5626463547 | San Jose, CA 95126

For God so loved the world: A Calvinist Response to Richard Brian Davis

This piece is co-authored by Guillaume Bignon and James A. Gibson. Order is alphabetical. 

In “Calvinism’s Gospel Tautology,” Richard Brian Davis argues that John 3:16 is evidence against the Calvinist doctrine of limited atonement. According to that doctrine, the death and resurrection of Christ is intended for only a subset of humanity, the elect. How, then, is a Calvinist to understand “world” in this passage? Davis considers one possibility suggested by R.C. Sproul.  “The world for whom Christ died cannot mean the entire human family. It must refer to the universality of the elect (people from every tribe and nation) or to the inclusion of Gentiles in addition to the world of the Jews” (Sproul, cited in Davis). Davis argues that it is a fundamental mistake to take the meaning of “world” to be the elect. By understanding “world” in that way, Davis argues, the Calvinist is faced with the dilemma of either Jesus saying something false or Jesus saying something trivially true. We argue that the Calvinist can resist the dilemma.

The Dilemma Presented

Assume that “world” does not refer to every individual that exists. Next, assume limited atonement is true. Then, Davis says, the following is a plausible Calvinist rendering of John 3:16:

  • (R1) God so loved the elect, that he gave his one and only Son, that whoever believes in him shall not perish but have eternal life.

He goes on to argue that the domain of the universal quantifier in the passage, “whoever”, may take either a wide or narrow interpretation. On the wide interpretation, the domain ranges over both the elect and non-elect. On the narrow interpretation, it ranges only over the elect. Before jumping into why both interpretations are problematic, it is useful to tidy up this argument a bit.

James White has responded that the participial phrase, πᾶς ὁ πιστεύων, is best rendered “everyone believing”. There is no universal quantifier “whoever” in the text. So it appears that Davis’ argument depends on the presence of a word that is not there. But Davis points out in a subsequent piece that his argument does not depend on that word. Let the rendering be:

  • (R2) God so loved the elect, that he gave his one and only Son, that everyone believing in him shall not perish but have eternal life.

It is still an open question what the domain of “everyone” is given the interpretation Sproul provides, and he cites the quote from Sproul (above) again. Davis’ argument remains mostly the same, only packaged slightly differently. We believe that the problems which beset Davis’ argument do not depend upon which rendering one prefers. But because the first rendering can be misleading to what the central issues are, we will continue with the second rendering in order to present the dilemma.

The dilemma arises from a disambiguation of the domain of “everyone”, which is implicitly of persons. The question about the domain is explicitly over whether this covers the elect only or both the elect and non-elect. This yields two precisifications:

  • (WIDE) God so loved the elect, that he gave his one and only Son, that everyone (of both elect and non-elect) believing in him shall not perish but have eternal life.
  • (NARROW)  God so loved the elect, that he gave his one and only Son, that everyone (of the elect) believing in him shall not perish but have eternal life.

We will first discuss the narrow interpretation and the problems that Davis believes affect it. Then we will turn to the wide interpretation, and finally close with some final remarks.  

The Narrow Horn of the Dilemma

Suppose the Calvinist opts for the narrow interpretation, such that it covers only the elect. Then the Calvinist is committed to:

  • (NARROW)  God so loved the elect, that he gave his one and only Son, that everyone (of the elect) believing in him shall not perish but have eternal life.

Davis discusses NARROW with predicate logic. We, however, do not believe that formalization adds to the clarity of this discussion. We will grant all of the entailments that Davis identifies between the following three principles:

  • (ELECT) For any x such that x is elect and x believes in Jesus, then x will not perish but will have eternal life.
  • (BELIEF) For any x such that x is elect, then x will believe in Jesus.
  • (TRIVIAL) For any x such that x is elect, x will not perish and will have eternal life

ELECT is supposed to represent the that clause – that everyone of the elect believing in him shall not perish but have eternal life. ELECT and BELIEF entail TRIVIAL. And TRIVIAL and BELIEF entail ELECT. This is bad according to Davis because it requires attributing a kind of semantic performance error to Jesus. It is worth quoting Davis in full:

This proposition [i.e. TRIVIAL], on Calvinism, is an empty tautology; it is utterly trivial. It is logically impossible that one be elect and yet perish and be lost. These predicates are included in the concept of being elect. Accordingly, TRIVIAL is about as informative as the analytic truth All bachelors are unmarried and male. It tells us nothing we didn’t already know by definition. Obviously, the good news Jesus came to preach isn’t anything like this. After all, nothing can be good news if it’s not news at all (and analytic truths aren’t news).

But there is something else. It turns out that TRIVIAL and BELIEF also entail ELECT. (I shall leave the proof as homework for the reader.) In the presence of BELIEF, therefore, ELECT and TRIVIAL are equivalent propositions. But then, curiously, what Jesus is telling Nicodemus is that God loved the elect and gave his Son to bring about TRIVIAL: that the elect have eternal life and won’t perish. But that scarcely makes sense. You might as well argue that God needed to enter human history to ensure that whoever is a bachelor and male is also unmarried. This is wrong headed. Bachelors would be unmarried and male even if there were no bachelors, no persons, and no world that God created.

The same thing goes mutatis mutandis for TRIVIAL. Even if the elect didn’t exist, and the Son wasn’t thereby given for them, it would still be true—and indeed necessarily so—that the elect believe, have eternal life, and don’t perish. To put (3:16b) into the mouth of Jesus either has him teaching falsehoods or empty tautologies, neither of which is befitting to the Savior.

We take this to strongly suggest that Davis regards TRIVIAL to be analytic. It is not clear how else it can be that the predicates of “not perishing” and “having eternal life” are “included in” the concept <elect>, and that TRIVIAL is necessarily true. We proceed by taking the sense of triviality used in the argument to be analyticity.

The central questions to ask, as we see them, are:

  1. On the assumption that TRIVIAL is trivial, would Jesus’ assertion of TRIVIAL thereby undermine the Calvinist interpretation?
  2. On the assumption that Jesus asserts NARROW and that TRIVIAL is part of NARROW, does Jesus thereby assert something trivial?
  3. Should we assume that NARROW is an appropriate interpretation of what Sproul says about how to understand John 3:16?

Our answers will be No, No, and No.

Infelicitous Trivial Assertions?

Two distinct problems Davis identifies, given that TRIVIAL is analytic, are that (i) the truth of the Gospel is not at all like an analytic statement; and (ii) it would be pointless for Jesus to tell “us nothing we didn’t already know by definition.” The former problem is more closely related to the question of whether Jesus asserts something trivial, and so we will return to that. Here we address the second issue regarding whether it would be pointless for Jesus to assert TRIVIAL if indeed it is analytic. To this, we say, it does not follow merely from the fact that Jesus makes a trivial assertion (which we do not grant), in the sense of being analytic, that what he asserts is thereby pointless. There are two reasons to doubt that they would be pointless.

The first reason concerns whether the audience understands the trivial assertion. Consider the fact that TRIVIAL is derived from two other principles: ELECT and BELIEF. Neither ELECT nor BELIEF is trivial for Nicodemus, since from his perspective, there is no known connection between election, belief in Jesus, and what follows from those pertaining to not perishing but having eternal life. Moreover, Jesus is giving a theology lesson to a Pharisee. The Pharisees had a theological dispute with another Jewish sect called the Sadducees. They disagreed over whether there would be a resurrection. In this context, you can see Jesus as taking sides in that debate. Thus Jesus’ assertion, even if it were trivial, would at least convey information to Nicodemus about how Jesus sees things. So even if Calvinists treated <belief in Jesus> as part of the concept of <election>, it would be irrelevant to Jesus’ assertion to Nicodemus. For it is he, not us, who counts as the primary standard by which we judge whether Jesus’ assertion in the text is pointless.

The second reason to be skeptical about the pointlessness of making an analytic assertion has to do with the pragmatics of our utterances. Suppose, for instance, that Millian descriptivism is true. That is the view that sentences involving names often pragmatically convey descriptively enriched propositions. For example, “Hesperus is a planet” might convey “Hesperus, the brightest object in the morning sky, is a planet.” (This view is advocated, for instance, by Scott Soames.) Then one will find plausible the suggestion that we may ignore what is semantically expressed by a sentence and focus instead upon what is pragmatically conveyed. Using an example from Jeff Speaks (“Millian Descriptivism Defended”, Philosophical Studies, 149 (2010):201-208), someone might have uttered during a conversation about the Clinton / Dole race in 1996, “Bill Clinton was a real politician.” Uninformative, yet what we would understand this to convey is something else, e.g. that Clinton is more capable than Dole, or that Dole is a fraud, or perhaps that Clinton is a fraud. It is a difficult question, in general, how to settle which proposition is conveyed. The upshot is that the mere fact that Jesus says something uninformative (which we do not grant) does not by itself make what Jesus says pointless.

TRIVIAL without Triviality

We deny that TRIVIAL is analytic. Although it is true that the elect will have eternal life, it is not an analytic truth. For if it were, then the debate between Sadducees and Pharisees would not be a substantive dispute at all, but a merely verbal dispute due to confusion over “elect”.

There is a further reason to doubt that TRIVAL is about as informative as an analytic statement. For most analytic statements (or all those we can think of), if someone knows the meaning of the terms of the statement, then probably not much argumentation would be required to convince such a person of its truth. In general, that a statement’s truth would require much argumentation for such a person signifies that such statements are not analytic. The use of “signifies”, that such is a sign of not being analytic, does not amount to a necessary or sufficient criterion of analyticity. There may be counterexamples; but again, this is not a comment about all analytic statements. It is about what is normally true of analytic statements.

With that in mind, consider the terms in TRIVIAL. Nicodemus may understand that “elect” denotes the group of persons God chose before the foundation of the world to save. He may understand to some extent what “belief in Jesus” means, even if he does not believe in Jesus in John 3. He may also understand the meaning of “eternal life” and “not perish”. What he apparently did not know, however, were the relations between belief in Jesus, election, and eternal life. That by itself does not prove that the statement is synthetic. But it is an indication that TRIVAL is not analytic so far as we consider what Nicodemus understands but does not believe.

We can go further with this line by considering a common criterion of analyticity, that analytic statements are necessarily true. One might doubt that it is a necessary truth that the elect believe in Jesus. Perhaps there is a possible world in which the divine persons, Son and the Spirit, have switched roles with respect to the incarnation. Then which divine person the elect believes in has a different singular object. We grant that this is speculative, but it is enough to question whether one is in a position to assert that there is a necessary connection between election and the object of belief-in. It may be thought a necessary truth that the elect believes in whichever divine person becomes incarnate. But even here someone might doubt that this is a necessary truth. So again TRIVIAL’s analytic-like character is doubtful.

The Narrow Interpretation and Frege’s Puzzle

Whereas the above objections discussed John 3:16b, our discussion will now focus upon the rendering in NARROW of John 3:16a, which we do not think is appropriate. We challenge the move by Davis of replacing “the world” by “the elect”, even if the elect is what ultimately is referred to by “the world”. Recall the narrow interpretation again:

  • (NARROW)  God so loved the elect, that he gave his one and only Son, that everyone (of the elect) believing in him shall not perish but have eternal life.

Davis alleges 3:16a as rendered is problematic: “suppose for reductio that Sproul is right: this principle [John 3:16] (taken at face value) is potentially misleading. Although (3:16) says God loved ‘the world’ and gave his Son for it, that is actually false. God didn’t (and doesn’t) love everyone; he loves only the elect.”

Does NARROW capture the information that Sproul intends? We do not believe so. Assume for the sake of argument that Sproul is correct, that in John 3:16, “world” refers to the elect. Still, it is appropriate to wonder why all his talk about Jews and Gentiles and different nations does not make it into the rendering Davis presents as the propositional content of Jesus’ utterance. Again, here is Sproul:

The biggest problem with definite or limited atonement is found in the passages that the Scriptures use concerning Christ’s death “for all” or for the “whole world.” The world for whom Christ died cannot mean the entire human family. It must refer to the universality of the elect (people from every tribe and nation) or to the inclusion of Gentiles in addition to the world of the Jews. It was a Jew who wrote that Jesus did not die merely for our sins but for the sins of the whole world. [1 John 2:2.]

Consider a well-known issue in the philosophy of language, Frege’s Puzzle.  In the case of identity statements involving singular terms, Frege wondered how it could be that two co-referring terms (e.g., proper names) might make one sentence trivially true and the other informative.

  • (A) Hesperus = Hesperus
  • (B) Hesperus = Phosphorus

Both terms refer to the same object, Venus. Only the second sentence appears to be informative. How can this be so? Frege’s answer was that names carried a sense, which was Frege’s term for a mode of presentation. The propositions expressed by A and B, on the Fregean view, are distinct propositions.

Assuming that Frege is correct about how referring terms contribute to a proposition’s content, then it straightaway follows that NARROW is not a proper rendering for John 3:16 because it ignores the sense of the definite plural term “the world” by substituting a term with a different sense. A more appropriate rendering of John 3:16 would be

  • (C) God so loved Jews and Gentiles, that he gave his one and only Son, that everyone (of the elect) believing in him shall not perish but have eternal life.

C is perfectly intelligible. C conveys that of the different ethnic groups, the persons (of those groups) believing in him will have eternal life. It happens to be true that all and only those who believe in him are elect. Because limited atonement is about God’s intended scope for and efficacy of the death and resurrection of Jesus, Sproul also accepts:

  • (D) God so loved the elect, that he gave his one and only Son, that everyone (of the elect) believing in him shall not perish but have eternal life.

But while on Sproul’s interpretation, “the elect” and “the Jews and Gentiles” ultimately refer to the same persons (the objects of God’s love), Jesus utters proposition C and not proposition D. 

Our defense of Calvinism pertaining to John 3:16a does not hang upon the truth of descriptivism (at least one of us denies it). In order to see why, a brief explanation of direct reference views is in order. Direct reference views hold that the semantic contribution of some singular terms is nothing more than the reference. Assuming that “the world” is directly referring and so is “the elect”, it is not available to the direct reference theorist to insist that C and D express different propositions. After all, given Sproul’s view, they both refer to the same persons. What is the direct reference Calvinist to say?

The Calvinist can say what other direct reference theorists have said. Propositions expressed are not the complete content of our attitudes. According to Francois Recanati, a direct reference defender, a proposition expressed is apprehended up a mode of presentation (“Direct Reference, Meaning, and Thought,” Nous 24(1990): 697-722).  The singular terms that refer directly can still be terms that have senses. For example, the singular term, “I”, in the sentence, “I am a software engineer” has a shared meaning whether said by James or said by Guillaume. Its meaning has something to do with the attribute of being a person. But which proposition is expressed or to whom “I” refers in an utterance depends upon which person speaks. Second, co-referential terms may be assented to, thought of, or apprehended under different modes of presentation. “Tully is bald” and “Cicero is bald” express the same proposition, but have different modes of presentation. 

With this background, one can see how the reductio against Sproul fails. That reductio, again, is that taking John 3:16a at “face value” results in Jesus saying something potentially misleading, or what is worse, something false. “Although (3:16) says God loved ‘the world’ and gave his Son for it, that is actually false. God didn’t (and doesn’t) love everyone; he loves only the elect.” Davis thinks Sproul is incorrect about the referent of “the world” because he (Davis) apprehends Jesus’ statement under a certain mode of presentation that is incompatible with “the world” having the reference Sproul identifies. But of course, one might wonder why “world” taken at face value means “every person, whether elect or non-elect” and not the mode of presentation found in C, and which is stressed in the Sproul quote above. The proper way to settle which mode of presentation is more appropriate to accept about a term is to read the surrounding text. Arguably, Davis’ view is incorrect on those grounds.

By attending to Frege’s puzzle, we have been able to show that Calvinist understandings of John 3:16a do not result in Jesus saying something false or misleading.

The Wide Horn of the Dilemma

Davis’ dilemma aimed to corner the Calvinist between a rock and a hard place. We have shown above there was no rock on the one side, and that’s where we feel like escaping. But let’s take another moment to show there was no hard place on the other side anyway. What is supposed to be the problem if a Calvinist grants a wide scope for the quantifier in John 3:16?

Davis tells us that:

On this reading of the quantifier, (3:16a) turns out to be false.

We disagree that it turns out to be false. With an explicit wide scope for the quantifier (and with the previously mentioned reformulation of “whoever” to avoid controversy), that statement which Davis called (3:16a) is what we called WIDE:

  • (WIDE) God so loved the elect, that he gave his one and only Son, that everyone (of both elect and non-elect) believing in him shall not perish but have eternal life.

If we thus suppose that the “everyone” is ranging over both elect and non-elect, it does not make the WIDE statement false. Rather, it just makes it a bit strange that God’s love in the antecedent would be affirmed of just the elect, and that the “everyone” in the consequent would still range over everybody. But that strangeness will be removed by the Calvinist who opts for a wide reading of the quantifier, by also adopting a wide reading of “the world” in the antecedent, thereby agreeing with Davis that it’s not just the elect this verse has in view here. If Calvinists do that, they do not need to find fault with much of anything that Davis presses them to say. Davis’ final, preferred rendering of John 3:16, that which he calls his “desiderata”, is something that such Calvinists can just as well affirm:

  • (3:16c)  God so loved the fallen world, that he gave his one and only Son, that any fallen human being who believes in him shall not perish but have eternal life.

Calvinists can affirm this, and they can do so even if they disagree that it is what John is saying!

First, it is clear that: “any fallen human being who believes in Jesus shall not perish but have eternal life” is consistent with Calvinism, and that it is true in virtue of the death of Jesus. So it’s not the conditional “if one believes, he shall not perish” that Davis must be saying Calvinists cannot affirm. Rather, Davis must be pressing Calvinists on the stated connection with God’s love. The verse would be saying God loved all these people (now supposed to be elect and non-elect) so much that he did something for them (all). And Davis says that on Calvinism, God either didn’t do anything for the non-elect, or whatever he did for them isn’t very significant:

To be sure, God’s loving and giving his Son for the elect is sufficient for their believing and having eternal life. But this has no bearing whatsoever on the non-elect. For abandoned to their fallen desires, they cannot believe in Christ. Surely, it would be a strange and pointless thing for Jesus to be at pains to tell Nicodemus that “whoever (of those who cannot believe) does believe: these persons won’t perish but have eternal life.”

But even on Calvinism, the giving of Jesus has a “bearing whatsoever” on the non-elect. It’s in virtue of the death of Jesus that “if they believe they will not perish”. It is true that they won’t believe (because on Calvinist premises they can’t: as the Bible puts it, they’re in the flesh and cannot please God, they cannot discern the things of the Spirit, they’re dead in sin and cannot raise themselves to spiritual life, they cannot believe). But the counterfactual “if they believed they would not perish” is still true, and is true in virtue of the death of Jesus. So it would not be false for John to make the affirmation about everyone (elect or not). We understand that the mere truth of this counterfactual in the case of the non-elect won’t get the libertarian very excited, but it is not nothing, and it is not a “strange and pointless thing” for Jesus to tell Nicodemus. Why? Because Jesus (on this reading) is saying it in one single statement, of both the elect and the non-elect. So for Jesus’ statement to be true and very significant (which it is), it only needs to be true of the non-elect (which it is), while true and very significant of the elect (which it is).

Given all this, we do not believe there is a problem for a Calvinist who (unlike us) feels inclined to read a wide scope in the quantifier of John 3:16. But let us finish the job with a tu quoque response as we believe Davis is a Molinist.

If Molinism is true, God knew in his middle-knowledge that the non-elect would not freely believe in the circumstances in which God is actually going to place them. Then what is the Molinist to say that John 3:16 is affirming about the non-elect? That God loved them so much that he sent Jesus to die for them, a thing which he knew full-well would not save any of them unless he placed them in different circumstances? If the alleged absurdity of saying of the non-elect “Because God loves the world, he goes ahead and does something that won’t save them anyway” takes down Calvinism, it takes down Molinism too.

The Molinist might reply that God is at least trying his best to save them. This point is not satisfactory, because God would be trying to do something that he knows cannot happen – that they be saved – given which world God decrees to actualize. To the extent that the Molinist is willing to accept this is something meaningful, the Calvinist can point to the general call of the Gospel as something equally meaningful.

The bottom line is that Calvinists and Molinists, if they opt for a wide reading of the quantifier in John 3:16, must affirm the same thing about the non-elect, whom God (on this reading) is said to love: for them, God gives Jesus, so that if they believed, they would be saved, but God knows they won’t believe, which means the net result for them in terms of actual salvation is exactly as if God hadn’t done anything for them.


We have argued that the dilemma argument from John 3:16 fails. Calvinists attracted to Sproul’s view that “the world” refers to the elect do not face even a prima facie problem with understanding the meaning of this text. The verse is not trivial, but even if it were, that feature would not by itself make Jesus’ assertion absurd. Furthermore, if it were true that the Calvinist were forced into accepting the wider interpretation of “everyone believing”, the Calvinist need not concede that Jesus would be asserting something false or insignificant. Finally, if we Calvinists faced a problem here, then so would the Molinists. But they do not, and so we do not. Both parties can agree that John 3:16 is true, and that its truth is significant. For the Son not only came with the intention to give eternal life to the world, but the Son succeeds in taking away their sin. (John 1:29)

2018-04-08T13:32:05+00:00 April 7th, 2018|2 Comments


  1. Rich Davis April 8, 2018 at 9:41 am - Reply

    This is an excellent piece of philosophical analysis, guys!

    • James A. Gibson April 8, 2018 at 11:54 am - Reply

      Thanks, Rich!

Leave A Comment