The aims of this paper were to examine the usage of some verbs and the division of verbs into different types from a semantic point of view. Grammar and syntax were only investigated from a semantic point of view in order to establish whether they were compatible with and how they affected semantics. The investigation was carried out in three parts. The aims of the first part, an investigation of previous research, was performed in order to establish a theoretical background against which to test the findings of the paper. In this part, the divisions of verbs by grammar were examined in order to establish which parts of this division derived from semantic aspects. Three of Chomsky’s works were investigated in order to establish to what extent Transformational Grammar contradicts, prohibits or is compatible with semantics and the current investigation. It was found that TG is compatible with semantics, a fact that was pointed out by Chomsky himself. The semantically based division of verbs proposed by Jackson was outlined and also categorisation models as presented by Taylor. The findings of these works as they had a bearing on the paper were presented in part 2, Theoretical Background and Previous Research. The second part was an analysis of the semantic aspects of the verbs be, hear, ready, sing, accumulate and in addition the semantic aspects of the –ing form (V-ing), primarily in connection with the aforementioned verbs. The aim of the second part was to establish in what manner verbs differ semantically, whether they can be divided into different types on a semantic basis and also if Jackson’s model is adequate. Effects deriving from syntactic co-location with words of different lexical categories on the meaning of the verbs were investigated. Also, temporal aspects were investigated in order to establish whether these were derived from the verb itself or if the verb was non-temporally delineated and, in the latter case, the means by which temporal aspect was given. Furthermore, each verb was investigated in order to establish whether it was monosemous or polysemous. [BE] was found to be a duosemous, non-temporally delineated state verb and an indicator of intrinsic qualities of its subject. [HAVE] was identified as a polysemous, non-temporally delineated state verb and an indicator of properties or qualities of its subject which derive from extrinsic influence. The –ing form (V-ing) was found to be a linguistic device by which action verbs can be made to express non-temporally delineated states with a temporal duration limited by the action of the verb in question. Also, (V-ing) permitted action verbs, as opposed to the state verbs [BE] and [HEAR], to function in the grammatical position of adjectives with the underlying meaning ‘state resulting from action’. This includes verbs identified by grammar as being adjectives. [HEAR] was identified as a polysemous action verb with forty differing meanings and the temporal aspect of hear is defined by the action, event or process which we hear. [READY] was identified as a monosemous action verb, the semantic equivalent of which is ‘superordinate term for the required actions to change the state from [-desired or required state] to [+desired or required state]. The temporal aspect of [READY] is that it is transitory, all other temporal aspects being derived from context. [ACCUMULATE] was identified as an action verb > process or activity and temporally delineated by its (implied) beginning and end. As the sense of temporal duration is acquired through context or specific expressions for time, it was concluded that [ACCUMULATE] is non-temporally delineated albeit the fact that there is a sense of its not being momentary but extending over time. The investigation of Jackson’s division of verbs on a semantic basis showed that the model was inadequate in several respects and as a consequence a revised model was proposed. Jackson’s categories ‘state > quality’ and ‘state > temporary state’ were found to be adequate. ‘State > private states’ was found to consist of action verbs that were either used creatively or in lieu of proper, intrinsic states that could not be perceived until an action of perception was performed. The category was allowed, but only with the proviso ‘Creatively expressed private states’, and the sub-groups retained. ‘State > stance’ was found to be wholly erroneous as it defined actions by animate agents as states but did not deal with inanimate locations. The latter was proposed as a category in its place in the revised model. Jackson’s division of ‘non-states’ into ‘events’ and ‘actions’ based on whether there is an agent or not was found to be impracticable, the creativity of human language allowing us to use most verbs to describe ‘actions’ as ‘events’ and the reverse. It was proposed that these two categories were amalgamated into one, ‘actions’. A first subdivision of ‘actions’ based on degree of transitivity, that is, the action placing (or not placing) its subject and / or object in a spatial relation, was also proposed whilst Jackson’s subdivision based on temporal duration and conclusiveness was retained. (See p. 55 for a schematic representation of the revised system of division of verbs on a semantic basis). The aim of the third part was to establish whether it is possible to incorporate the findings of this investigation into a scheme of semantic sentence analysis. Although the scheme of semantic sentence analysis was not a fully-fledged scheme, it was demonstrated that one is possible and that valuable information that is not evident from a purely syntactic analysis may be gained.