A writer under the pseudonym of “Gregory MacDonald” has written an “evangelical” defense of universalism: The Evangelical Universalist (SPCK 2008).
The core appeal of universalism is emotional, and that’s one of his leading arguments. Even at that level, the emotional appeal of universalism is quite one-sided. We’d like to see our loved ones saved. But that doesn’t mean we feel the same way about Josef Mengele. I’ve addressed his emotional arguments at length.
In addition to the ad misericordiam fallacy, MacDonald also tries to cobble together an exegetical argument for universalism. Let’s review his major arguments.
In chapter 2, he tries to find a prooftext for universalism in Colossians. He says “the ‘all things’ that are reconciled in [Col 1:]20 are, without any doubt, the same ‘all things’ that are created in v16. In other words, every single created thing” (45).
But there are a couple of basic problems with this claim, even on his own grounds. For one thing, he goes on to say, in reference to 1:22, “clearly, the reconciliation spoken of here is the restoration of a harmonious relationship between the believers and God” (45).
Needless to say, that falls far short of “every single created thing.” To the contrary, it’s only a tiny subset of “every single created thing.” It excludes all inanimate things. Yet the “all things” that God created in v16 would include inanimate things. Hence, Paul’s sweeping language is hyperbolic.
I’d add that “pacification” is not the same thing as conversion. The Pax Romana didn’t make the subjugated races fall in love with their Roman overlords.
In addition, when commenting on Paul’s statement that the gospel has been proclaimed to every creature under heaven (1:23), MacDonald admits that “Of course, this is not a description of the actual state of affairs” (52).
But that’s another admission that, in Colossians, Paul uses hyperbolic expressions. His statement is apparently universal in scope, but in reality, that can’t be the case. So MacDonald is forced to concede that, in Pauline usage, universal expressions are not literally universalistic.
MacDonald also skips over 2:15. But here we have a classic taunt-song, where the victor humiliates his enemies. That’s a very different concept than saving your enemies.
As one commentator explains, “The general sense of the word is clear: it is a metaphor from the Roman Triumph, in which a victorious general led his troops through the city, with the spoils of war displayed for all to see and the defeated enemy paraded before his chariot,” R. McL. Wilson, Colossians & Philemon (T&T Clark 2005), 212-13.
That’s not a very good way of winning hearts and minds. That’s not how you befriend your enemies. Rather, that’s rubbing their nose in defeat through public humiliation. And, remember, this was a shame culture. Soldiers lived and died for honor, for reputation. So the disgrace would cut to the quick.
As such, the theme of cosmic reconciliation in chap. 1 does not imply universal salvation.
In chapter 3, MacDonald admits, when all is said and done, that the OT doesn’t teach universalism (72-73).
Earlier, he quoted Isa 45:23. But the sweeping language of v23 is immediately qualified by the v24.
MacDonald tries to get around this by saying that v23 refers to Israel, while v24 refers to the nations, and if we took v24 at face value, that would exclude the salvation of the Gentiles—contrary to Isaiah’s teaching elsewhere.
But all this demonstrates is his wooden handling of Isaian usage generally. Isaiah has oracles of salvation and judgment for Israel as well as the nations. Yet, when speaking of Israel, an oracle of salvation doesn’t mean that every Israelite will be delivered—just as an oracle of judgment doesn’t mean that every Israelite will be condemned or punished. By the same token, oracles of salvation and judgment for the nations do not apply to the same set of individuals.
For one thing, oracles of salvation and judgment often refer to discrete historic events like the Babylonian Exile and the post-exilic restoration. As such, they address the generation living at the time of the event. So you couldn’t have the same set of people. Due to mortality, there’s a turnover from one generation to the next. These oracles aren’t directed at the same set of people throughout time. By birth and death, the population undergoes continuous change.
Of course, you can say historical events prefigure eschatological events. But, in that case, you have to project both oracles of salvation and oracles of judgment into the eschaton. The division remains: some are finally saved while others are finally condemned.
MacDonald also draws attention to the note on which Isaiah ends. In 66:23, you have a passage which, taken by itself, looks like a prooftext for universalism. Yet that is immediately followed by v24, where a clear line of demarcation is drawn between the worshipers of v23 and the rebels of v24.
Hence, in two of MacDonald’s OT prooftexts, Isaiah uses hyperbolic language. So you can’t infer universalism from Isaiah’s sweeping expressions.
In chapter 4, MacDonald cites Rom 5:18-19 as a prooftext for universalism. But one of the problems with this interpretation is that, according to Paul, justification is contingent on faith. So justification only applies to believers.
MacDonald tries to get around this by saying that “Paul needs only to believe that one day all will believe (and I shall argue later that he did)” (80). But there are two problems with this appeal:
i) That would turn on postmortem conversion. But Romans doesn’t teach postmortem conversion. Hence, Romans qua Romans doesn’t teach universalism. Therefore, MacDonald can’t very well cite Rom 5:18-19 as a prooftext for universalism. You can’t get that from Romans, for Paul has a doctrine of sola fide in Romans. To isolate Rom 5:18-19 from justification by faith does great violence to the teaching of Romans. And there’s nothing in Romans to offset that delimitation.
ii) Instead, MacDonald will try to invoke other Pauline epistles to establish universal salvation. But there are two additional problems with that move:
a) His other Pauline prooftexts do not, in fact, teach universal salvation.
b) Even if they did, they don’t teach postmortem conversion. Universal salvation is not equivalent to postmortem conversion. MacDonald needs a specific, Pauline prooftext for postmortem conversion if he’s going to use Rom 5:18-19 to prove universal justification consistent with sola fide.
MacDonald then discusses the use of “all” in Rom 5:18 (81ff.). One reason for the repetition of “all” is that the parallel structure, of itself, invites the use of parallel terminology. That’s a way of creating and reinforcing a parallel. You repeat certain catchwords. So the repetition of terms is partly rhetorical: a literary device.
Paul is setting up an analogy between Adam and Christ. Between Adam’s deed and Christ’s, as well as their respective consequences. A is to B as C is to D. Adam is to Adamites as Christ is to Christians. Each set is exhaustive within its domain, but that doesn’t make one domain conterminous with another domain.
At the same time, he’s also setting up a disanalogy. Comparison and contrast. The headship of Adam over against the headship of Christ. There are discontinuities as well as continuities.
And that, too, is part of the rhetorical structure. It’s easier to see the dissimilarities once you lay out the similarities.
Another concern, lying close to the surface, is to emphasize the ethnically inclusive character of the Gospel. God justifies Jews and Gentiles like—on condition of faith.
MacDonald then cites 1 Cor 15:22 as another prooftext for universalism. But there are several problems with that appeal:
i) V23 is epexegetical, delimiting the scope of v22 to Christians.
ii) Paul himself, in this very passage, points out that universal expressions in Scripture can be hyperbolic (v27; cf. Ps 8:6).
iii) In v26, Paul says that God will “destroy” (i.e. overthrow, dethrone) his enemies. That’s the language of conquest, not conversion. MacDonald tries to evade this by limiting the reference to fallen angels. But there are two problems with that move:
a) It could be a reference to human authorities, or human and angelic enemies alike
b) In any case, MacDonald’s universalism extends to fallen angels as well as fallen men. That’s how he construes Col 1:15-20.
MacDonald then cites Rom 11:26. Of course, that’s is a very controversial passage. For now I’d make the following observation: In OT usage, “all Israel” is an idiomatic phrase. It’s not just a case of adding a universal quantifier to a free-floating noun. “All Israel” doesn’t mean “all Israelites.” If you study OT usage, you’ll see this phrase is used in a representative sense, such as chieftains who stand for their respective clans.
MacDonald quotes Richard Bell’s claim that it would be “unthinkable that an Israelite could be excluded from final salvation” (96). Of course, it’s that sort of spiritual presumption that John the Baptist rails against in Mt 3.
Finally, he appeals to Phil 2:10-11. But here he makes no effort to interact with the detail exegesis of O’Brien in The Epistle to the Philippians (Eerdmans 1991), 243-48.
On p102, he tries to dispose of the Reformed doctrine of special redemption in one paragraph. That isn’t a serious attempt to rebut Calvinism.
In chapter 5, MacDonald appeals to Revelation to prove universalism. The bulk of his argument turns on the conversion of the nations (15:4; 21:24,26; 22:2). But this reiterates his wooden handling of OT prophecies.
In reference to the nations, John alternates between oracles of salvation in judgment. In that respect he repeats the pattern of the OT prophets, to whom he’s indebted.
This only creates a tension if you treat “the nations” as a monolithic unit. But “the nations” is just a conventional synonym for gentiles. And there’s no reason to treat every reference to the salvation or damnation of gentiles as denoting a numerically identical people-group. That would be quite unhistorical.
All men don’t live and die at the same time. The enemies of the church aren’t identical from one decade to another.
And Revelation isn’t limited to one-time events. It’s partly concerned with endtime events, but it’s also concerned with a perennial battle between God and his people, on the one hand, over against the enemies of God—throughout OT history and the church age. Both history and eschatology figure in revelation. Past, present, and future.
In Revelation, its philosophy of history is both linear and cyclical. The church age will come to an end. There will be a final judgment. But every Christian generation may have to face persecution.
John uses this designation (“the nations”) because it’s a traditional, OT designation, and he’s incorporating OT oracles of salvation and judgment into his own prophecy. So he alternates between the two, just as OT prophets alternate between the two—depending on the time and place. OT oracles don’t have a uniform referent. That’s historically variable. Israel didn’t have the same enemies from one decade to the next. And the pattern repeats itself in church history.
Unless MacDonald is either a pure preterist or a pure futurist, he can’t limit all the action in Revelation to a one-time event—with the same cast of characters. The immortal actors remain in place (the Trinity, angels, devil, demons), but not the mortal actors (human beings).
MacDonald also appeals to the description of the 144,000 as the “first-fruits” to argue that “the nations” represent the rest of the harvest (188).
i) However, that doesn’t necessarily follow. As Beale points out, the first-fruits were dedicated to God. So that can imply a separation between the first-fruits and the remainder of the harvest, which is profane. Cf. G. Beale, The Book of Revelation (Eerdmans 1994), 744.
ii) In addition, MacDonald’s argument assumes that the 144,000 are a subset of the redeemed. But that, too, is debatable. Cf. Beale, ibid., 416-23.
MacDonald then says of 22:17, “But to whom does she speak? One plausible audience is those in the lake of fire (after all, who else is there?)” (119).
How that’s plausible, he doesn’t say. It would only be plausible to a universalist. Commentators generally identify Christ as the target of this invitation.
MacDonald misses another point. The invocation in 22:17 isn’t limited to an end-time speaker. It applies throughout the church age.
This coda (22:6-21) is not a part of the narrative (4:1-22:2). It lies outside the narrative. It doesn’t come at the end of the chronological sequence, as a final event within the chronological sequence.
MacDonald says “this universal vision of salvation is confirmed again by the proleptic vision in 5:13” (119).
Actually, it’s just a literary antithesis. As Aune points out, “This is a verbal repetition of 5:3, where no one in the entire universe was able to open the scroll,” Revelation 1-5 (Word 1997), 366. So it’s a literary device: from “no one” to “everyone.”
This is another example of MacDonald’s wooden exegesis. He’s a heretical version of Tim LaHaye.
MacDonald must also attempt to defang the verses in Revelation which speak of eternal damnation. One move is to claim that “the expression literally means ‘unto the ages of ages’ (not ‘forever and ever’)” (128).
How he arrives at that conclusion he doesn’t say. It’s as if he were merely transliterating the word—then inferring the sense from the transliterated term: aion>eon>age.
But words often have an idiomatic meaning, conferred on them by popular usage. It’s not as if the literal meaning is the real meaning. Rather, meaning is assigned by linguistic convention.
There are passages in Scripture which employ a two-age schema: this world and the world to come. However, that doesn’t work for Rev 14:11 and 20:10, where the world to come is the only world in view. Otherwise, you’d have a world to come after the world to come.
I also notice, both here and elsewhere (e.g. Mt 25:46), that MacDonald doesn’t quote from any standard lexicons to support his semantic claims. Instead, he quotes Edward Fudge, who is—of course—an annihilationist.
He then runs through a number of “possibilities”: “First, argue that John was simply adopting stereotypical descriptions of the postmortem life, which formally contradict 21-22 but which are subverted by 21-22 and are thus not intended to be taken strictly literally” (128-28).
There are two problems with this move:
i) The point of exegesis is to offer the best interpretation, not to deflect attention away from the best interpretation by compiling a list of barely possible interpretations.
ii) His logic is reversible. Assuming a formal contradiction, why should we take 21-22 literally? Why should we conform the other descriptions to 21-22 and not vice versa?
Another move is to depersonalize the damned: “the focus of 20:10 is the utter defeat of the systems and not the individuals” (129).
But “systems” are composed of persons, from top to bottom. And no one punishes a “system.”
“One could maintain, as some recent theologians have, that the devil is not a personal being but something more akin to a personification of evil” (130).
I see. So, when Jesus was tempted by the Devil in the wilderness (Mt 4), he was tempted by a personification. Yeah, that makes a lot of sense.
To depersonalize the devil, you have to demythologize the Bible. And once you start down that path, you should demythologize the Book of Revelation. To consistently implement that program, you have to secularize the Book of Revelation. You wouldn’t end up with universalism—even if Revelation taught universalism. For that teaching would be part and parcel of an antiquated worldview: the very thing we need to demythologize.
On this view, Revelation wouldn’t be about the world to come. About God and heaven and angels and demons. No, it would be about this world. A metaphor for the immanent, recurrent battle between good and evil here and now. There is no hereafter. So this move is fatal to MacDonald’s thesis.
“One could maintain that the devil will be punished forever, but that Lucifer will ultimately be saved…The devil, like the ‘flesh,’ must be destroyed…But he dies, and Lucifer is reborn as a redeemed angel. It would still be possible to speak of the devil being tormented forever and ever to symbolize this defeat even though no actual being is still in the lake of fire” (131).
Sounds to me like MacDonald has been reading too much Alister Crowley or Anton LaVey.
“It ought to be noted that a debate has arisen within recent Gospels scholarship about whether Jesus actually spoke of punishment in the afterlife at all…According to Wright all the passages that warn of the fires of Gehenna speak not of any postmortem punishment but of the premortem events of AD 70 when Jerusalem was destroyed” (141).
One problem with this move is that it would also apply to whatever Jesus said about salvation in the afterlife. The Gospels can’t teach universal salvation if they don’t teach eschatological salvation.
And what about apocalyptic imagery in the other NT writings? MacDonald can only undercut some of the prooftexts for everlasting punishment by undercutting some of his prooftexts for universal salvation in the process.
Indeed, MacDonald goes on to mention that Andrew Perriman applies Wright’s approach to the rest of the NT. But that’s a double-edged sword. It isn’t limited to prooftexts for postmortem judgment.
“The strongest argument against a universalist interpretation of Jesus’ teaching starts by arguing that any adequate interpretation of Jesus’ words about final punishment must begin by reading them against the background of beliefs held by his contemporaries. Second Temple Jewish beliefs on the postmortem fate of those outside salvation are not at all uniform…However, none of them expected any kind of universal salvation. Thus, when Jesus spoke about the fires of Gehenna, almost everyone who was listening to him would interpret his words as a reference to the final state of the lost. Few, if any, of Jesus contemporary listeners would have understood his words as leaving any room for hope for those who find themselves in Gehenna…I think that it is quite clear that Jesus’ contemporaries would not have thought that he was a universalist of any variety” (144-45).
Sounds good to me.
In reference to Mt 25:46, he says, “the translation of aionios ahs been the subject of numerous studies in recent years, but there seems to be a strong case for maintaining that it means ‘pertaining to an age’ and often refers not just to any age but to ‘the age to come’” (147).
He doesn’t cite any lexicons to corroborate claim. For example, the entries on aion and aionios in the Exegetical Dictionary of the New Testament (1:44-46) contradict his blanket claims.
“Thus ‘eternal life’ may be better translated as ‘the life of the age to come’ and ‘eternal punishment’ as ‘the punishment of the age to come’” (147-48).
Suppose, for the sake of argument, that we play along with this rendering. When, in cursing the fig tree, Jesus says “May no fruit ever come from you unto the age [eis ton aiona] (Mt 21:19),” does he mean the fig tree will bear no fruit in the present age, but it will bear fruit in the age to come?
Likewise, when Jesus says that “whoever blasphemes the Holy Spirit has no forgiveness unto the age [eis ton aiona] (Mk 3:29a), does he mean the culprit is unforgivable in this age, but will be forgiven in the age to come? Not according to the Matthean parallel (Mt 12:32).
He then quotes another author’s claim that “the point is not that the fire will burn forever, or the punishment extend forever, or that the life continue forever, but rather than all three will serve to establish the rule of God” (148).
But assuming, for the sake of argument, that this is correct, then it doesn’t teach either everlasting life for the sheep or everlasting punishment for the goats. How does saying, “the point is not that the life continue forever” lend any support to universalism? How would that distinguish between universal salvation and universal annihilation?
“Any interpretation of Gehenna must be compatible with the claim that God is love and would never act in a way towards a person that was to ultimately compatible with what is best for that person” (148).
Of course, that merely begs the question in favor of universalism. MacDonald has abandoned exegesis, and is now insisting that Mt 25:46 can’t teach everlasting punishment since that would conflict with…universalism. Isn’t that a wee bit circular? The fact that MacDonald feels the need to make this last-ditch appeal betrays the weakness of his case against the traditional reading of Mt 25:46.
“The verse [Mk 9:49] ahs long perplexed commentators, but it seems to indicate that the firs of Gehenna function as a place of purification” (150).
Although he refers to commentators, he doesn’t cite any commentators. None of the commentators on Mark that I’ve consulted (Cranfield, Edwards, Evans, France, Gundry, Lane) construe the verse the way he does.
“In Romans 9 we saw the division within Israel between those ‘elect according to grace’ and those who are ‘objects of his wrath fitted for destruction.’ This division looks very final, but Romans 11 demonstrates it to be temporary. This serves as a warning to those who move too quickly from Paul’s claims about an apparently final division between the lost and saved to a traditional doctrine of hell” (151).
Of course, I don’t concede his interpretation of Rom 9-11. I favor the exegesis of Murray, Piper, and Schreiner. MacDonald doesn’t engage their exegesis.
So far from Rom 9-11 undermining the traditional doctrine of hell, it underwrites the traditional doctrine of hell.
MacDonald then tries to neutralize 2 Thes 1:6-10. In the process, we’re treated to such gems as: “Were one able to sit down with Paul and discuss the issue with him, he would agree that the [universalistic] qualifications did bring out the fuller dimensions of his theology, even though he never had them in mind when he wrote. Talbott could answer the question, ‘Would Paul agree with your interpretation?' with the reply, ‘He would if I had an hour to discuss it with him’” (154).
Of course, this isn’t exegesis. It’s the abdication of exegesis.
But even if, for the sake of argument, we indulge in this imaginary scenario, then that exercise isn’t limited to the universalist. Suppose, ex hypothesi, that MacDonald’s Pauline prooftexts apparently teach universal salvation. But I could counter this impression by claiming that if I had hour to sit down with Paul and talk it over, he’d agree with my qualifications.
“One could interpret the Book of Life in a predestinarian sense: God, before the foundation of the world, chooses whom he will save and records their names. This could be supported by 17:8 and 13:8. If that is correct, then, as Beale notes, universalism will have a problem in Revelation; for the universalist needs a Book of Life with flexible contents—one in which names can be deleted and, more importantly, added…If the context of the book is fixed before creation, then this is impossible” (192).
MacDonald then tries to evade the force of this argument by appealing to Rev 3:5. But there are several problems with that line of argument:
i) MacDonald mentions Beale in passing, but completely disregards his Calvinistic interpretation of Rev 3:5. Cf. G. Beale, The Book of Revelation, 278-82.
ii) Rev 3:5 is, of course, a standard Arminian prooftext to undermine the Reformed doctrine of perseverance. But there’s no reason in the world why Rev 3:5 would also be a prooftext for universalism. In Arminianism, true believers can lose their salvation; in universalism, everyone will be saved. Christian apostasy has no logical place in universalism.
iii) If universalism were true, then why would the names be penciled in? If universalism is true, then logically, everybody’s name ought to be inscribed with indelible ink before the foundation of the world. No names ought to be added or erased.
MacDonald appeals to other Arminian prooftexts like Heb 6. But this assumes the Arminian exegesis of Heb 6. Moreover, even if the Arminian interpretation were true, that’s hardly an argument for universalism. Just the opposite.
MacDonald also has an appendix on Ephesians in which he tries to extract universalism from Eph 1:10 and 1:22. To some extent, 1:22 unpacks the content of 1:10. However, there’s nothing in the terminology of 1:22 that implies universalism.
As Hoehner points out, “The metaphorical language ‘under his feet’ has the idea of victory over enemies. It is used of the winner of a duel who places his foot on the neck of his enemy who has been thrown to the ground, like Joshua who had his generals place their feet on the necks of the five defeated Amorite kings (Josh 10:24; cf. 2 Sam 22:39). Similarly, everything is subjected under Christ’s feet, meaning that everything is currently under his control, both friends and enemies,” Ephesians: An Exegetical Commentary (Baker 2003), 283-84.
In sum, MacDonald’s exegetical argument for universalism is no more successful than his sentimental argument for universalism.