1 A preliminary version of this essay was presented at a conference on “Saint Anselm of Canterbury and His Legacy” (Canterbury, 2009), organized under the aegis of the Centre for Medieval and Renaissance Studies, Durham University; the Canterbury Centre for Medieval and Early Modern Studies; and the Internationale Gesellschaft für Theologische Mediävistik. I am grateful to Giles E. M. Gasper and Ian Logan for their personal thoughtfulness in the coordination of the conference. An incipient form of the argument of this study appears in Jon Whitman, “The Transformation of Allegory in the Middle Ages” (Ph.D. diss., Harvard University, 1978) 72–104 and 535–54, although aspects of that early discussion are significantly revised in the analysis below.

2 See Ambrose, , Epistola 50, in PL 16Google Scholar, 1155B–1156A: “Quid ergo ei impossibile? Non quod virtuti arduum, sed quod naturœ ejus contrarium. Impossibile, inquit, est ei mentiri. Impossibile, istud non infirmitatis est, sed virtutis et majestatis ….”

3 See Augustine, , Serm. 214, 4Google Scholar, in PL 38, 1067–68 [emphasis in edition]: “Omnipotens id solum non potest, quod non vult. … Non enim potest justitia velle facere quod injustum est, aut sapientia velle quod stultum est, aut veritas velle quod falsum est…. Deus omnipotens non potest mori, non potest mutari, non potest falli, non potest miser fieri, non potest vinci. Hœc atque hujusmodi absit ut possit omnipotens…. Si ergo potest esse quod non vult, omnipotens non est ….” In a broad sense, of course, the view that divine ability is conditioned by divine will appears long before Augustine; see, e.g., the brief statement of Tertullian, , Carn. Chr. 3Google Scholar, in PL 2, 756A–B. But Augustine's repeated tendency to link the authority and the irresistibility of the divine will has a formative role in approaches to divine power in the early Middle Ages. See, e.g., Enchir. 95–96, in PL 40, 276; Symb. I, 2, and II, 3, in PL 40, 627 and 629; and Civ. V, 10, and XXI, 7, in PL 41, 152 and 719; compare, in diverse contexts in the early Middle Ages, Fulgentius of Ruspe, Ad Monimum I, 12, in PL 65, 161D–162A, and Epistola 15, 15, in PL 65, 441B; Paschasius Radbertus, De corpore et sanguine Domini I, 2, in PL 120, 1269C; Ratherius of Verona, De translatione sancti Metronis, in PL 136, 466B; and Peter Damian, De divina omnipotentia, discussed below.

4 See Damian, Peter, De divina omnipotentia, in Lettre sur la toute-puissance divine (ed. and trans. André Cantin; SC 191; Série des textes monastiques d'Occident 11; Paris: Cerf, 1972) 390–91Google Scholar, 396–97 (PL 145, 597C and 598D–599A): “Illo plane modo dicitur Deus non posse aliquid quo et nescire; uidelicet quicquid malum est, sicut non potest agere, ita nescit agere…. Hoc ergo quod dicitur Deus non posse malum aliquod uel nescire, non referendum est ad ignorantiam uel inpossibilitatem, sed ad uoluntatis perpetuae rectitudinem.” My translation draws partly on the work of Cantin. For the date of De divina omnipotentia (perhaps 1065 rather than 1067, the date sometimes assigned to it), see Holopainen, Toivo J., Dialectic and Theology in the Eleventh Century (Studien und Texte zur Geistesgeschichte des Mittelalters 54; Leiden: Brill, 1996) 6Google Scholar n. 1. On the emphasis on the will and nature of God in Christian treatments of omnipotence and impossibility before Damian, see Resnick, Irven Michael, Divine Power and Possibility in Saint Peter Damian's De Divina Omnipotentia (Studien und Texte zur Geistesgeschichte des Mittelalters 31; Leiden: Brill, 1992) 30–36Google Scholar, who also notes (77) the importance assigned to divine will in the Berengarian controversy of Damian's own period.

5 See Proslogion, ch. 7, in S. Anselmi Cantuariensis Archiepiscopi opera omnia (ed. Franciscus Salesius Schmitt; 6 vols.; 1938–1961; repr. in 2 pts. with introd. by Schmitt; Stuttgart-Bad Cannstatt: Friedrich Frommann, 1968) 1.105–6 (hereafter Opera). For a translation, see Complete Philosophical and Theological Treatises of Anselm of Canterbury (trans. Jasper Hopkins and Herbert Richardson; Minneapolis, Minn.: Arthur J. Banning, 2000) 96–97 (hereafter Complete Treatises), available online via the website of Jasper Hopkins, http://cla.umn.edu/sites/jhopkins/ (from which citations are taken). In my quotation (below) of Proslogion, ch. 7, I have inserted in both the original text and the translation bracketed numbers according to the sequence of Latin sentences in Opera. In both the original text and the translation I have also presented certain expressions in italics. In the translation the brackets that enclose English words appear in Complete Treatises itself. For an estimate about the date of the Proslogion (late 1070s), see Jasper Hopkins, A Companion to the Study of St. Anselm (Minneapolis, Minn.: University of Minnesota Press, 1972) 5 n. 5.

6 See Hopkins, and Richardson, , Complete Treatises, 96Google Scholar. For the original text, see Schmitt, , Opera 1.105Google Scholar (paragraph division between sentences 2 and 3 omitted): “[1] Sed et omnipotens quomodo es, si omnia non potes? [2] Aut si non potes corrumpi nec mentiri nec facere verum esse falsum, ut quod factum est non esse factum, et plura similiter: quomodo potes omnia? [3] An hœc posse non est potentia, sed impotentia? [4] Nam qui hœc potest, quod sibi non expedit et quod non debet potest.”

7 See Hopkins, and Richardson, , Complete Treatises, 96–97Google Scholar. For the original text, see Schmitt, , Opera 1.105–6Google Scholar: “[5] Quœ quanto magis potest, tanto magis adversitas et perversitas possunt in illum, et ipse minus contra illas. [6] Qui ergo sic potest, non potentia potest, sed impotentia. [7] Non enim ideo dicitur posse, quia ipse possit, sed quia sua impotentia facit aliud in se posse; sive aliquo alio genere loquendi, sicut multa improprie dicuntur. [8] Ut cum ponimus ‘esse’ pro ‘non-esse,’ et ‘facere’ pro eo quod est ‘non facere,’ aut pro ‘nihil facere.’ [9] Nam sœpe dicimus ei qui rem aliquam esse negat: sic est quemadmodum dicis esse, cum magis proprie videatur dici: sic non est quemadmodum dicis non esse. [10] Item dicimus: iste sedet sicut ille facit, aut: iste quiescit sicut ille facit, cum ‘sedere’ sit quiddam non facere et ‘quiescere’ sit nihil facere. [11] Sic itaque cum quis dicitur habere potentiam faciendi aut patiendi quod sibi non expedit aut quod non debet, impotentia intelligitur per potentiam; quia quo plus habet hanc potentiam, eo adversitas et perversitas in illum sunt potentiores, et ille contra eas impotentior. [12] Ergo domine deus, inde verius es omnipotens, quia nihil potes per impotentiam, et nihil potest contra te.”

8 On the acute sense of “power possessed by another agency” in this passage, see William J. Courtenay, “Necessity and Freedom in Anselm's Conception of God,” in Anselms Bedeutung für die Geschichte von Philosophie und Theologie (ed. Helmut Kohlenberger; pt. 2 of Die Wirkungsgeschichte Anselms von Canterbury. Akten der ersten Internationalen Anselm-Tagung, Bad Wimpfen, 13. September bis 16. September 1970; Analecta Anselmiana. Untersuchungen über Person und Werk Anselms von Canterbury 4.2; Frankfurt am Main: Minerva, 1975) 39–64, at 46–48, who comments that adversitas seems to have a “stronger” force than the Neoplatonic notion of a “turning away from God” and that in Anselm's period “the term ‘adversitas’ implied hostile opposition of a most positive kind” (47 n. 18).

9 Perhaps nothing shows more suggestively how Anselm's argument both draws upon and differs from earlier strategies than the orientation of this sentence (sentence six), which responds to the question about potentia and impotentia in sentence three. Sentence six itself is reminiscent of a remark by Philosophy in Boethius's Consolation of Philosophy IV, pr. iv: “But evil men, you will say, are able to do things. Not even I would deny that, but this ability of theirs is derived not from their strength but from their weakness” (“ ‘Sed possunt,’ inquies, ‘mali.’ Ne ego quidem negaverim, sed haec eorum potentia non a viribus sed ab imbecillitate descendit.”). For the text and translation, see Boethius, , The Consolation of Philosophy (trans. S. J. Tester; in The Theological Tractates; The Consolation of Philosophy; rev. ed.; LCL; Cambridge, Mass: Harvard University Press, 1973) 326–27Google Scholar. General relations between Proslogion, ch. 7, and Consolation of Philosophy IV, pr. iv, are noted by Paul Henry, Desmond in The Logic of Saint Anselm (Oxford: Clarendon, 1967) 150–54Google Scholar. But while both texts indicate that the power to act defectively is powerlessness, Anselm applies the argument not just to an account of human activity, but to the defense of divine power itself; he stresses not just the slipping of a defective agent, but the strengthening of a counteragent; he dramatizes the dialectic between these converse movements; and he elaborates a concept of divided reference in language in an effort to clarify distinctions regarding power; see below. In this regard, I think the view that in Proslogion, ch. 7, Anselm states “in a traditional manner” that “many of the alleged counter-examples to omnipotence are signs of impotence” needs considerable adjustment; for that view, see Knuuttila, Simo, “Anselm on Modality,” in The Cambridge Companion to Anselm (ed. Davies, Brian and Leftow, Brian; Cambridge: Cambridge University Press, 2004) 130Google Scholar n. 33. On relations between Anselm's treatment of power and distinctions of modal logic, see Knuuttila's discussion at large (111–31) and the references in n. 25 below.

10 See De casu diaboli, ch. 26, in Schmitt, , Opera 1.274Google Scholar: “Cum autem dicimus quia iniustitia facit rapinam, aut cœcitas facit hominem cadere in foveam, nequaquam intelligendum est quod iniustitia vel cœcitas aliquid faciant; sed quod si iustitia esset in voluntate et visus in oculo, nec rapina fieret nec casus in foveam.” For the translation, see Hopkins, and Richardson, , Complete Treatises, 259Google Scholar.

11 See De casu diaboli, ch. 12, in Schmitt, , Opera 1.253Google Scholar, with Hopkins, and Richardson, , Complete Treatises, 237Google Scholar. Compare De casu diaboli, ch. 1, in which Anselm explicitly raises a question of scriptural exegesis and uses a formulation suggestive of the terminology of integumenta (verbal coverings) that develops in different interpretive contexts in the early twelfth century: “But we ought not to cling to the verbal impropriety concealing the truth (veritatem tegenti) as much as we ought to attend to the true propriety hidden beneath the many types of expression” (Hopkins, and Richardson, , Complete Treatises, 217Google Scholar). For the original text, see Schmitt, , Opera 1.235Google Scholar: “Sed non tantum debemus inhœrere improprietati verborum veritatem tegenti, quantum inhiare proprietati veritatis sub multimodo genere locutionum latenti.”

12 See De veritate, ch. 8, in Schmitt, , Opera 1.188Google Scholar, with Hopkins, and Richardson, , Complete Treatises, 177Google Scholar. For related expressions of the two sides of power, see, e.g., De casu diaboli, ch. 12 (Schmitt, , Opera, 1.253Google Scholar), and Cur Deus homo II, chs. 10 (Schmitt, , Opera 2.107Google Scholar) and 17 (Schmitt, , Opera 2.123Google Scholar).

13 A brief, early sketch of Henry's general argument appears in his “Remarks on Saint Anselm's Treatment of Possibility,” in Congrès international du IXe centenaire de l'arrivée d'Anselme au Bec (Spicilegium Beccense 1; Le Bec-Hellouin, Eure: Abbaye Notre-Dame du Bec, 1959) 19–22. The argument is developed in detail in his Logic of Saint Anselm (cited above in n. 9), a study that includes a range of important points but that is problematic in its overall design and orientation. On the Hector/Achilles example, see Logic of Saint Anselm, 155.

14 The question about restoring virginity appears first in Jerome, who argues that “although God is able to do everything, he is not able to restore a virgin after her fall” (cum omnia Deus possit, suscitare uirginem non potest post ruinam); see Epist. 22, Ad Eustochium 5, in Lettres (ed. and trans. Jérôme Labourt; 8 vols.; Paris: Les Belles Lettres, 1949–1963) 1.115. Much later, Aquinas, for example, agrees; see Quodlibetum 5, q. 2, a. 3. The broad question of whether God can make the past not to have been has a long history before and after Damian. It has normally been argued that God cannot annul the past; see, e.g., Aristotle, Eth. nic. VI, 2, 1139b; Augustine, Faust. XXVI, 5 (PL 42, 481); Anselm (cited above and below in this essay); Honorius Augustodunensis, Elucidarium II, 8 (PL 172, 1139C); and John of Salisbury, Policraticus II, 22 (PL 199, 453A, 454D). For expressions of the argument in the thirteenth century, see Bonaventure, Commentarius in quatuor libros Sententiarum Petri Lombardi I, 42, un., 3, Concl., discussed by Gilson, Etienne, The Philosophy of St. Bonaventure (trans. Illtyd, TrethowanFrank J., Sheed; London: Sheed and Ward, 1938Google Scholar; repr., Paterson, N.J.: St. Anthony Guild Press, 1965) 152–53, and Aquinas, Summa theologiae I, q. 25, a. 4. Those after Damian with whom the opposing view has been associated include Gilbert of Poitiers, William of Auxerre, and (in the fourteenth century) Thomas Bradwardine and Gregory of Rimini; see Courtenay, William J., “John of Mirecourt and Gregory of Rimini on Whether God Can Undo the Past,” Recherches de théologie ancienne et médiévale 39 (1972) 224–56Google Scholar, and 40 (1973) 147–74, repr. in Courtenay, William J., Covenant and Causality in Medieval Thought: Studies in Philosophy, Theology and Economic Practice (London: Variorum Reprints, 1984Google Scholar) sections VIIIa and VIIIb.

15 On the tension in De divina omnipotentia between the views that God can and cannot make the past not to have been, see Cantin's detailed analysis in his 1972 edition, cited above in n. 4, esp. 129–31 and 178–80, with his remark (140) about a “jeu de ruptures.” For diverse later assessments of this tension, see, e.g., Resnick, , Divine Power, 77–111Google Scholar, and Holopainen, , Dialectic and Theology, 6–43Google Scholar.

16 See Cantin, , Lettre, 434–40Google Scholar (PL 145, 608C–610B). For detailed analyses of varying forms of Damian's general argument (beyond this specific variation), see the studies cited in the previous note.

17 For the citation of the original passage (Cantin, , Lettre, 390–91Google Scholar, 396–97; PL 145, 597C and 598D–599A), see n. 4.

18 Compare the similar formulation slightly later (Cantin, , Lettre, 400Google Scholar; PL 145, 600A), where God's inability to do or will evil is to be referred (referendum) not to impossibility (inpossibilitatem) but to his goodness (bonitatem).

19 See Cantin, , Lettre, 416Google Scholar (PL 145, 604B).

20 It should be stressed that Damian himself works with dialectical techniques in De divina omnipotentia, but his overall approach to the role and practice of dialectic is marked by deep reservations. On this question see Cantin, , Lettre, 35–37Google Scholar, 40, 184–264, 270–75, 301–11; Cantin, André, Les sciences séculières et la foi. Les deux voies de la science au jugement de S. Pierre Damien (1007–1072) (Spoleto: Centro italiano di studi sull'alto Medioevo, 1975Google Scholar) 417 n. 123, 433, 437–50; and Holopainen, , Dialectic and Theology, 1–43Google Scholar, 156–57.

21 See Hopkins, and Richardson, , Complete Treatises, 377Google Scholar. For the original text, see Cur Deus homo II, ch. 17, in Schmitt, , Opera 2.123Google Scholar: “Et sicut cum deus facit aliquid, postquam factum est, iam non potest non esse factum, sed semper verum est factum esse; nec tamen recte dicitur impossibile deo esse, ut faciat quod praeteritum est non esse praeteritum …. Quotiens namque dicitur deus non posse, nulla negatur in illo potestas, sed insuperabilis significatur potentia et fortitudo. Non enim aliud intelligitur, nisi quia nulla res potest efficere, ut ille agat quod negatur posse…. Nec dicimus deum necessitate facere aliquid, eo quod in illo sit ulla necessitas, sed quoniam est in alio ….” In both the original text and the translation I have presented certain expressions in italics; “(facit)” in the first line of the translation appears in Complete Treatises itself. On the dating of Cur Deus homo (completed in the late 1090s), see Hopkins, , Companion, 5Google Scholar n. 5.

22 In concentrating on the role of divine will in Cur Deus homo II, ch. 17, both Courtenay, “Necessity and Freedom,” 58–60, and Serene, Eileen, “Anselm's Modal Conceptions,” in Reforging the Great Chain of Being: Studies of the History of Modal Theories (ed. Knuuttila, Simo; Synthese Historical Library: Texts and Studies in the History of Logic and Philosophy 20; Dordrecht: D. Reidel, 1981) 117–62Google Scholar, at 138–39, seem to me to give inadequate attention to the intensive dialectic in this chapter between the capacity of God's own will and the constraint of the created world; see Schmitt, , Opera 2.122–26Google Scholar; Hopkins, and Richardson, , Complete Treatises, 376–81Google Scholar. Perhaps the closest approach to such a sense of interplay in De divina omnipotentia is a passage (Cantin, , Lettre, 450Google Scholar; PL 145, 612C–D) in which Damian indicates that the “necessity of nature” does not rebelliously resist God but, submissive to his laws, serves him as a handmaid (ei naturae necessitas non rebellis obsistat, sed eius substrata legibus uelut ancilla deseruiat). Here, however, necessitas is used not to explain (from the other side) what seems impossible for God, but to stress (given God's omnipotence) what is possible for him, and the counterpart to divine power is scarcely cited before it basically collapses: “Indeed,” Damian immediately continues, “nature itself has its own nature—that is, the will of God” (Ipsa quippe rerum natura habet naturam suam, Dei scilicet uoluntatem). For some remarks on approaches to necessity at large in Damian and Anselm, see Corvino, Francesco, “Necessità e libertà di Dio in Pier Damiani e in Anselmo d'Aosta,” in Analecta Anselmiana 5 (1976) 245–60Google Scholar, and Holopainen, Toivo J., “Necessity in Early Medieval Thought: Peter Damian and Anselm of Canterbury,” in Cur Deus Homo: Atti del congresso anselmiano internazionale, Roma, 21–23 maggio 1998 (ed. Gilbert, Paul, Kohlenberger, Helmut, and Salmann, Elmar; Studia Anselmiana 128; Rome: Centro Studi S. Anselmo, 1999) 221–34Google Scholar.

23 See De casu diaboli, ch. 27, in Schmitt, , Opera 1.275Google Scholar: “M. Cur ergo quœris unde venit iniustitia quœ nihil est? D. Quia quando iustitia inde recedit ubi erat, dicimus accedere iniustitiam. M. Dic ergo quod magis proprie et apertius dicitur, et quœre de abscessu iustitiœ.” For the translation, see Hopkins, and Richardson, , Complete Treatises, 260Google Scholar: “T. Then, why do you ask from where injustice, which is nothing, comes? S. Because when justice departs from where it was, we say that injustice approaches. T. Then, express yourself more properly and clearly, and ask about the departure of justice.”

24 For the text of Philosophica fragmenta, see Memorials of St. Anselm (ed. R. W. Southern and F. S. Schmitt; Auctores Britannici Medii Aevi 1; London: Oxford University Press, 1969; repr., Oxford: Oxford University Press, 1991) 334–51 (hereafter Philosophica fragmenta). For a translation, see Hopkins, and Richardson, , Complete Treatises, 390–414Google Scholar; the bracketed expressions in English appear in Complete Treatises itself. The translation in Complete Treatises is based on an earlier edition of the text by Schmitt (in which, as in the translation, the fragments are presented in an order different from the order in Memorials of St. Anselm, which follows the order of the manuscript). See Complete Treatises, 390 n. 1, which also indicates Schmitt's view that the “main sections” of the text postdate Cur Deus homo. The dating of the Philosophical Fragments remains a matter of debate.

25 On these tensions, including diverse approaches to the conditions of possibility, see, e.g., Serene, “Anselm's Modal Conceptions,” 117–62; Knuutilla, “Anselm on Modality,” 111–31; and the chapter on “Modality” in Visser, Sandra and Williams, Thomas, Anselm (Medieval Thinkers, Great; Oxford: Oxford University Press, 2009) 149–69Google Scholar.

26 For building the house, see Southern, and Schmitt, , Philosophica fragmenta, 347Google Scholar: “… ut cum quis construere domum dicitur qui nihil operatur sed praecipit …,” with Hopkins, and Richardson, , Complete Treatises, 404Google Scholar: “… when someone who orders [a house to be built] but does not do the actual work is said to build a house …” For “common parlance,” see, e.g., Philosophica fragmenta, 337 (usu loquendi; Complete Treatises, 393), with the following note.

27 On the relation between the terminology of per se / per aliud and the differentiation between speaking “properly” (proprie) and speaking “improperly” (improprie), see the references to Anselm's work at large (including Proslogion, ch. 7, quoted near the beginning of this essay) and the discussion in Henry, , Logic of Saint Anselm, 12–24Google Scholar, 128, 202–3, although his stress on the linguistic distinction does not seem to me to give adequate attention to the causal distinctions presented in the Philosophical Fragments. In this regard, compare Serene, “Anselm's Modal Conceptions,” 118–30, on the Fragments as an effort to analyze “the bases for ‘proper’ and ‘improper’ ascriptions” (119) and the persisting problem of clarifying “truth-conditions for ‘improper’ ascriptions” (129–30).

28 For causing a house to be lighted by making a window, see Southern, and Schmitt, , Philosophica fragmenta, 350Google Scholar, with Hopkins, and Richardson, , Complete Treatises, 400Google Scholar. For the general classification of six “modes” of causation (elaborated elsewhere by detailed examples), see Philosophica fragmenta, 343–44, with Complete Treatises, 395–96.

29 See Southern, and Schmitt, , Philosophica fragmenta, 343Google Scholar: “Ferrum enim gladii proxima causa est et suo modo per se … facit illum; et terra, ex qua ferrum fit, est eius longinqua causa faciens eum per aliud …. Habet enim omnis causa causas usque ad supremam omnium causam deum….” For the translation, see Hopkins, and Richardson, , Complete Treatises, 410Google Scholar.

30 See Whitman, Jon, Allegory: The Dynamics of an Ancient and Medieval Technique (Cambridge, Mass.: Harvard University Press, 1987) 161–262CrossRefGoogle Scholar, and my recent essay, “Twelfth-Century Allegory: Philosophy and Imagination,” in The Cambridge Companion to Allegory (ed. Rita Copeland and Peter T. Struck; Cambridge: Cambridge University Press, 2010) 101–15.

31 See the discussion and bibliography in Whitman, , Allegory, 58–217Google Scholar.

32 Not everyone, of course, endorses such dialectical designs. See, for example, Rupert of Deutz's De voluntate Dei (PL 170, 437–54) and De omnipotentia Dei (PL 170, 454–78), written less than a decade after Anselm's death. The De omnipotentia Dei nearly ends by quoting Jerome on God's inability to restore lost virginity, only to dismiss the dilemma as lacking scriptural authority, while finally referring this inability, in Damian's way, to God's unalterable justice (inflexibilem justitiam). With his wariness of dialectica (PL 170, 474A, with 469A–C and 473C), Rupert argues in these works against William of Champeaux and Anselm of Laon, whose attempt to explain evil by a twofold divine will—such an empty distinction (tam inertem divisionem; 473C), Rupert calls it—is suggestive of Anselm's technique of divided reference. On the “distinction” and the controversy at large, see Chenu, M.-D., Nature, Man, and Society in the Twelfth Century: Essays on New Theological Perspectives in the Latin West (trans. Jerome Taylor and Lester K. Little; Chicago: University of Chicago Press, 1968) 270–72Google Scholar.

33 The exact dating of works by these and other writers cited below is frequently uncertain. For chronological estimates of the works of Abelard, who repeatedly revises his writing over many years, see the overview of Eligius M. Buytaert in his edition of Petri Abaelardi opera theologica I (CCCM 11; Turnhout: Brepols, 1969), “General Introduction,” xxiii–xxiv; the adjustments proposed by Mews, Constant J. in “On Dating the Works of Peter Abelard,” Archives d'histoire doctrinale et littéraire du moyen âge 52 (1985) 73–134Google Scholar; and the later overview of Mews, in Petri Abaelardi opera theologica III (ed. Buytaert, E. M. and Mews, C. J.; CCCM 13; Turnhout: Brepols, 1987Google Scholar), “General Introduction,” 20–23. Despite some differences, these overviews broadly assign the formative period of Abelard's Theologia christiana to the 1120s, with later forms of the work and Abelard's Theologia “scholarium” assigned to the period from the mid-1130s to 1140. In citing passages from Theologia christiana, I refer in parentheses to passages from Theologia “scholarium” that are nearly identical to the initially cited texts. With regard to Abelard's Dialectica, chronological estimates have ranged from 1117–1119 to the early 1140s. For estimated times of composition of other twelfth-century works cited below, see Marcia L. Colish, “From the Sentence Collection to the Sentence Commentary and the Summa: Parisian Scholastic Theology, 1130–1215,” in Manuels, programmes de cours et techniques d'enseignement dans les universités médiévales. Actes du colloque international de Louvain-la-Neuve (9–11 septembre 1993) (ed. Jacqueline Hamesse; Louvain-la-Neuve: Institut d'Études Médiévales de l'Université Catholique de Louvain, 1994) 9–29, repr. in Colish, Marcia L., Studies in Scholasticism (Variorum Collected Studies Series CS838; Aldershot, Hampshire: Ashgate, 2006Google Scholar), section XII. With regard to works specified below, this overview offers the following approximate dates: Hugh of St. Victor, De sacramentis (1137); Summa sententiarum (1138–1142); Peter Lombard, Sententiae (1155–1157); Alan of Lille, Summa “Quoniam homines” (estimates ranging from 1155–1160 to 1170–1180) and Regulae caelestis iuris (1170–1185); Peter of Poitiers, Sententiae (perhaps 1167–1170, but possibly extending to 1175); and Gandulph of Bologna, Sententiae (1170–1185). The points in the discussion below do not depend upon these specific estimates.

34 See Abelard, Peter, Theologia christiana V, 20, in vol. 2 of Petri Abaelardi opera theologica (ed. Buytaert, Eligius M.; CCCM 12; Turnhout: Brepols, 1969) 354Google Scholar (with the corresponding passage in Theologia “scholarium” III, 20, in Buytaert, and Mews, , Opera theologica 3.508Google Scholar), indicating that the ability to sin (peccare) is to be attributed not to power (potentiae) but to weakness (infirmitati); Hugh of St. Victor, De sacramentis I, ii, 22, in PL 176, 214B, indicating that if God were able to destroy himself, this would be not an ability (posse), but an inability (non posse); and Peter Lombard, Sententiae I, dist. xlii, 3, in PL 192, 636, indicating that the ability to sin is a matter not of power (potentiœ) but of weakness (infirmitatis). For similar discussions, see Abelard, Peter, Dialectica (ed. de Rijk, L. M.; 2d ed.; Assen: Van Gorcum, 1970) 97Google Scholar; Peter Lombard, Sententiae I, dist. xlii, 5, in PL 192, 636; Alan of Lille, Summa “Quoniam homines,” in La somme “Quoniam homines” d'Alain de Lille (ed. P. Glorieux; Archives d'histoire doctrinale et littéraire du moyen âge 20 [1953] 113–364, at 227; Alan of Lille, Regulae caelestis iuris 54, in PL 210, 647B; Peter of Poitiers, Sententiae Petri Pictaviensis (ed. Philip S. Moore and Marthe Dulong; 2 vols.; Publications in Medieval Studies 7, 11; Notre Dame, Ind.: University of Notre Dame Press, 1943–50) 1.48; and Gandulph of Bologna, Magistri Gandulphi Bononiensis sententiarum libri quatuor (ed. Ioannes de Walter; Vienna: Aemilius Haim et Soc., Bibliopolae Academici, 1924) 108–9 (I, 164).

35 See Theologia christiana V, 18, in Buytaert, , Opera theologica 2.354Google Scholar (with Theologia “scholarium” III, 18, in Buytaert, and Mews, , Opera theologica 3.507Google Scholar): “Nemo enim hoc potentiae hominis deputat quod ille superari facile potest, immo impotentiae et debilitati ….” Compare the similar discussion in Abelard's Dialectica, ed. de Rijk, 97–98.

36 See Summa sententiarum I, 14, in PL 176, 69D–70A, a discussion similar to the argument of Hugh (capere ipsa non potest) in De sacramentis I, ii, 22, in PL 176, 215D–216B. Hugh is arguing against the necessitarian implications of parts of Abelard's treatment of God. While Hugh and Abelard differ in this regard, Abelard himself, working with different approaches to potentiality and modality, seeks to relieve constraints from the Christian God by exploiting the split references of language. See, for example, Theologia “scholarium” III, 57–64, in Buytaert, and Mews, , Opera theologica 3.524–27Google Scholar, where Abelard argues that though God is not able (non potest) to be incarnated now (modo), the limitation associated with “now” properly applies not to his ability, but to the changing circumstances in which he operates. From this perspective, his capacity in the past “to be incarnated” (incarnari) and his capacity in the present “to have been incarnated” (incarnatum esse) are not divergent kinds of possibilitas (nec diuersa possibilitas).

37 For these formulations, see Peter Lombard, Sententiae I, dist. xliii, 3–4, and I, dist. xliv, 2–3, in PL 192, 638 and 640. His terminology tends to associate linguistic questions of divided reference with broader interpretive inquiries; see, e.g., involuta, suggestive of the terminology of involucra (verbal wrappings) in the twelfth century; compare n. 11, above, on integumenta. On divided reference with regard to “better,” compare Peter of Poitiers, Sententiae, ed. Moore and Dulong, 1.58–59.

38 See the citations and comments correlated with nn. 4 and 17, along with n. 18, above.

39 See Peter Abelard, Theologia christiana V, 21, in Buytaert, , Opera theologica 2.356Google Scholar (with Theologia “scholarium” III, 21, in Buytaert, and Mews, , Opera theologica 3.509Google Scholar).

40 See Summa sententiarum I, 14, in PL 176, 68B–C.

41 See Peter Lombard, Sententiae I, dist. xlii, 2 and 7, in PL 192, 636–37.

42 See Alan's Summa “Quoniam homines,” ed. Glorieux, 227–29. Compare the similar discussion in Alan's Regulae caelestis iuris 56–60, in PL 210, 647D–650C.

43 See Aquinas, Thomas, Summa contra gentiles III, 69, 2445, in Liber de veritate catholicae fidei contra errores infidelium seu “SUMMA CONTRA GENTILES” (ed. Pera, Ceslas, with Marc, Pierre and Caramello, Pietro [in the order of editors for volumes 2 and 3 of the set]; 3 vols.; Turin: Marietti, 1961–1967) 3.96Google Scholar: “Detrahere ergo perfectioni creaturarum est detrahere perfectioni divinae virtutis.” For the translation, see Aquinas, Thomas, Summa contra gentiles III; Providence, Part 1 (trans. Bourke, Vernon J.; Garden City, N.Y.: Hanover House, 1956Google Scholar; repr., Notre Dame, Ind.: University of Notre Dame Press, 1975) 230.

44 See Summa theologiae I, q. 25, a. 3, Resp., in Summa theologiae (ed. Pietro Caramello; 5 vols.; Turin: Marietti, 1952–1956) 1.140: “… cum Deus omnia posse dicitur, nihil rectius intelligitur quam quod possit omnia possibilia ….” Such possibilia, continues Aquinas, do not imply a contradiction in terms (contradictionem non implicant). For these translations, see Basic Writings of Saint Thomas Aquinas (ed. Anton C. Pegis; 2 vols.; New York: Modern Library, 1945) 1.263. On the logical and ontological conditions of possibility in Aquinas's treatment of omnipotence, see Summa contra gentiles II, 25 (in the edition of Pera, Marc, and Caramello, 2.136–38), and Summa theologiae I, q. 25 at large (in the edition of Caramello, 1.138–44)—discussions that nonetheless show (e.g., Summa contra gentiles II, 25, 1028–32, and Summa theologiae I, q. 25, a. 5) how divine will can remain a substantial factor even in the context of a general shift in focus. On the general shift, see Resnick, , Divine Power, 35–39Google Scholar, although I think his account would be enhanced by a greater recognition of the continuing role of divine will and a closer assessment of Anselm, who is in effect relegated to the position of the “early Church”; see my discussion below.

45 Though there are significant differences between the arguments on which this essay concentrates and late-medieval distinctions between two aspects of divine power, absolute (potentia absoluta) and ordained (potentia ordinata)—with God by his absolute power capable of altering the order that he wills by his ordained power—these discussions raise overlapping issues about the conditions of possibility and contingency. For an overview of the development of the absolute/ordained power distinction, see William J. Courtenay, “The Dialectic of Divine Omnipotence,” in his Covenant and Causality, section IV. On some late-medieval approaches to the foundations of possibility, see Wolter, Allan B., “Ockham and the Textbooks: On the Origins of Possibility,” Franziskanische Studien Vierteljahr-Schrift 32 (1950) 70–96Google Scholar, repr. in Inquiries into Medieval Philosophy: A Collection in Honor of Francis P. Clarke (ed. James F. Ross; Contributions in Philosophy 4; Westport, Conn.: Greenwood, 1971) 243–73.

46 See Summa theologiae I, q. 25, a. 3, in Pegis, , Basic Writings, 1.263Google Scholar [emphasis added]. For the original text, see the edition of Caramello, 1.141 [emphasis in edition]: “convenientius dicitur quod non possunt fieri, quam quod Deus non potest ea facere.”

47 For the citation of this passage from Cur Deus homo II, ch. 17 (Schmitt, , Opera 2.123Google Scholar; Hopkins, and Richardson, , Complete Treatises, 377Google Scholar [emphasis added]), see n. 21 and the discussion corresponding to it. Compare Aquinas's comment in Summa theologiae I, q. 25, a. 4, ad 2 (edition of Caramello, 1.142), on the impossibility of annulling things of the past: “dicitur Deus ea non posse, quia ea non possunt fieri”; for a translation, see Pegis, , Basic Writings, 1.266Google Scholar [emphasis added]: “God is said not to be able to do them because they themselves cannot be done.”