I'm a newbie too, so take what I say with a few grains of salt.
I wouldn't think of the ambiguity zone as "symbols", rather think of them as areas where your sensors haven't ruled the contact out as being in (doesn't know for sure they're within the area, but based on the data the odds are they are in that area and not in the unmarked area around it). The shape can be influenced by where the totality of sensors you have access to that are detecting that contact. If it's only one sonar on your unit and if you are moving in roughly the same direction as the contact, the sonar will likely be able to give you a pretty solid bearing since it's not changing much and all the signal is coming from one direction. BUT it will likely have a BIG range ambiguity since it has no angles to triangulate to determine the distance. This would likely result in a "line" as you've described. If you do a run perpendicular to the course or get data from another unit that has a different angle on it (another ship or sub with a sonar that you have communications with for example) you should cut down your range ambiguity since the 2nd contact will limit the range possibilities by triangulation so it might be a diamond at that point.
If it's a trapezoid or wedge it could be more than one sensor has a weak contact with it and some overlap that allows it to rule out some areas as known area where the contact is NOT there, but other areas where it might or might not be.
For a circle it could be that you just have limited amount of contact data or a weak contact that might be moving at an angle to you, so you're not getting a consistent range or bearing (bearing changing constantly due to your movement and target movement).
And yes, it's not unusual that "the contact must be within this area...doh...no it's not." -- that's why it's ambiguous, but you're not sure if it's within the zone, but that's where the odds say it should be based on current contact information. When it turns out a swirling current caused 1 of your first 4 bearings to be wrong, that zone can change drastically when you get to 10 bearing fixes and throw out the bad data point. No guarantees it's in that zone, that's the educated guess based on current contact information.