Rödl’s inquiry into self-consciousness proceeds as follows:
a) He asks ‘what is the logico-grammatical role of the term ‘I’ in thought and speech?
b) ‘I’ is not an individual concept nor a definite description since there are many things (people) that can be referred to using this first-person pronoun. In other words, ‘I’ is not a name, it is not a person you could locate on a map or by GPS satellite.
c) ‘I’ is not a rigid designator since I can designate myself rigidly as, for example, Oedipus did without knowing that I am the one to whom I am so referring.
d) ‘I’ is a form of reference that is not demonstrative because it is not perceptual since I can perceive myself, for example in a mirror, without knowing that it is I whom I so perceive.
e) Like demonstratives, however, first-person reference is identification-independent: my reasoning about ‘I’ like my reasoning about any ‘this/here/now’ puts me in a relation to the object that is referred to in this way that affords me a certain amount and variety of pure knowledge about how things stand with this object in virtue only of its being referred to in this way.
f) First-person reference allows me to know an object without having to pass through demonstrative reference to it. E.g. I can know that I am he who is writing this sentence without having to check whether or not ‘I’ is identical to ‘these hands which are typing’; identifying myself via demonstrative perception and connecting such a perception to self-conscious thoughts is a further step. ‘I-thoughts’ are more fundamental than demonstrative thoughts since the latter cannot bypass the former.
g) the closest relationship I can have to an object is identity. Being the object is the closest way of knowing it. An object I know only demonstratively is not identity because it is mediated by sensation. Thus, to perceive X is to know X as Other or as Not-Self, even if the object which I perceive in this way turns out to be me or my body. However, with ‘I-thoughts’ I know the object not by perceiving it but by being it. This is a logical category and so far the definition is strictly analytic. It remains to be seen whether or not there are such thoughts that satisfy this definition which can also be understood as thoughts which are self-constituting, or ways of knowing an object such that my knowledge of the object, my positing of its existence and determination constitutes its being as such.
h) Cases in which being = knowing will be peculiar: their truth will be indubitable since posing such thoughts guarantees their truth. The only concepts that seem to satisfy this definition (the indubitability or ‘a priori’ constraint in relation to the use of the first-person pronoun) are thoughts containing action concepts and belief concepts. Thus, there are two kinds of first-person thoughts: practical thoughts and theoretical thoughts. Such first-person thoughts Rodl will connect to two powers of Reason: the Will in relation to practical knowledge and the Intellect in relation to theoretical knowledge. If empiricism takes sensation as the fundamental source of knowledge, then showing that sensory knowledge is logically grounded in first-person knowledge will show that empiricism is unable to give an adequate account of epistemology.