LaTeX II: Thesis Workflows, Bibliography, and Advanced Tools

Session 1 Recap

Document Classes for Longer Works

For documents longer than a typical journal article, choose the right \documentclass:

Most universities provide a custom class (e.g. unikoeln-thesis) that presets all formatting.

Useful preamble options for longer documents:

\documentclass[12pt, a4paper, twoside, openright]{report}
\usepackage{hyperref}     % clickable cross-references and PDF metadata
\usepackage{cleveref}     % smart refs: \cref{fig:x} → "Figure 1"
\usepackage{microtype}    % improved line breaking and spacing
\usepackage{setspace}     % line spacing (\onehalfspacing, \doublespacing)

hyperref · cleveref · microtype · setspace

Multi-file Projects: \input and \include

For long documents, split your .tex source into multiple files.

\input{filename} — inline inclusion:

% main.tex
\documentclass{report}
\begin{document}
\input{chapters/introduction}
\input{chapters/methods}
\input{chapters/results}
\end{document}

• Inserts the file content directly — equivalent to copy-pasting
• Can be nested; no page break is forced

\include{filename} — chapter inclusion with page break:

\include{chapters/introduction}  % forces a new page before and after

• Supports \includeonly{chapter1,chapter3} to compile only selected chapters — faster iteration on large documents

LaTeX Engines

A LaTeX engine is the program that reads your .tex file and produces output. Several exist — choose based on your font and language needs:

Browser-based engines (no installation):

Bibliography: How It Works

LaTeX separates the bibliography database (a .bib file) from the formatting style (applied at compile time).

Why multiple compilation passes?

Each pass does something different:

pdflatex main.tex   → compiles document; writes citation keys to main.aux
biber main          → reads main.aux; looks up keys in main.bib;
                      writes formatted bibliography to main.bbl
pdflatex main.tex   → reads main.bbl; inserts bibliography into PDF
pdflatex main.tex   → resolves all cross-references (page numbers, labels)

Skipping passes is the most common reason references show as [?] or the bibliography is missing entirely.

The bibliography system has two independent layers:

• Backend — an external program invoked between LaTeX passes (e.g. biber or bibtex on the command line); reads the .bib database, resolves citation keys, and writes the formatted .bbl file that LaTeX then reads back in
• Style package — a LaTeX package declared in the preamble (the part of your .tex file before \begin{document}); defines the \cite commands available in your document and controls how the bibliography list is formatted when printed

The backend and the style package must be chosen as a matched pair — they communicate via the .aux and .bbl files and expect a compatible format.

Bibliography: Choosing Packages

Preamble examples — the style package is where you declare the pairing:

% Option 1: biblatex + biber + APA 7th (recommended for linguistics / social science)
% biblatex-apa requires babel (or polyglossia) and csquotes for correct
% localisation of quotes and date formats — load them BEFORE biblatex
\usepackage[american]{babel}
\usepackage{csquotes}                          % language-aware quotation marks
\usepackage[backend=biber, style=apa]{biblatex} % style=apa loads biblatex-apa
\addbibresource{references.bib}

% Option 1b: same but with polyglossia (XeLaTeX/LuaLaTeX)
\usepackage{polyglossia}
\setdefaultlanguage[variant=american]{english}
\usepackage{csquotes}
\usepackage[backend=biber, style=apa]{biblatex}
\addbibresource{references.bib}

% Option 2: natbib + BibTeX (journal template compatibility)
\usepackage[round, authoryear]{natbib}
% no \addbibresource — use \bibliography{references} at end of document

% Option 3: biblatex + BibTeX fallback (biber unavailable)
\usepackage[american]{babel}
\usepackage{csquotes}
\usepackage[backend=bibtex, style=apa]{biblatex}
\addbibresource{references.bib}

% Option 4: apacite — AVOID (APA 6th only, unmaintained since ~2013)
\usepackage{apacite}
% no \addbibresource — use \bibliography{references} at end of document

APA Manuscript Layout

Citation style (biblatex-apa) and document layout are separate concerns. APA journals require specific formatting — double-spacing, running head, author note, abstract — that goes beyond citation style.

Options for full APA 7th manuscript layout:

Minimal apa7 setup (for manuscript submission):

\documentclass[man, 12pt]{apa7}   % man = manuscript mode; also: jou, doc, stu
\usepackage[american]{babel}
\usepackage{csquotes}
\usepackage[style=apa, backend=biber]{biblatex}
\addbibresource{references.bib}

The .bib File

A .bib file is a plain-text database of references. Each entry has a type, a cite key, and fields:

@article{schepens2024prominence,
  author    = {Schepens, Job and Smith, Anna},
  title     = {Prominence in Language},
  journal   = {Language and Cognition},
  year      = {2024},
  volume    = {16},
  pages     = {123--145},
  doi       = {10.1017/langcog.2024.12},
}

@book{chomsky1965aspects,
  author    = {Chomsky, Noam},
  title     = {Aspects of the Theory of Syntax},
  publisher = {MIT Press},
  address   = {Cambridge, MA},
  year      = {1965},
}

Common entry types: @article, @book, @incollection, @inproceedings, @phdthesis, @misc

Zotero: Building Your Reference Library

Zotero is a free, open-source reference manager widely used in academia.

Collecting references:

• Browser connector imports references from publisher websites, Google Scholar, PubMed with one click
• Drag-and-drop PDFs → Zotero extracts metadata automatically
• Add by DOI, ISBN, or arXiv ID directly

Organising your library:

• Collections and sub-collections (like folders)
• Tags for thematic grouping; related items for linking connected works

Generating citations:

• Word processor plugins (Word, LibreOffice) for inline citations
• Supports thousands of citation styles (APA, MLA, Chicago, linguistics styles …)

Citation Commands in biblatex

With \usepackage[backend=biber, style=apa]{biblatex} and \addbibresource{references.bib}:

Example text:

\textcite{schepens2024prominence} argue that prominence is cross-linguistic.
Several studies support this claim \parencite{chomsky1965aspects, schepens2024prominence}.

Print the bibliography:

\printbibliography                        % full bibliography
\printbibliography[heading=bibintoc]      % include in table of contents
\printbibliography[type=article]          % filter by entry type

Citation in Practice: A Real Example

Using burknerBrmsPackageBayesian2017 from our .bib file:

The .bib entry (BibLaTeX format, auto-exported from Zotero):

@article{burknerBrmsPackageBayesian2017,
  author  = {Bürkner, Paul-Christian},
  title   = {brms: {An} {R} Package for Bayesian Multilevel Models Using Stan},
  journal = {Journal of Statistical Software},
  year    = {2017},
  volume  = {80},
  pages   = {1--28},
  doi     = {10.18637/jss.v080.i01},
}

In your .tex document:

% Narrative citation — author as subject:
\textcite{burknerBrmsPackageBayesian2017} introduced the brms package for Bayesian multilevel modelling in R.

% Parenthetical citation — supporting claim:
Bayesian multilevel models can be fitted using Stan via R \parencite{burknerBrmsPackageBayesian2017}.

% Multiple citations:
\parencite{burknerBrmsPackageBayesian2017, gelmanRegressionOtherStories2020}

Rendered output (APA 7th style):

Bürkner (2017) introduced the brms package for Bayesian multilevel modelling in R.

Bayesian multilevel models can be fitted using Stan via R (Bürkner, 2017).

Quarto: Literate Programming Meets LaTeX

Quarto is a scientific publishing system that produces PDF output via LaTeX while keeping analysis code and prose in a single source file.

A Quarto document (.qmd):

---
title: "My Analysis"
format: pdf
---

## Introduction

We analysed **N = 42** participants.

```{r}
#| label: fig-histogram
#| fig-cap: "Distribution of scores"
hist(rnorm(100))
```

As shown in @fig-histogram, scores are normally distributed.

• Write in Markdown (with LaTeX commands where needed)
• Embed R, Python, or Julia code chunks — output appears in the compiled PDF
• Cross-references (@fig-histogram), citations (@key), and equations all work out of the box

Quarto PDF Output and LaTeX Options

Quarto converts .qmd → LaTeX → PDF via a TeX Live installation.

Configure LaTeX in the YAML header:

---
title: "Thesis Chapter"
format:
  pdf:
    documentclass: report
    papersize: a4
    fontsize: 12pt
    geometry: margin=2.5cm
    keep-tex: true          # inspect the generated .tex file
    include-in-header: preamble.tex   # your custom packages
    cite-method: biblatex
    bibliography: references.bib
    biblio-style: apa
---

Why keep-tex: true?

• Lets you inspect and modify the generated LaTeX if something doesn’t look right
• Some journals require a .tex submission file — Quarto can generate it

Writing a Complete Thesis in Quarto

Quarto supports multi-file book projects — the natural structure for a thesis.

Initialise a Quarto book:

quarto create project book my-thesis

This generates a ready-to-use folder:

my-thesis/
├── _quarto.yml        ← project config: title, author, chapters, output
├── index.qmd          ← front matter / preface
├── chapters/
│   ├── 01-introduction.qmd
│   ├── 02-methods.qmd
│   ├── 03-results.qmd
│   └── 04-discussion.qmd
├── references.bib
└── _extensions/       ← optional: university theme extension

_quarto.yml for a PDF thesis:

project:
  type: book

book:
  title: "My Thesis Title"
  author: "Your Name"
  date: "2026"
  chapters:
    - index.qmd
    - chapters/01-introduction.qmd
    - chapters/02-methods.qmd
    - chapters/03-results.qmd
    - chapters/04-discussion.qmd

bibliography: references.bib

format:
  pdf:
    documentclass: report
    papersize: a4
    fontsize: 12pt
    geometry: margin=2.5cm
    toc: true
    number-sections: true
    cite-method: biblatex
    biblio-style: apa
    include-in-header: preamble.tex   # custom LaTeX packages

Render the whole thesis:

quarto render          # builds PDF (and HTML if configured)
quarto render --to pdf # PDF only

TikZ: Programmatic Figures

TikZ (“TikZ ist kein Zeichenprogramm”) is a LaTeX package for creating figures programmatically — directly in your .tex source.

Why TikZ?

• Figures are plain text — version-controlled alongside your paper
• Perfect vector output at any scale
• Consistent fonts and styling with the surrounding document
• Highly precise control over every element

Minimal setup:

\usepackage{tikz}
\usetikzlibrary{arrows.meta, positioning}  % load useful libraries

Drawing shapes and paths:

\begin{tikzpicture}
  \draw[thick, fill=blue!20] (0,0) rectangle (3,2);
  \draw[red, ->, line width=1.2pt] (0,0) -- (3,2) node[right] {diagonal};
  \node[circle, draw, fill=green!30, minimum size=1cm] at (1.5,1) {center};
\end{tikzpicture}

TikZ: Flowcharts

\usetikzlibrary{shapes.geometric, arrows.meta, positioning}
\begin{tikzpicture}[
  node distance=1.5cm,
  box/.style={draw, rectangle, minimum width=2.5cm,
              minimum height=0.8cm, align=center},
  decnode/.style={draw, diamond, aspect=2,
                  minimum width=2.5cm, align=center}
]
  \node[box] (A) {Step 1};
  \node[box, below of=A] (B) {Step 2};
  \node[decnode, below of=B] (C) {Decision?};
  \node[box, below right=0.8cm and 1.2cm of C] (D) {Yes};
  \node[box, below left=0.8cm and 1.2cm of C] (E) {No};
  \draw[->] (A) -- (B);
  \draw[->] (B) -- (C);
  \draw[->] (C) -- node[above right, font=\small] {yes} (D);
  \draw[->] (C) -- node[above left, font=\small] {no} (E);
\end{tikzpicture}

Named styles keep the node definitions clean and reusable. The decnode style uses TikZ’s built-in diamond shape from shapes.geometric.

TikZ: Plots with PGFPlots

\usepackage{pgfplots}
\pgfplotsset{compat=1.18}
\begin{tikzpicture}
  \begin{axis}[
    xlabel={$x$}, ylabel={$f(x)$},
    grid=major, grid style={dashed, gray!30}
  ]
    \addplot[blue, thick, domain=-3:3, samples=60]{x^2};
    \addlegendentry{$x^2$}
    \addplot[red, thick, domain=-3:3, samples=60]{x^3/5};
    \addlegendentry{$x^3/5$}
  \end{axis}
\end{tikzpicture}

PGFPlots can also read data from CSV files (\addplot table {data.csv}), making it suitable for publication-quality plots directly from experimental data.

TikZ: Syntax Trees

\usepackage{tikz-qtree}
\begin{tikzpicture}
  \Tree [.S
          [.NP Det N ]
          [.VP V
            [.NP Det N ]
          ]
        ]
\end{tikzpicture}

tikz-qtree uses a compact bracket notation. For more control (node styles, movement arrows, coindexation), use the full tikz library directly or forest — a popular alternative for complex phrase structure trees.

TikZ: Dependency Graphs and Glosses

The tikz-dependency package draws dependency arcs over interlinear glossed text — standard in typological and computational linguistics:

\usepackage{tikz-dependency}

\begin{dependency}[theme=simple,
    label style={above, font=\small\itshape}]
  \begin{deptext}[column sep=1.2em]
    Der \& Mann \& gibt \& der \& Frau \& das \& Buch \\
    \textsc{det.m.nom} \& man.\textsc{nom} \& gives \&
      \textsc{det.f.dat} \& woman.\textsc{dat} \&
      \textsc{det.n.acc} \& book.\textsc{acc} \\
    `The \& man \& gives \& the \& woman \& the \& book' \\
  \end{deptext}
  \deproot{3}{root}
  \depedge{3}{2}{nsubj}
  \depedge{3}{5}{iobj}
  \depedge{3}{7}{obj}
  \depedge{2}{1}{det}
  \depedge{5}{4}{det}
  \depedge{7}{6}{det}
\end{dependency}

Linguistics: Interlinear Glosses — gb4e

The gb4e package is the standard for numbered linguistic examples with interlinear glosses following the Leipzig Glossing Rules:

\usepackage{gb4e}

\begin{exe}
  \ex \label{ex:german-dative}
  \gll Der Mann gibt    der   Frau   das   Buch. \\
       the.M man  gives the.F woman  the.N book \\
  \trans `The man gives the woman the book.'

  \ex[*]{
  \gll *Das Buch gibt der Frau der Mann. \\
        the.N book gives the.F woman the.M man \\
  \trans (ungrammatical word order)
  }

  \exi{(1a)}
  \gll \textsc{nom} \textsc{dat} \textsc{acc} \\
       subject indirect.object direct.object \\
  \trans Case roles in the German ditransitive construction.
\end{exe}

% Cross-reference later:
As shown in example (\ref{ex:german-dative}), ...

• \gll starts the word-by-word alignment; \\ separates tiers; \trans adds the free translation
• \ex[*] marks ungrammatical examples; \exi allows custom labels
• Alternative packages: linguex (more compact syntax), expex (most flexible, recommended for complex data)

Linguistics: IPA Symbols

\usepackage{fontspec}
\setmainfont{Charis SIL}   % free SIL font with full IPA coverage
% alternatives: Doulos SIL, Gentium Plus, Linux Libertine

% In running text — paste Unicode IPA characters:
The high front vowel /iː/ contrasts with the lax /ɪ/ in English.
The retroflex stops [ʈ ɖ] are contrastive in Hindi.
Tone is marked with diacritics: /má/ (high) vs /mǎ/ (rising).

% Phonological rules:
/p, t, k/ → [pʰ, tʰ, kʰ] / _V  (aspirated before vowels)

\usepackage{tipa}

\textipa{/\textphi \textbeta/}   % bilabial fricatives [ɸ β]
\textipa{[\textrtailt \textrtaild]}  % retroflex stops
\textipa{/m\'{a}/}               % high tone on /a/

% tipx extends tipa with additional symbols:
\usepackage{tipx}

Linguistics: Optimality Theory Tableaux

The ot-tableau package produces formatted OT constraint tableaux:

\usepackage{ot-tableau}
\usepackage{pifont}  % for \HandRight winner symbol

% Column spec: c = solid border, : = dashed, | = solid between
\begin{tableau}{c:c:c|c}
\inp{/ɾ\textipa{E}to/}
  \const{*Complex} \const{Onset} \const{NoCoda} \const{Max}
\cand[\HandRight]{[ɾe.to]}
                              \vio{*}
\cand{[ɾet]}
               \vio{*}                           \vio{*}
\cand{[e.to]}
  \vio{*!}
\cand{[ɾ\textipa{E}.to]}
               \vio{*}        \vio{*!}
\end{tableau}

Linguistics: Feature Structures (AVMs)

The langsci-avm package typesets attribute-value matrices used in HPSG, LFG, and related frameworks:

\usepackage{langsci-avm}

\avm{
  [ phon & < the, cat > \\
    synsem & [ local & [
      cat & [ head & noun \\
              valence & [ spr & < > \\
                          comps & < > ] ] \\
      cont & [ index & i \\
               restr & < [ reln & cat \\
                            arg0 & i ] > ] ] ]
}

Linguistics: Formal Semantics

Lambda calculus and model-theoretic notation use standard LaTeX math mode — no special package needed for the core notation:

\usepackage{stmaryrd}
\usepackage{amsmath,amssymb,mathtools}

\newcommand{\sv}[1]{\llbracket #1 \rrbracket}

% Lexical denotations:
\[
\sv{\textit{cat}} = \lambda x_e .\ \text{cat}(x)
\]
\[
\sv{\textit{runs}} = \lambda x_e .\ \lambda e_v .\
    ext{run}(e) \wedge \text{Agent}(e,x)
\]

% Quantification:
\[
\sv{\textit{every cat}} =
  \lambda P_{\langle e,t \rangle} .\
  \forall x[\text{cat}(x) \rightarrow P(x)]
\]

% Type reminder:
\sv{\textit{runs}} : \langle e,\langle v,t\rangle\rangle

Debugging LaTeX: Reading Error Messages

LaTeX error messages appear in the log file. The key pattern:

! LaTeX Error: File `missfont.fd' not found.
l.42 \begin{document}

• ! marks the error type
• The next line describes the problem
• l.42 tells you the source line where LaTeX stopped

Common errors and fixes: