Sometimes I still enjoy listening to my old, analog LP records, even with their snaps, crackles and pops.
If we are going to argue that language is a system for harnessing attention, we ought to be clear which of the two general theories of attention we are talking about, information oriented or consciousness oriented.
Information oriented attention was proposed by Donald Broadbent in the 1950s and is still favored by artificial intelligence investigators who seek to model attention on a computer. It defines attention as a filtering process that buffers some input before it moves on to short term storage; however, it is very different from the sort of attention considered on this blog.
The nature of consciousness oriented attention is explored in Chapter 2 of Attention and Meaning, "Attentional Semantics: an overview," by Giorgio Marchetti, one of the book's editors. He proposes that attention and consciousness are closely related and Marchetti argues, "there are never cases of consciousness in complete absence of some form of attention." [p 34, note 1] Consciousness itself is the "privileged way [i.e., available exclusively to each of us individually] for us to acquire and construct our knowledge of [an] object." 
This view of attention is much more active than the information-oriented one in which data is passively received and held. This difference is completely in keeping with the mechanical v humanist differences in the psychological sciences. Mechanical approaches seek to explain all activity in terms of responses to environmental conditions, while humanists allow for some things to arise at the level of the individual. Both have a certain a priori quality. Humanists point out that people routinely produce objects, explanations and plans that are both apt and novel, so they assume humans have some active power that makes them different from rocks and machines. Mechanists tend to argue that although they cannot explain all human output, it must all arise from either external causes or random processes.
Another difference is in the information theorist's emphasis on concepts, while humanists stress perceptions. Predictably, concepts are objective while perceptions are subjective. Concepts are expressed by symbols and are abstract, that is they are defined by pointing to other concepts. Thus, even a concrete word like "tree" is defined in terms of annual/perennial, grassy/woody and stalk/crown. Meanwhile, perceptions are formed by sensations and spatial-temporal distribution, known only to the perceiver.
Another difference is that mechanist attention tends to be digital, whereas humanist attention tends to be analogical. Computers can only process digital information, so even if an input arrives in the form of a wave, it is coded as digits. Analogical attention "accounts for the quantitative, qualitative and dimensional variability"  of the input. Thus, with analogical attention a person can notice something never previously attended to, and draw a colleagues attention to the phenomenon, drawing an, "Oh, yeah!" in surprised reply. There seems no way for digital data to attend to something unprecedented, unexpected, and not-yet-conceptualized.
Perhaps the clearest illustration of this difference in approach lies in the different way mechanists and humanists talk about language. Both agree that it is always possible to generate new sentences. To prove their point, mechanists point to the way structures can be embedded endlessly in sentences: John ate; John ate a meal; John ate a meal made of pasta; John ate a meal made of pasta bought that morning…, etc. Humanists prove the same point by noting the possibility of having new things to say based on new insights, new metaphors, and new points of view. Henry James wrote many an unprecedented sentence without resorting to the trivial recursions of the mechanists.
Consciousness-oriented attention is able to acknowledge many features of language that, at best, seem problematic to information-oriented theories, but many people assume that scientific explanations cannot be grounded in consciousness, no matter how many observations challenge the mechanistic ones. In my recent paper, Attention-Based Syntax, I closed with a defense of grounding the study of language in consciousness:
Physics, chemistry, biology, and medicine have all benefitted enormously from Galileo's ban on references to subjective qualities. The effort to apply the method to the study of human activities, however, has been far less successful. The humanities—literature, history, art, music, rhetoric, law—have scarcely been touched. It is true that those fields are complicated, but alchemy once seemed hopelessly mysterious and yet progress proved possible.
It is important to remember that Galileo's ban did not succeed because there is no such thing as beauty, pain, a claret-colored dye, or a perfumed scent. For us, these subjective experiences are as much givens as gravity or electricity. Those things, however, have nothing to do with physical actions and by excluding secondary qualities, Galileo chased naïve realism out of science. The approach gave us a way of investigating the natural world in natural terms.
Yet secondary qualities have not gone away. Anybody who attempts to speak only of physical things quickly discovers that most of the subjects that people talk about are out of bounds. Ernst Mach wrote that he wanted to develop a way of talking that did not require him to speak one way about physics and another way about everything else. Einstein told the wife of his good friend Max Born that although he supposed it was technically true that everything could be reduced to physical measurements, the ambition was ridiculous because it missed all that was meaningful in one's life. Human sciences are not going to progress by pretending that secondary qualities play no role in human activity. Any science that cares to investigate the product of human behavior must face up to the facts that human artifacts have not been produced by physical laws alone. A clay pot dug up in an Athenian ruin is a physical artifact subject to physical laws, but physical laws cannot explain why the decoration is so different from a similar pot of about the same age dug up near Beijing. For those explanations, we need to study the culture and tastes of the two localities.
At this point we have reached a fork in the road. We can say either that there is much about human production that can never be understood scientifically, or we can say that science must include some recognition of the subjective experience that produces so many human artifacts.