I have not seen this in specs before. I've posted a similar point in one of the forum that can a non-plastic soil have a liquid limit? Interesting answers - you should do a search for it. It might mean that they do not want the fines to be "sensitive" to water, e.g., erosion or movement when saturated - but this would be handled in a subbase if the subbase is filter graded within itself (see Terzaghi, Peck and Mesri, 1995). Sometimes, the subbase specs are so widely graded and non-uniform that the fines of the subbase can wash through. The specifyer might be trying to determine "how much water" it will take to move the fines. I think that in such a case, they are mis-using the test. I also find it interesting in that they have specified the LL to a 0.5 number when, according to test procedures, LL is to be rounded to nearest hole number. I have a suspicion on why, but . . . and, if someone can honestly say they thay can distinguish differences in hard geotechnical properties from the liquid limit of LL = 20 or LL = 21, I'd like to meet that person! Very interesting to say the least. Doubtful this spec was written by a geotechnical engineer or at least one with experience.