A great deal of critical enquiry bases itself around the notion that rational discourse is the key method to reach consensus. If all participants act with rationality, if they are allowed to voice and hear other voices, consensus can be reached. This forms the basis of some current popular social theories as well as the more liberal strand of thought which permeates discussion. Ideally, such a basis cannot really be faulted. In practise however the flaws in such an approach become apparent.
Rationality assumes that one will both reason correctly and choose the optimal method for resolving the object of discussion. If the issue is concern for one’s son’s schooling, one could reason attendance at a better school would aid his education. As limits on access to schools in the area may be in effect one may instead choose to pay for private tutoring (for example). Thus a rational conclusion to the initial problem is reached. The proposed problem though conceals the normative bias which led to the question being raised in the first place: the desire to better educate one’s son. Whilst this may seem natural, and perhaps is, the rationality presented here is contingent on an unvoiced concern which is considered as self-evident.
Rationality operates on these normative biases, voiced or not. When we raise rationality then as a mode of consensus-forming, to reach consensus implies that those taking part must either share the same norms or be brought round to sharing, or acknowledging, them. If there is no shared basis for rational discussion then there can be no consensus formed, only a tacit agreement. Furthermore the normative bias one has completely colours any other argument emanating from the otherwise rational discussion. Take smoking. One can agree with the economic and health arguments against it whilst still doing it because it is enjoyable. This is often described as hypocritical or ‘stupid’ but is neither: it is merely the highlighting of transitory joy as a bias for action rather than the more hard-nosed, and quantifiable, drawbacks to the action. All the arguments for smoking here are well reasoned and can be optimally formed to resolve the point at hand. What cannot be escaped though is the bias which colours one’s action; the only way to resolve this fully is to switch one’s observation of the problem to the normative standpoint of the rest.
This problem has hounded political thought since the days of Rousseau. Rousseau considered that society was led by a “general will” of the people which transcended mere individual desire to consider what is best generally. If one did not follow the general will one would have to be “forced to be free”. J. L. Talmon in his excellent book The Origins of Totalitarian Democracy, published in 1952, highlights how the demand for rationality and consensus quickly becomes totalitarian in this Rousseauian fashion. A demand for rationality is thus in actual practise a demand for a “general will”; a demand for all to follow the same normative basis without question. If consensus can only be reached under these conditions it seems difficult to place rationality as the foundation of meaningful discussion.
What is left then? Rationality evidently exists but it exists, in my opinion, closed off within the minds and systems that make up society. I can be brought round perhaps to another’s argument but only through understanding it on my own terms. In wider society, at the level of social system, the economy doesn’t understand climate change science but in terms of money; the law doesn’t understand governmental spying but through its legality. The systems that create society have their own rationalities through which they understand the world and the same problems and these rationalities are irreducible. A rose may smell just as sweet by any other name but to a man with a hammer every problem will always remain a nail.
ssptmusing 
This might be more an issue stemming from your examples but I think what’s at issue here is values, rather than norms.
The generalised injunction against smoking is a norm (specifically a purity norm) but since it’s conventional the decision to smoke is one based on value – the smoker values the enjoyment they get over the supposed health benefits of refraining from smoking (in part this is also a matter of risk and judgement, and we can still say their values are short-sighted and their assessment of the risk is poor, when we consider these things objectively). On the other hand, if it were a legal norm – as in the case of not smoking in enclosed public spaces – the decision might be based on the likelihood of getting away with it. Based on their values and resulting objectives, their rationality and reasoning is more or less fine, although we could say that if they applied their reason to their values they ought to be expected to revise those values.
Norms are externals which can be bent or broken, or applied contextually, whilst to value something is to have something interal to you (of course, values can be internalised) which affects the way you perceive the world and prioritise your possible courses of action. Haidt’s ‘The Righteous Mind’ is pretty interesting on this, since he says certain social conflicts might be impossible to solve since they are based on deeply psychological differences in the kinds of things people value (e.g. “security” vs. “inclusiveness”).
Similarly the mother who makes decisions about her child’s education is value-based. She decides to move to a better catchment area or hire a tutor based on a value she holds rather than a normative bias. She desires a certain outcome (and is pretty open about that) rather than holds a view of how the world ought to be which alters her assessment of it (that said, in desiring something she will very likely have a view of how the world ought to be, built out of that desire, the difference being that this is thus something she works towards). In realising these value-based desires she is being perfectly rational, but in both these cases it is instrumental rationality that is at work.
I still think it’s a point that need to be said; that rationality isn’t as many people imagine it to be. But I think what you’re pointing to here is that rationality as it actually exists, outside of the abstractions of philosophers (and I especially blame Kant for this), is of an instrumental kind. It doesn’t look like there is any kind of categorical or abstract rationality anywhere in the known universe. (And for me this is simple to explain – rationality is a problem-solving tool we evolved. We can turn it into general principles for thought because we are capable of highly complex thought, but it still needs goals and defined contexts to work with, because that’s the kind of thing it was developed to work on.)