On Philosophy

February 25, 2009

An Event Ontology

Filed under: Ontology — Peter @ 6:49 pm

It is natural to ground a metaphysical system in objects. The English language, along with most other western languages, is an object centric language. Sentences are constructed primarily by describing nouns and connecting them together. Expressions that describe or refer to some indefinite thing (“the thing in the box”, “something is in the box”) are common, but similar expressions that describe relationships or actions (“the relationship between the watch and the man”, “the effect that the car had on the man”) sound odd. Indeed, in such constructions the relationship or action is often treated grammatically as an object (“the relationship”, “the effect”). This is not necessarily a bad thing, but because it is so natural we are often blind to other alternatives; we reach so readily for objects and properties as tools that we never stop to think what else might be there. Here I want to take a look at one of those other possibilities, and then see whether there are cases where it might be a better fit than our customary objects and properties. This is doing philosophy in reverse, as I understand it, since here we have a solution looking for a problem rather than the other way around, but so be it.

Besides objects what other options do we have? Some possibilities can be found in our language, such as actions, relationships, and properties. These are all viable alternatives, but there is a tendency with all of them to fall back, perhaps unconsciously, into an object ontology. Actions and relationships are between two or more objects, and properties are things that objects have. Thus here I will build my alternative to object ontologies out of events. It is possible to describe all the things we ordinarily think of as objects as events of a very boring kind. A chair, for example, can be described as the event of the chair bring or existing. In this event the components of the chair take only very minimal actions, such that they remain basically the same throughout. A chair doing nothing is an event in the same way that silence is a sound; it is an event where nothing much is happening. Still, it is an event. An event is a characterization of change. Since the absence of change is itself a kind of change (a limiting case) we can talk about an event that describes it.*

Of course just because we can describe the world in terms of events, that we can point at the chair and say it is one kind of event, and at the man falling and say it is another, doesn’t mean that we have accomplished anything significant. If this ontology is to do any meaningful work we need the ability to say things about events. The temptation is, of course, to add properties into the system (or classes of events, event kinds, which amount to the same thing), and thus to say that an event has the property of being such and such. But, however intuitive that may be, there is good reason to avoid it. If we add properties in this way we will have fallen back into an object ontology, albeit under another name. By adding properties we make our events work exactly like objects. Instead of the object chair having the property green we would now have the event of the chair’s being having the property green. Thus events essentially are objects (following the philosophical principle that whatever walks like a duck, and quacks like a duck, is a duck); the only difference is that besides the normal set of objects (tables, chairs, cars, etc.) we have what can be thought of as situational objects as well (the car turning a corner, the chair being moved, etc).

So properties aren’t the way to go. What then can we say about events? As usual the place to start looking is with how we ordinarily talk and think about events. One common observation to make of an event is that it produces or causes some other event or events. For example, the car’s failure to start (an event, it failing) caused the professor to arrive late (another event, his arriving). In a similar way we also talk of the interaction between two or more events, often in connection with causation. For example, the forest fire (the forest burning) interacts with the rain (raining); it is caused by the rain to go out. Finally, we often describe an event as being composed of a number of interactions or causal connections, or both. The car’s failure to start, for example, was the event of the key being turned which caused certain events within the engine that interacted to cause the engine to stop. At this point the temptation is, naturally, to go farther along this road and to formalize these relations (and hopefully find a few more in the process). But I think that is going too far for our purposes here. The above examination of the possibilities is enough to show that we can have rich descriptions in an event ontology without resorting to properties.

With some general idea of what an event ontology would be like we can now turn to more substantial questions, such as “what is it good for?” In other words, what can we do with an event ontology that we couldn’t with an object ontology, or at least not easily. Well, consider how we would go about describing a river in an object ontology. Naturally we want to find some object “the river” to which we can ascribe properties, such as fast-flowing, or in-North-Dakota. But this river object is an elusive one; what is it exactly? We can’t identify it with the water because the water flows away. If we did we would never find the same river in the same place, and we would find many rivers in the sea. Indeed this is the problem understood by Heraclitus when he said “one cannot step into the same river twice”. On the other hand, if the river is not the water it is not clear what it is, and its relationship to the water becomes complicated. If the water is hot so is the river, but how can this be if the river is not the water. If the water is only part of the river then what is the rest? There is no uncontroversial object which can be the river for us. At best we will end up with an object that is part abstract, since it isn’t tied to any one bit of stuff, but part physical as well, since it somehow depends on that stuff for some properties. (It is not purely abstract in the same way the number 4 is.)

Of course this problem isn’t limited to rivers. The nature of the mind creates similar issues, since we think that the same mind could, in theory, be embodied in different physical stuff. And, at the same time, it is not independent of the physical stuff it is embodied in. There are also more objects like rivers if we go looking for them. Every organism, for example, is like a river in that its composition is constantly in a state of flux (due to growth, healing, and digestion). Although slower than a river, it is equally impossible to identify a living being with some particular stuff, and equally impossible to separate it from the stuff that composes it.

The problem that object ontologies encounter in such situations, as I understand it, is that an object is, fundamentally, a unified thing. An object is one thing, and the same thing over time. Problems arise, however, when “objects” such as rivers and organisms don’t display the necessary unity and sameness over time. A river is not always some one unified clump of stuff. Events, however, don’t presuppose any sort of unity. Thus it is easy to identify a river with a flowing of water (the event of the water moving). We could say that a particular river is really a particular flowing of water, in a certain way, and in a certain area. This solves Heraclitus’ problem, because even though the river, so understood, flows, it never goes anywhere. It is quite possible to step into the same flow or flowing twice, if not the same water.

To speak about the river (to describe its “properties”) is as simple as speaking about any event. Being hot and being cold, for example, are two events. To say that the river is hot or cold is to say that it has certain interactions with other events of being hot or being cold (to say a river is hot is to say it warms cool things). Organisms and minds can be described similarly. A particular organism is a particular kind of living (biological process). A particular mind is a particular kind of thinking, a particular kind of interactions between its components (events). Thus the same mind (the same event) can be found wherever there are the right interactions. This has exactly the right sort of relationship between the mind and its components; the mind is not identical to them, but nor is it independent of them, since it is their interactions that “produce” it. (Produce understood in the sense of determine it to be, not in the sense of one object causally spitting out another object.)

At this point weaker souls than ourselves might be tempted to conclude that an event ontology is strictly better than an object ontology. After all, there are so many problematic “objects” that it can handle better. The measure of an ontology, however, is not the number of object that we can fit into it. It is more important whether the objects we are interested in fit into it, and whether it is a helpful way to think about them. By that standard it is far from obvious that an event ontology is a superior option. The fact that an object ontology is so natural is not some accident; an object ontology is a generally useful tool to understand the world with. Yes many “objects”, such as living beings, aren’t quite objects proper. But, in most cases, little is lost be considering them to be objects proper in spite of their falling somewhat short of the ideal.

Indeed the defect of the object ontology that we have pointed out as generating a need for an event ontology, its identification of an object with some particular stuff and its strict conditions of identity, is also its greatest strength. Often when we mean to talk about something we want to talk about one particular thing. And we want to be able to re-identify that thing in the future. If I speak about my kitchen table, for example, I mean to talk about that one particular table. And if I mention it again tomorrow I mean to talk about the same table. Think about what would happen to the kitchen table in an event ontology. Instead of talking about that table, the object, we would talk about the being of the table. However, this being of the table, no matter how narrowly specified, is not restricted to just one instance. It is always possible for another table to exist that is another instance of the same event. This is an intrinsic feature of events. It is because they only care about what is happening, and not what is doing the happening, that they can describe things such as rivers in which the substances that are caught up in the event constantly change. But this also means that they are necessarily blind to those substances, such that it is always possible for the same event to occur in multiple places at the same time. From this also follows another unusual consequence: an event ontology is blind to what we normally understand as the continuity of objects over time. Yes, in an event ontology you could say that the being of the table at one time is the same as the being of the table at a subsequent time. But the same could be said of two tables that exist at the same time. Indeed the table could exist, be destroyed, and a copy could be built to replace it and we could end up saying that the being of the first table is the same event as the being of the replacement.

Do these problems “refute” an event ontology? Of course not. But they illustrate that an event ontology is not a universally superior option. It seems to me wisest to stick with an object ontology by default, in part because of its simplicity and familiarity, and to resort to an event ontology only when needed. Specifically when the ever changing nature of things cannot reasonably be ignored.

* Some philosophers like to talk about being. When they do they usually conceive of it as an object or a property (or possibly an indeterminate cause). However “being” is simply “to be”, a verb or event, turned into a noun, which is a product of our natural object ontology. But, given this event ontology, we are free to treat being as an event. Thus the being of the chair would naturally be understood as the event that we equate with the chair. The same goes for existing, subsisting, and so on.


Blog at WordPress.com.

%d bloggers like this: