The specific communicative needs within the field of engineering and their reflection on the linguistic surface have only been studied so far either in strictly qualitative terms or based on very small sets of data. In order to fully grasp the underlying semantic structures and characteristic ways of expressing them in engineering texts, large-scale corpus studies are required. The present article takes a step towards a reliable description of semantic markers in engineering texts, which also allows to draw conclusions about the underlying semantic-functional profile. Machine-readable expressions that encode certain semantic relations, in particular connectors and prepositions, are the main subject of the analysis and are investigated mainly within our own corpus Gingko. However, these expressions are not examined in isolation but within their linguistic context, ranging from their verbal embedding to surrounding pattern-like structures. Features which will prove to be possible specifics of engineering jargon throughout the analysis will be compared to four other corpora (composed of scientific and non-scientific texts) to identify formal and functional traits of engineering jargon which differ from those in scientific texts in general.