Is knowledge a natural concept?

Consider the following argument:

1. The best accounts of "x knows p" are conjunctive, i.e., of the form R₁(x,p) & R₂(x,p) & ... (with at least two conjuncts).
2. Therefore, probably, knowledge is a conjunctive concept. (See this post and this one)
3. No conjunctive concept is natural.
4. Therefore, knowledge is not a natural concept.

Alternately, one could claim that knowledge is a natural concept, and hence the best accounts of "x knows p" are false.