The authors deal with a paradigm of scientific discovery defined within a first-order logical framework, based on the notion of detectability. Model theoretically a sentence \(\theta\) is detectable in a class of countable models \(\text{MOD}_{\text c}(T)\) of theory \(T\) iff both \(\theta\) and \(\neg\theta\) are equivalent over \(T\) to existential- universal sentences. The main result is a proof that for every countable first-order language thee exists a universal scientist, a scientist which solves every problem that any scientist can solve, i.e. a Turing machine \(M^ T\) equipped with oracle for \(T\) such that \(M^ T\) detects \(\theta\) in \(\text{MOD}_{\text c}(T)\) if \(\theta\) is detectable in \(\text{MOD}_{\text c}(T)\). It is shown that first-order logic is a maximal regular logic which has a universal scientist and the Löwenheim-Skolem property.