Do Not Cry Out on Behalf of Plants

Plants are crying out. This sentence goes wrong too quickly: before it has even dealt with sound, it has already fitted sound with a throat. Under drought or cutting stress, plants may indeed emit airborne ultrasound, roughly between twenty and one hundred kilohertz. The problem is not that there is sound, but that the moment we hear sound, we hurry to hear a stress event as an expressive event.

This urgency is not entirely ridiculous. To say “crying out” makes people pause; it cracks open the smooth quiet of the greenhouse, so that a plant is no longer merely background, ornament, yield, or greenery. A rough metaphor can sometimes draw attention toward something. The trouble is that it draws it too close, and too humanly. It makes us think that as long as an event is given a throat, concern will become more accurate.

Standing in a greenhouse, what human ears still hear is quiet. The sounds recorded by microphones have to be lowered into the range of human hearing before they resemble brief clicks, like a few sudden pops of popcorn. Even these clicks are already displaced versions: first recorded, then frequency-shifted downward, and finally placed where human ears can catch them. Hearing is not arrival at the original sound, only a successful act of transport.

“Crying out” gives too much. It pushes sound toward throat and intention, as though wherever there is sound, there must be a center that is speaking. Yet these sounds may be related to the formation or rupture of bubbles in the water columns of the xylem. The sound does not come from a place that wants to speak, but from a flow somewhere inside the plant that has been pulled taut and interrupted. Calling it “acoustic emission” is more accurate: the name pulls sound away from the mouth and returns it to the lumen, the water column, the bubble, and the stress event.

Machines can assign some sounds to categories such as drought, cutting, or normal state. State matters, enough to change action; but classification is not translation. What it outputs is a state, not a line of speech. The greenhouse is still not entirely quiet. It is just that the string of downshifted clicks is still not a sentence.