Nida-Rümelin's approach to practical philosophy is, perhaps, best summarised by his account of “Structural Rationality.” As an alternative to consequentialism, it avoids the dichotomy between moral and extra-moral rationality that is typical in Kantian approaches, and is thus able to integrate a vast complexity of practical reasons that can result in coherent practice. Whereas for Kant, rule orientation is constitutive for moral agency, the structural account of rationality extends to the more general idea that rationality consists in embedding momentary or point-wise optimization within the broader structure of agency. The intimate connection between morality and rationality that Kant postulates becomes one aspect of the broader, all-encompassing account of acting in such a way that fits into a desirable structure of agency. The practice of giving and taking reasons is understood as aiming at both interpersonal and intrapersonal structural coherence. In this way, the account of structural rationality avoids the dichotomy of reasons - moral versus extra-moral - and allows us to make use of the conceptual frame of decision and game theory in order to clarify some essential aspects of practical coherence. For example, the postulates of the von Neumann/Morgenstern utility theorem are now interpreted as rules of practical coherence and not as axioms of consequentialist optimization. The utility function becomes a mere representation of coherent preferences and expected utility maximization can no longer be interpreted as optimizing the consequences of one’s actions. The term “utility” is now misleading and should be replaced by “subjective valuation.” The deontological character of structural rationality is compatible with using the conceptual framework of decision theory. This may come as a surprise, but is only due to a logically stringent interpretation of the utility theorem and other theorems of decision and game theory. The usual economic interpretation is only one among many others and in fact, this interpretation is incompatible with most of the practical reasons that we take to be indispensable. The structural account of rationality does not differentiate itself from the everyday practice of giving and taking reasons.
These considerations led Nida-Rümelin to discuss the relation between philosophy and Lebenswelt (life-world) and Lebensform (life-form). His position is inspired by Wittgenstein's philosophy. The difference is primarily that the later Wittgenstein seems to think that different language-games are to a large degree autonomous, whereas Nida-Rümelin emphasizes the unity of practice, viz. the unity of the person (the agent) and societal interactions. We strive for coherence regarding our beliefs or epistemic attitudes, our actions and emotive attitudes. This gives us a reason to criticize any incoherencies and these incoherencies are the starting point for philosophy in general, and for ethical theory more specifically. Philosophical theory has to be careful not to leave the common ground of the human practice of giving and taking reasons. It cannot reinvent reasons; it cannot postulate some principle and deduce moral duties from it. Ethical principles can only be systematizations of a given practice of reasoning. Here, Nida-Rümelin takes the side of pragmatism in its conflict against rationalism.
Humanism is taken as a philosophical perspective that starts from the conditio humana, the common elements of the human condition over time and between cultures. Humanism has both an anthropological and an ethical dimension. The anthropological dimension is presented in core normative concepts like reason or rationality, freedom and responsibility. Humanists think that the ability to act and believe and feel based on reasons is essential in order to understand the human condition. This does not however directly imply humanistic ethics. Human agency requires the ability to weigh reasons and act upon the result of weighing reasons, but it does not ensure morally acceptable agency. Even an officer in a Nazi concentration camp may act by weighing reasons. Humanistic ethics must discriminate between good and bad reasons, good and bad forms of reasoning, good and bad forms of emotive attitudes. Hating somebody because he lives a different life is irrational, as seen in the hatred of homosexuals in a majority heterosexual community, or hatred based on skin colour. The structural account of rationality is optimistic insofar as it assumes that clarifying reasons that strive towardsat intra- and interpersonally coherent agency and belief allow the elimination of bad, misleading reasons of all the three kinds mentioned above (practical, theoretical, and emotive reasons). Therefore, the relationship between anthropological humanism and ethical humanism is not deductive, but pragmatic. Those who take anthropological humanism seriously tend to embrace a humanistic ethos and those who reject humanistic principles of agency tend to fight against anthropological humanism, expressed in different forms: social Darwinism, racism, reductionist naturalism, chauvinist nationalism, discriminating sexism and other forms of anti-humanism. Since communication plays a central role in this form of humanism, Nida-Rümelin presented an account of humanistic semantics. It takes the Gricean model as the starting point, i.e. a speaker’s intentions constitute meaning, but he extends this model with a deontological, more specifically a structural account of reason giving. It is not the effect that the speaker wants to cause in the hearer, but it is the reason the speaker intends to give, leaving it up to the autonomous agent, the hearer, whether he will act on these reasons or not.
Nida-Rümelin defends a basic, non-ontological, non-metaphysical realism against instrumentalism in the philosophy of science, positivism, and post-modernism in practical philosophy, the humanities and the social sciences. Realism is not a metaphysical postulate but it is part of our everyday practice of giving and taking reasons. This practice would not make sense if we did not presuppose that there are facts that we want to find by developing arguments for and against certain hypotheses. Realism is part of our way of life and this makes it implausible to assume that the sciences and humanities can be understood in an anti-realist manner, as is assumed by instrumentalists, constructivists and post-modernists. It is the continuity between everyday life and science that speaks in favor of the proposition that existence is not constituted by belief, neither by the beliefs of an individual (solipsism), or by the beliefs of collectives (cultures) or even ideal discourse communities as seen in the work of Jürgen Habermas, or in Hilary Putnam’s so-called internal realism, which is in fact a version of idealism.
Since naturalism is the dominant metaphysical attitude in the natural sciences as well as in a good part of the social sciences and even in the humanities (sometimes seen in a combination of post-modernism and naturalism), Nida-Rümelin had to show that this implicit or explicit point of view cannot be maintained in spite of its predominance in analytical philosophy. The driving force behind most of his arguments against naturalism is pragmatist: he takes the constituency of human agency as given and tries to show that these constituents are incompatible with naturalism. In other words, it is shown that there is no plausible naturalistic interpretation of reason, freedom, and responsibility. Nida-Rümelin first presented his arguments in a trilogy of books: the first about practical reason (2001), the second about freedom (2005), and the third about responsibility (2011). Practical reason, epistemic and practical freedom, epistemic, practical and emotive responsibility are interpreted as three aspects of the same phenomenon, being affected by reasons. This does not mean that the chain of reasoning doesn’t come to an end. On the contrary, here again Nida-Rümelin is close to Wittgenstein in claiming that all reasoning ends in the indisputable elements of our shared form of life. It is irrational to doubt everything, or as Wittgenstein claims there are some things that a sensible man will not doubt. Since there can be no reasonable doubt that our reasoning is relevant for what we do, this means that what we do cannot be fixed in advance, before deliberation takes place. Whatever that means for the old mind-body problem, it cannot be doubted. To doubt it would mean that we step outside our human condition. As Peter Strawson argued in an influential article ‘Freedom and Resentment,’ being reasonable means that we do not exclusively depend on factors beyond our control; that what we think to be right is relevant for what we do. There cannot be any form of naturalistic determination that excludes reasoning itself. Nida-Rümelin takes the argument against the reduction of logics to psychology (Frege, Husserl) as an argument in favor of the gradual autonomy of reason. In being able to reason logically we participate in logical laws that cannot be identified with psychological or neurophysiological facts. Additionally, he takes the results of Alonzo Church and Kurt Gödel from the 1930s regarding non-computability as showing that our reasoning cannot be exclusively algorithmic. As a naturalist position takes causal processes to be algorithmic, reasoning cannot be naturalized. Humanism in Nida-Rümelin’s sense therefore excludes deductionist naturalism.