If your author refers to something as being like, say, the Litany of the Saints in a Catholic mass, should you, out of concern for the possible opacity of the reference for some readers, replace it with a more general reference such as to “a litany in a Christian mass or liturgy”?
I say, with exceptions, no. In general, prefer the more concrete image, even if it’s not necessarily familiar. If you use a more general descriptor, from being a distinct image (even if one with a referent that needs to be filled in) it becomes a sort of hand-waving. It’s like saying “She had a smile like a famous male or female movie star” rather than “She had a smile like Carol Channing.” And while some readers may not know what the Litany of the Saints is like, since it’s specified, they can find out. (Anyway, not all litanies have the same pattern.)
I’m put in mind of a summer student we had at my company a few years ago who, in defiance of my instructions, replaced “bevelled-edged” with “with slanted-edge” throughout dozens of descriptions of pills. Her position was that not everyone knows what bevelled is. My position was that they could look it up easily, whereas “with slanted-edge” – aside from having an unnecessary hyphen – was not perfectly clear either, nor for that matter entirely accurate, and if you’re not sure what it means you can’t_look it up, because “slanted” is simply too vague a word – no dictionary will give you a picture of a bevelled edge under “slanted.”
This is not an earnest plea in favour of opaque language; if you have a choice between terms that most readers will understand and terms that few readers will understand, then you do need a good reason to use the less common terms (e.g., they give a substantially clearer or more accurate image – if looked up, and they have to be look-up-able – or the use of low-frequency words adds a certain flavour, such as abstruseness or reconditeness, that you actually want). But always remember: a clear image that you can’t see yet but will be able to see if you want is, as a rule (i.e., unless you specifically want unclarity for effect), better than an image that you can see but that’s not clear and never will be.