A few weeks ago I discussed one of the problems with trying to use 'real' to differentiate between video game worlds and our world. There, my discussion was based around issues of constitution and factuality. Now, I want to pick up a different issue in the same vein, one that I touched on when discussing Unwind; the fact that video games don't take place in our space.
In fact, it's tempting to say that video game events don't take place in (a) space(s) at all. They represent spaces (better: they present visual patterns which represent spaces), sure, but that's not the same as involving a space. Or is it? Depends what 'a space' is.
At the most basic level, Barry Dainton (full disclosure: my boss) identifies two functions we attribute to space rather than its occupants. Constraint is the function of limiting the ways we can move; our space is three dimensional, so any motion we can make is a combination of movements in those three dimensions. Connection is the function of allowing space-occupants to interact; wherever you are, you and I are spatially connected because there are logically possible spatial movements that would bring us into contact.
It should be obvious that video game spaces meet these criteria. Connectedness is almost trivial when it comes to game objects, since a game cannot present an object unless it is possible within the game to view that object (and thus connect it at least to the camera/viewpoint). Dimensional constraint is such a significant factor in some games that we use different dimensionalities to partly define some genres – 2D and 3D platformers, for example.
Of course, this is a broad definition. It allows as a space anything that has both dimensions and occupants; the standard windows and HTML colour charts become spaces, whose dimensions may be rendered either as red, blue and green or hue, saturation and value. This shouldn't seem too counterintuitive, since we often and productively represent the colour chart spatially.
The colour chart is a kind of space we might naturally think of as abstract, but it's quite hard to pin down what this means. Its occupants are abstract in that they are colours abstracted from any particular object possessing them, but this is a fact about the occupants of the space, not the space itself.
Video game spaces are in a similar situation; we know they don't take place in our space, since we are not conventionally connected to their occupants (nor, indeed, conventionally constrained by their dimensions). But if there is a difference in kind between our space and game spaces – if video game spaces are not real in the way that our space is – it is difficult to articulate.
The simplest way is to adopt the substantival construal of space put forward by Newton to fit his classical mechanics. According to this view, space is a thing that exists independently of its occupants; list the fundamental ingredients of the universe and in addition to a set of physical atoms, there will also be a set of spatial points or regions. Empty a room of all its physical contents, and a substance remains: space, separating the walls and permeating the intermediary void.
Despite the utility of the physics that go alongside it, Newton's theory of space is notoriously difficult to prove. Newton's proofs rely on demonstrating absolute motion and rest – motion which is relative not to other objects but to space itself. The bucket argument is the best-known of these. Whether the bucket argument proves Newton right about substantival space, though, it cannot differentiate systemically between our space and game spaces; any game may be programmed to mimic the effect on which the argument trades.
Newton, then, cannot help us differentiate between 'real' and 'virtual' spaces. The chief opponent of substantivalism, the relationism of Leibniz and Descartes, does even less to distinguish the two. According to relationism, space is merely a matter of the spatial relations among objects; no objects, no space. So any set of real objects that are connected and subject to shared dimensional constraints will constitute a space.
This passes the question of whether game spaces are real or not back to the question of whether game objects are real, and as I argued last time out, there's little ground for saying game objects are less real than the conventional objects we deal with on a day-to-day basis.
Perhaps it's better to view this as a reductio ad absurdum on the common use of the term 'real'. We can intuitively understand that there's a difference between our relation to game objects and our relation to regular tables and chairs without needing to dismiss game objects as lesser.
Here, one final philosophical notion may be useful. In his work on possibility, David Lewis developed the theory of modal realism, which holds that every possibility is a universe in its own right, on a metaphysical par with our own. His reasons for this extremely profligate version of parallel world theory aren't important; what's useful is how he understands one particular possibility's being 'actual'.
For Lewis, 'actual' just means 'in our universe'. It's an indexical term, like 'here' or 'now' – its meaning depends on who says it. When I say 'here' it (almost always, because I don't get out much) means my house in Liverpool. I can say 'It's warm here', and absolutely not contradict you when you say 'It's cold here', so long as you and I are in different places.
We can use 'real' to distinguish between our space and video game spaces in much the same way; a 'real' space is a space we're in. Our space is real to us; the space of Hyrule is real to Link.
I've got one more sense of 'real' that I want to discuss, the one that I think is both most important and most clearly shared by game events and objects, but that will have to wait for another time.
: This puts aside, for simplicity's sake, string and M theories which suggest our space may have more than three dimensions, as well as holographic theories which suggest it has fewer.