Ive been writing an article on Carolingian mirrors for princes (which is why the blog has been a bit neglected), which has got me worried about the whole concept of mirrors for princes and lay mirrors as genres. The problem is reading several articles which seems to be to define the genre so tightly as to exclude a large number of examples. Alain Dubreucq, for example, argues that a lay mirror must theorise the lay condition, which seems to me to mean that theres only one lay mirror in the Carolingian period, Jonas of Orleans De institutione laicali (instead of four or five texts, which I would argue for). This got me worrying more generally about whether a modern imposition of genre labels is useful or not.
Historians are always, of course, discussing the past in terms that contemporaries wouldnt have used. But in this case, I wonder whether it is meaningful to talk about texts as being in a particular genre/tradition if the authors arent consciously writing in this genre. Its only when an author is conscious of writing a text that belongs to a particular category, and then chooses either to conform or modify/subvert the conventions of the genre, that the genre idea really becomes useful as an analytical tool. If youre writing a history of science fiction, you may want to trace it back to Mary Shelleys Frankenstein, but Im not sure how much extra insight labelling the text as SF actually gives you.
There are some early medieval genres which authors at the period were conscious of, for example, history writing, hagiography, Biblical exegesis. But Im not sure that lay mirrors or mirrors for princes come into this category. They seem to me to be an unrealistically simplified way to pin down something that has at least three aspects: the aim of a text, the intended audience and the form of a text. These are all aspects that are willed (if sometimes not actually realised successfully) by any author. In that sense you can partially define a lay mirror/mirror for princes as one addressed to a specific layperson or ruler and intended for their moral instruction (at least till you get to Machiavelli, the aim is explicitly morally uplifting). The problem is that, at least for princes, there are an awful lot of texts in the Carolingian period of varying forms that have this aim and audience.
What has tended to happen, therefore, is that a particular subset of these texts in form (moderately substantial prose works) have tended to be designated as mirrors for princes. Then the content of these is analysed, in the hope of finding some coherent themes. (The problem is, even then, there isnt much commonality). Some scholars have wanted to limit the subset even further, but I think they can end up tending to put the cart before the horse and (possibly unconsciously) choosing a group of texts as the representatives of a genre and then producing a definition that includes only them.
Is the concept of mirrors for princes useful at all? It does work as a shorthand for texts across the ages which share an audience and function, and thus allows comparative studies. (The collection Im writing this article for looks at mirrors from antiquity to modernity). But, at least for the Carolingian period, its a fairly arbitrary label, and thus potentially misleading. I started doing my thesis on lay moral instruction focusing on the lay mirrors, but soon realised that it was unrealistic to exclude so many other relevant texts. So my studies expanded to fill a vast field of works for a lay audience with a moral message, and I still had to use arbitrary limits to exclude some categories of material. I worry that maybe the only purpose of the label mirrors for princes is to produce a manageably small corpus of texts: its noticeable that Hans Hubert Antons work on the Carolingian ruler ethos, which has a fairly wide definition of mirrors, is a long book. In my article I dont problematise the label lay mirrors or mirrors for princes – instead I try and show that some mirrors for princes have a different tone from discussions of ordinary lay morality. But Im not sure (though I dont say so), whether its just the particular texts Ive chosen that fit this pattern, or whether if you include all the relevant texts you can detect any patterns/themes at all. The articles due in soon and is already bumping against the word limit, so I will have to cop out of discussing these issues in it. Another on my long list of issues to consider in further research, I guess.