I am working on a document in pandoc, converting to LaTeX for pdf output. For some reason, it seems that \ldots is not rendering properly, always giving tighly spaced periods (...). During the conversion, I see that pandoc correctly converts all input ... to \dots, but it seems to render as unspaced periods. So LaTeX appears to understand \dots as three dots, but the spacing is strange. I can't seem to find something in one of my packages that would be causing this to break:

\usepackage[top=2.5cm,bottom=2.5cm,left=2.75cm,right=2.75cm]{geometry} % margins
\usepackage[english]{babel} % language of headings, etc.
\usepackage[notes,backend=biber]{biblatex-chicago} % for chicago style footnotes 
\usepackage[marginal,bottom,ragged]{footmisc} % for hanging footnotes
\usepackage{fnpct} % footnote kerning


\usepackage{graphicx} % for including graphics
\usepackage{booktabs} % for tables
\usepackage{multirow} % for tables
\usepackage[justification=raggedright,font=footnotesize,labelfont={sc}]{caption} % for caption settings

\usepackage[all]{nowidow} % orphan / widow control
\usepackage{setspace} % for line spacing
\usepackage{enumitem} % for creating lists 
\setlist{nosep} % removes too much vertical spacing in lists
\frenchspacing % single space after periods

\usepackage[hyperfootnotes=true,colorlinks=true,allcolors=blue]{hyperref} % hyperref colour

\pagestyle{myheadings} % page number top right




The criteria of \emph{harmonia} {[}harmony{]} are hearing and reason,
not however in the same way\ldots Hearing is concerned with the matter
and the modification, reason with the form and the cause, since it is in
general characteristic of the senses to discover what is approximate and
to adopt from elsewhere what is accurate, and of reason to adopt from
elsewhere what is approximate, and to discover what is
accurate\ldots.Since matter is determined and bounded only by form, and
modifications only by the causes of movements, and since of these the
former\ldots belong to sense perception, the latter to reason, it
follows naturally that the apprehensions of the senses are determined
and bounded by those of reason, first submitting to them the
distinctions that they have grasped in rough outline\ldots and being
guided by them towards distinctions that are accurate and accepted. This
is because it is a feature of reason that it is simple and unmixed, and
is therefore autonomous and ordered, and is always the same in relation
to the same things, while perception is always involved with
multifariously mixed and changeable matter, so that because of the
instability of this matter, neither the perception of all people, nor
even that of the very same people, remains the same when directed
repeatedly to objects in the same condition; but it needs, as it were as
a crutch, the additional teaching of reason.


You can even see that I tried to make the ellipsis spacing huge (12 pt!) to see if it would have an effect, but nothing. Any ideas?

Here is a screenshot of the rendered blockquote above: enter image description here

  • Please provide a full example others can copy and test as is. We cannot do much with a sniplet and an image.
    – daleif
    Commented Jun 18, 2020 at 13:31
  • @daleif I have included the full code for creating the blockquote. Commented Jun 18, 2020 at 13:37
  • Interesting. Exactly which engine are you using (given fontspec)? Note that fontenc should not be used with engines that use fontspec
    – daleif
    Commented Jun 18, 2020 at 14:07
  • @daleif ah good catch, I am using XeLaTeX --- remnant from a previous template I have been expanding. Even when I remove fontenc, it still does not seem to render. I need XeLaTeX as I am accessing a font with special music symbols. Commented Jun 18, 2020 at 14:18
  • It is more interesting than that. There seems to also be an interaction with hyperref, I'm trying to make your MWE as minimal as possible.
    – daleif
    Commented Jun 18, 2020 at 14:21

2 Answers 2


\ldots in text mode expands to \textellipsis which is defined in the format by


Which means it is an encoding-specific command and uses three . separated by tex spaces if the font does not provide something else.

xelatex and luatex use TU (Unicode) encoding , and tuenc.def has

\DeclareTextSymbol{\textellipsis}        \UnicodeEncodingName{"2026}

So it will use U+2026 which looks different in different fonts but in the current font here it looks like: …

  • I see, this is what I would have thought. However, it seems to be a bizarre interaction with hyperref package as @daleif has discovered above! Commented Jun 18, 2020 at 14:25
  • @ThomasNicholson it seems to be the order og hyperref and ellipsis
    – daleif
    Commented Jun 18, 2020 at 14:29
  • Yes, I just found a reference to that here tex.stackexchange.com/a/43959/202950 Commented Jun 18, 2020 at 14:31
  • 1
    @ThomasNicholson the author of ellipsis really ought to mention that in the manual.
    – daleif
    Commented Jun 18, 2020 at 14:33
  • I will write an email! Commented Jun 18, 2020 at 14:34

Apparently hyperref must be loaded before ellipsis (tex.stackexchange.com/a/43959/202950)

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .