对敦煌归义军曹氏族属问题的讨论,学界持完全不同的两种观点:或出自中原汉族大姓“亳州谯郡”曹氏,或为中亚粟特后裔。其中持曹氏为汉族者重要的证据,即是洞窟壁画中大量出现的曹氏供养人画像之面貌特征,无一例为胡貌人物。我们认为,要单一从洞窟壁画中的供养人画像面貌特征上来确定其族属虽有一定的道理,但从方法论上是值得进一步讨论的,实要注意“图像证史”的“陷阱”。本着这样的思路,我们首先分析了曹氏供养像之写真性概念,然后逐一揭开附着于曹氏供养像表面的非写实因素,最后参证以洞窟相关资料,结果表明从供养像的角度,归义军曹氏粟特胡人的族属关系似有可能。 The discussion of the clan genocide of the Guiyijun in Dunhuang has two completely different views: one is from the Central Plains Han’s surname “Caozhou”, or the Sogdia of Central Asia. Among them, the important evidence of holding Cao as the Han nationality is the appearance of the large number of Cao’s dependents in the murals of the cave, all without exception. In our opinion, although it is reasonable to ascertain the affiliation of the dependents in the mural paintings of the cave, it is worthy of further discussion on the methodology. Therefore, we must pay attention to the trap of “the history of the images” ". In this way, we first analyze the photographic concept of the portrait of Cao’s support, and then uncover the non-realistic factors attached to the surface of Cao’s support image. Finally, we verify the cave-related material and the results show that from the aspect of supportive image It is possible that the clan relations between the CJ Sogdians and the Jiyi Army should be possible.