Hostname: page-component-745bb68f8f-l4dxg Total loading time: 0 Render date: 2025-02-06T04:00:30.832Z Has data issue: false hasContentIssue false

Luther's necessitarian argument in De servo arbitrio

Published online by Cambridge University Press:  28 January 2015

ANDERS KRAAL*
Affiliation:
Department of Philosophy, University of Calgary, 2500 University Drive NW, Calgary, AB T2N 1N4, Canada e-mail: akraal@ucalgary.ca
Rights & Permissions [Opens in a new window]

Abstract

In De servo arbitrio (1525) Luther famously argues that the divine attributes of will, power, foreknowledge, and immutability are incompatible with (human) free will, and hence that free will is a ‘name with no reality’. I survey some earlier explications of Luther's argument in the literature, and reject them as exegetically unsound. I then go on to propose a new explication. On the proposed explication, Luther's argument turns out to be theologically cogent, provided that we follow Luther in understanding the relevant divine attributes in accordance with Augustinian theology.

Type
Articles
Copyright
Copyright © Cambridge University Press 2015 

At an early stage of the Reformation, in his Disputatio Heidelbergae habita (1518), Luther denied that humans have ‘free will’ (liberum arbitrium) in their ‘fallen state’ (post peccatum).Footnote 1 Following Leo X's condemnation of this claim in his papal bull Exsurge Domine (1520),Footnote 2 Luther expanded on the claim, saying now, in his Assertio omnium articulorum (1520), that free will is a ‘name with no reality’ (titulus sine re), and, indeed, that ‘all things occur by absolute necessity’ (omnia . . . de necessitate absoluta eveniunt).Footnote 3 Call this ‘Luther's necessitarian thesis’.

When Erasmus decided to take a public stance against Luther with his book De libero arbitrio diatribe sive collatio (1524), he made Luther's necessitarian thesis one of his chief grounds of complaint.Footnote 4 Taking free will as a theological given, Erasmus branded inquiries into ‘whether God's foreknowledge is contingent’ (an deus contingenter praesciat aliquid) or ‘whether everything we do is done out of mere necessity’ (an quicquid facimus . . . mera necessitate faciamus) as ‘irreligious curiosity’ (irreligiosa curiositate).Footnote 5 In so doing, Erasmus was alluding to inquiries into the relation between God's foreknowledge and free will that had been going on for centuries in Christendom, and which can be found in well-known theological treatises such as Augustine's De civitate dei, Boethius's Philosophiae consolationis, Peter Lombard's Sententiae, and Thomas Aquinas's Summa theologiae, to mention only some better-known examples.

Luther strongly disagreed with Erasmus's way of branding the relevant inquiries.Footnote 6 In his long and spirited response to Erasmus in De servo arbitrio (1525), Luther affirmed that it is both ‘wholesome’ (salutare) and ‘fundamentally necessary’ (imprimis necessarium) for Christians to know that God ‘foreknows nothing contingently’ (nihil praescit contingenter) but instead ‘foresees, purposes, and does all things according to His own immutable, eternal and infallible will’ (omnia incommutabili et aeterna infallibilique voluntate et praevidet et proponit et facit).Footnote 7 Luther then goes on to defend his necessitarian thesis by means of ‘argument’ (disputatione).Footnote 8 The argument proceeds from a few premises about God's ‘nature’ (natura) and arrives at the necessitarian thesis as its conclusion. Call this argument (to be looked at more carefully in the body of this article) ‘Luther's necessitarian argument’.

Luther never retracted his necessitarian thesis or necessitarian argument. Indeed, in his later years he referred to De servo arbitrio as one of his very best books.Footnote 9 Modern Lutheran theologians would for the most part probably disagree with this estimate of De servo arbitrio.Footnote 10 Nevertheless, De servo arbitrio has attracted a lot of scholarly attention, perhaps especially from German and Scandinavian Luther scholars.Footnote 11 Indeed, as Robert Kolb has noted, there appears to be a ‘modern fascination’ with the book, at least amongst Protestant theologians.Footnote 12 However, and as we shall see in more detail further on, there is little agreement on exactly how Luther's necessitarian argument in De servo arbitrio is supposed to work, and the main explications of the argument found in the literature leave us with an argument which is, on the whole, quite unimpressive, at least from a logical point of view.

I believe this lack of agreement on the structure of Luther's necessitarian argument is due to the fact that commentators haven't paid sufficiently close attention to what Luther actually says in the passages of De servo arbitrio in which the argument is developed. Once we pay close attention to what Luther actually says, I believe the basic structure of the argument becomes reasonably clear, and some of the main alternative explications of the argument in the literature can be seen to be manifestly exegetically unsound. Interestingly, and perhaps surprisingly, Luther's argument, once clarified, is quite impressive, at least in view of his largely Augustinian theological starting point. My aim in this article will be to substantiate these claims.

The argument in outline

As was noted above, Luther took it to be fundamentally necessary and wholesome for Christians to know that God ‘foreknows nothing contingently’, but ‘foresees, purposes, and does all things according to His own immutable, eternal and infallible will’. What Luther means in saying that God foreknows nothing contingently can be explained as follows. Suppose that a certain man will go fishing at some future point, and that God in his foreknowledge knows this. Now if it were solely in the man's power to decide whether to go fishing, then, of course, this wouldn't be in God's power. So if it is solely in the man's power to decide whether to go fishing, God's foreknowledge would be contingent on a certain man's decision or action. Hence, in denying that God's foreknowledge is contingent, Luther is in effect denying that there are scenarios in which people have power of the relevant sort. And in denying this, Luther wishes to vindicate the contrary belief that all power resides in God's unchangeable, eternal, and infallible will.

Luther goes on to characterize the doctrine of the non-contingency of God's foreknowledge as a ‘thunderbolt’ (fulmine) which ‘knocks “free will” flat, and utterly shatters it’ (sternitur et conteritur penitus liberum arbitrium).Footnote 13 As becomes clear further on in the same paragraph, the way in which Luther takes this doctrine to ‘shatter’ free will is by means of a logical implication. Luther's claim is that the doctrine of God's foreknowledge and will, or some relevant aspect of this doctrine, logically implies that humans don't have free will.

What Luther means by ‘free will’ (liberum arbitrium) is explained a bit further on, where he takes the term to signify a ‘power of freely turning in any direction, yielding to none and subject to none’ (libere possit in utrunque se vertere, neque ea vis ulli caedat vel subiecta sit), which is also taken to be the ‘ordinary sense’ (natura vocabuli) of the term.Footnote 14 Clearly, if humans possess free will in this sense, then not even God will have power over their free choices, for the expression ‘subject to none’ excludes also God. Luther's contention, then, is that humans don't possess a freedom to turn in any direction without being subject to God in so turning.

The argument that Luther adduces in support of his necessitarian thesis can be divided into two steps. In the first step he argues that Erasmus's own words carry commitments to the unchangeability (and hence non-contingency) of God's will and foreknowledge. In a second step he argues that the unchangeability of God's will and foreknowledge entail that humans don't have free will.

The first step of Luther's argument runs as follows:

[18] . . . Nonne tu es mi [19] Erasme, qui asseruisti paulo ante, Deum natura iustum, natura clementissimum? [20] Si hoc verum est, nonne sequitur, quod incommutabiliter sit iustus [21] et clemens? ut quemadmodum natura eius non mutatur inaeternum, ita nec [22] eius iusticia et clementia. Quod autem de iusticia et clementia dicitur, etiam [23] de scientia, sapientia, bonitate, voluntate et aliis divinis rebus dici oportet. [24] Si igitur haec religioso, pie et salubriter de Deo asseruntur, ut tu scribis, [25] Quid accidit tibi, ut tibi ipsi dissidens, irreligiosum, curiosum ac vanum nunc [26] asseras, dicere, Deum necessario praescire? Scilicet voluntatem [27] immutabilem Dei praedicas esse discendam, immutabilem eius vero praescientiam [28] nosse vetas. An tu credis, quod nolens praesciat, aut ignarus [29] velit?Footnote 15

. . . Surely it was you, my good Erasmus, who a moment ago asserted that God is by nature just, and kindness itself? If this is true, does it not follow that He is immutably just and kind? that, as His nature remains unchanged to all eternity, so do his justice and kindness? And what is said of His justice and kindness must be said also of His knowledge, His wisdom, His goodness, His will, and the other Divine attributes. But if it is religious, godly and wholesome, to affirm these things of God, as you do, what has come over you, that now you should contradict yourself by affirming that it is irreligious, idle and vain to say that God foreknows by necessity? You insist that we should learn the immutability of God's will, while forbidding us to know the immutability of His foreknowledge! Do you suppose that He does not will what He foreknows, or that He does not foreknow what He wills?Footnote 16

Luther's line of reasoning in [18]–[23] of this passage appears to be as follows: God has attributes like justice, kindness, foreknowledge, wisdom, goodness, and will; if one concedes that any of God's attributes are unchangeable, then one is committed to holding that all of God's attributes are unchangeable;Footnote 17 Erasmus concedes that God's justness and kindness are unchangeable; hence, Erasmus is committed to holding that God's foreknowledge is unchangeable. In [24]–[29] Luther proceeds to expose the arbitrariness of Erasmus's view that the justice and kindness of God can be affirmed to be unchangeable while leaving out the foreknowledge of God.

Luther proceeds next to the second and more important step of his argument, urging straightforwardly that humans lack free will. Says Luther:

[29] Si volens praescit, aeterna est et immobilis [quia natura] voluntas, [30] si praesciens vult, aeterna est et immobilis [quia natura] scientia.

[31] Ex quo sequitur irrefragabiliter, omnia quae facimus, omnia quae fiunt, [32] etsi nobis videntur mutabiliter et contingenter fieri, revera tamen fiunt [33] necessario et immutabiliter, si Dei voluntatem spectes. Voluntas enim Dei [34] efficax est, quae impediri non potest, cum sit naturalis ipsa potentia Dei, [35] Deinde sapiens, ut falli non possit. Non autem impedita voluntate opus [36] ipsum impediri non potest, quin fiat loco, tempore, modo, mensura, quibus [37] ipse et praevidet et vult.Footnote 18

If He wills what he foreknows, His will is eternal and changeless, because His nature is so. From which it follows, by resistless logic, that all we do, however it may appear to us to be done mutably and contingently, is in reality done necessarily and immutably in respect to God's will. For the will of God is effective and cannot be impeded, since power belongs to God's nature; and his wisdom is such that He cannot be deceived. Since, then, His will cannot be impeded, what is done cannot but be done where, when, how, as far as, and by whom, He foresees and wills.Footnote 19

In seeking to get clear about the structure of this argument, I will first identify what appear to be the core claims and assumptions of the argument, and then proceed to identify what appear to be the core inferences of the argument. Having done that, in the next section I will propose a systematization of the deductive structure of the argument.

Among what appear to be the core claims of the argument are the following ten. A first claim, found in [29] and [30], is that God's essence (natura) is changeless (immobilis) (‘claim 1’). Three related claims, or assumptions, manifest in [29], [30], and [34], are that God has will (voluntas) essentially (‘claim 2’), that God has knowledge (scientia), and hence foreknowledge, essentially (‘claim 3’), and that God has power essentially (‘claim 4’). Two further claims, found in [29] and [30], are that God's will is changeless (immobilis) (‘claim 5’) and that God's foreknowledge is changeless (immobilis) (‘claim 6’).Footnote 20 (This last point is neglected in Packer and Johnston's translation.) In [29] it is claimed that God ‘wills what he foreknows’ (volens praescit) (‘claim 7’). In [31] and [36]–[37] we have the central claim that everything is brought about by him who foresees and wills (quin fiat loco, tempore, modo, mensura, quibus ipse et praevidet et vult) (‘claim 8’); and in [33]–[37] we have the claim that the will ([v]oluntas) of God is effective (efficax) and cannot be impeded (impediri non potest) (‘claim 9’). A final point, which isn't brought out explicitly in the above passage but is clear from Luther's statement of the outcome of his argument considered earlier, is that humans lack free will (‘claim 10’).

The above claims are related to each other inferentially. The main inference clearly occurs in [31], and is marked by the expression ‘[f]rom which it follows, by resistless logic’ ([e]x quo sequitur irrefragabiliter). This inference appears to consist in the inference of claim 8 from claims 5, 6, 7, and 9. Related to this is the inference of claim 10 from claim 8. Three further inferences that appear to occur in the argument are the inferences of claim 5 from claims 1 and 2, of claim 6 from claims 1 and 3, and of claim 9 from claim 4. It would appear that any exegetically sound explication of Luther's argument would need to do justice not only to the above-surveyed claims, but also to these inferences.

At first sight it isn't entirely obvious how Luther's argument is supposed to work. The crucial inference of claim 8 from claims 5, 6, 7, and 9 appears to draw heavily on considerations pertaining to the unchangeability of God's will and foreknowledge, but it isn't clear what role this consideration is supposed to play in the relevant inference. Nevertheless, it appears to me that we have, under the surface structure of the above passage, an argument with a respectable deductive structure.

Problems with earlier explications

Given the above outline of Luther's argument, it would seem that any explication of the argument that can qualify as exegetically sound must do justice to the fact that the argument uses as premises the claims that God's essence is unchangeable (claim 1), that God has will essentially (claim 2), that God has foreknowledge essentially (claim 3), that God has power essentially (claim 4), and that God wills what he foreknows (claim 7). The explication must also do justice to the fact that the argument regards as logical consequences of these premises the claims that God's will is unchangeable (claim 5), that God's foreknowledge is unchangeable (claim 6), that God brings about whatever he wills (claim 9), that all things take place by the God who foreknows and wills (claim 8), and that humans don't have free will (claim 10).

None of the main explications of Luther's argument in the literature comes anywhere near to meeting these criteria of exegetical soundness, however. To support this, I shall in what follows briefly survey the explications of the argument found in three major studies: Harry McSorley's Luthers Lehre vom unfreien Willen (Reference McSorley1967), the standard modern study of Luther's necessitarian argument; Linwood Urban's ‘Was Luther a thoroughgoing determinist?’ (Reference Urban1971), perhaps the most important study of De servo arbitrio after McSorley's book; and Robert Kolb's Bound Choice, Election, and the Wittenberg Theological Method (Reference Kolb2005), the most recent major study of De servo arbitrio.

McSorley states Luther's necessitarian argument in a preliminary way and in a more systematic way. His preliminary statement of the argument is as follows:

Luthers necessitaristisches Argument gegen den freien Willen kann folgendermaßen ausgedrückt werden: Gott sieht nichts kontingenterweise vorher; er sieht alles vorher, nimmt sich alles vor und tut alles mit unveränderlichem, ewigem und unfehlbaren Willen. Deshalb wird der freie Wille ‘ganz und gar vernichtet’.Footnote 21

McSorley's more systematic statement of the necessitarian argument takes the argument to proceed from a major and a minor premise:

Luthers Obersatz . . . ist: Gott weiß alle Dinge vorher. / . . . / Es ist jedoch ein Untersatz in diesem Argument mit inbegriffen . . . nämlich die Behauptung: Was immer Gott vorhersieht und vorherweiß muß sich mit Notwendigkeit ereignen, andernfalls könnte Gott irren.Footnote 22

Both statements of the argument are clearly exegetically unsound. There is no mention in either statement of the premises that God's essence is unchangeable (claim 1), that God has will essentially (claim 2), that God has foreknowledge essentially (claim 3), that God has power essentially (claim 4), or that God wills what he foreknows (claim 7), for example. Nor is any real effort made to show how the various premises of Luther's argument are related inferentially. Indeed, McSorley's statement of Luther's necessitarian argument looks more like a brief summary of the argument than as an explication of it.

Urban seeks to make sense of Luther's necessitarian argument by stating it in two allegedly equivalent ways. His first way of stating the argument is as follows:

In this view a complete determinism is thought to follow deductively from the nature of God. The argument can be sketched as follows:

If God wills an event, then the event necessarily takes place.

God necessarily wills the event.

Therefore, the event necessarily takes place.

. . . The first premise is established by the additional claim that God's will cannot be impeded.Footnote 23

Urban's second way of stating the argument is as follows:

If God knows p, then p is necessarily true.

God necessarily knows p.

Therefore p is necessarily true.Footnote 24

The second of these statements of Luther's necessitarian argument is clearly exegetically unsound, for, and as we have seen, it is crucial to this argument that the necessity of all events follows not simply from God's foreknowledge of all events, but from God's unchangeable foreknowledge of all events, and this is something that Urban's second statement leaves out. Urban's first statement of the argument is better: it is true that Luther's argument involves claims equivalent to the claims that ‘if God wills an event, then the event necessarily takes place’ and that ‘God necessarily wills the event’. However, there is much that is missing here too, such as, once again, any reference to the crucial role that the unchangeability of God's foreknowledge and will plays in Luther's argument.

Kolb's explication of Luther's necessitarian argument is quite complex. It involves distinguishing between what Luther said and what Luther intended to say in the relevant passage of De servo arbitrio. Talk of the necessity of events in the relevant passage is viewed as an ‘experiment’ on the part of Luther ‘in expressing his views’.Footnote 25 What Luther hereby intended to say, we are told, is ‘what it means that God continues to be Creator with absolute and unconditioned power as befits a Creator’.Footnote 26 There are at least two problems with this suggestion, however. A first problem is that it is grounded in a claim for which no adequate exegetical support is provided; namely, that in saying that all things take place of necessity Luther only intended to express what it means for God to be the creator.

A second problem is that the interpretation doesn't harmonize with what Luther actually says in the passage, for, and as we have seen, Luther talks in the relevant passage of the necessity of events as something that ‘follows’ (sequitur) from God's will and foreknowledge by means of an ‘argument’ (disputatione), and nowhere suggests that this is just a way of ‘expressing’ what it means for God to be the creator. Accordingly, Kolb's explication is exegetically unsound.

A new explication of the argument

I now proceed to offer an explication of Luther's necessitarian argument that aims at doing justice to the fact that the argument uses claims 1, 2, 3, 4, and 7 as premises, and which treats claims 5, 6, 8, 9, and 10 as logical consequences of these premises. By meeting these aims, I believe we can bring out the logical structure of Luther's necessitarian argument in a way that is exegetically sound.

My explication takes the argument to consist of five premises and ten consequences. Using ‘P’ for ‘premise’ and ‘C’ for ‘consequence’, I suggest the argument be understood as follows:

  • (P1) God's essence is unchangeable.

  • (P2) God has will essentially.

  • (P3) God has foreknowledge essentially.

  • (P4) God has power essentially.

  • (P5) God wills what he foreknows.

  • (C1) Hence, God foreknows all things. (From P3.)

  • (C2) Hence, God brings about whatever he wills. (From P4.)

  • (C3) Hence, God's will is unchangeable. (From P1, P2.)

  • (C4) Hence, God's foreknowledge is unchangeable. (From P1, P3.)

  • (C5) Hence, God foreknows all things unchangeably. (From C1, C4.)

  • (C6) Hence, God wills what he foreknows unchangeably. (From P5, C4.)

  • (C7) Hence, God wills all things. (From C5, C6.)

  • (C8) Hence, God wills all things unchangeably. (From C3, C7.)

  • (C9) Hence, God brings about all things unchangeably. (From C2, C8.)

  • (C10) Hence, humans lack free will. (From C9.)

As can be seen, P1 accords with claim 1, P2 with claim 2, P3 with claim 3, P4 with claim 4, P5 with claim 7, C3 with claim 5, C4 with claim 6, C9 with claims 8 and 9, and C10 with claim 10. Moreover, the explication captures the inferences of claims 8 and 10 from claims 5, 6, 7, and 9, of claim 5 from claims 1 and 2, of claim 6 from claims 1 and 3, and of claim 9 from claim 4. The above explication accordingly fulfils our criteria for exegetical soundness.

The cogency of the argument in view of Augustinian theism

As I will be using the term ‘logical validity’ (or just ‘validity’), an inference is logically valid if and only if it proceeds from a proposition to another proposition by means of a relation of logical consequence, where a relation of logical consequence holds if and only if the inferred proposition necessarily follows from the proposition or propositions from which it is inferred. I take this usage to conform to traditional usage. Aristotle, the founder of the study of logical validity, fleshed out logical validity as a relation according to which ‘when certain assumptions are made, something other than what has been assumed necessarily follows from the fact that the assumptions are such’.Footnote 27 This understanding of logical validity passed on into the mediaeval logical tradition,Footnote 28 a tradition with which Luther was thoroughly acquainted.Footnote 29

In the mediaeval logical tradition, validity was fleshed out in two ways: in terms of ‘formal consequences’, which depend only on the forms of propositions as determined by the syncategorematic or non-referring terms, and in terms of ‘material consequences’, which depend also on the content of propositions as determined by the categorematic or referring terms. The latter kind of validity can be said to hold by virtue of the meanings of referring terms, or, alternatively, by virtue of the nature of things, i.e. by virtue of the nature of the things referred to by the referring terms in the relevant propositions.Footnote 30

Even a quick glance at Luther's necessitarian argument makes it clear that Luther must have viewed the various inferences of his argument to be valid in the material sense rather than in the formal sense, for the argument obviously turns on what it means for God to have ‘foreknowledge’, ‘will’, ‘power’, and so on. In other words, Luther is assuming that the various inferences of his argument depend on the meanings of the categorematic terms used in the argument.

In what follows I shall offer an assessment of the validity of the various inferences occurring in Luther's necessitarian argument, asking, in each case, whether the proposition inferred is a logical consequence of the proposition from which it is inferred. Since, as was noted above, Luther's argument depends on relations of material consequence, it will be crucial, in assessing the validity of the argument, to understand how Luther understood the key terms of the argument.

But how shall we attain such an understanding? In what follows I shall assume that we can arrive at a fairly adequate view of Luther's understanding of the relevant terms by looking at the way these terms are understood in the theological tradition in which Luther was firmly rooted: the Augustinian tradition.Footnote 31 In line with this I will be making frequent reference to the way the relevant terms are used in Augustine's main works. Given that Luther was an Augustinian monk for over a decade prior to the outbreak of the Reformation, that he appealed to Augustine's theology in key Reformation statements like the Disputatio contra scholasticam theologiam (1517) and the Disputatio Heidelbergae habita (1518), and that he explicitly invokes Augustine as his theological ally in De servo arbitrio (as we shall see further on), the present methodology would seem to make good sense. I will accordingly assume that Luther's understanding of key theological terms of his argument is in substantial agreement with Augustinian theology, provided, of course, that we don't have good reason to think that Luther was departing from Augustine.

I proceed now to my assessment. The first inference of the argument, made in C1, is the inference of ‘God foreknows all things’ from ‘God has foreknowledge essentially’. The validity of this inference turns on what it means for God to have ‘foreknowledge’; more precisely, on whether the meaning of the term ‘foreknowledge’ is such that if God has foreknowledge he foreknows all things. In Augustinian theology this is clearly part of what the term ‘foreknowledge’ is taken to mean: we read in De civitate dei, for example, that ‘one who does not know all future things is surely not God’ (enim non est praescius omnium futurorum, non est utique Deus).Footnote 32

The second inference, in C2, is the inference of ‘God brings about whatever he wills’ from ‘God has power essentially’. The validity of this inference turns on whether it is part of the meaning of the term ‘powerful’, when applied to God, that God brings about whatever he wills. This appears to be what the term is taken to imply in Augustinian theology: in De civitate dei, once again, the ‘omnipotence’ (omnipotens) of God is said to mean that God ‘does what he wills, and does not suffer to be done what he does not will’ (faciendo quod vult, non patiendo quod non vult).Footnote 33

The third inference, in C3, is the inference of ‘God's will is unchangeable’ from ‘God's essence is unchangeable’ and ‘God has will essentially’. The validity of this inference depends mainly on whether it is part of the meaning of the term ‘unchangeable’, when applied to God's essence, that all attributes comprised in God's essence are unchangeable. That this is the understanding of the term ‘unchangeable’ in Augustinian theology is evident. For it is fundamental to Augustinian theology not only that God's ‘essence’ (essentia) is ‘unchangeable’ (immutabilem), as is affirmed, for example, in De trinitate,Footnote 34 but also, and as is affirmed in De civitate dei, that God's essence is ‘simple’ (simplicia), meaning that God's essence is ‘identical’ (non aliud) to God's ‘attributes’ (qualitas).Footnote 35 Now if God's essence is unchangeable, and if it pertains to God's essence to have will, then God's will must also be unchangeable. But even apart from this simple deduction it is easy to verify that the belief in the unchangeableness of God's will is part of Augustinian theology. It is asserted in various passages in De trinitate, for example, that ‘the will of God’ (uoluntas dei) is ‘unchangeable’ (incommutabilis).Footnote 36

Similar remarks apply to the fourth inference, in C4, the inference of ‘God's foreknowledge is unchangeable’ from ‘God's essence is unchangeable’ and ‘God has foreknowledge essentially’. Since God's essence and attributes are held to be identical, it follows that if it pertains to God's essence to be unchangeable and to have foreknowledge, God's foreknowledge must be unchangeable. But even apart from this it is clearly part of Augustinian theology that God's foreknowledge is unchangeable, as is clear, for example, from De civitate dei 2.11.21, a section devoted to the topic of ‘the eternal and unchangeable knowledge of God’ (aeterna et incommutabili scientia Dei).Footnote 37

The fifth inference, in C5, is the inference of ‘God foreknows all things unchangeably’ from ‘God foreknows all things’ and ‘God's foreknowledge is unchangeable’. That this inference is valid can be seen as follows: if God's foreknowledge is unchangeable, but God didn't foreknow all things unchangeably, it wouldn't be part of the meaning of ‘foreknowledge’ that God foreknows all things. But Augustinian theology takes it to be part of the meaning of the term ‘foreknowledge’, when applied to God, that God foreknows all things; hence, if God's foreknowledge is unchangeable, God foreknows all things unchangeably. But even apart from this deduction it is plainly part of Augustinian theology that God foreknows all things unchangeably: in De civitate dei 1.5.9, for example, we are told that ‘he knows unchangeably all things that will come to pass’ (novit incommutabiliter omnia quae futura).Footnote 38

The sixth inference, in C6, is the inference of ‘God wills what he foreknows unchangeably’ from ‘God wills what he foreknows’ and ‘God's foreknowledge is unchangeable’. The validity of this argument can be seen from the fact that if God wills what he foreknows, but changeably, then God's foreknowledge, the object of God's will, is changeable. Hence, if God's foreknowledge is unchangeable, God wills what he foreknows unchangeably.

The seventh inference, in C7, is the inference of ‘God wills all things’ from ‘God foreknows all things unchangeably’ and ‘God wills what he foreknows unchangeably’. The validity of this inference is rather straightforward: if God wills what he foreknows, and foreknows all things, both of which are held in Augustinian theology, it follows that God wills all things.

The eighth inference, in C8, is the inference of ‘God wills all things unchangeably’ from ‘God's will is unchangeable’ and ‘God wills all things’. The validity of this inference can be seen from the fact that if God wills all things but only changeably, then God's will would be changeable. But Augustinian theology holds that God's will is unchangeable; so if God wills all things, it follows that God wills all things unchangeably.

The ninth inference, in C9, is the inference of ‘God brings about all things unchangeably’ from ‘God brings about whatever he wills’ and ‘God wills all things unchangeably’. The validity of this inference can be seen from the fact that if God wills all things unchangeably but there are things that God brings about changeably, it couldn't be that God brings about whatever he wills. But Augustinian theology holds that God brings about whatever he wills. So if God wills all things unchangeably, it follows that God brings about all things unchangeably.

The tenth inference, in C10, is the inference of ‘humans lack free will’ from ‘God brings about all things unchangeably’. The validity of this inference turns on an acceptance of Luther's earlier mentioned definition of (human) ‘free will’ as a power to bring about things independently of God and other external agents. Clearly, if humans bring about things independently of God, then God won't bring about all things unchangeably. So if Luther's definition is accepted, the tenth inference seems valid. Interestingly, the Augustinian tradition doesn't understand the term ‘free will’ in this way, as is clear, for example, from Augustine's claim in De civitate dei that ‘human free will’ (libera hominis voluntate) is consistent with ‘a definite pattern of causation’ (certus . . . omnium ordo causarum) foreordained by God and comprising all history.Footnote 39 Luther clearly thought that this way of understanding the term ‘free will’ departs from ‘ordinary usage’ (natura vocabuli).Footnote 40 What people ordinarily mean by ‘free will’, Luther holds, is a ‘power of freely turning in any direction, yielding to none and subject to none’ (libere possit in utrunque se vertere, neque ea vis ulli caedat vel subiecta sit).Footnote 41 Given this understanding of free will, Augustine's view that God has foreordained a definite pattern of causation comprising all history would clearly undermine the claim that humans have free will.

If the above assessment of Luther's necessitarian argument is sound, it would seem that the various inferences on which the argument depends are valid, given that we understand the key terms of the argument in line with Augustinian theology, and provided that we understand the term ‘free will’ in what Luther takes to be its ordinary sense. It would seem, then, that Luther's necessitarian argument is logically valid.

But the logical validity of an argument doesn't entail that the argument terminates in a true conclusion; this requires that the premises of the argument are true. To determine whether Luther's necessitarian argument is not only logically valid but also terminates in a true conclusion is not my task in this article. I shall conclude by raising a related question, though, namely: is Luther's necessitarian argument theologically cogent, in the sense that adherents of Augustinian theology (or, more precisely, of those aspects of Augustinian theology delineated above) are committed to its conclusion?

That Augustinian theology accepts the first four premises of the necessitarian argument – i.e. the claims that God's essence is unchangeable, and that it pertains to God's essence to have will, foreknowledge, and power – has already been made evident by quotations from De civitate dei and De trinitate. What hasn't been made evident is whether Augustinian theology also accepts P5, i.e. the claim that ‘God wills what he foreknows’. If so, the argument would appear to be not only logically valid but also theologically cogent in the relevant sense.

I believe a strong case can be made for taking Augustinian theology to be committed to P5. For suppose that the Augustinian takes P5 to be false. If so, the Augustinian will believe that there are things that God foreknows but nevertheless doesn't will. If so, then since the Augustinian holds that God brings about everything he wills (see C2), the Augustinian will be committed to holding that some things are not brought about by God. If so, then since the Augustinian holds that God foreknows all things unchangeably (see C5), the Augustinian will be committed to holding that there are things that God foreknows unchangeably but which nevertheless are brought about by factors distinct from God. But it is part of the core claims of Augustinian theology that God alone is unchangeable, and that factors distinct from God are contingent. So on the present supposition the Augustinian will be committed to holding that there are things that God foreknows unchangeably but which nevertheless have contingent causes. But this is contradictory, for whatever is unchangeably foreknown to take place cannot but take place, and so must take place of necessity, and nothing that takes place of necessity can have a contingent cause (for in that case it wouldn't take place ‘of necessity’). Accordingly, the Augustinian is committed to rejecting the supposition that P5 is false. That is to say, the Augustinian is committed to the claim that God wills what he foreknows.

It would seem, then, that Luther's necessitarian argument is theologically cogent in the sense that adherents of Augustinian theology are committed to the conclusions of the argument. The only point where there might seem to be a substantive difference between Luther and the Augustinian is that whereas Augustine allows that people have ‘free will’, Luther denies this. However, once it is realized that Luther's denial of ‘free will’ is mere shorthand for the affirmation, made also by the Augustinian, that ‘all we do . . . is in reality done necessarily and immutably in respect to God's will’, the positions of Luther and the Augustinian can be seen to amount to the same thing. It would seem, then, that Luther was right in claiming, against Erasmus, that ‘Augustine . . . is entirely with me’ (Augustinus . . . meus totus est).Footnote 42 What Luther's necessitarian argument in effect does is push Augustinian theism to its logical conclusion with respect to the idea of human free will.Footnote 43

Footnotes

1. WA, Abt. 1 Schriften, vol. 1, 354. (‘WA’ abbreviates D. Martin Luthers Werke. Kritische Gesammtausgabe, the standard scholarly edition of Luther's works.)

2. For Leo X's bull, see Denzinger (Reference Denzinger1965), 1486.

3. WA, VII, 94–151. For the English translation, see Trinkaus (Reference Trinkaus1999), 301–310.

5. Footnote Ibid., 12.

6. I will not at present be concerned with the relation between Luther's argument and discussions of similar arguments found in such thinkers as Augustine, Boethius, Lombard, and Aquinas, but will instead be focusing more narrowly on Luther's argument as such. For my take on the relation between Luther's argument and earlier discussions in the history of philosophy and theology, see Kraal (forthcoming).

7. WA, XVIII, 615. The English translation is from Luther (Reference Luther, Packer and Johnston2006), 80.

8. WA, XVIII, 615; Luther (Reference Luther, Packer and Johnston2006), 80.

9. WA, Abt. 4 Briefe, VIII, 99.

10. This is true even of many conservative Lutheran theologians; see, for example, Bernhard Lohse's comments on the ‘exaggerated remarks’ of De servo arbitrio in Lohse (Reference Lohse and Harrisville1999), 168.

11. For an overview of some main German scholarship on De servo arbitrio, see Stayer (Reference Stayer2000). For an overview of some main Scandinavian scholarship, see Kraal (Reference Kraal2013).

12. Kolb (Reference Kolb2005), 9.

13. WA, XVIII, 615; Luther (Reference Luther, Packer and Johnston2006), 80.

14. WA, VIII, 637; Luther (Reference Luther, Packer and Johnston2006), 105. Note that Luther doesn't take free will to involve simply the absence of subjection to God, but also a ‘power’ of ‘turning in any direction’. If we allow that the ‘power’ of ‘turning in any direction’ entails an ‘ability’ to ‘do otherwise’, then Luther is here implying that ‘free will’ (as ordinarily understood) involves what is today known as ‘the principle of alternate possibilities’: humans have free will only if they could have done otherwise (see Frankfurt (Reference Frankfurt1969), 829). It would thus seem that Luther's denial of (what he takes to be) free will entails a denial also of free will as modern libertarians or incompatibilists typically understand it.

15. WA, XVIII, 615.

17. In making this claim Luther might have been presupposing the Augustinian doctrine of divine simplicity, according to which God's intrinsic attributes and being are absolutely identical (which entails that if one of God's intrinsic attributes is unchangeable, they all are). For more on the Augustinian understanding of divine simplicity, see Kraal (Reference Kraal2011).

18. WA, XVIII, 615.

20. Luther also says in the relevant passages that God's nature, will and knowledge are ‘eternal’ (aeterna), but this claim would seem redundant inasmuch as anything that is unchangeable is also eternal.

21. McSorley (Reference McSorley1967), 286.

22. Footnote Ibid., 286–287.

23. Urban (Reference Urban1971), 115–116.

24. Footnote Ibid., 116.

25. Kolb (Reference Kolb2005), 26.

26. Footnote Ibid., 52.

27. Aristotle (Reference Aristotle and Tredennick1962), 201 (An. Pr. 1.1, 24a 20).

28. See e.g. Peter of Spain's account of logical validity in his influential Summulae logicales, in Dinneen (Reference Dinneen and de Rijk1990), 38.

29. Luther studied logic under the famous German logician Jodocus Trutfetter while an undergraduate at the University of Erfurt in 1501–1505, and some of Luther's earliest professorial duties at Wittenberg University included teaching logic classes. For an in-depth study of Luther's relation to the medieval logic tradition, see White (Reference White1994).

30. For an overview of this aspect of medieval logical theory, see Boehner (Reference Boehner1950), 57–58. Contemporary logicians typically reserve the term ‘logical consequence’ for what medieval logicians called ‘formal consequence’; see e.g. the definition of logical validity in Bergmann et al. (Reference Bergmann, Moor and Nelson2004), 447. I know of no theoretically compelling reasons for restricting logicality to formality, however. Indeed, Alfred Tarski, the main pioneer of this approach, took the restriction to rest in part on an arbitrary distinction between logical and extra-logical terms; see Tarski (Reference Tarski2002), 188. In what follows I shall go along with the medieval logical tradition in regarding both formal and material consequences as cases of ‘logical consequence’.

31. For more on Luther's Augustinian background, see Oberman (Reference Oberman and Martin1981), 64–110; Idem (Reference Oberman1986), 66–80.

32. Augustine (Reference Augustine and Green1963), 180–181 (De civ. dei 1.5.9). An alternative reading would be to take this passage as claiming (or assuming) that it is implicit in the concept of ‘God’ (rather than in the concept of ‘foreknowledge’) that whatever is God foreknows all things. (Similar remarks apply mutatis mutandis to the reasoning leading to C2 and C5.) In view of the Augustinian doctrine of divine simplicity, either reading would appear to validate the relevant inference (since on this doctrine God and God's foreknowledge are identical). Thanks to Kaj B. Hansen for bringing this alternative reading to my attention.

33. Augustine (Reference Augustine and Green1963), 184–185 (De civ. dei 1.5.10).

34. Augustine (Reference Augustine and Mountain1968b), 464 (De trin. 15.5.34).

35. Augustine (Reference Augustine and Wiesen1968a), 465, 469 (De civ. dei 2.11.10). I am here assuming that non aliud amounts to identity. This is a common assumption in Augustine scholarship; see e.g. Portalié (Reference Portalié1960), 128; O'Daly (Reference O'Daly1999), 140; Ayres (Reference Ayres, Dodaro and Lawless2000), 67.

36. Augustine (Reference Augustine and Mountain1968b), 134 (De trin. 3.3.38); cf. also De civ. dei 2.22.2.

37. Augustine (Reference Augustine and Wiesen1968a), 503 (De civ. dei 2.11.21).

38. Augustine (Reference Augustine and Green1963), 174–175 (De civ. dei 1.5.9).

39. Augustine (Reference Augustine and Green1963), 167–168, 174–175 (De civ. dei 1.5.11).

40. WA, XVIII, 637; Luther (Reference Luther, Packer and Johnston2006), 105.

41. WA, XVIII, 637; Luther (Reference Luther, Packer and Johnston2006), 105.

42. WA, XVIII, 640; Luther (Reference Luther, Packer and Johnston2006), 109.

43. Many thanks to Professors J. J. MacIntosh (University of Calgary) and Kaj B. Hansen (Uppsala University) for valuable commentary on an earlier version of this article. Thanks also to two anonymous reviewers for Religious Studies for feedback.

References

Aristotle, (1962) Prior Analytics, Tredennick, H. (tr. & ed.) (Cambridge MA: Harvard University Press).Google Scholar
Augustine, (1963) The City of God against the Pagans, II, Green, W. (tr. & ed.) (Cambridge MA: Harvard University Press).Google Scholar
Augustine, (1968a) The City of God against the Pagans, III, Wiesen, D. (tr. & ed.) (Cambridge MA: Harvard University Press).Google Scholar
Augustine, (1968b) De trinitate, Mountain, W. J. (ed.) (Turnhout: Brepols).Google Scholar
Ayres, L. (2000) ‘The fundamental grammar of Augustine's trinitarian theology’, in Dodaro, R. & Lawless, G. (eds) Augustine and his Critics (London and New York: Routledge), 5176.Google Scholar
Bergmann, M., Moor, J., & Nelson, J. (2004) The Logic Book, 4th edn (New York: McGraw-Hill).Google Scholar
Boehner, P. (1950) Medieval Logic (Manchester: Manchester University Press).Google Scholar
Denzinger, H. (ed.) (1965) Enchiridion symbolorum definitionum et declarationum de rebus fidei et morum, 33rd edn (Freiburg im Breisgau: Herder).Google Scholar
Dinneen, F. (ed.) (1990) Peter of Spain: Language in Dispute: An English Translation of Peter of Spain's ‘Tractatus’ called afterwards SUMMULAE LOGICALES, based on the Critical Edition by de Rijk, L. M. (Philadelphia: John Benjamins Publishing Company).Google Scholar
Erasmus, (1969) De libero arbitrio diatribe sive collatio, Lesowsky, W. (ed.) (Darmstadt: Wissenschaftliche Buchgesellschaft).Google Scholar
Frankfurt, H. (1969) ‘Alternate possibilities and moral responsibility’, The Journal of Philosophy, 66, 829839.CrossRefGoogle Scholar
Kolb, R. (2005) Bound Choice, Election, and the Wittenberg Theological Method (Grand Rapids MI: Eerdmans).Google Scholar
Kraal, A. (2011) ‘Logic and divine simplicity’, Philosophy Compass, 6, 282294.CrossRefGoogle Scholar
Kraal, A. (2013) ‘Free choice, determinism, and the re-evaluation of Luther in twentieth-century Swedish theology’, Studia Theologica, 67, 2842.CrossRefGoogle Scholar
Kraal, A. (forthcoming) ‘Valla-style determinism and the intellectual background of Luther's De servo arbitrio’, Harvard Theological Review.Google Scholar
Lohse, B. (1999) Martin Luther's Theology: Its Historical and Systematic Development, Harrisville, R. A. (ed. & tr.) (Edinburgh: T&T Clark).Google Scholar
Luther, M. (1883–2009) D. Martin Luthers Werke: Kritische Gesammtausgabe, 121 vols (Weimar: Böhlau).Google Scholar
Luther, M. (2006) The Bondage of the Will, Packer, J. I. & Johnston, O. R. (trs & eds) (Grand Rapids MI: Revell).Google Scholar
McSorley, H. J. (1967) Luthers Lehre vom unfreien Willen nach seiner Hauptschrift ‘De servo arbitrio’ im Lichte der biblischen und kirchlichen Tradition (Munich: Hueber).Google Scholar
Oberman, H. (1981) Masters of the Reformation: The Emergence of a New Intellectual Climate in Europe, Martin, D. (tr.) (Cambridge: Cambridge University Press).CrossRefGoogle Scholar
Oberman, H. (1986) The Dawn of the Reformation: Essays in Late Medieval and Early Reformation Thought (Edinburgh: T&T Clark).Google Scholar
O'Daly, G. (1999) Augustine's City of God: A Reader's Guide (Oxford: Clarendon Press).CrossRefGoogle Scholar
Portalié, E. (1960) A Guide to the Thought of Saint Augustine (London: Barnes & Oats).Google Scholar
Stayer, J. M. (2000) Martin Luther, German Saviour (Montreal and Kingston: McGill-Queen's University Press).CrossRefGoogle Scholar
Tarski, A. (2002) ‘On the concept of following logically’, History and Philosophy of Logic, 23, 155196.CrossRefGoogle Scholar
Trinkaus, C. (ed.) (1999) Collected Works of Erasmus, LXXVI (Toronto: University of Toronto Press).Google Scholar
Urban, L. (1971) ‘Was Luther a thoroughgoing determinist?’, Journal of Theological Studies, 22, 113139.CrossRefGoogle Scholar
White, G. (1994) Luther as Nominalist: A Study of the Logical Methods used in Martin Luther's Disputations in the Light of their Medieval Background (Helsinki: Luther-Agricola-Society).Google Scholar