Here one can see that the letter groups خ ح ج and several others are identical if not for the dots. In this pair, two are vaguely H-sounding, the other was formerly some form of palatalised G, which I would expect to instead pair with one of gh, k or q, not x or ħ. Particularly perplexing are the medial forms of ـبـ, ـتـ , ـثـ, ـنـ, ـيـ, such that a governor sent a letter by Caliph Umar is recorded as having been unable to tell whether to اقبل (accept!) or اقتل (kill!) one guest.
Do we have any information as to how different letters with different sounds converged on the same latter? Do we have any evidence of a complexity loss in the script due to some factor, or of the particular forms of the letters in Arabic prior to their orthographic collapse?