REFERENCE AND ESSENCE:

A Response to Devitt and Sterenly’s Language and Reality

The Naturalistic language model

If a dog bit me, and I wished to communicate this event, I would most probably construct an English sentence like “That dog bit me.” According to the “naturalistic” philosophy of Devitt and Sterelny, ‘dog,’ ‘bit,’ and ‘me’ all “stand for” something in the real world, and the meaning of the specific utterance “That dog bit me” could be determined by assembling the parts into a truth condition about the world. Their philosophy of language is based on the notion of reference, which provides a direct link between words and the real world objects they represent. It is the dog and the bite in reality that underpin the statement’s meaning and guarantee its truth-value.

How does reference provide language users with a link to reality? Let’s explore the word ‘dog,’ a common noun that Devitt and Sterelny would also call a natural kind term. A natural kind is a thing that exists in nature, as opposed to an artifact, which is manmade. Natural kinds are discovered, and we create terms for them.

The process of introducing the term ‘dog’ into the language reminds me of Adam giving names to the animals. Let’s say that Adam is in contact with many of God’s creations in the Garden of Eden, and he begins to notice certain surface distinctions among the things he experiences. He calls some things that appear similar ‘dogs, ’ thus grounding the reference in these things. Cain and Abel hear Adam use ‘dog’ to refer to these things and they borrow the reference. This process of reference borrowing continues until I finally use the word ‘dog’ in the sentence “That dog bit me.”

The extension of the term is then all those objects, or all those examples of stuff, that are the same kind as the ostensively given samples, that share the underlying essential nature of the samples.

(D & S, pg. 88)

Why did Adam notice similarities among dogs? Devitt and Sterelny say that all dogs are ‘dogs’ by virtue of an essential nature. In other words, dogs have various attributes that allow them to be isolated and indentified as such. Adam’s perception of the natural category dog led to his creation of the word ‘dog.’ In addition, references are often grounded multiple times by further individuals who are able to isolate the factors that determine what makes something a ‘dog.’ Through scientific observation, we have discovered that dog is an entity with a certain “evolutionary origin.”

In other words, since Adam, we have always sensed distinctions, and have created words to describe specific things as natural kinds. Nowadays, science has the capacity to justify this claim by revealing a specific genetic makeup that makes a dog a dog, and so we are able to refine our reference to insure that our linguistic representations are accurate. For example, after modification of reference, we can talk about black swans even if our original sample was composed of only white swans.

This philosophy is founded on a specific worldview—that language is a human phenomenon that reflects facts about the world; words are necessarily representations that emerge from encounters with the world. This philosophy of language is built on an underlying ontology that assumes the world is capable of being described via words. They believe that the world consists of distinct entities with a set of inherent characteristics. In other words, their philosophy of natural kind terms presupposes the existence of natural kinds, with essences. If their theory of essence is correct, then their theory of reference follows. The question then becomes, how valid is their belief in essence?

My dream model

This model of the world, although it has science and some ordinary experience on its side, still cannot leave me free from the doubt planted by the possibility that this worldview is incorrect or at least dangerously incomplete.

I had an absurd kind of dream, which described a world very different from our own, and despite its apparent absurdity, it cemented this feeling of doubt in my mind. In my dream, I awoke in a world with no distinct substances, or essences. Instead, what I saw appeared continuous and indistinct. I found myself terribly confused and confounded by an inability to describe my experience. In the midst of this frightening and inexplicable world, I was surprised to discover a strange pair of glasses in my pocket, a grid superimposed on the surface of the lenses. To my delight, as soon as I placed these lenses over my eyes, I was given a sense of confidence that the previously obscure amorphous field was now a collection of shapes, each clearly identifiable if I defined them within this new geometric system. In this dream world, I spent my time discovering new shapes, and creating “constellations”—listing the series of coordinates that each shape seemed to fall within. Some of these shapes were other people, like me, wearing glasses, and yelling out coordinates to each other. I can only conclude that they were wearing similar glasses, as they shared my geometric system, and we quickly formed a consensus about forms.

I was glad that a world once so inexplicable was now available to me in the form of specific units of information (a series of coordinates). With a clear understanding of this world cemented in my mind, I decided to take off the glasses. To my surprise, the world, made up of seemingly unique components, once again reverted to the same vast expanse of formlessness as before. I was now unable to perceive the objects I held so much belief in.

Comparing the two

How does this dream model compare to the previously clear, naturalistic theory of reference? Both systems reveal attempts by humans to understand and describe the world around them. The inhabitants of my dream world scenario share grid glasses, which act as a filter for reality, and allow them to create a comprehensible world of things out of the void on a seemingly arbitrary basis. Devitt and Sterelny’s naturalistic philosophy also seems to posit language as an intermediary filter between the world and humanity, but with arbitrary words standing for actual things in reality.

In other words, the only difference is that whereas my dream’s model is hopelessly subjective, Devitt and Sterelny’s model contains external essences, which allows for scientific objectivity. What if the inhabitants of the dream world, thousands of years in the future, after my set of coordinates had entered into common usage, had to defend their own system? Their system, formed on a basis of consensus about forms due to a shared mode of interpretation of the world, would be as consistent as ours. Members of this system couldn’t help but see the world in terms of the objects that had once been imposed. It is the only reality they had ever known. Of course we know it was a biased view of reality, and what if we pointed out to a participant the arbitrariness of the system and inherent subjectivity underlying it?

Suppose a defender of the system went one extra step, and claimed that the objects in reality themselves justified their reference.

In fact, Devitt and Sterelny’s insistence on the existence of essences (water is fundamentally H2O) reveals a similarly overt attempt to defend a particular, scientific mode of interpretation, against this threat of arbitrariness. According to Devitt and Sterelny, it is not enough that things exist in natural categories—they believe that things are ultimately distinct and identifiable due to their “evolutionary origin” or “chemical composition.” Any such point of view which privileges one true quality of things over another can only reveal a bias toward one worldview. And, like the man who proclaims that beauty is truth, and truth beauty, saying that science is truth ultimately reveals only the inter-subjectivity of a specific, historically grounded bias of apparent objectivity. There may be little pragmatic difference between an objectivity formed by the consensus of observers and an objectivity grounded in truths about the things-in-themselves, but the ontological difference is vast.

In the end, we are left with a multiplicity of possible natural kind terms that can be defined through any number of constructed modes of interpretations—ways of slicing and dicing reality to suit our needs, whatever they may be. These modes of interpretation are what we pass down through language when we borrow references. Essences act as the touchstones of linguistic order that we leave in the world in an attempt to comprehend it, defining the things of our experience in terms that can be understood. After all, the world was not created with our descriptions in mind. Rather, if we are to describe the world, we must do so by constructing a coherent system, and essences impose this specific order on the variety of real world objects we choose to describe. If we choose to believe that these essences exist in the world rather than in the language, we are led simply towards a more limiting view of the world. After all, if the world can be seen in many ways, rather than one, isn’t the reliance on the bias of one kind of essence leading us away from truth—the more inclusive understanding of the world that may be gained by other modes of interpretation, or at least by questioning the basis of our own?

Restoring objectivity

My dream of an illusory world with order imposing glasses led me to question whether our own language system was as hopelessly subjective. After careful thought, I have determined this cannot be so. Let’s begin again with the dream model, which I will then contrast with some realistic examples of “language in action.” These examples, which draw from causal reference theory, will help to assuage the concern that the only distinction between our language and the glasses is that one may be removed.

In the dream model, I created forms by listing a series of coordinates on a grid, thus defining reality arbitrarily. This system was refined by a gradual consensus with other glasses-wearing observers based on our similarly filtered vision of the world. When I removed my glasses, the world again appeared indistinct and without form, because all the essence and substance in this world was contained in my own mental constructions, which I had justified with the help of an apparently consistent and shared geometric system.

How does this system compare to ours? Our system provides many examples of objective verification that the dream model cannot emulate. For example, let’s say that a painting is discovered in someone’s attic, and there is controversy over the identity of the painting’s creator. In this case, if the painting is a genuine work of a Dutch master, it may be valuable. As is often the case, an art expert is flown in, as well as a scientist. The art expert believes (based on his considerable knowledge of the Dutch masters) that the painting may indeed be valuable—a newly discovered, hitherto unknown Rembrandt. The scientist performs chemical tests on the work, and determines that it cannot be a Rembrandt, because it contains certain fibers not used to manufacture paint canvases until the 19^th century. In this case, the art critic would most likely concede that the painting is not, in fact, a Rembrandt. Both the art expert and the scientist, coming from separate disciplines, know the criteria that would make something a Rembrandt, and can agree on what would consist of proof of Rembrandt-ness. If the art expert does not concede to the fact that the painting is not a Rembrandt due to this scientific evidence, he runs the risk of being called a liar, or worse.

If the chemical test reveals that the painting was in fact from the 17^th century, it would still not determine whether the painting was in fact a Rembrandt. In this case, the art expert would step in, and he would begin looking for aesthetic hallmarks rather than chemical compositions.

By attempting to determine if this painting is indeed a Rembrandt, the art expert and the scientist both reveal the main goal of language: to speak truthfully about the world. Of course, with this example, the beliefs of the art expert and scientist are not enough to determine Rembrandt-ness. The truth-value of the art expert’s statement “That is a Rembrandt,” can be only be verified through observation of the painting itself. In this case, various tests may be able to determine certain historical conditions that would enable a more or less convincing confirmation of Rembrandt-ness. The value of the painting would reflect the degree of certainty one way or the other.

Of course, this example is a debate between two experts, who bring plenty of knowledge about Rembrandt to the table. Many would not know the first step in attempting to identify a Rembrandt, and thus one decision among experts may well appear as true as the next.

However, Devitt and Sterelny’s causal theory of reference negates the “Cartesian assumption” that we must carry around knowledge of the thing with us. This was formed as a rebuttal to a pure-descriptive theory of reference, which can more easily be taken to the extreme of my dream example. In description theory, definitions are formulated to correspond to objects, and then we identify objects by somehow matching them to these descriptions. Like the dream example, description theory seems to place criteria for verification in the lenses of the observer.

Devitt and Sterelny point out that language simply doesn’t work this way. If our definition of tiger, for instance, included a clause about having four legs, and we came across a three-legged tiger, we would be forced to say the thing was not a tiger. However, this does not happen. Our language continues to refer to objects even as our descriptions change. A scientist and a little girl can both refer to a tiger, though they doubtless have different descriptions in mind. This could only make sense if words referred to objects rather than concepts. In fact, we can all speak about the same objects in different ways.

Thus, there must be something unchanging in objects that allows us to continue to observe and refer to them even as our understanding of them changes, and matures. Here’s another example from the world of art that illustrates this point. Let’s say that another painting is discovered in yet another attic. In this case, it is a genuine painting by Cézanne that was hitherto unknown. This new painting would now be included in the recognized oeuvre, or body of work, of the artist. If the extension of the word oeuvre is all known works by a particular artist, then this new discovery represents a new grounding of the reference of the term. According to Devitt and Sterelny’s causal reference theory, users of the term oeuvre needn’t have specific knowledge about the specific composition of Cézanne’s oeuvre, or even to be able to identify Cézanne’s paintings from a post-Impressionist line-up. Even if a user first borrowed the reference before the new painting was discovered, the painting would be included in the extension of his use of ‘Cézanne’s oeuvre’ because the phrase refers, not to a concept, but the kind itself, which always referred to the painting by virtue of the historical fact that Cézanne painted it.

Natural Kind Terms

The art examples above are not equivalent to the examples in the first paper, because they deal with artifactual kinds rather than natural kinds. People did have a lot to do with the creation of artifacts, and thus are ultimately responsible for the criteria of reference, which can be summed up with a definition, like “A Rembrandt painting is any painting created by Rembrandt.” Therefore, we can be infinitely more certain about our knowledge of Rembrandts than our knowledge of tigers.

How can we be sure that that the grounding of natural kind terms takes a true essence into account in the same way as the groundings of Rembrandt and Cézanne’s oeuvre? Artifactual kinds are easier to account for because of their reliance on description and the plausibility of the grounding situation. It is more difficult to imagine the discovery of a new natural kind. What was the situation that prompted the introduction of the word “tiger” in the language? What was the experience of the object like before the name had been grounded—before we grasped significant knowledge about its nature?

The grounding of natural kind terms can also be understood with the help of an example. Let’s take the phenomenon of AIDS, an interesting example of a natural kind grounded only recently, and thus a situation in which the entire process of grounding, up until this point, has been carefully documented.

Approximately 20 years ago, a group of doctors identified a new pattern. Identifying a set of common symptoms in a group of seriously ill young people, they sensed the common presence of a previously unknown disease. Isolating the known facts about this disease and the individuals it seemed to affect, they labeled it GRID, or Gay Related Immune Deficiency. Some people referred to it as the “gay cancer,” as it seemed as kill only homosexuals. Some people referred to it as “God’s wrath” for the same reason.

The disease eventually became known as AIDS (Acquired Immune Deficiency Syndrome) when people realized that it didn’t only affect gays. Someone isolated the virus HIV through scientific research and determined that it was the cause. Through additional research, mostly at the microscopic level, scientists discovered more and more about the circumstances leading to the epidemic.

Ten years later, we have many terms that derive from the newly discovered natural kind HIV. People use it in general conversation, due to its political and social importance. We talk about being HIV positive or negative, we use phrases like “full blown AIDS.” Everybody knows what HIV refers to, but not everyone knows what HIV is.

When we use the phrase HIV in a sentence, we don’t have to be experts of disease, we are simply borrowing the reference from the grounder of the term. In this case, based on a limited sample of gays, the grounder even had false information about the true nature of the disease, but he was still able to ground the phrase in the natural kind due to the necessity of the situation. This case reveals how a phrase can be multiply grounded in order to identify the true essence of the natural kind. Our initial conceptions of essence were no doubt as often wrong as in this case. When we go back and observe the discovery and initial research of the doctors and scientists who first grounded ‘AIDS,’ it becomes clear that although our conceptions of GRID/AIDS/HIV no doubt changed over time, the natural kind HIV has changed very little.

Conclusion

In order for our language to operate as it does, it must be objective, and distinct from our collective or individual subjective thoughts about things.

Language assumes that we can disagree on the causes of things and their accurate descriptions, but that we must agree on the objects of discourse, and on what would constitute a true statement. A true statement is one that can be empirically verified as being in the thing-in-itself.

In the first example, I mentioned that the art expert would be forced to concede based on the scientific research that the painting was not, in fact, a Rembrandt. If not, he would be deemed a liar. Eventually, we may doubt that he even knows what a Rembrandt is. Likewise has been the fate of those that have insisted that AIDS is “God’s wrath.”

Language assumes a common quest for the true essences of objects, and neither the art expert nor the religious fanatic fulfill this requirement, they rely on descriptions that cannot be empirically verified.

Through the process of grounding, one version of reference always emerges over another, and this is the version that most closely matches the true essence of the thing that it describes, rather than our concepts. This can only mean that language is truly objective, saying things about the world, and not simply about consensus, or the infinite multiplicity of possible subjective interpretations.

The objectivity of natural kinds

In Language and Reality, Devitt and Sterelny formulate a comprehensive theory of language, cemented in objectivity by fixing reference to the outside world and not simply our subjective interpretations. Their theory of reference is based on an ontology—the notion that things in nature have essential characteristics that allow us to classify and name them. By observing a group of swans, for instance, we are able to identify the existence of a natural kind swan and ground the natural kind term ‘swan’ to the species. According to Devitt and Sterelny, we do with the knowledge that we are grounding this particular kind of term.

As our knowledge about the natural kind swan increases, we are able to gain a truly accurate understanding of the species’ biological structure, and can determine the truth about “swan-ness.” The natural kind terms is itself fixed to the kind and its essence, not to the individual swans, with their individual characteristics. Thus, ‘swan’ has the ability to refer to black swans even if all swans present at a grounding of the term were white—language has the power to isolate true forms in the nature and allows us to refer, despite some ignorance of the true nature of the forms themselves. Our words continue to refer to the things as they really are.

Devitt and Sterelny take language out of the subjective realm and firmly place it back in the realm of objective reference. However, this evaluation of language, if true, prompts many questions about the nature of language itself.

Primarily, if natural kinds terms are as a fundamental a part of language as Devitt and Sterelny suggest, how does our reliance on natural kinds manifest itself in our social and political culture? Does the grounding and subsequent borrowing process play itself out as we might expect from Devitt and Sterelny’s explanations?

Looking back at the outset of the AIDS phenomenon certainly provides us a situation in which a natural kind term was grounded, thus creating a context for talking about the world. This term, ‘HIV’, clearly identified a virus as the cause of the epidemic. It linked directly to a thing, allowing us to speak of the epidemic without judgment. Before it was identified as such, people were confused and terrified by the effects of this virus, and they used other words to describe it, based on a wide range of explanations for the events we now know it caused. They called it GRID, or Gay Related Immune Deficiency, for instance, or “God’s revenge.” However, ‘HIV,’ grounded in the thing-in-itself, does not directly refer to these unfounded beliefs, and thus reveals the objectivity that natural kind terms provide.

In this case, the grounding of the term helped us on the road towards fully understanding the disease. Now that we can use ‘HIV,’ we are in a better position as a society to formulate beliefs about the AIDS epidemic, as the truth value of these beliefs can be verified through empirical observation.

Complications of natural kind terms

Take another example, the natural kinds measles and mumps. These kinds were once no doubt as misunderstood as HIV until we grounded them as natural kinds, providing a context for understanding them. The germ theory of disease was a first step in identifying a biological basis for the many epidemics that strike society, and measles and mumps are an example of how our knowledge about these natural kinds can accumulate, making us more capable of controlling their spread.

One technique, in the case of measles and mumps, was to encourage laws forcing immunizations upon each child before he was allowed to enroll in school.

Doctors and scientists understood the statistical probability of someone contracting the diseases and decided that the total benefit to the society as a whole was worth the risks of immunization upon the individuals. In this case, it seems that the classification of natural kinds, with their biological essences, gives the experts of biology an advantage over the general population. They are who we look to for the definitions of our terms such as ‘measles,’ ‘mumps,’ and ‘AIDS.’

In cases such as diseases, natural kinds terms were not necessarily grounded as such to begin with, despite Devitt and Sterelny’s insistence. Instead, it seems more probable, as with the germ theory of disease, that this understanding sometimes came later, a term being re-grounded as we finally discovered biologically or chemically classifiable natural kinds. The case of ADD provides an example. Nowadays, ADD (Attention Deficit Disorder) is in the DSM-IV and is classified as a mental disorder, with a classifiable, neurological basis. Doctors used to treat the observable symptoms of this disorder, but now they see the symptoms as signs of a specific imbalance, which is what they attempt to treat. In this case, the term ADD, once referring to a set of traits, was re-grounded as a natural kind—a particular brain problem.

However, in this case, it seems that the behavior-based disorder was carried over to the description of the re-grounded natural kind, as if to give doctors a biological basis to justify their treatments despite a relative ignorance about the causes and outcomes of the brain “imbalances” they linked to the previous category through empirical observation.

Teachers currently believe that children diagnosed with ADD will have difficulty performing in a school environment, and many encourage parents to administer powerful drugs like Ritalin to their children to negate its effects. If a simple test could be performed to determine whether a student has the brain “imbalance” that causes the behaviors, would school districts, with the scientific and medical disciplines on their side, require students with ADD to take Ritalin, as in the case of those other natural kinds, measles and mumps?

How about another long perceived behavior disorder—the phrase ‘homosexual’? Traditionally the homosexual has been seen as a social type, like bachelor. It has referred to individuals that engage in amorous relations with the same sex, and was usually seen as a behavior disorder. Now, some scientists believe that there may be a gene that determines sexual preference, thus re-grounding the term—changing it from an “un-natural kind” to a natural kind. Or would it?

According to Devitt and Sterelny, if we discovered an animal that looked suspiciously like a unicorn, it would not be a unicorn, it would be a new natural kind. It seems to me, however, that this supposition does not come close to describe the situation that would actually occur. Newspaper headlines would most probably state something like “Mythical creature turns out to be real.” Nothing could stop individuals from referring to this new creature as a ‘unicorn.’ Likewise, nothing would stop us from referring to a brain imbalance as ‘ADD’ and the presence of a specific gene as ‘homosexual.’

This complication, which Devitt and Sterelny dismiss in a simple sentence, seems immense. Not only can terms be re-grounded as natural kind terms, thus negating the notion that natural kinds are uniquely perceivable as such, but the original descriptions surrounding the previous uses of terms would no doubt remain. Thus, ‘ADD,’ re-grounded as a natural kind, is still a group of behaviors as well as a brain imbalance. ‘Homosexual,’ re-grounded as a natural kind, no doubt still refers to sexual practices as well as a specific gene. Both descriptions are sure to be carried through the language, thus thwarting Devitt and Sterelny’s hybrid theory, which attempts to negate the Qua-problem by simply adding a descriptive element to a causal reference theory. They simply do not address a case in which the same term is used to refer to two kinds, as in the case of ‘unicorn.’

Imagine that scientists could test for the homosexual gene, now assumed to be the root of homosexual behavior. In a society in which homosexuality was shunned (most societies fulfill this criteria), we could now have decisive proof of what determines homosexuality. Most societies engaged in pseudo-scientifically established racism have attempted to quantify the exact specifications of ‘Jew’ or ‘black’ in similar ways. However, the problem of classification becomes infinitely more complicated in this case, as individuals with the gene may have never engaged in ‘homosexual’ activities.

In this case, what would ‘homosexual’ refer to? The scientific, biologically based natural kind would not match up with the social group under attack. In this case, it seems that like ‘unicorn,’ Devitt and Sterelny’s philosophy has no real-life answer.

Above this discrepancy in theory, the most apparent problem of these examples seems to be the filter of objective distance that biologically based essences place upon our view of the world. By calling ADD a chemical imbalance, it takes the problem out of society’s hands. We don’t need to know what causes it, only how to properly medicate its victims. Likewise, calling homosexuality a natural kind has the possibility, like HIV, to promote an understanding of its true nature among individuals, though, like measles, it may promote some to attempt to “cure” it scientifically.

What these complications reveal is that a language formed on biologically based essences gives “experts” control over natural things. Putnam’s “linguistic division of labor” may help proliferate references, but it does not question the outcomes of this socially enforced classifying machine, which seem to ground understandings based on facts about the real world, but can then manipulate these facts in a variety of ways.

This leads one to question the true basis for grounding natural kind terms. If Devitt and Sterelny are correct, and we intentionally ground natural kind terms as such, there may be motives behind these groundings that reflect certain points-of-view. The scientist who looks for a biological basis for ADD finds it. Perhaps the problem with language isn’t that it is subjective, rather that its objectivity can be used to manipulate truth.