The study of the role of language in the formation of reasoning and the recording of observation has a difficult and undignified past. On the one hand, a record must be made, which, if it is to be validated, must be recordable. For the same reason, this record of the record must in turn be recordable, which must in turn be recordable and so on, so that the act of recording is repeated to the point of becoming meaningless. This trap of infinite repetition is countered by the argument that, if a record is useful according to any specified criteria, it is nevertheless in some way meaningful. Communication is either destroyed or limited, or according to some theorists, both together. There is however a growing realisation that the functions of language, recording and observation are wrongly prioritised, and that this whole debate simply asks the wrong questions. This understanding really began in the mid 20th century when some physicists began to question the assumption that the only really important aspect of quantum theory was that it worked. Interpretation had became a beta noir (and for many that is what it has remained ever since). However, it occurred to some scientists that it might be useful to describe and therefore interpret what is going on in quantum science, even if there were little consensus about meaning. In order to do this, they began with the principle that the most important aspect of quantum science is that it is necessarily probabilistic. The reason for this, put simply, is that it is a science of things that are physically impossible to observe, because they are too small and too fast. In science, the understanding of probable behaviour is as useful and as successful, it turns out, as the understanding of actual behaviour. It began to dawn on philosophers and critics as well as some scientists that this high handed dismissal of the powers of observation, in situations where it can’t be used as a means of understanding, can teach us some useful lessons about what we really need to know. The most useful and widely applicable of these lessons is about accuracy and certainty. To an interesting degree, quantum science works without either. And it has begun to occur to theorists that if scientists don’t need it, they might not either. The realisation has been that this might suggest some useful middle ground in the observation stand off. If we know something for certain, we immediately question it. That is where the trouble starts. But make that knowledge probabilistic and the whole proposition changes. It is difficult to oppose a fact that may or may not be true. If two, or as in physics actually an infinite number of possibilities need to be considered in order to comprise a single certainty, there is no way of considering the observation, if made, as potentially incomplete, for the simple reason that any revelation of new knowledge must then be incorporated as part of a possibility.  This continual process of incorporation, it is now being suggested, is exactly the function of the human brain. So language, insofar as it is an interpreter, does not exist independently, any more than any other independent form of observation or interpretation exists. Instead, language and observation are a part of a process, which we might call consciousness, which is both subjective and universal because its only means of negotiating reality is probabilistic. Language does not define these probabilities, but neither is it specified by them. It is unrestricted. There is no information to record, but at the same time there are many sources. Or to put it another way, there is no output to start from, but there are many inputs. Language is predictive and speculative. These are its only real characteristics.