I noticed there generally seem to be two ways to look at functions.
On the one hand you have those that see functions as some sort of process; it gets some input, does some processing and gives some output. For instance, Hitta sees Fi as some sort of evaluation, a judgement process (here). He gives the example of cheese where given certain characteristics of cheese, Fi can evaluate and come to the conclusion: "I like cheese". Another example is when people try to map functions to brain regions, assuming that the processing by each function must be visible somewhere when people are using it. Yet another example I think is when people talk about functions like Ti structures things or breaks things down or Fe manipulates the emotional atmosphere and so on. The idea here is that functions and type directly determine your behaviour and state of mind.
The other way to see functions is only as "giving awareness". In a given context each function tells you these are the Ni aspects, these the Fi aspects, these Se and so on. And the position of each function in model A determines how you tend to deal with this information. So if you find yourself in a situation where you don't particularly value something, which could be an Fi aspect, then your type determines whether you actually care a lot about that fact or not, whether you actually trust yourself on seeing such aspects correctly, or would rather have someone assist you with it and so on. The idea here is that functions are only a socionics invention to describe how the brain might filter raw data into different information categories. Of course behaviour is then largely influenced by what information you are presented with, but also by mood, emotional state, personal knowledge, values, beliefs, history, experience and so on. Socionics attempts to model perception and awareness, not complete human reasoning and behaviour.
I was wondering what you guys think about this, because it's a rather fundamental point.