The problem with yermack here is that he's ascribed the functions as strict behaviours instead of thought processes, but it's not his fault he's primarily a static comprehender. Yes he is correct to note that ILIs:
*"observe logicality" (verification of logic to make sure the models aren't absolute nonsense)
"compliance with logicality in any details" (verification.... )
"orientation on clarification of details" (verification.... )
"Concrete usage of rules, order, consistency (legislation, resolution, regulations, rules, instructions, etc.)." ( related to verification this is looking at the objective circumstance)
"strict logic" (standard logic)
"non-recognition of illogicality" (yes this is true when suppressing the subjective side of their logic)