Why Translations Differ

The last two lessons were about the Greek text: which words the apostles wrote. This lesson is about something that causes far more of the differences you will meet between good translations, and that is what to do with Greek words that everybody agrees on. At Romans 3:25, for example, the editions differ in only one small way. The SBL Greek New Testament (SBLGNT) reads διὰ πίστεως (dia pisteōs, through faith), and the Tyndale House edition reads διὰ τῆς πίστεως, with the article, which no translation would render differently. Yet the English translations of this verse differ sharply, and they differ over a word whose spelling nobody disputes.

The Hermeneutics lesson Translation Is Interpretation, in the module called The Toolbox, has already taught the spectrum from formal to functional translation and how to compare translations without knowing Greek. Do not expect that here. What this lesson adds is the view from the Greek side: when two translations differ, you can now usually say which of five kinds of decision lies behind the difference.

The text. Sometimes translations differ because they translate different Greek, as at Mark 1:1 or the end of Mark. You learned to check that in the last two lessons, and it should always be the first thing you rule out.

The word. A Greek word has a range of meanings, and English rarely has one word with the same range. Hilastērion in Romans 3:25 is the classic case, and it is the subject of sections 2 and 3.

The grammar. Greek constructions can be ambiguous in ways English cannot keep. The genitive in πίστις Χριστοῦ (pistis Christou, faith of Christ), which the module on nouns and cases weighed in its lesson on the genitive, is the best-known example: faith in Christ or the faithfulness of Christ.

The punctuation. The oldest manuscripts have almost no punctuation, so every comma and full stop in a Greek edition, and therefore in every translation, is an editor's decision. At Romans 9:5 the Greek runs ὁ χριστὸς τὸ κατὰ σάρκα, ὁ ὢν ἐπὶ πάντων, θεὸς εὐλογητὸς εἰς τοὺς αἰῶνας (ho christos to kata sarka, ho ōn epi pantōn, theos eulogētos eis tous aiōnas, the Christ according to the flesh, the one being over all, God blessed for ever). With a comma after flesh, Christ is called God over all, as in the BSB, 'Christ, who is God over all'. With a full stop, the last words become a separate blessing of God, which is how the RSV printed it. The Greek letters are the same either way.

The register. Here belong the decisions about how formal or plain the English should be, how to render idioms, and how to render Greek words that are grammatically masculine but refer to men and women together. That last question is the subject of section 4.

Learning to name the kind of decision is worth more than it looks, because the kinds call for different responses. A difference of text is settled, so far as it can be, by manuscripts. A difference of word or grammar is settled, so far as it can be, by usage and context. A difference of register is often not a disagreement about meaning at all, but about the English reader.

Here is the clause in the SBLGNT: ὃν προέθετο ὁ θεὸς ἱλαστήριον (hon proetheto ho theos hilastērion), whom God presented as a hilastērion. Parse it. Ὅν (hon, whom) is the relative pronoun, accusative masculine singular, referring back to Christ Jesus in 3:24. Προέθετο (proetheto) is aorist middle indicative, third person singular, of προτίθημι (protithēmi), which some tools list under its middle form προτίθεμαι (protithemai), a verb that can mean to set something before the public, to display, or to set something before oneself, to purpose. Paul uses the same verb in Romans 1:13, where the BSB renders it 'I planned'. Most translations take it here in the sense of public display, and the BSB's 'presented' leans that way, but you should know that the sense God purposed him as a hilastērion has defenders. Ἱλαστήριον (hilastērion) is accusative neuter singular, standing as a second accusative alongside ὅν: God presented him as this thing. That is how most grammarians and the free tools take it. A minority read it instead as a masculine adjective agreeing with ὅν, presented him as propitiatory, since the masculine accusative would be spelled the same way.

Now the word itself. The inquiry sent you to its family, and the data are worth setting out in one place. The noun ἱλαστήριον appears in the New Testament only here and in Hebrews 9:5, where the cherubim are 'overshadowing the mercy seat', τὸ ἱλαστήριον (to hilastērion). The noun ἱλασμός (hilasmos) appears in 1 John 2:2 and 4:10, where Christ is the hilasmos 'for our sins', and the BSB renders both 'atoning sacrifice'. The verb ἱλάσκομαι (hilaskomai) appears in Hebrews 2:17, εἰς τὸ ἱλάσκεσθαι τὰς ἁμαρτίας τοῦ λαοῦ (eis to hilaskesthai tas hamartias tou laou, to make atonement for the sins of the people), with sins as its object, and in Luke 18:13, where the tax collector prays ὁ θεός, ἱλάσθητί μοι (ho theos, hilasthēti moi), which the BSB renders 'God, have mercy on me'. Ἱλάσθητι is aorist passive imperative, second person singular. The BSB's footnotes offer 'as a propitiation' at Romans 3:25 and 1 John 4:10, so even the house translation of this course tells its readers the word is disputed.

The Greek Old Testament adds the decisive background. In the Septuagint, ἱλαστήριον is the regular word for the golden lid of the ark, the kapporet, in Exodus 25 and Leviticus 16, the mercy seat you read about in the inquiry. Outside the tabernacle, the word appears in a Jewish book of the period, 4 Maccabees, which speaks of the death of martyrs in the persecution under Antiochus as atoning for the nation, and there it has nothing to do with the ark. So the word could name the place, and could also describe a death as atoning.

This course has already told the history of the scholarly argument twice. The Hermeneutics lesson Word Studies Done Properly sets out the debate between C. H. Dodd, who in a 1931 article and in The Bible and the Greeks (1935) argued for expiation, and Leon Morris, who in The Apostolic Preaching of the Cross (1955) argued for propitiation. The Systematic Theology lesson The Work of Christ shows what follows from the debate for the doctrine of the atonement, which is less than partisans on either side suggest. Go back to them for the history. What this lesson does is put the three main readings side by side as translation choices, each with the Greek evidence its defenders rely on, so that you can see why careful committees have printed four different English words.

Propitiation. On this reading the hilastērion is the means by which God's wrath against sin is turned away. Its defenders, of whom Leon Morris and Roger Nicole in the 1950s were the most influential, begin with the argument of the letter. Romans 1:18 opens the case with 'The wrath of God is being revealed from heaven', and 2:5 speaks of 'the day of wrath'. Three chapters establish that the whole world stands under that wrath before 3:21 announces the remedy, and 5:9 then says that we shall be 'saved from wrath through Him'. A word in 3:25 that dealt only with sin and not with wrath, they argue, would leave the problem of the first three chapters unanswered. They point to Luke 18:13, where the verb is addressed to God himself, and to the ordinary meaning of the word group in Greek outside the Bible, and in some Jewish writers, which is to turn away anger. And they answer the charge that propitiation pictures humans bribing an angry god by pointing to the text itself: here it is God who presents the propitiation, out of love, as 1 John 4:10 says. The KJV and ASV print 'propitiation', and the NASB keeps it.

Expiation. On this reading the hilastērion is the means by which sin itself is removed, cancelled or cleansed. Its defenders, beginning with C. H. Dodd, begin with the Greek Old Testament, where the verbs of this family very often have sin or defilement as their object and God as their subject, so that God is the one who deals with sin, not the one who is dealt with. Hebrews 2:17, with 'the sins of the people' as the object of the verb, and 1 John 2:2, 'for our sins', fit that pattern. Romans 3:25 itself has God as the subject who presents. Defenders of this reading do not deny God's wrath; many hold that wrath in Paul is God's settled opposition to sin, and that when sin is removed the cause of wrath is removed with it. They argue that the English word propitiation, to a modern ear, carries pagan overtones of an angry deity being placated, which Paul did not intend. The RSV of 1946 printed 'expiation'.

Mercy seat. On this reading the hilastērion in Romans 3:25 is what it is in Hebrews 9:5 and throughout the Greek Old Testament: the mercy seat, the place of atonement. Its defenders include Martin Luther, whose German Bible rendered the word Gnadenstuhl, mercy seat, and in recent scholarship Daniel Bailey's Cambridge doctoral study. They argue that the word's settled biblical meaning should be the starting point, that 'in His blood' calls up exactly the sprinkling of blood on the mercy seat in Leviticus 16, and that προέθετο in the sense of public display makes a deliberate contrast: the mercy seat was hidden behind the veil and seen by one man once a year, but God has set Christ forth in public. On this view Paul is saying that Christ is the place where God meets sinners and where atonement is made, which can contain both the turning away of wrath and the cleansing of sin. The NET Bible prints 'the mercy seat', and so does the CSB, with a footnote offering propitiation and place of atonement.

Each reading has a serious objection, and each has an answer. Against propitiation, defenders of expiation point to the many places where the word group takes sin as its object. Morris and Nicole answered that their critics had read the evidence selectively, and that the context in Romans settles what the word is doing here. Against expiation, defenders of propitiation point to the wrath that frames the whole argument of Romans. Dodd's followers answer that removing sin removes the cause of wrath, so that nothing is left unanswered. Against the mercy seat, critics note that Romans 3:25 has no article, unlike Hebrews 9:5 and most of the Septuagint uses; that it seems a strange image to make Christ both the place of sprinkling and the blood sprinkled on it; and that Paul does not use tabernacle imagery elsewhere in Romans. Its defenders answer that a noun used as a complement, as here, often stands without the article, that Paul's images regularly overlap, and that a letter so soaked in the Old Testament assumes readers who knew Leviticus.

And then there is the rendering in front of you. The BSB's 'atoning sacrifice', like the rendering sacrifice of atonement in the NIV and the NRSV, chooses a phrase broad enough to sit with more than one of these readings, and adds 'Or as a propitiation' in the margin. Its defenders call that wise restraint, a translation declining to settle what scholars have not settled. Its critics call it a loss, since it makes the English reader unable to see either the specific claim about wrath or the specific echo of the mercy seat. Notice that this, too, is a translation choice with a case on each side.

What the Greek settles and what it does not: it settles that Romans 3:25 belongs to the language of Israel's atonement, that the word family is used both of God being merciful and of sins being dealt with, and that God is the one who acts. It does not settle by itself which of the three readings Paul intended, because that depends on how one weighs the word's Old Testament home against the flow of Paul's argument. Careful scholars who share the same faith, the same Greek and the same commitment to scripture come out in different places. The honest preacher says so.

Start with what is agreed, because a great deal is. In Greek, ἄνθρωπος (anthrōpos) is normally the word for a human being, and its plural ἄνθρωποι means people, men and women together, unless the context narrows it. A different word, ἀνήρ (anēr), is used for a male or a husband. The plural ἀδελφοί (adelphoi), grammatically masculine, was regularly used of a group that included women, as the inquiry showed you at Philippians 4:1-3, where the brothers include Euodia and Syntyche. And Greek, like older English, uses masculine grammatical forms for a mixed group. None of that is in dispute among translators. The dispute is about English: whether words like man, men and brothers are still heard by ordinary English readers as including women, and what a translation should do if they are not.

The case for what its defenders call gender-accurate translation runs like this. A translation must carry meaning, not grammatical form. If a congregation in which most of the hearers are women hears 'brothers' and thinks Paul is addressing the men, the translation has misled them, however literal it looks. English has changed: many readers, especially younger ones and those for whom English is a second language, hear man as male. So where the original refers to men and women together, the English should too: 'brothers and sisters' for ἀδελφοί when the context shows that women are included, 'people' or 'everyone' for ἄνθρωποι. D. A. Carson made this case in The Inclusive-Language Debate (1998), and Mark Strauss in Distorting Scripture? (1998). The NIV in its 2011 revision and the NRSV render the address in Romans 12:1 as brothers and sisters. And at 2 Corinthians 4:2, which the next lesson studies, the CSB, the NIV and the NET all render πᾶσαν συνείδησιν ἀνθρώπων with the word everyone, where the BSB and the KJV keep every man's.

The case for keeping the traditional forms runs like this. A translation should be as transparent to the original as English allows, and the original uses a masculine word. 'Brothers' with a footnote, as some translations print it, tells the reader the truth about both the Greek and the meaning, while 'brothers and sisters' adds words that are not there. More importantly, changing singulars to plurals or generic man to people can lose real features of the text. Psalm 1:1, 'Blessed is the man who does not walk in the counsel of the wicked', pictures one representative man, a picture the church has long read as fulfilled in Christ, and a plural rendering loses it. Psalm 8:4, 'what is man that You are mindful of him, or the son of man that You care for him?', is quoted in Hebrews 2:6 and applied to Jesus in the verses that follow; a translation that renders it with human beings and mortals, as the NRSV does, hides the link. Vern Poythress and Wayne Grudem made this case in The Gender-Neutral Bible Controversy (2000), and the Colorado Springs Guidelines of 1997 set out principles along these lines. The BSB generally keeps brothers and man, as the KJV and the NASB of 1995 kept brethren and man. Defenders of gender-accurate renderings reply that Hebrews 2:8 first applies the psalm to humanity in general, before 2:9 turns to Jesus, and that a footnote can keep such links visible without making ordinary readers hear men where the writer meant everyone.

Notice what both sides agree on, because it is often misreported. No major evangelical translation on either side changes the masculine language used of God, the Father, or of Jesus, the Son, and none changes language that refers to particular men. The debate concerns generic language, references to people in general or to mixed groups, and it is a debate about how English works today as much as about Greek. Both sides want the reader to hear what the original reader heard. They disagree about what English readers now hear.

One verse shows how a translation decision and a gender question can meet. In Romans 16:7 Paul greets Ἀνδρόνικον καὶ Ἰουνίαν (Andronikon kai Iounian, Andronicus and Junia). With the accent the SBLGNT prints, the second name is the woman's name Junia, a common Latin name; with a different accent it could be a shortened man's name, Junias, but no example of that man's name has been found in ancient sources, and early commentators such as John Chrysostom took her to be a woman. Most translations now read Junia. The pair are then called ἐπίσημοι ἐν τοῖς ἀποστόλοις (episēmoi en tois apostolois). The BSB renders that 'outstanding among the apostles', as the NIV does, and the KJV has 'of note among the apostles'. The NET renders it instead 'well known to the apostles', and the CSB 'noteworthy in the eyes of the apostles', each with a footnote giving 'among'. Michael Burer and Daniel Wallace argued for that sense in an article in New Testament Studies in 2001; Richard Bauckham in Gospel Women (2002), Eldon Epp in Junia: The First Woman Apostle (2005) and others replied that the Greek more naturally means 'among'; and Burer answered them with a much larger set of Greek parallels in the Journal of the Evangelical Theological Society in 2015. Both sides are still arguing from the same evidence. And even the reading 'among' leaves open what 'apostles' means here, since the word was used both of the Twelve and of envoys sent out by churches. What follows for women in ministry is weighed in the Homiletics lesson Who May Preach, and this lesson will not settle it. The point here is narrower: an accent mark and a preposition have carried a large debate, and you can now see both on your phone.

When you meet a hard translation choice in a passage you are preaching, you have three honest options and one dishonest one.

The first honest option is to preach what is common to all the good renderings. At Romans 3:25 every reading agrees that God himself acted, that the cross belongs to the language of Israel's atonement, that Christ's blood deals with sin, and that God's justice is shown, not set aside. A sermon that preaches those things preaches the text, whichever reading is right.

The second honest option is to name the choice briefly. A sentence or two will do: the word Paul uses here is the word the Greek Old Testament used for the mercy seat; some translations say propitiation, some say atoning sacrifice, and here is why it matters. Ordinary believers can follow that perfectly well, and they will trust you more, not less, for being shown the workings.

The third honest option is to argue for one reading, openly, as your reading. You are entitled to a view, and you should have one on a word this important. Say what the other readings are, give their best evidence, and then say why you are persuaded otherwise.

The dishonest option is the one you have often heard: the Greek really means. Said of a word on which the best scholars disagree, that phrase claims an authority the speaker does not have and hides the evidence from the very people who most need to weigh it. It also teaches a congregation that their Bible cannot be trusted without a Greek expert to tell them what it secretly says, which is the opposite of what this track is for.

Open your notebook to the pages for your capstone passage and start a new page headed Translation. On your phone, read your passage in the BSB and in at least two other translations, ideally one more formal and one more functional, using Bible Gateway, which lets you set versions side by side.

Mark every place where the translations differ in a way that changes the meaning, not merely the wording. For each one, find the Greek word or phrase behind it on STEP Bible or Blue Letter Bible, and decide which of the five kinds of decision from section 1 lies behind the difference: text, word, grammar, punctuation or register. If it is text, go back to your Text and Weighing pages.

Choose the one difference that matters most for preaching. Write the renderings of each translation, then write two or three lines for each giving the best case its defenders would make, using the Greek evidence you can see: the word's other uses, the grammar, the context. Do this for the rendering you dislike as carefully as for the one you prefer.

Finally, write the sentence you would say from the pulpit about this choice, using one of the three honest options in section 5, and check it for the phrase the Greek really means. If it is there, rewrite the sentence.