Objective: Brain trauma evidences suggest that the two grammatical categories of noun and verb are processed in different regions of the brain due to differences in the complexity of grammatical and semantic information processing. Studies have shown that the verbs belonging to different semantic categories lead to neural activity in different areas of the brain, and action verb processing is related to the activity of motor and pre-motor areas of the brain. Researchers use different tasks to evaluate action verb processing. The most common tasks are action naming and action fluency tasks. Although these types of tasks are sensitive to deficits in action verb processing, they do not specify the nature of the injury. To understand whether dysfunction in action verb processing is due to difficulty in lexical access or specific impairment in semantic processing, it is necessary to design a specific test to evaluate lexical-semantic processing. Semantic Similarity Judgment (SSJ) test targets the lexical-semantic encoding at a deep and controlled processing level. The purpose of the present study was to develop a SSJ test for Persian action verbs and non-action nouns and determine its content validity. Materials & Methods: In this methodological study, 70 Persian action concrete verbs and 80 Persian non-action concrete nouns were first selected. For each word, a semantically related word based on functional, physical, categorical features and similarity in action was selected according to the opinion of 4 experts (3 speech-language pathologists and one linguist) using a 7-point scale. For semantic similarity rating, only the pairs of words with a high semantic similarity score (5 to 7) remained and the rest were omitted. Then, for each pair of semantically related words, a semantically unrelated word was selected. After determining content validity qualitatively by three experts and removing inappropriate items, for matching the two sets of nouns and verbs, the lexical and psycholinguistic characteristics of the remaining words (207 nouns and 156 verbs) including frequency, number of syllables, phonemes, letters, phonological and orthographic neighbors, action association, imageability, familiarity and age of acquisition were extracted by 18 volunteers (13 speech-language pathologists and linguists and 5 parents selected by a convenience sampling method) based on a 7-point scale. The verbs with low action associations and the nouns with high action association were removed and then, the two sets of words were matched for other lexical and psycholinguistic characteristics. Finally, 34 triples of verbs with high action association and 34 triples of nouns with low action association were selected. In both noun and verb sets, the words were chosen in such a way that, in order to judge, the semantic features of the words need to be carefully considered. Data analysis was performed using descriptive statistics and independent t-test.
Reducing redundancy in multi-document summarization using lexical semantic similarity
