What you are seeing is to do with the way that biblatex tries to identify whether names are "unique" or not. It has nothing to do with the number of citations. What triggers the change is the arrival of "R J Arnott" on the scene, to join "R Arnott".
You don't say what style you are using, but I'll demonstrate using the standard authordate
style. Suppose we have the following MWE:
\documentclass{article}
\usepackage{filecontents}
\begin{filecontents}{\jobname.bib}
@book{jones:1,
author = {Jones, R.},
title = {Titular Sees},
publisher = {PubCo},
date = {1900}}
@book{jones:2,
author = {Jones, R. L.},
title = {Titles of Nobility},
publisher = {PubCo},
date = {1901}}
\end{filecontents}
\usepackage[style=authoryear]{biblatex}
\addbibresource{\jobname.bib}
\begin{document}
\cite{jones:1}
\end{document}
We have a .bib
file with two works by someone called Jones, one with the initial "R" and one with the initials "R L". These may or may not be different people.
However, so long as we cite only one of them, we get just the surname:
Jones 1900
Now if we change our citation so that we cite
\cite{jones:1, jones:2}
We get something different:
R. Jones 1900; R. L. Jones 1901
That's perfectly reasonable. There's no way for biblatex to "know" that this is the same person (even we can often not be sure).
However, if we think it's wrong, there are two things we can do. The first is to set the option uniquename=false
when we load biblatex. With that option set in the standard authoryear style, the same citation now gives us
Jones 1900; Jones 1901
In an author/year style biblatex will still be sensible enough to add necessary disambiguation. For instance, if we add an entry for Dr B Jones's work, also published in 1900, and uniquename
is false, we would then get labels of "Jones 1900a" and "Jones 1900b" (whereas if uniquename
is true, we don't need and won't get the additional letters, because the initials already disambiguate).
The second possibility of course is to change our bib
entries so that they are absolutely identical, at which point biblatex
will no longer attempt to distinguish them. So if we change the author of jones:1
to Jones, R. L.
, then even with uniquename=true
we get
Jones 1900; Jones 1901
In general, though, I think that a bibliographer would be chary of this solution. Strictly speaking a bibliography should accurately record the information as contained in the published book, and "silently" adding or removing initials is suspect. I daresay however that you can get away with it.
uniquename=false
.\documentclass{...}
and ending with\end{document}
.