I've got to jump in here and 'side' with JBD.
_Voltage_ is always measured between two points. When you have an AC supply, you can graph the voltage versus time. In the ideal case for power distribution, your voltage versus time graph will be a nice sine wave.
If you have more than two nodes in your system, then you can measure the voltage between _any_ pair of nodes. In a system with three nodes, there are _three_ voltages that you can measure; in a system with four nodes then there are _six_ possible pairs of nodes and thus _six_ voltages that you can measure. For each of these possible voltages, you can plot voltage versus time, and get a _different_ sine wave.
When you talk about phase angles in a polyphase system, you are simply describing the timing difference of these sine waves. The phase angles only have meaning if you define your reference zero (which of the possible sine waves is considered as being at zero degrees) and if you define which _pairs_ of points you are measuring between.
In a common three phase wye system, the phase angle of the voltages measured between line and neutral are all 120 degrees apart. In other words, if you were to graph the voltage as measured from neutral to leg A, and graph the voltage as measured from neutral to leg B, you would find a 120 degree phase difference. If, on the other hand you were to measure the voltage from leg A to leg B, and then measure leg A to leg C, you would only see a _60_ degree phase difference between your two measurements. Finally, if you compared the A to neutral measurement with the A to B measurement, you would see both a difference in amplitude, and only a 30 degree phase difference.
If you only ever measured between the A leg and the B leg, say because you had a single phase load connected A to B, then you would only have a single sine wave to look at, and thus _no_ phase angle difference to consider.
If you have two legs of a three phase wye system, then _referenced to neutral_ the two legs will be 120 degrees apart. But if you only connect a single phase load between A and B, at any instant in time the current flowing from phase A will have to exactly match the current flowing to phase B. Referenced to neutral the _voltages_ may be 120 degrees apart, but referenced to each other the current flow is exactly the same (meaning a 0 or 180 degree phase difference depending upon your sign conventions).
Because of this, if you have an imbalanced three phase load, even if it is perfectly resistive, you will find some 'power factor' (difference in phase angle between current and voltage) when you compare the voltage phase angles and the current phase angles.
Finally, the 30 degree phase angle difference mentioned above shows up when you compare primary and secondary phasing on a delta to wye transformer.
-Jon
