In mental imagery, time often represents itself. This is clearest in perception, where experience of duration is duration of experience. You experience the interval of time that a play or symphony occupies by the length of time it takes to experience it. Much the same is true of recollection or anticipation. Mentally to experience the length of a sequence of events is to have a mental experience of it playing in “real time.” In the absence of real-time imagery, you have an experience of temporal sequence, but not an experience of duration. For example, you can replay Hamlet in the theatre of your mind, but you can’t experience its duration without running each scene at its actual pace. There is no scaling in time perception or time imagery. (Time is special in this way. You don’t experience size by the size of your experience. Or, mutatis mutandis, anything else. A question: is this just a trivial fact about time, or does it say something about how we experience duration?)
In discussing the experience of time passed, Mark Lance and Eric Winsberg suggested that the experience of time passed implied the existence of one or more inner clocks. This seems right to me. Responding to their suggestion, I wrote: “you can't . . . know elapsed time simply by having a memory of moments in the interval being measured—some moments [in the interval] or all moments. You need an elapsed time clock to measure elapsed time; nothing simpler will do.”
In fact, the idea of an internal clock is relatively new. As C. R. Gallistel tells the story (in his wonderful 1980 book, The Organization of Action), it was unknown to Charles Sherrington. For Sherrington, the nervous system connects sensory inputs to motor outputs in a system of “reflex arcs”, which consist of “at least three separable structures—an effector organ, gland cells or muscle cells; a conducting nervous path or conductor leading to that organ; and an initiating organ or receptor whence the reaction starts.” The function of the nervous system was merely to connect receptors to effectors in a complicated way; it did not initiate movement on its own.
Sherrington knew of certain rhythmic motions of animals: scratching, stepping, etc. He also knew, of course, of the rhythmic action of the heart muscle. However, Gallistel explains: “neurobiologists of the first half of [the twentieth] century were unwilling to entertain the idea that components of the nervous system were endogenously active.” Consequently, they (and, in particular, Sherrington) tended to explain rhythmic motions by a “refractory phase” of certain synapses: once excited, these synapses would transmit a signal and then shut down for a while, thus breaking “afferent excitation into rhythmic bursts.” This mechanism functions, in effect, as an internal clock, but it is driven by external stimulation. The nervous system could not, in their view, drive bodily processes autonomously.
Sherrington and others pursued their researches into the nervous system by means of terribly cruel device. They had laboratory dogs whose spinal cord was cut so as to isolate autonomic reflexes. One such researcher, Gordon Brown demonstrated in 1914 that the rhythmic scratching and stepping of the “spinal dog” required no sensory input. These reflexes were driven by an endogenous mechanism governed by an internal clock. Brown’s work was taken over and expanded by Erich von Holst, who in the late 1930s posited an elaborate system of autonomous timekeepers within the nervous system.
The role of purely endogenous clocks is still somewhat controversial. In a 2007 review article (kindly sent to me by Julian Kiverstein), Dean Buonamono (writing in Nature Chemical Biology) reviews some models. The fastest “clock” is probably a Sherringtonian device (i.e., one driven by an external stimulus). Sound location depends on timing the difference between the arrival of an auditory signal at our two ears. Obviously, this interval is tiny—a couple of microseconds. The signal from each ear is sent in opposite directions down a neural circuit to determine where they meet and cancel each other out. Where they meet is a function of the inter-ear time lag. Inner time experience depends on much longer time periods—including circadian rhythms. These longer periods cannot be measured by conductance patterns in the nervous system. They seem to rely on biochemical feedback loops.
The full story is still quite obscure.