Skip to main content

On the Concept of Theodicy


My purpose in this paper is to clarify or explicate the concept of theodicy. More specifically, I shall provide an account of the concept that takes its logical aspects seriously into consideration as well as satisfies the basic intuitions philosophers of religions have had about it. This shall be done by systematically analysing the several theodical conditions found in the literature. As it shall be seen, these conditions are logically related to one another; collectively, they point not to one, but to several concepts of theodicy. Thus, as by-product of this explicatory endeavour, I shall provide a logically guided, I may say, analysis of such conditions as well as a systematization of the theodical concepts arising from them. I shall follow what might be termed a semiformal approach; despite not developing a full logical theory, I use the standard notation and some important results from the field of formal logic.

This is a preview of subscription content, access via your institution.


  1. That these reasons should give a moral justification is clear. The purpose of a theodicy is to conciliate the existence of evil with the existence of an omnipotent, omniscient, and perfectly good being. If the reasons given do not have a moral character, then one feature, namely perfect goodness, would be left out. It is not, as one might think, an asymmetry regarding the treatment given to the three features; instead, it is an exigence of the problem of evil itself.

  2. This taxonomy conflicts with a traditional one according to which a theodicy is (and only is) the theist’s answer to the evidential problem of evil; a defence would be the theist’s answer to the logical problem of evil. According to this tentative taxonomy of mine, there can be a theodicy aimed at answering the logical problem of evil; Plantinga’s so-called free-will defence (1974, 1977) would be an example of this. On the other hand, a defence, as the very name indicates, would be an answer to (or a tentative refutation against) a specific argument from evil, be it logical or evidential. While, for instance, Pike (1963) would be a defence against Hume’s logical argument, Wykstra (1984) well-known CORNEA critique would be a defence against one of Rowe’s evidential arguments (1979). Although I am convinced this is a much more satisfactory and elegant taxonomy, for space reasons, I shall not elaborate on it any further; besides, it is in no way an essential part of what I intend to accomplish in this paper.

  3. There are different uses for the terms ‘problem of evil’ and ‘theodicy’ in the literature. As I made clear in the beginning, my use of them follows the standard in analytic philosophy of religion and assumes the concept of God as encompassing an omnipotent, omniscient, and perfectly good being. Therefore, a ‘solution’ that denies that God is omniscient, for instance, is not a solution at all.

  4. For space reasons, I shall not discuss the standard objections to the project of theodicy. Despite all drawbacks that have been pointed out, it is this project, as traditionally conceived since Hume, say, that is the object of my analysis.

  5. The term ‘meta-theodicy’ was used for the first time by Peterson (1983: 321) to refer to the structural analysis of arguments from evil.

  6. See Salmon (1989) and Ruben (19900, for instance.

  7. In the course of the paper, I shall use the logical symbols , ¬, and with their usual meaning. As is customary, I omit the brackets in the case of unitary sets.

  8. Notice the similarity with Plantinga’s well-known characterization of how a defence is supposed to answer the logical problem of evil (1977: 26). The only formal difference is R⊬S, which is there to guarantee that G is indeed a relevant component of (T); Γ⊬α is an abbreviation for ‘it is not the case that Γ⊢α.’

  9. I am here relying on an intuitive understanding of the notion of evidential support. This however does not mean that specific aspects of it, like the two above mentioned, do not find support in existing theories of non-deductive inference. More on that later.

  10. For a discussion on the history and rationale behind the project of formalization in philosophy See Williamson (2017), Horsten and Pettigrew (2011), Van Benthem (2006), and Hansson (2000).

  11. Two comments on my use of propositions in this paper. First, I shall overlook the conceptual differences that exist between propositions on the one hand and sentences and statements on the other. Second, I shall use the same symbol to refer to both a proposition and the state of affairs expressed by it. Given a proposition named A, for instance, I use the symbol ‘A’ not only to refer to the proposition it names, as when I say that A is true, but also to refer to the state of affairs expressed by the proposition, as when I say that A is the case.

  12. As the word ‘factual’ tries to point out, the content of h refers to the world, in a wide sense of the term; in particular, h possibly says something about the moral value of human actions.

  13. This condition has appeared in Hick (1981: 39) and Trakakis (2007: 243). Although neither Hick nor Tratakis elaborate on the rationale of this condition, it is not difficult to see why it must be there. First, if is classical logic inferential relation, then if (2) is not satisfied, {G, R, h} would be an inconsistent set and (1) true by vacuity (in classical logic, an inconsistent set Γ is such that Γ⊢α for every α). From a more general perspective, and considering the case where might be an inductive inferential relation, (2) means that G disproves R; but what is the point of bringing R into scene to help G explaining S if R is proved to be false in the presence of G?

  14. Needless to say, in the case is classical logic’s inferential relation, (2) and (3) are equivalent. Although (3) is not found in the literature, for the sake of symmetry, I think there must be an evidential counterpart to (2). For its rationale, I can restate the justification given for (2) in terms of an evidential relation: (3) means that G serves as evidence against R; but what is the point of bringing R into scene to help G explaining S if R is made implausible in the presence of G?

  15. This condition has appeared in Hasker (1988: 5, Van Inwagen (1991: 139), and Tooley (2002: 22).

  16. Here, while h would roughly correspond to R; ‘theism’ means the same as G.

  17. This is the cut property, which in general holds in deductive logics and is seen as a desirable property of commonsense and non-monotonic reasoning. See Makinson (1994: 39).

  18. See Hempel (1945: 105). For a historical exposition on the debate on this principle See Silvestre (2010: 41–69).

  19. The fourth condition, corresponding to S 1, is condition (1).

  20. See, for instance, Rowe (1988) and Trakakis (2007: 233).

  21. The most notable exception is of course relevant logic.

  22. The numbering inside the brackets is mine.

  23. Theodicies which make use of this idea are called greater-good theodicies. Since this is by far the most traditional type of theodicy, for space reasons, I have decided to focus exclusively on it and neglect other types of theodicies which follow different paths in their attempts to justify why God allows evil and suffering in our world.

  24. See Rowe (1979: 336).

  25. We use the expression ‘a good is expressed in R’ in the usual sense of a state of affairs A being expressed in a proposition P; if P is true then A is necessarily the case.

  26. This condition has appeared in Schuurman (1993: 210–211), Swinburne (1995: 75–76), Langtry (1998: 4), and Trakakis (2007: 237).

  27. This condition has appeared in Rowe (1979: 336), Swinburne (1995: 760, Langtry (1998: 5–9), and Trakakis (2007: 237–238).

  28. See, for instance, Plantinga (1979: 7), Schuurman (1993: 210), Chisholm (1990: 60–62), and Trakakis (2007: 234).

  29. By using this presupposition, I am adopting a strategy similar to the one I have adopted regarding . In the same way that my analysis does not depend on an existing satisfactory theory of inference, it does not depend either on a satisfactory theory for explaining the expression ‘good x is greater than an evil y.’ As far as my analysis is concerned, and ‘good x is greater than an evil y’ can be understood, for instance, as unanalysed notions whose use relies on their common sense meaning. The same applies to the expressions ‘worse than or as bad’ and ‘necessarily true’ in condition (10). About this later, although I am deliberately avoiding technical discussions such as which logical framework better captures a specific notion, a minimal interpretation for ‘necessarily true’ is one that takes α as necessary if α is true in all possible worlds under consideration.

  30. This condition has appeared in Adams (1999: 307) and Trakakis (2007: 236–237). It is worth noting that both Adams and Trakakis speak of this condition exclusively in the context of horrors.

  31. Hick’s criterion of possibility is our internal consistency condition (2).

  32. It might be that despite G⊬¬R and W⊬¬R, when taken together G and W evidentially imply ¬R.

  33. This is due to the non-monotonicity property of non-deductive inferential relations. See Makinson (1994) and Silvestre (2010).

  34. Recall that (2) is implied by (3).

  35. Abbreviation for ‘if and only if.’

  36. Due to the already mentioned semi-formal character of my approach (2) is only partially fulfilled, and I do believe it is simple enough to say (4) has been fulfilled.


  • Adams, M. M. (1999). Horrendous evils and the goodness of God. Melbourne: Melbourne University Press.

    Google Scholar 

  • Adams, R. M. (1985). In J. Tomberlin & P. Van Inwagen (Eds.), Plantinga on the problem of evil, in Alvin Plantinga (pp. 225–255). Dordrecht: D. Reidel Publishing Company.

    Chapter  Google Scholar 

  • Carnap, R. (1950). Logical foundations of probability. Chicago: University of Chicago Press.

    Google Scholar 

  • Chisholm, R. M. (1990). In M. Adams & R. Adams (Eds.), The defeat of good and evil, in the problem of evil (pp. 53–68). Oxford: Oxford University Press.

    Google Scholar 

  • Hansson, S. (2000). Formalization in philosophy. The Bulletin of Symbolic Logic, 6, 162–175.

    Article  Google Scholar 

  • Hasker, W. (1988). Suffering, soul-making, and salvation. International Philosophical Quarterly, 28, 3–19.

    Article  Google Scholar 

  • Hempel, C. (1945). Studies in the logic of confirmation. Mind, 54(1–26), 97–121.

    Article  Google Scholar 

  • Hick, J. (1981). In S. Davis (Ed.), An Irenaean theodicy, in encountering evil: Live options in theodicy (pp. 39–52). Edinburgh: T and T Clark.

    Google Scholar 

  • Horsten, L., & Pettigrew, R. (2011). In L. Horstein (Ed.), Mathematical methods in philosophy, in the continuum companion to philosophical logic (pp. 14–26). London: Continuum Publishing.

    Google Scholar 

  • Langtry, B. (1998). Structures of greater good theodicies: the objection from alternative goods. Sophia, 37, 1–17.

    Article  Google Scholar 

  • Makinson, D. (1994). In D. Gabbay et al. (Eds.), General patterns in non-monotonic reasoning, in handbook of logic in artificial intelligence and logic programming (Vol. Volume 3, pp. 35–110). Oxford: Oxford University Press.

    Google Scholar 

  • Peterson, M. (1983). Recent work on the problem of evil. American Philosophical Quarterly, 20, 321–339.

    Google Scholar 

  • Pike, N. (1963). Hume on evil. The philosophical Review, 72, 180–197.

    Article  Google Scholar 

  • Plantinga, A. (1974). The nature of necessity. Oxford: Oxford University Press.

    Google Scholar 

  • Plantinga, A. (1977). God, freedom, and evil. Grand Rapids: Eerdmans.

    Google Scholar 

  • Plantinga, A. (1979). The probabilistic argument from evil. Philosophical Studies, 35, 1–53.

    Article  Google Scholar 

  • Rowe, W. (1979). The problem of evil and some varieties of atheism. American Philosophical Quarterly, 16, 335–341.

    Google Scholar 

  • Rowe, W. (1988). Evil and theodicy. Philosophical Topics, 16, 119–132.

    Article  Google Scholar 

  • Rowe, W. (Ed.). (2001). God and the problem of evil. Oxford: Blackwell Press.

    Google Scholar 

  • Ruben, D. (1990). Explaining explanation. New York: Routledge.

    Book  Google Scholar 

  • Salmon, W. (1989). In P. Kitcher & W. Salmon (Eds.), Four decades of scientific explanation, in Minnesota studies in the philosophy of science, vol XII, scientific explanation. Minneapolis: University of Minnesota Press.

    Google Scholar 

  • Schuurman, H. (1993). Two concepts of theodicy. American Philosophical Quarterly, 30, 209–221.

    Google Scholar 

  • Silvestre, R. (2010). Induction and plausibility. Berlin: Lambert Academic Publishing.

    Google Scholar 

  • Swinburne, R. (1995). Theodicy, our well-being, and God’s rights. International Journal for Philosophy of Religion, 38, 75–91.

    Article  Google Scholar 

  • Swinburne, R. (1996). In D. Howard-Snyder (Ed.), Some major strands of theodicy, in the evidential argument from evil (pp. 30–48). Bloomington: Indiana University Press.

    Google Scholar 

  • Tooley, M. 2002. The Problem of Evil, in The Stanford Encyclopaedia of Philosophy, ed. E. Zalta, Winter 2002 edition.

  • Trakakis, N. (2007). The god beyond belief: in defense of William Rowe’s evidential argument from evil. Dordrecht: Springer.

    Google Scholar 

  • Van Benthem, J. 2006. Logic in philosophy, in philosophy of logic, ed. D. Jacquette, Amsterdam: North-Holland: 65–100.

  • Van Inwagen, P. (1991). The problem of evil, the problem of air, and the problem of silence. Philosophical Perspectives, 5, 135–165.

    Article  Google Scholar 

  • Van Inwagen, P. (2000). The argument from particular horrendous evils. Proceedings of the American Catholic Philosophical Association, 74, 65–80.

    Article  Google Scholar 

  • Williamson, T. (2017). In R. Blackford et al. (Eds.), Model-building in philosophy, to appear in: philosophy’s future: the problem of philosophical progress. Oxford: Wiley.

    Google Scholar 

  • Wykstra, S. (1984). The Humean obstacle to evidential arguments from suffering: on avoiding the evils of appearance. International Journal for Philosophy of Religion, 16, 73–93.

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Ricardo Sousa Silvestre.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Silvestre, R.S. On the Concept of Theodicy. SOPHIA 56, 207–225 (2017).

Download citation

  • Published:

  • Issue Date:

  • DOI:


  • Concept explication
  • Theodicy
  • Problem of evil
  • Logical analysis