article. For Mark the clearest definition of Jesus identity is provided only by the very voice of God, not by the author of the gospel in his introduction, and not by any character in the gospel. The confession of the centurion must be understood in some other way.
2. The conclusion that Mark must have had divi filius (as a Roman description of the Emperor) in mind in 15,39 in order to challenge Roman Imperial theology and Roman belief in Augustus as a god is also highly questionable10. It cannot be assumed that the indefinite divi filius must have been taken over into Marks Greek. The Latin is an imprecise guide to Marks intentions: since Latin does not have definite and indefinite articles to correspond to those in Greek, the meaning must always be determined by context11.
Equally unconvincing is the contention that emperors after Augustus deliberately avoided the use of son of god to describe themselves, keeping that honor intact for the greatest emperor of them all, Divus Augustus. In fact, the Roman practice of granting divine status to the emperors was far more complex