In "Law is Code," Artur Bergman discusses a program that (somehow) turns text into these nifty visualizations with lines and dots and such. When they ran the US Code through the program, it churned out this very, very dense picture of criss-crossing lines. Mr. Bergman's comment:
Legalese is a massively structured dialect. Symbols appear in very distinct patterns that are more reminiscent of machine code than text.
Which isn't entirely fair to machine code. Yes, legalese contains more structure and repetition than normal English. But that doesn't mean it achieves precision. In machine code, a word may denote some tangible reality ("widget"), whereas the common words in legalese denote abstract concepts ("reasonable").
Or so it seems to me it seems to me it seems to me it seems to me.