Frege on Thoughts and Their Structure*

Thoughts are essentially structured. That much is agreed by almost everybody who finds it useful to talk about thoughts. Consequently, any account of the nature of thoughts must incorporate an account of their structure. Frege's is the most comprehensive and worked-out account of the nature of thoughts, but recent work has cast doubt on whether he can be credited with a coherent conception of how they are structured. In an extremely interesting series of exchanges David Bell and Michael Dummett have investigated Frege’s views on the relation between thoughts and the concepts of which they are composed.1 Both authors have identified tensions in Frege’s views in this important area and proposed emendations to smooth out the apparent inconsistencies. The difficulties stem from Frege's simultaneously holding both that the structure of a thought is isomorphic to the structure of a sentence and that two structurally different sentences can express the same thought. In the case of Bell, the proposed cure looks as if it might well kill the patient. Bell ends his latest contribution by proposing a distinction between thoughts and the senses of sentences such that the analysis of a sentence no longer reveals the intrinsic nature of a thought (Bell 1996). The consequence of this, as Bell willingly admits, is that thoughts no longer have a determinate, intrinsic structure. Dummett proposes, on Frege's behalf, to abandon the idea that two structurally different sentences can express the same thought. This is also a drastic revision of Frege's expressed views.