Home » Library » Modern Library » Plantinga’s Selective Theism: The Circular Reasoning at the Heart of Where the Conflict Really Lies

Plantinga’s Selective Theism: The Circular Reasoning at the Heart of Where the Conflict Really Lies


 Alvin Plantinga. 2011. Where the Conflict Really Lies: Science, Religion, and Naturalism. New York, NY: Oxford University Press. xvi+376 pp.

Plantinga’s Selective Theism and Argument that P(R/T&E) is High
Plantinga’s Dilemma
A More Balanced Theism and Argument that P(R/T&E) is Low


In Where the Conflict Really Lies: Science, Religion, and Naturalism (hereafter WCRL), philosopher and Christian apologist Alvin Plantinga invites his readers to conclude that there is “deep concord” between theistic religion and science.[1] In fact, according to Plantinga, the only way that we can have confidence that our cognitive faculties and beliefs are reliable, and thus that many of our beliefs (including scientific beliefs) are true, is to assume that God structured the process of evolution by natural selection with that goal in mind. Stated as a conditional probability, Plantinga holds that P(R/T&E) is high (i.e., “the Probability that our beliefs are Reliable, assuming Theism and Evolution” is high). In contrast, naturalistic, unguided evolution could not be expected to produce reliable beliefs, including scientific knowledge[2], according to Plantinga’s evolutionary argument against naturalism.[3] Therefore, advanced knowledge of the world in the form of science could not have emerged without God’s hand guiding the evolution of human beings so as to produce cognitive faculties that generate reliable beliefs.

Plantinga’s argument for the “deep concord” between theism and science relies on a version of Christian theism in which God’s very nature ensures that humans would be expected to have reliable, sophisticated knowledge of the world. Starting with axiomatic concepts in Christian theism, namely that God is omniscient and that humans are created in God’s image, Plantinga argues that, under theistic evolution, one would expect human cognitive faculties to be reliable, such that the majority of human beliefs about the world would be true. Plantinga’s definitions pertaining to the reliability of human cognitive faculties and beliefs are both vague and simplistic, but there’s a larger problem that’s fatal to his argument.

What if there is a more balanced view of Christian theism in which God has another, higher priority that is in opposition to any interest that God might have in humans having a majority of true scientific beliefs about the world, so that P(R/T&E) would be low rather than high? In the face of two plausible interpretations of Christian theism leading to opposite conclusions, it would become impossible to estimate P(R/T&E); the conditional probability statement would be inscrutable, and one of the main pillars of Plantinga’s argument for the “deep concord” of theism and science would fail. (The other pillars of his argument are critiqued in Richard M. Smith’s review of Where the Conflict Really Lies.[4])

In this paper, I describe such a mainstream, plausible version of Christian theism as an alternative to Plantinga’s version. Plantinga is guilty of circular reasoning in that he chose a narrow, highly selective representation of Christian theism to support his preferred conclusion. In doing so, Plantinga failed to fully consider other valid representations of Christian theism, as well as empirical facts that might undermine his conclusion.

Plantinga’s Selective Theism and Argument that P(R/T&E) is High

Plantinga frames his argument by focusing on one of God’s fundamental characteristics, omniscience: “God is a knower, and indeed the supreme knower. God is omniscient, that is, such that he knows everything…. We human beings, therefore, in being created in his image, can also know much about our world, ourselves, and God himself” (WCRL, p. 268).

To flesh out these theological assertions, Plantinga maintains that God structured evolution by natural selection so that human beings have knowledge of the world in the form of “reliable beliefs,” the concept represented by “R” in P(R/T&E). It is surprising, then, that Plantinga’s definitions of “beliefs” and what it means for them to be “reliable” are so vague and so uninformed by scientific knowledge available at the time the book was written.

Plantinga provides an idiosyncratic list of human “cognitive faculties” that can generate different types of beliefs: “perception … memory … a priori insight … broadly inductive procedures … sympathy … and perhaps still others” (WCRL, p. 270). Later in the book he adds “introspection,” “testimony,” the “moral sense,” the “sensus divinitatis,” and “the whole process of theory building, which may or may not be reducible to the previous abilities” to his list of cognitive faculties (WCRL, p. 312).

Perception and memory are indeed fundamental cognitive faculties that generate beliefs, but there’s no mention of human executive functioning, the faculty by which beliefs are evaluated and decisions about action are made, a fundamental concept in cognitive psychology. Likewise, there’s no mention of the trait of social intelligence, which is a defining characteristic of our species that plays a central role in the discovery, refinement, and transmission of knowledge through collaborative effort, as exemplified in scientific research.

Plantinga’s discussion of what it means for beliefs to be “reliable” is both vague and simplistic. Plantinga does not specify which cognitive faculties and resulting beliefs must be reliable in order to draw the conclusion that our beliefs are (or aren’t) reliable. Must they all be reliable, or just a few key faculties and associated categories of beliefs? Are some cognitive faculties more important for general belief-reliability than others? Plantinga doesn’t say; he simply asserts that “We can speak of the reliability of a particular faculty—memory, for example—but also of the reliability of our whole battery of cognitive faculties” (WCRL, p. 313). Presumably, then, “R” in P(R/T&E) refers to “our whole battery of cognitive faculties,” leaving us in the dark about whether they are weighted or must all meet the same criterion of reliability. Plantinga defines “reliable” cognitive faculties by the proportion of true beliefs produced by that faculty. “What proportion of my memorial beliefs must be true for my memory to be reliable? Of course there is no precise answer; but presumably it would be greater than, say, two-thirds” (WCRL, p. 312-313).

The problem here goes beyond vague definitions to, more importantly, Plantinga’s failure to take into account even the most basic scientific understanding of how human beliefs are generated, evaluated, and refined. Typically, beliefs come into consciousness based on the unconscious activation of one or more previously acquired beliefs, hopefully ones that are true and applicable to the situation. Beliefs typically carry with them a degree of certainty or uncertainty, which can be expressed as the degree of confidence that we have in the accuracy or relevance of a belief for the situation at hand. For example, have I, at first glance, correctly identified a tree in my yard as a black locust tree? I may be quite certain of its classification, or I may be less confident, and choose to utilize a book or a computer program to help me assign the tree to the right category.

In cognitive psychology, much human activity can be described as decision-making under uncertainty. Becoming confident enough in a belief to act on it often requires an iterative, reflective process. Therefore, it’s painfully simplistic to refer to our beliefs as reliable in terms of the proportion that are true, without saying something about the point in the process when a belief’s truth is assessed. Is Plantinga referring to our initial, first-impression beliefs in a situation, beliefs that are often evaluated and corrected before action is taken, as in the tree-classification example? Human executive functioning allows for beliefs in which people have low confidence to be refined or replaced. This executive functioning is crucial in medical practice, for example, where a substantial proportion of diagnostic first impressions may be incorrect or incomplete. Judging the reliability of human beliefs by the accuracy of first-impression beliefs is likely to understate the accuracy of the beliefs of individual human beings after they expend additional effort.

At a group level, collaborative knowledge-building efforts, whereby beliefs are compared, refined, and combined over time by groups of people in order to produce bodies of knowledge, have been fundamental to the success of our highly social species. When he speaks of reliability, is Plantinga referring to the proportion of human beliefs produced by a particular cognitive faculty that are true after a persistent group effort over time, utilizing self-correcting processes such as those found in the scientific research community, in a sustained attempt to produce true beliefs and useful, generalizable knowledge? Plantinga doesn’t say, leaving us in the dark about how to measure and apply his criterion for the reliability of our beliefs.

Plantinga’s reputation as one of the more scientifically informed Christian apologists is not borne out in Where the Conflict Really Lies. One could reasonably argue that Plantinga’s inadequate explication of cognitive faculties, beliefs, and the criterion for them to be judged as reliable are flawed enough to make P(R/T&E) inscrutable rather than high. However, for purposes of the present discussion, let’s assume that the concept of the “reliability of our beliefs” is meaningful enough to merit a continued analysis of Plantinga’s claim that P(R/T&E) is high.

Plantinga claims that unguided evolution by natural selection only selects for behavior, in that behavior is caused by neurophysiological properties that have no necessary connection to specific beliefs, which have no causal power of their own (WCRL, p. 312). Therefore, unguided evolution would select for the neurophysiological properties that produce adaptive behavior, not for true beliefs and reliable cognitive faculties; this is a central, controversial claim in his evolutionary argument against naturalism.

In Plantinga’s view, God’s intervention in the workings of evolution by natural selection was therefore necessary for true beliefs and reliable cognitive faculties to emerge. To accomplish this, God had to connect adaptive behavior, which Plantinga sometimes refers to as “successful action,” to reliable cognitive faculties that produce true beliefs. Plantinga reasons as follows: “if, as most theists think, God has created us in his image, including the ability to have knowledge, then God would presumably establish psychophysical laws of such a sort that successful action is correlated with true belief” (WCRL, footnote on p. 339). In other words, since God would see to it that the human brain evolved so that successful, adaptive action is “correlated with true belief,” then the tendency to form true beliefs would be selected and strengthened over time, and human beings would develop reliable cognitive faculties and the ability to learn a great deal about the world.

With God’s intervention in the evolutionary process, many generations of natural selection would produce cognitive faculties that generate far more true beliefs than false ones, satisfying Plantinga’s threshold for beliefs to be reliable. Therefore, P(R/T&E) is high.

Plantinga’s Dilemma

Plantinga’s argument glosses over the fact that he faced a consequential dilemma in framing it. On the one hand, to only emphasize God as the supreme knower, with humankind as lesser knowers in his image, would serve Plantinga’s goal, but it would also understate the characteristics and perfection of God, and the range of goals that God has for human beings. On the other hand, to more fully explore what it means to be made in God’s image, which might reveal that God has one or more higher priorities than human beings possessing advanced knowledge of the world, could undermine Plantinga’s goal of subsuming science under theism.

In trying to finesse this dilemma, Plantinga criticizes Thomas Aquinas for making the same error of emphasis that is central to Plantinga’s own argument. Plantinga first quotes Aquinas: “Since human beings are said to be in the image of God in virtue of their having a nature that includes an intellect, such a nature is most in the image of God in virtue of being most able to imitate God…” (WCRL, p. 268). Plantinga then gently chides Aquinas for possibly overemphasizing the human ability to acquire knowledge as an aspect of being made in God’s image: “Here Aquinas says that a nature including an intellect is most in the image of God, in virtue of being most able to imitate God. Perhaps he exaggerates a bit in thinking that understanding, the ability to know, is the chief part of the image of God. What about being able to act, what about having a grasp of right and wrong, what about being able to love one another, what about being able, in some way, to experience God?” (WCRL, p. 269).

Plantinga thus gives the appearance of seriously considering and taking into account God’s other goals for humankind without actually doing so. He knew that to pursue that broader discussion would undermine his argument that Christian theism conclusively entails the high reliability of human beliefs, so he broached the topic of other implications of being made in the image of God and promptly dropped it. In fact, Plantinga’s own argument relies on a narrow focus on God’s omniscience and his desire for human beings to have knowledge of the world that is every bit as imbalanced and incomplete as the position attributed to Aquinas. I broaden the discussion to include the most theologically plausible candidate for God’s highest priority for humankind in the next section of the paper.

A More Balanced Theism and Argument that P(R/T&E) is Low

Having advanced knowledge of the world can help humans to flourish during our brief lives on Earth, but being in the right sort of relationship with God is much more important, for it has eternal consequences, not just fleeting earthly consequences. In Christian theism, God’s highest priority for humankind is that human beings come to know him in such a way that they can spend eternity in God’s presence: “For God so loved the world that he gave his one and only Son, that whoever believes in him shall not perish but have eternal life” (John 3:16, NIV). Compared to eternal life with God, earthly things such as possessions and knowledge of the world are unimportant. As Jesus said, “What good will it be for someone to gain the whole world, yet forfeit their soul?” (Matthew 16:26, NIV).

Coming to know God in the right way is so important that, according to Plantinga and others, God gave human beings a special cognitive faculty, the sensus divinitatis, to draw people to God (WCRL, p. 60). The sensus divinitatis operates through common experiences, such as seeing a beautiful sunset or experiencing guilt. For example, consider the experience of famed scientist Francis Collins, who was an atheist in his early 20s but began to read C. S. Lewis and consider Christianity during his medical training. He committed to Christianity after the sort of emotional, subjective experience that could be attributed to a sensus divinitatis, which functions so that advanced knowledge of the world is not necessary in order to come to know God.

On a beautiful fall day, as I was hiking in the Cascade Mountains … the majesty and beauty of God’s creation overwhelmed my resistance. As I rounded a corner and saw a beautiful and unexpected frozen waterfall, hundreds of feet high, I knew the search was over. The next morning, I knelt in the dewy grass as the sun rose and surrendered to Jesus Christ.[5]

It makes sense that God would design the sensus divinitatis in this way, for God would know that having advanced knowledge of the world is not only difficult to acquire—it also tends to undermine theistic belief. As philosopher Barbara Forrest said, “The metaphysical adequacy of supernaturalism is inversely proportionate to the explanatory power of science. The more science successfully explains, the less need or justification there is for the supernatural as an explanatory principle.”[6]

There is, therefore, a fundamental tension between God’s highest priority—the salvation of as many people as possible for an eternity in Heaven with God—and human possession of advanced knowledge of the world. In national surveys conducted from 2006-2009, 83% of the American public expressed belief in God, compared to only 33% of scientists.[7] Among the most accomplished American scientists (members of the National Academy of Sciences), a 1996 survey showed that only 7% believed in a personal God.[8] These data represent a strong negative linear relationship between scientific knowledge and theistic belief among Americans.

The same negative relationship between scientific and technological mastery and theistic belief holds among nation-states. Countries with high levels of scientific knowledge tend to be prosperous, supporting a high standard of living for the country’s citizens, accompanied by relatively low levels of theistic belief. Less prosperous nations tend to have the highest levels of religiosity compared to more prosperous nations.[9]

God, being omniscient, would know that having advanced knowledge of the world would produce, on average, lower levels of theistic belief. Therefore, God’s goal would be to establish the reliability of human beliefs at a modest level sufficient for survival and reproduction while minimizing the acquisition of advanced knowledge about the world, thus maximizing the proportion of humans who would come to know God and spend eternity in Heaven.

As previously noted, in Plantinga’s view, God has the power to structure and guide the process of evolution to produce the desired level of reliability of the beliefs generated by each human cognitive faculty. Under Christian theism, basic cognitive faculties such as perception and memory would be expected to have adequate reliability to support human survival and reproduction; however, advanced cognitive faculties such as those necessary for scientific knowledge would have low reliability. Therefore, P(R/T&E), at least in relation to scientific knowledge, is low.


The latter conclusion concerning the low expected reliability of our beliefs under Christian theism is the opposite of Plantinga’s conclusion about theism and the high expected reliability of our beliefs. Plantinga’s argument for the high reliability of beliefs given theism relies on a selective version of Christian theism that predetermines the outcome of the argument; his argument is therefore circular, and P(R/T&E) is inscrutable rather than conclusively being low or high.

It’s a bit ironic that Plantinga wants so badly to bring science under the rubric of theism when the scientific age has been such a disaster for theistic belief in the world. A higher proportion of human beings would likely hold theistic beliefs if knowledge of the world was fixed at the modest level found in the universally religious Middle Ages. Indeed, one could use the facts about the negative relationship between scientific knowledge and theistic belief at both individual and societal levels of analysis, along with the high and still increasing level of scientific knowledge in the world today, to construct an argument against the existence of the Christian God. Why would an omniscient, omnipotent, loving God make his favorite creatures too smart for their own good, reducing their chances of spending eternity in his presence and increasing their chances of spending eternity in a lake of fire after the dead are judged by God? (Revelation 20:15) And why would a loving, all-knowing God structure evolution so that the smartest humans, who are therefore most like God in terms of intellect and knowledge, would be the least likely to believe in him?

In addition to the specific argument in this paper, which relies on contrasting versions of Christian theism leading to different conclusions about the expected reliability of human beliefs, the paper provides an example of a more general point about Christian theistic arguments: their extreme flexibility. God is held to be perfect, and therefore to have every desirable characteristic to its ultimate degree: omnibenevolence, omniscience, omnipotence, and other characteristics that approach contradiction, such as transcending space and time, but also being immanent in all things. In addition, scripture provides an abundance of lofty sounding but ambiguous and even contradictory starting points from which to build a theistic argument to a specific end. Many scriptural passages attribute behavior to God that seems to contradict his fundamental attributes, providing even more interpretive flexibility. Christian apologists like Plantinga have plenty of theological room in which to maneuver.

Where the Conflict Really Lies is couched in the lofty language of theology and the formalism of analytic philosophy. However, the book’s appearance of rigor cannot hide the fact that Plantinga gave brief consideration to a mainstream view of God’s wide range of goals for humankind, but then, in a display of circular reasoning, chose to focus on a narrow version of Christian theism consistent with his desired conclusion.


I thank an anonymous reviewer for comments on this paper. The reviewer suggested that a more relevant approach than Plantinga’s would be to compare the conditional probabilities that our actual cognitive faculties (C), with all of their well-known biases, would have evolved (E) under naturalism (N) compared to theism (T). In other words, compare P(C/N&E) to P(C/T&E). The reviewer further suggested that discussion of such phenomena as the “hyperactive agency detection device” (HADD) would contribute to such a comparison.


[1] Alvin Plantinga, Where the Conflict Really Lies: Science, Religion, and Naturalism (New York, NY: Oxford University Press, 2011).

[2] Alvin Plantinga, “Naturalism vs. Evolution: A Religion/Science Conflict?” (September 1, 2007). The Secular Web. <https://infidels.org/library/modern/alvin-plantinga-conflict/>.

[3] Paul Draper, “In Defense of Sensible Naturalism” (September 1, 2007). The Secular Web. <https://infidels.org/library/modern/paul-draper-naturalism/>.

[4] Richard M. Smith, “Alvin Plantinga Can’t Say That, Can He? A Review of Where the Conflict Really Lies” (February 11, 2016). The Secular Web. <https://infidels.org/library/modern/richard-smith-plantinga/>.

[5] Francis S. Collins, The Language of God: A Scientist Presents Evidence for Belief (New York, NY: Free Press, 2006), p. 225.

[6] Barbara Forrest, “Methodological Naturalism and Philosophical Naturalism: Clarifying the Connection.” Free Inquiry Vol. 37, No. 5 (August/September 2017).

[7] David Masci, “Scientists and Belief.” Pew Research Center (November 5, 2009). <https://www.pewresearch.org/religion/2009/11/05/scientists-and-belief/>.

[8] Edward J. Larson and Larry Witham, “Leading Scientists Still Reject God.” Nature Vol. 394 (July 23, 1998): 313.

[9] Steve Crabtree, “Religiosity Highest in World’s Poorest Nations.” Gallup (August 31, 2010). <https://news.gallup.com/poll/142727/Religiosity-Highest-World-Poorest-Nations.aspx>

Copyright ©2023 Doug Mann. The electronic version is copyright ©2023 by Internet Infidels, Inc. with the written permission of Doug Mann. All rights reserved.

all rights reserved