Introduction

What follows is an assessment of A.D. Smith’s direct realism (henceforth DR) as it is presented in The Problem of Perception (Smith). In this paper, I will argue that DR is unfeasible because it is incoherent. This position will be presented in the following way. In the initial part of the paper, I will develop the following premises of my argument: (P1) To adopt a realist view involves adherence to certain metaphysical and epistemic theses. (P2) A coherent account of realism must ensure that its metaphysical claims are consistent with its epistemic claims. This section is followed by an exposition and analysis of DR’s primary metaphysical and epistemic assumptions as they can be derived from Smith’s refutation of the argument from illusion. In this section, I will develop the final premise of my argument which states that (P3) DR’s epistemic claims are inconsistent with its metaphysical claims. Finally, the paper ends with some observations on the perils of adopting direct realism when we attempt to understand perceptual knowledge.

At the onset, it should be noted that there are good reasons for developing the aforementioned argument. At face value, it analyzes to what extent how we know the world and what we can know about it can be based on our common sense. DR supports our common sense view of the world in the sense that Smith maintains that we have direct access to the purely physical stratum of the world and the purely physical objects within it (Smith). In a way, DR can be seen as epitomizing the view that philosophical inquiry is an extension of our common sense. Common sense dictates that we are capable of directly perceiving the objects in our environment. At the same time, it also dictates that we are capable of having knowledge of our surroundings. We are then confronted with the question as to what extent a philosophical theory of perception can uphold our common sense view of our perceptual capacities and their ability to help us discern and navigate our environment. On a more fundamental level, this inquiry deserves further discussion since it is crucial to determine how we can formulate an adequate theory of perception without falling prey to the errors of existing theories. After all, as Smith himself claims, an adequate theory of perception is necessary for it is through perception that we have our initial access to the world. On a different note, the argument presented in this paper also merits exposition for most of the criticisms against DR immediately focus on how it fails to give a sound solution to the problem of illusion (e.g. Fish, and Meadows). This paper, on the other hand, aims to assess DR in terms of its foundational assumptions. In a sense, this paper aims to uphold philosophy’s focus on clarity and consistency as it shows how crucial one’s assumptions are to the cogency of the position one wishes to uphold.

The Metaphysical and Epistemic Theses of Realism

In this section, I will develop the first two premises of my argument. Recall that they are as follows: (P1) To adopt a realist view involves adherence to certain metaphysical and epistemic theses. (P2) A coherent account of realism must ensure that its metaphysical claims are consistent with its epistemic claims.

Let us begin by laying down the two primary metaphysical theses that are either denied or affirmed by the different versions of realism and antirealism in philosophy. They are as follows: the existence thesis (henceforth ET) and the independence thesis (henceforth IT) (^{Brock and Mares}; ^Devitt). ET claims that entities or facts exists outside of a domain whereas IT states that the existence of these facts or entities is objective and mind-independent (^{Brock and Mares}). A realist about the external world can thereby adopt both ET and IT while at the same time deny both when it comes to the objects in another domain (e.g. mathematical objects) (e.g. ^Leng). On a different note, he may also choose to adopt ET and IT when it comes to the objects of two different domains (e.g. the domains of science and mathematics) (e.g. ^Colyvan). One may even be more specific by stating that one only adopts ET and IT when it comes to some of the objects in a domain. For instance, ET and IT may be adopted only for the observable objects in the external world and not for its non-observable objects (e.g. neutrinos) (e.g. ^Leng).

Apart from these metaphysical theses, the various forms of realism and anti-realism in philosophy can also be characterized by their adherence or non-adherence to the following epistemic theses: the confidence thesis (henceforth CT) and the insecurity thesis (henceforth ST). CT states that “(a)lthough it may in some circumstances be difficult, we are always capable of coming to know about the existence and nature of the domain we are realists about” (^{Brock and Mares} 5-6). ST, on the other hand, claims that since we can be ignorant or erroneous about the domain we are realists about, there is no guarantee that we can have knowledge of a domain regardless if we make appropriate contact with it (^{Brock and Mares}). At face value, there seems to be a conflict with CT and ST. For instance, a negative formulation of their conjunction leads to the view that since we have no guarantee that we can have knowledge about the domain that we are realists about then we are incapable of knowing the existence and nature of the objects in that domain and the domain itself. In effect ST negates CT. This leads to extreme skepticism. However, a closer inspection shows that ST and CT can support each other. A positive formulation of their conjunction leads to the view that even if we are fallible, we are still capable of knowing the objects in the domain that we are realists about as well as the domain itself. Brock and Mares refers to this later formulation as the confident but insecure thesis (henceforth CST).

As we can see, both the metaphysical and epistemic theses adopted by a theory not only provides its presuppositions. It also supplies us with the general attitude it adopts regarding our capacity to know a domain. In addition, it also hints, and one may even say dictates the methodology that a theory will adopt. Since the domain that we are concerned with is that which gives descriptions of the external world, let us presuppose that all the formulations of the realist theories below will adopt an empiricist methodology. Such a presupposition is warranted since an empiricist methodology coincides with both ET and IT. It should be clear that at this juncture I will only adopt a very broad description of an empiricist methodology. It is to be understood as a methodology that harnesses our perceptual capacities to gain knowledge about the external world. Such a broad characterization is warranted since, as Prinz notes, regardless of the different versions of empiricism at hand (i.e. epistemological empiricism, concept empiricism, and semantic empiricism), they remain similar as they maintain that perception gives us our primary access to the external world and the objects within it.

At this point, let us consider the repercussions of adopting ET and IT in conjunction with CT, ST, and CST respectively. Doing so will allow us to see whether there are any inconsistencies in adopting a combination of these theses. Let us begin with a theory T ₁ which adheres to ET, IT, and CT. T ₁ , in this case, claims that the existence of the objects in a domain are mind-independent. Also, given that we are capable of accessing this domain, we can have knowledge of it. All seems well and good at this point. A problem however arises when we see that their conjunction leads to the view that so long as an object is unknowable then it does not exist in the domain that we are realists about. However, it is not the case that so long as an object is unknowable then it does not exist in the domain that we are realists about. Consequently, CT is false. In effect, CT is inconsistent with ET and IT. Consequently, T ₁ is false.

A seemingly good counter-example to the abovementioned argument may point out cases where we came to have knowledge of what was previously unknown to us (e.g. existence of Neptune, creation of black holes). However, such examples are misleading for they fail to capture the sense in which we are using the term unknowable earlier. To elaborate, the unknowability of the existence of Neptune as well as the creation of black holes at one time is due to the temporary limitation of the knower. At one point, he did not have the means to observe the planet and the phenomena, yet it is possible for him to perceive it with the later development of an instrument. That is, both the object and the phenomena can be accessed by the subject at a later time since what was only needed was a magnification of his existing perceptual capacity. It would be inaccurate, in this sense, to use these cases as counter-examples to the argument for the inconsistency of T ₁ for the counter-argument emphasizes that the combination of the three theses forces us to accept that there are objects in the domain that cannot exist so long as we cannot access them. For instance, it forces us to accept that the brute external world does not exist since it is unknowable to us.

A similar problem arises for ST when it is combined with ET and IT. Consider a theory T ² which maintains that a domain has objects that have a mind-independent existence yet what we claim to know about it, regardless if we make appropriate contact with it, is uncertain since we are fallible. T ₂ is also prone to a problem for it is forced to maintain that so long as something is known with certainty in a domain then it is false. Yet, there are some things that can be known with certainty in a domain. Consequently, T ₂ is false. The problem really arises in ST itself for it will only hold so long as we maintain that we are certain that we are fallible.

At this point, the remaining option available for a realist position given the metaphysical and epistemic theses we have mentioned earlier is the conjunction of CST, ET, and IT. Let us then consider a theory T ₃ which claims that although we are fallible, we can still have access and knowledge about the objects and facts in the mind-independent domain that we are realists about. At first glance, it seems that we have arrived at a tenable position. After all, T ₃ is a very modest position. It recognizes our fallibility while at the same time acknowledging that we can know objects and facts with certainty. Yet, T ₃ is also untenable for the conjunction of IT and CST leads to a contradiction. CST emphasizes the role of our minds in how we access and come to know the objects and facts in a domain. By doing so, it leads to the view that what we know about a domain is dependent on the proper relationship between our mind and the objects within it. If such is the case, then the objects in a domain are mind-dependent. This is contrary to what IT states. That is, the objects and/or facts in a domain are mind-independent.

At this juncture, I have shown (P1) and (P2) of my argument. In doing so, I have laid down the basic metaphysical and epistemic theses adopted by any conceivable form of realism about a domain. In addition, I have also shown why it is crucial to ensure the consistency of one’s metaphysical and epistemic assumptions. Simply put, an inconsistency leads to a theory’s incoherence. Let us now proceed to an assessment of a realist theory that adheres to the main assumptions of T ₃ .

On the Metaphysical and Epistemic Assumptions of Smith’s DR

In this section, I will develop (P3) of my argument. Recall that it takes the following form: DR’s epistemic claims are inconsistent with its metaphysical claims. The discussion of (P3) allows us to do the following. First, it allows us to address a possible criticism to what may be mistakenly inferred as a generalization made in the previous section. That is, any realist theory which follows the pattern of T ₁ -T ₃ is incoherent. It should be clear that I am merely saying that if one simply adopts T ₁ -T ₃ as is then one’s theory is incoherent. As such, I am open to the fact that theories can make qualifications. Second, developing (P3) will also enable us to see an instance where a theory remains incoherent even if it has made qualifications to the assumptions of T ₃ . Finally, the discussion of (P3) will allow us to see both the appeals and perils of DR’s claims especially when they are applied to the problem of illusion (henceforth PI) in philosophy.

Let us begin by laying down the assumptions of DR. As I have noted earlier, DR is a form of T ₃ . This is evident even in Smith’s initial description of DR in The Problem of Perception where he states:

(T)he philosophical position known as “Direct Realism” ... holds that the physical world has an existence that is not in any way dependent upon its being “cognized” ... The physical world is not ... dependent on “consciousness” ... It is a position that states that) there is a purely physical aspect or stratum to a world. Direct Realism is the claim that we can be directly aware of objects that themselves possess such a stratum. (^Smith 1-3)

We can immediately see ET and IT above. That DR adopts CST, on the other hand, can be seen in the fact that it recognizes that we are prone to misperception, yet we are still capable of having veridical perceptual experiences. That is, we are still capable of having experiences that imply the truth of our belief about the content of our perceptual experiences. Such a recognition immediately shows the two components of CST: the confidence component and the insecurity component regarding our ability to know the objects in the external world.

Let us now proceed to the qualifications that DR makes to the conjunction of ET, IT, and CST. These qualifications can be seen in Smith’s account of a perceptual experience (henceforth PE). Let us begin by giving Smith’s description of a PE. For Smith, a subject S has a PE of a normal physical object¹O so long as (1) S is perceptually aware of O and (2) S’s perceptual apparatus (e.g. eyes, ears, etc.) makes intentional contact with O. In this case, a PE differs from mere sensation because S’s awareness of O has intentional content. He further claims that this intentional content is non-conceptual in character because what causes S to perceive O is not O but S’s perceptual apparatus whose function is dependent on S’s “distinctive and irreducible type of psychological state” (^Smith 185). In effect, S’s awareness of O arises from non-conceptual and non-sensory factors. These being: (a) phenomenal three-dimensional spatiality, (b) kinetic movement or constancy, and (c) the Anstoss or the non-sensory awareness of one’s self (^Smith, 2002). Based on these descriptions, a PE can be seen as being enabled and at the same time delimited by our psychological state. Since it is determined by our psychological state then it counts as a conscious state that has sensible qualities as its content. It is at this point that we can see the uniqueness of Smith’s position. By maintaining that a PE can only occur as a conscious state, he is able to claim that direct perception of O is possible because S’s experience of O remains constant since the intentional content of his experience of O will always be determined by one or more of the non-sensory factors that characterize perceptual awareness mentioned above. In effect, the novelty of Smith’s position lies on the primacy his DR attributes to the phenomenological aspect of PEs. Such a position is unique since the PI uses phenomenological grounds to show that it is difficult for S to differentiate veridical from illusory perception. Smith, in this case, uses the PI’s blade against itself.

To further understand Smith’s position, let us consider the case of Angus seeing a rabbit. Angus sees a rabbit because he is in an intentional state (i.e. he is aware of his environment and the objects within it). His intentional state also has content because he directed his eyes to the rabbit. It is his eyes then which causes him to see the rabbit and not the rabbit itself. Now, there is something distinct in Angus’ PE of seeing a rabbit because he may have merely seen the left side of the rabbit thereby giving him access to only one of the rabbit’s ears, the spots visible on that side of its body, and so on. However, this does not mean that what Angus can see of this rabbit is utterly delimited to its left side. It is possible for Angus to move on any side of this rabbit (i.e. any side of this rabbit allowed in the dimensions of space that they both occupy). It is also possible for him to redirect his eyes from one part of this rabbit to another (e.g. focus on the whiskers, the fur near its tail). By doing these, Angus is able to see his companion animal from different perspectives. The same effect is given if he chooses to literally take a few steps back from his rabbit. His rabbit looks smaller because he is viewing it from a different distance. Now, he can also lose sight of his companion animal as he moves his head or even simply his eyes towards the feeding bowl that the rabbit is approaching. Yet, despite of this, when he turns back his head or his eyes, he can see his rabbit once again. In other words, there are certain aspects of his experience that remain constant. It is these constant qualities of the rabbit that Angus can directly perceive. Now, it is also important to note that when Angus looks at his companion animal, there seems to be an intuitive distinction between himself and his rabbit. As such, there is an immediate separation between himself and the object of his perception. Smith claims that this “collision” between one’s body or sensory organs and the objects in the external world shows the perceiver’s non-sensory awareness of his self, the Anstoss (^Smith 155). Going back to our example, Angus is enacting his agency in the process of perceiving (i.e. especially if he is touching) his rabbit for it involves an implicit recognition that he is distinct from the object of his perception. That is, there is a space between them. This is not only the case when he perceives his companion animal. This remains to be the case regardless of what object he chooses to perceive in his environment.

Now that we have given Smith’s characterization of a PE, let us now proceed to show how it can be used to address the PI. The PI poses a qualitative identity problem as it maintains that there should be no discrepancy between the object of perception and the primary object of awareness. Its force lies not only in the pervasiveness of illusions in our daily lives but also in its use of the most fundamental law in identity theory, Leibniz’s Law. Leibniz’s Law takes the following form: “x=y if, and only if, x has every property which y has, and y has every property which x has” (^Tarski 50-3). The problem posed by PIs to DR maybe presented in the following way: (1) If we can directly perceive the objects in the external world then all our PEs of these objects must have every property which these objects possess. (2) It is not the case that all our PEs of these objects have the same properties that these objects possess as can be seen in cases of illusions. (3) We have difficulty distinguishing between cases of veridical perception and cases of illusions due to their phenomenological similarity. (4) Due to (3), we are incapable of directly perceiving objects even in cases of veridical perception. Consequently, we cannot directly perceive the purely physical objects in the external world.

It is important to note that Smith’s formulation of the PI gives emphasis to what he terms as the “sense-datum infection” (Smith 26). Simply put, the infection forces us to claim that in both cases of veridical and illusory perception, what we perceive is something mediate between the object and ourselves, that is sense-data. In the formulation of the PI above, Smith’s DR can be presented as targeting (3) as he notes that it is in fact the phenomenological aspect of PEs that shows us that we are capable of having veridical PEs. To reiterate what we have mentioned earlier, the non-sensory and non-conceptual features of PEs shows us either that an object remains constant regardless of how we perceive it or there is an immediate distinction between the perceiver and the object of his perception in the self-affirming act of perceiving an object. The sense-datum infection, in this case, is avoided since sense-data are relegated to mere sensations and mere sensations are not characterized by phenomenal three-dimensional spatiality and kinetic constancy. Nor do they give rise to the Anstoss. In other words, solely using sense-data to explain PEs is insufficient for they cannot explain the phenomenological character of PEs. In effect, Smith’s account of PEs still follows Leibniz’s Law for S’s PE of O still contains the relevant properties of O which are made available to S by the non-sensory component that affects how his perceptual organs access O.

Returning to the previous example of Angus seeing his companion animal, the fact that the Anstoss is at play in the act prevents it from being a perceptual illusion. In the case that it is a perceptual illusion (e.g. it only seems as if he is touching his rabbit due to some kind of illusion), the mere fact that his experience of his companion animal’s position, size, color, shape, etc. remains constant regardless of how and where they are both located in the room shows that he can still directly perceive his rabbit.

At this point, we have already fleshed out Smith’s DR. Let us now proceed to the assessment of whether his qualifications to T3 will not lead to a contradiction. Recall that the problem we posed to T3 earlier notes that it is inconsistent due to the conflict between IT and CST. The inconsistency arises because adopting CST leads to the view that what we know about a domain is dependent on the right relationship between our mind and the objects within it. In effect, the objects in a domain become mind-dependent. This contradicts IT’s metaphysical claim that the objects in a domain are mind-independent.

At first glance, Smith’s position seems to overcome this hurdle due to his usage and characterization of the perceptual constancies. He states:

The term “perceptual constancy” ... refer(s) to any veridical perceptual situation in which an unchanging physical feature of an object gives rise, because of its changing relation to the perceiver, to changing proximal stimulation at our sense-organs, while the perceived feature of the object appears unchanged ... (A)s with all genuine constancies, (it) is not a matter merely of “judging” ... (I)t is a matter of simple perceptual appearances... Such constancy is not the function of any “judgement” that supplements the deliverances of the senses; it characterizes what we are aware of in the most “basic” and “immediate sense.” (Smith 170-8)

As we can see, Smith seems to avoid the contradiction between CST and IT as he claims that our PEs are not mind-dependent in the sense that they do not require the use of concepts or their formation does not involve judgements. He may also further claim that even if our PEs are determined by the limitations of what we can be perceptual aware of, we can still have knowledge about the objects in the external world since we are aware of the constancy of these objects as well as their constant properties. In this account, the existence of the objects in the external world are not determined by our mind nor are they determined by our perceptual awareness. It is just the case that we are intrinsically capable of perceiving the objects in our environment.

As I see it, there is something amiss in such an account of PEs. In order for S to state that O continues to persist or that certain properties remain constant in O, S must make an inference to the best explanation (henceforth IBE) that O and/or its relevant properties remain constant. To explain further, let us look at Gilbert Harman’s characterization of IBE. He states:

In making this inference one infers, from the fact that a certain hypothesis would explain the evidence, to the truth of that hypothesis. In general, there will be several hypotheses which might explain the evidence, so one must be able to reject all such alternative hypotheses before one is warranted in making the inference. Thus one infers, from the premise that a given hypothesis would provide a “better” explanation for the evidence than would any other hypothesis, to the conclusion that the given hypothesis is true. (^Harman 89)

Applied to our recurring example, Angus’ PE of seeing his rabbit is a byproduct of an IBE. Such is the case since in order for him to claim that the position of his rabbit in space remains constant, he needs to compare this PE with his previous PEs. At the very least, he needs to compare it with at least one instance where an object’s position in space remained constant under similar conditions. The process need not be a conscious one in the sense that he is actively aware of his premises and the inference process involved in the formation of his perceptual belief. In this case, the fact that establishing the constancy of an object or its relevant properties involves an inference (i.e. a judgement) shows that it is still a mental act. In other words, it is still dependent on the mind.

At this point, one may emphasize Smith’s description of the constancies that we quoted above. That is, our phenomenological and physiological make-up allows us to become immediately aware of these constancies prior to judgment. It is at this point that Smith’s distaste towards the theories of perception that emphasize how it involves the relay of information in our sensory systems becomes detrimental to his own position. To further debunk Smith’s position, we only need to present an instance where a sensory system functions in such a way that shows that constancies are a byproduct of cognition.

Consider, for example, the case of the visual system. There is general agreement in the existing literature that its function can be understood in terms of high-level and low-level processes as well as in terms of top-bottom and bottom-up processes (e.g. ^Cavanagh and ^{Poltoratski et al.}). To see these processes at work, let us return to our well-loved example: Angus perceiving his rabbit. How Angus perceived his visual environment and the objects within it depends on the “interplay of (the) top-down guidance of spatial attention and (the) bottom-up processing of visually salient information” (Poltoratski 564). Now, the visual salience of an object refers to “the degree to which it can ‘grab’ attention… (It is a) subjective percept (which) is strongly influenced by the local feature differences that occur within the visual scene (^Poltoratski 564).

As we can see, even if we exclude the other top-down factors that affect visual perception (e.g. Angus wants to check if his rabbit is in the room because it is his favorite companion animal) in our description of Angus’ PE, we can still demonstrate that his PE is only possible because it involves a judgement. The constructed description that he initially derived from the retinal input that he received from his surroundings is dependent on previously stored descriptions of his environment. In other words, when Angus’ retina were hit by light waves that bounced on the surface of the objects in the room, he was only able to focus on one of the objects in the room (i.e. his rabbit) because at the higher level of his visual system he recognized the visually salient features that coincide with his prior existing representations of his rabbit. Given that it is not the visual system itself that stores prior representations of the objects in our surroundings then it follows that establishing the constancy of an object or its relevant properties cannot be performed by the visual system alone. In effect, Smith’s claim that we are immediately aware of this constancy does not hold for this constancy can only be established based on prior experience. This only supports what we mentioned earlier that to establish the constancy of an object and its relevant properties requires judgment for it involves an inference.

So far, we have shown that Smith’s version of T ₃ fails to solve the conflict between CST and IT. This is so because an analysis of Smith’s constancies based on his descriptions as well as a description of how these constancies are arrived at by the visual system shows that they are mind-dependent. In effect, I have developed (P3) of my argument. At this point then we have good reasons to hold that DR is incoherent.

Conclusion

I began this paper by noting that one of the appeals of DR lies in its closeness to our common sense view of how we experience the external world. It is not difficult to think of a scenario where a man in the street provides an account of his PEs that is close to DR’s descriptions. However, it is very seldom that our common sense is reliable. More so in the case of how we explain what we know and how we know the world.

We have already noted the problem with Smith’s position in the previous section. However, as I see it, there is a more fundamental problem with his view and this problem is also applicable to any possible version of DR. Smith seems to adopt a foundationalist position that is grounded on empiricism and the methodology that goes with it. It is a form of foundationalism because it is purely grounded on the assumption that there is an external world that has a mind-independent existence. As objects in the external world, we are immediately granted existence as per his foundational assumption. Our existence is affirmed because we are capable of perceiving the external world and the objects within it. In fact, he takes pride in showing that it is the phenomenological aspect of our perceptual awareness which grants us access to the external world. The problem here lies in the dependence of his empirical claims on his metaphysical claims. An empirical claim must be based on some form of generality in our experiences. There must be an accumulation of data before we can state that an empirical claim is justified. No such data can be given by a metaphysical claim. It is just a presumption that something is the case or that something exists. To put it bluntly, it seems that Smith is saying that I have a PE because I am just taking it to be the case that the world exists. He cannot resort to saying that his claims about PEs are consistent with a schema because he seems to be averse to the idea that a PE can be best understood within the context of a conceptual scheme (Smith 12). The perils of DR then lies not only in the inconsistency of its metaphysical and epistemic assumptions but also in its tendency to derive matters of fact from metaphysical claims.

Now, if a version of T ₃ that adopts Smith’s qualifications is untenable, we are left to ponder if there are available versions of T ₃ that may address the inconsistency between CST and IT. As we have shown in the second section of this paper, neither T ₁ , T ₂ , or T ₃ seem to offer a consistent account of realism about a domain yet amongst them, T ₃ seems to be most tenable position (i.e. provided that a cogent argument supports it) largely because it recognizes our fallibility and our capability to have knowledge about the external world. In addition, we currently have reliable empirical evidence that we can attain this kind of knowledge despite of our physical and mental constraints as can be seen in the progress in the physical sciences. What I am suggesting then is to adopt a version of T ₃ that follows the most feasible version of scientific realism. A Quinean route, for example, seems highly feasible.²

Within this context, perhaps we can still salvage T ₃ if we adopt a Quinean route. As per Quine’s view, our existence and truth claim about the external world should be made and assessed in the context of the most successful discipline that describes it at the moment (Quine). At this point, I would like to point out a common misinterpretation of Quinean naturalism: that it amounts to the “worship” of the physical sciences. To clarify that such is not the case, it is helpful to introduce the following analogy. Suppose we follow the practices of one our peers because in all accounts, he really is a virtuous individual. That is, his demeanor, decisions, and actions show that he is an exemplary individual. If we choose to follow his actions (e.g. adopt his methodology when we make decisions), this does not necessarily mean that we “worship” this individual. It also does not necessarily mean that we are now devoid of our agency and any criticism that we can make about his actions will be devoid of value. It merely shows that we recognize that it is good to be virtuous and he is an example of a virtuous individual. In a similar manner, adopting the natural sciences’ representation of the external world does not necessarily mean that we “worship” science. Rather, it shows that we value the goals of science (e.g. we value its aim to provide us with the most accurate explanations of physical phenomenon) and we recognize the reliability of the methodology that it adopts (e.g. we recognize the reliability and utility of the scientific method). To make further use of our analogy, it is still possible to make well-founded criticisms of science’s methodology, its results, and even its history since we can assess the internal standards of justification of the discipline itself and whether its theories adhere to these standards of justification (e.g. whether it makes existence claims that are incoherent with its standards of justification as is the case with mathematical objects). In other words, philosophy and science can co-exist happily and productively side by side as equals.

As a final note, if we attempt to salvage T ₃ by adopting the most feasible form of naturalism that can be available to us, it is interesting to note how this shows that both our scientific and philosophical accounts of the external world will still be closely knit with our common sense view of our environment. After all, both science and philosophy are byproducts of the growth and development of our common sense view of the world.