Most scientists' immediate impulse when looking for answers to the Important Questions about feel is to start by experimentally seeking the neural correlates of feel, and hoping that these will shed light on the problem of how feel is generated and what it is like. It is argued that whereas today we have no idea how feel might be generated, a new discovery may await us in the future.
The sensorimotor approach claims on the other hand that any explanation that appeals to a special mechanism that generates feel in the brain will be pushed into an infinite regress in which appeal is made to other mechanisms.
For example if a particular neural structure was found whose activation correlated perfectly with the occurrence of feel, then we would actually be in a worse situation than before: now we would have to explain exactly what special thing it was about that structure that allowed it to give rise to feel. Also we would have to explain how modulations in activity of that structure explained the different qualities of feel between and within sensory modalities. In order to provide such explanations we would be obliged to appeal to other mechanisms in addition to the perfect neural correlate that we had initially singled out.
It would seem that ultimately any appeal to neural mechanisms generating feel will have to appeal to some irreducible concept like reverberations in the global brain state, quantum gravity or synchronous oscillations whose efficacy in generating feel is left unexplained.
This is of course an admissible step taken at some point in every scientific theory. In physics, for example, one could argue that the concepts of mass, time and space are irreducible concepts that are made use of without seeking further explanation. For consciousness, such a step is the one taken by most current thinkers. However, for reasons of economy, taking the step of postulating unexplainable irreducible concepts for generating feel surely should be a last resort in making a scientific theory, particularly as one would probably have to postulate as many such irreducible concepts as there are multitudinous different feels that one can experience.
The solution proposed by the sensorimotor approach is to rethink what is meant by feel in such a way that the question of what generates it is no longer a sensible question. It turns out that choosing to elude rather than to answer Question 1 (the neural origin) miraculously unveils the answers to the other Questions.
If eluding Question 1 seems more like eluding rather than solving the problem of phenomenal consciousness, then one should reflect that it is the same kind of trick that has previously been used in solving problems in science, in particular the problem of the origin of life. At the beginning of the twentieth century scientists were seeking for the "vital spirit" that generated life. Now at the beginning of the twenty-first century we know that the question of what generates life is no longer a sensible question: Life is not generated. It is a word used to describe the potential that an organism has to interact with its environment.
The sensorimotor approach proposes a solution to the problem of phenomenal consciousness which takes a similar strategy. Instead of saying that phenomenal consciousness is generated by a neural mechanism, we say that it involves a very particular kind of potential that the organism possesses to interact with its environment. This way of eluding Question 1 (how feel is generated) then allows Questions 2 ("presence"), 3 (quality) and 4 (ineffability) to be addressed in a more satisfactory way than before. Taking this new stance also leads to new empirical predictions and interesting experiments.