A dialogue is a strategic activity characterized by an intentional structure of competition and/or cooperation. The truth of the statements is far from being the most important thing in many dialogical situations. It is a commonplace to stress that many linguistic constructions, even if they provoke agreement or disagreement (and thus argumentation) have nothing to do with truth or falsity. In particular the ≪prescriptive≫ or ≪projective≫ formulas express that an action must be done (by the speaker, his counterpart or a third party) or that it is forbidden. These statements are not upon the world but upon a project of action that cannot in a strict sense be said to be ≪true≫ or ≪false≫. Nevertheless, it appears to be a permanent or at least a recurrent temptation as well as a strategic attitude for a tenant of the ethical naturalism to maintain that some prescriptions are true or false. Fanaticism can be regarded as the assumption that it is true that certain acts of speech are to be morally condemned or forbidden. I consider that every asserted projective statement can be legitimately denied in four different ways: one can reject the normative dimension, the proposed action, the potential actor(s), the time table. Leaving aside the last two, I use a simple formula to describe the first two moves: A weak form of negation is to deny that there is a prescription to do: ¬JP(A) ≡ I do not judge prescribed to do ‘A’. A stronger opposition is to judge that there exists a norm commanding not to do the action: JP(¬ A) ≡ I judge prescribed not to do ‘A’. As every judgement is itself a speech act exposed to denegation, the action ‘A’ (≪neustik≫) can very well be another judgement and, in particular, another prescriptive statement. For example: JP(≪JP(A)≫) ≡ I judge as prescribed that you judge as prescribed that ‘A’. We find here one of the expressions of fanaticism. Following the logical type of judgement exposed to condemnation, — do they belong to the category of prescriptions, evaluations, or observations? — I suggest the following hierarchy: 1a. JP(≪¬ JP(A)≫) 1b. JP(≪JP¬ (A)≫) 2a. JP(≪¬ JE(E)≫) 2b. JP(≪JE¬ (E)≫) 3a. JP(≪¬ JO(O)≫) 3b. JP(≪JO¬ (O)≫) In the most ≪fundamentalist≫ of its forms (3a & 3b), fanaticism invests a lot of energy to sanction a series of norms concerning observations. Note the apparent paradox: If you consider that this paper defends the prescription not to speak like a fanatic, does it illustrate a weak expression of fanaticism? Hint: My answer would be NO.