We have become skilled at counting appearances. A signal arrives, is assigned a place, and joins the evidence that something exists. Its absence is harder to catalog. The empty line can mean that nothing happened, that an instrument failed, that a witness turned away, or that the record was altered. The line looks the same in every case. An archive that refuses to distinguish them is not economical. It is careless.
The specimen called 237 entered these materials as an omission. No one reported seeing it arrive. Instead, two otherwise ordinary records ceased to agree. The first inquiry attempted to remove the discrepancy. The second attempted to explain it. The third, whose papers survive here, asked what operations would allow another reader to observe it without trusting either inquiry.
This distinction is small enough to be neglected and large enough to organize a life. An explanation can be repeated without its evidence. An observation can be repeated only if its conditions remain available. Our work is concerned with those conditions: the copy before correction, the apparatus before adjustment, the witness before a catalog has decided what they must have meant.
The cicada is our chosen figure not because it explains the network, but because it interrupts a bad habit of thought. We mistake the interval in which we do not observe a thing for a history in which nothing takes place. A figure is useful when it makes an error visible. It becomes dangerous when we ask it to replace an inquiry. The following pages should be read with that limitation in mind.