bibTeX and homonymy
How comes BibTeX is unable to spot homonyms?! Namely, if I quote two of my 1996 papers in the same LaTeX document, they will appear as Robert (1996a) and Robert (1996b). However, if I quote two different authors (or groups of authors) with the same surname, Martin as in the above example, who both happened to write a paper in 2014, BibTeX returns Martin (2014) and Martin (2014) in the output, hence it fails to recognise they are different authors, which is just weird! At least for author-year styles. I looked on Stack Exchange TeX forum, but the solution I found did not work with the IMS and Springer styles.
Xi'an's Og 
October 31, 2015 at 2:43 pm
The problem is easily solved by not using bibTeX.
What is supposed to be the benefit of bibTeX when using (as you should) the Author (Year) citation style? You only occasionally need to resort to manually qualifying dates as 1994a, 1994b, etc., and as you see, bibTeX doesn’t handle that properly anyway. It can’t really handle it properly, since fixing problems may require that you forego “et al” in favour of listing authors explicitly. (And you don’t want bibTeX to generally decide whether to use “et al”, since you don’t want to use that on first reference, or when it would result in concealment of self-reference.)
Some people like the idea that bibTeX can automatically update their citations to reference new versions of papers. But that’s not actually a benefit if you haven’t read the new version – quite the reverse!
October 31, 2015 at 3:59 pm
Thanks, Radford. Points taken. My use [if not uttermost defence] of bibTeX is related with (a) handling a single bibliography file that I can use and update for every single paper, book, referee’s report, letter, talk, course that I write, (b) the general availability of a bibTeX entry for papers I want to quote, (c) a simplicity in handling reference quote in my document since \cite{neal:1996} is as easy as writing Neal (1996), without to cut and past the entire reference from a source file, (d) the possibility to update references from preprints to published versions. (I do agree that quoting a new version that one has not read opens a Pandora box for misquoting and missupporting a paper.) There are cons, obviously, as the exotic difficulty I mention in that post, albeit the very first occurrence for me, as the possibility for journals to impose a counter-productive Number citation style [as I agree with you that Author (Year) is the most rational way to quote papers in a readable manner] or to cut the references by removing the titles of the papers, as the need to carry a humongous bibTeX file when joining a co-authorship, or as the difficulty to merge the bibTeX files of two or more co-authors, even though there are merging tools…
October 31, 2015 at 11:30 am
I ran into a similar problem the other day — and had similarly poor experience with the various TeX hacks. In the end I went with a simple solution — replacing the authors field in the bibtex with “Author, First et al.” instead of the full author list. This forces TeX to treat them as a conflict and use the letter identifier in the citations.
October 31, 2015 at 2:42 pm
My solution was to quote the first two authors and the date by \citeyear{martin:1994a}.