The paper introduces horizontal membership functions (HMFs) which define a fuzzy set not in form of commonly used vertical membership functions of type μ = f 1 ( x ) but in the horizontal form x = f 2 ( μ ). Until now, constructing HMFs had seemed impossible because of horizontal ambiguity of this function. Now, however, it became possible thanks to the multidimensional, RDM-interval arithmetic based on relative-distance-measure variables. HMFs enable direct introducing uncertain, interval or fuzzy variable-values in usual mathematical formulas of type y = f ( x 1 ,…,x 2 ) together with crisp values, without using Zadeh’s extension principle. Thus, a relatively easy aggregation of crisp and uncertain knowledge became possible. The paper shows application of HMFs, first on example of a classical mathematical function y = f ( x 1 ,x 2 ) and next, on example of a computing with words challenge problem.