Paul confirms the rumor the Philippians had heard (v. 26) and raises the stakes: Epaphroditus was not merely unwell but παραπλήσιον θανάτῳ, right up against death. The verse then pivots on ἀλλά to God’s mercy, and Paul characteristically widens the circle of that mercy to include himself — a small but revealing move: the apostle who has just written that he would count being poured out as a libation a joy (v. 17) does not pretend to be indifferent to loss.

Three things carry the weight:

1. παραπλήσιον θανάτῳ — a NT hapax adverbial construction; the idiom measures nearness, not arrival, and prepares for μέχρι θανάτου ἤγγισεν in v. 30.
2. ἠλέησεν — recovery from illness is named as mercy, not as luck or as answered technique. The verb belongs to the LXX ḥesed/raḥam domain.
3. λύπην ἐπὶ λύπην — layered grief. Paul already has grief (imprisonment, the situation of vv. 20–21, perhaps the church’s distress); Epaphroditus’s death would have stacked another course on top. The rendering must keep the stacking image, not resolve it to ‘more sorrow.’

Note also what Paul does not say: no explanation of why God spared this man and not others, no theology of healing, no claim to have prayed him well. The reticence is data.

Built onobrien-nigtc · fee-nicnt · bdag

παραπλήσιον is the neuter of παραπλήσιος (παρά + πλησίος, ‘near alongside’), here used adverbially with the dative θανάτῳ: ‘very nearly to death,’ ‘close alongside death.’ It is a NT hapax. BDAG classifies the neuter as adverbial and glosses the phrase ‘nearly to the point of death.’

In classical and Hellenistic prose παραπλήσιος normally means ‘similar, comparable’ (Thucydides, Polybius passim), and the medical writers use it of conditions resembling one another. Here, however, the dative and the context force the spatial-proximity sense rather than mere similarity: not ‘something like death’ but ‘right up next to death.’ The distinction matters — Paul is reporting severity, not describing a death-like state.

The construction anticipates v. 30, μέχρι θανάτου ἤγγισεν (‘he drew near as far as death’), and both stand deliberately short of v. 8, where Christ is ὑπήκοος μέχρι θανάτου and actually dies. Epaphroditus comes to the boundary; Christ crosses it. Any rendering that makes Epaphroditus ‘at death’s door’ in the same register as Christ’s obedience unto death blurs a line Paul keeps sharp.

Some manuscripts read the accusative παραπλήσιον θάνατον (D F G and part of the Latin tradition), which flattens the construction into a cognate-accusative-like ‘he was ill a near-death illness.’ The dative is better attested (𝔓⁴⁶ א A B C) and is the harder, more idiomatic reading; NA28 prints it without brackets.

Built onbdag · na28 · obrien-nigtc · silva-becnt

Aorist of ἀσθενέω, repeated from v. 26 where it stood in indirect discourse (‘they heard that he had fallen ill’). Here καὶ γάρ confirms the report as fact: ‘for he really did fall ill.’

The binding gloss for this corpus is ‘be weak,’ since Paul’s dominant usage is figurative (Rom 14:1–2; 2 Cor 11:29; 12:10). This verse forces the physical sense — nearness to death, divine mercy, recovery implied — and so the deviation is flagged rather than silent (§4). The aorist is best read as ingressive/complexive: he fell ill and was ill, with the illness now past.

Greek has more specialized illness vocabulary available (νοσέω, κάμνω, ἀρρωστέω), and Paul’s choice of the broader ἀσθενέω keeps the word inside his own weakness-vocabulary. It would be over-reading to make that a deliberate theological pun here; it is not over-reading to notice that the same verb Paul applies to himself in 2 Cor 12 is what nearly killed his messenger.

Built onbdag · obrien-nigtc · fee-nicnt

Aorist of ἐλεέω, ‘show mercy, have compassion on.’ In the LXX ἐλεέω/ἔλεος is the standard rendering of the ḥesed/raḥam family (Exod 33:19; Ps 6:3; Isa 49:13), so the word carries covenantal loyalty and not mere pity. Paul’s other uses are overwhelmingly soteriological (Rom 9:15–18; 11:30–32; 1 Cor 7:25; 1 Tim 1:13, 16).

The theological point is that Paul names a recovery from illness with the same verb he uses of God’s saving mercy. Physical deliverance is not a lesser category of act; it is ἔλεος. At the same time Paul does not universalize: he says God had mercy on this man, which implies nothing about cases where the illness runs its course. The verb reports a particular act, not a rule.

Built onbdag · muraoka-lxx · obrien-nigtc

A οὐκ … ἀλλὰ καί construction: not exclusion of the first term but addition of the second — ‘not him only, but me as well.’ The postpositive δέ stands in third position after the negated pronoun, which throws emphasis onto αὐτόν; the effect is close to ‘and it was not just him God spared.’

The accusatives αὐτόν and ἐμέ are both objects of ἠλέησεν. Paul thus makes himself a recipient of the mercy shown to Epaphroditus — a striking claim, since the mercy consisted in another man’s survival. This is of a piece with the whole letter’s logic of shared life (κοινωνία, 1:5; 2:1; σύμψυχοι, 2:2): benefit to one member is benefit to Paul, and Paul says so without embarrassment.

Note also the pastoral function. Epaphroditus is returning early and possibly under suspicion of having failed his commission (cf. v. 30 and the emphatic honors of v. 29). By folding his own relief into God’s mercy, Paul makes the man’s recovery a matter of apostolic interest, not congregational disappointment.

Built onbdf · obrien-nigtc · silva-becnt

λύπη is grief, pain of mind — the same root that yields ἀλυπότερος in the next verse, and the pair must be visible in English (‘grief … less grieved’). The construction ἐπί + accusative with a repeated noun is a Semitic-flavored idiom of accumulation, well attested in the LXX: Isa 30:1 ἁμαρτίαν ἐπὶ ἁμαρτίαν, Ezek 7:26 ἀπώλεια ἐπὶ ἀπώλειαν, Ps 68:28 LXX ἀνομίαν ἐπὶ τὴν ἀνομίαν αὐτῶν. The image is one layer set on top of another.

So ‘grief upon grief’ is not a stylistic flourish but a live metaphor of stacking (§8), and it presupposes grief already present. Paul does not say what the first layer is. The natural candidates from the immediate context are his imprisonment and uncertain outcome (1:20–24), the failure of others to seek Christ’s interests (2:21), and the Philippians’ own distress at the news (2:26). Reading the first λύπη as Epaphroditus’s illness and the second as his death is also possible.

Also worth naming: Paul admits to grief. A man who commands rejoicing throughout this letter (2:18; 3:1; 4:4) is not describing a life without sorrow. The rejoicing and the λύπη coexist, as in 2 Cor 6:10, ὡς λυπούμενοι ἀεὶ δὲ χαίροντες.

Built onbdag · muraoka-lxx · bdf · fee-nicnt

The combination is confirmatory-ascensive: γάρ grounds the statement in what precedes (the report of v. 26), and adjunctive καί intensifies it — ‘for indeed,’ ‘for in fact he really did.’ BDF treats καὶ γάρ as a strengthened γάρ where the following statement both explains and confirms.

Per §10, epistolary connectives are rendered, not dropped. ‘For indeed’ preserves both particles; a bare ‘for’ loses the confirmation that the rumor was accurate, which is the point of the clause. The Philippians had heard something; Paul is telling them it was worse than a rumor.

Built onbdf · bdag

The verse is stable in substance. Two variants deserve note, neither doctrinally consequential:

1. παραπλήσιον θανάτῳ (dative) vs. παραπλήσιον θάνατον (accusative). The dative has the earlier and broader support; the accusative in the Western tradition looks like an assimilation to a more familiar cognate-accusative pattern. NA28 prints the dative.
2. Some witnesses add ἐν before λύπην ἐπὶ λύπην or vary the second λύπην to a dative λύπῃ, smoothing the doubled accusative. The harder repeated accusative is original and is the reading that preserves the LXX accumulation idiom.

Metzger’s committee does not treat this verse in the Textual Commentary, which is itself an index of how little is at stake. Per §4, the variants are reported rather than silently resolved, but neither changes the sense of the verse.

Built onna28 · metzger-tc

Chrysostom (Hom. in Phil. 10) fastens on οὐκ αὐτὸν δὲ μόνον ἀλλὰ καὶ ἐμέ as evidence of Paul’s love: the apostle counts another man’s health as his own mercy. He also notes that Paul does not conceal his sorrow, and takes this as licence for Christians to grieve without shame — a reading well anchored in the text’s own admission of λύπη.

Theodoret understands ἵνα μὴ λύπην ἐπὶ λύπην σχῶ as referring to Paul’s bonds as the first grief, bereavement as the threatened second — the ‘imprisonment’ option in the contested note above has patristic pedigree, not just modern commentary.

The Latin tradition renders the key phrase usque ad mortem (‘even unto death’), which is idiomatic but slightly stronger than the Greek ‘near to death’; it also anticipates the actual μέχρι θανάτου of v. 30. Worth noticing, since Latin readers of this verse would have heard Epaphroditus and Christ (v. 8, usque ad mortem) in closer verbal parallel than the Greek permits.

Built onobrien-nigtc · silva-becnt

• ἠσθένησεν rendered ‘fell ill’ rather than the corpus default ‘was weak,’ context-forced and flagged (§4).
• παραπλήσιον θανάτῳ rendered ‘close to death’ / ‘very near to death,’ not ‘at the point of death,’ so that the boundary Epaphroditus reached stays distinguishable from the death Christ actually died (v. 8) — and so that v. 30’s μέχρι θανάτου ἤγγισεν can be rendered with its own vocabulary.
• ἠλέησεν rendered ‘had mercy on,’ keeping the LXX mercy-domain audible rather than ‘spared’ or ‘was gracious to.’
• λύπην ἐπὶ λύπην kept as ‘grief upon grief’ — the stacking image is live and must not be resolved to ‘still more sorrow’ (§8).
• καὶ γάρ rendered ‘for indeed,’ both particles surviving (§10).
• οὐκ … ἀλλὰ καί rendered ‘not … only but also/as well,’ additive rather than corrective.