putting in a 100 ohm resistor between the two terminals, and setting the probe type to PT100, should make the PID read 32*F.
its not going to read ambient temp just by shorting two terminals... thats why a sensor is required in the first place...
Thermocouples work based on a thermoelectric effect known as the Seebeck effect, wherein a conducting wire with a temperature gradient along its length develops a potential difference. Electrons diffusing and all that. Take two dissimilar conducting wires, join them at the "measurement junction," and you can measure a voltage difference on the "reference" ends due to a temperature gradient between the measurement site and the reference ends. However, in the absence of a temperature difference, there is no Seebeck effect and therefore no voltage or potential difference at the reference ends.
Shorting the thermocouple terminals ensures there is zero measured voltage, which the controller interprets as being no difference in temperature between the measured point and the controller. It has a cold-junction correction (typically an onboard thermistor or PRT) so that it can determine the ambient temperature, and it displays that. It would only display 32F/0C if there was zero voltage and the reference was being held in an ice bath.
Here are a couple of pictures illustrating what I am talking about. Notice the orange wire shorting terminals 9 and 10. Okay, its an Auber 2362, but I am willing to wager the 2352 works the same.
You
DO NOT need a 100 ohm resistor to run the PID in manual mode, only a bit of spare wire.