Thursday, November 20, 2014

Desperately Seeking Relevance: Music Theory Today (6.2)

Rubbish Theory
and Music Theory Today

2. Maverick Integration
[Previous Post: 1. Tripping over Rubbish]

There are many ways to jump into this. The most direct way is to replace music-theory-today's presumed equal sign with a question mark and start from scratch.

Diagram 1

Preliminarily, analysis is how theory probes the real world. As much as theory needs analysis if it wishes to stay grounded, analysis needs theory or it is nonsense. Without theory, analysis is at best a collection of disconnected hunches. The moment a connection appears, we know that theory is lurking in the background.

Theory with no view to application is similarly sense-less. (Nevertheless, theory can develop in the abstract, and this is often desirable and even necessary when navigating possibility's garden of forking paths. But for now, put this garden out of your mind. For the present, while it lasts, there is only certainty's garden path.)

Theory applied yields a necessary ground for analysis; each analysis provides theory with a confirmation. Each confirmation strengthens our confidence in theory as it edges toward "truth," i.e., appears incontestable. (Diagram 2.)

Diagram 2

If we leave it here, we're stuck in a loop: apply–confirm–repeat. We have yet to ask: Just what is it that analysis analyzes? The theory–analysis transaction (Diagram 2) assumes a supply of motivating objects or artifacts we call musical works. Think of this supply as a library of objects available for analysis to draw from to perform its theory-mandated analytical operations. (Diagram 3.) This draw can be made in one of two ways, either randomly or selectively.

Diagram 3

If the draw is selective, it can bias in favor of the theory, i.e., analysis is free to select just those works (or portions of works) that confirm the theory (cf. "procrustean intonations"). This is always tempting, but it's cheating. We assume no user of the model would want analysis to cheat.

But a work randomly drawn (analysis is blindfolded for the draw) might not confirm the theory. In that case analysis looks for ways to "save the theory" by suggesting adjustments to theory in order to accommodate the new information supplied by the maverick work. The theory is not replaced – it is effectively the same theory[*] corrected, improved, expanded, with essential core invariants remaining untouched. Now on future draws the model can handle similar works previously considered as mavericks.
[*] In the present context, two theories are the "same" in the sense that they are both versions of one abstract theory defined by a core set of invariant features (propositions, rules, objects, patterns, etc.) "Core feature" is left undefined at this point. Some may think of it as common sense, others a consensus of experts, others a structural sine qua non, still others audience expectation, cultural norm, etc. The important thing is that, were theory's core to be breached, either it would morph into something unrecognizable or unpalatable, or it would collapse entirely taking its model down with it (see previous comments on Euclidean"axioms").

This is evolutionary adaptation. Theory has built-in room to grow. A constant supply of works chosen randomly fuels (challenges) the analytical pump which in turn confirms or improves or corrects the theory which then allows analysis to encompass a wider, more varied selection of works. While the model in Diagram 2 represents a homeostatic system, the model in Diagram 3, simply by including the adjustment function, represents a homeorhetic[1] system.

Note posted at the library's circulation desk: The works library is a public lending library. After a work is removed and analyzed it must be returned to the library. No exceptions.

Continuing to work backward, the next question emerges: Where do all those works come from? This consideration might appear to be redundant for the model. We sense that adding unnecessary weight is asking for trouble, and, as we shall see, adding a third role does add a big dose of complexity and discomfort. It would be easier  to stop here. Using a wave-of-the-hand strategy, we could call the question an irrelevant nuisance. But relevance is the game we've decided to play. We cannot not look. So....

If we add a composition (synthesis, assembly) arrow to those arrows labelled application, confirmation and adjustment, we still have to decide on composition's source and that source's relationship to the model. Composition commonly implies a composer[2], but the model encourages us to continue to focus on function and process over flesh and blood actors such as theorist and analyst. This suggests that the work source in the model ought to focus on the more technical craft of composition rather than how the craft is employed or who employs it or what the employer's motives might be. Drawing on a venerable tradition, let's call the source of any work techne.[3] (Diagram 4.)

Diagram 4

Tracing back from theory-analysis through the work and then to techne suggests the next obvious question: What is the source of techne?

Within the confines of the model, techne has only one real option for a source: theory. This is represented by the arrows and overlapping green box in Diagram 5.

Diagram 5

Theory might be "informed" by techne directly, short-circuiting the work→analysis→theory chain by going right to the work's source for confirmation and adjustment. This would suggest that techne in turn might apply theory as its source of acceptable patterns and objects – an authority/guide to be followed in order to compose a work that stays within the status quo and maintains the consistency of the model. 

So techne might draw from theory in ways roughly parallel to the two possible ways analysis draws from the works library, but with reverse effects. On the one hand, techne might slavishly accept everything in theory's list of objects and patterns – somewhat akin to analyzing anything a random draw proffers. Or techne might pick just those items in theory's list that are "relevant" to its work-in-progress and ignore or reconfigure the rest in ways that still leave theory's core untouched – acceptable variation somewhat akin to a biased analytical draw. 

This possibility, despite a circumscribed freedom of choice given to both analysis and techne within their domains, envisions theory as a Janus-faced overlord within the model, one face governing analysis, the other governing techne. In principle, any time analysis comes across a maverick work, it would no longer be a mystery where the maverick came from – only a puzzle whose solution lies within the model. Either techne misunderstood or unintentionally misused theory (a mistake), or theory in its navel-gazing abstract mode ("speculative theory") spontaneously expanded within its self-imposed bounds of internal consistency and directly encouraged techne to choose freely from the new menu of possibilities provided by this expansion.

Strangely though, closing the circuit – staying within the confines of the model – implies that theory and techne aren't really compelled to talk to one another directly at all (which can be helpful when they're separated by three centuries). Certainly techne can "read" theory directly, but a lecture from theory – especially one that does no more than rehearse techne's own past – is hardly conducive to discovering new procedural approaches. (For some reason, techne is averse to running in ruts.) For its part, theory can always get at techne the long way around via analysis of techne's works. (For some reason, theory finds analysis' company more amenable.) Similarly, given that techne stays within the model, there is no need for a direct path between techne and analysis. That particular transaction – if it is ever necessary at all – is adequately mediated by theory.

In another world we might expect to find a triangular model with three co-dependent and interacting roles. In music theory today the model more closely resembles the interaction between two pairs of roles that share one of those roles. Let's now separate those pairs to see if this is really the case.

With time (four or five centuries ought to be more than enough), theory reasonably comes to assume that it can always count on a relatively well-behaved techne to express itself within theory's rules or helpfully expand those rules via the enrichments of "clarifying violations"[4] which will sooner or later be caught by analysis as they appear in maverick works whence, it is assumed, these violations can be folded back into the model. Theory, again quite reasonably, comes to assume that its model is the only model: there are no other models that techne can escape to – indeed, why should techne even want to escape a benevolent dictator? And even if techne is unhappy with its prospects within a fully mature model, there are no other options. Resistance is futile.

Theory now feels safe in ignoring techne altogether because it can get all its developmental needs from the work via analysis, with no particular need to know anything more than it already does concerning the work's source. At this stage, the works library has become huge, and it continues to grow in quantity if not in quality. Analysis detects no more mavericks. The overstuffed library now appears to consist almost entirely of centuries of artifacts that conform to a maximally expanded theory that can take no more adjustments if it is to maintain an intact core (i.e., if it is to be itself). Obviation of the need for analysis' adjustment function means the model has effectively collapsed back into its archetype (Diagram 3 above minus the adjustment function).[5] Apply–confirm–repeat. There are plenty of artifacts to go around to support a homeostatic model without techne having to produce any more of them. This has made life easier for analysis whose only task now is to confirm the theory. In the now vanishing possibility that analysis comes across a stray nonconforming work, analysis will simply reject it as a monster.[6]

Diagram 6

Diagram 6 shows the situation. Rather than overlapping pairs we see one pair and a detached singleton. Techne – know-how – is now unemployed – or, to be accurate, can find no meaningful employment beyond music's version of a fast-food chain, which is to say that techne can continue to produce as many model-conforming works as it likes. New-to-my-generation is fine and may get a pat on the head from theory, but new-to-the-world is no longer in the cards for the model or, evidently, for the library it has created. A submissive techne can only look back nostalgically at the days when it was able to produce those reinvigorating mavericks – when it was relevant to the model.

Knowledge (knowing-that) is now encapsulated in the perfectly self-contained, teachable red box. Pedagogically at least, theory and analysis can go on forever, perennially renewed as succeeding generations are presented with the same library of new-to-them works to learn from. New-to-my-generation requires no more than a good curator to pull off the illusion of newness. And if perchance theory, in its navel-gazing mode,  comes up with a new "approach" to some of the artifacts in the library, verification of that approach means no more than instructing analysis to do a "search of the literature."

The model is perfected.
The world is one.

Ite, missa est?

[1] Homeorhesis: "The condition of a flow process which remains canalized within limits in a growing system. ... [A homeorhetic] system turns homeostatic (i.e., acquires dynamic stability) when it reaches its full development, in accordance with its archetype[!]" More at the on-line Encyclopedia of Systems and Cybernetics.
[2] Since music is a performing art, a work may also be construed as a performance-of-a-work, and so a performer may be construed as a composer. This identification, which is my own position, still raises practical issues that remain open for discussion and development. Whichever position one takes on this would not significantly change the model; but such a discussion here would obscure the issue at hand by taking us off into a maze of sidebars.
[3] Other candidates in naming this role are the more highfalutin poiesis and synthesis. My preference for techné comes from a desire to maintain the earthbound "workshop" essence of the model which is intended to emphasize the toolbox nature of theory and the tool-use nature of techne. Both art and kitsch emerge from tool-users whose activities are circumscribed by the tools available within the model. Also, identifying the work's source within the model as techne has the added advantage of interpreting the model as yet another variation on the old philosophy–etymology game played between genesis (doing, making, craft, techné) and knowledge (theory-analysis, epistémé). Knowing-how vs. Knowing-that.
[4] The phrase is Charles Wuorinen's ("Toward Good Vibrations" originally pub. in Prose, reprinted in Elliott Schwartz' Electronic Music; a listener's guide, p. 257). While Wuorinen was making a point about the contribution of interpretation in the composer→score→performer context ("clarifying violations of the text"), I find it also neatly summarizes techne's contribution to enriching the closed normative model presented here.
[5] "Growth of a system is normally homeorhetic, because it is the only way to maintain its identity[!] The system turns homeostatic (i.e., acquires dynamic stability) when it reaches its full development, in accordance with its archetype." (Encyclopedia of Systems and Cybernetics)
[6] "Monster" is Thompson's word for such a reject. It is particularly appropriate in our present discussion given it's etymology. In the 14th century "monster" was associated with any creature afflicted with a birth [read "compositional"] defect. As the discussion proceeds it is important to keep in mind that such a work rejected by analysis as irredeemably nonconforming was nevertheless conceived in the womb of techne. The monster's story will be told in the next post.


Carsonics said...

Enjoyed reading your latest essay Stephen. Analysis is necessary in a modern complex world in any technical discipline - but we are constantly discovering nature's laws and hidden secrets so that the very process of trying to define laws or "theory" is fraught with limitations and dead ends. Art [good art] is too complex and dimensional to be described by theory alone and one must constantly keep an open mind and look at any complexity with an innocent sense of awe and critical eye for patterns and logic that are comprehensible. I want my own students to discover things that they comprehend and find their own way into the meaning of music without me telling them what to find. My job is to point out things that I am capable of comprehending they may have overlooked or not considered as an added part of the discussion and dialogue.

stephen soderberg said...

Thanks for your comment. And I don't disagree with you entirely, but I'm trying to come at this from a different direction. Art comes from someplace, it doesn't just suddenly appear out of nowhere, leaving the "theorist-analyst" to make up theories to describe it after the fact. The composer isn't simply God's amanuensis, taking dictation from angels. (I have heard many variations on that theme - some are quite subtle.) I don't see theory's *primary* goal as one aspect of description - description is secondary. I see theory's primary goal (and historically this was the case until relatively recently I think) as a kind of "tool kit" for the *composer*. And lately, composers aren't talking much about their "personal techne," and with good reason (for another day, perhaps). But I don't want to say any more right now. This is getting into my next blog post. I'll just leave it at my personal observation that theory has nudged the creative act out of the picture - it's weird!
-- Steve