Adams: Middle Knowledge and the Problem of Evil

Philosophical

American Volume

Quarterly

14, Number

2, April

1977

III. MIDDLE

ROBERT

MERRIHEW

had not been shot, would President Kennedy he have bombed North Vietnam? God only knows. Or does He? Does even He know what Kennedy would have done ? There is a little known but interesting literature on the general issue exemplified by this question. TF *

In

the

a

i58o's

fierce

controversy

between

erupted

about the relation the Jesuits and the Dominicans between God's grace and human free will. The Jesuits held, among other things, that many human are

actions

not

in

free

the

sense

that

their

are

agents

or causally determined to do them. always be used in this sense in the then does God maintain essay.) How over human history? Not by causally

logically ("Free" will

present control

human determining to believe,1 seemed

in which

He

accordance

actions, but by

with

as

the

causing

Dominicans

His

freely answer

This

plans.

conditional

in

a middle

ledge?between

status God's

between

other

knowledge

kinds of

of the

sort

the

are

that

that sup?

of

creatures

who

ever

whom

made

made

free

choices, choices?

wrong

but The

relevance of this question to the problem of evil is obvious and well known. If He could have, why didn't He? If He couldn't have, that's a good enough reason why He didn't. He could not have done it by causally determining the choices of

was

for

creatures,

then

their

choices

and

acts

would

not

have been free in the relevant sense. But it might seem that if God has middle knowledge, He could have

know?

allowed

creatures

secured

creating

merely

or and His knowledge of the actual; possible of necessary truths, which between His knowledge all follow from the divine nature, and His know? that is ledge of His own will and everything causally determined by His will.2 This paper is about two questions. The first is even for is possible, whether middle knowledge

of

propositions

had

He

none

developed with great ingenuity by Luis de Molina, and defended by other Jesuit theologians, notably Suarez. Their theory includes the by Francisco thesis that God knows with certainty what every possible free creature would freely do in every in which situation that creature could possibly find himself. Such knowledge was called "middle by the Jesuits, because they thought it knowledge" had

that it is not, on the ground

posed to be known by middle knowledge cannot be true. I will examine (in section II) the attempts of Molina and Suarez to explain how God can have and then (in section III) the middle knowledge; account recently offered by Alvin Plantinga, who has reinvented the theory of middle knowledge. Two objections to my position will be discussed in section IV. The idea of middle knowledge emerges in recent discussion because of its philosophical chiefly to the second question relevance that I shall dis? is whether God could have made free cuss, which creatures who would always have freely done right. More precisely: Could God have brought it about that

act

PROBLEM

ADAMS I shall argue

God.

circumstances

that we would

knew

THE

AND KNOWLEDGE OF EVIL

those to

act

sinless

that He freely.

knew

In

section

but

free

V,

by

not

would

therefore,

just

sin

if

we

shall see what light the discussion of middle know? ledge may shed on the question whether God could have

arranged

to have

free

creatures

who

were

all

sinless.

But first of all (in section I) I will try to explain why there seems to me to be a problem about the possibility of middle knowledge.

1An to the debate contribution is Diego De auxiliis divinae gratiae et acutely argued Dominican Alvarez, O.P., (Didacus) humani arbitrii viribus, et lib?rtate, ac legitima eius cum efficacia eorundem auxiliorum concordia (Rome, the seventh 1590) ; see especially disputation. 21 believe Molina the term "middle I have given a very simplified account of his originated knowledge" (scientia media). reasons for thinking it appropriate. See his Liberi arbitrii cum gratiae donis, divina praescientia, et reprobatione providentia, praedestinatione concordia [hereafter abbreviated, Concordia], ed. by John Rabeneck (O?a and Madrid, 1953), qu. 14, art. 13, disp. 52, m 9-10, and disp. 53, memb. 1, n. 6, and memb. 4, n. 4 (pp. 339f., 360, 394). log

IIO

AMERICAN

PHILOSOPHICAL

QUARTERLY

I In the twenty-third chapter of the first book of Samuel it is written that after David had rescued and the Jewish city of Keilah from the Philistines, settled his men there, Saul made plans to besiege David Keilah in order to capture David. When heard of Saul's plans, he consulted God by means of of

an

which

ephod, divination

was

apparently

that

a

yielded

an

instrument answer

yes-or-no

to

questions. David asked, "Will Saul come down, as The Lord answered thy servant has heard?" men of David Then "Will the asked, affirmatively. Keilah surrender me and my men into the hand of Saul?" And the Lord replied, "They will surrender David his men from evacuated you." Thereupon out in with the result that hid the and Keilah, hills, to besiege him Saul did not have the opportunity in Keilah, did not have and the men of Keilah occasion to betray him to Saul. (I Samuel 23:1 -14,

RSV)

passage was a favorite proof text for the theologians. They took it to prove that God to be true: the following two propositions Saul would be? stayed in Keilah, (1) If David the siege city. (2) If David stayed in Keilah and Saul besieged the city, the men of Keilah would surrender to Saul. David

This Jesuit knew

for it is assumed This is a case of middle knowledge; in (1) and (2) would that all the actions mentioned have been free, in the relevant sense, if they had occurred. If we

versy whether

suppose

that

as

Therefore, comes

determinate

say that (i) and

Suarez would

are

God

we

is omniscient,

cannot

back

those truth."

Suarez to

says,

this, conditionals 3

"the that have

contro?

whole we

should a

see

respectively, pairs that one member fore

(2) and and be

(4),

there? true.

He

thus affirms what has been called the law of Con? But this is a mistake. To ditional Excluded Middle. obtain

the

tion

is not

must

negate

enough the whole

a

of

contradictory

conditional

to negate

the

consequent; as was

conditional,

proposi? one pointed

out by Alvarez.4

Dominican Suarez's opponent, Diego It is true that in everyday speech we might deny (1) by asserting (3), as we may deny a by asserting any belief we hold that is proposition obviously enough inconsistent with it. But we might also deny both of them by asserting, "If David Saul might or might not besiege stayed in Keilah, the city." I believe the case of what Saul would or might have done ifDavid stayed inKeilah provides a

plausible

counterexample

Conditional

Excluded

have

even

found

to

more

the

law

proposed

and

Middle;

of

philosophers counter?

convincing

examples.5

I do not understand what it would be for any of to be true, given that the (i)-(4) propositions actions in question would have been free, and that David I will explain my did not stay in Keilah. incomprehension.

First we must

note

simple yb^knowledge.

that middle knowledge is not The answers that David got

will the ephod?"He surrender you"?are as categorical theologians

from

predictions,

categorical have about ponding

contingent to the

they predict. true

that

supposed

come not

categorical can events,

actual

occurrence

But propositions

in this way.

For

there

never

and

down," understood

"They the

by

were they false. Most

If

predictions. they would

Suarez

(including

philosophers

knowable

But I do doubt that propositions (1) and (2) ever were, or ever will be, true. This is not because I am inclined to assert the truth of their opposites, not Saul would stayed in Keilah, (3) If David the city. besiege

(3), and

of contradictories, must of each pair

will

had this middle that He doubt consistently we that doubt and unless (1) (2) were knowledge true.

(4) If David stayed in Keilah and Saul besieged the men not the city, of Keilah would to Saul. surrender David

be

but

not Molina)

predictions, true be by of

the

(1) and was

nor

even corres?

event

that

(2) are not will

be

an

besieging of Keilah by Saul, nor an actual to David to of Saul by the men of Keilah, betrayal which those propositions might correspond.6 Some other grounds that might be suggested for actual

3 to De vol. 7, p. 85. (All my page references Suarez, De gratia, prol. 2, c. 7, n. 1, in his Opera omnia (Paris, 1856-1878), and volume.) gratia will be to this edition 4 Alvarez, op. cit., Bk. 2, disp. 7, n. 30 (p. 74). See Suarez, De gratia, prol. 2, c. 7, n. 24 (p. 95). 5David of Conditionals," American Philosophical "Four Kinds Lewis, Counterfactuals (Oxford, 1973), p. 79f.; John H. Pollock, was defended in "A Theory law of Conditional Excluded C. Stalnaker, Middle Quarterly, vol. 12 (1975), p. 53- The by Robert Rescher American Philosophical ed. by Nicholas of Conditionals," 2, Studies in Logical Theory, Series, No. Quarterly Monograph (Oxford, 1968), p. io6f. 6 Suarez see his "De scientia Dei futurorum saw this point pretty clearly; abbreviated, DSDFC], contingentium" [hereafter Bk. 2, c. 5, n. 6 {Opera omnia, vol. 11, p. 357).

MIDDLE

AND

KNOWLEDGE

of (i) and (2) are ruled out by the that the actions of Saul and the men of assumption Keilah are and would be free in the relevant sense. The suggestion that Saul's besieging Keilah follows staying there is by logical necessity from David's in It would be more case.7 any implausible to that Saul's suggest besieging Keilah plausible from David's follows by causal necessity staying there, together with a number of other features of in fact obtained. But both of the situation which the

these tion

truth

are

suggestions that Saul's

Since

action

of

have

would

is

necessitation sort

relevant

with

inconsistent

free

the

been

with

incompatible we

will,

assump? free.

seek

might

the non

grounds for the truth of (1) and (2) in necessitating the actual intentions, desires, and character of Saul It does appear and the Keilahites. from the that Saul actually Biblical narrative intended to in Keilah if he could. Perhaps besiege David true is virtue of its correspond? by (1) proposition ence with Saul's intention. One might also suppose that (2) was true by virtue of correspondence with the desires and character of the leading men of if not their fully formed intentions. Maybe Keilah, they

were

cowardly,

and

untrustworthy,

ungrateful.

And I take it that neither the Jesuits nor Plantinga would say that Saul's intentions, or the desires and character of the Keilahites, necessitated their actions or interfered in any way with their freedom of will. But the basis thus offered for the truth of (1) and it is not because (2) is inadequate precisely A

free

agent

may or

necessitating. or change

his

Therefore

the propositions

virtue

intentions,

of correspondence

with

act fail

which the

out to

of character, on them.

act

may

be true by

intentions,

THE

EVIL

III

or what

II I trust that it is clear by this point that there is reason to doubt the possibility of middle knowledge. Those who believe it possible have some explaining to do.

In Molina's the superiority of God's explanation cognitive powers bears the heaviest burden. He holds "that the certainty ofthat middle knowledge comes from the depth and unlimited perfection of the divine knows intellect, by which [God] 9This came to certainly what is in itself uncertain." be

as

known

the

of

theory

it "supercomprehends" more stands about to

merely

pointed

them?that them

than

in rejecting to

comprehension,

than

even anyone, that.11 Molina

free

creatures

that

of

theory

comprehend

already to understand there to be understood,

necessary as Saurez

But

the

sur? that

it under?

is, be

would

them.10

comprehend

out

"supercomprehension."

intellect so immensely all created free wills,

to it God's According in its passes, perfection,

super is

something

that is about it everything and it is absurd to suppose could

God, seems

would

do

more

understand

to want

to say various

under

that

what

possible

to be known, conditions is not there, objectively, but that God's mind is so perfect that He knows it anyway. But that is impossible. The problem to be solved is how the relevant subjunctive conditionals can be true, and nothing that may be said about the

of God's

excellence

Suarez

what

offers

of that problem.

seems

to me

the

type of explanation unsatisfactory of middle possibility knowledge. to

effect,

a

no analysis,

of what to

conditionals

c, who

creature, free

primitive be

may a, which

action,

in a possible

cmay situation

a

Consider ever

exist,

freely s.We

clearly

which

understanding,

not

least

for the alleged in He appeals,

is for the relevant true.

contributes

powers

cognitive

to the solution

anything

doing

to act on, if he is (5) and (6) are enough for David but not will prudent; they satisfy the partisans of middle It is part of their theory that knowledge. God knows infallibly what definitely would happen,

OF

and not just what would probably happen free creatures would be likely to do.8

desires

and character of Saul and the men of Keilah are not (1) and (2) but in Keilah, Saul would (5) If David stayed the probably besiege city. (6) If David stayed in Keilah and Saul besieged the city, the men of Keilah would probably to Saul. surrender David

PROBLEM

do

needs

subjunctive free

possible a and

or

are

possible refrain from

to consider

c,

not as actually existing, but as having "possible being" in the cause (God) that is able to produce c. So

considered,

property

7 a similar point: DSDFC, Suarez makes Bk. 2, c. 5, n. 11 (p. 358). 8 See Bk. 2, c. 1, n. 1-2, and c. 5, n. 9 Suarez, DSDFC, (pp. 343f., 357f.). 9 Molina, Concordia, qu. 14, art. 13, disp. 53, memb. 3, n. 10 (p. 389^). 10 Ibid., qu. 14, art. 13, disp. 52, n. 11, 17 (pp. 341, 345). 11 Bk. 2, c. 7, n. 6 (p. 366L). Suarez, DSDFC,

according

to

(a habitudo, as Suarez

Suarez,

c

puts

it) which

has

a

is

PHILOSOPHICAL

AMERICAN

112

the property of being a possible agent who in s freely do a, or the property of being a in s freely refrain from possible agent who would

either would

a. c has one doing is nothing either

itself, which

property to

these

of

one

have

than

rather

properties

c the

of what cwould other. God has middle knowledge do in s, because God knows which of the two pro? c has.12

perties

one

to Suarez's but perhaps

would object philosophers of merely possible entities,

Many ontology could

a similar

develop

account

of

relevant

the

such an ontology. God's idea conditionals without an existing subject is presumably of c, for example, ascribe to it, as a of properties. And one might the property of being an idea primitive property, if itwere satisfied by anything in s, would be which, satisfied by an agent that freely did a in s. This would have the disadvantage, however, of implying that

c would

whether

in s depends,

of c, but on a property

property That

a

do

consequence

to

on

a

idea of c.

of God's

seem

might

not

?'s

compromise

My principal objection to Suarez's defense of the is not based on of middle knowledge possibility I do not think however. considerations, ontological I have

any

conception,

the sort of habitudo or property to possible

possible other for My

the

primitive relevant

reason

for

least clearly difficult

primitive

with

agents

conditions.

to

to

of

otherwise,

that Suarez

respect

their

ascribes

acts

under

subjunctive that saying

unsatisfactory someone

understanding

it would

of what

to be

conditionals Suarez's

type who

claims

which

I

be

the

it is very to have

seem

not

a

to

have.

In his several published discussions of the "free will defense" to the problem of evil, Alvin Plan tinga in effect, that God can have middle has assumed, in the most recent of these dis and knowledge;

For

proposition

knowledge.

the extent

(A) To worlds

crucial

that it is plausible, not

does

explanation

to our problem

solution

conditionals.

up-to-date solutions

form

for

we

may

that

really

give

It merely offers the expression have

us

truth

the

about

already

the possible

us

a

new

of the

a new

and

of

attempted and considered

rejected. (In fairness it should be said that Plantinga Two points must be does not claim otherwise.) here.

is to be plausible, the kinds that are possible worlds to the truth and falsity of

(i) If the explanation of similarity among to be relevant allowed conditionals

counterfactual that

would

must in any

mirror case

con?

the

our

determine

judgment of their truth and falsity. Some similarities cannot plausibly be allowed any relevance at all. the possible worlds in which David stays in Among for

example,

I suspect

the most

similar

to

is one in which Saul does not world besiege Keilah, and inwhich the subsequent history of David, Saul, and of Israel and Judah goes very much as it did in the actual world. Perhaps in such a world Saul has a slightly different character, or the actual

acts

Ill

counterfactuals."14

There are two important reasons for denying that this analysis establishes the possibility of middle

Keilah,

true.

is of

defense

is that

of

to Plantinga's (i) to be true, according theory, is for the following to be the case : is more similar to some (7) The actual world in which David stays in possible world Keilah and Saul besieges the city than to any in which David stays in possible world Keil?h and Saul does not besiege the city.

siderations

do I think that I have any

Nor

understanding

refute

or

primitive

he has defended this assumption.13 Robert and David Stalnaker Lewis, adopts what he calls "the possible worlds

Following Plantinga

made

of will.

freedom

eussions

explanation

or determine

make

would

these

of

there although to c, except the

properties, or external

internal

QUARTERLY

out

of

character

in a way

that

he

does

not

in

the actual world ;but I doubt that that is as great a dissimilarity as the dissimilarity between a world in there is a siege of Keilah which by Saul (and a world in perhaps a killing of David by Saul) and not not. I would is which there conclude, certainly not have Saul would that therefore however,

12 I believe in De gratia, prol. 2; c. 7, n. 21, 24, 25 (pp. 94-96). views come to, as they are found Suarez's this is what 13The God and Other Minds in Alvin (Ithaca, 1967), ch. 6. In his The Nature of Plantinga's assumption passed unquestioned vol. 70 The Journal of Philosophy, Created?" Could God Have (Oxford, Necessity 1974), ch. 9, and less fully in "Which Worlds I think) to free a part of his larger has attempted it is defended. At the same time Plantinga (successfully, ?1973)? PP- 539~552> has not used the term on the assumption from dependence op. cit., pp. 182-184). Plantinga {The Nature of Necessity, argument of his views. it seems to me very apt for the expression "middle although knowledge," 14The Nature I shall disregard paper op. cit. In the present op. cit., and Lewis, op. cit., p. 178. See also Stalnaker, of Necessity, as all the conditionals that will concern us have are impossible, to antecedents whose do with conditionals having complications antecedents. possible

MIDDLE

AND

KNOWLEDGE

if David had stayed in the city.15 besieged Keilah is in That a world in which Saul besieges Keilah that respect unlike the actual world, is irrelevant to the question what Saul would have done if David the Some similarities between stayed in Keilah. actual to

world

and

other

that

question?for similarities in causal

are

worlds

possible

example, laws and

in people's

dis?

characters.

But we have already considered and rejected the idea of founding the truth of our crucial conditionals on

causal

more

given accord

truth

we

than

would

that

considerations

the

must

of counterfactuals

decisiveness

to

that are allowed

to be not

be

A

mirror.

in which David and Saul stays in Keilah more to the is the similar besieges city perhaps

world actual

in

world

respect

of

Saul's

character

decisions

the

former

than

the

therefore

actual

world worlds

possible to the

explanation. us more reason

is more

latter for purposes if we

explanation, For this to

to

similar

of the

to adhere

conclusion

the

reject

mean

the

would in

analysis

terms

give of

similarity of possible worlds than to abandon our that Saul might have acted out previous judgment of character and so would only probably, not have laid siege to Keilah if David had definitely, stayed in the city. The issue here is a general one, and important. We have a well entrenched belief that

under

person

might

many have

he probably

counterfactual

would is to be

explanation decisiveness

to

acted

out

not have. plausible, similarities

conditions of

a

many

character,

If the possible worlds it must of

not

give

character

such and

behavior as to be inconsistent with this belief. the (B) On the possible worlds theory, moreover, truth of the crucial conditionals cannot be settled soon enough to be of use to God. The chief as for Plantinga importance of middle knowledge, well as Molina and Suarez, is that God is supposed to be guided by it in making decisions about the creation

and

providential

governance

of

the world.

And as Molina and Suarez insist, ifGod is tomake such use of it, His middle knowledge must be prior, 15

not

let us

create.16

as

of God

For

no

some

and

(8) If God created Adam be more moral good of

history

all.

to

many and them, to

According

Adam

are to be explained

particular, of

among

God's

knowledge,

creatures,

are

We

Adam creating instead of at

creatures

free

free

from

were

Eve

and

made.

creatures

free

decisions.

for the sake of

Adam

God

were

them

other

creating

that

choosing

; among

making

creative

that

suppose creatures

free

think

make

God's

(fallen or Unf?llen)

angels

argument, the first

Eve, and

on

depend

Ignoring

decisions

to

and

in

Eve

in part by the truth and Eve, there would than moral evil in the

the world.17

This explanation would be viciously circular if the truth of (8) were later in the order of explanation than the decisions it is supposed to help explain. Here we are dealing with a type of subjunctive that we may call deliberative conditionals. conditionals They factual.

ought For

in

not,

in asserting

strictness, one

of

to be

called

them

one

counter does

not

commit oneself to the falsity of its antecedent. That is because a deliberative is asserted (or conditional a context in of about deliberation entertained) to (try to) make true or its antecedent whether false.

although

to

creatures

what

about

reasons the truth of the conditional pro? positions which are the object of middle knowledge

a

than

Il?

similar

world inwhich David stays inKeilah and Saul does not besiege the city. But we had better not conclude that

EVIL

the theory of middle

otherwise

they

OF

if not temporally, at least in the order of explana? tion (prius ratione, as Suarez puts it), to His

alternatives

characters.

people's

the similarities

to the

relevant

on

or

laws

(ii) Even

PROBLEM

must

relevant and

similarities

THE

In

a

such

asserting

one

conditional

commits

oneself rather to the view that its truth is independ? ent of the truth or falsity of its antecedent. There

is a problem,

not been discussed to deliberative possible

Consider

of

explanation

a deliberative

(9) If I did x, y would Is (9) true? According explanation,

as

conditionals,

worlds

so far

which

in the literature,

that

depends

as

I know

has

about applying

does, Plantinga counterfactuals.18

the

conditional, happen. to the possible on

whether

the

worlds actual

world ismore similar to some world inwhich I do a: and y happens than to any world in which I do # and y does not happen. That in turn seems to

are discussed Similar problems The Nature of Necessity, and Lewis, Counterfactuals, by Plantinga, op. cit., pp. 174-179, op. cit., 72-77. 91-95 16 See Bk. 2, c. 4, n. 6, and c. 6, n. 3, 6 (pp. 355, Suarez, DSDFC, especially 361, 363). 17 I have simplified in the antecedent. is supposed to have known God that there would here, particularly be more moral good than moral evil in the world a ifHe executed with the creation of Adam and Eve. Many of these long series of actions, beginning actions would be occasioned in part by responses He supposedly knew creatures would to earlier actions in the series. freely make 18 Stalnaker would it to deliberative too. Lewis might conditionals apply not; see his Counterfactuals, op. cit., p. 4. PP-

PHILOSOPHICAL

AMERICAN

114

is the actual world. And depend on which world which world is the actual world ? That depends in I do x. Thus the truth of (9) seems part on whether to depend on the truth or falsity of its antecedent. the truth of (8) will depend on whether Similarly creates

God

Adam

and

Eve.

I think it may be possible for a possible worlds to overcome this conditionals theory of deliberative in general, but not in such a way as to difficulty rescue the doctrine of middle knowledge. There is, I presume, a large class, K, of possible worlds that are more

to some

similar

x and

I do

in which

world

being

member

of K,

on which

not

but

member

of K it is. In asserting (9) in the context of delibera? in effect, to the view that the tion I commit myself, actual world is a member of K and that itsmember? ship in K does not depend on which I choose of the This I am deliberating. alternatives among which view

may

well

are

linked

by

for

be a

correct?if, causal strict

x and

instance,

y

law.

and

Adam

and

evil

in the history

how

on

decision

God's some

to

world the

possible

world's in the

earlier

settled

whether free

other

moral

moral

truth of (8)

The

being worlds

moral

than

good

of the world.

actual

according can the actual

been

or

the

is more

there

and

more

is not

there

Eve

depends of K*,

Eve

evil in the history of the world in which God creates Adam and

good than moral than to any world

some

member But

theory. in K*

have

membership of explanation order to create and Adam or none

creatures,

than

we

?Here

Eve, face

and all the old difficulties about middle knowledge, the possible worlds theory does nothing to help us answer this question. At most it explains why (8) is true,

given

that

some

of K*

member

us

Let

decision.

say

that

one

of

God's

is represented inK* if and only if there is alternatives that some world in which He in K* chooses If any

alternative.

the

of

alternatives

among

which

God was choosing is not represented inX*, then the in K* depends on His actual world's membership that

rejecting

prior 19

alternative,

and

Plantinga

insists on

this point

cannot

therefore

in the order of explanation

was

unrepresented no free to make a world

see how

of God's

in K*.

at

creatures

in which

are

actions

that

possible

world,

all, no

are

there

be a member

at all could

alternatives one

For

of K*.

is

alternative not

I do

and

creatures

free

it is free

Since

no evil,19 good morally be a member of K* unless and

morally w, will

a in

there is some feature of w by virtue of which in the free actions of free creatures difference some u

more

are

similar

And the

involve no

v would

u and

worlds

as

free

a reason

be v

than

in w

creatures

at

creatures.

of free

in w, what

all

counting relevant

(in

of w must

such feature

any

existence

for

w

to

surely If

would

there make

w more like a world in which most free creaturely decisions are good ones than like a world in which are bad ones? I most decisions free creaturely in K* conclude that the actual world's membership cannot

be

God's

in

earlier

order

the

some

to make

decision

of free

than

explanation creatures.

There?

fore the truth of (8), on the possible world's analysis, to that cannot be prior in the order of explanation Perhaps it will be objected to me that the parti? sans of middle knowledge need not claim that the truth of (8) precedes God's creative choices in the is enough

It

of explanation.

order

for

their

explana?

the choices. believed (8) prior to making on a in (8) if acted is that belief God My reply before it was settled that (8) is true, then the fact (if it is a fact) that there ismore moral good than tions ifGod

moral

evil

good

luck

in the history than rather

the world

of His

to God's

is due

the

wisdom?whereas

of the theological chief motivation theory of middle that has been the desire to maintain knowledge such happy results of God's dealings with created and that He had freedom are due to His wisdom, no need at all of luck.

is actual.

there is reason to believe that the Furthermore inK* cannot have been actual world's membership than settled earlier in the order of explanation God's

indeed

decision.

there is a class, K*, of possible worlds Similarly that are more similar to some world in which God creates

least one

I think at

But

respects).

y happens than to any world in which I do x and y to the possible worlds does not happen. According of the truth (9) depends on the actual world theory some

QUARTERLY

to His

(The Nature

IV be that may the philosophical Of objections raised against my critique of the theory of middle (A) I have cumstances

be

(i)

If David besiege

op. cit., p.

i66f.).

the most

in our

assumed

and

intentions

to me

seem

relied on the claim

and Saul at Keilah,

decision.

of Necessity,

two

knowledge,

what

character

stayed the city,

example

important.

that

in the cir?

about

David

is true by virtue of Saul's is not

in Keilah,

Saul

would

MIDDLE

AND

KNOWLEDGE

to

but in Keilah, stayed (5) If David probably besiege the city. Suarez has an interesting objection He

THE

in effect,

argues,

(1) is probably commits

that

(5)

can

Saul

would

mean

only

albeit

oneself,

some

with

that

(5) one to

trepidation,

of (i).20 Certainly it would be prag? 1 to assert inconsistent that is matically ( ) probably true and deny (as I do) that there is any way in which (1) can be true. to (1), how? In proposing (5) as an alternative as a not I it understand do claim that (1), or ever, the

any

such

other

is probable.

proposition,

It

a

is rather

that

(10) Saul will besiege Keilah would be probable, given facts that would (definitely, not just probably) obtain ifDavid stayed inKeilah. is an epistemological While term, "probably" it in is used moreover, (5) primarily to characterize dispositions or tendencies toward the truth of ( 1o) that there would be ifDavid stayed in Keilah. (5) does not imply that anyone would know the facts that would probabilify (10), but only that they would obtain, ifDavid stayed in Keilah. This view is consistent with treatment that (5) might receive under either of the two major types of theory of counterfactuals distinguished by Lewis. to a metalinguistic According "a counterfactual is true, it,

only

if its antecedent,

further

premises,

as Lewis theory, or assertable,

with

together its

implies

puts if and

suitable 21

consequent."

Hold?

say that (5) is ing a theory of this type, we might true if and only if (10) would be probable on total evidence constituted of (5), by the antecedent with

together

suitable

able further premises about

Saul's

proposed similar

further

intentions

and

character.

be partly Lewis

has

for the possible worlds theory an essentially treatment

of

counterfactuals

that

in the way that (5) does.22 probability the most serious grounds (B) Probably givings in which

suit?

The

premises.

in this case would

about we

like a piece

my seem

argument to have

of middle

be may confidence

knowledge.

found in what

Suarez

and

the

of

part

ordinary us

to provide

endeavors

Plantinga

thinks

Plantinga

answer

"the

seems

clear:

fairly

indeed

the larger [Smith] would have" accepted bribe;24 and I agree. But what makes (11) true ? Let us note that it to of the class belongs subjunctive conditionals with antecedents

have

that

makes

false have

and

consequents called semi

been

(11) true, I think,

makes is true

consequent not would

be

which

true,

factuals. What

event

to

assumed to be

assumed

and

the

truth

or made prevented, the consequent

is that its

its antecedent

of

less true.

the

likely, My

view

here is in accord with Nelson Goodman's claim that "in practice full counterfactuals affirm, while semifactuals tains

a

that deny, antecedent

between

certain

connection

ob? 25

and

consequent."

My

account of what makes (11) true does not suggest a or in which (1) (2) could be true, since they

way

not

do

true

have

to

consequents

help

make

them

true. account

if my

Furthermore,

is

right,

it was

not settled

that (11) is true before (in presumably the order of explanation) it was settled that Smith was his

going actual

true.

God

I

to be

it as He

for mis?

and

offered,

is part

acceptance see no reason,

could have

enough

perplexes to be any

looks

on

$36,000.

ledge. Another

appeals

II5

(11) Smith would still have accepted a bribe to if the bribe had been drop his opposition,

involve

in cases

EVIL

with convincing examples of it. In one of Plantinga's fictitious examples Curley Smith, a mayor of Boston, has accepted a bribe of to a proposed free? $35,000 to drop his opposition case route. In this is the way following true ?

truth

claim

OF

confidence

speakers,23

to this claim.

true, and that in accepting

PROBLEM

in

the

is supposed

me

therefore,

known

order

of

to make

at

to

since (11) that

suppose

of the truth of (11) early explanation

not type of case, more. There

uncertainty

accept, $35,000, of what makes

all

to make

use of middle

presented not does about

use

of

know?

by Plantinga, seem normally a butcher, what

for example, would have done if I had asked him to sell me a pound of ground beef, although we suppose he would have had free will in the matter.

20 Bk. 2, c. 5, n. 9 (p. 357f.). I am simplifying Suarez, DSDFC, here, but I think not in such a way as to make less plausible. 21 Lewis, Counterfactuals, op. cit., p. 65. 22David "Counterfactuals and Comparative Journal of Philosophical Lewis, Probability," Logic, vol. 2 (1973), 23 Bk. 2, c. 5, n. 8 (p. 357). Suarez, DSDFC, 24 The Nature of Necessity, Plantinga. op. cit., p. 177. 25Nelson Goodman, Fact, Fiction, and Forecast(London, 1954), p. 15.

this argument

P- 437f

AMERICAN

Il6

PHILOSOPHICAL

We say he would certainly have sold me the meat, if he had it to sell. What makes us regard it as certain? Chiefly his character, habits, desires, and and

intentions,

the

me.) There

are

three

dis?

countervailing

have had no motive

(He would

positions.

of

absence

one

views

alternative

to refuse take

might

of this case. One might say that if I had asked the to sell me the meat, butcher (i) he would only probably have sold it to me, though we normally but real chance there would ignore the minute or have been that he would refuse; (ii) he would sold me

have

certainly

I have the

an

rested

of my part a person's

important that what

assumption

that

however,

?although content

with

causal

Humean

For

and

sound

nature

of

theory

views?

and

impossible;

that

would

render

it

V Could

would

to have

arranged

free actions

perform

us

Let

have

the

consider

but first

question

creatures

only on

the

who

right ones? assumption

In that case, we has middle knowledge. sinless free think, He could have obtained might creatures only those that He simply by making knew would always freely do right in those situa? tions inwhich He would permit them to act freely.26 that God

response

Plantinga's

to

this

argument,

he develops with much which than I have space to reproduce not

could creatures

26This vol.

do who

this

there

unless

would

in

fact

are

a

response

greater elegance here, is that God some

behave

is crisply stated by argument replaces (1966), p. 93L* "Will" 27 The Nature of Necessity Plantinga, 28 Bk. 2, c. 4, n. 4 Suarez, DSDFC, 63

(or their speak? has

Roughly its essence)

if and

sense,

trans-world

must

therefore

inescapable

have of

price

have

and

knowledge, some

known

that

created

freedom.

it

that

must

God

true,

does not claim

even

or

true,

is

depravity

it is true, ifHe had middle

known

was

evil

the

that the hypothesis

is plausible.27

He

argues

is

only

that it is logically possible, because he is using it to the view that it is logically possible defend that both God and evil exist. I do not doubt that the latter is logically possible; but religious thought an

seek

account

the

of

between

relations

God

and evil that is credible, as well as logically possible. It is worth asking, therefore, whether the hypo? thesis

of

universal

trans-world

is plaus?

depravity

ible, on the assumptions about truth of conditionals that Plantinga shares with the Jesuit theologians. I think Molina and Suarez would deny that any creature

free

possible

has their

support Suarez

that

. . . and

to

and

omnipotence, to

enough,

depravity with persuasive to "it is alien

denial

holds

the

divine

creature's could

arguments. common the and

perfection

cannot

God

they

of itself

is therefore that

say

free

any

(or

; and

trans-world

incredible

predetermine

free act, in particular [praedefinir?] an honorable and with all [its] circumstances, by His absolute and effective will, the freedom of the created will still being preserved."28 God uses his middle knowledge

intelligible.

God

free creatures

if God only if that creature would do some wrong created it and permitted it to act freely, no matter what else God did. If the hypothesis of universal

doctrine

and

the

proposes

Plantinga

have trans-world essences) depravity. a creature free possible ing, (or in trans-world depravity, Plantinga's

essence)

(iii), if it is

probability

(iii) is plainly

any

nature

is the

some

On

the

of

know

what

in alternative

determination?

alternative

not

them.

certain

views?of

causality, I do

is

assumption

of

either

of the rendering not

my

than (i) or (ii) (iii) ismore implausible I must admit that I am not altogether

alternative

would.

that all possible

hypothesis

must

argument character

and dispositions do not causally determine, they do not render absolutely certain. Alternative (iii) is It still seems to inconsistent with this assumption. me,

none

perhaps

Plantinga

to do so by his have been causally determined or (iii) his character and character and dispositions; not have causally determined dispositions would his action, but they render it absolutely certain that he would have complied with my request. on

that

he would

because

the meat,

QUARTERLY

possible so well,

free and

to make

such

elicit

a

under

which

favorable He

able possible

free

act

outward

divine

the that

course, of

of grace

response, that knows to

act

according of supposes,

effective,

predeterminations

and helps

conditions

every

for

possible

circumstances,

that He

choosing

knows will

and the

avoiding creature would

purpose. every free

This

not pre? honor?

possible creature, there

those

are

in any some

incentives or helps of grace that God could supply, to which the creature would respond favorably he could have responded unfavorably. But though ifwe assume, this is a very plausible presupposition as Suarez does, that the theory of middle knowledge

on the Free Will Defense: A Reply," Nelson The Journal of Philosophy, Pike, "Plantinga "would" in Pike's formulation, but it is clearly middle that is involved. knowledge p. 165 with p. 189. (op. cit.). compare (p. 354).

AND

KNOWLEDGE

MIDDLE

is correct, and that there is an infinite variety of natural and supernatural ways in which God can work

us

on

affecting desires,

without

our

assisting

inwardly,

our feelings

causally

reasoning,

our beliefs

and perhaps

and

our

determining

res?

ponse.29

And possible

to suppose

if it is plausible

of

occasion

particular

that for every

action

are

there

a to

that would elicit divine operations possible favorable free response, is it not also plausible suppose

that

for many

free

possible

and

creatures,

THE

not

for whole worlds full of them, there are to which series of divine those operations possible creatures would respond by always freely doing held that both right, never doing wrong? Molina and

from

preserved

all

elicit

a favorable

could

He

Presumably

free

response

have

done

from

the

(14)

If the hypothesis trans-world of universal is implausible, itmight seem that I offer depravity I deny the possibility theodicy a better alternative. I deny that the of middle because knowledge, are

conditionals

subjunctive

true.

In

I deny that the following is true: particular, If God had acted differently in certain (12) ways, made

He

would

made

wrong

free

In other words,

have but

choices,

had

creatures

none

of whom

who ever

choices.

I deny

that God

could have made

made

free

made

wrong

none

but

choices,

behaved

free

of whom

choices.

If God had acted differently ways, He would probably have He

for

U7

If God had acted differently in certain ways, He would probably have had creatures ever

them.30

same

EVIL

true:

who

sin

others.

relevant

be

(13)

their whole lives by God supplying and aids that He knew would gifts

throughout them with always

were

Mary

OF

free creatures who would always have freely done right. The supposition that He could have done so is burdened with all the difficulties about truth of that afflict conditionals the theory of middle Since (12) is not true, a reproach knowledge. against God cannot rightly be based on its truth. And God cannot know that (12) is true, and cannot rightly be blamed for not using such knowledge. how? My views about the truth of conditionals, ever, do not tend to show that the following could

even

Jesus

PROBLEM

creatures,

on

in certain had better than

the whole,

has.

actually

In fact ( 13) seems to me rather implausible. With? out middle knowledge God must take real risks if He

makes

free

creatures

or

millions,

(thousands,

trillions of risks, if each free creature makes thousands of morally significant free choices). No matter so in running how shrewdly God acted on not risk His would be many risks, every winning But I think (14) is very antecedently probable. These the that plausible. suggest judgments necessity free will

of permitting in creatures

some may

play

evil a

in order part

in a

to have theodicy

put cannot bear the whole weight of it, even is rejected.31 possibility of middle knowledge

if the

University of California, Los Angeles Received January 14, igj6

29 Cf. ibid., Bk. 2, c. 4, n. 5 (p. 355). 30 Molina, Concordia, qu. 14, art. 13, disp. 53, memb. 4, n. 15-24 (pp. 399-405). 31 I am to several, indebted and especially David Lewis and Alvin Plantinga, for discussion and for including David Kaplan, comments on an earlier version of this paper, which was read to an American An abstract Association Philosophical symposium. of the earlier version, "Middle on the in The Journal of Philosophy, vol. 70 (1973), pp. 552-554. Work Knowledge," appeared Endowment for the Humanities. present version was supported by the U.S. National