Levelt's model of speech production

levelt.tex

\documentclass[12pt]{standalone}
\usepackage{tgtermes}
\usepackage{tgheros}
\usepackage[T1]{fontenc}

\usepackage{tikz}
\usetikzlibrary{arrows}
\usetikzlibrary{arrows.meta}
\usetikzlibrary{calc}

\begin{document}
\begin{tikzpicture}
% Box 1:
\draw (0,7.1) rectangle (4.4,10);
\draw (2.2,9.65) node {\textsc{Conceptualizer}};

% Message generation
\draw (0.2,8.2) rectangle (2.2,9.3);
\draw (1.2,8.7) node[align=center] {Message\\generation};

% Monitoring
\draw (1.7,7.3) rectangle (4.1,8);
\draw (2.9,7.6) node {Monitoring};

% Monitoring to message generation
\draw[-Latex] (3,8) -- (3,8.4) -- (2.2, 8.4);
% ..and vice versa
\draw[-Latex] (0.8,7.65) -- (1.7,7.65);
% Down and out
\draw (0.8,8.2) -- (0.8,6.8);

\draw (1.6,6.6) node {Preverbal message};

% Box 2:
\draw (0,6.2) rectangle (4.4,1.3);
\draw (2.2,5.8) node {\textsc{Formulator}};

% Continuation of the line
\draw[-Latex] (0.8,6.4) -- (0.8,5.3);

% Grammatical encoding
\draw (0.2,5.3) rectangle (2.7,4.2);
\draw (1.5,4.7) node[align=center] {Grammatical\\encoding};

\draw (0.8,4.2) -- (0.8,4);
\draw (1.1,3.5) node[align=center] {Surface\\structure};
\draw[-Latex] (0.8,3) -- (0.8,2.6);

% Phonological encoding
\draw (0.2,2.6) rectangle (2.7,1.5);
\draw (1.5,2.0) node[align=center] {Phonological\\encoding};
\draw[-Latex] (2.3,2.6) -- (2.3,4.2);

% Down and out
\draw (0.8,1.5) -- (0.8,1);
\draw (1.4,0.5) node[align=center] {Phonetic plan\\(internal speech)};
\draw[-Latex] (0.8,0.1) -- (0.8,-0.3);

% Articulator
\draw (0,-0.9) rectangle (4.4,-0.3);
\draw (2.2,-0.6) node {\textsc{Articulator}};

% Lexicon
\draw (6,3.5) circle (1.3cm);
\draw[dashed] (4.7,3.5) -- (7.3,3.5);
\draw (6,3.74) node[align=center] {\textsc{Lexicon}\\lemmas\\forms};

% Lines to the lexicon
\draw[dashed,shorten <=1.3cm] (6,3.5) -- (2.7,4.7);
\draw[dashed,shorten <=1.3cm] (6,3.5) -- (2.7, 2.05);

% Box 3:
\draw (7.6,6.2) rectangle (12,1.3);
\draw (9.8, 5.3) node[align=center] {\textsc{Speech-}\\ \textsc{comprehension}\\\textsc{system}};

\draw[dashed] (7.3,4) -- (7.6,4);
\draw[dashed] (7.3,3) -- (7.6,3);

% Audition
\draw (7.6,-0.9) rectangle (12,-0.3);
\draw (9.8,-0.6) node {\textsc{Audition}};

\draw[Latex-] (9.8,1.3) -- (9.8,0.8);
\draw (9.8,0.5) node[align=center] {Phonetic string};
\draw (9.8,0.3) -- (9.8,-0.3);

% Articulator to audition
\draw[-Latex] (2.2,-0.9) -- (2.2,-1.5) -- (9.8,-1.5) -- (9.8,-0.9);

% Phonetic plan to speech comprehension
\draw[-Latex] (2.9,0.6) -- (8.2,0.6) -- (8.2,1.3);

% SCS to Monitoring
\draw (9.8, 6.2) -- (9.8, 6.4);
\draw (9.8,6.6) node {Parsed speech};
\draw[-Latex] (9.8, 6.8) -- (9.8, 7.65) -- (4.1, 7.65);

% Message generation to background knowledge
\draw (9.8,8.95) node[align=center] {Discourse model,\\ situation knowledge,\\ encyclopedia, etc};
\draw (9.8,9) ellipse (2.1cm and 1cm);
\draw[dashed,shorten >=2.1cm] (2.2,9) -- (9.8,9);
\end{tikzpicture}
\end{document}

Here's the image produced by this LaTeX document: