What is Formal Ontology

After briefly addressing the origin of the word ontology, I will ask what formal ontology might be. I will first argue that formal ontology is not the same as formalized or general ontology. In order to understand the meaning of the epithet “formal,” I will then look for a suitable way of distinguishing form from matter. Formal ontology will be shown to be formal insofar as it does not refer to concrete, singular entities. This implies that it cannot be the study of singular things, but also not the study of our knowledge of them. For in both cases, it would depend on empirical facts about concrete things. Rather, formal ontology is about the objectivity of objects, and in this sense, it is the study of the form that objects have only insofar as they are objects. 1. Ontology and its Title “Ontology” is a Greek neologism coined in early modern times. It is not difficult to grasp its general meaning: “on” is the present participle of Greek “einai,” which means “to be,” and “logos” derives from legein, “to talk about” or “to give an account of” something. Ontology is accordingly the discourse that has being as its subject matter. This is what Aristotle describes as first philosophy: “a discipline which studies that which is insofar as it is, and those features that it has in its own right.” But “ontology” cannot just refer to any discipline that studies being. Every philosophical or scientific discipline studies being, for each studies something that is. There are special sciences, such as physics or differential psychology, and special philosophical disciplines, such as ethics and logic. These sciences and disciplines study certain kinds of things that exist, insofar as these things exhibit certain special features. Two different kinds of restrictions are involved in circumscribing what a special science is. First, a special science may apply only to a limited 1 Metaphysics 1, 1003a21–2. Boris Hennig What is Formal Ontology? 27.08.2006 2 range of things, and second, it need not study every single aspect of the things to which it applies. Physics, for instance, studies the physical properties of everything that has such properties. Biology only studies living beings, and this only insofar as they are alive, not insofar they are sheer physical objects. Differential psychology studies human beings only insofar as they differ from other human beings in ways that are psychologically measurable. Two different special sciences may very well have overlapping domains; that is, domains that include the same members. For example, the claims of physics and chemistry apply to the very same things, except that the former investigates their physical properties, the latter their chemical properties. Ontology differs from such sciences as physics and differential psychology, but not by considering another special range of things. There are no objects studied by ontology that would not also be studied by at least one other discipline. Ontology studies a different aspect of those things. According to Aristotle, ontology is concerned with everything that exists only insofar as it exists. Existence itself is thus the aspect relevant to ontology. Hence, ontology will only be possible if there are features that every existing thing has only because and insofar as it exists. I will ask momentarily what sorts of features these may be. The objective of this section, however, is to give a preliminary impression of what ontology is by considering the history of the discipline and its title. Aristotle’s Metaphysics deals with questions of ontology. The title “ontology,” however, is much younger than this work. As a title for a philosophical discipline, “ontologia” has been in use since about the seventeenth century. Jacob Lorhard, rector of a German secondary school, uses this term in his Ogdoas Scholastica (1606) as an alternative title for metaphysics as it was taught in his school. However, he does not further explain the term. More prominently, the German philosopher Christian Wolff uses “ontologia” in 1736 as a name for the discipline introduced by Aristotle in the passage quoted above. The list of topics that Wolff discusses under this heading resembles the one given by Lorhard. It includes the notion of being, the categories of quantity and quality, the possible and the impossible, necessity and contingency, 2 The second edition appeared 1613 under the title Theatrum Philosophicum. 3 Christian Wolff, Philosophia Prima sive Ontologia, Leipzig 1736. Boris Hennig What is Formal Ontology? 27.08.2006 3 truth and falsehood, and the several kinds of causes distinguished in Aristotelian physics. This choice of topics certainly derives from Aristotle’s Metaphysics and such works as the Metaphysical Disputations by Francisco Suárez. We can gather some additional facts about the early use of the term “ontologia” by considering the first known appearance of the corresponding adjective in the Lexicon Philosophicum by Rudolph Goclenius (1613). A foray into his use of “ontological” will provide insight into how the term came to be used as it is, but as we will see, there are some important respects in which his usage differs from contemporary usage (and thus from the usage in this volume). Goclenius uses “ontological” in his entry on abstraction, where he discusses abstraction of matter. As everywhere else in his lexicon, he does not present a unified account of the phenomenon in question, but rather lists several definitions and other findings from the literature. In the present context we are not concerned with what Goclenius means by abstraction and matter, although the concept of matter will become important later in our discussion of formal ontology. Provisionally, matter can be taken to be the stuff out of which a thing is made. To abstract it from a thing simply means to take it away from that thing, in our imagination or in reality. For the time being, we are primarily interested in the sense in which Goclenius uses the epithet “ontological.” In science, he says, there are three different ways of abstracting matter from given things. First, one may think away the particular lump of matter out of which a given thing is made, but still conceive of the thing as being made up of some matter or other. According to Goclenius, this is what natural scientists do: they investigate particular samples, and they study their material nature. But they are not interested in one sample rather than another, unless the samples differ with respect to their general properties. In studying a particular diamond, for instance, scientists think away its particularity and consider only those of its features that any other diamond would have as well. Scientists abstract from a particular thing’s matter in order to grasp those general features of a thing in virtue of which it falls under a certain category; but the fact that things of its type are made of matter remains a factor in their account. This is what Goclenius calls physical abstraction. Second, we may think away all matter whatsoever from a given thing, in such a way that no matter at all will figure in our account of it. This kind of abstraction is practiced in geometry and Boris Hennig What is Formal Ontology? 27.08.2006 4 accordingly, Goclenius calls it mathematical abstraction. But he also calls it ontological abstraction, glossing the latter term as: “pertaining to the philosophy of being and of the transcendental attributes.” I will explain this phrase in due course. Finally, Goclenius continues, one may abstract matter from a given thing in reality as much as in thought. The result will be that the entity in question literally no longer possesses any matter. This Goclenius calls transnatural abstraction, of which he claims that only God and the so-called divine Intelligences are capable. There are at least three important things to note here. First, Goclenius identifies ontological abstraction with mathematical abstraction. He thereby implies that ontology in general, as much as mathematics, is concerned with abstract entities and formal structures. For instance, geometry is concerned with the properties that physical objects have only by virtue of their shape and location. Their other properties, such as color, weight, smell, etc. are irrelevant. In this sense, geometry abstracts from the matter that is shaped and focuses on the shapes themselves. Whether a triangle is made of iron or wood makes no geometrical difference. If formal ontology is abstract in the same sense, it should also abstract from certain properties of things and focus on their more general features. So far, we do not know what these more general features are that ontology is supposed to study. What is important here is that ontological abstraction goes farther than mere physical abstraction. The physicist is not interested in singular samples, but still in material things insofar as they are material. Ontology is not interested in matter at all. I will discuss later what that might mean. Here, however, we can already note that according to Goclenius, since concrete things are composed of matter, ontology is not concerned with concrete things at all, not even in a general way. Second, Goclenius equates ontology with the philosophical doctrine of the transcendental attributes. These attributes include being, oneness (or unity), goodness and truth. Being and oneness are mentioned by Aristotle; goodness and truth are introduced by later authors. These attributes are called “transcendental” because they apply to every existing thing, regardless of any categorial boundary. That is, they surpass (or transcend) the general categories which distinguish 4 Goclenius 1613:16 s.v. abstractio. 5 Metaphysics I, 1053b; cf. Aertsen et al. 1998. Boris Hennig What is Formal Ontology? 27.08.2006 5 things of different kinds. According to neo-platonic authors like Pseudo-Dionysius the Areopagite, the transcendental notions not only surpass the categorial boundaries between things, but also the realm of the things to which they apply; that is, the world. For instance, Dionysius writes that “the name being extends to all beings which are; and it is beyond them.” In other words, the transcendental attributes are manifested by everything in the world, but they do not apply to themselves. The transcendental attribute “being” is not itself something that exists. In any case, if ontology studies the features that everything that exists has insofar as it exists, then it will also be concerned with the transcendental attributes. Third, Goclenius does not use the epithet “ontological” in order to indicate something that really or actually happens. When we ontologically abstract matter from a thing, we do not really take away its matter. We do so only in thought. Real abstraction, by contrast, is what Goclenius calls “transnatural abstraction:” it occurs when God separates the human soul from its body. In this regard, Goclenius’ use of “ontological” is directly opposed to some of the contemporary uses of this word. When contemporary writers call something ontological, they often mean to indicate that it really obtains, or at least that it has implications for what exists independently of our thoughts. Ontology is thus often opposed to epistemology. The former is often said to be about what there really is, whereas the latter is only about what we know. A common view, for instance, is that ontology is concerned with the “level of things” rather than the “level of truths.” That is, ontology concerns objects in the world, not our thoughts as they are expressed in true propositions or true sentences. Goclenius, however, does not distinguish between things and truths. For him, both being and truth are transcendental attributes that apply to everything that is, on every conceivable level. Accordingly, he has no reason to suppose that it makes a difference whether we study things or truths, and ontology may be the study of both. In fact, Goclenius’ use of the epithet “ontological” differs from the modern one in all three respects that I have emphasized. First, ontology is no longer considered to be as abstract as mathematics. It does not abstract from all matter whatsoever, since it must also discuss the general features that things possess when they are material and particular. Second, at least in the 6 Pseudo-Dionysius the Areopagite, The Divine Names 5, 816B. 7 Smith and Mulligan 1983:73. Boris Hennig What is Formal Ontology? 27.08.2006 6 tradition of analytic philosophy, ontology does not include a treatment of such transcendental attributes as goodness and truth. These topics are dealt with instead within ethics and epistemology. Third, as we have seen, the opposition of real and ontological abstraction appears odd from a modern perspective. However, we will see that there is also some continuity between Goclenius’ and our uses of “ontology.” We will see later that ontology is still an abstract discipline in the sense that it avoids dependence on singular references. Further, the idea is still upheld that there are at least two transcendental attributes which surpass the categorial boundaries: oneness and being. Finally, we certainly would oppose ontological features and happenings to the transnatural ones; that is, features and happenings that surpass the realm of nature. 2. Some things that are not formal ontology So far I have introduced a rough notion of ontology, according to which it is the study of features that existing things have insofar as they exist, but not insofar as they are concrete objects consisting of this rather than that matter. Since ontology, conceived in this way, seems to abstract from matter in the same way in which mathematics abstracts from matter, all ontology seems to be formal ontology. But what is formal ontology? Edmund Husserl, who introduced that term into philosophy, describes it as the “eidetic science of the object as such.” Eidetic derives from Greek eidos, which means form. I will therefore approach Husserl’s formula by means of a (necessarily brief and sketchy) discussion of the general distinction between matter and form. This will lead to a discussion of judgments and their objects, thus enabling us to understand the second part of Husserl’s description (“the object as such”). Husserl deliberately uses the term eidetic instead of formal, because he wants to avoid misleading connotations. He is well advised in doing so, since there are at least two common – and mistaken – accounts of what it means for a discipline to be formal. First, a discipline is sometimes called formal merely because its claims are expressed by 8 Edmund Husserl, Ideen zu einer reinen Phänomenologie §10, Husserliana 3/1: 26–7. 9 Op. cit., Introduction, p. 9. Boris Hennig What is Formal Ontology? 27.08.2006 7 means of formal symbolism or even only a shorthand notation, as when one writes “ x:MAN(x) MORTAL(x)” instead of “all men are mortal.” However, shorthand notations are merely short, and sometimes not even that. There is no particular reason for calling them formal. Logic and mathematics are indeed formal, and they also often use shorthand notations. But logic and mathematics are not formal because they use this kind of symbolism. For one thing, mathematical and logical truths can be expressed perfectly well in prose, although this would often take up more space. For another, any old body of knowledge can be expressed by short and rigorously defined symbols without thereby turning into a formal discipline. Logic and mathematics are properly called formal only because they are about formal structures and features, for instance those of shorthand symbolisms. Hence, formal ontology may indeed use symbolic shorthand notations as far as they are helpful. But it need not do so, and it will not be formal by virtue of doing so. Second, formal ontology has sometimes been opposed to “regional ontology.” Both labels, formal and regional ontology, were introduced by Husserl. Roughly, a regional ontology is a cognitive partition of the world, much like a special science: it gives accounts of the nature of a given, limited range of entities. There are separate regional ontologies for the domains of physics, biology, differential psychology, and so forth. It has been claimed that formal disciplines are “set apart from ‘regional’ or ‘material’ disciplines in that they apply to all domains of objects whatsoever, so that they are independent of the peculiarities of any given field of knowledge.” According to this view, formal ontologists should only advance judgments that hold true of all objects in general. This is not far from the truth, but some qualifications are in order. For it is not the case that every claim that is made within formal ontology applies to everything that exists. Formal ontology can also study the formal features of a limited range of entities, in the same way in which geometry can study the shapes of a limited range of entities. In order to show this, I consider two quite different ways in which judgments may refer to specific entities in the world. (1) Judgments like “Marlene Dietrich was beautiful” or “that child over there is intelligent” 10 Husserl, Ideen ... §9, Husserliana 3/1:24. 11 Smith & Smith 1995:28. Boris Hennig What is Formal Ontology? 27.08.2006 8 are singular judgments. Singular things are concrete, discrete, and they exist only once. Singular judgments refer to such things by using proper names or demonstrative expressions like “Marlene,” “this,” or “over there.” Further, their truth depends on the state of exactly those singular things to which they refer. (2) The other way in which a judgment may depend on peculiarities consists in its being particular. Particular judgments hold true only of a limited range of entities, such as the judgment that some actresses are beautiful. This judgment holds true of actresses, and not of other persons or things. The truth of a particular judgments also depends on the state of individual and concrete things, but only of anonymous ones, as it were. Particular judgments do not refer to singular things by using a name or demonstrative. Although they do not apply to everything in general, they still refer to their objects by means of a form that can be specified in a general way. The distinction between singular and particular judgments is important because it will turn out that formal ontology may rely on the use of particular judgments, but not on the use of singular judgments. The point is thus not that formal ontological insights apply to all objects alike, but rather that they apply to certain ranges of objects that may be specified in general terms. Whereas formal ontology must not refer to singular beings like Marlene Dietrich or that child over there, it can still refer to particular entities like organisms in general, or anonymous children and actresses. Hence, formal ontology need not be independent of the specific entities within a limited domain of knowledge, as long as it does not depend on singular judgments and references. Ontology is formal as long as it picks out and applies to particular entities solely by means of their general form. Formal ontology is thus not the same as general ontology, and hence it is not opposed to regional ontology. Rather, ontology may be a formal and regional at the same time. A regional ontology deals with a limited range of entities, but as long as it does not advance any singular judgments, it can still deal with them in a sort of general way. For instance, the ontology of events studies only a limited range of entities, namely those that occur or unfold in time, but it does not study singular processes in particular, such as the death of Socrates or the great 12 We adopt the terminology used by Kant, Kritik der reinen Vernunft, B95. Boris Hennig What is Formal Ontology? 27.08.2006 9