Response to Dr. Olson on Middle Knowledge

, posted by Godismyjudge

Dr. Roger Olson has repeatedly publicly objected to the doctrine of middle knowledge.[1]   His basic objections are that middle knowledge amounts to determinism, makes God the author of sin and is a form of Calvinism rather than Arminianism.  This is a response to Dr. Olson’s two most recent blog posts criticizing middle knowledge (here and here).

Arminius and the Remonstrants on God’s Middle Knowledge

Dr. Olson argues that God’s having middle knowledge is fine, but His using middle knowledge amounts to determinism.  But what Dr. Olson suggests is equivalent to divine amnesia since to totally get rid of middle knowledge from God’s decision making; middle knowledge must not be part of God’s conscious thought.  Now perhaps Dr. Olson might say, God uses His middle knowledge in His decision making, but He just doesn’t give it much if any weight.  But this sounds like irrational or impulsive decision making and God is Wisdom personified.  So Dr. Olson leaves us guessing what he has in mind here.

In any event, Arminius said God both has and uses middle knowledge.  In Arminius’s response to Perkins, he says:

  • God issues decrees based on what He knows via middle knowledge and He would not make those decrees had the creature in God’s middle knowledge acted differently (p. 283)[2];
  • Contra Supra-lapsarianism, God could not have decreed to glorify Himself through the salvation of sinners without first considering man as fallen (via middle knowledge) (p. 295-296);
  • God’s permission is defined as His being unwilling to hinder an event He foresees would occur without His hindering it – yet we cannot conclude from this that God willed the permitted event (p. 287);
  • God has middle knowledge about whatever He permits (p. 394);
  • God can prevent our actions by knowing how we would respond to various persuasions and providing them (p. 423);
  • God’s prevention is successful, not because of His power, but because of His middle knowledge (p. 425);
  • God does not harden sinners unless He knows via middle knowledge that the person will not respond to warnings (p. 329);
  • God decided to prolong Hezekiah’s life and to spare Nineveh, because He knew they would repent (p. 342-343);
  • God knows who will freely assent to His grace (p. 481).

This all is just from Arminius’ Review of Perkins.  Middle knowledge is foundational to Arminius’s thought rather than being a foreign body. Indeed, one wonders just how Dr. Olson defines Arminianism.  Arminius taught middle knowledge.  So did Episcopius (Works of Simon Episcopius. Page 303); Grotius (Opera p. 351-354); Corvinus (Against Peter Moulin, page 73 Chapter 5, Section 5); Vorstius (Tractatus theologicus de Deo. Page 47) and Grevinchovius, Dissertatio theological du duabus questionibus – cited in Turretin.  Institutes.  Volume 1 page 213).  One would be hard pressed to find any original Remonstrant who didn’t teach middle knowledge, so to say middle knowledge is a foreign body to “Classic Arminianism” is to not allow history to define “Classical Arminianism”.

Does God use Middle Knowledge for Evil?

Dr. Olson contests that middle knowledge undermines the classic Arminian distinction between God’s antecedent and consequent will.[3]  The antecedent/consequent will distinction is a fancy way of saying there’s a difference between what God wants you to do and what God wants to do about what you have done.  For example, when commenting on Matthew 23:37-39, Arminius said by God’s antecedent will, He wanted to gather the children of Jerusalem to Him, but by His consequent will, after He was rejected, He judged Jerusalem. (p.463) As Limborich pointed out, the antecedent/consequent will distinction is basically a conditional decree.[4]  Thus God provides man with significant freedom in that God ends up acting differently than He would have acted, had His creation chosen otherwise.

Theologians sometimes explain God’s use of middle knowledge in ways that do undermine the antecedent/consequent will distinction.  For example, Thomas Flint said God unconditionally elects some for salvation and then consults His middle knowledge to find a way to save only His elect.[5]   I call this “Supra-lapsarian Molinism”, because God decides the ends first and then decides on the way to get there.  This is highly inconsistent with Arminianism and the antecedent/consequent will distinction[6].  So the real question is not if God uses middle knowledge, but rather how He uses it.

To avoid this problem, Arminius explained the order of decrees as logically progressing forward in time (as opposed to supra-lapsarianism, which start from the end of time and works backwards against time).  By God’s antecedent will, God desired that Adam not sin.  But given God knew Adam would sin; God consequently decreed: that Christ be the foundation of salvation, and that faith be the condition for salvation, and to provide prevenient grace to enable us to believe.  By God’s antecedent will, God desires all men to believe and be saved, but given God knows who would believe and be saved, God consequently decrees to save those who would believe.

This provides the global view but Arminius followed this line of reasoning in more detailed matters as well.  By God’s antecedent will, God desires Pharaoh to repent.  But given God knows Pharaoh has not repented and would not repent, God consequently wills to harden Pharaoh.  Consider how Un-Arminian the reverse would be – if God first decides to harden Pharaoh without considering Pharaoh’s obstinacy.  Had Arminian-Molinists reversed this order, we would be guilty of the error Olson supposes.  But God does not do such things, not because He lacks knowledge but because He is good and loves all mankind and desires to communicate His goodness to all mankind.

Does Middle Knowledge Amount to Determinism?

Dr. Olson objects to God using middle knowledge “to render certain that every creature does what they do by creating them and placing them in circumstances where he knows they will “freely” do something”.  So does Arminius. The word ‘certain’ can either mean “free from doubt” or it can mean “fixed and determined” ( entry).    Using “fixed and determined” as a definition, Arminius objected to the idea that future events are certain – because he rejected determinism.  But given the definition of “free from doubt”, Arminius says God’s foreknowledge is certain – because God foreknows all things.  Dr. Olson seems to be equivocating these two definitions of certain.

For Arminius, it does not follow that, given God’s “creating them and placing them in circumstances where he knows they will “freely” do something” that God has “rendered that event certain”.  Rather, God has no doubt that the event will happen.  Arminius carefully distinguished between a certain or logically necessary conclusion to a syllogism and causal necessity.  Arminius understood that given God foreknows an event, that event will happen and the opposite will not happen.  (James Arminius. Works of James Arminius. 1875 London Edition. Examination of Gomarus on Predestination. Volume 3. p 547-549)  One might see a contradiction in Arminius’ views if they mix certainty with necessity, which apparently is what Olson has done.

Dr. Olson quotes a guess commenter on his blog who argues if God knows, say “you would accept the manager job if offered” then, you “couldn’t act so as to bring about its falsity”.  I have little doubt this commenter is an open theist, since such an objection works just as well against Dr. Olson’s simple foreknowledge view as it does against middle knowledge.  But that aside, following Occham, we can act in such a way that if we did, God’s middle knowledge would have been different.  Libertarian freedom is defined in causal terms.  It’s about what agents can do and our ability to cause events.   Thus libertarian freedom is the ability to choose something or not given all the causal forces influencing use.  So while libertarian freedom requires the ability to cause alternative events; it does not require the ability to change future tense propositions from true to false or undefined to false.  It’s enough to have the ability to cause an alternative event, such that if you do cause it, predictions about that event would be true.

Book Illustration

As we have seen, Dr. Olson’s objections to middle knowledge come under two distinct headings: 1) God would use middle knowledge in immoral ways and 2) the philosophical objection that any use of middle knowledge would amount to determinism.  Let’s take Dr. Olson’s example of a Professor recommending a book to a student, while knowing the student will misunderstand the book and fail.  Ironically, this example illustrates Dr. Olson’s two main misrepresentations of middle knowledge. First, people do not choose to misunderstand things so Olson substitutes determinism for libertarian freedom.  Second, the Professor’s goal was both selfish and set prior to his decision to recommend the book. He wants to flunk someone and his knowledge is just a means to that end.  But this is like supra-lapsarian Molinism and God’s love and goodness to His creation are the reasons He doesn’t do this. Arminian-Molinism is more like a professor recommending a book because it’s the best way for all his student to pass and he wants them all to pass. But with sorrow, he knows that even if he recommends the book, some will not pass anyways, because they would choose not to study.  Can we blame the Professor?


My primary objection to Dr. Olson’s position is that the bible simply teaches God knows what we would choose in various settings (Exodus 3:19, 1 Kings 11:2-9, 1 Samuel 23:6-10, Ezekiel 3:6-7, Matthew 11:21-23, Matthew 23:27-32, 1 Corinthians 2:8) .  Secondarily, I am concerned that so called simple foreknowledge is providentially useless and therefore unable to account for the biblical texts on God’s providence (Genesis 50:20, Proverbs 16:9, Acts 4:28, Ephesians 1:11).[7]  The concern is Christians holding Dr. Olson’s position will read their bibles, find things that don’t fit, and end up troubled and primed for Calvinists to come swoop them up.

The reality is Calvinism, Molinism and Open Theism are philosophically robust systems and simple foreknowledge is not.  For example, one can go to and buy dozens of resources on Calvinism, Molinism or Open Theism.  Good luck finding resources on simple foreknowledge.   One wonders if the simple foreknowledge position is under-articulated because it’s just not a well thought through position and it cannot stand up to criticism.  It’s perfectly OK and in some cases applaudable to simply say “I don’t know” but please don’t use ignorance as a sort of one way glass to hide behind and criticize others without subjecting your own position to scrutiny.

Middle knowledge vs. simple foreknowledge is one of the many lively “in house” Arminian debates.  But to some extent, I think Dr. Olson is putting in house Arminian business out on the street, by publicly attacking middle knowledge.  Never-the-less I am very grateful Dr. Olson says “I’m certainly not going to say that one cannot be an Arminian and a Molinist” and such comments give me hope that Arminian-Molinsits and Arminian-NonMolinists can work together even without full agreement.

[1] Olson. Arminian Theology. P 196 and Against Calvinism.  P 183-186

[3] Arminius explains that God’s antecedent will (His will for us logically prior to our will) is a velity or mild desire rather than a final decision or mental resolution.   God does not employee His omnipotence to accomplish His antecedent will.  Yet it still reveals God’s character and intentions.

[5]  Flint, Thomas.  Divine Providence The Molinist Account.  Principle of Predilection. p. 117.  Flint is following Fransisco Suarez and Robert Bellarmine’s Congruism – see New Advent Catholic Encylopedia entry on Congruism.

[6] In Molinism, unlike Compatible Determinism, God cannot make you do something.  He cannot decree things contrary to what He knows you would choose.  So on Supra-Lapsarian Molinism, what if God elects some who would not repent?  So Supra-Lapsarian Molinism is fundamentally inconsistent with William Lane Craig’s theory of trans-world damnation, wherein the lost are lost in ever feasible world.

[7] The typical argument is that simple foreknowledge is logically “too late” to change the foreknown event.  This is sometimes illustrated by the so called Grandfather Paradox in which someone goes back in time and kills his own Grandfather.