Contour theory is a particular account of what it means for music to be expressive. The idea, usually associated with the earlier work of Peter Kivy and with the work of Stephen Davies, is that music (or at least some basic kind of music) is expressive not because it symbolically represents emotion, but because it in some way presents it, that music can have a structure that we recognize as having the same 'contour' as our own physical expressions of emotion. For instance, music can have a tempo much like our heartbeat, or our gait, when we are excited; it can have a directionality that's like the direction our body takes when it slumps or rises; and so forth. Roger Scruton argues in The Aesthetics of Music that contour theory cannot account for the importance of musical expression, and also that it seems to confuse means and end. I think there's something to both criticisms; they are a good reason to think that contour theory, simply on its own, does not adequately characterize the full richness of what we experience music as expressing. But this is very far from saying that there is nothing to the theory; even if we must add something to sort out important or interesting expressions from unimportant or uninteresting ones, and even if we take the contour or shape of the music only to be something that disposes sound to be expressive, rather than the expression itself, there still does seem to be a way in which contour plays a role. I mean, listen to Prokofiev's Op. 67 (narrated here by Basil Rathbone) and tell me that the music in "Peter and the Wolf" is not expressive by resemblance at all; obviously there is some purely conventional element to it, but we can recognize the appropriateness of Prokofiev's choices for it. And when we look at pure music, as opposed to music like this in which we are aided in our interpretation by words, we still find the same kinds of appropriateness. In any case, for the purposes of this post, we don't need to consider contour to be the only or even the primary account of expressiveness, as long as we can recognize it as a major means of expression.
One of the interesting things about contour theory is that if it's true, it is plausible to extend it beyond music to other things. I think Davies uses the example of the weeping willow being 'sad'. It's not purely a matter of arbitrary convention; the willow has the same demeanor, one might say, as someone who is sad. Of course, the willow is not itself expressing any emotion, but we can say it is in some sense expressive of it, or suitable to express it, or some such thing. (We want to be able to say in general that something can be expressive of an emotion without anyone expressing that actual emotion, because of acting and the like.)
Marta Benenti and Cristina Meini had a paper a little while back in Philosophia ("The Recognition of Emotions in Music and Landscapes: Extending Contour Theory") in which they extend contour theory specifically to landscape painting. (They differ from some standard contour theories, so they distinguish their view from those, but it can be considered a contour theory in a broad sense.) They note that contour theory doesn't require that we ourselves perceive the similarity itself, as long as we perceive a particular manifestation or characteristic that is similar; we don't have to perceive the similarity of smiles to perceive smiles as smiles, and neither do we have to put our finger on exactly what makes the music and the emotional behavior have the same shape. We just perceive the shape and react to it in the same way in both cases. We could in a sense just as easily say that the emotional behavior expresses the same thing as a bunch of music; indeed, people do occasionally talk as if music sometimes sheds light on the emotional behavior rather than vice versa. But this will also be true of depicted landscapes. They give the example of Caspar Friedrich's Der Nachmittag:
You can see this landscape as calm, perhaps even somber (Benenti and Meini suggest sad and melancholy, these are perhaps a little strong for the quiet expressiveness we find here, but you could suggest, perhaps, that there is a something of a tendency toward these things): there's a lot of gray and dark earth colors, not much indication of energetic action. If you, seeing this painting, were then asked what kind of painting would express joyful excitement, you could give a clear and coherent answer: brighter colors, more suggestion of motion, perhaps something about the painting's lines drawing you up and forward, and so forth.
It's clear that the expressiveness can't be quite the same, because painting doesn't allow for the actual movement that music does, the acceleration and deceleration, the vibrational characeristics (church organs don't just play, a good church organ in some sense plays you, by sending its sound through you), and so forth. But Benenti and Meini note that facial expressions and the like can be expressive even when static. But even setting that aside, paintings and drawings can, of course, be suggestive of motion.
If we take this to be true of depicted landscapes, though, it is surely true of the landscapes themselves. The depiction can add or mute features, of course, but many of the perceptible features that make the painting expressive will be shared by the actual scene itself (and the scene, of course, can be in actual motion that is only suggested by the painting). And this fits our experience. The world is expressive to us. The sky may be lowering, the flowers may be joyous, the mountains may be dignified, having that very shape.
Davies, if I recall correctly, actually suggests that saying that a willow is sad is not metaphorical but literal, although in a kind of secondary way; I think this perhaps draws the line between literal and figurative in the wrong place. But I do think it shows that emotive metaphors, like saying that a meadow is joyful or a sky is sad, can be argued to be non-arbitrary and to be describing something that is genuinely in the thing itself. And, of course, these emotive metaphors are the basis for others again, which are building directly on them, like the metaphor of a smiling meadow. Regarding the world around us as expressive is a part of our rational interaction with it.
There is a tendency to think that we get dryads by taking nondryadical trees and dryadizing them, personifying them. But the evidence really suggests that it works the other way. It would generally be closer to the truth to say that we start with the dryad and then reduce her to a bare tree. But for all that, the tree has a sort of recognizable expressiveness and will always seem a little like a dryad.