Universals
This essay is concerned with the problem of universals, in particular the problem of quality universals. The main focus will be on exploring the possibilities and problems of Realist theories of quality universals in an attempt to dismiss Realism in favour of a Conceptualist account of quality universals. Although there are several versions of Realism, for the purpose of this essay, it will suffice to keep descriptions as general as possible, and as I shall be criticising the essential aspect of realist ontologies I will not intentionally be focusing on any particular version. Essay discussion will be based on the colour universal redness though it is believed the arguments provided could be applied to at least all quality universals involving sensory perception. After an introduction to the problem of quality universals, I will briefly acknowledge and discuss two arguments against Realist theories of quality universals. I shall then commence the central thesis of my essay by considering the ontological implications of employing a noun to denote a quality universal, before delving into the main argument, in which I will firstly consider what physics has to say about colour and why it undermines simple Realist ontologies that hold a language-implied, commonsensical view of colour as being a property inhering within objects. Resultant realisation coerces me to shift from treating redness as an actual property of red objects to treating redness as being related to red qualia. It is then argued that this modification, an attempt to maintain a viable Realist ontology, results in a Realist ontology that is still vulnerable. The approach to this argument will be to create a hypothetical, yet probable scenario, involving the notion of perceiver relative qualia, which I intend to show, cannot be adequately accommodated by this modified Realism. I then finally pursue a case for Conceptualism by showing how it can adequately address the problems put forward against Realism throughout the essay. Before I commence, let me mention that this essay exemplifies the maxim that "metaphysical enquiry shall not expect any certainty" .
The problem of universals for the realist is the question of whether universal abstractions in the mind can be said to be objectively representative of anything in reality. Before continuing, I must specify that if universals are to be considered real, it would be appropriate to adopt Bertrand Russell's terminology and refer to them as being subsistent rather than existent, for it is clear that in a system which admits of particulars and universals, universals are neither spatially nor temporally located . I shall therefore adopt this terminology in my essay. Now general definitions of the universal 'redness' given are that of 'something which is common to all red objects' or that of 'something instantiated by all red objects'. Consider any red objects; some red apple, some red car and some red sweater. It is visually evident that these three objects all possess one common quality, namely the quality of being red. Unlike these particular red objects though, 'redness', may be said to be in many places at once: we may say, in a sense, that 'redness' is in every place where there is a red thing . To the Realist, this common quality, labeled by the noun 'redness', identifies some entity, a universal, objectively subsisting independent of human minds. This ontological status attributed to 'redness' is certainly not beyond doubt. Before I commence expounding the central thesis of my essay, I shall acknowledge and very briefly consider two arguments often made against realism, the second perhaps stronger than the first. It should be noted that my central argument is independent of the following two.
Of the three red objects previously considered, let us say that the apple is light red, the car is medium red and the sweater is dark red. The hues of these three colours are clearly different, so it seems problematic to say that these three objects instantiate the same colour. The Realist may be able to avoid this problem by assuming that there are hierarchies of qualities, of which redness is the only essential quality universal in the case of all shades of reds and quality universals such as lightness and darkness are accidental. Antagonists will persist and point out that since not all light reds are the same, this hierarchy of light red would have to extend to another level and since colours come in a continuum, this hierarchy would have to include at least an infinite number of levels. "But Realism would require that such a structure of categorization exist as part of the structure of reality prior to our undertaking such a process in the first place. To account for all the richness found in the world, the structure could end only at the level of particulars, the inhabitants of everyday experience, which constitute a concrete reality against which this approach falls into a kind of reductio ad absurdum" .
A Realist may object to this by saying that although the portion of the electromagnetic spectrum corresponding to visible light is theoretically continuous, the colours that can be distinguished from it are discrete. As there is a finite set of shades of red derivable from this discontinuous spectrum, this hierarchy would be satisfied by a finite number of levels. The hierarchy could end at a base level of amalgamations of universals and still accommodate 'all the richness found in the world'. Each of these different shades of red (amalgamations of universals) would all contain the essential quality universal redness and be instantiated by a significantly small portion of all the red objects. For example, consider a pencil manufacturing company that as part of their colour pencil range manufacture a pencil of a particular red shade. This shade is instantiated by all such pencils made by the company and the collection of all such red pencils is a significantly small subset of all red objects. It is not necessary to end at the level of particulars (each pencil) to account for this portion (all such pencils) of the world as suggested in the quote ending the previous paragraph.
Another, more significant problem for the realist is known as 'the borderline problem'. Blue and green are obviously distinct qualities and it seems reasonable to assert that both blueness and greenness cannot be instantiated at the same spatiotemporal location. However there exists a region between these two separate qualities which cannot be clearly identified with either. This region would include the variations of blue-green, greenish blue and bluish green. If realism were true, this 'borderline region' would not be expected to exist. According to realism, colours should clearly be the same qualities or different and mutually-exclusive qualities, and there should be no borderline cases. But borderline cases clearly do exist, and there exists an entire region which is composed of them. If the line is to be drawn (something which may seem arbitrary), to which quality universal does it belong, blueness or greenness?
As it is relevant to my argument, I shall now briefly consider the semantics of a noun denotative of a quality universal, such as 'redness', before expounding my hypothetical argument against the reality of quality universals.
Realists may claim that "Realism is an exercise in common sense. It is that theory which comes closest to explaining everyday speech" . The human capacity to group red objects together leads to the accepted assertion that redness is common to all red objects. Although grammatically acceptable, the adjective-noun translation of a predicate such as 'is red' into a corresponding predicate such as 'exemplifies redness' is problematic. The apparent meaning attached to the word redness seems to imply that there is something real to which it actually refers. To intransigently maintain this belief would merely be reification, a case of being misled by verbal form into thinking simply because some noun has a use, there must be something to which it refers. Surely "the existence of a general word does not imply the existence of a general thing".
One considered move to deprive 'redness' of any ontologically real status is the elimination of the word from our vocabulary altogether and translation of sentences involving the noun 'redness' into sentences involving the adjective 'red'. Sentences such as 'redness is common to objects x, object y and object z' can be replaced by 'objects x, y and z are red'. But this is unnecessary and any attempt at complete translation may not be satisfactory. It could be that the existence of the word is solely a matter of linguistic necessity. It would be improper to fill a descriptive predicate with an adjective, so if we want to describe the sensation produced by our observation of a red object, we must express this description by employing the word 'redness'. It seems though, that this sensation is something and perhaps the word 'redness' can be acceptably employed to express awareness of perceived commonality in particular qualitative sensations. This is of great significance as it alludes to the possibility that a conceptualistic interpretation of the word redness may still render it meaningful, yet not commit us to consider it to be denotative of something objectively subsisting. I will further discuss this notion in the closing stages of the essay.
One possible objection to a Realist's interpretation of redness is that physics does not consider colour to be an intrinsic property of an object in the first place, a 'red' object is not actually red. To say that object A is red or object A exemplifies redness is just a way of explaining our conceptualization of the world; redness is a subjective interpretation that does not actually inhere in 'red' objects. Therefore the problem of quality universals is an unnecessary confusion, the result of realist philosophers neglecting the fact they are represented by secondary qualities. A realist might suppose though that "surfaces are red if and only if they answer to a certain physical formula having to do with the emission of light waves" (?0.000075cm for the physicists) and "might take it that this formula tells us what redness is" . Although perhaps worthy of consideration, this interpretation will not do for the purpose of my essay, as I am considering redness to be a word associated with red qualitative sensations (this formula, or such-and-such an atomic configuration, may be universals, although as they lie beyond the scope of direct empirical experience they cannot be what the words red and redness directly refer to if these words are employed to express empirical experiences).
Associating redness with red qualitative sensations and not treating it as something that actually inheres within or is instantiated by 'red' object does not demolish Realism. In a possible alternative Realist ontology, it would be unnecessary that a quality universal inhere within or be instantiated by an object for that universal to objectively subsist. Rather, the 'vessel' in which redness instantiates itself is the actual qualitative sensation experienced by a perceiver. Basically, in this theory, a red quale is considered to be an instantiation of redness. So comments such as 'red objects instantiate redness' or 'red objects exemplify redness' can be converted to 'when I observe 'red' objects, a red quale instantiates redness. This speculation leads to me to consider a model suggested by David Armstrong, in which it is "the case that redness is distinct from the physical property of an object's surface, although exactly correlated with it" . So redness is distinct from, although exactly correlated with, the physical property of a 'red' object's surface. A relation can be derived from this model, involving redness the universal, an external 'red' object (and technically also the electromagnetic radiation emitted by the object) and a red quale. This relation may perhaps also be considered a universal. So for a ball to be 'red' is not for it to have the property of redness , but for it to have a certain physical property that is correlated with redness being instantiated in a red quale.
Following are some abstract functions for this model, devised to hopefully help me provide a more lucid explanation: fi is the instantiation function -fi(colour universal) = quale of colour fe is the emission function -fe(colour object) = electromagnetic radiation corresponding to colour fp is the cognitive perception function -fp(electromagnetic radiation corresponding to colour)
So fp( fe(colour object)) ? fi (colour universal) or fp( fe(colour object)) ? quale of colour
The use of material implication is important, as it does necessitate the observation of an object with a certain physical property associated with a colour quality universal, for that universal to be instantiated. This signifies that our model can accommodate 'brain in a vat' scenarios and more concrete examples such as scientific tests involving cerebral stimulation as well as the bizarre psychological phenomenon known as synaesthesia, a confusion between the senses, as when musical notes are experienced as coloured.
I shall also devise a class definition to help better understand a Realist's notion of the statement 'redness is instantiated by all red objects'. This definition requires two corresponding classes. The first class contains all 'red' objects that have been observed. The second class contains all red quale that have been experienced (instantiations of redness). For each member of the former there is some member of the latter to which it is associated.
A = {x | x is a 'red' object} B = {y | y is a red quale}
If redness possesses objective subsistence and is not merely a subjective mental abstraction, then I believe that for the class of all 'red' objects there will have to be one unique corresponding class of all red qualia and vice versa.
The broad definition that has been expounded may seem metaphysically speculative to an immoderate degree, though it is a deliberate attempt to exhaust a remaining, albeit extreme, possibility for a realist theory of quality universals. I shall now expound my reasons for thinking it untenable.
Now membership of this second class, and hence determination of the second class directly and the first class indirectly(although objectively it is a 'red' object that leads to a red quale, it is a red quale which makes us call an object red), is based on each element being a red quale. As qualia are ineffable, I need to rely on the reader's visual experience to better define this requirement for class membership. I shall provide a negative example. Excluding exceptions, when a person looks at a 'red' apple first then a green apple second, the red quale of their first experience is not similar to the green quale of their second experience. Therefore the first quale is not to be included in the same class as the second quale and vice versa. Let the equality operator in the first line be interpreted as meaning 'is qualitatively similar to'.
Red quale ? Green Quale Red quale & Green Quale
I shall assume for the following thought experiment the not unreasonable possibility of perceiver relative qualia, which as we will see poses a problem for the existence of a required unique second class. As the trite explanation goes, it is possible that when person A and person B both look at what is mutually referred to as a 'red' apple, the quale of person A's visual experience (quale A), is different to the quale of person B's visual experience (quale B). This difference is considerable, to the extent that if person A had experienced quale B, they would not identify it as red and vice versa. For qualia to be considered identical, it is imperative that they have a sufficient qualitative resemblance, not that they be merely related to the same external object. If both people look at a collection of red objects and agree that 'redness is instantiated by each object in the collection', their respective denotations of redness do not refer to the same thing and three classes result.
The first class contains all 'red' objects mutually observed by person A and person B, the second class contains each quale associated with each 'red' object for person A and the third class contains each quale associated with each 'red' object for person B. The second class (class A) and third class (class B) are not identical, as classes are identical if and only if their members are identical and this is not the case.
In accordance with our criteria, the universal denoted by is not identical to the universal denoted by . The first class, that of all 'red' objects, is the same for both person A and person B, but instead of there being a one-to-one relationship between this class and a unique class of corresponding qualia, there is a one-to-many relationship between this class and classes of qualia. If we are to select a unique corresponding class of red qualia, there is no objective reason for selection of the second or third class. Applying this though experiment more generally, the statement 'redness is instantiated by all red objects' implies the subsistence of one unique redness universal. If there does not subsist one unique redness universal, then it is not clear there subsists a redness universal, or that the subsistence of a universal can be inferred from such a statement.
A statement such as 'rednesses are instantiated by all red objects' would be problematic, least of all because of its grammatical improperness. It is also ambiguous. The plural noun 'rednesses' in this statement can be interpreted as referring to differing shades of red, whereas in this context it refers to completely different colours. Although several people may all experience different qualia when observing a 'red' object, it is acceptable in everyday use of language for them to assert that the object is red, as long as they are consistent in their identification. This is metaphysically problematic though as the physical property of a 'red' object cannot be exactly correlated with a redness quality universal. Hence there can be no rigid ternary relation universal involving a quality universal Xness, an external X object and an X quale, for any colour X.
Discussion so far has focused on the potential impossibility of an objectively subsisting quality universal, such as redness, that is exactly correlated with all 'red' objects. This is equally a possibility though, as the basis of my argument against real quality universals, namely perceiver relative qualia, is not beyond doubt. It was argued that under the assumption of perceiver relative qualia it is not the case that one unique class of 'red' qualia corresponds to the class of 'red' objects.
It can be argued with a greater degree of certainty that this condition does not hold the other way around. That is, for all 'red' qualia, there will have to be one unique corresponding class of 'red' objects. If both person A and person B were to state that 'redness is common to all red objects' and person B was afflicted with a partial color blindness that consisted of an inability to differentiate between what person A experienced as red and what person A experienced as green, so that person B only perceived the red, then these statements would not have metaphysically identical meanings. The class of all 'red' objects for person B (RB) would have a greater cardinality than the class of all 'red' objects for person (RA), therefore these classes will not be equal and it will not be the case that for all 'red' qualia there is one unique corresponding class of red objects. Would elements of the class RB\RA instantiate redness or greenness?
Although the arguments expounded in this essay against Realism are far from being irrefutable or even certain, I believe they are significant. This is because they can be more adequately addressed by Conceptualism and any Realist arguments against Conceptualism are not as cogent as the arguments against Realism contained within this essay. Furthermore, I do not see any advantage to Realist theories of quality universals over Conceptualist theories. Conceptualism can adequately address all the problems regarding quality universals identified in this essay. Conceptualism is a theory of universals that is midway between Nominalism (which is the one theory of the three I am least likely to consider) and Realism, and sees universals as existing in the minds of individuals, as concepts, which are grounded in experience. These concepts do not exist independently of human minds but are mental entities resulting from perceived commonalities in different objects . The initial two arguments acknowledged in this essay pose no threat to Conceptualism, as Conceptualism makes no assertion that quality universals possess an objective existence/subsistence. Furthermore, according to Conceptualism, humans, such as person A and person B, unconsciously create mental categories, which deal with a similar property abstracted from a set of objects. The identification of these categories or concepts with words or other symbols enables their cognitive manipulation. A statement such as 'redness is instantiated by all red objects', makes no assertion as to any real and objective ontological status for redness, but rather refers to a conceptual universal abstracted from the set of all objects that are associated with 'red' qualia. If completely differing qualia between two observers are consistently associated with the same objects, it is not problematic for these observers to employ a mutual noun to denote the universal associated with each respective quale. As such nouns denote conceptual subjective universals abstracted from a perceived commonality in these qualia, it is not necessary that the conceptual quality universal which person A denotes by redness be identical to the conceptual universal that person B denotes by redness. Meaning can still be attached to quality universal nouns for each individual, as they refer to something, namely mental categories. Furthermore, if we are to delve into the problem of universals, as conceptual universals are mental categories, they are more tangible than real universals and it is possible that by identifying a conceptual universal with a particular cognitive structure a discipline such as neuroscience can give us a more substantial and demystified theory of quality universals than metaphysical realist ontologies can.
In conclusion, I believe this essay to have sufficiently established that it is problematic to consider quality universals as having a real subsistence. Of two established arguments against Realism briefly considered in this essay, the second, known as the 'Borderline Problem', is quite cogent. It was argued that the physical sciences have undermined classical or extreme realist ontologies and the task to redefine an adequate realist ontology compatible with scientific interpretations of sensory perception was pursued. It was shown that this redefinition was merely an attempt to comply with the interpretations of physics and had not evaded the metaphysical problem of universals. The central argument of this essay was based on a hypothetical yet probable scenario in which it was shown to be very difficult or impossible to identify a unique colour universal, such as 'redness' as there was no definitive correlation with this universal and the objects which are associated with its instantiation (red qualia). A more satisfactory theory for quality universals was found in Conceptualism, which is better able to explain the presence of quality universals that appear to exist in a world of particular objects by treating these universals as conceptual mental abstractions with some kind of subjective existence. Conceptualism is able to adequately address the problems raised against Realism in the essay and is able to provide an explanation for the meaning attached to a noun denotative of a quality universal, an ability which initially seemed exclusive to Realist ontologies. In closing, to capture the conclusion of conceptualism it may be fitting, with regards to quality universals, to negate a line of Bertrand Russell's, and state that quality universals are not the object of thoughts, they are thoughts .