本研究では，曖昧な文章や多義的な構造を含む文脈依存的文章理解のために，人が理解している文法や概念の包含関係を考慮したネットワーク構造について検討を行った．人が使う辞書を用い，文法構造を反映したネットワークを生成する過程で形成される際の文章特異的な分岐や循環構造など構造 （トポロジー） の可視化を試みた．We have been proposed and demonstrated a tree structure to discriminate the right answer in the word-sentence English matching test, which is built as the network of "principal nouns" obtained from the tight link referring the English dictionary and thesaurus. In this model, we hypothesized that the level of complexity can be used for judgements to determine which is the right answer, however the evaluation and accuracy could not achieve its desired level. In this report, we attempt to visualize individual differences of the tree structure depending on the depth to dig in the dictionary and step into the improvement of the proposed model.