James Douglas Baird
’15, Covenant College
Deciphering the nature of knowledge and epistemic warrant is a sticky task that has received many avid participants in recent years. In this paper, I examine the contours of three proposed accounts of epistemic warrant: justificatory, proper functionalist, and causal. I conclude that a causal theory of warrant has promise because of its intuitive appeal, flexibility, and lack of vulnerability to the problems of the justificationist and proper functionalist accounts of warrant.
Paul K. Moser defines epistemology as the “philosophical study of the nature, origin and scope of knowledge.” From the beginning of modernity until the mid-20th century, the origin or source of knowledge predominantly occupied the thoughts of epistemologists. Some of the greatest debates of the modern age revolved around the question of whether knowledge was gained through the a priori principles of reason or primarily from the deliverances of sense experience. Contemporary philosophy has concerned itself with quite another issue: the task of providing an analysis of the nature of knowledge. The focus of this contemporary conceptual analysis has been on propositional knowledge, or knowledge of the truths which that-phrases express. A bundle of knotty problems have arisen as philosophers have attempted to identify the conditions which are individually necessary and together sufficient for propositional knowledge to obtain.
The general consensus among epistemologists is that knowledge requires both a truth condition and a belief condition; in order for S to know that p, it is necessary for S to believe that p and for that p to be true. Knowledge, however, is more than mere true belief. As Alvin Plantinga puts the issue, “There is wide agreement that knowledge requires true belief; but as far back as Plato’s Theaetetus, there is also recognition that it requires more. I may believe that I will win a Nobel Prize next year; by some mad chance my belief may be true; it hardly follows that I know the truth in question.” At least a third condition must be added to true belief to constitute knowledge. Since the beginning of the modern age (initiated by the philosophical Meditations of Rene Descartes), the justified-true-belief (hereafter, JTB) analysis has dominated the epistemological playing-field. On this account, the third necessary condition in the analysis of knowledge is justification; that is, justification as typically construed in terms of evidence. The JTB thesis asserts that one must have true belief and evidence supporting one’s belief in order to have knowledge. In 1963, Edmund Gettier dropped a three page bombshell onto the landscape of epistemology. In his short article, “Is Justified True Belief Knowledge?” Gettier presented hypothetical cases wherein an individual has a justified true belief and yet fails to have knowledge. The import of Gettier-type cases is that the JTB thesis is an insufficient analysis. Debate has since swirled around whether the JTB thesis is insufficient because knowledge requires a fourth condition or requires a replacement concept for justification.
The search for warrant—defined by Plantinga as whatever it is which must be added to true belief in order to yield knowledge—continues to draw its impetus from Gettier’s illuminating article and the numerous responses given by philosophers. Alvin I. Goldman set forth a causal theory of warrant early on in response to the Gettier-problem. According to the early Goldman, “S knows that p if and only if . . . the fact p is causally connected in an ‘appropriate’ way with S’s believing p.” On this view, I know that there is a table in front of me if and only if I believe that there is a table in front of me, it is true that there is a table in front of me, and the table in front of me caused my perceptual belief that there is a table in front of me. According to causal accounts, warrant is an appropriate causal connection between a subject’s true belief and the fact that makes her belief true. Causal analyses of knowledge have fallen by the wayside because of two main challenges: (1) their supposed inability to account for our a priori knowledge; and (2) Gettier-type counter-examples where a subject, S, supposedly has an appropriately caused true belief but fails to attain knowledge. I will argue that the above two objections are inconclusive against a properly formulated causal theory of warrant, and that a causal theory of warrant has much promise compared to some of its fellow accounts. First, I will examine the justificationist and proper functionalist accounts of warrant and their relevant problems. Second, I will present a causal account of warrant and exposit the ways a causal account avoids the problems of the JTB and proper functionalist accounts as well as incorporates their insights. Finally, I will respond to the two main objections given against causal accounts of epistemic warrant.
I. The Rationale and Difficulties of Justificationist and Proper Functionalist Accounts
Plantinga seems to think that a belief which is true by a “mad chance” does not constitute knowledge. Robert Audi makes this explicit: “a belief true just by good luck does not constitute knowledge.” Plantinga and Audi do not mean that the fact a belief is about must obtain by divine providence, fate, or some other principle of determinacy before said belief can constitute knowledge. Instead, they are asserting that one’s assenting attitude towards a true proposition must be formed correctly and not inappropriately, as by a guess. Answering what it means to form a true belief correctly and not by chance/luck is the task before those presenting an account of warrant. The JTB analysis of knowledge asserts that to have epistemic warrant for one’s true belief means to have good reasons for believing a particular proposition is true. Knowledge, according to the justificationist, is a true belief which is not formed luckily, because it is held according to the evidence one has for it. Unfortunately for the JTB theorist, the post-Gettier philosophical community can present a plethora of cases wherein a person has a justified true belief but fails to attain knowledge. Robert K. Shope summarizes one of Gettier’s own examples:
Brown in Barcelona: S has strong evidence for a proposition, which S does not realize is false, namely, that F2: ‘Jones owns a Ford.’ S picks at random a city name, ‘Barcelona,’ and recognizes that the proposition that F2 entails that P2: ‘Either Jones own a Ford or Brown is in Barcelona.” Not having any idea of Brown’s whereabouts, S proceeds to accept that P2 on the grounds of the proposition that F2.
As it turns out, Brown is indeed in Barcelona; and, therefore, P2 is true. In the above case, S meets all the conditions of knowledge set forth by the JTB thesis: S believes that P2, that P2 is true, and S has evidence supporting her belief that P2. Most people, however, intuitively recognize that S does not know that P2 in the case of Brown in Barcelona. Justification simpliciter is then insufficient as a warrant-condition for knowledge. Some justificationists have responded to the Gettier-problem by adding a fourth condition, like indefeasibility. It seems, however, that both the traditional and modified versions of the JTB analysis of knowledge fail to adequately capture the nature of warrant, and, subsequently, of warranted true belief (i.e., knowledge). According to the traditional and modified JTB analyses of knowledge, one must have cognitive access to the evidence for one’s belief, the ability to mentally reflect on the ground for one’s justified belief, in order to have knowledge. As a result, the JTB analyses of knowledge are too restrictive; they exclude from the circle of knowers those who do not have the ability to reflect on the evidence for the beliefs they hold. Consider the following case:
Robert the Laborer: Robert is a hard-working construction laborer. He works from 7:00 a.m. until 3:30 p.m. five days—sometimes six—a week. One particular Wednesday, he formulates the belief that P3: “I should pick up flowers for my wife on my way home.” He and his wife had a spat the night before, and flowers are a tried and true way to communicate to his wife his enduring love. Later that same Wednesday, Robert remembers that P3. He is particularly exhausted, however, and affected by the heat of the sun. Such trauma disables Robert from identifying (the reasons or evidence for) why he believes that P3; he does not remember the spat with his wife or any of the other premises from which he inferred that P3 earlier.
In the case of Robert the Laborer, Robert clearly knows that P3, even though he has no cognitive access to the evidence for his belief that P3. Our rational intuitions recognize that Robert knows that P3; common sense dictates, therefore, that both the traditional and modified JTB analyses of knowledge—as well as any other analysis of knowledge that requires significant cognitive access—are overly restrictive.
To use contemporary jargon, our basic intuitions about Robert the Laborer evince that internalism is misguided. Plantinga describes epistemic internalism as the idea that
what determines whether a belief is warranted for a person are factors or states in some sense internal to that person. . . . Warrant and the properties that confer it are internal in that they are states or conditions of which the cognizer is or can be aware; they are states of which he has or can easily have knowledge; they are states or properties to which he has cognitive or epistemic access.
The alternative to internalism is, unsurprisingly, externalism—i.e., the idea that “warrant need not depend upon factors relevantly internal to the cognizer; warrant depends or supervenes upon properties to some of which [sic] the cognizer may have no special access, even no epistemic access at all.” Plantinga’s own proper functionalism offers an externalist account of warrant. According to Plantinga, epistemic warrant obtains when the following four conditions are met by (in?) a person, S:
- The relevant cognitive faculties are functioning properly
- The relevant cognitive faculties are aimed at truth (i.e. the production and/or sustenance of true beliefs)
- The relevant cognitive faculties are reliable with respect to truth
- The operation of the relevant faculties that results in S’s believing p is in a suitable cognitive environment
According to the proper functionalist, epistemic chance/luck, detrimental to warranted belief and knowledge, arises when an individual forms a belief with improperly functioning cognitive faculties (or by a faculty not aimed at truth, or by an unreliable cognitive faculty, or in an inappropriate cognitive environment, or any combination of the four). Conversely, epistemic chance/luck is eliminated by the proper functioning of our cognitive faculties (and their truth-aim, etc.). The proper functionalist is in a much better position than the justificationist to make sense of our intuitions about Robert the Laborer because proper functionalism does not require a condition of cognitive access for warrant, i.e. it is externalistic.
There is a nagging issue, however, for the proper functionalist which springs from cases like Robert the Laborer. The heat of the day caused Robert’s memory to malfunction, disabling him from accessing the reasons why he believes that P3, and this malfunction happens at the same time that Robert remembers that P3. The cognitive faculty which is sustaining Robert’s belief that P3 is the same cognitive faculty that is not properly functioning with respect to Robert’s beliefs which evidentially support that P3: memory. There is a problem here for the proper functionalist to specify how broadly or locally the concept of proper function must apply to a cognitive faculty for it to be thought to produce warranted beliefs. Must there be no malfunction at all or no malfunction in some more specific aspect of a faculty for it to be considered “properly functioning” and produce and sustain warranted beliefs? This is a form of the Generality Problem for proper functionalism. One could hold that a cognitive faculty must function properly in all its aspects in order to confer warrant on any one belief, or one could hold that the relevant cognitive faculty must function properly in only one of its aspects in order to confer warrant on all its beliefs, or one could fall somewhere in between these two poles. The proper functionalist account could run rough shod over our intuitions about whether Robert the Laborer knows that P3 or it could conform to them, depending on how one specifies the relevant facet(s) of a cognitive faculty which must be properly functioning in order to confer warrant. Proper functionalism as such does not clearly explain our intuitions about Robert the Laborer. The Generality Problem encourages a search for an alternative externalistic account of warrant.
II. The Rationale and Benefits of a Causal Account
Causal theories of warrant gain their impetus from reflection upon Gettier-type cases. Consider, once more, the case of Brown in Barcelona: why exactly is it that S does not know that P2? The causal theorist sees it as immediately intuitive that S lacks knowledge because a falsehood—namely, that F1—caused S to believe that P2, and not the truth—namely, that P4: Brown is in Barcelona. What eliminates epistemic chance/luck, according to the causal theorist, is an appropriate causal connection between S’s true belief and the fact that makes S’s belief true. Had S known, by the testimony of Brown, that P4 and then seen that the truth of that P4 entailed the conjunction that P2, then S would have known that P2. More specifically, S’s belief that P2 would have been warranted if the fact of Brown’s being in Barcelona caused, through Brown’s perceptual faculties (let’s say he saw some signs and Barcelonans outside the airport), Brown’s belief that P4, and, through S’s faculty of credulity (upon hearing Brown’s testimony over the phone), caused S’s belief that P4, and, through a process of inference, caused S’s belief that P2.  A causal account of warrant provides an altogether satisfactory explanation of Brown in Barcelona.
A causal theory of warrant also satisfactorily explains our intuitions about Robert the Laborer. Robert’s belief, gained from the testimony of the Bible, that P5: “I should love my wife,” his belief, gained from past experience of fights with his wife and making up with her by buying flowers, that P6: “Buying my wife flowers is a good way to love her after a fight,” his belief, gained from his experience of a fight with his wife the night before, that P7: “My wife and I had a fight last night,” his derived belief (inferred from that P5, that P6, and that P7) that P3, and then Robert’s memory of that P3 is the appropriate causal chain which warranted his belief that P3. The integrity of this warrant-conferring causal connection between the facts that make that P3 true and Robert’s belief that P3 remains, even though Robert’s memorial faculty is malfunctioning with respect to the evidence for that P3 and he is disabled from having cognitive access to such evidence. This causal account of warrant, unlike the JTB thesis, requires no condition of cognitive access, and, unlike proper functionalism, satisfactorily explains why Robert has the warranted memorial belief that P3 at the same time his memory is malfunctioning with respect to the evidential beliefs which support that P3—it is externalistic and is not subject to the Generality Problem.
It is easy to see the appeal of the JTB and proper functionalist accounts of warrant. Even though neither justification nor proper function is warrant, they are both valuable and closely associated with warrant. Justification often involves appealing to the causal chains which confer warrant on our beliefs as evidence that our beliefs are grounded in reality. For example, S may justify S’s simple perceptual belief that P8: “The Mountain is high” by stating, “I see the height of the Mountain.” When unpacked in the way I have suggested, this statement is equivalent to, “The height of the Mountain caused me to believe that P8 through my faculty of sight.” Justification does not, however, imply that our beliefs are warranted. Often, we overestimate the status of our beliefs, mistakenly thinking they are warranted as the result of a legitimate causal chain. In such cases, the causal chains we take to have rationally secured our beliefs in the facts, and which we present as justification for our beliefs, are in some way deficient and do not allow for warrant, even though they may provide ample justification. Other times, we have knowledge (i.e. warranted true belief) and yet we, like Robert the Laborer, are unaware of the causal chain which warranted our beliefs. Nevertheless, when we justify our beliefs we do indeed draw a causal connection from our beliefs to the facts we take to be true. Justification can therefore be easily confused with warrant because both deal with causal chains and our beliefs, although in significantly different ways.
Proper function can also be easily misidentified with warrant. This is because although warrant and proper function are not identical, our cognitive faculties, which are a key focus of proper functionalism, are significantly involved in warrant-causation. The causal connection which provides for warranted belief always involves our cognitive faculties; it is through our cognitive faculties that warrant-causation takes place. In order for the height of the Mountain to cause my belief that P8, someone must be able to see how high the Mountain is; in order for the relevant facts to cause Robert the Laborer to believe that P3 he must have memory, reasoning, moral, and perceptional faculties. Even though proper function is not warrant, warrant-causation does involve our cognitive faculties, and thus has much in common with a central concept to the whole proper functionalist project.
III. A Response to the Two Primary Objections to Causal Accounts
The strength of causal analyses of knowledge is their ability to respond to Gettier-type cases in a way that has intuitive appeal while at the same time incorporating the strengths of other analyses of knowledge, like the JTB and proper functionalist analyses. There are, however, two supposed problems for causal analyses which must be addressed: (1) their supposed inability to account for our a priori knowledge; and (2) Gettier-type counter-examples where a subject, S, supposedly has an appropriately caused true belief but fails to attain knowledge. I will address these two objections in order.
Audi articulates the first objection as follows: “How might what underlies the truth that if one tree is taller than another then the second is shorter than the first be causally connected with my believing this truth?” In other words, how can an abstract object, like a law of logic, be involved in a causal relation? Audi’s objection is that a causal analysis of knowledge cannot account for our a priori knowledge, because much (if not all) of our a priori knowledge is of abstract objects which are most likely incapable of causal relationships; a causal analysis of knowledge is subsequently too narrow. The burden of proof, however, is on the objector to identify why we should consider abstract objects causally inert. Why cannot abstract objects appropriately cause us to believe that they are the case through rational intuition? The answer is at least not immediately obvious. The first objection against a causal account of warrant needs further elaboration and is therefore inconclusive.
Epistemologists have used a number of examples to show that a causal theory of knowledge is too broad. We will only address the following “Barn Facades” thought experiment used by Goldman as an objection to his own earlier causal understanding of warrant:
Henry is driving in the countryside with his son. For the boy’s edification Henry identifies various objects on the landscape as they come into view. “That’s a cow,” says Henry, “That’s a tractor,” “That’s a silo,” “That’s a barn,” etc. Henry has no doubt about the identity of these objects; in particular, he has no doubt that the last-mentioned object is a barn, which indeed it is.
Goldman points out that most people would affirm that Henry does, in the above case, know that P10: “That’s a barn.” Goldman continues:
Suppose we were told that, unknown to Henry, the district he has just entered is full of papier-mâché facsimiles of barns. These facsimiles look from the road exactly like barns, but are really just facades, without back walls or interiors, quite incapable of being used as barns. They are so cleverly constructed that travelers invariably mistake them for barns.
Even though Henry correctly identified a real barn on the landscape of the countryside, Goldman states that since this district is filled with numerous other barn facades, “we would be strongly inclined to withdraw the claim that Henry knows the object is a barn.” Is it true, however, that Goldman’s earlier “causal analysis cannot handle this problem”?
The causal theorist might implement some of the insights of proper functionalism and hold that a cognitive environment for which our cognitive faculties are not attuned dissolves the causal connection between the fact p and S’s belief that p. The causal theorist might then hold with Plantinga that Henry’s perceptual belief is unwarranted because a district filled with barn facades constitutes an inappropriate cognitive environment for his cognitive faculty of sight. This is certainly a live option for the causal theorist. I am inclined to think (contra the intuitions of most contemporary epistemologists) that Henry does know that P10, even though the district in which he perceived the barn is filled with barn facades. I am not at all convinced that Henry’s cognitive environment somehow ruptured the causal connection between his belief that P10 and the barn. Clearly, Henry’s belief that P10 was caused by the barn through his faculty of sight; that P10 is warranted for him. At least two possible escape routes from the Barn Facades counter-example are subsequently available to the causal theorist: 1) affirm that Henry does not know that P10 and make use of the proper functionalist’s idea of “cognitive environment” to give a causal explanation of why not; and 2) affirm that Henry knows that P10 and appeal to the obvious causal connection between the barn and Henry’s belief that P10. Both 1) and 2) seem initially plausible, although I think 2), even if controversial, is the better option more likely to illuminate the causal nature of epistemic warrant. Regardless, there are at least two legitimate responses to the Barn Facades counter-example available to the causal theorist. Thus, the second objection against a causal account of warrant is inconclusive.
In the present paper, I have set forth the basic issues plaguing those attempting to provide an analysis of knowledge. In doing so, I examined the contours of two proposed accounts of epistemic warrant: justificationism and proper functionalism. I have shown how both accounts have difficulties which a causal understanding of warrant lacks. Additionally, I have laid out how a causal theory explains the appeal of both the justificationist and proper functionalist accounts, and how a causal theory can integrate their insights. Finally, I have defended a causal theory of warrant from its two most common objections: the first that it is too narrow and the second that it is too broad. I have found that a causal theory of warrant has promise because of its immediate intuitive appeal, flexibility, and lack of vulnerability to the problems of the justificationist and proper functionalist accounts of warrant.
 Paul K. Moser, “Epistemology,” in Philosophy of Meaning, Knowledge and Value in the Twentieth Century, ed. John V. Canfield (New York: Routledge, 2004), 197.
 In contemporary philosophical parlance, for something to obtain means for it to be the case.
 Alvin Plantinga, Warrant: The Current Debate (New York: Oxford University Press, 1993), vi.
 Edmund Gettier, “Is Justified True Belief Knowledge?,” Analysis 23, no. 96 (1963): 121-123.
 Alvin I. Goldman, “A Causal Theory of Knowing,” The Journal of Philosophy 64, no. 12 (1967): 357-372.
 Ibid., 169. Emphasis in original.
 Michael Huemer helpfully notes that “Goldman’s causal theory is most intuitive for the case of perceptual knowledge. . . . [But] Goldman also allows more complicated sorts of ‘causal connections’ (note that being ‘causally connected’ to a belief does not simply mean causing the belief). He allows cases in which the fact that makes my belief true causes evidence for the belief to be present, and that evidence, in conjunction with background knowledge I already have, causes my belief. He also allows cases in which my belief that p and the fact that p both have a common cause, as cases of there being an appropriate ‘causal connection’ (this is needed in order to secure knowledge based on induction)” (“The Analysis of ‘Knowledge’,” in Epistemology Contemporary Readings, ed. Michael Huemer [New York: Routledge, 2002], 437. Emphasis in original).
 Plantinga, The Current Debate, vi.
 Robert Audi, Epistemology: A Contemporary Introduction to the Theory of Knowledge (3rd Edition; New York: Routledge, 2011), 249.
 Moser, “Epistemology,” 234. See also Alvin Plantinga, “Respondeo,” in Warrant in Contemporary Epistemology: Essays in Honor of Plantinga’s Theory of Knowledge, ed. Jonathan Kvanvig (New York: Rowman & Littlefield Publishers, 1996), 309.
 Robert Shope, “Conditions and Analysis of Knowing,” in The Oxford Handbook of Epistemology, ed. Paul K. Moser (New York: Oxford University Press, 2002), 30.
 See Keith Lehrer and Thomas Paxson, “Knowledge: Undefeated Justified Belief,” in Epistemology: Contemporary Readings, ed. Michael Huemer (New York: Routledge, 2002): 464-474.
 In Plantinga’s words: “[According to the JTB analysis of knowledge] there is an internalist component to justification….the believer must have cognitive access to something important lurking in the neighborhood—whether or not he is justified, for example, or to the grounds of his justification (that by virtue of which he is justified…), or to the connection between those grounds and the justified belief” (Plantinga, The Current Debate, 10).
 Coincidentally, I developed this test-case independently of the similar example which Goldman drew from Gilbert Harman (Goldman, “Causal Theory,” 307).
 Plantinga, The Current Debate, 5. Emphasis in original.
 Ibid., 6.
 John C. Wingard, Jr., “Analysis of Knowledge,” (2014), 10. See Alvin Plantinga, Warrant and Proper Function (New York: Oxford University Press, 1993), 3-20.
 See John Wingard, Jr., “Reliability in Plantinga’s Account of Epistemic Warrant,” Principia: An International Journal of Epistemology 6, no. 2 (2010): 249-278 for a discussion of the Generality Problem as it applies to proper functionalism’s reliability condition.
 Wherever the proper functionalist may take her stand, such specifications run the risk of serious arbitrariness.
 A similar appropriate causal chain from fact, through our cognitive faculties, to belief would have also sufficed for warrant to supervene.
 This causal account escapes the Generality Problem of proper functionalism because it casts “proper function” in terms of whether a cognitive faculty causes a true belief in a way that links that belief with the fact that makes it true. Thus, the causal theorist can specify a non-arbitrary way Robert’s belief that P3 is both warranted and the product of a properly functioning cognitive faculty (i.e. memory) while Robert’s memory is malfunctioning with respect to the reasons why he believes that P3.
 Audi points out that “justified belief is important for knowledge because at least the typical things we know we also justifiedly believe on the same basis that grounds our knowing them” (Audi, Epistemology, 4).
 Consider Audi’s case of the Lifelike Photo of Jane: “Suppose that when I first visit the Smiths I have no idea that they have a photographic collection which includes very realistic, life-size pictures of themselves. When I approach the door to their living room I see, just twelve feet before me, and constituting all I can see through the doorway, a life-sized picture of Jane, standing facing me and smiling like the good hostess she is, with the background looking just like the living room’s rear wall. I say ‘hello’ before I get close enough to realize that I see only a photograph of her. I discover that the picture is so lifelike that this happens to everyone who knows Jane and enters unaware of the photograph. . . . As it happens, however, Jane is standing opposite me—in the next room, right behind the wall on which the picture is hung” (Robert Audi, Belief, Justification, and Knowledge [Belmont: Wadsworth, 1988], 103-104. Emphasis in original). Should a subject, S, in a situation similar to the one above formulate the belief that P9 (“Jane is standing across from me”) and have cognitive access to the pseudo-causal chain from what S took to be Jane through S’s cognitive faculty of perception to S’s belief that Jane is standing across from S, S would have justification for S’s belief and yet S would not have knowledge, only unwarranted true belief (because S’s true belief that P9 was not caused by the fact P9).
 Space does not permit the lengthy discussion required to map out what changes would need to be made to proper functionalism in order to accommodate a causal account of warrant. Suffice it to say, as above note 20 indicates, proper functionalism should not be incorporated into a causal account wholesale; some changes must be made to reorient the idea of proper function around the concept of appropriate causation.
 Goldman recognized early on the necessity of our cognitive faculties for warrant when he used the qualifier “appropriate” to signify that “knowledge-producing causal processes” like perception, memory, and inference are a required aspect of a warrant-conferring causal chain (Goldman, “Causal Theory,” 369-370).
 I may see the Mountain, or someone else may see the Mountain and relay its height to me by testimony—either way, under normal circumstances, someone must see the height of the Mountain for me to have knowledge that it is high.
 Audi, Epistemology, 255. Audi goes on, “This truth is not (in general) perceptually known, nor is its status dependent on any particular object in the world, as is the case with the (empirical) knowledge to which the causal theory best applies” (Ibid).
 Notice that the type of causation needed for warrant is doxastic causation, not material causation. It seems fairly clear that abstract objects cannot cause physical objects to move, but it is not at all equally clear that abstract objects cannot cause us to believe them.
 In, Warrant and Proper Function, Plantinga asks the question, “Why think propositions, properties, sets, states of affairs, and their like [i.e. abstract objects] cannot stand in causal relations?” (120); after finding no substantial reason, he concludes: “It is therefore quite possible to think of abstract objects as capable of standing in causal relations, and in causal relations with us” (121). Originally, Goldman was also inclined to think that the (abstract) process of inference is a causal process (“Causal Theory,” 362). The strong analogy between logical and mathematical intuition and sense perception additionally supports the idea that abstract objects can cause beliefs (or otherwise be causally related to beliefs) like physical objects can (see Octavio Bueno, “Logical and Mathematical Knowledge,” in The Routledge Companion to Epistemology, ed. Sven Bernecker and Duncan Pritchard [New York: Routledge, 2014], 360-362 and William G. Lycan, “Epistemology and The Role of Intuition,” in The Routledge Companion to Epistemology, ed. Sven Bernecker and Duncan Pritchard [New York: Routledge, 2014], 813).
 As with all topics in metaphysics, there is much debate and little agreement about the nature of abstract objects (see Peter van Inwagen, Metaphysics [Boulder: Westview Press, 1993], 13-14). Therefore, the causal theorist has grounds for pushing back against the objection that abstract objects are causally inert.
 I think it likely that the following response to Goldman’s counter-example will also apply quite well to other counter-examples (like that presented by Audi in Epistemology, 255-256) with some tweaking.
 Alvin I. Goldman, “Discrimination and Perceptual Knowledge,” The Journal of Philosophy 73, no. 20 (1976), 772.
 Ibid., 773.
 Ibid. Emphasis in original.
 To use Plantinga’s proper functionalist language more precisely, a district filled with Barn facades constitutes an inappropriate mini-environment (Plantinga, “Respondeo,” 313-317).