Legal philosophers who espouse an expressivist (or non-cognitivist) meta-ethics may fairly be said to have adopted an anti-normativist position in law. This is so because they hold that moral statements do not aim to report (moral) facts or to describe the (moral) world, but aim instead to express the speaker’s attitudes, preferences, or emotions and thereby to influence the audience, and because they typically also hold that there simply is no normativity in the world. Scandinavian realist Karl Olivecrona, for example, maintains that legal rules are best conceived as independent imperatives, which cannot establish legal relations but which can cause human behavior, and that they can do this because the citizens respect the constitution and so are disposed to obey rules that can be traced back to the constitution. He also maintains that there is no such thing as the binding force of law, arguing that the reason why so many people nevertheless believe in the binding force of law is that they misunderstand the nature of legal statements and fail to see that they do nothing more than express the attitudes (preferences, emotions) of the speaker.

      One could, however, object to Olivecrona’s position, and to the expressivist position more generally, that it cannot account for our practice of discussing questions of legal interpretation and application. For it is clear that we often make use of fact-stating language and assume that in many cases there are right and wrong answers to legal questions. Although I accept moral anti-realism and am sympathetic to Olivecrona’s position, I agree with the critics that the expressivist position in general is a bit problematic in this sense; and I would therefore like to consider in this article Simon Blackburn’s quasi-realism, which aims to account, on the basis of expressivism, for any objectivist (or realist) features of moral (or legal) discourse and to earn in this way the right to discuss moral (or legal) questions using fact-stating language.

      I am going to argue (1) that in order to understand legal reasoning one needs to distinguish between first-order and second-order legal statements and between two types of first-order legal statements, namely, between committed statements and detached statements, and to be clear about which type of statement one is concerned with; and (2) that, on an expressivist analysis, first-order legal statements are not normative at all. I am also going to argue (3) that Scandinavian realist Karl Olivecrona’s anti-normativism is indeed problematic, (4) that expressivism in general is problematic, too, and (5) that Simon Blackburn’s quasi-realism can provide us with an interesting, and in many ways attractive, account of our practice(s) of legal reasoning, conceived as resting on an expressivist foundation. However, I am also going to argue (6) that the modest version of quasi-realism is problematic, because it assumes that clashing attitudes (which are assumed to lack truth-value) are inconsistent; and (7) that ambitious quasi-realism is even more problematic, primarily (i) because it assumes that there is such a thing as a best possible set of attitudes, but also (ii) because it might be self-defeating.