Temperature — additive?!

No.

At least one wrong premise — the scale of Celsius, and Fahrenheit — relative. +80 C 400% more than +20 C, but only by about 25% (too lazy to count exactly), for sure +40 will not be.

1. The temperature of the additive. At least if you count in degrees Celsius or in Kelvins. About Fahrenheit will not tell, did not use.
\r
2. The amount of the exhaust cooling system heat is proportional to the Delta temperature. The non-linear dependence of thermal conductivity of materials, resistances of transistors and other usually small and can be ignored.
\r
3. Therefore, theoretically do with the same system parameters cooling temperature of the cooled CPU will decrease by the same Delta temperature, which decreased the temperature of the environment.
\r
4. But in practice, it is better not to check. Humidity in winter is large, frequent temperature changes -> the condensate.

non-linear dependence of the CPU temperature from the ambient temperature
It is necessary to consider the thermal conductivity of the cooler, the airflow in the case, etc.

The issue of thermal conductivity.
\r
In an ideal — Yes. Actually — no. When blowing from the awesome speed wind temperature will be infinitely small value to differ from the surrounding.

when blowing the intensity of heat transfer is dQ/dt = k*ΔT, where ΔT is the difference between the temperature of the cooler (air) and the temperature of the processor.
The heat transfer intensity equal to the intensity of the heat (so the processor heats to a certain temperature, and it is further not creeps), so we can assume dQ/dt = const.
we can assume that k in a small range of conditions does not depend on temperature (the configuration of the cooling system depends heavily... but nothing changes).
The resulting dQ/dt = k*ΔT1 = k*ΔT2 and ΔT1 = ΔT2 — I mean, Yes, we can assume that the "additive".

>The CPU temperature becomes a +40?
No, because the greater the temperature difference between the CPU+heatsink and the environment ⇒ faster heat transfer (assuming that CPU power is one and the same). Dependence afaik exponential ⇒ cools faster ⇒ temperature is below +40 (but not dramatically, because the count from absolute zero). In the presence of the cooler exact figure is hard to consider as with the radiator, in principle, actually.