Post History

I must admit that your question (the described problem) is an interesting one.
Speaking about the circuit (without consideration of the parts values), there are two feedback effects:
* At first, the resistor R1 provides **negative** feedback resulting in a low closed-loop gain slightly below unity. 
* Secondly, the path R3-R2 conncts the emitter with the base (no phase shift) thereby providing a certain (rather small) amount of **positive** feedback.
* Thus, we have a "mixed" feedback (positive and negative) - provided that the parts values are correspondingly chosen.
* In addition, we also have feed-forward because the "output node" (emitter) is directly connected to the input (via R2-R3).
* Such a circuit is also known from opamp applications (positive and negative feedback simultaneously).
* However, the mathematical treatment of the circuit is not a simple matter because there is no real voltage output (zero output impedance) and no high-resisitve signal input (base node). 
* Thereforte, the classical feedback block diagram (as shown in the first post) cannot be applied.