具体化语义观点（embodied semantic view）认为概念分散表征在与其语义特征相关的感知觉运动系统内。使用图片、声音等材料的研究显示：客体与动作概念知识分布式表征在大脑中，包括基本感知觉系统、运动系统以及高级的计划、目标整合和客体功能知识系统。当这些区域受到损伤时，人们无法正常理解相关概念。大量行为研究和脑成像研究发现语言理解与感知觉系统也存在密切联系。具体名词、肢体动词的理解会激活基本视觉皮层和运动皮层。但是，高级的感知觉运动系统，诸如运动计划、目标整合、客体功能知识等系统是否会额外参与相关词汇理解（工具动词理解），目前的研究很少而且不一致。本研究通过对比汉语单字手动词和工具动词来回答该问题。并通过操纵任务，考察语言理解中具体化表征的性质。研究一使用了被动看任务。结果显示工具动词在左侧额叶、顶叶和颞叶等工具使用相关区域均有更多激活，并且在区域之间的功能连接更加复杂。而手动词更多激活左侧前运动皮层，且仅仅显示出背侧联系。所有差异均集中在左半球。这些结果显示高级的感知觉运动系统会更多参与工具动词的理解，这种参与包含了区域之间的联系。此外，工具动词之间显示出与形旁差异一致的效应。研究二使用了动作想象任务，考察研究一看到的动词的具体化表征是否由主动想象引起。结果显示工具动作想象在双侧梭状回，左侧额中回（BA 9）, 双侧顶下小叶（BA 7）, 和左侧颞中回/颞下回后部（BA 21/37）激活更强。而手动词在双侧前运动皮层、辅助运动区效应更强。这显示工具动词与手动词之间的分离效应不再局限于左半球，而是表现在双侧半球。此外，与形旁相关的差异消失。该结果表明主动想象与被动看的机制存在差异，具体化表征并非由主动想象引起。研究三使用了声调判断任务，考察当任务强调语言学信息加工时，具体化语义表征是否受到影响。结果显示工具动词与手动词主要的差异呈现双侧分布，并且集中在双侧额叶盖部和右侧颞叶等与声调或韵律加工有关的区域。而在工具使用相关的区域，没有发现明显差异。但在基本手运动区上仍然观察到激活。这表明，声调判断任务削弱了工具使用相关语义的具体化表征，而对基本手运动运动的具体化表征干扰较小。总之，上述研究显示高级的感知觉运动系统，包括运动计划、目标整合和工具功能知识也参与语言理解。语言理解中的具体化表征与主动想象机制不同，并且受到语言学信息加工的影响。

Embodied semantic theories suppose that conceptual knowledge represents in the corresponding sensory-motor systems. Previous studies using pictures or sounds found that conceptual knowledge is represented in the brain as the distributed network. When these areas are destroyed, the corresponding conceptual comprehension is impaired. Lots of behavioral and neuroimaging studies found language comprehension has close relations to sensory-motor system. Concrete nouns and verbs describing body-part motion could activate basic visual and motor areas. However, for high-level sensory-motor system, such as skill-motor plan system, goal-integration system, and object knowledge system that involve in tool-use action, whether they will play roles in language comprehension, e.g. tool-use verbs comprehension, so far little is known. The current study employed single-character Chinese hand verbs and tool verbs to explore this issue. Besides, we manipulated the language processing task to examine the task influence on the embodied representation of tool verbs. Study one used passive reading task to see the automatic involvement of tool-use network in comprehending tool verbs. Results showed tool verbs elicited stronger activation in areas of left frontal, temporal, and parietal areas, and had more complex connections between these areas. The hand verbs, in contrast, elicited stronger effect in left premotor areas. All contrast effects were found in left hemisphere. These results indicated the involvement of tool-use network in tool verb processing, and the involvement was dynamic, containing the connections among areas. Besides, differences that consistent with semantic radical information were found between tool verbs. Study two used action imagination task to test if the embodied representation has similar mechanism with imagination. The results showed that tool-use action imagination elicited stronger activations in bilateral FG, left MFG (BA 9), bilateral IPL (BA 7), and left posterior ITG/MTG (BA 21/37). The hand verbs, in contrast, elicited stronger activation in bilateral premotor cortex, supplemental motor areas. These results indicated that the trade-off effects between tool verbs and hand verbs were not limited in left hemisphere as they were in passive reading task, but spread in bilateral hemisphere. Besides, the differences between tool verbs disappeared. Therefore, the imagination showed different mechanism from embodied representation. . Study three used tone judgment task to see if the embodied representation is modulated by attention. The results showed that the main differences between tool verbs and hand verbs bilateral spread, and mainly in bilateral frontal lobe and right temporal lobe, which play important roles in tone processing or prosody processing. As for the tool-use related areas, no significant difference was found. However, all verbs elicited activation in premotor areas that relate to basic hand motions. These results showed that the tone judgment weakened the embodied representation of tool-use related semantics, but had little influence on basic action information. In a word, the current studies indicated high-level sensory-motor system involved in language comprehension. This embodied representation was different from imagination and was influenced by the linguistic information processing.