The Chosen Son: Luke 3:22

At Jesus's baptism, a voice from heaven declares his identity. Every major English Bible translates it: "You are my beloved Son." The UPDV renders it differently: "You are my chosen Son."

The rendering follows the word to its source — through Greek, back into the Hebrew prophecy the voice is quoting — and finds that "beloved" obscures the concept the voice actually conveys.

What the voice is quoting

Scholars widely recognize that the baptism voice weaves together two Old Testament passages. The opening — "You are my Son" — echoes Psalm 2:7, the coronation formula spoken to Davidic kings. But the title applied to the Son and the second clause — "in whom I am well pleased" — come from Isaiah 42:1, the opening of the first Servant Song:

Behold my servant, whom I uphold; my chosen one, in whom my soul delights: I have put my Spirit on him. (Isaiah 42:1)

The match between the second half of the baptism voice and Isaiah 42:1 is unmistakable. God expresses delight in his servant. God bestows his Spirit. At the Jordan, the Spirit descends on Jesus and the voice expresses pleasure — fulfilling Isaiah's promise "I have put my Spirit on him." The majority of scholars recognize this allusion, and even the standard critical text notes the connection.¹

The title modifying "Son" — the word translated "beloved" or "chosen" — belongs to the Isaiah half of the formula, not the Psalm 2:7 half. Psalm 2:7 has no such modifier; it simply says "You are my Son." The modifier comes from Isaiah 42:1, where the corresponding word is unambiguous.

The Hebrew word in Isaiah 42:1 for "my chosen one" is bechiri (בְּחִירִי). The Septuagint translated it correctly as eklektos (ἐκλεκτός) — "chosen" or "elect." The Aramaic Targum also has "chosen." The Syriac Peshitta at Isaiah 42:1 has gabya (ܓܒܝܐ) — "chosen." Every ancient version of Isaiah 42:1 agrees: the word is "chosen."²

The question is what happened to that word on its way into the Gospels.

How "chosen" became "beloved"

The Greek word at Luke 3:22 is agapetos (ἀγαπητός). Modern translations render it "beloved," and for most Greek contexts that would be correct. But agapetos has a second life in biblical Greek that standard dictionaries acknowledge but English Bibles ignore.

In the Septuagint, agapetos does not primarily translate a Hebrew word for "beloved." Its most common Hebrew equivalent is yachid (יָחִיד) — "only." When Genesis 22:2 says Abraham should take his yachid son Isaac, the Septuagint renders it agapetos — not because Isaac is "beloved" in some generic sense, but because he is the only son of the promise. C.H. Turner demonstrated this in 1926 in the Journal of Theological Studies, and BDAG — the standard Greek-English lexicon — classifies the baptismal uses of agapetos under sense 1: "only, only beloved, in ref. to an only son," not under sense 2: "dear, beloved."³

This explains how agapetos entered the divine voice formula. It was a theologically loaded word evoking the Isaac tradition — the unique son of promise — and it displaced the original "chosen" (eklektos) from Isaiah 42:1 because it carried higher christological freight. The standard Greek lexicon thus confirms that the default English rendering "beloved" does not reflect the sense BDAG assigns to this word in the baptismal formula.

But agapetos did not simply mean "only" in the baptism formula. When it was installed into the slot previously occupied by eklektos, it became a de facto Greek equivalent for Hebrew bachir ("chosen") in early Christian usage. The mechanism is visible at Matthew 12:18, where the Gospel quotes Isaiah 42:1 directly. Where the Septuagint has eklektos ("chosen"), Matthew writes agapetos. The translator kept "chosen" as a verb in a relative clause — "whom I have chosen" (hairetisa) — but replaced the title itself. "My Chosen One" became "my Beloved." The concept was demoted from a title to a verb, with agapetos installed in the title slot.⁴

This is where the substitution became self-reinforcing. Once agapetos stood in the baptismal formula, it was imported backward into quotations of Isaiah 42:1. And once it stood in the Isaiah quotation, it looked as though "beloved" had always been the right word. The loop closed, and "chosen" disappeared from the divine voice tradition everywhere — almost.

Where "chosen" survived

Luke preserves "chosen" at two other points where God or the narrator identifies Jesus.

At the Transfiguration (Luke 9:35), the NA28 critical text prints eklelegmenos (ἐκλελεγμένος) — "chosen" — as the original reading of the heavenly voice. The Old Syriac Sinaiticus has gabya (ܓܒܝܐ), the same Syriac word used at Isaiah 42:1. Later manuscripts changed this to agapetos to match Matthew 17:5 and Mark 9:7, but the editors of the standard critical text judged "chosen" to be what Luke wrote.⁵

At the Cross (Luke 23:35), the crowd mocks Jesus as "the Christ of God, the Chosen One" (ho eklektos). This reading is uncontested in every manuscript.

Luke therefore uses "chosen" (eklektos / eklelegmenos) for Jesus at both the Transfiguration and the Cross — the two other grand moments where Jesus's identity is declared. The only place Luke has "beloved" is the baptism (3:22), which is also the verse under the most intense harmonization pressure from Matthew and Mark. The pattern is clear: "chosen" is Luke's word for Jesus. "Beloved" is the outlier, and the outlier sits at the one verse most susceptible to scribal conformity with the other Gospels.⁶

The lexical double-bind

Even if a reader accepts agapetos as the correct Greek text at Luke 3:22, the case for translating "chosen" does not depend on emending the Greek. It depends on what the Greek word means in this formula.

Joachim Jeremias, in the Theological Dictionary of the New Testament, addressed this directly. Discussing the heavenly voice at the baptism and transfiguration, he wrote:

"when the text of the saying from heaven at the baptism and transfiguration vacillates between agapetos and eklelegmenos (Lk. 9:35), these are presumably alternative renderings of bachir in Is. 42:1, which is sometimes translated eklektos (LXX), sometimes agapetos (Mt. 12:18)."⁷

That is: even the Greek word agapetos, in this specific formula, is functioning as a translation of Hebrew bachir — "chosen." The two Greek words are not competing meanings; they are alternative renderings of the same Hebrew source. Translating "chosen" is not correcting the Greek. It is translating what the Greek itself represents.

Jeremias went further, arguing that the baptism voice was originally based on Isaiah 42:1 alone — not a blend of Psalm 2:7 and Isaiah 42:1, as commonly supposed. On his reconstruction, the original title was pais mou ("my servant"), which the ambiguous Greek word pais — meaning both "servant" and "child" — allowed to be reinterpreted as "my Son" (huios) in the Greek-speaking church. The UPDV rendering does not depend on this fuller reconstruction; it requires only the widely accepted premise that the modifier in the baptism voice derives from Isaiah 42:1.⁸

John 1:34 — "chosen" at the Jordan

There is one more witness. In John's Gospel, John the Baptist concludes his testimony about the baptism with a declaration about Jesus. The standard text reads: "This is the Son of God" (ho huios tou theou). But the earliest manuscript evidence — including Papyrus 5, the original hand of Codex Sinaiticus, Old Latin manuscripts, and the Old Syriac — reads: "This is the Chosen One of God" (ho eklektos tou theou).⁹

Jeremias noted the significance: "ho eklektos tou theou is a Messianic title deriving from Is. 42:1. Jn. 1:34 shows plainly that the voice at baptism must have been originally a uniform quotation from Is. 42:1."

This early form of the Baptist's testimony independently corroborates the pattern: at the Jordan, Jesus is identified with the Isaiah 42:1 title — "the Chosen One of God." The same word. The same source. The same event.

The UPDV rendering

The UPDV translates Luke 3:22: "You are my chosen Son; in you I am well pleased."

This follows the methodology of Semitic Substrate Restoration: when converging evidence shows that a Greek rendering has obscured the underlying Semitic concept, the translation restores the original meaning. The same methodology produces "Speech" for logos in John 1 (from Aramaic Memra) and the active voice at Matthew 1:16 (from the Syriac reading). No Greek manuscript reads eklektos at Luke 3:22. But the Hebrew of Isaiah 42:1 does. The Septuagint does. Luke himself does at 9:35 and 23:35. The earliest witnesses to John 1:34 do. And the standard theological dictionary of the New Testament says that agapetos in this formula is a rendering of Hebrew bachir — "chosen."

Whether Luke originally wrote eklektos at 3:22 (as he did at 9:35) and scribes altered it, or whether he wrote agapetos directly from his Markan source, the underlying Semitic concept remains the same. The UPDV does not claim to reconstruct a lost Greek reading. It translates what the Greek word represents in this formula: the Hebrew bachir of Isaiah 42:1, confirmed by the standard theological dictionary, supported by Luke's own vocabulary at the Transfiguration and Cross, and independently attested in the earliest witnesses to John's testimony at the Jordan.

The voice said "chosen."¹⁰

---