We know that the voltage over an inductor is defined by the formula:
\$V = L * \frac {di}{dt} \$
So in the case where the current flow is suddenly interrupted (like when a mechanical contact is opened), voltage spikes occur in real life.
However, this is not always the case: we don't see arcs happen in small inductive loads. (By small inductive loads I mean a toy car motor, for example.) However, the formula says that the \$ \frac{di}{dt} \$ term should approach infinity when mechanical contacts are opened, therefore the \$L\$ term (which should be small in small inductive loads) shouldn't have a significant effect. Simply, we should be able to see sparks any time we open any inductive load - independent of the inductance.
What are the practical factors that stop the voltage from reaching infinity? Does the current flow actually decrease slower, or is the formula perhaps insufficient for such a "discontinuity"?