Introduction: The Methods of Ontology

Gregg H. Rosenberg

[1] Ontology is the study of what exists, with special attention paid to the different ways of existing possessed by different kinds of things. One obvious way of going about doing ontology is science. Scientists gain information about what exists by interacting with and observing different features of the world, and searching for the best explanations of the information they gain that way. This method of studying what exists is empirical and largely pragmatic. It lends itself to a practice that places a premium on operational definitions, and competing explanations that lead to differences in possible observations and interactions. The successes of science are the envy of philosophy, and they naturally raise the question, "Does philosophy have a distinctive contribution to make to ontology?"

[2] Answering this question is not easy, and it sometimes leads to internal conflict among philosophers. Part of the problem is deciding what is left for philosophical ontology to do, outside of science. There are, of course, the meta questions about science itself. Does even science yield knowledge about ontology, or merely instrumental fictions? These debates may be interesting, and they soon yield to subtle philosophical issues, but they are no comfort to the philosopher who wishes to do ontology. If science does not yield ontological insight about its subject matters, it seems unlikely that philosophy will. Alternatively, if it does yield ontological insight, then it is not clear what insight of its own philosophy may add.

[3] The subjects left for philosophy to cover seem very elusive. Pervasive topics remain, such as causation, that seem assumed by all the sciences but explained by none of them. Ancient topics remain, such as mathematics, that are used by science while seeming to outrun it. Elusive topics remain, such as modality, that are not clearly either subjective or objective. Controversial topics remain, such as ethics, that are not clearly in the domain of the ontologist at all. Ephemeral topics remain, such as consciousness, that should be part of science and yet somehow present special difficulties. And there still remains the overarching puzzle of how all these different ontologies, the scientific and non-scientific alike, exist within a single world.

[4] How can philosophers even begin to approach such topics? Through what avenue can the philosopher discover whether these things exist, much less study the nature of their existence? One answer to this problem is the skeptical one given by Carnap, and discussed by Huw Price in his essay, "Carnap, Quine, and the Fate of Metaphysics." Price defends the Carnapian anti-metaphysical view that all ontology is relative to a conceptual framework. From within a framework, one may use an ontology but only question it in certain limited ways. From outside of a framework, one may merely ask whether or not the framework is useful for one purpose or another. According to Price's version of Carnap, the philosophical ontologist makes the mistake of trying to discover truths, say about mathematical entities, by stepping outside the framework in which truths about those entities may be meaningfully investigated. Quine is widely thought to have undermined Carnap's thesis by showing that one cannot individuate conceptual frameworks, forcing Carnap's local pragmatic decisions into a global arena that includes metaphysical questions. By appealing to Ryle, Price defends Carnap's view against Quine, suggesting that the existence of individuated conceptual frameworks hinge on empirical questions about linguistic function. If different functions exist, these functions may delineate different conceptual "worlds". Price suggests that the places to look for this distinct functioning are the traditional M worlds of contemporary metaphysics: meaning, modality, morality, and mind.

[5] Roger Gibson, in his paper "Quine on Matters Ontological", performs the service of more fully articulating Quine's view of the matter. Gibson focuses especially on one puzzling feature of Quine's view: Quine champions a preferred ontology, the physical, at the same time that he champions a structuralist attitude of ontological indifference. How, Gibson asks, can Quine have it both ways? Gibson engages in a succinct and fascinating discussion of Quine's views on how we come to posit entities, on reification, and on his famous dictum, "No entity without identity". This discussion leads to an explanation of how Quine's naturalism yields an ontology of epistemology, as well as an epistemology for doing ontology. This ontology of epistemology requires the entities of science as a preferred basis, but only tentatively and pragmatically.

[6] Like Huw Price, Jill Dieterle endorses a framework relative view of ontological commitment in her contribution to our special issue, "Julius Caesar and the Number 2". Unlike Price, Dieterle is optimistic that one may still gain philosophical insight into ontology even once one has conceded this sort of pluralism. Using Frege as her starting point, Dieterle approaches existence questions through language, endorsing Frege's thesis that a saturated expression in a true sentence achieves reference to an object. She proposes to avoid the contradictions that plagued Frege's program by examining more closely the notion of 'object', restricting the property of "objecthood" to conceptual frames. Something may be an object from within one frame, and not be an object from within another frame. Her project, then, dovetails with Huw Price's in that they both need to clarify the identity conditions for conceptual frames.

[7] In "Ontological Categories and How to Use Them", Amie L. Thomasson takes on the ambitious task of actually providing a framework for systematizing the multifarious ontologies that we use. She focuses on the wide variety of dependence relations that hold between different kinds of existents. She erects a skeleton framework for mapping these relations, and sketches a way that plotting these relations may help clarify ontological issues. She explains how we could see that a variety of common philosophical dichotomies actually are ends of a continuum, if only we had a proper map.

[8] Taking a more focused view, Barry Smith and Achille Varzi provide a model of how to do ontology in a very specific domain by investigating the formal ontology of boundaries. Departing from Dieterle's linguistic model, Smith and Varzi appeal to the phenomenology of our experience of ordinary objects, and of regions of space, to address paradoxes about the nature of boundaries. What are boundaries? Are they parts of the objects that have them, and what relation do they have to those objects? If every object has a boundary, does that prohibit any two objects from actually making contact, as their boundaries will always be in the way? Approaching the problem from the perspectives of topology and mereology, Smith and Varzi formalize a logic of objects and their boundaries. With appropriate definitions in place, they argue that we may begin to clarify the nature of their existence and their relations to their hosts.

[9] Finally, our discussion piece, "CYC: A Case Study in Ontological Engineering" by B.J. Copeland shows why the study of ontology, and the methods of ontology, is more than a merely academic exercise. The engineering pursuit of intelligent systems requires dealing with the world in an intelligent way, and that requires an equally intelligent approach to the world's ontological richness. At this date, the most ambitious attempt to build an intelligent system is the huge computer knowledge base called CYC, a project spearheaded by Douglas Lenat and R.V. Guha. Copeland examines the way that Lenat and Guha have tackled the ontological questions that CYC has had to confront, convicting CYC's creators of accepting hasty and inadequate resolutions. Copeland argues that the ontological naivete of CYC will be its undoing as it grows, as a proper understanding of ontology is crucial to the proper operation of intelligence.

[10] I hope that readers will consider, with me, the variety of viewpoints implicit and explicit in the contributions to this issue. The plurality of methods here presents a problem as severe for ontologists as does the pluralism in our ontologies. Could Quine even communicate fruitfully with a philosopher like Dieterle, or philosophers like Smith and Varzi? Similarly, Price's Carnap seems to arrive at a similar destination as Dieterle, but their different starting points lead to vastly different characterizations of that terrain. Carnap claims the pluralistic terrain vitiates the search for philosophical insight into ontology, while Dieterle claims it yields philosophical insight after all. How does this sort of disagreement really arise, and, more importantly, how may it be adjudicated? Thomasson has conceived of an interesting, overarching framework for ontologists to use, but which ontologists? If ontologists are to flesh out the skeleton she provides, they must make many difficult decisions about specific cases. How are these decisions to be made? These are the issues that need more thought, as philosophers need to better understand how it is that philosophy may do ontology.



1997 Gregg Rosenberg

EJAP is a non-profit entity produced at Indiana University. EJAP may not be printed, forwarded, or otherwise distributed for any reasons other than personal use. Do not post EJAP on public bulletin boards of any kind; use pointers and links instead. EJAP retains the right to continuously print all the articles it accepts, but all other rights return to the author upon publication.

EJAP, Philosophy Department, Sycamore Hall 026, Indiana University, Bloomington, IN 47405.

http://www.phil.indiana.edu/ejap/1997.spring/rosenberg976.html