Formal Definition:

Let me introduce some more notation:

$$\mathcal{X} := \{\: (X_1,x_1), \; (X_2,x_2), \; \ldots , (X_n,x_n)\:\} = ( \vec{X},\vec{x})$$

$$\mathcal{Y} := (Y,y)$$

where $X_1,\ldots,X_n$ and Y are variables, $x_1,\ldots,x_n$ and y are specific values (or value-sets) from their respective domains. $\vec{X}$ and Y can be seen as discrete random variables. We define a rule as:

$$\mbox{If } \vec{X}=\vec{x}, \mbox{ then } Y=y\mbox{ with transition-probability }c$$ (25)