The interpretation of words in natural language is partially determined by the context in which they are used. Context can refer both to the immediate physical context and to the world knowledge of the speakers. It is known that both of these aspects of context play a role in language interpretation, but the contribution of each aspect, how the two aspects interact with each other and what constitutes world knowledge are open questions. The main goal of this dissertation is to examine these three questions. I focus on relative adjectives, such as 'big' and 'small', whose meaning is typically evaluated relative to a contextually determined "comparison class''. Their explicit context sensitivity makes these adjectives a good test case examining the role of context in language interpretation. To accomplish this goal, I present a series of experiments aimed at showing the interaction of the immediate visual context and world knowledge in the interpretation of relative adjectives, as well as the interplay between different types of world knowledge in this process. The data from these experiments show that world knowledge is integrated with information from the immediate visual context and that at least two different types of world knowledge are involved. I argue that manipulating the stimuli present in the visual context affects the interpretation of the adjectives by shifting the salience of the two types and thus shifting the comparison class relative to which the adjectives are interpreted. The secondary goal of this dissertation is to examine the role context plays in the acquisition of relative adjectives, focusing on a particular subset: dimensional adjectives, such as 'big', 'tall' and 'wide'. To address this question, I present an elicitation study of dimensional adjective antonyms. Dimensional adjectives all designate the physical extent of an object; each specifies a particular dimension (e.g., verticality for 'tall') and polarity. Finally, each is associated with an antonym of opposing polarity; forming this association requires understanding the contribution of all of the adjectives' meaning components. Therefore, studies of antonym acquisition can provide a window into the structure of the semantic field. Through an experiment designed to elicit dimensional antonyms from Hebrew speaking children using pictures for contextual support, I show that context helps anchor children's association of one member of an antonym pair with the other. Their responses reflect the stages of acquisition of dimensional adjectives demonstrating progressive and structured knowledge of these meaning components. Together, the two investigations in this dissertation provide new insight from multiple perspectives on the interaction of context, world knowledge and language.