\begin{abstract}
In this dissertation, a computational analysis of 
Turkish in the Government-Binding (GB) domain has been realized. 
The first
 principles and parameters (P\&P) parser for Turkish 
has been developed by using
 a P\&P development environment, namely PAPPI system. 
The language-specific and universal features of Turkish have been investigated
and the values of Universal Grammar (UG)
parameters for Turkish have been determined. 
All major syntactical structures and many interesting features of Turkish,
including  the embedded sentences,
 passivization, causativization,
questions, pro-drop, complement-drop, empty categories and anaphors,
have been analysed and 
explained by the interaction of universal principles in P\&P framework.
Thus, the parser achieves a substantial linguistic coverage.

This thesis also attempts a GB analysis of word order variation 
 in Turkish. Pragmatic functions 
(topic, focus, background) are associated with
syntactic positions in Turkish. 
The interconnection between two independent systems of (subject, direct\_object,
indirect\_object) and (topic, focus, background) has been investigated,
and a new X-bar structure in which particular structural positions are
associated with topic, focus, and background has been discussed.
The effects of scrambling on
binding and quantifier scopes have also been considered.

During the analysis, the existing mechanisms of UG have been utilized and 
it has been tried to obey all universal principles. However, it is needed 
to re-evaluate, extend or override the universal principles in some cases.
This research provides a supporting evidence and
a critical evaluation for GB's conception of UG.
\end{abstract}