The psychophysical problem is long standing and, probably, intractable.
We have a corporeal body. It is a physical entity, subject to all laws of physics. Yet, we experience ourselves, our internal lives, external events in a manner which provokes us to postulate existence of a corresponding, non-physical ontos, entity. This corresponding entity ostensibly incorporates a dimension of our being which, in principle, can never be tackled with instruments and formal logic of science.
A compromise was proposed long ago: soul is nothing but our self awareness or way that we experience ourselves. But this is a flawed solution. It is flawed because it assumes that human experience is uniform, unequivocal and identical. It might well be so - but there is no methodologically rigorous way of proving it. We have no way to objectively ascertain that all of us experience pain in same manner or that pain that we experience is same in all of us. This is even when causes of sensation are carefully controlled and monitored.
A scientist might say that it is only a matter of time before we find exact part of brain which is responsible for specific pain in our gedankenexperiment. Moreover, will add our gedankenscientist, in due course, science will even be able to demonstrate a monovalent relationship between a pattern of brain activity in situ and aforementioned pain. In other words, scientific claim is that patterns of brain activity ARE pain itself.
Such an argument is, prima facie, inadmissible. The fact that two events coincide (even if they do so forever) does not make them identical. The serial occurrence of two events does not make one of them cause and other effect, as is well known. Similarly, contemporaneous occurrence of two events only means that they are correlated. A correlate is not an alter ego. It is not an aspect of same event. The brain activity is what appears WHEN pain happens - it by no means follows that it IS pain itself.
A stronger argument would crystallize if it was convincingly and repeatedly demonstrated that playing back these patterns of brain activity induces same pain. Even in such a case, we would be talking about cause and effect rather than identity of pain and its correlate in brain.
The gap is even bigger when we try to apply natural languages to description of emotions and sensations. This seems close to impossible. How can one even half accurately communicate one's anguish, love, fear, or desire? We are prisoners in universe of our emotions, never to emerge and weapons of language are useless. Each one of us develops his or her own, idiosyncratic, unique emotional language. It is not a jargon, or a dialect because it cannot be translated or communicated. No dictionary can ever be constructed to bridge this lingual gap. In principle, experience is incommunicable. People - in very far future - may be able to harbour same emotions, chemically or otherwise induced in them. One brain could directly take over another and make it feel same. Yet, even then these experiences will not be communicable and we will have no way available to us to compare and decide whether there was an identity of sensations or of emotions.
Still, when we say "sadness", we all seem to understand what we are talking about. In remotest and furthest reaches of earth people share this feeling of being sad. The feeling might be evoked by disparate circumstances - yet, we all seem to share some basic element of "being sad". So, what is this element?
We have already said that we are confined to using idiosyncratic emotional languages and that no dictionary is possible between them.
Now we will postulate existence of a meta language. This is a language common to all humans, indeed, it seems to be language of being human. Emotions are but phrases in this language. This language must exist - otherwise all communication between humans would have ceased to exist. It would appear that relationship between this universal language and idiosyncratic, individualistic languages is a relation of correlation. Pain is correlated to brain activity, on one hand - and to this universal language, on other. We would, therefore, tend to parsimoniously assume that two correlates are but one and same. In other words, it may well be that brain activity which "goes together" is but physical manifestation of meta-lingual element "PAIN". We feel pain and this is our experience, unique, incommunicable, expressed solely in our idiosyncratic language.
We know that we are feeling pain and we communicate it to others. As we do so, we use meta, universal language. The very use (or even thought of using) this language provokes brain activity which is so closely correlated with pain.
It is important to clarify that universal language could well be a physical one. Possibly, even genetic. Nature might have endowed us with this universal language to improve our chances to survive. The communication of emotions is of an unparalleled evolutionary importance and a species devoid of ability to communicate existence of pain - would perish. Pain is our guardian against perils of our surroundings.
To summarize: we manage our inter-human emotional communication using a universal language which is either physical or, at least, has strong physical correlates.