As part of a probabilistic reconstruction of quantum theory (QT), we show that spin is not a purely quantum mechanical phenomenon, as has long been assumed. Rather, this phenomenon occurs before the transition to QT takes place, namely in the area of the quasi-classical (here better quasi-quantum) theory. This borderland between classical physics and QT can be reached within the framework of our reconstruction by the replacement $p \rightarrow M (q, t)$, where $p$ is the momentum variable of the particle and $M(q, t)$ is the momentum field in configuration space. The occurrence of spin, and its special value $1/2$ , is a consequence of the fact that $M(q,t)$ must have exactly three independent components $M_{k}(q,t)$ for a single particle because of the three-dimensionality of space. In the Schrödinger equation for a "particle with spin zero", the momentum field is usually represented as a gradient of a single function $S$. This implies dependencies between the components $M_{k}(q,t)$ for which no explanation exists. In reality, $M(q,t)$ needs to be represented by three functions, two of which are rotational degrees of freedom. The latter are responsible for the existence of spin. All massive structureless particles in nature must therefore be spin-one-half particles, simply because they have to be described by $4$ real fields, one of which has the physical meaning of a probability density, while the other three are required to represent the momentum field in three-dimensiona