As i have heard, there were 2 types of e's, one got creating through connecting "a" + "i" (which i think is the e for 已然形) and one through "i + a" (which i think is the e for 命令形),
And since the suffix り apparently developed through fusion of auxiliary verb あり + 四段 in 連用形 (like 書き + あり = 書けり), that means it should be "i (as in kaki) + "a" (as in ari), which means its a fusion of 命令形 right? am i correct on this?
It is safe to say that the suffix り is connecting to 命令形 and not to 已然形 because of two types of e.
Until the beginning of the last century, there are many variants of hiragana (including current ones) and people pretty randomly choiced among them to express each mora. Also in 万葉仮名 scripts, people used variety of chinese characters for the same pronunciation. However, scholars found in ancient documents there were two groups of characters not interchangeable (referred as 甲類 and 乙類) representing the same vowel (i, e and o).
Because of this discovery, nowadays it is widely accepted that there were two e's pronounced differently but there are, as far as I know, no established theory what are their actual pronunciation or origin. (a+i and i+a is one of the leading hypothesis, though.)
Therefore, the evident part of the statement is that the suffix り comes after e甲, which is the same as 命令形（e甲） and different from 已然形（e乙）.
