106 Timo Flink (1), (2) and (4)? Such a case would require transcriptional activity that produced three independent abbreviated readings, each dropping out a different part of the phrase. Doctrinal concerns (anti-Adoptionism) or an inter-textual harmonisation to Ps 2,7 can explain the phasing out of κλεκτ ς. In Ps 2,7 God addresses his son with the words υ ς μου. That can easily become υ ς το θεο in the hands of the scribes, because the context in John 1,34 requires such an alteration to the phrase of Ps 2,7. Similarly, an inter-textual harmonisation to Isa 42,1 LXX ( κλεκτ ς μου) can explain the dropping out of υ ς. Both harmonisations are possible if the scribes were consciously thinking of one Old Testament background for John 1,34, but why would anyone drop out το θεο ? It is more likely that it is added than omitted. This leaves υ ς κλεκτ ς as the reading from which other readings derive by the aforementioned scribal harmonisations. The reading υ ς κλεκτ ς has much to commend itself. Such a reading is awkward and abrupt, so it is easy to see why scribes would add το θεο to it. Thus, such a reading has left secondary footprints into the textual tradition. If one allows this, the reading has secondary support of some witnesses in Latin speaking Western Rome and in Greek and Coptic speaking Egypt. It is an early, second century reading. It fits the contexts by allowing seven unique honorific titles for Jesus: (1) μν ς το θεο , (2) υ ς κλεκτ ς, (3) αββ , (4) μεσσ ας, (5) υ ς το θεο , (6) βασιλε ς το ‘Ισρα λ, and (7) υ ς το νθρ που. It satisfies the chiastic structure suggested by Ellis, balancing υ ς κλεκτ ς with υ ς το θεο . It combines two Old Testament backgrounds: the sonship motif from Ps 2.7 with the servant motif in Isa 42,1. It also contains a reference to the “chosen one” motif from Exod 4,22; 19,5-6; Deut 7,7-8. The latter passage is especially interesting, because Israel is likened to God’s beloved, which motif recurs in the baptism of Jesus. This defuses the stalemate seen in the arguments of previous proposals. It explains the rise of other variants due scribal activities, such as textual harmonisations to the immediate context and to the author’s style. It is lectio difficilior and liable to heterodox interpretations. It is not the only unique expression in the Fourth Gospel. The phrase γιος το θεο in John 6,69 is another one. Its uniqueness is not a good argument against it. I agree with Rodgers’ approach that scribal activity has produced the plethora of readings but I disagree with his solution. In my opinion, a singular reading is a better candidate for originality than a conjectural emendation. Singular readings are factual while conjectural emendations are hypothetical. Thus it seems safer to choose the singular over the emendation, in this case at least. It is my view – based on the findings of this study – that the Fourth Evangelist wrote originally κ γ ρακα