Brazil’s Bolsa Família (PBF) is the largest conditional cash transfer (CCT) in the world. CCTs are policies that aim to fight poverty by transferring money to households in poverty, under the condition that they fulfill certain behaviors (e.g., making sure their children attend school and visit the doctor periodically). CCTs supporters claim that conditionalities are justified as they contribute to fighting intergenerational transmission of poverty by inducing parents – above all, mothers – in poverty to invest in their children’s human capital. We call this the common rationale for conditionalities. In this paper, our first aim is to show how the empirical basis for the common rationale is on a weak footing, both for CCTs in general and PBF in particular. However, there are further reasons for conditionalities. Our second aim is to give a set of normative – moral and political – reasons that justify PBF and its conditionalities. Our analysis mixes non-ideal political philosophy and micro-level public policy analysis, taking Brazil’s historical and political context as well as social beliefs about the policy as normatively relevant. We assess PBF and its conditionalities according to three dimensions: programmatic performance, legitimacy, and feasibility. They fare well in all three. While this is a conditional and contextual defense of conditionalities, it provides general lessons for the basic income debate. Conditional transfers can be superior to unconditional ones when they connect basic income with the offer of universal basic services. This is particularly important if we think basic income can undermine welfare state universalism.

This paper develops an account of normative validity and illustrates it through poverty measures. Many ideas addressed below are legatees of Anna Alexandrova’s reflections on what she called “value-aptness” in measuring well-being. To my knowledge, she coined the term “normative validity.” Still, my goal is to treat normative validity more generally than she did and, for that matter, to say something useful for researchers who work with concepts other than well-being––like poverty but not only.  Normative validity is the property of a measure that adequately captures its targeted concept’s normative (moral and political) aspects. Part 1 starts with an outline of the general features of this definition (1.I). Then, I discuss the accounts of validity from a version of the representational theory of measurement and psychometrics and explain why normative validity can be better understood within their framework (1.II). Normative validity is conceptually-driven, while other commonly recognized types of validity are data-driven (1.III). I detail how normative validity can be achieved (1.IV) and discuss its objections (1.V). Part 2 shows how normative validity is pursued in poverty measurement, even without being explicitly called so (2.I). Researchers employ informal and formal tools when identifying individuals in poverty (2.II.a) or creating a poverty index (2.II.b) to argue that a poverty measure captures the normative features of its concept. Lastly, I explore how normative validity can help us better understand the conflict between normative and data-driven poverty measures (2.III).

This is a critical review of Herman Cappelen’s Fixing Language (2018), an excellent and thought-provoking introduction to a hot topic in metaphilosophy: conceptual engineering, which defines the process of evaluating and improving/revising our representational devices (popularly known as concepts). Here, I first present an overview of the book, summarizing his General Theory of conceptual engineering. Second, I point out some limits of the General Theory, in particular the putative consequence of his semantic externalism, the Lack of Control thesis. According to it, the processes behind changes in meaning are too complex and amorphous for revisionary projects to be generally successful. However, I claim that a proper investigation of Lack of Control demands us to look at the sciences, especially the social sciences, something absent in Cappelen’s book. Furthermore, I remember that many conceptual engineers do not put forward large-scale revisions, but rather local ones – i.e., restricted to specific scientific and institutional contexts, and therefore more feasible and potentially relevant. 

"Em vez disso … o que esperamos fazer nesta postagem de blog é fornecer uma introdução básica aos problemas de tomada de decisão sob extrema incerteza. Estamos usando a formulação de políticas para a covid-19 como um exemplo prático disso. Abordamos o assunto porque tem havido alguma preocupação pública sobre o tipo de ação dramática que muitas nações estão adotando, embora as evidências específicas que temos sobre o vírus sejam muito provisórias ou baseadas apenas em modelos. E as políticas governamentais anteriores do Reino Unido foram criticadas por cientistas precisamente porque elas foram consideradas como riscos muito grande a serem tomados à luz da escassez de evidências. Como dizem os cientistas nessa carta aberta, “não estamos convencidos de que se saiba o suficiente sobre “fadiga comportamental” ou até que ponto essas ideias se aplicam às atuais circunstâncias excepcionais. Tal evidência é necessária se quisermos nela basear uma estratégia de saúde pública de alto risco”. Se quisermos entender por que as pessoas podem pensar que essa é a abordagem errada, precisamos entender: como devemos raciocinar sobre escolhas de alto risco quando faltam as evidências ou elas não são claras?"

"Ao avaliarem a ciência, a tendência das filósofas e filósofos é se concentrar nas leis gerais e em sua respectiva verdade.  Defendo que haja uma mudança de nosso foco em direção à confiabilidade da panóplia de produtos [panoply of outputs] fornecidos pela ciência (“o emaranhado da ciência” [‘the tangle of science’]) e, de maneira concomitante, do geral para o particular. Apresento aqui cinco argumentos em defesa disso, dentre os quais (a) muitos, muitos produtos científicos (medidas, instrumentos, estudos empíricos, organismos modelo e assim por diante) que precisamos supor ser confiáveis se quisermos garantir [to warrant] nossos princípios gerais, não são aptos à verdade [truth-apt] ; e (b) a confiabilidade nos convida a uma pergunta crucial: “Confiável para quê?” Ter clareza quanto aos propósitos particulares é algo essencial para ações eficazes e avaliações justas."