Model-theoretic dividing lines via posets

Abstract: We show that for each property $$\mathsf{P}\in {\mathsf{OP}, \mathsf{IP}, \mathsf{TP}<em>1, \mathsf{TP}_2, \mathsf{ATP}, \mathsf{SOP}_3}$$ there is a poset $\Sigma</em>{\mathsf{P}}$ such that a theory has property $\mathsf{P}$ if and only if some model interprets a poset in which $\Sigma_{\mathsf{P}}$ can be embedded. We also introduce a new property $\mathsf{SUP}$, consistent with $\mathsf{NIP}_2$ and implying $\mathsf{ATP}$ and $\mathsf{SOP}$.