If a claim is made that the nature of truth consists in having attribute y, then that claim must itself have the same attribute y (if it is to be true) in which case it is potentially recursive.
If the attribute y must apply prior to the condition of truth in order to substantiate the claim then the attribute y requires a further prior condition of truthfulness — and so to infinite regress.
Claims about the nature of truth are vulnerable to recursion and regression, and therefore unsustainable by conventional logical standards.
(By analogy: If I need confirmation that a particular object is exactly one metre in length I can apply a ruler and measure it. The truthfulness of the claim about the length depends of the veracity of the metre length I compare it to, and the veracity of this claim depends on comparison to a prior metre length, and so on. And although numerous attempts have been made to eradicate the uncertainty over the exact length of one metre, from the eighteenth century to the present day, there remains a degree of uncertainty (now at the sub-atomic level).
That the length of one metre is established by convention is not in doubt. What remains doubtful is how long one metre is. Therefore, the truthfulness about the claim that a certain object is one metre long refers only to other claims about the length of a metre. There is no final fact to establish the truth.)
By example: If I make the claim that the nature of truth lies in there being a correspondence between a state of affairs in my mind and a state of affairs in the world, then that claim must itself be subject to the same conditions of truth, such that there must be a correspondence between the nature of truth (being a correspondence between mind and world) and something else. What would that something else be? If it is something in the mind or the world then we risk self reference and regression. If it is something not in the mind or world then where would it be?
Note: This is not to argue that no truthful claims can be made, but that claims about the nature of truth cannot be made (at least not without recursion or regress).
