Part 1 argued that replicability, an important facet of scientific research, is also found in philosophical thought experiments. Indeed, philosophical thought experiments are easier to replicate than most natural or social science research.
Here, in Part 2, I apply this idea to interpreting texts, whether in the history of political thought, in philosophy, or anywhere else.
My key claim is that when we make an empirical claim about a text – for example, what an author meant by a word or phrase – we should provide our evidence, so that other interpreters can replicate our reading to see if they agree or not. In other words, we should give precise references (e.g. page numbers) so that other people can find the passage, read it for themselves, and see if they share our interpretation.
Aside from replicability, there are two more self-interested reasons to give precise references . First, it forces us to try to be careful. I can think of several occasions where I find that I have misread or misremembered an argument when I look for the page number. Second, it shows our readers that we have tried to be careful. I’m more likely to trust an interpretater if I think that the author has been careful with her evidence, although there are exceptions in both directions, of course.
Unfortunately, sometimes we cannot give precise references, because we have not read the source we are citing, or not read it closely enough, or not read it recently. We don’t always give precise references in informal contexts (e.g. on blogs!) but where possible we should do so in published academic writings. One reason we don’t is the bad academic convention of giving precise references for direct quotations but not necessarily when only citing ideas without quoting them. I believe we should give precise references in both situations.
To change the convention, journal editors and publishers should make us give precise references where we can. I remember one journal editor of a leading political theory journal who considered forcing people to give page numbers in order to get away from slapdash references to “Rawls 1971” and the like. I note with great pleasure that the American Political Science Review now requires authors to give ‘precise page references to any published material cited’. My only caveat to that is where page numbers are not helpful: for example, there are so many different editions of Rousseau’s Social Contract that chapter numbers are probably more helpful there.
But the basic principle stands: ideally, other people should be able to replicate what we have done to see if they agree with our claims. This principle is as important in textual interpretation as it is in the natural sciences.