I will attempt to answer Cyril's question: "What is this dialectic?".
For me, dialectics is an extension of the concept of Logic beyond the limits within which formal logic is adequate; that is, when the objects referred to are "not equal to themselves" or when the complexity of the relations between them transcend a qualitative limit. I think it may be called alternatively "the logic of development".
Formal logic becomes known to us as the system of relations which reflect the validity of propositions, but it turns out that this system of relations is the same as that which reflects all objects of the same kind as propositions, and human beings acquire sensori-motor intelligence of these relations long before we learn logic in secondary school. The recognition of a thing as being "the same thing" as something perceived a moment earlier is an early stage in this process. But the relation of identity has an objective existence and is reflected in countless natural processes long before human beings come to understand that relation of identity as a concept; I have in mind for example chemical reactions which retain elements through different compound reactants, or the trajectory of bodies other than sub-atomic particles.
We acquire sensori-motor intelligence of dialectical relations, for example when we respond to the "first signs" of something as if the thing were already in existence, in all sorts of aphorisms, maxims and turns of speech such as "up to a point", "it is and it isn't" and so on. Hegel gave us a systematic exposition of this Logic, but great as it is, just like formal logic, it can only be a more or less satisfactory reflection of relationships in the objective world which are infinitely more complex.
Because dialectical logic is concerned with transition rather than identity, it is inherently more creative than formal logic, but like logic it is not absolutely creative; George Berkeley showed fairly adequately that formal logic cannot prove the existence of a material world, let alone any particular fact. I think dialectical logic can do somewhat better, but I think Sartre had a point in introducing the concepts of "dialectical dogmatism" and "dialectical scepticism" - only a problem that his "Critical dialectic" somewhat predictably returns to Kant.
In defining dialectics as logic, one could easily draw the conclusion that therefore dialectics is indeed something to do only with human thought. In the same way, it would be easy to draw subjectivist conclusions from Einstein's revolutionising of physics by meticulous "operational" definitions of distance and duration, of how length and duration are measured - but inanimate objects "measure" distance and duration in their interaction with each other; putting oneself "into" an object is a means of understanding it. How can we understand nature other than through precision in handling our concepts of nature?