I can exectue a metanode that does not have the input port connected (e.g. the first node in the metanode is a "R source (Table)" node, however, the metanode has an input port anyway), but this does not seem to be the case for wrapped metanodes. If this is true, is this simply a philosophical difference between wrapped metanodes and metanodes? Or, is there an important functional reason for this difference? Thanks.
yes this is correct (and intended). The main difference is as follows: a meta node is a wrapper for many nodes. For example for cleaning up your workflow. A wrapped node is a fully functional unit. You can consider it as a node you "programmed" yourself using nodes.
And as in all nodes if one of the inputs is not connected the node cannot be executed.
However, you can use wrapped node as sources but you need to make sure there are no inports configured. You can edit the inports via rightclick -> Wrapped Node -> Reconfigure
Best regards, Iris