Dranorter: thanks for the reference. Johanna Drucker was a visitor to my bookshop long before her catalogue raisonne was published (we also share an interest in the independent press). I bought Alphabetic Labyrinth when it came out and have been familiar with its contents for years. My archive of studies on script is quite extensive, supporting 20 years of intensive research in the subject.
In other words, I'm acquainted with the insights and issues raised by earlier scholars of the alphabet (and likewise conversant with the categorical qualifications inherent in novel hypotheses: eg, intentional perception). But the discovery I made led to so many corroborating insights (substantiated mathematically) that readers are in fact called upon to evaluate a broad new interpretation of ancient thought; requiring consideration beyond the dismissive resort to prevailing consensus (with which, as it turns out, I am already all too familiar).
For instance, the significance of this insight is enhanced by an understanding of lunar mechanics. Ancient cultures, as has long been established, sought to determine the measures of lunar recurrence. There are three main cycles:  lunation (whose mean measure is 29.530588 days);  course cycle (230 lunations);  phase cycle (235 lunations). Empirically determining extended measures (such as the latter two), involves records maintained over considerable intervals of time (18.6 years for a single course cycle; 19 years for one phase cycle).
The course cycle marks the recurrence of the rise of the moon at its extreme northern declination on the horizon. The phase cycle, the recurrence of the same phase on the same solar date (eg, full moon at equinox).
The obvious method for keeping track of the moon over extended periods is to ascribe a different character to each phase, and copy that figure into the appropriate square on a blank calendar grid, when observed (or on the night it should have been observed). To most people this may seem redundant because the common impression is that the lunation doesn't change from month to month. But in fact, there may be as much as a 36-hour difference between focal phases (such as crescent and half-moon) in different months of the same year. Which means that one month appears to contain one less phase than another.
In fact lunar mechanics are so complex that it wasn't until 1919 that we were finally able to predict the appearance of the moon with reliability. And the equation employed to this end, requires 1500 variables!
It is my contention that the alphabet facilitated the maintenance of these ancient records, each letter marking a different phase on the putative grid. The benefit in employing a serial mnemonic which is also used as a script, is that the means of keeping the records are effectively secured against disaster, both natural (quake or plague) and cultural (invasion).
Although the lunar priests might meet their end, their system is readily recoverable by observant heirs astute enough to recognize the convergence of focal phases of the lunation with pivotal characters in the script. And the surviving records (which would appear to intruders, or to the uninitiated, merely as tables of letters) regain coherence, restored to the utility of determining an accurate calendar of lunar recurrence over extended intervals. A precaution hypothetically established by the priests against wasting their lengthy labours, in leaving the heirs the fatal legacy of having to recount an archive of course cycles (over centuries) from the beginning again, to recover the sacred insights.
If ancient man believed that the moon was the dominant deity (presiding over him at his most vulnerable, during sleep), the question of what his god may have been trying to convey in showing a different face every night, might inspire him to keep track of the number of phases (faces) which arose between the recurrence of the same phase. Abundant evidence of such lunar counts exist (scratched on bone or stone), dating back to the Palaeolithic.
In pursuing this hypothesis, I conjectured that ancient cultures (such as the Egyptian and Hellenic), may further have equated their deities with phases of the lunation. Identifying the focal deities of Egypt with the focal phases of the lunation, in turn, exhumed insights into numerous enigmas of Egyptian history.
One example is the Eye of Horus (which gave rise to the Hekat fractions) [http://www.archive.org/details/TheLunarContextOfTheHekatFractions]. Supplementary insight appears among my replies to an earlier correspondent [http://www.allempires.com/forum/forum_posts.asp?TID=29164], including a brief explanation (further along) of the Egyptian Great Year theophanies of the three sacred bulls.
Curator of the Palatine Library, Gaius Julius Hyginus (ca 64 BC - 17 AD), left one of the more memorable ancient accounts of the origin of the alphabet, in which Mercury drew the first seven letters from the flight of cranes. Cranes are white like the moon. The letters appeared to him in flight (like the phases on high). Seven letters for the seven focal phases of the lunation: opposing crescents; opposed half-moons; twin full moons; first waning phase. This ancient 'fable' accords 'coincidentally' with the hypothesis of lunar origin (not with 'proportions of the human body', the 'diagram of a flute', or 'phonetic patterns', not to mention many of the other ingenious projections conjured throughout history).
Mercury was known as Thoth in Egypt, the god of writing, not only credited with the invention of writing (and measure) but also called upon by the council of deities to restore the Eye of Horus in the myth cited above. A myth whose solution reveals a remarkable ancient computation of the mean lunar cycle, previously unsuspected by scholars.