I think the original question to be valid and Ciarán Taaffe's answer to be useful: Consider a diagonalisation on Turing machines, computing total computable functions from N to N. Then, G would be given a number, select the corresponding TM from the assumed enumeration, calculate the value and so on. But by assuming G was part of this list, there is a recursion (if G is given its own number in the enumeration) and therefore G does not terminate on that number and, in conclusion, does not compute a total function. With this contradiction, only the assumption that G was on that enumeration, is wrong (sinceespecially since it doeswould not compute a total funtcionfunction if it were) and does not help proving more than that.