Several approaches have been proposed to improve meaning negotiation in the pragmatic web. This section presents some of these works. To begin, the approaches will be classified into four different groups according to their representation in context: logic, ontology, neural network model, and hybrid. These approaches use ontologies to model the context. It represents a conceptual group in a well-defined field and the relationship between the different concepts 16.The first Approach is that of B. Magnini in 2002 17, it presents an algorithm of automatic meaning negotiation that allows the semantic interoperability (to find correspondences) between the local ontology and the heterogeneous ontology for different autonomous communities. The authors use the semantic dictionary in order to improve and disambiguate the meaning of the concepts, to eliminate irrelevant concepts in a context. They did not take into consideration relationship labels. In 2005 A. De Moore 3 established a pragmatic model to place ontology in context and operationalize the pragmatics of the Web. It allows presenting a negotiation scenario (sale) between a cat mat seller (matmaker) and an association of cat lovers (CLAW). It represents the basic model of the meaning negotiation that we will take into consideration in our future work. It is very difficult to find a good granularity of context at the pragmatic level with domain ontology.Then, A. De Moore in 18 improved the ambiguities of communication in the meaning negotiations between communities by its DOGMA model. The negotiation process becomes very complex when the community of practice is large especially in the step of interaction with the ontology engineering layer.The authors of 19 implemented 3 protocols in 2007, for ontological negotiation in the Internet News area of the Anemone system that implements normal communication, ontology alignment and a transition between these ontologies. These protocols are evaluated according to the criteria of strength and lossless, soundness, laziness and the minimal construction of the common vocabulary. These three protocols will become complex when several agents interact. Another approach presented in 20, where the authors have developed a formal framework that offers a negotiation strategy to compare the set of contexts of two Backgrounds domain theories by calculating the relationships between concepts, instance, properties and the constraints on the properties while using different measure of similarity. This process can be very long if the SD is still highly evolved, the orphans still existed.Finally, the paper 21 integrates ontological negotiation into a communication mechanism for a multi-agent system with the use of the notion of translation between ontologies; and algorithms to compute these translations; it guarantees a meaningful communication. For successful communication, it allows agents to exchange factual and terminological knowledge in an individual domain.