A holds the 'sibling-in-law' relation to B only in the case when:
(1) A is a sibling of C and C is married to B; or
(2) A is married to C and C is a sibling of B.
What is common to (1) and (2) is that they both involve a complex relation, i.e. a relation composed of two relations, is a sibling of and is married to, but that one is substituted for the other, keeping their arguments constant. But then one may wonders why use a single notion i.e. 'sibling-in-law' (as is the cas in at least English, French, Italian, and countless natural languages) to refer to individuals when it makes quite a difference to those individuals' identity, whether the order of two subrelations is (1) or (2).
Hence this question: Is there a plausible hypothesis as to why many natural languages mark no lexical distinction between sibling-in-law as determined by (1) and sibling-in-law determined by (2)? And does it even make sense to assign sibling-in-law what seems to be a disjunctive meaning, i.e. either (1) or (2)? Are there known examples of the same phenomenon in the category of common nouns denoting individuals?
edit: I just noticed that some dictionaries add another condition, (3): A has a spouse who has a sibling who has a spouse B. This does not affect the point or the question above.