It is clear that lying is wrong in general. And there seem to be good reasons for saying that this is true without exception. In the first place, as was said in the linked post, lying always harms the common good by taking away from the meaning of language.
This is also related to St. Thomas’s argument that lying is always wrong:
An action that is naturally evil in respect of its genus can by no means be good and lawful, since in order for an action to be good it must be right in every respect: because good results from a complete cause, while evil results from any single defect, as Dionysius asserts (Div. Nom. iv). Now a lie is evil in respect of its genus, since it is an action bearing on undue matter. For as words are naturally signs of intellectual acts, it is unnatural and undue for anyone to signify by words something that is not in his mind. Hence the Philosopher says (Ethic. iv, 7) that “lying is in itself evil and to be shunned, while truthfulness is good and worthy of praise.” Therefore every lie is a sin, as also Augustine declares (Contra Mend. i).
The idea here is that just as “killing an innocent person” is always wrong, so “speaking against one’s mind” is always wrong, and the harm consistently done to the language and to the common good is a sign of this wrongness. Still, there are cases where it is right to do something that involves the death of an innocent person incidentally, and likewise there could be cases where it is right to do something that incidentally involves speech apparently contrary to one’s thought. But just as such incidental cases are not murder, so such incidental cases are not lying. This post and the previous past are good examples, since I appear to be saying things contrary to my mind.
Even if someone does not accept St. Thomas’s manner of argument, there are reasons for thinking that lying is always harmful even in terms of its consequences. One should consider the consequences not only of the individual act, but also of the policy, and the policy “never tell lies,” seems more beneficial than any policy permitting lies under some circumstances. We can consider the Prisoner’s Dilemma. If everyone has the policy of cooperating, everyone will be better off. Likewise, society will be better off if everyone has the policy of never lying. Of course, not everyone has this policy. Nonetheless, the more people adopt it, the more other people will be willing to adopt it, and the better off everyone will be. Even the typically discussed case of the Nazi and the Jews may not change this. If you tell the truth to the Nazi, it will be bad for the Jews in the particular case, but the world as a whole may be better off because of your policy of consistent truth-telling.
On the other hand, it is also easy to argue that we should make an exception for cases like that of the Jews. In the first place, almost everyone would in fact make an exception in this case, and simply say that there are no Jews. Yes, you could respond with a verbal evasion, if you happened to think of one. But suppose that you are on the spot and one does not occur to you. Your real choice here is simply to say, “Yes, there are Jews,” or “No, there are none here.” If you do not respond, your house will be searched, which will have the same effect as giving an affirmative response. In practice most people would lie. Nor can this be dismissed as moral weakness, the way we can dismiss people’s tendency to overeat as moral weakness. For people regret overeating; they will say things like, “I wish I didn’t eat so much.” But in the case of the Nazi and the Jews, most people would lie, and would never regret it. They would never say, “I wish I had admitted the Jews were there.” This indicates that almost everyone agrees that it is ok to lie in that case: regardless of how they describe this situation philosophically, at a deep level they believe that lying is justified in this case. If you attempt to justify it by saying that it isn’t really lying in that case, then you are simply confirming the fact that you believe this.
And insofar as this is a practical matter, we can make a strong argument for their conclusion as a matter of practice, regardless of the theoretical truth of the matter. Suppose you are 95% certain of the arguments in the first part of this post. You think there is a 95% chance that lying is always wrong, even in the case of the Nazi and the Jews. Now the Nazi is at the door, asking about the Jews in your house. You can tell the truth. In this case, according to your opinion, there is 95% chance that you will be doing the morally right thing, and incidentally allowing the death of some innocent persons. But there is a 5% chance that you will be doing the morally wrong thing. That is, if you are wrong, you will not only be doing something morally neutral: you will be doing something morally wrong, namely allowing the death of innocent persons for no good reason. And the 95% chance is of telling a useful lie, and saving lives. If it is morally wrong, it is a small wrong. The 5% chance, however, is of pointlessly allowing deaths. If it is morally wrong, it is extremely evil.
And it is easy to argue that in practice there is only one good choice here: a certainty of saving lives, together with a 95% chance of a slightly wrong act, seems much better than the certainty of allowing deaths, together with a 5% chance of an extremely evil act.