When using biblatex with the bibtex backend and the option giveninits=true, I have the issue that if any author I cite has a given name beginning with a character including a diacritic, the diacritic is sometimes not rendered correctly in the output. Is there some way of typing the character 'safely' in the .bib file to prevent this from happening?

Update: For most accented characters, the answer is to use something like {\'A} instead of \'{A}, but this does not work for the Turkish character İ (see below), since bibtex replaces {\.I} by {\\bibinitperiod I} in the bbl file. There may be other such exceptions, so a general solution would be nice, but something that deals with this single character would also be ok.

Here is a MWE.

author={Character, \'{A}ccented},

The name of the 'author' is then given as "A. Character" in the output, instead of "Á. Character". In this example, the problem can be solved by using {\'A} instead (as I gather is recommended), but this does not work for all characters -- using \.{I} gives "I." in the output, whereas using {\.I} produces the even worse "bibinitperiodI.", when what I want is "İ.". (It appears that examples where the TeX command is just the backslash and a letter, like \O, are ok, and one gets "Ø." in the output as desired.)

I should remark that I am aware of the obvious solution of using biber instead -- this is what I am doing in practice. (I discovered the problem when switching temporarily to bibtex to work around a different issue on my university's system, which has since been resolved.) I am making my question about bibtex specifically because if this cannot be easily solved by appropriately formatting the .bib file, then it probably counts as a bug in bibtex which perhaps the developers should be aware of. (I am told by a coauthor with an accented first initial that this accent has sometimes disappeared in the bibliography of their own published articles, so it is possible that some journals are actually using biblatex with bibtex -- having a 'robust' .bib file may then reduce the risk of such errors occurring and being missed at the proofing stage.)

  • Did you try with Áccented directly? Default input encoding has been utf8 for more than three years. Also, don't forget toload \usepackage[T1]{fontenc}.
    – Bernard
    Apr 15 at 12:04
  • Yes: this produces à instead of Á, and Ä instead of İ with fontenc loaded, otherwise it produces nothing. Bibtex is introducing these errors when producing the bbl file. (Of course it is fine with biber, as this is part of the point.) Apr 15 at 12:24
  • Some further experimenting suggests that, providing everything is surrounded by braces (like {\'A}), the only problem is with a dot, which bibtex replaces by \bibinitperiod in the bbl file. But there may be more possible accents I have not yet thought of... Apr 15 at 12:32
  • {\.I} is the correct way with BibTeX and should work also with biblatex. But there seems to be a bug in biblatex.bst. I think I have an idea for a fix, but I will have to do a more in-depth investigation later. (Of course with Biber, the UTF-8 character would be nicer but since BibTeX doesn't do unicode that is not an option if you want to use your files with classical BibTeX. [Usually you don't notice that BibTeX doesn't do Unicode as the characters are just passed through, but with initials it becomes important.]).
    – moewe
    Apr 15 at 12:35
  • Refer to tex.stackexchange.com/q/57743/35864 for how to input non-ASCII characters in BibTeX. All of those should work also with biblatex (no matter if you use BibTeX or Biber). If not that could be considered a bug (except maybe the old \i vs i issue.)
    – moewe
    Apr 15 at 12:38

The correct input for İ when you use BibTeX to process your .bib files (regardless of whether you are using biblatex or the classic BibTeX/\bibliographystyle approach) is


or {\.{I}}. See also How to write “ä” and other umlauts and accented letters in bibliography?.

The fact that this does not work at the moment is a bug in biblatex (more specifically in biblatex.bst). It was fixed in commit 3a7fdeb on GitHub. This fix will be included in the next biblatex release (probably v3.17).

The problem of accented letters disappearing or becoming mangled with BibTeX may be due to wrong input. If you are using BibTeX stick to the input discussed in How to write “ä” and other umlauts and accented letters in bibliography?. Missing braces may lead to issues with sorting, case change or initial generation. Using Unicode characters directly will also not work with BibTeX, because standard BibTeX can only deal with ASCII. That means that all Unicode characters not in ASCII range will be read as several characters. Often this is not a problem, because they are passed through unchanged, but it is a problem when BibTeX counts letters, for example for initial generation.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.