Yes, valuing happiness context-dependently is more general and to be distinguished from fearing its "irrationality". Interestingly, until people reflect to distinguish these cognitive phenomena, they might not yet be operating separately in their brainware, in which case how they reflect can affect the development of their values and hence the conclusion. If one has been reflecting (consciously or unconsciously) with the bias that "rationality" is the only considerable meta-value of feelings, then one is more likely to end up in the second category.