The author reviews two opposite traditional positions on the role of values and value judgements in sociological research and theory: treating values as a bias interfering in research, or treating values as ideology providing privileged access to knowledge. He traces the recent revival of the debate about valuations, focusing particularly on the claims of the so-called 'public sociology'. Then the author's own position is outlined based on the fundamental particularity of the social sciences as contrasted with the natural sciences. The old argument that values do not follow from facts is acknowledged as true in the sense of logical deduction, but in the social sciences we encounter different mechanism of implication, which may be called 'sociological syllogism': values may follow from facts, and facts may imply values because, on the one hand, people act on their axiological beliefs, and human actions constitute social facts, and on the other hand, social facts (e.g. about poverty, inequality, degradation, crime, terrorism) mobilize moral impulses and valuational commitments. In other words values shape meanings of human actions and resulting social facts, and the knowledge of facts acquires valuational meaning by mobilizing human axiological impulses. The strict separation of facts and values does not work in the social sciences; there is a two-directional link between the two. This opens the possibility for 'sociological ethics' deriving normative standards of social life from the research results of sociology.