Introduction: Direct realism and the problems it faces.
The philosophical theory of perception most akin to our commonsense approach is known as direct or naive realism. This theory maintains that through perception we gain immediate, direct access to the physical mind independent world surrounding us, and therefore rejects the postulation of entities such as sense data  that act as intermediaries in the perceptual process. Throughout its life as a philosophical doctrine, direct realism has come under much scrutiny and has been accused of many inadequacies. Many have argued that simple reflection on our perceptual lives produces counterexamples to the direct realist's key tenets.
One such criticism is presented by Hume and concerns "perspectival variation"  , something we experience frequently in some form. Suppose you change your position in relation to an object by moving towards it; most would agree that your perception of the object in question changes. Hume uses the example of walking away from a table, claiming that as one does this "the table which we see seems to diminish"  . From this observation he concludes that we cannot engage directly with physical objects through perception, and instead what we perceive directly are "representations"  of those objects. Hume is therefore advancing a version of indirect realism, a rival theory which claims perception can be understood as a relation consisting of three parts; the perceiver, an intermediary and the object.
Another problem frequently posed towards the direct realist is known as the time lag argument  which runs as follows. Consider your perception of a star in the night sky. Due to the nature of the way light travels and the distance involved, the star you think you're perceiving will actually no longer exist. Critics of direct realism argue that because what you perceive here cannot be the same as the actual object of perception, direct perception in this case is simply not possible. The two things can be thought of as "numerically distinct"  . This particular case may seem somewhat farfetched, but critics argue the reasoning involved can be extended to ordinary everyday visual perceptions also. Whenever I visually perceive an object, the occurrence of the "reflection of light"  involved has always happened before "the light waves arrive at the retina"  . Again then, there is an element of time delay included within such perceptions, which some argue suggests that what we directly perceive cannot be the object from which the "physical process originates"  .
Whilst the above problems cause the direct realist some notable concerns, perhaps the most troubling perceptual scenarios pitted against them are those concerning illusion and hallucination. Many critics of direct realism claim such arguments alone to be sufficient to disprove it as a theory of perception. It is for this reason my project will focus upon the following two arguments.
Let us first consider the argument from illusion. Although we experience perceptual illusions with a fair amount of regularity, as Howard Robinson states we usually "treat them as inconsequential"  and carry on with our lives. Within our present philosophical debate however they are of key importance and so need to be clearly articulated. A perceptual illusion, A. D. Smith maintains, can be described as "a perceptual situation in which a physical object is actually perceived, but in which that object perceptually appears other than it really is, for whatever reason"  . A popular example often used within the philosophy of psychology is the Muller-Lyre illusion:
Here we are presented visually with two arrows that are actually identical in length. A very quick use of a ruler would establish this. When one looks at them however the lines seem to be two differing lengths. We therefore are experiencing an illusory perception, one which is not veridical in nature. How then do such cases supposedly force us to deeply question the nature of direct realism? After noting this element of our perceptual lives, critics argue that direct realism simply doesn't possess the necessary machinery to deal with them coherently.
The object which appears to us in an illusory case has certain properties which the true object does not, as is the nature of an illusion. In the Muller-Lyre example above the illusory object appears to posses the property of lines differing in length, although the true object does not have this property. Considering this, those who oppose direct realism argue that what appears to us directly through perception cannot be the same as the true object, as can be confirmed by use of the uncontroversial principle of Leibniz's Law. This holds that "no two objects have exactly the same qualities"  . If one agrees with this reasoning, it appears that the direct realist has great trouble in claiming we are immediately aware of the physical objects of perception in illusory cases.
Alternative theories of perception often use the above argument to illustrate how they supposedly can coherently account for illusory cases and should therefore be considered superior. Some, such as Howard Robinson, claim to do so by postulating sense data which can be described as the "mind- dependent objects we are directly aware of in perception"  . According to the sense data theorist, what we directly perceive in a visual perception is a sense datum that acts as an intermediary between the object and the perceiver. In the case of an illusion then, the sense data theorist claims we are aware of a sense datum which exhibits the illusory properties, rather than the actual object itself. With the illusion above it would be argued that what I am directly perceiving is a non physical sense data exhibiting the property of lines differing in length.
The argument from hallucination at first glance may seem relatively similar to that of illusion, but an important distinction must be made. A hallucination occurs when a person appears to have a true perception of an object in their environment, but in actual fact there is "no mind independent object of the relevant kind being perceived"  . Here we find the difference with illusory cases as then an object is present during the perception. Hallucinations do not happen as frequently as cases of illusion. They can be induced by the use of certain drugs such as LSD, for example. A well used example of a hallucination occurs in the play Macbeth  . Before a murder occurs Macbeth has the hallucination of visually perceiving a dagger. He experiences the same "visual sensation"  that would occur had there been a real dagger before him. Perceptual cases such as these are even more troubling for the direct realist. How is one to claim that through perception we have direct access to reality if there is no actual object involved in the perception? The argument from hallucination is used by the opposition to stimulate a further question. When we experience a hallucination, there appears to be no difference phenomenologicaly between the hallucinatory perception, and what we would perceive in the presence of an actual object. One might argue that, taking this into account, it is reasonable to propose that even in veridical cases of perception we are not immediately aware of the physical object  . It might be proposed that the true essence of perception cannot rely on the physical mind independent object in our environment, as we experience the same thing in hallucinatory cases  .
The naive realist is clearly put in a very difficult position by the perceptual occurrences I have outlined above. Considering this, is it possible to still formulate a coherent account of direct realism? My project here is to consider two responses to the arguments from illusion and hallucination respectively which claim such a task is indeed possible. In doing this I will arrive at a theory of perception known as disjunctivism. Through an analysis of this position I will establish whether or not it is able to offer an account of illusion and hallucination that does indeed preserve the essence of naive realism. Ultimately I will conclude that....
Bill Brewer's account of illusion.
One philosopher who is keen to retain a theory of perception that is direct and immediate in nature is Bill Brewer. In his recent paper "How to account for Illusion", Brewer offers a way of approaching the issues involved in illusion and hallucination which he hopes will enable him to achieve this. His motivation for providing such an account rests on his adherence to the principle that the "core subjective character of perceptual experience is to be given simply by citing the object presented in that experience"  . Brewer believes such an idea to be crucial to a coherent understanding of perception and argues it provides us with an "indispensible insight"  ; that the true nature of our perceptual experiences are "constituted"  or in a sense formed by the perceived object. The relation is therefore direct between the perceiver and the object concerned, and no intermediaries are involved. Brewer aims to show that indeed one need not appeal to sense data or intentional theories of perception to explain illusion as he thinks both of these to be inadequate and flawed. He maintains one should instead adopt an approach which he labels the "Object view".
The Object view which Brewer proposes consists of the following notions. Reinforcing the ideas contained within the principle mentioned above, Brewer states that we should think of our perceptual experiences as "a relation between a perceiving subject and the object presented"  . The immediacy of such processes is again being made clear; there is purposefully no mention of sense data or other such intermediaries. The notion which then follows is the key to understanding Brewer's proposed response to occurrences of illusion. He argues that when one looks at an object in differing ways and within differing circumstances, the object in question bears "visually relevant similarities"  to other objects.
In order to see how this suggestion applies to illusory cases, let us consider an example set out by Brewer. Imagine a white piece of chalk is lit up with red light in such a way that it appears red to the perceiver. This is an instance of an illusory perception. The Object view responds to the situation by stating that the "core subjective character"  of this experience, even though it involves illusion, is formed by the actual piece of chalk existing in our environment. Considering this though, how are we to then explain the fact that in this particular case the chalk appears red to the perceiver? Brewer here agrees that from the way in which the perceiver is visually engaging with the white piece of chalk, and noting the fact that a red light is involved in the perception, the chalk does indeed look red. Crucially though, he claims this is because the piece of chalk at that time has "visually relevant similarities with paradigm red objects"  . The perception therefore involves likenesses with examples of other objects that are actually red such as a post box, for example. Illusions then, Brewer proposes, can be explained as being cases of perception in which what one is directly experiencing has "similarities with paradigms of a kind of which it is not in fact an instance"  .The white piece of chalk is not itself an instance of a red object, but in the illusory perception described above it is sharing a similarity with an example of an actual red object. Importantly though the essence or core of the perceptual experience is still provided, and indeed formed Brewer maintains, by the actual object of your perception and this occurs in a direct relation.
Another example might be the Muller-Lyer illusion I looked at previously. Using the same method as before, Brewer's direct object approach would be to claim that the lines do indeed look to be differing lengths, but this is because the object of perception bears visual similarities to a "paradigm of inequality in length"  . Brewer therefore states it to be perfectly understandable why someone might assume through visual perception that one line is longer in length than the other. He does not think, however that this provides reason to postulate intermediaries within our perceptual processes. A satisfactory explanation, he believes, can be provided by the proposed object view which still embraces naive realist notions.
Within his paper Brewer does also mention cases of hallucination, but the detail within this discussion is very limited in comparison. Brewer agrees with other theories of perception in that when a person experiences a hallucination, there is no physical object present in that perception. Such instances, Brewer believes, should be described by presenting a "qualitative description of a mind independent scene"  and stating that the person experiencing the hallucination is unable to determine, through inward reflection only, whether or not the objects they are perceiving are constituting the perception. It is clear therefore that Brewer's main concern is illusory cases, so it is this element I shall concentrate on.
We have therefore seen how Brewer proposes to account for instances of perceptual illusion whilst still preserving the key elements of naive realism. My concern at this point is to establish whether or not Brewer is successful in providing the naive realist with the necessary tools with which to answer the argument from illusion. After considering Brewer's account, it actually seems that rather than producing a direct rebuttal which dissolves the arguments problematic conclusion, he is simply stipulating a view which disagrees with it.
Recall the argument from illusion as explained in my introduction. The argument states that during an illusion, one cannot be in a direct perceptual relation with the true object of perception, as what you are perceptually aware of possesses different properties to this object. It is not clear that Brewer's view actually responds directly to this line of argument. Brewer might argue he is trying to explain why we seemingly acknowledge a difference in properties between the actual object and what is perceived by making reference to the visually relevant similarities the direct object possesses in those circumstances, but this appears to be sidestepping the issue highlighted. Brewer does not seem to be able to avoid the critics notion that if two objects possess differing properties, as occurs within an illusion, they cannot be the same thing.
In order to coherently provide an answer to the proponent of the argument from illusion, and therefore ultimately preserve naive realism, I would argue that Brewer is required to do more than simply presenting his proposed object view. Brewer might retaliate to this criticism by stating that the object view he proposes has great appeal due to its remaining faithful to our common sense approach to perception, and retaining the idea that our perceptual experiences are constituted by the very objects of our perception. This may be true, but the question here is whether or not Brewer's view allows the naive realist who upholds such points to escape the worries presented by the argument from illusion. From analysing the object view he proposes, I would argue such a goal is not achieved, and therefore illusory perceptions still present a major concern to the naive realist. Brewer's view certainly does embrace the key elements of direct realism, but fails to eliminate the threat of the argument from illusion which ultimately therefore still undermines the position. I do not therefore believe that direct realism is "fully vindicated"  by the object view as Brewer claims it to be.
Michael Martin and the disjunctivist approach.
There are however other philosophers also keen to defend direct or naive realism from the objections raised through illusion and hallucination. Michael Martin is a key player in discussions on this topic and is motivated in a similar way to Brewer, although his theory offers a slightly different approach. Martin is keen to promote what he calls the "transparency of experience"  , something which he believes accurately describes the true nature of our perceptual experiences. Naive realism, Martin claims, is able to coherently encapsulate the idea that our perceptual experience involves a "phenomenal transparency"  between the perceiver and the object perceived. Not only is the relation direct between the perceiver and the object, but in a similar vein to Brewer, Martin proposes this view entails that the objects in question are "constituents of the experiential situation"  . The thought is that when, for example, I visually perceive the room in which I am currently writing, I am aware of all the different objects in my environment and these form or constitute that particular "sensuous experience"  . Martin acknowledges the proposed objections brought against defenders of the direct realist tradition involving illusion and hallucination but has a proposed strategy in place with which to respond to them. Martin argues these arguments are in fact flawed due to a mistaken assumption. This assumption involves the idea that cases of illusion, hallucination and veridical perception are all of a "common kind"  , that is, they all involve the same sort of "conscious experience"  . Martin argues strongly that we should reject this claim, and does so for the following reasons. In the case of a veridical perception, as we have seen, a naive realist claims that the objects of perception are "constituents"  of the perceptual experience. With hallucinations however, this simply cannot be the case, as by the nature of a hallucination there is no mind independent object actually being perceived. It is therefore impossible for anything to form that particular perceptual experience as would happen in veridical cases. From this we can see that hallucinations and veridical perceptions must therefore involve different "states of mind"  , Martin claims.
Martin proposes it is use of this "common kind assumption"  which allows critics of direct realism to make the "spreading step"  from hallucinatory cases to veridical cases. This is because if, as is the case with common kind theorists, you maintain all three perceptual categories involve the same phenomenal character or mental state, then a statement about hallucinations can be applied to veridical cases also. Martin hopes therefore, that in denying this assumption, he will be able to preserve the notion that in a veridical case we have an immediate relation to mind independent objects in our environment. This move by Martin enables him to argue that an explanation of what he calls the phenomenal character in a veridical case, and an explanation of the phenomenal character in a hallucinatory case are in fact different. The key point Martin wishes to stress is that when one experiences a veridical perception,that particular experience is of a "fundamental kind"  . If one were to experience a visual perception in which the object of one's perception was not present, such as occurs within a hallucination, then this experience cannot be of the same fundamental kind.
After an analysis of both of their accounts, it has become clear that both Martin and Brewer are supporters of a position known as disjunctivism. The key element which distinguishes this theory of perception from others is its rejection of the common kind assumption as is articulated by Martin above. Disjunctivists acknowledge that there are three different types of perceptual experience but argue that one should not maintain, as many theories of perception do, that they involve the same kind of mental state. The theory as endorsed by Martin can therefore be seen as "negative"  in a sense as it involves this particular rejection. The aim of the disjunctivist account is to highlight the importance of the veridical perceptual case, and support a "positive"  explanatory account of it which retains the naive realist elements we looked at previously. It is important at this point to note that the version of disjunctivism associated with Martin is of a specific kind. It is known within the relevant literature as phenomenal or reflective disjunctivism due to its focus upon the phenomenological aspects of perceptual processes and its points concerning introspection. There are other ways of formulating the disjunctivist programme including, for example, the epistemological disjunctivism championed by McDowell, and so care should be taken to make a distinction here. It should also be noted that Martin's focus is predominantly upon hallucination, so this will shape the following discussion.
Disjunctivism gains its name from the following way in which it attempts to deal with perceptual experiences. The disjunctivist claims that a perceptual experience is either veridical, that is it involves a mind independent object which is constitutive of that perceptual experience, or it is just a hallucination of that particular object. The second disjunct therefore involves an experience in which the "conscious character"  is not "derived" from the mind independent physical objects in the percipients environment. This particular way of encapsulating disjunctivism comes from one of its first notable supporters, J. M Hinton. Hinton explained perceptual disjunctivism through an analysis of the sentences that can be formed concerning perception. Consider the sentence "I seem to perceive an apple on the table"  for example. Prior to the disjunctivist way of thinking, one might argue this sentence could be true both if someone were veridically perceiving an apple on the table, or indeed if someone were experiencing a hallucination/illusion to the same effect. This in turn may lead one to thinking there must be common element to the two different types of perceptual occurrence such that they can both be true. Hinton however argued we should think of the sentence "I seem to perceive an apple on the table" in an entirely different manner to this. He maintained that actually it should be thought of as a condensed way of stating something like "either I see an apple on the table, or I have an illusion/hallucination of an apple on the table". It is this "reinterpretation" by Hinton of the perceptual statement which provides disjunctivism with its name. By thinking of these statements in such a manner Hinton also proposed that we should no longer assume a common element between the two disjuncts. This is because the disjunctive statement can be "made true in two different ways"  ; either I do veridically perceive an apple on the table and so the first disjunct is true, or I don't actually perceive an apple on the table but it appears to me as if I do and so the second disjunct is true. Hinton's work here enables later philosophers like Martin to develop the rejection of the common kind assumption as it introduces the notion that veridical and hallucinatory experiences do not share a common component.
Disjunctivists do not therefore reject the possibility of hallucinations occurring. Supporters of the position, as we have seen albeit briefly with Brewer previously, acknowledge when undergoing a hallucination, one is not be able to determine through introspection or inward reflection whether or not that perception is indeed a hallucination or is veridical in nature. Consider the hallucination within the play Macbeth. If the character were only able to use the tools of introspection when assessing his sensory experience, then he would not be able to determine whether he actually was perceiving a dagger, or was merely hallucinating. The two would therefore be known as "subjectively indistinguishable"  . Considering this notion, one might assume that veridical perceptions and hallucinations must therefore involve the same kind of conscious experience. With the aid of the rejection of the common kind assumption however, the disjunctivist is keen to emphasise that one mustn't think that the two are the same experientially. As Martin is eager to point out, they both have a very different "phenomenal character".
Assessing the disjunctivist position.
The formulation of disjunctivism I have explored through Brewer and Martin's accounts could therefore be said to offer an approach to perception that retains the key ideas of direct realism, whilst at the same time offering a way in which to possibly account for illusions and hallucinations. As we have seen, disjunctivists maintain that it is the mind independent objects of our perceptual experience that form or constitute that experience. Indeed, as another disjunctivist Peter Snowdon claims, when a veridical perception occurs, the experience "reaches out to and involves the perceived external objects"  . In the case of veridical perception then, the conscious nature of the experience is "derived" from the parts of our physical environment which we perceive. By appealing to disjunctivist notions, one is able to remain faithful to the naive realist thought that the objects in our environment "shape the contours of consciousness"  in a direct and immediate manner. It is the denial that veridical cases and hallucinatory cases involve the same phenomenology which seemingly allows the disjunctivist to formulate an answer to certain problems usually pitted against the direct realist. This is due to the proposed notion that the generalisation from cases of hallucination to veridical perception is halted, so the argument from hallucination supposedly does not threaten the naive realist's conception of veridical experiences.
Unfortunately for the disjunctivist however, their position is not without its criticism. One such objection is presented by Jonathan Dancy in "Arguments from Illusion". Within this article, Dancy explores some responses to the arguments from illusion, one of which takes the form of disjunctivism. The core of Dancy's argument against the disjunctivist is that their accounts of both hallucinatory and illusory perceptual experiences are rather lacking in detail and therefore should be considered unsatisfactory. As we have seen, the disjunctivist maintains that a perceptual experience is either veridical in nature or hallucinatory/illusory, and the two separate types of experience, or disjuncts, do not share a common experiential phenomenology. Dancy proposes that such an approach to perception actually has very little in the way of explanatory power, as all it actually achieves is to present the opposition of one of the steps of the argument from illusion in rejecting the common kind assumption. Very little, Dancy claims, is actually said about the nature of hallucinatory cases in particular. He criticises the disjunctivist for attempting to explain cases of hallucinatory experiences "solely by saying that it is like what it is not"  . One might indeed argue at this point that the account of perceptual experiences provided by the disjunctivist is somewhat unbalanced, as one of the two disjuncts appears rather more detailed than the other. Martin and Brewer seem to take great care in giving a detailed description of what happens in veridical cases (Brewer places illusions on the same disjunct as veridical); when one undergoes a veridical perception, that particular experience is of a "fundamental kind"  as it involves an immediate relation between the perceiver and a mind independent physical object of perception which is constitutive of that experience. It might be suggested however that all that is established concerning hallucinatory occasions is that such cases are subjectively indistinguishable from the veridical ones. This means that one cannot determine, through introspection, or reflective processes whether or not a particular experience is hallucinatory or veridical. Other than this characterisation, very little else appears to be said by the disjunctivist. We are given very little guidance as to the actual phenomenology of a hallucinatory perception; what happens experientially in such cases. Brewer's account in particular does not give much attention at all to this category of perceptual experience and the small amount he does mention relies heavily on Martin's thoughts.
Dancy maintains that the disjunctivist has simply formulated a "structured way of expressing opposition to the argument from illusion"  , and has not actually provided anything more than this. Their account could therefore be described as "incomplete"  . Dancy proposes that a disjunctivist theory could only be considered coherent and explanatorily sufficient if it articulates a "more direct"  account of the phenomenological nature of hallucinatory experiences. Dancy does seem to have a point worthy of consideration here. One might indeed argue that hallucinations are a kind of perceptual occurrence that surely merit as detailed an explanation as veridical cases if we are to properly understand them and include them within a theory of perception.
One might initially assume such an amendment to the disjunctivist theory to be easily achieved and commonsensical. Surely, one might think, if we are to truly understand the nature of our perceptual experiences, we need to provide a full account of the phenomenology of all the differing types? There are some disjunctivists that maintain that the second disjunct should indeed be further described and articulated so as to provide "an account of the experiential event"  which occurs in such perceptual situations. Such a position can known as positive disjunctivism  , as it proposes that a positive description of hallucination be specified. A species of this position can be found in disjunctivism about intentional content. Those who support the intentionalist account argue that our perceptions can be thought of as "representational states"  . Such theorists therefore claim that the conscious character of our perceptions is provided by intentional content which is how objects within our environment are represented as being  .
Disjunctivists who endorse intentionalist notions may therefore maintain that in cases of veridical perceptions, the mind independent world is presented to us by virtue of intentional content in a representational manner. In hallucinatory cases however they might claim that it is not possible for that experience to possess the same intentional content. Hallucinations could therefore be described as perceptual cases in which the intentional content is lacking  . Illusions, on this view can be explained as occasions involving "false representational content"  . The thought here would be that the physical environment of the perceiver is visually represented to them to be in a state that it simply isn't.
Martin is however highly critical of the requirement Dancy places on disjunctivism at this point and rejects attempts to provide a positive account of the second disjunct such as the one suggested above. He argues that if the disjunctivist wishes to remain faithful to the key tenets of naive realism, which he believes to be of great importance, then attempting to achieve Dancy's suggestion is actually hugely problematic. Martin maintains that if we try to provide a more positive and direct account of the phenomenology of hallucinations, we will ultimately need to identify an element that is "present not only in all cases of illusion or hallucination but also in the case of (veridical) perception"  . Martin's thought here is that if we wish to retain the disjunctivist notion that veridical and hallucinatory perceptions are of different fundamental kinds, then attempting to further describe the second disjunct as Dancy requires will in actual fact be damaging. To say any more than the characterisation given above would be to fall into providing descriptions of properties which veridical perceptions also posses, Martin argues. To add more explanation to cases of hallucination, or cases which involve "non privileged circumstances"  , Martin believes, would ultimately involve establishing the similarities between the two categories of perceptual experience, thus weakening the disjunctivist's position as this is the precise thing they are keen to reject. The way in which the disjunctivist proposes to describe perception would therefore become "explanatorily redundant"  . Taking this into consideration, Martin therefore continues to maintain that the only coherent way in which to characterise a hallucination that remains faithful to naive realist commitments he is do keen to uphold, is to state that it is indistinguishable from a veridical perceptual experience. This is the only way the disjunctivist can accurately describe the "fundamental mental character of the event"  whilst simultaneously continuing to uphold the rejection of the common kind assumption. Ultimately then, all that can be said about hallucinations is that they do not involve the same mental state as veridical perceptions, but that they are indiscriminable from them.
Even though Martin has provided reasons as to why this negative characterisation of hallucination is necessary in order for his account to remain coherent, I find it difficult to be convinced by the joining together of the two claims central to Martin's position. That two categories of perceptual experience can be subjectively indistinguishable yet at the same time have different phenomenal characters appears rather puzzling. The first of the two claims, that hallucination and veridical perceptions be indistinguishable through introspection alone seems accurate; it appears as if you would indeed need external factors to make a proper distinction between the two. The problem arises when this claim is combined with the notion that veridical and hallucinatory perceptions do not share a common experiential core. The two claims do not appear to be able to sit comfortably with one another, yet it seems that if Martin attempted to modify either, his disjunctivist stance would be jeopardised.
Martin is also put in a difficult position by the criticism from Dancy, which argues that his description of the hallucinatory disjunct is insufficient and incomplete. Martin is unable to amend or develop his description without wavering his adherence to the rejection of the common kind assumption, which as we have seen is the key factor in attempting to answer the argument from hallucination that plagues the naive realist.
Martin may attempt to argue the plausibility of his account by making reference to a particular motivation for the disjunctivist position. This involves the notion that the approach is preferable to other theories of perception because it is the least "revisionary"  . Here Martin is arguing that by rejecting the assumption that a veridical and illusionary or hallucinatory experience are of the same fundamental kind, the disjunctivist is providing a way of answering the problems of illusion and hallucination that doesn't stray too far away from our "pre theoretical conception of perception"  .The thought here is that disjunctivism as a theory of perception remains the most faithful to the common sense approach we have towards perceptual experience; that we experience an immediate relationship with the mind independent objects in our environment. Whilst some may think that maintaining that the three categories of perception we have looked at have a common phenomenological core is intuitive  , Martin claims that the process of rejecting this claim actually remains substantially closer to the "phenomenological data" available to us than rejecting the notion that perception involves a relationship between the perceiver and mind independent objects  , a notion for example at the heart of the sense data theory.
This motivation however does little to retrieve Martin from the difficult position he is in. It also relies on the disjunctivist being able to produce a strong response to the arguments from illusion and hallucination so as to relieve the naive realist from their problematic conclusions, and after considering their proposed strategy it is not clear that this has actually been provided. Echoing Dancy's thoughts, it appears to me as if in rejecting the common kind assumption the disjunctivist is merely stating the opposition a stage of the argument rather than giving a response which eliminates its force. Martin does give reasons for his rejection by arguing that there must be a difference in phenomenal character in cases of hallucination as there is no object of perception to constitute that perceptual experience. This particular claim however is motivated by his commitment to naive realism so it is as if he is using naive realist notions to ground his argument against the critic of naive realism, a move which is somewhat questionable. It appears as though Martin is basing his disjunctivist strategy on the very thing which the arguments from illusion and hallucination threaten in order to try and answer them.
It has been argued, that the rejection of the common kind assumption on which the disjunctivist places so much emphasis is itself questionable. Some are of the opinion, Scott Sturgeon included, that actually embracing the common kind assumption has much more explanatory power than rejecting it. He offers several points to illustrate that veridical, hallucinatory and illusory perceptions must have more in common than the disjunctivist allows. One such observation concerns our behavioural responses to particular perceptions  . Consider, for example, that I appear to perceive a car coming towards me at high speed. Whether this perception is veridical, illusory or hallucinatory it seems that I would still form a belief about the oncoming danger and move out of the way hastily, Sturgeon argues. From this, one might argue, that even if the perception of the car is hallucinatory, you still have a sense of the phenomenal character of the experience as you respond in the same way as you would if it were veridical. Another example might be watching a film whilst wearing three dimensional glasses. When it appears as if a character moves out of the screen towards you, you automatically flinch or move in the same manner that you would if this were actually occurring, even though the three dimensionality of the character is illusory. Surely, Sturgeon argues, the best way to explain such phenomena would be to accept that the three categories of perception involve "mental states of the same kind"  , and this is why our behaviour can be the same irrespective of the category to which the perceptual experience belongs. This observation by Sturgeon does appear to lend support to postulation of the common kind assumption, and if this is the case then it questions the necessity of the rejection of this which the disjunctivist adheres to.
Martin may respond to this objection by saying that the same behavioural response can result from either a hallucination or a veridical perception simply because the two are subjectively indistinguishable to the perceiver. Sturgeon however argues that the truth of this claim adds further wait to the notion that the two types of perceptual experience share a common phenomenal character, rather than strengthening the disjunctivist position.
At first glance the theory of disjunctivism is extremely attractive to the naive realist. At the core of the theory, as we have seen through the accounts of Brewer and Martin, is the notion that in veridical cases our perceptual experience is constituted by the mind independent objects of our perception in an immediate relation. It also, through the strategy of rejecting the common kind assumption, attempts to offer sanctuary to the naive realist from the damning conclusion of the argument from hallucination and illusion. On further inspection however, such merits appear to dissolve. It is not clear that the disjunctivist offers a way in which to avoid the arguments, other than simply denying one of the premises - a denial which is arguably not sufficiently warranted anyway. There are also elements of the theory itself which are rather dubious. Martin's inability to offer a full explanation of the second disjunct leaves him in an unsatisfactory and difficult position, but the point of greatest concern is his attempt to postulate two claims that seem incompatible in order to secure the rejection of the common kind assumption.
I therefore conclude that the theory of disjunctivism as I have explored ultimately fails to explain occurrences of hallucination and illusion in a way that preserves the essence of direct realism.