Last week I started questioning the "standard terminology" we tend to use when talking about music, not only in the domain of music theory but also in the talk that arises in the course of trying to prepare a performance. I suggested that the concept of syntax might be relevant when we want to talk about any structure that involves more than a linear ordering. such as order based on elapsed time. I offered, by way of an example, Heinrich Schenker tried to view embellishment from a hierarchical point of view (meaning, basically, that one can think in terms of embellishing the embellishments). Since then, I have been having second thoughts about whether or not it really makes sense to talk about a hierarchy of embellishments as a syntactic structure.
When we talk about language, it is easy to get seduced into thinking about syntax in terms of tree structures. This is particularly true of those of us who grew up having to diagram sentences in secondary school. The idea of having such a diagrammatic representation at all is predicated on the premise that every word can be classified according to some syntactic category, such as noun or verb. These categories admit of relationships that can be defined among them. Thus, a noun may have the relationship of being the subject of a verb, while an adverb may relate to that same very by modifying it. The tree structure that one builds in the course of a syntactic analysis therefore does a "double duty" of representing both the categories of the words and the relationships across those categories.
Do Schenkerian hierarchies of embellishment make sense in this framework? I have my doubts. The elements of elaboration are defined in terms of patterns based on convention. Carl Philipp Emanuel Bach tried to catalog these patters in the second chapter of his Essay on the True Art of Playing Keyboard Instruments. He did a rather impressive job for his time; but it is important to remembers that his focus was on playing an instrument, rather than composing for one (a position he shared with this father, Johann Sebastian). Schenker's system amounted to a revival of Emanuel Bach's perspective but applying it to analysis by execution; and, in terms of how we have come to think about music theory, he may have opened a Pandora's Box in the process.
I would like to suggested that embellishment does not readily fit into either the syntactic categories used to classify words or, for that matter, the semantic categories postulated by different approaches to discourse structure. To use Schenker's own terminology, embellishment is not about the modification of one construct by another; rather, it is about prolonging the time-span of a "musical event." Prolongation has less to do with entities that might be construed as the primitives of some syntactic structure and far more do to with the underlying phenomenon of time-consciousness, without which we can neither make nor listen to music. As I previously suggested, time-consciousness must be with us, even at the lowest level of filling silence with sound. Thus, it may well be that, if we wish to talk about structure at all, we need to talk in terms of constructs applicable to time-consciousness, rather than any of those marks on paper that may abstract those constructs but, through abstraction, have become too remotely removed from the domain of the music itself.