Introduction

The problem of sense perception together with the theory of sense data was one of the most hotly debated philosophical issues of British and American philosophy in the first quarter of the 20th century. Eminent thinkers like Moore, Russell, Broad, Price and others grappled with issues relating to the nature and relationship of sense data with   physical objects and the external world. In this paper I shall try, first, to state how the philosophical problem of sense perception arises and the different lines of solution to the problem before Wittgenstein arrived on the scene. I shall then very briefly explain the rationale of the Linguistic Analytical approach to philosophical problems and theories in general. Having done so, I shall apply, in some detail, the linguistic analytical approach to the theory of sense data. This will be followed by some concluding remarks to round off the paper.

Commonsense Realism and the Statement of the Problem

According to commonsense, physical objects like chairs and tables and the entire panorama of nature exist independently of being perceived by any mind, and the act of veridical perception reveals the objectively real qualities of physical objects. Illusions and hallucinations of different types do occur, but this an aberration of normal sense perception. Commonsense Realism is quite satisfactory for all practical purposes. What then are the difficulties, which the philosopher notes, and which prompt him to formulate the different theories of perception?

We believe we perceive objects directly and non-inferentially. But our actual perception is confined at any moment of time to a part of the visible surface of the object rather than the whole object in the literal sense. It may be said that what we perceive directly and non-inferentially may not be the whole object, but it is after all a part of the same. But the occurrence of illusions and of plural perspectives raises the following difficulty. An optical illusion is not a case of misperception due to carelessness or lack of training on the part of the subject. The content of the illusion persists no matter how careful or how repeated the act of perception. The observer clearly and directly sees something as given to ones sense organs and the determinate features of the given data differ from the independent physical object. If, for example, what we directly perceive is something with an elliptical shape, this something cannot be deemed to be a part of a known circular object, like a penny. Again, if what is directly and clearly perceived has a crooked shape this something crooked cannot be a part of a straight stick. This something, which, is directly and immediately perceived in a single act of perception, has been called the sense datum.

It may be said that though in the case of illusions the sense data may not be equated with the parts of the physical object, this is the case in veridical perception. Since, however, the perceptual process and perceptual assurance are common in both veridical and non-veridical perception (involving one particular sense organ) how can we justify the belief that sense data are literal parts of physical objects in the case of veridical but not in the case of non-veridical perception.

If the reports given by different sense organs fail to converge, this suggests that one or the other sense organ is giving a non-veridical report. But if we concede this, we will also have to concede the basic theoretical possibility that some other test or tests might not confirm the perceptual report now accepted as veridical. Consequently, all that we can say with absolute certainty is that something, which is not of our own making or choice, is the sense datum rather than the whole object. Whether the sense datum is part of the physical object is a matter for inquiry, and cannot be taken as a settled matter of fact as one is certain about the sense data, as such.  Philosophers thus thought that the concept of sense data was a neutral concept, which did not presuppose any theory of the universe apart from the view that our sense organs gave us reports about some independent data. The concept of sense datum left them free to determine the exact nature of sense data as direct objects of perception without committing them to any further ontological or epistemological theory. They retained the fundamental commonsense belief that something external to the perceiving subject exists independently of sense perception, but they questioned the commonsense view that sense perception provided a reliable photographic copy of the real world. This line of thinking was facilitated by a growing knowledge of the physiological basis of perception, which science showed to be a highly complex psychophysical process instead of being a simple or purely mental or physical phenomenon. 

Difficulties in the Sense Data Theory:

This concept immediately raises the problem of the relation between the sense data and the physical object, indeed the very existence of the latter. If the commonsense position raises prima facie puzzles about the objects of direct perception, the sense data theory raises puzzles about the relation between sense data, as the direct objects of perception, and physical objects, as posited by commonsense as well as Newtonian Physics and Cartesian metaphysics. It is as if the cure of jaundice led to the onset of paralysis.

At the commonsense level we believe that the immediate object of direct visual perception is a part of the total physical object, just as the hand rest of a chair is part of the chair. But could sense data be said to be constituent parts of the physical object in the above sense? No, since sense data vary according to the internal and external conditions of the perceptual act, while physical objects have relatively fixed and stable constituents, according to both commonsense and Newtonian Physics. Now different sense data cannot all be the parts of one and the same object. Secondly, sense data are by the definition mind dependent, according to some philosophers, while the physical object exists independently. Now entities, which are at least partly dependent upon the mind and which, therefore, exist only intermittently cannot be a pat of a relatively permanent and independent physical object.

I think both the above difficulties could have been removed in either two ways, (a) by accepting that sense data have not merely epistemological primacy over the physical object but also ontological primacy, that is, sense data are the primary entities that constitute the ultimate furniture of the world, (b) the term ‘sense data’ does not stand for or refer to any entity over and above physical objects, but is only an abbreviated expression for the full descriptive phrase/expression, ‘physical object or a part thereof as it appears to a particular observer at a particular point of time’.

The Neo-Realists did adopt the first half of the solution and declared that physical objects are logical constructions out of sense data. But then another problem was generated. There were different types of sense data themselves. Clearly, a double image or after image was as much a sense datum, in the literal sense, as a mirror image, but there were obvious differences between them. Now, if we believed in the existence of physical objects in the realist sense, we could say that there were two types of sense data namely those, which were parts of the physical object, and those, which were not parts as such. Since the Neo-realists gave ontological primacy to sense data and held that physical objects were logical constructions, they had no criterion left for distinguishing between all manner of countless sense data that constituted the furniture of the universe. Thus, the Neo-realists produced another puzzle in their efforts to solve the puzzle of the relation between sense data and physical objects.

It appears the Neo-realists fell into the trap because they confused the reality of the perceptual experience qua experience with the reality of the object as such. The illusion is a real occurrence just like veridical perception. But this does not mean that the experienced datum is itself real. The Neo-realists did not analyze the different uses of ordinary words and went about constructing a conceptual model of the perceptual situation, no matter how startling to commonsense, provided it could enable them to solve the puzzle of the relationship between physical objects and sense data. Perhaps they thought that if scientific theories could take such extreme liberties with commonsense concepts and beliefs, philosophical theories could also do likewise. But they ignored the fact that scientific theories were subjected to the discipline of empirical verification, while philosophical theories did not have this built in check, and thus could go wild.

It may be said that the futility of the speculative excesses of the Neo-realists, on the one hand, and the honest puzzlement of Moore about the proper analysis of indubitable judgments on the other, conspired to bring the focus of enquiry upon philosophical method rather than Epistemology or Ontology. Moore had accepted the concept of sense data because of its promise of solving the difficulties created by the occurrence of illusions, and fragmentary perception etcetera and also because the concept seemed to be clear, non-controversial or non-speculative in character. But when he asked how was the sense datum related to physical objects whose independent existence he could not doubt, he got bogged down into insuperable difficulties, which he had the moral and the intellectual courage to admit. He never believed in a therapy that cures the disease but kills the patient. It was Moore’s honest puzzlement and the honest admission that he was stuck up both in ethics and epistemology that stimulated and paved the way for Wittgenstein’s linguistic analysis (much after his earlier Tractatus) and much different from the method of Russell’s logical atomistic analysis. 

Brief Explanation of the Linguistic Analytical Approach to Philosophical Problems:

According to the school of Linguistic Analysis philosophical controversies arise when we, quite unknowingly, use words and expressions in different senses and thus land ourselves in puzzles and perplexities that seem to be insoluble. For example, one may say that if nothing happens without God’s will, why should criminals be punished for their crimes?  This question may trigger a debate that never ends because no one is able to clinch the issue. The reason is that there is no prior agreement about the exact meaning and use of words and expressions used in the controversy, such as Creator, Divine Will, crime, justice and so on. All words of a natural language have fluid uses and meanings in different contexts.

There is yet another major source of confusion in our thinking and reasoning; our natural tendency to think that all meaningful words that are grammatical nouns must be names for some existent or entity of some kind or other.  For example, noun words such as ‘justice’, ‘love’, ‘truth’, ‘government’, ‘Indian Navy’, ‘winter’,  ‘forest’, ‘storm’, and so on, must refer to some specific entity or states of affairs constituting reality. Linguistic analysts maintain that this unconscious assumption generates philosophical puzzlement and controversies that can never be solved through abstract reasoning or argumentation.

Let us suppose we accept the above basic approach of Linguistic Analysis. The question now arises: Will accepting this approach suffice, by itself, suffice to clear up the mess created by numerous philosophical theories since the dawn of language and systematic reasoning in human society? The answer is a plain and emphatic no. What is further needed is the rigorous and sustained application of this type of linguistic analysis of the numerous problems of philosophy and of life, as such, with a view to pointing out in detail the origin and genesis of various conceptual illusions generated by specific semantic confusions, assumptions, Paradigms, analogies and the like. Russell or Moore did not practice this type of analysis. Both rejected and outgrew the classical assumption that metaphysics was the ultra-scientific anatomy of reality, while natural science dealt only with phenomena.

Is classical metaphysics, then, merely a language game and nothing more? I hold that the metaphysical quest, leads to a critical existential interpretation of the human situation and this brings about inner integration and peace of mind to the individual. John Wisdom himself points out that metaphysical theories draw our attention to unsuspected facets of human experience even though they also mislead us. This paradox of human language makes him say that the proper method of doing philosophy is eirenics, not polemics. We must strive to show how different theories simultaneously illuminate and mislead. Thus, for example, we must show how or in what sense the view that human willing is free, and the view that human willing is determined are both true and also false in some sense or other. The same remarks apply to the view that there is purpose in the universe and the view that there is no purpose of the universe. Contrary theories become acceptable when one asserts them as true and in the same breath qualifies them to point out how they mislead. This joint affirmation and negation reveals the complexity of the universe and the limitations of human communication. This type of dialectical reasoning is not required in the case of scientific theories because they can be empirically verified.

The Linguistic Analysis of the Theory of Sense Data

How does linguistic analysis of the problem of perception resolve the difficulties of common sense realism without the introduction of the concept of sense data? Let us examine in some detail the argument from illusion, which is supposed to provide the grist to the mill of the theory of sense data, and let us concentrate on the well-known illustration of the stick, which is seen as bent when it is placed in water.

The sense data philosopher wants to know what precisely is the direct and immediate object of our perception when we see the stick in water. It cannot be the stick, since what we see, and see most clearly and directly, is bent, while the stick is straight. But we do see something. This something is the sense datum, which cannot be a part of the object, since the sense datum is bent or crooked, while the parts of the stick are all straight. 

Now the linguistic philosopher points out that the way in which the sense data philosopher poses the problem is most misleading, and that if we look at the commonsense view carefully, no problem or puzzle arises. In both cases of seeing the stick outside water or inside water we see the straight stick. In the latter case we see a straight stick, part of which is above water and part of which is inside water, and the immersed part appears to be bent or crooked. We know that often things appear to have features, which they do not, in fact, have or do not appear to have the features, which they in fact have. But this does not constitute any problem, and is merely a feature of our experience, which we should take note of and accept as a given fact.

The statement that in the above case we do not see the stick but only the sense data is just the initial blunder, which vitiates our entire way of looking at the matter. The concept of sense data is introduced because we use the word, ‘see’ in both the cases. If we had used the word see in the first case and the word ‘appears’ in the second, there would be no trouble at all.

What is the puzzle or problem in the statement that we see a straight stick, which appears to be crooked in that part which is immersed in water? Similarly, what problem or paradox is there in the statement that an object is round or ten feet high, but appears to be elliptical or only ten inches high from such and such a distance?

The trouble only arises when we say that while the stick is straight the sense data of the stick or what we directly see is bent or crooked, or, while the penny is circular, its ‘sense datum’ is elliptical. There would be no trouble if the philosophical expression; ‘the sense datum is elliptical’ is avoided and we continue using the ordinary expression the penny appears elliptical. This would suffice to give us a clear and consistent conceptual picture of the perceptual situation without recourse to sense data language. Even using the expression sense data would not create any mischief, if it were understood, that sense data is just a philosophical expression for referring to a physical object or part thereof and not a noun to designate any new entity, that was unsuspected by the common man but discovered by some philosophers.

The term sense datum would then mean the ‘physical object or part thereof, as it appears to the sense organs of a perceiver in a particular situation’. The point of introducing this long descriptive phrase would be that the ordinary expression physical object and the long descriptive phrase in question do not have the same connotation, although they have   same denotation or refer to the same entity, just as the expression; ‘the present Prime Minister of India’ and ‘the daughter of J.N. Nehru’ had different connotations but exactly the same denotation at a certain point of time. A linguistic analysis of the words ‘appears’ and ‘is’ removes all confusion. For example, the sentence; ‘the penny is circular, but appears to be elliptical’, makes perfect sense. Again, ‘the penny is circular, but the appearance of the penny is elliptical’ is also correct. But in the second case the word ‘appearance’ has been used as a noun and this suggests that it referent must be some entity or object. But whereas the word, ‘is’ entails existence or predication (among several other uses) the word ‘appears’ does not refer to sheer existence of something but also to an act of ‘perceiving’ or a process. The moment the act stops the word appearance ceases to be applicable any more. In other words, the something to which the word ‘appearance’ is supposed to refer is not an independent object but refers to an intermittent relationship involving three terms; subject, object, and act of perceiving.  If so, no contradiction is involved if the object is ten feet high but its appearance is only ten inches high.

The appearance of an independent object is not another independent object, but how the subject perceives the object in non-veridical perception.  We may say that the physical object consisting of parts exists independently.  This independent object, at times, becomes the object of perception. When this happens its parts may be called sense data. We can also say without any theoretical difficulty that in one sense the sense data are parts of the physical object as such, without involving the subject, and, in another sense, they are part of the perceptual situation of the subject during the act of sense perception. When there is no observer there are no sense data in one sense, but in the other sense the sense data of an object are co-terminus with the object as such. In this sense the philosophical expression ‘sense data’ becomes almost equivalent with the ordinary language expression ‘physical object’.   

The point in introducing the expression, sense datum is that it can function as an abbreviation for the rather lengthy and cumbersome expression; ‘a physical object or part thereof as it appears to an observer in a specific situation’. Thus, we can say that a circular physical object exists all the time whether it is perceived or not, but the circular object appears as an ellipse or the elliptical sense datum exists only at a particular time. This type of appearance is however, different from mal-functional appearances that arise due to some reason or other but can be prevented from occurring. The sense datum, in the case of an optical illusion, however, persists no matter how careful the act of perception and how repeatedly the observer attempts to get at the real thing. In this sense the appearance is a hard fact of life.

But when we say that appearance is a hard fact, what we mean is that the appearing is a fact, and not that the contents of the appearance correspond with the contents of the world. The appearance is presentationally real though not objectively real. Thus, all we are justified in claiming is that there are many uses of the word appearance. In some contexts the appearance is not fleeting or arbitrary, but is a function of definite rules of perspective.

Let us further examine the logic of ‘appears’. The sense data philosopher takes a fancy for the word ‘is’. Instead of saying that, ‘the penny is red but appears to be brown’ he prefers to say, ‘the penny is red but its sense datum is brown’. Similarly, he says that the sense datum of the stick is straight, but the sense datum of the stick in water is bent. The divergent sense data cannot be parts of the same object. But this difficulty is partly similar to the difficulty that the thing remains the same, and yet its shadow waxes big or small at different times. This phenomenon remains intriguing so long as we do not understand the laws of light and optics, but once we do come to know them, we are not mislead or deceived by the changing shapes or sizes of the shadows of a stable physical object.

The phenomenon of mirror images, double images and after images is also one of the sources of the sense data controversy. We know that mirror images are neither just like physical object not just like mental images. Mirror images cannot be touched but they are seen and can be photographed. They can certainly mislead the unwary and they play havoc with birds or animals, which cannot discriminate them from physical objects.

The argument from mirror images for the existence of sense data parallels the argument from illusion. That which is seen when we see a mirror image is not the physical object, which is on this side of the mirror, while the something seen is on the other side. That something is the sense datum. Now what is the difficulty in saying that a mirror mage partly resembles and partly differs from both of them? As regards the relationship between the mirror image and the physical object it may be said that the mirror image is the image of the object on the surface of the mirror but this ‘on’ is different from the sense of ‘on’ when the book is on the table. The reasons for this is that the image is projected or formed as much behind the mirror as the object is in front, and yet looking directly on the mirror is necessary for seeing the image behind the mirror. If we focus our attention behind the mirror, in the literal sense, and not on the mirror, we will not see any image at all. The image is, thus, sui generis and to call it mental is as misleading as to call it a physical object.

Let us now examine some further aspects of the puzzle whether we do or do not perceive physical objects directly or only sense data are directly perceived. It is true that physical objects like, chairs and tables are not fully perceived in one perceptual act but require a number of perceptual acts from different angles and sides to make the perception complete. Our perception is avowedly partial or fragmentary. But this does not imply that our perception is indirect. We may, if we like, indicates this feature of our perception by saying that we perceive sense data directly and we perceive physical objects indirectly. But such language becomes extremely misleading, indeed.

What, if any, are the conditions in which the statement; ‘we perceive sense data directly and physical objects indirectly’, would have been true in the non-trivial sense? This statement would have been significant in a profound sense if we could point out to some perceptual experience which is direct and immediate, and in contrast with which the perception of physical objects could be held to be indirect. The sense data philosopher thinks that the experience of sense data is such a direct or privileged perception. But we find that this is not the case, and perceiving a physical object and perceiving a sense datum is the same type of experience and the same process, even though there is some distinction between how the different expressions are used. When I say, ‘I lend you this book’, while handing it over to you, and when I say, ‘I present this book’ while handing it over, the process of handing the book is the same. Now, the point is that the process of seeing a sense datum or seeing a physical object is the same, although our expectations and subsequent behavior differ in some though not all situations, when I say ‘I see a brown patch’, and when I say ‘I see a table’. But because of this difference in some cases, it does not follow that the act of seeing sense data is different from seeing a table or chair. We just cannot give a clear and positive sense to the expression ‘direct perception’. And the expression ‘indirect perception’ remains an empty phrase. Seeing the mirror image of a chair or seeing a chair enclosed in a glass chamber etcetera, could be deemed to be cases of indirect perception of the chair in contrast with normal perception of the chair. But normal perception of a chair itself cannot be deemed to be indirect, when there is no direct perception to contrast with.

There is, thus, no contrast of direct/indirect knowledge of physical objects and knowledge of sense data. If I say, ‘I see a brown surface over there’, and if I say, ‘I see a table over there’, in both cases I report what I directly see or confront and not what I infer, predict, dream, or remember. In point of directness, immediacy and even the certainty that some ‘not-self’, as object, is confronting myself, as subject, there is no difference between the two perceptual situations. But does this mean there is no difference at all between the two statements?  No, there is a difference.

The difference between the two statements lies in the degree of their specificity of truth claims rather than in their objective referents. Austins example of the difference between kicking Jones door and kicking painted wood is illuminating. In other words, statements about physical objects need greater and more varied tests for confirming them, while statements about sense data do not need the same tests for their confirmation. The same feature could be expressed by saying that statements about physical objects are more vulnerable than statements about sense data. Suppose a person claims that he has intense pain in his hand or head but shows no signs of it at all. We will be justified in inferring that he is joking or lying. If, however, he persists in claiming that he does have pain sensations, but that he has the capacity to bear them with a smile on his lips, it would be difficult for us to refuse him. On the other hand, if a man claimed that there was a table over there, but the table was visible to none, could be touched by none, would support no light objects on its surface etcetera, we would be justified in referring his claim and refuting his statement. In the same manner verification procedures of statements about physical objects and statements about sense data partly differ. The sentence, ‘I see a brown patch over there’ entails far less than what the sentence, ‘that is a table’ entails. This latter sentence entails such truth-claims statements as, ‘if you touch it, your hand will have such and such sensation’, ‘if you walk in that direction or try to put another table in that region, there will be a collision’, and ‘if you place a book or ash tray on the surface they will remain supported’, etcetera. On the other hand, a sentence like ‘I see a brown patch over there’ entails very little. 

The more the scope or complexity of the principal truth-claim the greater is the risk of its being false. On the other hand, the verification of each sub-claim tends to confirm the truth of the principal claim. But the fact is that even the truth of all the sub-claims does not logically prove the truth of the principal claim. It is logically possible for the principal claim to be disconfirmed by the nth test or in the nth instance, though right up to the n-1 instance, all sub-claims had been converging upon the truth of the principal truth claim. This is the real point behind the statement that our knowledge of physical objects is indirect, while our knowledge of sense data is direct and immediate. But then to put this point in terms of direct and indirect perception is extremely misleading indeed.

The upshot is that there is some difference between statements of the form I see a brown patch and of the form I see a brown table: but the difference is not such as the sense data philosophers suppose it to be, and this difference does not imply that our knowledge of physical objects is indirect or imply that our knowledge of physical objects is indirect or inferential while that of sense data direct and immediate. The inference of fire from the sight of smoke or the inference of a motorcar from the engine sound is indirect knowledge as compared with actually seeing the fire or the car. But when we see the fire or the car as such, knowledge cannot be said to be indirect, inferential and dubitable, though of course it is contingent. Moreover, even the knowledge of sense data is dubitable and liable to be mistaken not only in the trivial sense of being given a wrong label or name, but in the more serious sense that the sense data may be misperceived due to carelessness or lack of training etcetera. It is, therefore, very misleading to contend that while the existence of sense data is indubitable, the existence of physical objects is only probable, or, to contend that, the reasons for the belief in the existence of physical objects are not as compelling or cogent as are the reasons for the belief in sense data. This suggests that the belief in physical objects could have been more compelling or could have possessed a higher certainty than it actually does have. But then just as in the case of indirect perception, there must be a positive sense of direct perception to act as a foil for indirect perception; there must be a sense of complete certainty in contrast with which we could understand incomplete or near certainty.

It may be said that we do have the concept of logical certainty, which we attribute to mathematical and logical truths. This is indeed the model or Paradigm of certainty, which the sense data philosopher has in mind or which rather holds his mind captive, as Wittgenstein aptly puts it. The philosopher forgets that the model, at least in principle, should be applicable to the belief concerned. If the model is inapplicable, in principle, to the domain of the truth-claim in question it will not make any sense to demand that its level of certainty should be the same as logical certainty. There would be no point in the lamentation that the belief in question lacks certainty. This is precisely the error into which the philosopher falls. It just does not make sense to say that a factual truth-claim should be certain in the logical sense of certainty.

The factual model of certainty is quite different from the logical model. Indeed we could even say that factual certainty and logical certainty are distinct concepts, which are clumsily lumped together due to our general tendency to ignore subtle differences and seek unity in variety. This is the craving for unity or the search for essences in the language of Wittgenstein. Consequently, if factual truth and logical truth are different concepts, each with its own appropriate reach or area of application, the demand for logical certainty in an area where it is not applicable, by definition, is like the demand that ethical judgments be verifiable like the scientific, or that the general laws of science or other law-like truth-claims be confirmed like particular truth-claim of the type, ‘this cat is on the mat’. The demand for this type or model of certainty is rooted in our tendency to suppose that the word ‘certainty’ must designate or stand for some essence of certainty behind all cases or instances of real certainty.

When however we free ourselves from one particular model of certainty, that is, the logical, we realize that in the area of factual truth-claims the demand for this particular model is misplaced. We, then, no longer feel prompted to say that the existence of physical objects is either not certain, or less certain than that of sense data. We do not feel prompted to say that while we have direct and immediate knowledge of sense data we have indirect knowledge of physical objects. Thus the puzzle of the relationship between physical objects and sense data and other related philosophical problems are not generated at all. This constitutes their dissolution instead of their solution (in the classical speculative or analytical tradition).   

Conclusion

The above linguistic analysis of the theory of sense datum has removed some of the confusions that arise in our thinking due to our almost inevitable tendency to reify words specially those that are used as nouns in our discourse. This creates artificial problems that demand true answers or solutions when in fact there is no problem and no true answer that excludes other answers. This means that the so-called philosophical problem of perception and the so-called sense datum theory to solve the problem were both pseudo arguments. At the same time the expression sense datum, as a philosophical or theoretical construct, had a limited use as a convenient short hand expression for the much longer expression, ‘the physical object or part thereof as presented to the observer at a particular space-time moment’. Genuine factual knowledge relating to human perception involving different human sense organs and the nervous system and the human subject as such lies in the domain of anatomy, physiology and psychology rather than of logic or philosophy. Wittgenstein has clearly stated that linguistic analysis does not add anything to our knowledge but just dissolves our confusions or illusions that give birth to the different theories of perception or knowledge or ethics or metaphysics as the case may be.  The proper way to acquire knowledge is to tap the door of factual knowledge or to investigate the truth according to the scientific method. However, the removal of confusions and logical errors is, by itself, no mean achievement. In fact, it amounts to human liberation from artificial theoretical perplexity.

The above linguistic analysis of the problem of perception also shows very clearly the value of John Wisdom’s concept of the paradox of linguistic communication. This is the paradox that all acts of communication illuminate as well as mislead the communicatee.  Thus, to say that ‘one perceives physical objects indirectly’, or to say that ‘physical objects are logical constructions out of sense data’, or to say that ‘physical objects are nothing but permanent possibilities of perception’, or to say that ‘all perceptual knowledge is confined to our own perceptions’ and so on, at one and the same time, are penetrating remarks that make a valid point but also mislead in specific ways. In what follows I shall give several examples of this paradox in different theories of perception and knowledge.

The sense data theory draws our attention to the different ways in which words are used and different types of truth-claims are verified. It also helps us to understand the logic of our language when we deal with different types of nouns, verbs, adjectives or prepositions even when they are grammatically the same. The realization that though the grammatical structure of the sentence, ‘this rose is red’, and the sentence, ‘this rose is beautiful’, is exactly the same, their functional logic and methods of verification are entirely different. The uses or functions, implications and methods of verification of words and expressions that refer to physical objects, individuals, corporate bodies and processes and so on are all different. These issues are barely understood or known at the common sense level of the users of ordinary language.

          

According to the Representative Realism of Descartes and John Locke the primary qualities of matter or Substance‚Äîshape, size location and motion are objective and inhere in the external world independently of being perceived by any observer or not, while the secondary qualities ‚Äì colour, taste smell and sound are subjective and they appear only when the external substance impacts some human percipient in the act of perception. The secondary qualities do not inhere in the external world but emerge when the object and the perceiving subject are suitably placed.  The Phenomenalist approach to knowledge and the external world holds that both primary and secondary qualities refer to reality as it appears to some subject, but one has no access to Reality as it is by itself.    All such views or theories have some ‘point’ to them, but none is free from some objectionable inner twist or knot. In other words, each theory illuminates and misleads at the same time. Locke’s theory appears impregnable so long as we operate in the Newtonian universe of space, time and motion. But it seems to break down when some one begins to question the objectivity of space and time as such. Indeed this is what Einstein did. Kant had done it much earlier. Phenomenalism is, perhaps, logically irrefutable. On this view even the most accurate and rigorous factual knowledge based on the scientific method becomes tainted by human subjectivity. 

The objective idealist position is hardly a clear statement and can be interpreted in diverse ways. Moreover, there is no compelling or logically coercive support for accepting it. However, both the phenomenalist and idealist positions are logically possible though they cannot be established or clinched.  The same applies to Realism as a theory of knowledge.

The upshot of the above linguistic analysis is that the demand to prove any of the above philosophical theories of knowledge and of Reality is misconceived. This is the case not because of any limitation in our powers of reasoning, or the scientific method. It is due to our asking wrong questions because of semantic confusions in regard to the nature, types and functions of human language in action. This was Kant’s point of departure. Despite his tremendous genius, the Western world had to wait for almost two centuries before Wittgenstein could formulate his mature views on the diverse functions of language and his key concept of language games to bring out the hidden nature of pompous and forbidding philosophical theories which can never be proved or disproved one way or the other.

To conclude, if theories such as Realism, Phenomenalism, at bottom, be alternative language systems or games, that are neither true nor false, but only helpful or illuminating (in part) and misleading (in part) why not play the game of ordinary language and avoid technical philosophy as such. This is what the famous Oxford thinkers, J.L. Austin and Gilbert Ryle recommend.  All said and done ordinary language is not more misleading/illuminating than philosophical theories. Moreover, we are all familiar with its standard expressions, unlike the pompous expressions philosophers love to coin, every philosopher taking delight and pride in ones own theories or ‘isms’.