diff --git a/terminology/terminology.tex b/terminology/terminology.tex
index 5bcff53..c4d70f6 100644
--- a/terminology/terminology.tex
+++ b/terminology/terminology.tex
@@ -11,7 +11,7 @@ \section{Grapheme}
 The term \emph{sign} and \emph{grapheme} are considered to be equivalent and 
 interchangeable throughout this work.
 
-\section{Allograph}
+\section{Graph and allograph}
 \textcquote[204]{crystal2010}{Graphemes are abstract units which may adopt a variety of forms 
 \elide Each of these possible forms is known as \emph{graphs}\elide
 There is a vast amount of physical variation in the shapes of graphs that does not affect the 
@@ -94,38 +94,56 @@ \section{Cataloging of signs}
 (e.g. A for animals, M for signs with hands etc.) whereas the third digit is an arbitrary number
 sequencing the different graphemes.
 
-Mayawoerterbuch (classicmayan.org)
+TODO Mayawoerterbuch (classicmayan.org)
 
 
 \subsection{Problems and limitations}
-Having all these sign catalogs are huge help to systematically analyze the Maya writing system.
-Identifying graphemes are crucial to decipher the meaning of the signs.
-One way to do this, is by segmentation of the texts into distinct graphemes.
+Having all these sign catalogs are a huge help to systematically analyze any writing system.
+It is especially important when the writing system cannot be read.
+As one could see above, researchers assign codes or numbers to address individual or 
+even groups of signs.
+Therefore, identifying graphemes are crucial to systematically build up a sign catalog.
+Yet, determine them in an unknown writing system is challenging.
+One way to approach this, is by segmenting the texts into distinct \emph{graphs} .
 Researchers hereby followed the assumption that graphemes of a script are considered the same if 
 they resemble each other in more features than either resembles any other.
 \textcquote[34]{knorozov1967}{Two [signs] are identical when they are both composed of the same 
 graphic elements\elide, whose drawing and disposition is sufficiently similar to allow them to 
-be identified.}
+be identified}.
 However, if there is no control in terms of linguistics and content, 
-identifying graphemes can be challenging and even be problematic.
+this approach can be problematic.
+Three major issues can occur when segmenting signs from an unknown writing system:
+\begin{itemize}
+    \item Allographs are interpreted as separate graphemes.
+    \item Graphemes with distinct phonemes and meanings are interpreted as allographs.
+    \item Complex graphemes are split into its sub-graphemic components.
+\end{itemize}
 
 Especially in writing systems with many allographs like the Maya hieroglyphs,
 allographs are sometimes not recognized and, instead, are interpreted as separate graphemes. 
-Another problem is that some signs are considered to be separate graphemes, 
-but, as later progress in decipherment has shown, were actually allographs.
-Eric Thompson (\cite[12\psq]{thompson1962catalog}) also recognized the method of segmentation as 
+Another problem is that some signs are considered to be allographs because of their similarities, 
+but, as later progress in decipherment has shown, were actually distinct graphemes.
+Eric Thompson (\cite[12\psq]{thompson1962catalog}) recognized the method of segmentation as 
 a potential source of false conclusions.
 David H. Kelley (\cite{kelley1962}) showed in his review of Thompson's sign catalog that
 some T-numbers represent more than one grapheme (e.g. TODO) and some 
 T-numbers are allographs of another (e.g. TODO).
-
-
-
-
-
-
-PAS example for over-splitting (Tikal Temple IV, Lintel 2 A7)
-
+Despite merging unrelated graphs or separating allographs which actually belong to each other,
+the Maya writing system also utilized graphemes which consist of two or more subgraphemic 
+components.
+Those complex graphemes might not be recognized and therefore only its components are registered
+as graphemes.
+One of those complex graphemes, is the grapheme \grapheme{pas} ``dawn'' 
+(\ref{fig:terminology-glyphs-pas}) which is built from
+grapheme \grapheme{chan} ``sky'' (\ref{fig:terminology-glyphs-chan}), 
+grapheme \grapheme{k\glottalstop} ``k\glottalstop{}in'' (\ref{fig:terminology-glyphs-kin}) and
+grapheme \grapheme{kab} ``earth'' (\ref{fig:terminology-glyphs-kab}).
+It can be found, for example, on Tikal Temple IV, Lintel 2 A7.
+All three components are graphemes themselves, but in combination they form the complex 
+grapheme \grapheme{pas} with its phoneme and meaning.
+This grapheme doesn't show up in Thompson's sign catalog.
+Later revisions and new catalogs like Macri and Looper (\cite{macrilooper2003}) added it as
+separate grapheme and assigned it the code ZX2.
 \begin{figure}
     \centering
     \begin{subfigure}[b]{0.24\textwidth}
@@ -158,8 +176,4 @@ \subsection{Problems and limitations}
              (\cite[139]{prager2018}).}
 \end{figure}
 
-
-\section{Characteristics in Maya writing}
-
-
 \end{document}