The vlan-tagging statement is used so you can configure logical units aka layer 3 subinterfaces.
That way you can configure ge-0/0/1 unit 0
ge-0/0/1 unit 0
ge-0/0/1 unit 1
ge-0/0/1 unit 2
ge-0/0/1 unit 4094 (if your device supports it)
etc
When you configure the subinterfaces, you will then add a vlan-d tag to each of the subinterfaces created. This allows these sub-interfaces to recieve untagged packets.
How is this any different from simply configuring set interfaces family ethernet-switching interface-name port-mode trunk?
Ethernet-switching is a layer 2 concept. With the trunk port, it will only accept frames that have a vlan tag identifier.
The nomenclature vlan-tagging maybe what is misdirecting the understanding