In the field of textual affect sensing many methods have been proposed. These methods vary from keyword spotting techniques, lexical affinity, statistical natural language processing and hand-crafted models. Based on a large scale survey, two profounding theories have been selected for investigation. The first is the proposed work of (Kamps & Marx, 2001) which states that the lexical relations found in WordNet (Fellbaum, 1998) can be used to measure the activation and evaluation of words. This theory has been investigated, by implementing various search algorithms, including a multi-threaded bidirectional search algorithm, which enables us to compare the results with manually annotated word sets. Improvements to this theory have been made so that for more words the activation and evaluation values can be calculated, without compromising the results. Secondly the theory of (Liu, Lieberman, & Selker, 2003) has been investigated. This theory is based on a novel technique, by inferencing commonsense knowledge to reason about the emotional content of a given text. No full implementation has been made, but a basis has been created for future implementation. Finally, we have implemented a natural language resource toolbox for affective NLP research, called the NLP Affect Toolbox. This toolbox can be used as a programming library to support and fastly implement future research. It can also be used to conduct experiments and to explore the possibilities of state-of-art (affective) natural language processing by experienced programmers, and through a graphical user interface for others.