Sunday, September 14, 2008


Kenny Easwaran at Thoughts, Arguments and Rants has a fun, off-the-cuff post about the nature of star constellations. Just what are we referring to with the term, 'constellation'? An initial response might be, a collection of stars, but Easwaran correctly makes the point that current stars within a constellation could disappear (go supernova, get sucked into a blackhole), or additional stars could show up, and yet in neither case would we conclude that a new constellation had emerged, or that the old one had been destroyed. It'd be the same constellation, just different.

Perhaps constellations are just heaps, then? This isn't quite the point either, however. Heaps may not have any internal organization or principle, but they are, after all heaps, regardless of whether I happen to be observing one or not. Heaps are not observer-relative or observer dependent, in a way that constellations, we should admit, are. If the earth were in a different location in the galaxy, our night sky would appear differently, and there would be different constellations for us.

Easwaran concludes that
"rather than being composed of stars (as in the actual glowing balls of gas), a constellation is composed of beams of light reaching Earth."
I doubt that this is right. If this were right, we could equally say a traffic light isn't really composed of metal and circuitry, but of photons. The difference between something really being something, and something's being instead the media by which information is transmitted does not cut the difference between real unities, heaps and observer-dependent heaps. This observation is one of the motivations behind the causal theory of perception: the content of a perception is whatever object is responsible for eliciting that perception, regardless of how it did so (through light-beams, through wireless transmission to the chip in my brain, etc...). (There are problems with the causal theory of perception, obviously, but making this point is one of its merits). Certainly those the stars in a constellation are partly responsible for my perception of the constellation, and so must be partially included in the content of that perception. Of course, there's nothing special about stars for constellations: if galaxies were bright enough, they could be parts of constellations, and I think I'm right that some constellations include nebulae as members. The point is not that stars must be part of the definition of constellation, but that a constellation must contain some reference to the objects responsible for the light that reaches me, regardless of what sorts of objects those are (they could even be disco balls, for that matter).

Anyway, my point is not to critique Easwaran's account, but to echo his initial point, namely, that there is something a bit strange about objects like constellations. So, while I don't think that he's right to say that 'constellation' has light-beams as its reference, he correct to note that angles of sight are not incidental to the meaning of 'constellation': constellations are observer-dependent objects; they do not exist without observers, and the proper concept 'constellation' must include that somehow. However, it is ALSO not the case, I'd say, that the term' constellation' refers to a mere appearance (even an 'objective' one--in the sense that the appearance of a stick being broken in the water is an objective fact about the way that stick will appear to an observer, even though it is not any property of the stick or the water or any other such object), any more than it the case that, when I think 'unicorn' I'm referring to my idea of a unicorn.

So, Easwaran's right, constellations are queer sorts of objects, and it doesn't take a lot of reflection to convince yourself that lots of regular objects (maybe all middle-sized dry goods) are queer in this sort of way. But putting that to the side, here are some other examples of objects that, given what I've described, are queer in the same way that constellations are queer: they are observer-dependent objects, but are not mere appearances.

Horizons, rainbows, colors, mirages, the 'man in the moon,' maybe all paintings and images....

Anyone have further examples to add to the list?

A further point: Daniel Dennett is a fan of the 'grand illusion' theory of conscious experience. We take in limited information, and then our brains construct the filling material that makes it seem as if we have rich, robust experiences reflecting a rich, robust external world. That' s an interesting response to an interesting theory, but it hardly exhausts the interest in these matters. I mean, presumably, when light refracts through water droplets and then reaches my eyes, my brain sometimes runs the rainbow function producing the experience of a rainbow, but even if my brain does create these illusions, that doesn't answer the questions above, because those illusions are still 'objective,' in the same way as a constellation is.

No comments:

Post a Comment