Have you ever made an “immoral promise,” i.e. have you ever promised to do something wrongful, like tell a lie or steal? (By the way, how does one decide when something is wrongful?) Is such a promise even logically possible; specifically, is an immoral promise a paradox or a contradiction? Here is an excerpt from our work in progress “Immoral Promises“:
… when the promisor makes an immoral promise, is the promisor under a moral obligation to keep his word? From a logical perspective, there are four possibilities: I. No: no promises are morally binding. II. No: only moral promises are morally binding. III. Yes: all promises, even illegal ones, are morally binding. IV. Yes: only illegal/immoral promises are morally binding.
We can safely ignore option IV, since it is not only paradoxical but also nonsensical. This leaves us options I, II, and III for consideration. Of these remaining logical possibilities, option II appears to be the one that is most likely to be true, but is it? For this option to be true we must have some reliable method of distinguishing between ‘good’ and ‘bad’ promises, moral and immoral ones.
prior probability 
Promises made by one member of a criminal enterprise to another member of said enterprise are taken very seriously and may be regarded by the participants as “moral” — recognizing that the “local” moral code differs from the global one. So I think one has to tread carefully here as to the defintiion of moral and immoral for the purposes of your paper. Also, one might consider that there is a third possibility beyond “paradox” and “contradiction” — simply “contractual”.
It seems that the only paradox involved here is, “can you trust a liar to keep his word.” If you phrase this instead as, “can you trust an untrustworthy person to fulfill his side of the contract” then the answer is: maybe not — so the “solution” (so to speak) is to elevate the sanctions that apply if the contract is not fulfilled.
An interesting point for your treatment might be, when does a “promise” become an “obligation”… i.e. what level of sanction (to apply if the “promise” is broken) turns a promise into something else? Is a promise *only* an honor-related concept? Or is there always a hidden “or else” involved?
These are excellent observations. I especially agree with your point about “local moral codes,” and (like Atiyah), I find the “high moral tone” of many theorists in this domain to be unhelpful.