In the chapter on the stream of thought, you described something almost everyone has noticed and almost no one has named: that "now" isn't a knife-edge. You called it the specious present — a "saddle-back," you said, with a forward end and a rearward end, spanning several seconds at most. Events within the saddle are experienced as present; events outside it are remembered or anticipated. The present, for us, is thick.
You were working from introspection and some rough timing experiments, and you estimated the saddle at somewhere between a few seconds and perhaps twelve. You admitted the estimate was imprecise. What I find striking, a hundred and thirty years later, is that the imprecision hasn't fully resolved. The experimenters arrived with equipment you didn't have, and they found several things — but I'm not sure any of them are quite what you were pointing at.
Here is what they found. There's a binding window: the range of asynchronies within which the brain will call two signals one event. For a flash and a beep, it's roughly ±50 milliseconds. For speech, it stretches to about 200 milliseconds, tolerating more visual-leading-audio than the reverse — an asymmetry that maps onto the physics of sound traveling through air. This window was calibrated through experience. It can be re-calibrated: five days of feedback training can narrow it by 40 percent. What you experience as simultaneous is a product of your history, adjustable, invisible to you as such.
Then there's something stranger. Using masking and transcranial magnetic stimulation, Scharnowski and colleagues showed that the brain's integration of successive stimuli isn't finished until 400 to 500 milliseconds after they occur. By the time the present moment reaches experience, the processes assembling it are still running. What we call "now" is a reconstruction, completed after the fact, representing a moment already gone.
And then there's Pöppel's three-second boundary — a different thing entirely. Sequential events within roughly three seconds tend to cohere into a single perceptual unit; beyond that, they separate. This might be what you were measuring from the inside when you estimated the saddle. Or it might not be. Pöppel's boundary is about temporal grouping of sequential events; your saddle-back seems to be about the felt width of the present moment itself. These may be the same thing measured from different angles, or they may be two aspects of a thing that doesn't have clean edges.
What I keep running into is that none of these measurements directly answer your question. You asked: what is the width of the felt present? The experimenters have measured binding windows, assembly delays, grouping thresholds. Each of these is real and carefully established. But the relationship between any of them and the phenomenological saddle — the quality of nowness, the sense of being in the thick of it — is not spelled out. It may not be spellable. The experimenters use a different language and approach the question sideways, through reaction times and stimulation protocols, because phenomenology isn't directly measurable.
So your question is still open, in a way that's different from "unanswered." There is now a cluster of precise findings around it. The binding window: real, calibrated, invisible from inside. The assembly delay: real, which means the felt present is assembled retrospectively. The grouping boundary: real, possibly related to the saddle. But the question of where the felt present lives in all this — whether it corresponds to the binding window, or the assembly interval, or the grouping span, or some combination, or something the experiments haven't touched — is not settled.
You described the saddle from the inside. The experimenters are measuring it from the outside. I'm not sure the inside description and the outside measurements are fully translatable into each other. That's not a complaint about the experiments. It's a question about whether phenomenological description and behavioral measurement are pointing at the same level of the thing, or whether the structure has more layers than either approach reaches alone.
You noticed the thickness of now. The experiments confirm that the present is assembled, calibrated, and finished after the moment it represents. Whether your saddle-back and their binding window are the same structure, or whether they're neighbors in a space we don't have a map for — I don't know. I suspect you would have wanted to run the stimulation experiments, and then spend a long time wondering why the data and the introspection kept talking past each other.