The object of this work is the interface between the grammar and the lexicon of Old Turkic, a language which clearly differs from e.g. Republican Turkish in such domains as morphonology and the causative. Rich formational structures are shown to correlate with syntactic and pragmatic behaviour and nominalization turns out to follow a mostly ergative pattern. The author accounts for the transparency and creation processes of words in runic, Uigur and, as a complement, Qarakhanid Turkic. Special attention is given to the problems of describing wordformation in a dead language. The published corpus, which has grown enormously in recent years, is critically scrutinized; a coherent treatment of derivation and phraseology leads to the discovery of the meaning and use of many lexemes and to fresh interpretations of numerous passages. This, in turn, makes the book an indispensible tool for the study of as yet unedited (mostly Buddhist) Uigur manuscripts. Most occurrences of derivates are listed, which adds up to a dictionary for this domain of the Old Turkic lexicon. The work should also serve as the base for scientific etymological reconstruction among the Turkic languages on the one hand, and for (areal and/or genetic) Altaic comparison on the other.