Typically when we encounter passages in Scripture that seem to stand in tension, we instinctively try to find a way to alleviate that tension. Nowhere is this more evident than in relation to the tension we feel between the texts of Scripture that seem to indicate the universal extent of God’s saving will and those that assert the sobering reality of the eternal damnation of all who, for whatever reason, die without placing their faith in Christ through the preaching of the gospel. Usually attempts to resolve this difficulty end up affirming a version of limited atonement (Christ died ultimately to save the elect alone) or universal possibility (Christ died to save all but not all will freely believe) or universalism (Christ died for all and so all will be saved).
While they are all logically coherent, it seems to me that each of these positions, in one way or another, downplays certain texts at the expense of others. To all of us who wrestle with these issues, Bruce McCormack offers some wise words of counsel in an essay he contributed to the book Karl Barth and American Evangelicalism. Although McCormack does not offer us a tidy solution, he does encourage us to stay rigorously faithful to the whole range of biblical teaching, giving equal weight to all of its parts. He writes:
Conspicuous by its absence from Paul’s theology is any mention of “hell.” One might well think that the talk of the wrath and fury, tribulation and distress which awaits those who do evil in Rom. 2:8-9 is equivalent to “hell” but Paul never uses that word. The concept of “hell” does play a sizable role in the teaching of Jesus and in the Book of Revelation, of course. Taken by itself, this difference between Jesus and Paul would not be a problem. The two views could easily be harmonized simply by regarding what Jesus has to say as a further expansion upon what Paul knows to say. The problem, however, is that Paul is committed to a universal atonement [e.g. 1 Tim. 2:3-6] – as well as the understanding that faith in Jesus Christ is effected in human beings by God’s grace alone. And the combination of these two elements creates a difficulty of no small proportions. For if grace is irresistible, if faith is God’s to give as He wills and Christ died for all, then, logically, God’s will ought to be to give the gift to all and universal salvation should be the result. Alternatively, we could take up our starting point in the “two-group” eschatologies of Jesus and the Seer and look back through the lens provided by the Pauline understanding of the relation of grace and faith and the only logical option would be to affirm a limited atonement. Universal salvation on the one side and limited atonement on the other; those are the only two logical possibilities which arise on the soil of the Pauline understanding of faith as a sovereign work of God. And because Jesus’ teaching on hell, especially, was taken to be the fixed pole, Reformed theology in its orthodox expressions always concluded to a limited atonement. The net effect of that decision was, of course, that Paul’s commitment to a universal atonement had to be negotiated out of existence.
I would suggest that there is a better way of dealing with this, the most profound and important of the tensions found in the New Testament. I am certainly conservative enough in my understanding of biblical inspiration to believe that if something appears in the New Testament, it is there because God wanted it there. So if a tension exists, there must be a reason for it. And if I had to guess, I would say that the reason has to do with the fact that those awakened to faith in Jesus Christ in this world are still sinners. If God told us the answer to the problem in advance of the eschaton, we would harm ourselves on one side or the other. If He were to tell us that a universal salvation will be the final outcome, we would very likely become lax, antinomian even. The sense of urgency that is pervasive in Paul’s Christian existentialism would be lost. If, on the other hand, God told us that limited atonement is the true resolution of the tension, we would very likely despair of our salvation. How could anyone be certain that the atoning death of Christ was really intended for him or her? And so I would venture to guess that the tension I have described is divinely intended – in order to protect us from ourselves.
In short, I think it was a mistake for the Westminster Assembly to seek to resolve this question on the side of limited atonement in advance of the return of Christ in glory – just as I think that it would be a mistake for any church today to teach universalism. Again, these are simply the logical possibilities that arise on the soil of the Reformed understanding of the relation of grace and faith. As such, they constitute the walls within which we are to live in this world. All of us will tilt more to one side than the other. And if individual theologians wish to conclude to one or the other – for the sake of exploring implications and relationships among the various Christian doctrines, they should be allowed to do so. That belongs to their unique calling. But churches need to be responsible for all the faithful. And for that reason, I would say, neither limited atonement nor universalism should ever be made church dogma.
As I mentioned before, McCormack does not offer any simple solutions or clean ways of reducing certain tensions in the full scope of biblical teaching, especially those in relation to human salvation. But he does offer sage advice, namely, that however difficult or uncomfortable, we must allow Scripture to dictate to us how we should think, reason, theologize, preach, and evangelize, even when we cannot understand how all of the pieces fit together. Indeed, we see in a glass darkly, and we must walk by faith rather than by sight. Therefore, let God’s Word be God’s Word, and may we bow all of our logic and systematic categories in humble submission to it.
 McCormack, B.L., 2011. ‘So That He May Be Merciful to All: Karl Barth and the Problem of Universalism’ in Karl Barth and American Evangelicalism. Grand Rapids: Eerdmans, pp.239-241.