Peter Hobson's chapter on Emotion, Self-/Other-Awareness, and Autism: A Developmental Perspective in Peter Goldie's (Ed) Oxford Handbook of the Philosophy of Emotion posits identification as a core driver of our developing understanding of one another. Against simulationists he holds that imagination, as it is normally understood, is not what it takes for us to grasp another. It is not that we understand one another because we imaginatively project ourselves into their shoes. Importantly, it is not as if we start with a viable understanding of ourself, and then go on to make use of this in understanding other people. Against simulationist's main rivals, the theory theorists, he holds that the relevant abilities are far more developmentally primitive and non-cognitive than the idea that they are a function of deploying a tacit theory would suggest.
And that all sounds very attractive. But what is identification, and how does it help us to make sense of one another? Well, I suspect the latter part of that question may be badly put - in so far as it suggests that we know what it is to understand one another but are just unsure about the route we take to come to this understanding. At any rate, here is my answer to the question.
First, understanding someone is here not something other than being able to unreflectively anticipate their next gesture or movement or utterance. This pre-reflective non-predictive expectation is what it is for us to be able to find our feet with one another, to be able to 'make sense of' how someone is behaving.
Next, emotions are object-relations, where these intentional relations are primitive forms of social understanding. When I feel hurt, this feeling is my grasp of your hurtfulness, your slight of me. The emotion discloses to me the social meaning of your action. Emotion is a window onto the social world, and emotion - in all its own dispositional glory - is an embodied action-guiding grasp of others: without it there is no such thing as true social understanding, since this more cognitively elaborate forms of this understanding are always embedded on the most primitive emotional form.
It is true that to understand your anger I need not myself be angry. More likely I will be afraid. It is in my fear that your anger is disclosed to me. So there can be no straightforward story about identification leading to similar emotional experience leading (somehow) to understanding. That way of thinking about it in any case is highly external - once again we have the identification leading to understanding, rather than being part of its very form.
So what is it for me to be able to 'feel' or 'come to know' your anger? It could be that you remain blankly on the outside of my understanding, as may be the case for the autistic child. For this coming to know to obtain, two processes must, I believe, take place.
First we have the identificatory entering into the space of shared experience. This is a kind of loss of self, a merging. Your arousal becomes my arousal at this point. Identification is in a sense a pre-psychological process - part of something which makes for a psychological experience of understanding, but not itself something which is a psychological experience. (It doesn't happen within a psyche, one might say, to borrow the Cartesian spatial metaphor: rather it structures the psyche itself in a part of itself.)
Second we have the differentiating 'cut' of self and other, in which the directionality of emotion is registered. For example, your arousal is directed outwards rather than inwards: you are angry rather than fearful. In the conjoint intentional field, the reinstating of self-other differentiation leads to me feeling afraid. My fear is my recognition of your anger.
I've talked about 'first' and 'second' here, but I suspect that we don't have to do with two temporal segments of experience, but rather two merely-formally-separable aspects of interpersonal understanding.
Of course identification need not be accompanied by differentiation. For example, the little boy does the things his father does, adopts his mannerisms, all quite unconsicously. What it is to pattern yourself on someone in this pre-reflective manner is to allow your neurological system to reverberate with theirs. Well, actually the difficult thing is preventing this from happening, rather than allowing it: identification usually happens automatically. This is a powerful learning mechanism, but it does not come with understanding unless we also have the cut of differentiation.
Yet of course the cut can be drawn in the wrong place. Where it is drawn - how the arousal in the interpersonal sphere comes to be divided into the complementary doublets of anger and fear, for example (or sadness and pity) - will depend not only on the identifier but also on the identified-with. Forms of emotional micro-interaction can affect, skew, the placing of the 'cut' (or the 'chiasm' as Merleau-Ponty calls it). Hence we get projective and introjective forms of identification, where I take in some of you or lose some of myself in you. The placement of the chiasm is also, as we know all too well in therapy, under largely unconscious motivational control.
And we have the work of therapy itself where I manage, metabolise, and then reflectively or otherwise feed back your projections into me.