Menu

Improvisation

for classical musicians

Buxtehude 2 - Fugues & Fughettas

Some more (tricky) fugue subjects and answers from Dietrich Buxtehude!

Fughetta subject in G major

answer

Also this one in G minor and in compound time (12/8)

answer

Buxtehude - preludes

Using the Friesach sample set of Piotr Grabowski (see here for more details), am enjoying several break-throughs in improvising Buxtehude-style Preludes and free-style forms. These forms tend to mix short sections of declamatory, fantasy phrases with choppy fughettas and contrapuntal sections. The Friesach set gives me the right sound and mix I need to understand the articulation of Buxtehude's figurations. This 'choppiness' is important - more than can be conveyed in theoretical terms such as decoration or embellishment. This choppiness is what first strikes me when I hear this music, as I walk into the church or imagine the movement of the music. It is jagged, complex, angular music - fleeting, moving and pursuing. It captures the flight element of the Fugue - the chase of the subject as it appears and disappears through the texture.

Registration is crucial but not tricky. Again, it depends a great deal on the type of organ used. The Baroque organs of N-Germany were characteristically bright with clear, sculpted sound structures. A typical 8', 4', 2' is mobile, clear and authoritative. Adding a mixture stop results in a strong, even complex sound which might prove too complex for fugal work.

Focus on David Dolan

David Dolan stands out in the world of classical improvisation. Why?

He's an important pedagogue who specialises in teaching non-improvisers (classically-trained musicians at the highest level, who have experience only in interpreting written scores) to improvise. This means that not only has he devised ways of overcoming ingrained habits of musical practice, ideological beliefs associated with these habits, and the consequent creative barriers of classical training, but - and this particularly appeals to me - his teaching stays within the repertoire of classical music: that is, he does not resort to another type of music such as jazz, or free-style improvising, but takes classical models as the basis for experimentation.

This post is based on an article he wrote in 2005, entitled Back to the Future: towards the revival of extemporisation in classical performance. Dolan's teaching approach starts from the consideration of improvising knowledge (what does one need to know in order to improvise). This knowledge base he defines through the idea of 'schemes' - the 'learned musical schemes' of a particular musical style and culture, and the 'natural schemes' which are more universal aspects of human perception, being related to speech, expressive body language, sense of proportion, variety, etc. Often these two different types of schemes meet in musical language, for example many musical structures take the form of an 'arch' which describes a natural trajectory through space and time, comprising a point of departure, and a highest point, from which the journey back to earth begins.

To improvise then, one needs to acquire not only theoretical and practical knowledge - knowledge of musical elements and instrumental technique - but also knowledge of how to manipulate this knowledge, how to mobilise one's skills towards expressive ends.

Dolan, thus puts refutes a common concept in improvising pedagogy: that one needs to first learn 'facts' or patterns through abstract exercises which are then (somehow) transformed into expressive, fluent improvised music. Instead he puts expressive gesture and communicative intent at the forefront of learning. Thus, students aim to make a broad gesture or communicative intention without reflecting too much over the details; once the gesture is made then one look back and consider what one did, and how to do it differently or better. As communicative gesture is movement based, there is a close afinity between Dolan's approach to learning and Jacques-Dalcroze' use of eurythmics, and insistence on learning through action first. Both men recognise the dangers of being lost in thought, the hesitation of conscious calculation, and how fear of errors can lead to lack of flow.

My feeling is that this type of learning is greatly assisted by the presence of an encouraging teacher as Dolan or Jacques-Dalcroze themselves. I can imagine the way in which Dolan constructs each exercise, establishes the criteria (rules of the game), gives critical or non-critical feedback, and so on, is very much part of the success of his teaching. Such a method might be more difficult to achieve on one's own, or where surrounding feedback is more critical..!

Dolan's work fills a much needed gap in knowledge of how to start improvising. While many theorists and writers agree on the kind of knowledge and skills needed to improvise, the actual process of acquisition is often left to the reader's imagination. Dolan lists a number of strategies by which nervous novices can begin to improvise and mobilise their learning experience. For example, he recommends setting improvising on given structures and frameworks - these can be created by reducing repertoire to skeleton scores in which just the main structural features are laid bare; the improviser then re-improvises the score on these features, creating new versions of the piece. Another technique is to set a structural goal point as the focus of the exercise and the focus of the student's attention, i.e., "improvise to this point in the piece" which helps the student to bypass the details of texture which tend to overwhelm one at the outset. This is significant in that Dolan recognises the role of critical feedback in novice improvisation - and how this must be suppressed to a degree at the outset in order for improvising (and learning) to flow.

Lastly, Dolan's articulation of improvisation as three elements - (i) a general knowledge base of patterns, moves and technical know-how, (ii) a 'referent' which is the stylistic knowledge structures used for a particular improvisation, and (iii) the ability to manipulate one's knowledge towards creative ends - is amazingly useful in considering the problems of improvisation pedagogy. All three elements are crucial to the improviser's skill, yet each element involves a different kind of learning approach. Whereas arguments over pedagogical methods usually rage over the question of rules versus freedom, Dolan's three elements (if we treat all three with equal respect) offer a varied and holistic approach to teaching. For example, one could work on developing the knowledge base through patterns, analysis, conscious control and even prescribed exercises; this work can also be extended to particular structures, forms and genres. Yet, important though this work is, it is not complete, for it's necessary also to approach the task in a completely different way, following one's imagination wherever it leads. Thus, neither 'rules' or 'imagination' is a complete pedagogy in itself, but is only a strand in the improviser's toolkit or armoury.

More on David Dolan's work can be seen through the METRIC organisation. http://metricimpro.eu/

 

Prelude schemas

I had a great time today writing out 'chord streams' for Preludes. These are rows of chords which might be used to structure a free Prelude. I've had many problems with improvising free Preludes - usually it's the first thing I do in church, just when I'm feeling a bit cold! I tend to have a good opening idea but then don't how to shape this into a bigger coherent form.

This morning, the time felt right to have a go at these exercises. As I wrote them I had a chance to focus on different things - the flow of pure harmony; the larger structural shapes, the sentences and paragraphs underlying Preludes. This was nice because usually when I improvise I'm forced to focus more on the surface details: how things fit together in a more detailed way. This stops me from understanding the larger structure and often forces me to go in directions which don't really make sense or feel musically satisfying.

While writing these I felt as if important connections were being made - a sense of timing of events which was important to me. Afterwards I improvised for a long time on the piano, not on these structures, but freely, drawing on the connections I'd made through the experience of writing

Managing consonance-dissonance in diatonic tonality

 When Joseph Fux (1660-1741) describes - in the Gradus ad Parnassum - that some consonances such as the sixth or third are less perfect than the fifth for example, I have at first to accept this distinction in good faith. The fact is, as a musician of the 21st century with ears accustomed to a high level of dissonance, it's difficult to perceive these subtle dissonant-consonant relationships without some kind of explicit sign-posting. Only through practicing the exercises, by following the rules explicitly and painstakingly do I begin to feel the governing rules of consonance and dissonance of this style. Over a period of time I begin to feel more sensitive to certain intervals, to recognise them as performing certain functions within the musical texture. Fux's exercises help me to label and organise a diatonic musical fabric in a rich and meaningful way. 

         The results of this work of absorbing Fux's rules is therefore incredibly useful. For how often have I pursued the products of imagination with the feeling that I lack a map with which to find my way home; how often do I try to pursue ill-conceived 'rules' of harmony and voice-leading at the expense of all imagination? I am led through these exercises to experience for the first time, the underlying structural principles on which tonal music is built. Fux assigns functional properties to each interval which dictate the possible subsequent movements, either towards establishing a new dissonance, or resolving an existing dissonance towards consonance and resolution (7 to 6; 4 to 3 for example). To become aware of these rules is, for me, a significant advance over my previous approach or conceptualisation of dissonance. For previously I wandered around in ignorance, reacting to certain superficial properties which I attributed to particular 'chords' which my hands performed rather easily. I was aware that these chords had expressive potential, as they contain dissonant intervals which sound good to my ears (also because they are extracts or distillations of musical moments that are significant to me). However, even while playing these same chords, I was not able to fulfill the expressive potential that I perceived in them: I moved speculatively from one harmony to another, without understanding the principles on which I moved. I could not therefore control either the music or my actions satisfactorily, for the environment was one of continuous change and impermanence. 

This disorganised and unsatisfactory way of proceeding rarely improved, for I faced the same problems in the same way each time I improvised. Though I found temporary (harmonic/countrapuntal) solutions as I went along, I had no framework for understanding how I arrived at these solutions; nor could I perceive if indeed the solutions are good or bad, as the halting, unsatisfactory context in which they are performed contained no clear point of reference against which I could assess the success or failure of my methods.

By absorbing the functions which Fux gives to dissonant intervals, and the rules of progress from one interval to another, I began to understand the power of restraint. In my previous search for expression I tended to amass dissonance; through species counterpoint I learn to honour the resolution of a dissonance before establishing another. It required the exercise of conscious restraint, but this restraint resulted not only in a greater clarity of expression, I sensed too that there was more power, more organisation in the music. It was (for want of a better expression) better music, and certainly more interesting to listen to than my previous unstructured rambling!

It is for this reason that I belive progress in controlling dissonance is a conscious and careful process, which has a gradual effect over time on improvisation. So, to summarise:

1. Wrong notes certainly do exist, in the mind of the improviser and in the perception of the listener, unless the improviser adopts (as is possible) an anarchic approach of uncontrolled dissonance.

2. If, however, the improviser adopts a intuitive approach of controlling dissonance this is a chancy and inexact solution, for the reason that our ears are so attuned to dissonant intervals there is tendency for improvisers to amass dissonance in the search for greater expression.

One final thought: Fux’s rules represent a method for understanding the relationship between dissonance and consonance when using a systme of diatonic tonality. In this sense it serves as a means of control over a particular musical system. By absorbing his fundamental rules one is able to assign functions to certain musical properties and thus perceive the stimulus and the consequences of one’s actions in a stable and secure way; to make (creative and generative) decisions with certainty and, in this way gain control (through understanding) over contrapuntal textures, harmonic movement and so on towards larger sections of musical structure.

The use of this ‘means of control’ through the clear and explicit explanation of conceptual relationships illustrates the need for such mental tools whenever one improvises. That is, there must be a musical system (of constraints) and a means of controlling or manipulating these constraints is the improviser’s actions are to be effectively organised and mentally represented.

 

 

Further discussion on learning (2)

All about learning - conscious? unconscious? What's the difference? Does it matter?

Previously I quoted from Nick Ellis's (2011) chapter: Implicit and Explicit SLA and Their Interface, which is found in Sanz, C., & Leow, R. P. (Eds), Implicit and explicit language learning: conditions, processes, and knowledge in SLA and bilingualism. Washington, DC: Georgetown University Press.

Ellis is one of the few writers who not only distinguishes between the two types of learning process (implicit and explicit) but also tackles the issue of what role each process plays in learning. I began by quoting his distinctions:

  • Implicit learning is acquisition of knowledge about the underlying structure of a complex stimulus environment by a process that takes place naturally, simply, and without conscious operations (p.38)
  • Explicit learning is a more conscious operation, where the individual attends to particular aspects of the stimulus array and volunteers and tests hypotheses in a search for structure (p.38)

But then proposed my own assertion that implicit (unconscious) acquisition of the rules necessary for improvisation was not sufficient for improvisation. Why not? If such processes are sufficient for learning the complexities of our mother tongue (L1), why not for music? Such a comparison is, after all, valid if we believe that the ideal improviser is one who can 'speak' music in the same way that one speaks one's birth language. Would implicit (unconscious) acquisition of the rules then be sufficient if we learnt to improvise in the same way as we are taught to speak a birth language? To reconstruct these conditions would be an interesting experiment! It would mean:

  • thousands of hours of one-to-one instruction.
  • the use of music to communicate one's basic emotional and physical needs.
  • the acquisition of skill during the years of greatest neurological plasticity and cognitive growth, i.e., one would learn to express oneself in music in parallel to the processes of adapting to and understanding the surrounding social and physical environment.

It's relevant I think to note that L1 acquisition is not as implicit as it is often implied. Explicit instruction need not be restricted to the teaching of abstract rules and theory, but can also apply to forms of guidance in which the individual receives 'explicit' (i.e., clear) feedback, guidance and correction. Successful learning could hardly occur through meer passivity, as Ellis remarks:

Whatever the domain, learning the patterns, regularities or underlying concepts of a complex problem space or stimulus environment with explicit instruction, direction, and advances clues, heuristics, or organisers is always better than learning without any cues at all. Ellis (2011, p.38)

It's my personal conviction that musical improvisation could be learnt in the same fashion as L1 under the optimal conditions outlined above. Quite possibly, improvisers who learn the skill of improvisation also benefit from some of these conditions - in particular the learning of improvisatory skills alongside other musical skill formations, and the development of these skills during periods of general cognitive and personal development. In a sense, young improvisers in common with L1 learners are equipped with a mental white sheet, whereas classically-trained performers have to adapt their existing systems of musical perception to the new ones of improvisation, in common with L2 learners. In this, second case, the contexts of learning (of both music and language) are usually not comparable, calling for a different approach to learning, as Ellis (2011) describes: 

In contrast to the newborn infant, the L2 learner’s neocortex has already been tuned to the L1, incremental learning has slowly committed it to a particular configuration, and it has reached a point of entrenchment where the L2 is perceived through mechanisms optimised for the L1. Thus the L1 implicit representations conspire in a “learned attention” to language and in the automatised processing of the L2 in nonoptimal, L1-tuned ways. (Ellis, 2011, p.40)

Let's look more closely at the contrasting nature of these two types of knowledge could be applied to learning situations, using the insights of Ellis (2011), Hiebert (1986), Carpenter (1986) and others.

EXPLICIT LEARNING.

  • Ellis (2011, p.38). When the material to be learned is simple, ... and the critical features are salient, then learners gain from being told to adopt an explicit mode of learning where hypotheses are to be explicitly generated and tested and the model of the system updated accordingly. As a result they are also able to verbalise this knowledge and transfer to novel situations. 

  • Hiebert & Lefevre (1986, p.3-4). Conceptual knowledge is characterised most clearly as knowledge that is rich in relationships. It can be thought of as a connected web of knowledge, a network, in which the linking relationships are as prominent as the discrete pieces of information. Relationships pervade the individual facts and propositions so that all pieces of information are linked to some network. In fact, a unit of conceptual knowledge cannot be an isolated piece of information; by definition it is a part of conceptual knowledge only if the holder recognises its relationship to other pieces of information.

  • Ellis (2011, p.41). ... conscious processing is spread wide over the brain and unifies otherwise disparate areas in a synchronised focus of activity. 

  • Ellis (2011, p.41). Conscious activity affords much more scope for focused long-range association and influence than does implicit learning. It brings about a whole new level of potential associations.

  • Ellis (2011, p.42). Compared with the vast number of unconscious neural processes happening in any given moment, conscious capacity evidences a very narrow bottleneck. But the narrow limits of consciousness have a compensating advantage: consciousness seems to act as a gateway, creating access to essentially any part of the nervous system. ... It is a facility for accessing, disseminating, and exchanging information and for exercising global coordination and control: consciousness is the interface.

  • Ellis (2011, p.43) ...where L1 experience has tuned the learner’s attention elsewhere. ... Instruction is thus targeted at increasing the salience of commonly ignored features by first pointing them out and explaining their structure and, second, by providing meaningful input that contains many instances of the same grammatical meaning-form relationship. 

  • Ellis (2011, p.45).Consciousness creates access: its contents are broadcast throughout the brain to the vast array of our unconscious sources of knowledge, and by these means, consciousness is the interface.

IMPLICIT LEARNING.

  • Ellis (2011, p.35): The acquisition of L1 grammar is implicit and is extracted from experience of usage rather than from explicit rules - simple exposure to normal linguistic input suffices, and no explicit instruction is needed.

  • Ellis (2011, p.38): When the material to be learned is more randomly structured with a large number of variables and when the important relationships are not obvious, then explicit instructions only interfere and implicit mode of learning is more effective. This learning is instance-based but, with sufficient exemplars, an implicit understanding of the structure will be achieved.

  • Ellis (2011, p.39): Implicit learning collates the evidence of language, and the results of this tallying provide an optimal solution to the problem space of form-function mappings and their contextualised use, with representational systems modularising over thousands of hours on task.

  • Ellis, 2011, p.41). ... implicit learning occurs largely within modality and involves the priming, or chunking, of representations or routines within a module; it is the means of tuning our zombie agents, the menagerie of sepcialised sensorimotor processors ... that carry out routine operations in the absence of direct conscious sensation or control.
  • Ellis (2011, p.41). ... our implicit systems automatically process the input, allowing our conscious selves to concentrate on the meaning rather than the form. Implicit, habitual processes are highly adaptive in predictable situations. But the more novelty we encounter, the more the involvement of consciousness is needed for successful learning and problem solving

THE ROLE OF CONSCIOUS KNOWLEDGE IN LEARNING. One thing that emerges from these useful definitions is that the role of conscious knowledge (declarative knowledge, conscious awareness or conscious control) becomes more circumscribed. It's clear that, although Hiebert & Lefevre (1986) describe conscious knowledge as 'rich in relationships' and 'a connected web of knowledge' these connections are conceptually different to the kind of multitudinous connections calculated by implicit knowledge, which, over time - Ellis mentions 'thousands of hours on task' - furnishes us with instincts for what is appropriate or fitting in form and gesture. There is so much of skilled learning and performance which cannot be ascribed to consciousness: most importantly (i) I cannot consciously control or train my muscles, (ii) I cannot consciously control, generate or even fully understand the functioning of the music I improvise in real time, (iii) I cannot consciously understand the processes of automatisation which are so essential to fluency, (iv) the very nature of automatisation is that it reduces demands on consciousness and allows the performance of parallel rather than serial processes.

Yet, Ellis (2011) claims that implicit learning is itself inadequate for the learning of a new language, due to effects of interference arising from already existing skills in the native language. In what way do the existing language skills interfere with the acquisition of new skills? Can these effects be comparable to learning improvisation? 

Problems seem to arise when existing language skills prevent the learner from noticing what is important (salient) in the new material: '...where L1 experience has tuned the learner’s attention elsewhere' (Ellis, 2011, p.43). Why should this occur? Presumably because individuals try to understand or process the new language in similar ways to that in which they habitually process or understand their native tongue. This feature alone I think is directly replicable in the learning of improvisation (= the acquisition of novel generative skills) by musicians trained in performance (= possession of interpretative skills). The fact that music is not learned implicitly in an L1 fashion should not disrupt the comparison because, for trained musicians, the initial declarative methods of learning music will have been long transformed into automatic procedures. (In other words, the trained and fluent performer possesses an intact and specialised knowledge base sufficient for the task of performance, and it is this specialised and intact knowledge base which is brought to the task of learning improvisation.) Of course, when we have great familiarity with a task we don't need to analyse the techniques we use to perform it, not do we think of the constituent parts of the task, or in what way these parts are combined to make greater structures. The step-by-step learning of our native tongue, so embedded in communicative needs and direct feedback, is long since transformed into unconscious performance; likewise, the hard-won technical mastery of a musical instrument and the skills of meticulous score-reading and interpretation are second nature to the trained performer. How then should a musician with such a training look anew at a score and understand it as a model, say, for improvisation? As Ellis suggests, ',,, our zombie agents, the menagerie of specialised sensorimotor processors' which characterise implicit learning are inadequate for these initial stages of learning, powerful instruments though they are. The processes of unconscious learning must be steered in a certain direction by consciousness, given specific goals to head for, and have these goals reiterated through a clear framework of instruction: 'Instruction is thus targeted at increasing the salience of commonly ignored features by first pointing them out and explaining their structure and, second, by providing meaningful input that contains many instances of the same grammatical meaning-form relationship (Ellis, 2011, p.43). 

In this way, consciousness, though restricted as a tool for the actual performance of improvisation, and inadequate in terms of perceiving intrinsic matters of form in the material (whether music or language) through experience, can be seen to play an essential, leading role in the early stages of adult learning. How does it do this? Ellis suggests that there is no part of the nervous system to which consciousness does not have 'access'. Through this right of access then, consciousness can act as an interface between disparate, automatic systems, as a force which coordinates, sychronises and ultimately exercises 'global coordination and control' (p.42).

PROCESS PURE: Are the two types of knowledge really distinct? Much of the research giving rise to these distinctions comes from tests on amnesiac patients (Green & Shanks, 1993, p.304): 'Although amnesics are severely impaired on tasks of explicit memory, their performance is often within the normal range on implicit memory tasks. Normal controls, as well as amnesics, exhibit this independence of performance on implicit and explicit memory tasks.' The same authors also warn that, with regard to implicit memory (recall of unconsciously learnt material) '... individual memory tasks are not process-pure. In other words, a single task may involve the contribution of more than one type of memory process. A given task, for example, may involve any combination of conscious and automatic processes, such as those that are conceptually and data driven (Green & Shanks, 1993, p.314).

Hiebert & Lefevre (1986), also warn that '... the distinction provides a classification scheme into which all knowledge can or should be stored. Not all knowledge can be usefully described as either conceptual or procedural. Some knowledge, seems to be a little of both, and some knowledge seems to be neither' (p.3).

 

Further discussion on learning (1)

SOME OPENING THOUGHTS

When I improvise, what is it that I learn?

1. I will experience learning through implicit processes, i.e. unconsciously.

2. I will experience learning through explicit processes, i.e. reflectively, of which I am aware.

What do I then retain from these experiences for the future? In what way have I developed?

 

VARIABLES (Can also be discussed under the 'goals' of the improviser)

1. I have a certain knowledge and disposition which I bring to the task.

2. I also have certain goals, set by myself, or my teacher. These goals, in all probability will form the focus of my attention while improvising. They will play an important role in the nature of feedback.

As this discussion has immediately raised the subject of implicit/explicit learning, here are some further thoughts about the distinctive characteristics of these different yet complimentary processes:

Nick Ellis (2011) makes the following observations:

  • Implicit learning is acquisition of knowledge about the underlying structure of a complex stimulus environment by a process that takes place naturally, simply, and without conscious operations (p.38)
  • Explicit learning is a more conscious operation, where the individual attends to particular aspects of the stimulus array and volunteers and tests hypotheses in a search for structure (p.38)

1. A NEUROLOGICAL DISTINCTION BETWEEN IMPLICIT AND EXPLICIT LEARNING:

Paradis (1994): ‘... that the memory system that subserves the formal learning of a second language (declarative memory) is neurofunctionally and anatomically different from the one that subserves the first language or a foreign language acquired in conversational settings (procedural memory)’ (p.393).

Ellis (2011): ‘Explicit learning is supported by neural systems in the prefrontal cortex involved in attention, the conscious apperception of stimuli, and working memory; the consolidation of explicit memories involves neural systems in the hippocampus and related limbic structures. In contrast, implicit learning and memory are localised, among other places, in various areas of perceptual and motor cortex’ (p.38).

Ellis (2011, p.40). There are different types of knowledge of and about language, stored in different areas of the brain, and engendered by different types of educational experience.

The preceding quotations propose that it is explicit functions which govern attentional systems, and involves emotional states. Implicit learning in contrast runs in an automatic fashion through processes of perception and motor coordination. Paradis (1994) has already implied that the essential elements of performance are learned implicitly and performed implicitly, i.e.,

It appears that what has been acquired incidentally is stored implicitly and can only be evidenced through behaviour (performance) (Paradis, 1994, p.401).

But what exactly do we retain from the engagement of implicit mechanisms? Presumably information which organises the action of improvisation, for after all, we know that we cannot coordinate our actions directly, but only in relation to a perceived point (i.e., Hommel et al. 2001, p.876, who conclude that: '... coordinating actions and events is much easier in terms of distal than proximal representations'). However, is the information resulting from implicit mechanisms sufficient for the task of improvisation? In my experience I would say no. I am currently practising various tasks in baroque improvisation. Tasks, such as harmonising chorale melodies in a German baroque style, or accompanying a phrase of gregorian chant as a cantus firmus in the French style, demand a knowledge (perception) of a given body of music to the extent that I can generate new music in this style on demand.' When I place my hands on the keyboard I need more than motor memory to see me through at this stage of learning! For this particular genre of improvisation the musical material must be highly organised to be a convincing representation of the style. I find that at the outset of each practice session I have very little idea of what I should do. After a few attempts at harmonising, my developing knowledge is engaged and certain patterns emerge. However, in general, much testing of choices and pauses for thought (and also for writing solutions) characterises my improvising, and any attempt to perform in this style (at this stage) would, I think, be doomed to failure.

Expert knowledge on the other hand is generally characterised by fluency. According the words of performers theselves, and the assertions of researchers into memory and human performance, their improvisation is generally conducted unconsciously. What then is the nature of their knowledge? Is it merely quantitative, i.e., does the expert simply remember more connections, more characteristic chord progressions (or contrapuntal textures, or cadences etc.) with their corresponding hand movements than I do? Or is the knowledge more qualitative in essence, being linked to, or integrated within a host of other cognitive, emotional and perceptive mechanisms? Why can't the knowledge of the expert be simply explained to a novice? Why does the novice themselves have to experience over time the nature of the musical material to attain expert insights? Note that Johnson-Laird (2002) asserts:

‘... whenever a creative process calls for working memory, it is likely to depend on a multistage procedure’ (p.239).

So, what are the various stages of a 'multistage procedure', and why is it necessary that human skill learning should be such a lengthy and subtle process? Clearly, by now, we can describe expert performance as primarily unconscious; presumably then, it is a necessary condition of unconscious performance that it takes considerable time to develop. In addition, we can assume that the material which we are learning to manipulate is poses problems which are extremely complex to master and to understand; that mastery of this material cannot be understood singly, but is the result of a plurality of experience, occuring sequentially (one step in the knowledge base must be assimilated before another can be perceived) over time; and that the resulting codes or representations of knowledge used by expert improvisers refer to a multilayered, multi-experiential database which cannot (for example) be transfered to a third party. 

 

Surely it makes sense then to look to general theories of cognitive development for illumination of this process, as (I have no doubt) that expert skill, resulting as it does from such a considerable investment of time, attention and cognitive resources, should be described in terms of cognitive growth. 

One aspect of learning which seems to me overlooked by researchers, is the fact that the novice often cannot bring his (implicit) critical or evaluative faculties to bear on his own work. That some rules of music construction are unconsciously perceived seems to be beyond question, i.e., Tillmann & Bigand (2004):

By mere exposure to musical pieces, Western listeners acquire sensitivity to the regularities underlying tonal music. This implicitly acquired knowledge allows listeners to perceive subtle relations between musical events and permits musically untrained listeners to behave as musically trained listeners ... (p.211) 

So why, then must the pupil Josephus, in Joseph Fux's (1725) famous treatise Gradus ad Parnassum, be explicitly instructed in rules of counterpoint which exist primarily to avoid unpleasant acoustic effects:

'This mi against fa you have written in the progression from the sixth to the seventh bar by a skip of  an augmented fourth or tritone which is hard to sing and sounds bad, for which reason it is forbidden in strict counterpoint (Fux, transl. Mann, 1971, p.35)

If these rules reflect all that is natural and pleasing to the ear, why cannot Josephus apply his listening and critical faculties to his own work? Nor do I need to look towards literature for illustrations of this problem, as in my own case, when I am harmonising a chorale: why is it that I can be utterly convinced of a solution while writing (a certain chord or distribution of voices), only to find that later such a solution is clearly wrong and inadequate? Why is 'doing' so far removed from 'hearing'? Is it simply a question of attention, which is claimed by selected goals of learning at certain stages? While desperately trying to fulfill certain goals, do we simply fail to notice other problems? I notice that Johnson-Laird also comments on this phenomenon when he refers to Perkins' (1981) description of '... the fundamental paradox of creativity: people are better critics than creators.' And that artistic creativity '... is based on unconscious knowledge acquired only by laborious practice in creating' (p.422). However, this offers no more than a description of the problem, which it is implied will resolve itself over time. The alignment of skill with the unconscious perception of rules (proportion, beauty etc.) is presumably a goal of skill learning, but more information is needed to (i) explain why this misalignment should occur in the first place, and (ii) through what processes or stages of development are these two different areas brought into alignment?

 

 

Problems of learning in improvisation.

My main motivation for starting this PhD was in order to improvise better. Better? Well, to learn how to express myself musically, to escape from the frustration and confusion I felt when I tried to improvise. The motivation was in this sense primarily musical. But, I realised that, in order to attain this musical goal, I had to make sense of my experience in some way, which was why I turned to psychology as the field of investigation rather than seeking another strategy, such as practising harder. I was unclear what to practise, so to increase the hours spent at the keyboard would have been to simply augment my confusion.

I remember that at the outset I had the impression that I might, through a certain type of investigation, discover discrete answers to the problems I faced. For example, as one of my main problems was in choosing the next chord, I thought I might discover a technique for encoding sequences of chords which would solve this problem. I suppose that the difficulties of improvisation occurred to me then to be difficulties of processing information: experts were somehow able to process musical material efficiently, whereas novices could not. 18 months later, I have discovered a host of different perceptions on a vastly more complex subject than I originally perceived. I have experience in intuition, in which the next chord is felt rather than (theoretically) known; my inner ear has developed, and I have explored different modes of relating to improvisation - through movement and through an increasing awareness of my own intentions. More recently I started lessons with a pedagog and expert improviser which caused me to profoundly question my own musical knowledge, and the ways in which music is known, (encoded, memorised, represented among other terms) so that it can be used for improvisation. In some ways, I suspect that I may come full circle, and perhaps indeed discover (for myself and under certain musical conditions) a discrete form of (for example) harmonic representation, but certainly this knowledge comes as part of a gradual and multimodal increase in experience, and could never be written down as a formula and communicated to a novice in a form in which they would then understand what I now understand. 

What do we learn, when we learn something?

My more recent interests could be formulated by this question: When J.S.Bach copied the works of Nicolas de Grigny in 1713, what did he learn from this process? Copying music is quite a common technique for learning, but what is it that we learn from copying which we don't learn from playing or reading? Johann Sebastian, from accounts of his life was earnestly seeking a deeper understanding of the ways in which music could be constructed. Apparently he was quite indefatigable in this respect. What did he then discover from his copying? Do I discover the same features of the musical material, or reach the same understanding of these pieces as Johann Sebastian when I copy them? Presumably, as it's not hard to imagine someone engaging in the same task and discovering nothing at all, then we could propose that what we might gain from the process is a very personal process, and that, for example, our motivations for copying may influence our perception of the musical material, just as much as mechanisms such as priming (what we are currently interested in) or in fact, any of our individual cognitive and emotionl states might dictate a different form of understanding. Are there any generalities then that can be learnt or surmised?

A specific type of knowledge for improvisation?

The ways in which individuals seek and discover knowledge are questions crucial to the subject of improvisation, because the improviser is a seeker after knowledge, but a knowledge of a particular type. We need a type of knowledge about music that enables us to generate music at speed. If I say it is not therefore a reflective knowledge suitable for discussing and analysing musical scores, but an action knowledge suitable for adaptable, generative output in real time, then it becomes clear that the difference may not be the knowledge or facts of the matter which are conceptually different, but the form in which those facts are represented, encoded or understood in a way which facilitates improvising. In this sense we can talk of the reflective knowledge being interpreted for action (see Anderson, 1982). Unfortunately, this results in forms of knowledge which may be unavailable to conscious awareness, eluding analysis, as Paradis (1994) reveals, when he discusses the development of speaking capacity in individuals' linguistic development:

The aspect of practice which is useful, i.e., conducive to improvement of competence, is not the aspect(s) of which the speaker is aware. The aspect of which the speaker is aware does not get automatised, transferred or converted into what is stored in procedural memory - and what does get stored is not within the speaker’s awareness. Practice of utterances is what is directly conducive to improvement of implicit competence. Knowledge of the rules is not (Paradis, 1994, p.404-405).

Knowledge of facts, or knowledge of action?

So, through action - perhaps the action of writing/copying, or perhaps the action of improvising, aspects of knowledge become secretly encoded. And it is this secretly encoded knowledge, unavailable to consciousness which pertains directly to the development of the action itself. Does this mean we should dispense with factual learning all together?Apparently not, as Paradis (1994) in the same article also argues for a role of explicit awareness of rules in forming productions (automatic performance): 'It appears that what has been acquired incidentally is stored implicitly and can only be evidenced through behaviour (performance). On the other hand, some deliberately learned tasks seem to gradually become automatic through prolonged practice' (Paradis, 1994, p.401). So, consciously learnt facts do play a part in some aspects of learning! As this may appear rather confusing I will quickly summarise some of Paradis's findings, and then use these to illustrate the problems facing novice improvisers:

  1. that the implicit or unconscious memory systems used for fluent performance (i.e. first language production or expert improvisation) are 'neurofunctionally and anatomically different' (p.393) from those used for conscious recall of declarative information (i.e. formal study of second language, or theoretical knowledge about musical improvisation).
  2. that, as a result, individuals have two types or sources of knowledge, one which is available to consciousness (explicit), and one which is not available to consciousness (implicit).
  3. that, when we act fluently (i.e., speak or improvise) we rely on implicit knowledge for this fluency, In the words of Paradis (1994, p.400): '... in natural conversational settings, the production of the utterance remains non conscious.' Explicit knowledge then can only be used to slowly construct performance (a rather painful process for the performer and listener alike!), but explicit knowledge can also be used to reflect or check on the output of performance. However, explicit knowledge cannot control performance; as a knowledge type it is entirely unsuited for this task.
  4. that, because of this unsuitability, the introduction or intrusion of one type of explicit knowledge during performance causes interference in the processes of implicit knowledge. (The effects of this interference cause well-known phenomena of slowing down or pauses during performance.)

From this, it would seem that, while facts play an unavoidable part of adult learning, it is only through the repetition and automatisation of these facts that anything approaching natural (fluent) performance can result. As to the nature of performance itself, Paradis's distinctions are important, because they describe in a particularly clear way an aspect of improvisation which is often difficult to understand and to experience: namely, that to be fluent, improvisation has to be unconsciously executed; that, in essence, we must act in a way that we cannot know or describe. 

Implicit knowledge gained through action.

However, this distinction between declarative and procedural knowledge, the functions of which are so clearly described by Paradis, is not a symptom purely of fluent performance, nor is it a distinction that can be understood only in terms of improvisational output. Rather, the relationship between unconscious and conscious is dynamically present at all stages of learning; or perhaps it is better conceptualised as the role of the unconscious in all stages of learning. For example, if we return to my opening question: When J.S.Bach copied the works of Nicolas de Grigny in 1713, what did he learn from this process? we can now conjecture that, in addition to what Bach saw in the musical material, and recognised in theoretical, analytical, or musical terms, he also learnt something from the action of writing. And that whatever perceptions were gained through this action, yet more learning would have occurred when Bach attempted to generate new music from this knowledge, either by writing or improvising. For, through the actions themselves, an awareness of action is created, and '... when often repeated, end up leaving proprioceptive kinaesthetic traces' (p.401) and it is this 'proprioceptive data' that is 'stored in implicit memory'. As Paradis is at pains to point out, it is not the facts of  improvisation, or music which are rehearsed, but their production in performance (of those facts), and the processes by which we interpret and act on these facts are not known to us. 'The speaker (or improviser) is only conscious of the result, not of how it is obtained' (p.401). 

Not that this aspect of learning can account for J.S.Bach's musical knowledge! Far from it... either that that he learnt from De Grigny, or from listening to Dietrich Buxtehude or any other aspect of his musical formation. But, it does I think account for an often overlooked aspect of learning which is particularly useful for illuminating the learning of improvisation. 

 

Competence vs. Performance

Distinction between competence - performance in language and improvisation; presence of rules - are rules a good aid to learning? - contrasting the rules of theory and musical construction, vs. rules of production.

(image, Noam Chomsky)

One idea which constantly recurs in theories of language learning is the idea of competence - a person's knowledge or awareness of language, as against their performance - the actual production of words, sentences, phrases etc. Can this distinction, which has come under considerable criticism in the field of linguistics, be usefully applied to improvisation? By usefully, I mean can the concept of two differing forms of knowledge and ability help in pedgaogical or performance situations of improvisation?

My first thought is that one form of ability - the performance, can be much more easily assessed than the other - competence. (Competence, as a concept, was originally formed to describe the linguistic knowledge of native speakers, who seem to miraculously gain knowledge of a language within the first five years of development without recourse to formal learning of grammatical rules.) Performance, for example, is the evidence of a person's ability: when musicians improvise, then it is the performance of their knowledge that we hear, enjoy, criticise and process as listeners. Conversely, a person's competence - what they know about improvisation - is hidden, revealed only through conversation or through the act of improvising itself. Perhaps for this reason, research and teaching often focuses principally on the performance of individuals, perhaps comparing improvisations to compositions (reflecting on the artistic value), or by teaching students as if their faulty performance was entirely and consciously willed - instead of the outcome of incomplete or insufficiently connected knowledge. 

In this sense, Chomsky's distinction is valuable, because it is on the knowledge base (competence) of musicians that improvised performance is constructed, and it is upon the nature of this knowledge base that research and teaching should (in my opinion) focus.

One aspect of language learning which is I think directly linked to improvisation- at least the rather strict style of baroque improvisation which is my present study - is the question of rules. Before discussing the acquisition of rules, let's first define what they are. Some improvisers may reject the very idea of artistic creativity being rule based. My own difficulty with rules comes from the limitations of working memory, which argues against the conscious use of rules while improvising, i.e., "Don't give me another rule to remember, just let me express myself!". Yet, we must accept that the 'language' or chosen genre in which improvisers express themselves musically or in which they construct their artistic productions, must, to be coherent and recognisable to a listener, follow certain rules.

My recent lessons with Jürgen Essl in Stuttgart taught me that my knowledge of certain rules - we could call them characteristics of J.S.Bach's style of writing - was shaky to say the least. For example, when harmonising a chorale melody I did not perceive that each phrase should remain in one tonality, defined by the final cadence. To understand this allowed me to construct a harmonisation which was more true, expressive and recognisable within the idiom of J.S.Bach's style of harmonisation.

Ideally, we would improvise as we speak our native language. Certainly I have imagined ideal scenarios similar to L1 acquisition in which improvisation occurs as a natural communicative device between mentors and novices, who through many hours of exploration and one-to-one guidance allow natural learning abilities to develop a rich resourceful capacity for improvising directly linked to emotional intention. Yet, such idealised forms of learning to improvise are unlikely to occur when we are exposed to the learning process as adults (in which we are considerably more self-aware and conscious of the learning process); or when learning to improvise occurs as an addition to our existing musical training (in which case new knowledge and new ways of understanding music must be connected with existing forms, a source of much friction and conflict). In these situations some similarities can be observed between learning a second language and learning to improvise, in which case the insights of researchers into L2 acquisition can be usefully employed as guidance to improvisers.

Sharwood Smith (1994) describes the strange dichotomy between child and adult language learning thus:

First language acquisition happens in the natural world in the natural course of things as children try to communicate with the outside world. Older learners rarely experience this both naive and successful way of language learning because they are older and more worldly-wise. They have some idea what language is. They know it is a system with rules and principles that have somehow to be acquired. They worry about their failure to work that new system. They want to understand more about the nature of the task in the hope that this understanding will help them more. At the same time, the increase in awareness about the language being learned often seems not to be accompanied by a corresponding increase in proficiency, hence the question: do rules help?

From this account, it might be reasonably assumed that adults then somehow obstruct natural learning capacities? Perhaps, because of embarrassment or self-consciousness, adults are unwilling to experiment or play with language; after all if you don't try then you can't learn. On the other hand, Reber (1993) warns against the assumption of mysterious or innate powers of language competence in the child, reminding us how, during the first seven years of the child's life '... just how much learning takes place' (p.151). If, then, a child consciously practices language for 3 hours a day, '... we still find a very large number, 8,214 hours of practice in one's native language. Cut the estimate to 1 hour a day and we still end up with 2, 738' (p.152). This number he argues is consistent with research which asserts the amount of time needed to '... learn the structure of any rich, complex, abstract domain such as those that underlie natural sciences like physic, social sciences like psychology, or performing arts like music or dance. Numbers between 1000 and 5000 hours keep cropping up, no matter what the discipline' (p.152).

To return to Sharwood Smith's question: 'Do rules help?' it seems sensible to explore the different meanings given to rules by individuals in different learning contexts. 

For example, when Sanz & Leow (2011) conclude: 'Whatever the domain, learning the patterns, regularities or underlying concepts of a complex problem space or stimulus environment with explicit instruction, direction, and advances clues, heuristics, or organisers is always better than learning without any cues at all' (p.38); they refer to rules as cues - a substantially more friendly term than rules. Sharwood Smith (1994) also widens the terms of reference: 'The idea of a rule - whether it is called a rule, a routine, a plan or a process is immaterial...' (p.37), to illustrate the plain existence of rules as a means of structuring or organising experience. Perhaps the main problem with rules then is our usage of them; the danger being that in some learning environments, theory takes precedence over performance and rules '... may be couched in very abstract and technical terms such that they do not appear to offer much of immediate benefit to teacher or learner' (Sharwood Smith, 1994, p.34). Taken to further extremes, rules can be couched in coercive terms - a characteristic of much musical teaching of the 19th century conservatoire approach. In this sense, rules can be more than cues for structuring perception and understanding; instead they are conditions under which music-making is accepted or rejected. Such rules or 'Rules', if taken literally by the student, dominate the learning process in particular through their influence on attentional processes and feedback, as the student seeks naturally to align the ouput of improvisation with the idealised (imaginary) product of the Rules. 

One interesting aspect of Sharwood Smith's article (1994) is his elaboration of the different kind of knowledge needed for speaking, rather than listening or thinking about language: '...we may assume that the on-line production and reception processes during actual language performance are governed by principles requiring theoretical explanations of quite a different character to those used to explain competence' (p.37). This interested me because, in my experience, my competence (in the Chomskian sense of implicit or idealised knowledge) in improvisation greatly exceeds my improvised performance. For example, while driving to church I can plan and imagine the most marvellous improvisations, which, when I approach the organ seem to fade, and when I start actually improvising evaporate entirely, leaving little if any trace behind. Clearly, when I start to play, I am under the authority of a different set of constraints altogether, and these constraints or principles, as Sharwood Smith suggests, require a different theoretical explanation. (This, being a private or inner experience takes the investigation of improvisation in a new direction: if we are to discover or describe the consciousness of improvisation then a new framework is needed, a phenomenological account which could (i) describe this experience in embodied, first-person terms, (ii) explore research into proprioceptive memory and experience, and (iii) analyse ways in which musical material is inwardly represented.) 

For Sharwood Smith (1994), what follows from this distinction between the knowledge we bring to language use (or improvisation) and the knowledge we actually use while  speaking (or improvising), is that rules of construction once more assert themselves into the process. Improvising and speaking take place in real time, and thus, events have to be organised into a serial stream. Ideally, at expert level, this stream is fluent and is enacted without breaks and pauses, meaning that the individual has to have the material-to-be-improvised sequentially organised. While we may theoretically adopt a relaxed attitude towards construction (for example, it is often said in improvisation circles that mistakes don't matter and can be turned to creative advantage), the fact is that the improviser must '... determine what takes place first, what second, and what takes place simultaneously' (p.37); if they cannot determine this, for what ever reason, then faults emerge in their performance. This necessity for organisation can, as Sharwood Smith suggests, best de described as rules, and, whether applied to language or musical improvisation, such 'performance rules determine what happens millisecond by millisecond, second by second, etc.' (p.37).

Thus, performance itself is governed mercilessly by its own rules - rules of production. The acquisition of these rules, through experience and learning is the subject of other posts.

 

 

 

 

 

  • rules and creativity
  • attitude to rules
  • acquisition of rules - contrast between theory and performance.

References:

Sharwood Smith, M. A. (1994). The Unruly World of Language. In Ellis (Ed.), Implicit and explicit learning of languages (pp. 33-44). Bodmin: Academic press.

 

New lessons!

I recently started a new course of study with the composer, organist and improviser Jurgen Essl in Stuttgart.

This is a wonderful and salutary experience for me, having worked so much on my own, to bring my development as an improviser to the attention of someone with this level of experience and deep musical knowledge.

My first lesson was, from the point of view of my own proficiency (i.e., as a demonstration of skill), a disaster! Clearly I was unable to piece together even the simplest material in the Baroque style which I had been studying for so long.. From a psychological point of view, perhaps it was more successful: rather than feeling cast down by my miserablt show of abailities, I was able to focus on Jurgen's instructions which clearly formed a basis for further development. I think if this kind of experience had happened to me as a young man, I would have been very upset and humiliated, so clearly I have learnt something from my psychological investigations!

In general, I knew it was not going to be easy, but still I was surprised at the extent to which I was unable to produce a simple, fluent improvisation. In spite of my recent work improvising in baroque style at home, it seemed as if a vast gulf exists between what Chomsky (in a linguistic context) would refer to as my competence (mental awareness, implicit knowledge and ideals) and performance (what I actually produce). Let's consider then what I had been doing:

  • I had been piecing together lots of ideas in a baroque style, practising figurations, making sequences, and trying to understand the various relationships of consonance and dissonance. 
  • I had been studying works by Bach and Buxtehude, analysing some of the structural plans, writing out the rhythmic motifs.
  • I had been practising fugue constructions, using a variety of different fugal themes.

This practice had left me with an unfocused and generalised knowledge base.

While playing at home, I could sometimes achieve nice results - but these results were momentary, interspersed amongst a great deal of trial and error. Probably my biggest achievement was emotional, in the sense that I was not afraid to improvise: I could enjoy the process of trial and error, and could play around with all different types of musical construction. My biggest problem was that, when it came to a 'product' moment - when I actually had to come up with something coherent, then I didn't know what to do. Playing in church for example was an ordeal, because I had no resources or techniques with which to organise my material. I was left with a very complex mental state, of incoherent goals, and (corresponding) emotional reactions. Note, that the incoherence of my knowledge (by this I mean the procedural knowledge of 'what to do') left me with other problems to confront: I questioned my relationship with the listeners, the context in which I found myself, my attitude towards music, performance, etc.

Working with Jürgen taught me the following:

  • To focus unrelentingly on one aspect of the texture - in this case the harmonisation of a chorale theme.
  • Jürgen told me to decide on a harmonisation of the chorale melody, and to keep this harmonisation for further treatment (variations). 
  • From this harmonisation, came tasks such as putting the melody in the alto, the tenor and the bass.
  • To harmonise with three voices in the right hand, just the bass in the left; to harmonise then with two voices in the left (alto and tenor), with an ornamented melody in the right, and the bass in the pedals; also to harmonise with two voices in each hand.

Deciding on a harmony - i.e., to have a memory of that harmony was something I found I couldn't do as my approach to harmonising was too random. My goal (to this point) when harmonising was to find something that fitted and then to move on. This meant that the harmony I selected had no real function or relation to something which I could define. It was ad hoc, - to hand.. For me, this ad hoc-ness was quite an achievement, as it meant I had progressed from searching for the 'correct', theoretical version, to something improvised, something which allowed for margins of error. I had stopped searching for the version, and was now searching for version; I had emerged from a product state of mind, to a process approach.

Jürgen's advice seemed at first to be a return to the product view: I had to define my harmonisation, and then to write it down. Why do this? To define more clearly what the harmony is? Actually, I think (I write now from the perspective of a week's fairly intensive experience) the process of writing down is to understand more clearly what the role of a harmonisation is - perhaps in terms of its possibilities, or functions. Remember that before I was using a bricolage or ad hoc technique. Anthing could be used so long as it sounded good in that moment. Once I'd played my choice of harmony, it was gone, because that moment had gone - it was not connected to anything else in the improvisation.

Writing down is an iterative technique, as one returns from writing to playing, and repeats this process until a version is arrived at. It is also an elaborative technique, as to write is to see one's harmonisation more objectively, perhaps too to see it with the same mechanisms as one would view a printed score. No longer is the harmonisation a thing of air and fantasy, the possibilities of a moment; it is corporeal, and can be repeated verbatim. After writing, I also see visually some connections within the harmony which I didn't perceive when playing; from this I find that some problems can be viewed graphically and perhaps solved analytically. Writing down the harmony does not produce a solution, as often new 'solutions' arrive through playing, in which case writing takes a form of dictation: writing usually takes its cue from the playing, but it helps to refine and instruct the process of playing.

Learning can be a difficult process, because one's faulty or incomplete knowledge leads one's instincts and capacities for doing astray. It's hard to arrive at a solution and be told that it is not a good solution! But, in this disciplined form of improvising there are clear and practical reasons for doing certain things (i.e., ways of harmonising) in certain ways. But this is the subjetc of another post..

Impro blog

Buxtehude 2 - Fugues & Fughettas

Some more (tricky) fugue subjects and answers from Dietrich Buxtehude!

Fughetta subject in G major

answer

Also this one in G minor and in compound time (12/8)

answer

Buxtehude - preludes

Using the Friesach sample set of Piotr Grabowski (see here for more details), am enjoying several break-throughs in improvising Buxtehude-style Preludes and free-style forms. These forms tend to mix short sections of declamatory, fantasy phrases with choppy fughettas and contrapuntal sections. The Friesach set gives me the right sound and mix I need to understand the articulation of Buxtehude's figurations. This 'choppiness' is important - more than can be conveyed in theoretical terms such as decoration or embellishment. This choppiness is what first strikes me when I hear this music, as I walk into the church or imagine the movement of the music. It is jagged, complex, angular music - fleeting, moving and pursuing. It captures the flight element of the Fugue - the chase of the subject as it appears and disappears through the texture.

Registration is crucial but not tricky. Again, it depends a great deal on the type of organ used. The Baroque organs of N-Germany were characteristically bright with clear, sculpted sound structures. A typical 8', 4', 2' is mobile, clear and authoritative. Adding a mixture stop results in a strong, even complex sound which might prove too complex for fugal work.

Focus on David Dolan

David Dolan stands out in the world of classical improvisation. Why?

He's an important pedagogue who specialises in teaching non-improvisers (classically-trained musicians at the highest level, who have experience only in interpreting written scores) to improvise. This means that not only has he devised ways of overcoming ingrained habits of musical practice, ideological beliefs associated with these habits, and the consequent creative barriers of classical training, but - and this particularly appeals to me - his teaching stays within the repertoire of classical music: that is, he does not resort to another type of music such as jazz, or free-style improvising, but takes classical models as the basis for experimentation.

This post is based on an article he wrote in 2005, entitled Back to the Future: towards the revival of extemporisation in classical performance. Dolan's teaching approach starts from the consideration of improvising knowledge (what does one need to know in order to improvise). This knowledge base he defines through the idea of 'schemes' - the 'learned musical schemes' of a particular musical style and culture, and the 'natural schemes' which are more universal aspects of human perception, being related to speech, expressive body language, sense of proportion, variety, etc. Often these two different types of schemes meet in musical language, for example many musical structures take the form of an 'arch' which describes a natural trajectory through space and time, comprising a point of departure, and a highest point, from which the journey back to earth begins.

To improvise then, one needs to acquire not only theoretical and practical knowledge - knowledge of musical elements and instrumental technique - but also knowledge of how to manipulate this knowledge, how to mobilise one's skills towards expressive ends.

Dolan, thus puts refutes a common concept in improvising pedagogy: that one needs to first learn 'facts' or patterns through abstract exercises which are then (somehow) transformed into expressive, fluent improvised music. Instead he puts expressive gesture and communicative intent at the forefront of learning. Thus, students aim to make a broad gesture or communicative intention without reflecting too much over the details; once the gesture is made then one look back and consider what one did, and how to do it differently or better. As communicative gesture is movement based, there is a close afinity between Dolan's approach to learning and Jacques-Dalcroze' use of eurythmics, and insistence on learning through action first. Both men recognise the dangers of being lost in thought, the hesitation of conscious calculation, and how fear of errors can lead to lack of flow.

My feeling is that this type of learning is greatly assisted by the presence of an encouraging teacher as Dolan or Jacques-Dalcroze themselves. I can imagine the way in which Dolan constructs each exercise, establishes the criteria (rules of the game), gives critical or non-critical feedback, and so on, is very much part of the success of his teaching. Such a method might be more difficult to achieve on one's own, or where surrounding feedback is more critical..!

Dolan's work fills a much needed gap in knowledge of how to start improvising. While many theorists and writers agree on the kind of knowledge and skills needed to improvise, the actual process of acquisition is often left to the reader's imagination. Dolan lists a number of strategies by which nervous novices can begin to improvise and mobilise their learning experience. For example, he recommends setting improvising on given structures and frameworks - these can be created by reducing repertoire to skeleton scores in which just the main structural features are laid bare; the improviser then re-improvises the score on these features, creating new versions of the piece. Another technique is to set a structural goal point as the focus of the exercise and the focus of the student's attention, i.e., "improvise to this point in the piece" which helps the student to bypass the details of texture which tend to overwhelm one at the outset. This is significant in that Dolan recognises the role of critical feedback in novice improvisation - and how this must be suppressed to a degree at the outset in order for improvising (and learning) to flow.

Lastly, Dolan's articulation of improvisation as three elements - (i) a general knowledge base of patterns, moves and technical know-how, (ii) a 'referent' which is the stylistic knowledge structures used for a particular improvisation, and (iii) the ability to manipulate one's knowledge towards creative ends - is amazingly useful in considering the problems of improvisation pedagogy. All three elements are crucial to the improviser's skill, yet each element involves a different kind of learning approach. Whereas arguments over pedagogical methods usually rage over the question of rules versus freedom, Dolan's three elements (if we treat all three with equal respect) offer a varied and holistic approach to teaching. For example, one could work on developing the knowledge base through patterns, analysis, conscious control and even prescribed exercises; this work can also be extended to particular structures, forms and genres. Yet, important though this work is, it is not complete, for it's necessary also to approach the task in a completely different way, following one's imagination wherever it leads. Thus, neither 'rules' or 'imagination' is a complete pedagogy in itself, but is only a strand in the improviser's toolkit or armoury.

More on David Dolan's work can be seen through the METRIC organisation. http://metricimpro.eu/

 

Prelude schemas

I had a great time today writing out 'chord streams' for Preludes. These are rows of chords which might be used to structure a free Prelude. I've had many problems with improvising free Preludes - usually it's the first thing I do in church, just when I'm feeling a bit cold! I tend to have a good opening idea but then don't how to shape this into a bigger coherent form.

This morning, the time felt right to have a go at these exercises. As I wrote them I had a chance to focus on different things - the flow of pure harmony; the larger structural shapes, the sentences and paragraphs underlying Preludes. This was nice because usually when I improvise I'm forced to focus more on the surface details: how things fit together in a more detailed way. This stops me from understanding the larger structure and often forces me to go in directions which don't really make sense or feel musically satisfying.

While writing these I felt as if important connections were being made - a sense of timing of events which was important to me. Afterwards I improvised for a long time on the piano, not on these structures, but freely, drawing on the connections I'd made through the experience of writing

Managing consonance-dissonance in diatonic tonality

 When Joseph Fux (1660-1741) describes - in the Gradus ad Parnassum - that some consonances such as the sixth or third are less perfect than the fifth for example, I have at first to accept this distinction in good faith. The fact is, as a musician of the 21st century with ears accustomed to a high level of dissonance, it's difficult to perceive these subtle dissonant-consonant relationships without some kind of explicit sign-posting. Only through practicing the exercises, by following the rules explicitly and painstakingly do I begin to feel the governing rules of consonance and dissonance of this style. Over a period of time I begin to feel more sensitive to certain intervals, to recognise them as performing certain functions within the musical texture. Fux's exercises help me to label and organise a diatonic musical fabric in a rich and meaningful way. 

         The results of this work of absorbing Fux's rules is therefore incredibly useful. For how often have I pursued the products of imagination with the feeling that I lack a map with which to find my way home; how often do I try to pursue ill-conceived 'rules' of harmony and voice-leading at the expense of all imagination? I am led through these exercises to experience for the first time, the underlying structural principles on which tonal music is built. Fux assigns functional properties to each interval which dictate the possible subsequent movements, either towards establishing a new dissonance, or resolving an existing dissonance towards consonance and resolution (7 to 6; 4 to 3 for example). To become aware of these rules is, for me, a significant advance over my previous approach or conceptualisation of dissonance. For previously I wandered around in ignorance, reacting to certain superficial properties which I attributed to particular 'chords' which my hands performed rather easily. I was aware that these chords had expressive potential, as they contain dissonant intervals which sound good to my ears (also because they are extracts or distillations of musical moments that are significant to me). However, even while playing these same chords, I was not able to fulfill the expressive potential that I perceived in them: I moved speculatively from one harmony to another, without understanding the principles on which I moved. I could not therefore control either the music or my actions satisfactorily, for the environment was one of continuous change and impermanence. 

This disorganised and unsatisfactory way of proceeding rarely improved, for I faced the same problems in the same way each time I improvised. Though I found temporary (harmonic/countrapuntal) solutions as I went along, I had no framework for understanding how I arrived at these solutions; nor could I perceive if indeed the solutions are good or bad, as the halting, unsatisfactory context in which they are performed contained no clear point of reference against which I could assess the success or failure of my methods.

By absorbing the functions which Fux gives to dissonant intervals, and the rules of progress from one interval to another, I began to understand the power of restraint. In my previous search for expression I tended to amass dissonance; through species counterpoint I learn to honour the resolution of a dissonance before establishing another. It required the exercise of conscious restraint, but this restraint resulted not only in a greater clarity of expression, I sensed too that there was more power, more organisation in the music. It was (for want of a better expression) better music, and certainly more interesting to listen to than my previous unstructured rambling!

It is for this reason that I belive progress in controlling dissonance is a conscious and careful process, which has a gradual effect over time on improvisation. So, to summarise:

1. Wrong notes certainly do exist, in the mind of the improviser and in the perception of the listener, unless the improviser adopts (as is possible) an anarchic approach of uncontrolled dissonance.

2. If, however, the improviser adopts a intuitive approach of controlling dissonance this is a chancy and inexact solution, for the reason that our ears are so attuned to dissonant intervals there is tendency for improvisers to amass dissonance in the search for greater expression.

One final thought: Fux’s rules represent a method for understanding the relationship between dissonance and consonance when using a systme of diatonic tonality. In this sense it serves as a means of control over a particular musical system. By absorbing his fundamental rules one is able to assign functions to certain musical properties and thus perceive the stimulus and the consequences of one’s actions in a stable and secure way; to make (creative and generative) decisions with certainty and, in this way gain control (through understanding) over contrapuntal textures, harmonic movement and so on towards larger sections of musical structure.

The use of this ‘means of control’ through the clear and explicit explanation of conceptual relationships illustrates the need for such mental tools whenever one improvises. That is, there must be a musical system (of constraints) and a means of controlling or manipulating these constraints is the improviser’s actions are to be effectively organised and mentally represented.

 

 

Further discussion on learning (2)

All about learning - conscious? unconscious? What's the difference? Does it matter?

Previously I quoted from Nick Ellis's (2011) chapter: Implicit and Explicit SLA and Their Interface, which is found in Sanz, C., & Leow, R. P. (Eds), Implicit and explicit language learning: conditions, processes, and knowledge in SLA and bilingualism. Washington, DC: Georgetown University Press.

Ellis is one of the few writers who not only distinguishes between the two types of learning process (implicit and explicit) but also tackles the issue of what role each process plays in learning. I began by quoting his distinctions:

  • Implicit learning is acquisition of knowledge about the underlying structure of a complex stimulus environment by a process that takes place naturally, simply, and without conscious operations (p.38)
  • Explicit learning is a more conscious operation, where the individual attends to particular aspects of the stimulus array and volunteers and tests hypotheses in a search for structure (p.38)

But then proposed my own assertion that implicit (unconscious) acquisition of the rules necessary for improvisation was not sufficient for improvisation. Why not? If such processes are sufficient for learning the complexities of our mother tongue (L1), why not for music? Such a comparison is, after all, valid if we believe that the ideal improviser is one who can 'speak' music in the same way that one speaks one's birth language. Would implicit (unconscious) acquisition of the rules then be sufficient if we learnt to improvise in the same way as we are taught to speak a birth language? To reconstruct these conditions would be an interesting experiment! It would mean:

  • thousands of hours of one-to-one instruction.
  • the use of music to communicate one's basic emotional and physical needs.
  • the acquisition of skill during the years of greatest neurological plasticity and cognitive growth, i.e., one would learn to express oneself in music in parallel to the processes of adapting to and understanding the surrounding social and physical environment.

It's relevant I think to note that L1 acquisition is not as implicit as it is often implied. Explicit instruction need not be restricted to the teaching of abstract rules and theory, but can also apply to forms of guidance in which the individual receives 'explicit' (i.e., clear) feedback, guidance and correction. Successful learning could hardly occur through meer passivity, as Ellis remarks:

Whatever the domain, learning the patterns, regularities or underlying concepts of a complex problem space or stimulus environment with explicit instruction, direction, and advances clues, heuristics, or organisers is always better than learning without any cues at all. Ellis (2011, p.38)

It's my personal conviction that musical improvisation could be learnt in the same fashion as L1 under the optimal conditions outlined above. Quite possibly, improvisers who learn the skill of improvisation also benefit from some of these conditions - in particular the learning of improvisatory skills alongside other musical skill formations, and the development of these skills during periods of general cognitive and personal development. In a sense, young improvisers in common with L1 learners are equipped with a mental white sheet, whereas classically-trained performers have to adapt their existing systems of musical perception to the new ones of improvisation, in common with L2 learners. In this, second case, the contexts of learning (of both music and language) are usually not comparable, calling for a different approach to learning, as Ellis (2011) describes: 

In contrast to the newborn infant, the L2 learner’s neocortex has already been tuned to the L1, incremental learning has slowly committed it to a particular configuration, and it has reached a point of entrenchment where the L2 is perceived through mechanisms optimised for the L1. Thus the L1 implicit representations conspire in a “learned attention” to language and in the automatised processing of the L2 in nonoptimal, L1-tuned ways. (Ellis, 2011, p.40)

Let's look more closely at the contrasting nature of these two types of knowledge could be applied to learning situations, using the insights of Ellis (2011), Hiebert (1986), Carpenter (1986) and others.

EXPLICIT LEARNING.

  • Ellis (2011, p.38). When the material to be learned is simple, ... and the critical features are salient, then learners gain from being told to adopt an explicit mode of learning where hypotheses are to be explicitly generated and tested and the model of the system updated accordingly. As a result they are also able to verbalise this knowledge and transfer to novel situations. 

  • Hiebert & Lefevre (1986, p.3-4). Conceptual knowledge is characterised most clearly as knowledge that is rich in relationships. It can be thought of as a connected web of knowledge, a network, in which the linking relationships are as prominent as the discrete pieces of information. Relationships pervade the individual facts and propositions so that all pieces of information are linked to some network. In fact, a unit of conceptual knowledge cannot be an isolated piece of information; by definition it is a part of conceptual knowledge only if the holder recognises its relationship to other pieces of information.

  • Ellis (2011, p.41). ... conscious processing is spread wide over the brain and unifies otherwise disparate areas in a synchronised focus of activity. 

  • Ellis (2011, p.41). Conscious activity affords much more scope for focused long-range association and influence than does implicit learning. It brings about a whole new level of potential associations.

  • Ellis (2011, p.42). Compared with the vast number of unconscious neural processes happening in any given moment, conscious capacity evidences a very narrow bottleneck. But the narrow limits of consciousness have a compensating advantage: consciousness seems to act as a gateway, creating access to essentially any part of the nervous system. ... It is a facility for accessing, disseminating, and exchanging information and for exercising global coordination and control: consciousness is the interface.

  • Ellis (2011, p.43) ...where L1 experience has tuned the learner’s attention elsewhere. ... Instruction is thus targeted at increasing the salience of commonly ignored features by first pointing them out and explaining their structure and, second, by providing meaningful input that contains many instances of the same grammatical meaning-form relationship. 

  • Ellis (2011, p.45).Consciousness creates access: its contents are broadcast throughout the brain to the vast array of our unconscious sources of knowledge, and by these means, consciousness is the interface.

IMPLICIT LEARNING.

  • Ellis (2011, p.35): The acquisition of L1 grammar is implicit and is extracted from experience of usage rather than from explicit rules - simple exposure to normal linguistic input suffices, and no explicit instruction is needed.

  • Ellis (2011, p.38): When the material to be learned is more randomly structured with a large number of variables and when the important relationships are not obvious, then explicit instructions only interfere and implicit mode of learning is more effective. This learning is instance-based but, with sufficient exemplars, an implicit understanding of the structure will be achieved.

  • Ellis (2011, p.39): Implicit learning collates the evidence of language, and the results of this tallying provide an optimal solution to the problem space of form-function mappings and their contextualised use, with representational systems modularising over thousands of hours on task.

  • Ellis, 2011, p.41). ... implicit learning occurs largely within modality and involves the priming, or chunking, of representations or routines within a module; it is the means of tuning our zombie agents, the menagerie of sepcialised sensorimotor processors ... that carry out routine operations in the absence of direct conscious sensation or control.
  • Ellis (2011, p.41). ... our implicit systems automatically process the input, allowing our conscious selves to concentrate on the meaning rather than the form. Implicit, habitual processes are highly adaptive in predictable situations. But the more novelty we encounter, the more the involvement of consciousness is needed for successful learning and problem solving

THE ROLE OF CONSCIOUS KNOWLEDGE IN LEARNING. One thing that emerges from these useful definitions is that the role of conscious knowledge (declarative knowledge, conscious awareness or conscious control) becomes more circumscribed. It's clear that, although Hiebert & Lefevre (1986) describe conscious knowledge as 'rich in relationships' and 'a connected web of knowledge' these connections are conceptually different to the kind of multitudinous connections calculated by implicit knowledge, which, over time - Ellis mentions 'thousands of hours on task' - furnishes us with instincts for what is appropriate or fitting in form and gesture. There is so much of skilled learning and performance which cannot be ascribed to consciousness: most importantly (i) I cannot consciously control or train my muscles, (ii) I cannot consciously control, generate or even fully understand the functioning of the music I improvise in real time, (iii) I cannot consciously understand the processes of automatisation which are so essential to fluency, (iv) the very nature of automatisation is that it reduces demands on consciousness and allows the performance of parallel rather than serial processes.

Yet, Ellis (2011) claims that implicit learning is itself inadequate for the learning of a new language, due to effects of interference arising from already existing skills in the native language. In what way do the existing language skills interfere with the acquisition of new skills? Can these effects be comparable to learning improvisation? 

Problems seem to arise when existing language skills prevent the learner from noticing what is important (salient) in the new material: '...where L1 experience has tuned the learner’s attention elsewhere' (Ellis, 2011, p.43). Why should this occur? Presumably because individuals try to understand or process the new language in similar ways to that in which they habitually process or understand their native tongue. This feature alone I think is directly replicable in the learning of improvisation (= the acquisition of novel generative skills) by musicians trained in performance (= possession of interpretative skills). The fact that music is not learned implicitly in an L1 fashion should not disrupt the comparison because, for trained musicians, the initial declarative methods of learning music will have been long transformed into automatic procedures. (In other words, the trained and fluent performer possesses an intact and specialised knowledge base sufficient for the task of performance, and it is this specialised and intact knowledge base which is brought to the task of learning improvisation.) Of course, when we have great familiarity with a task we don't need to analyse the techniques we use to perform it, not do we think of the constituent parts of the task, or in what way these parts are combined to make greater structures. The step-by-step learning of our native tongue, so embedded in communicative needs and direct feedback, is long since transformed into unconscious performance; likewise, the hard-won technical mastery of a musical instrument and the skills of meticulous score-reading and interpretation are second nature to the trained performer. How then should a musician with such a training look anew at a score and understand it as a model, say, for improvisation? As Ellis suggests, ',,, our zombie agents, the menagerie of specialised sensorimotor processors' which characterise implicit learning are inadequate for these initial stages of learning, powerful instruments though they are. The processes of unconscious learning must be steered in a certain direction by consciousness, given specific goals to head for, and have these goals reiterated through a clear framework of instruction: 'Instruction is thus targeted at increasing the salience of commonly ignored features by first pointing them out and explaining their structure and, second, by providing meaningful input that contains many instances of the same grammatical meaning-form relationship (Ellis, 2011, p.43). 

In this way, consciousness, though restricted as a tool for the actual performance of improvisation, and inadequate in terms of perceiving intrinsic matters of form in the material (whether music or language) through experience, can be seen to play an essential, leading role in the early stages of adult learning. How does it do this? Ellis suggests that there is no part of the nervous system to which consciousness does not have 'access'. Through this right of access then, consciousness can act as an interface between disparate, automatic systems, as a force which coordinates, sychronises and ultimately exercises 'global coordination and control' (p.42).

PROCESS PURE: Are the two types of knowledge really distinct? Much of the research giving rise to these distinctions comes from tests on amnesiac patients (Green & Shanks, 1993, p.304): 'Although amnesics are severely impaired on tasks of explicit memory, their performance is often within the normal range on implicit memory tasks. Normal controls, as well as amnesics, exhibit this independence of performance on implicit and explicit memory tasks.' The same authors also warn that, with regard to implicit memory (recall of unconsciously learnt material) '... individual memory tasks are not process-pure. In other words, a single task may involve the contribution of more than one type of memory process. A given task, for example, may involve any combination of conscious and automatic processes, such as those that are conceptually and data driven (Green & Shanks, 1993, p.314).

Hiebert & Lefevre (1986), also warn that '... the distinction provides a classification scheme into which all knowledge can or should be stored. Not all knowledge can be usefully described as either conceptual or procedural. Some knowledge, seems to be a little of both, and some knowledge seems to be neither' (p.3).

 

Further discussion on learning (1)

SOME OPENING THOUGHTS

When I improvise, what is it that I learn?

1. I will experience learning through implicit processes, i.e. unconsciously.

2. I will experience learning through explicit processes, i.e. reflectively, of which I am aware.

What do I then retain from these experiences for the future? In what way have I developed?

 

VARIABLES (Can also be discussed under the 'goals' of the improviser)

1. I have a certain knowledge and disposition which I bring to the task.

2. I also have certain goals, set by myself, or my teacher. These goals, in all probability will form the focus of my attention while improvising. They will play an important role in the nature of feedback.

As this discussion has immediately raised the subject of implicit/explicit learning, here are some further thoughts about the distinctive characteristics of these different yet complimentary processes:

Nick Ellis (2011) makes the following observations:

  • Implicit learning is acquisition of knowledge about the underlying structure of a complex stimulus environment by a process that takes place naturally, simply, and without conscious operations (p.38)
  • Explicit learning is a more conscious operation, where the individual attends to particular aspects of the stimulus array and volunteers and tests hypotheses in a search for structure (p.38)

1. A NEUROLOGICAL DISTINCTION BETWEEN IMPLICIT AND EXPLICIT LEARNING:

Paradis (1994): ‘... that the memory system that subserves the formal learning of a second language (declarative memory) is neurofunctionally and anatomically different from the one that subserves the first language or a foreign language acquired in conversational settings (procedural memory)’ (p.393).

Ellis (2011): ‘Explicit learning is supported by neural systems in the prefrontal cortex involved in attention, the conscious apperception of stimuli, and working memory; the consolidation of explicit memories involves neural systems in the hippocampus and related limbic structures. In contrast, implicit learning and memory are localised, among other places, in various areas of perceptual and motor cortex’ (p.38).

Ellis (2011, p.40). There are different types of knowledge of and about language, stored in different areas of the brain, and engendered by different types of educational experience.

The preceding quotations propose that it is explicit functions which govern attentional systems, and involves emotional states. Implicit learning in contrast runs in an automatic fashion through processes of perception and motor coordination. Paradis (1994) has already implied that the essential elements of performance are learned implicitly and performed implicitly, i.e.,

It appears that what has been acquired incidentally is stored implicitly and can only be evidenced through behaviour (performance) (Paradis, 1994, p.401).

But what exactly do we retain from the engagement of implicit mechanisms? Presumably information which organises the action of improvisation, for after all, we know that we cannot coordinate our actions directly, but only in relation to a perceived point (i.e., Hommel et al. 2001, p.876, who conclude that: '... coordinating actions and events is much easier in terms of distal than proximal representations'). However, is the information resulting from implicit mechanisms sufficient for the task of improvisation? In my experience I would say no. I am currently practising various tasks in baroque improvisation. Tasks, such as harmonising chorale melodies in a German baroque style, or accompanying a phrase of gregorian chant as a cantus firmus in the French style, demand a knowledge (perception) of a given body of music to the extent that I can generate new music in this style on demand.' When I place my hands on the keyboard I need more than motor memory to see me through at this stage of learning! For this particular genre of improvisation the musical material must be highly organised to be a convincing representation of the style. I find that at the outset of each practice session I have very little idea of what I should do. After a few attempts at harmonising, my developing knowledge is engaged and certain patterns emerge. However, in general, much testing of choices and pauses for thought (and also for writing solutions) characterises my improvising, and any attempt to perform in this style (at this stage) would, I think, be doomed to failure.

Expert knowledge on the other hand is generally characterised by fluency. According the words of performers theselves, and the assertions of researchers into memory and human performance, their improvisation is generally conducted unconsciously. What then is the nature of their knowledge? Is it merely quantitative, i.e., does the expert simply remember more connections, more characteristic chord progressions (or contrapuntal textures, or cadences etc.) with their corresponding hand movements than I do? Or is the knowledge more qualitative in essence, being linked to, or integrated within a host of other cognitive, emotional and perceptive mechanisms? Why can't the knowledge of the expert be simply explained to a novice? Why does the novice themselves have to experience over time the nature of the musical material to attain expert insights? Note that Johnson-Laird (2002) asserts:

‘... whenever a creative process calls for working memory, it is likely to depend on a multistage procedure’ (p.239).

So, what are the various stages of a 'multistage procedure', and why is it necessary that human skill learning should be such a lengthy and subtle process? Clearly, by now, we can describe expert performance as primarily unconscious; presumably then, it is a necessary condition of unconscious performance that it takes considerable time to develop. In addition, we can assume that the material which we are learning to manipulate is poses problems which are extremely complex to master and to understand; that mastery of this material cannot be understood singly, but is the result of a plurality of experience, occuring sequentially (one step in the knowledge base must be assimilated before another can be perceived) over time; and that the resulting codes or representations of knowledge used by expert improvisers refer to a multilayered, multi-experiential database which cannot (for example) be transfered to a third party. 

 

Surely it makes sense then to look to general theories of cognitive development for illumination of this process, as (I have no doubt) that expert skill, resulting as it does from such a considerable investment of time, attention and cognitive resources, should be described in terms of cognitive growth. 

One aspect of learning which seems to me overlooked by researchers, is the fact that the novice often cannot bring his (implicit) critical or evaluative faculties to bear on his own work. That some rules of music construction are unconsciously perceived seems to be beyond question, i.e., Tillmann & Bigand (2004):

By mere exposure to musical pieces, Western listeners acquire sensitivity to the regularities underlying tonal music. This implicitly acquired knowledge allows listeners to perceive subtle relations between musical events and permits musically untrained listeners to behave as musically trained listeners ... (p.211) 

So why, then must the pupil Josephus, in Joseph Fux's (1725) famous treatise Gradus ad Parnassum, be explicitly instructed in rules of counterpoint which exist primarily to avoid unpleasant acoustic effects:

'This mi against fa you have written in the progression from the sixth to the seventh bar by a skip of  an augmented fourth or tritone which is hard to sing and sounds bad, for which reason it is forbidden in strict counterpoint (Fux, transl. Mann, 1971, p.35)

If these rules reflect all that is natural and pleasing to the ear, why cannot Josephus apply his listening and critical faculties to his own work? Nor do I need to look towards literature for illustrations of this problem, as in my own case, when I am harmonising a chorale: why is it that I can be utterly convinced of a solution while writing (a certain chord or distribution of voices), only to find that later such a solution is clearly wrong and inadequate? Why is 'doing' so far removed from 'hearing'? Is it simply a question of attention, which is claimed by selected goals of learning at certain stages? While desperately trying to fulfill certain goals, do we simply fail to notice other problems? I notice that Johnson-Laird also comments on this phenomenon when he refers to Perkins' (1981) description of '... the fundamental paradox of creativity: people are better critics than creators.' And that artistic creativity '... is based on unconscious knowledge acquired only by laborious practice in creating' (p.422). However, this offers no more than a description of the problem, which it is implied will resolve itself over time. The alignment of skill with the unconscious perception of rules (proportion, beauty etc.) is presumably a goal of skill learning, but more information is needed to (i) explain why this misalignment should occur in the first place, and (ii) through what processes or stages of development are these two different areas brought into alignment?

 

 

Problems of learning in improvisation.

My main motivation for starting this PhD was in order to improvise better. Better? Well, to learn how to express myself musically, to escape from the frustration and confusion I felt when I tried to improvise. The motivation was in this sense primarily musical. But, I realised that, in order to attain this musical goal, I had to make sense of my experience in some way, which was why I turned to psychology as the field of investigation rather than seeking another strategy, such as practising harder. I was unclear what to practise, so to increase the hours spent at the keyboard would have been to simply augment my confusion.

I remember that at the outset I had the impression that I might, through a certain type of investigation, discover discrete answers to the problems I faced. For example, as one of my main problems was in choosing the next chord, I thought I might discover a technique for encoding sequences of chords which would solve this problem. I suppose that the difficulties of improvisation occurred to me then to be difficulties of processing information: experts were somehow able to process musical material efficiently, whereas novices could not. 18 months later, I have discovered a host of different perceptions on a vastly more complex subject than I originally perceived. I have experience in intuition, in which the next chord is felt rather than (theoretically) known; my inner ear has developed, and I have explored different modes of relating to improvisation - through movement and through an increasing awareness of my own intentions. More recently I started lessons with a pedagog and expert improviser which caused me to profoundly question my own musical knowledge, and the ways in which music is known, (encoded, memorised, represented among other terms) so that it can be used for improvisation. In some ways, I suspect that I may come full circle, and perhaps indeed discover (for myself and under certain musical conditions) a discrete form of (for example) harmonic representation, but certainly this knowledge comes as part of a gradual and multimodal increase in experience, and could never be written down as a formula and communicated to a novice in a form in which they would then understand what I now understand. 

What do we learn, when we learn something?

My more recent interests could be formulated by this question: When J.S.Bach copied the works of Nicolas de Grigny in 1713, what did he learn from this process? Copying music is quite a common technique for learning, but what is it that we learn from copying which we don't learn from playing or reading? Johann Sebastian, from accounts of his life was earnestly seeking a deeper understanding of the ways in which music could be constructed. Apparently he was quite indefatigable in this respect. What did he then discover from his copying? Do I discover the same features of the musical material, or reach the same understanding of these pieces as Johann Sebastian when I copy them? Presumably, as it's not hard to imagine someone engaging in the same task and discovering nothing at all, then we could propose that what we might gain from the process is a very personal process, and that, for example, our motivations for copying may influence our perception of the musical material, just as much as mechanisms such as priming (what we are currently interested in) or in fact, any of our individual cognitive and emotionl states might dictate a different form of understanding. Are there any generalities then that can be learnt or surmised?

A specific type of knowledge for improvisation?

The ways in which individuals seek and discover knowledge are questions crucial to the subject of improvisation, because the improviser is a seeker after knowledge, but a knowledge of a particular type. We need a type of knowledge about music that enables us to generate music at speed. If I say it is not therefore a reflective knowledge suitable for discussing and analysing musical scores, but an action knowledge suitable for adaptable, generative output in real time, then it becomes clear that the difference may not be the knowledge or facts of the matter which are conceptually different, but the form in which those facts are represented, encoded or understood in a way which facilitates improvising. In this sense we can talk of the reflective knowledge being interpreted for action (see Anderson, 1982). Unfortunately, this results in forms of knowledge which may be unavailable to conscious awareness, eluding analysis, as Paradis (1994) reveals, when he discusses the development of speaking capacity in individuals' linguistic development:

The aspect of practice which is useful, i.e., conducive to improvement of competence, is not the aspect(s) of which the speaker is aware. The aspect of which the speaker is aware does not get automatised, transferred or converted into what is stored in procedural memory - and what does get stored is not within the speaker’s awareness. Practice of utterances is what is directly conducive to improvement of implicit competence. Knowledge of the rules is not (Paradis, 1994, p.404-405).

Knowledge of facts, or knowledge of action?

So, through action - perhaps the action of writing/copying, or perhaps the action of improvising, aspects of knowledge become secretly encoded. And it is this secretly encoded knowledge, unavailable to consciousness which pertains directly to the development of the action itself. Does this mean we should dispense with factual learning all together?Apparently not, as Paradis (1994) in the same article also argues for a role of explicit awareness of rules in forming productions (automatic performance): 'It appears that what has been acquired incidentally is stored implicitly and can only be evidenced through behaviour (performance). On the other hand, some deliberately learned tasks seem to gradually become automatic through prolonged practice' (Paradis, 1994, p.401). So, consciously learnt facts do play a part in some aspects of learning! As this may appear rather confusing I will quickly summarise some of Paradis's findings, and then use these to illustrate the problems facing novice improvisers:

  1. that the implicit or unconscious memory systems used for fluent performance (i.e. first language production or expert improvisation) are 'neurofunctionally and anatomically different' (p.393) from those used for conscious recall of declarative information (i.e. formal study of second language, or theoretical knowledge about musical improvisation).
  2. that, as a result, individuals have two types or sources of knowledge, one which is available to consciousness (explicit), and one which is not available to consciousness (implicit).
  3. that, when we act fluently (i.e., speak or improvise) we rely on implicit knowledge for this fluency, In the words of Paradis (1994, p.400): '... in natural conversational settings, the production of the utterance remains non conscious.' Explicit knowledge then can only be used to slowly construct performance (a rather painful process for the performer and listener alike!), but explicit knowledge can also be used to reflect or check on the output of performance. However, explicit knowledge cannot control performance; as a knowledge type it is entirely unsuited for this task.
  4. that, because of this unsuitability, the introduction or intrusion of one type of explicit knowledge during performance causes interference in the processes of implicit knowledge. (The effects of this interference cause well-known phenomena of slowing down or pauses during performance.)

From this, it would seem that, while facts play an unavoidable part of adult learning, it is only through the repetition and automatisation of these facts that anything approaching natural (fluent) performance can result. As to the nature of performance itself, Paradis's distinctions are important, because they describe in a particularly clear way an aspect of improvisation which is often difficult to understand and to experience: namely, that to be fluent, improvisation has to be unconsciously executed; that, in essence, we must act in a way that we cannot know or describe. 

Implicit knowledge gained through action.

However, this distinction between declarative and procedural knowledge, the functions of which are so clearly described by Paradis, is not a symptom purely of fluent performance, nor is it a distinction that can be understood only in terms of improvisational output. Rather, the relationship between unconscious and conscious is dynamically present at all stages of learning; or perhaps it is better conceptualised as the role of the unconscious in all stages of learning. For example, if we return to my opening question: When J.S.Bach copied the works of Nicolas de Grigny in 1713, what did he learn from this process? we can now conjecture that, in addition to what Bach saw in the musical material, and recognised in theoretical, analytical, or musical terms, he also learnt something from the action of writing. And that whatever perceptions were gained through this action, yet more learning would have occurred when Bach attempted to generate new music from this knowledge, either by writing or improvising. For, through the actions themselves, an awareness of action is created, and '... when often repeated, end up leaving proprioceptive kinaesthetic traces' (p.401) and it is this 'proprioceptive data' that is 'stored in implicit memory'. As Paradis is at pains to point out, it is not the facts of  improvisation, or music which are rehearsed, but their production in performance (of those facts), and the processes by which we interpret and act on these facts are not known to us. 'The speaker (or improviser) is only conscious of the result, not of how it is obtained' (p.401). 

Not that this aspect of learning can account for J.S.Bach's musical knowledge! Far from it... either that that he learnt from De Grigny, or from listening to Dietrich Buxtehude or any other aspect of his musical formation. But, it does I think account for an often overlooked aspect of learning which is particularly useful for illuminating the learning of improvisation. 

 

Competence vs. Performance

Distinction between competence - performance in language and improvisation; presence of rules - are rules a good aid to learning? - contrasting the rules of theory and musical construction, vs. rules of production.

(image, Noam Chomsky)

One idea which constantly recurs in theories of language learning is the idea of competence - a person's knowledge or awareness of language, as against their performance - the actual production of words, sentences, phrases etc. Can this distinction, which has come under considerable criticism in the field of linguistics, be usefully applied to improvisation? By usefully, I mean can the concept of two differing forms of knowledge and ability help in pedgaogical or performance situations of improvisation?

My first thought is that one form of ability - the performance, can be much more easily assessed than the other - competence. (Competence, as a concept, was originally formed to describe the linguistic knowledge of native speakers, who seem to miraculously gain knowledge of a language within the first five years of development without recourse to formal learning of grammatical rules.) Performance, for example, is the evidence of a person's ability: when musicians improvise, then it is the performance of their knowledge that we hear, enjoy, criticise and process as listeners. Conversely, a person's competence - what they know about improvisation - is hidden, revealed only through conversation or through the act of improvising itself. Perhaps for this reason, research and teaching often focuses principally on the performance of individuals, perhaps comparing improvisations to compositions (reflecting on the artistic value), or by teaching students as if their faulty performance was entirely and consciously willed - instead of the outcome of incomplete or insufficiently connected knowledge. 

In this sense, Chomsky's distinction is valuable, because it is on the knowledge base (competence) of musicians that improvised performance is constructed, and it is upon the nature of this knowledge base that research and teaching should (in my opinion) focus.

One aspect of language learning which is I think directly linked to improvisation- at least the rather strict style of baroque improvisation which is my present study - is the question of rules. Before discussing the acquisition of rules, let's first define what they are. Some improvisers may reject the very idea of artistic creativity being rule based. My own difficulty with rules comes from the limitations of working memory, which argues against the conscious use of rules while improvising, i.e., "Don't give me another rule to remember, just let me express myself!". Yet, we must accept that the 'language' or chosen genre in which improvisers express themselves musically or in which they construct their artistic productions, must, to be coherent and recognisable to a listener, follow certain rules.

My recent lessons with Jürgen Essl in Stuttgart taught me that my knowledge of certain rules - we could call them characteristics of J.S.Bach's style of writing - was shaky to say the least. For example, when harmonising a chorale melody I did not perceive that each phrase should remain in one tonality, defined by the final cadence. To understand this allowed me to construct a harmonisation which was more true, expressive and recognisable within the idiom of J.S.Bach's style of harmonisation.

Ideally, we would improvise as we speak our native language. Certainly I have imagined ideal scenarios similar to L1 acquisition in which improvisation occurs as a natural communicative device between mentors and novices, who through many hours of exploration and one-to-one guidance allow natural learning abilities to develop a rich resourceful capacity for improvising directly linked to emotional intention. Yet, such idealised forms of learning to improvise are unlikely to occur when we are exposed to the learning process as adults (in which we are considerably more self-aware and conscious of the learning process); or when learning to improvise occurs as an addition to our existing musical training (in which case new knowledge and new ways of understanding music must be connected with existing forms, a source of much friction and conflict). In these situations some similarities can be observed between learning a second language and learning to improvise, in which case the insights of researchers into L2 acquisition can be usefully employed as guidance to improvisers.

Sharwood Smith (1994) describes the strange dichotomy between child and adult language learning thus:

First language acquisition happens in the natural world in the natural course of things as children try to communicate with the outside world. Older learners rarely experience this both naive and successful way of language learning because they are older and more worldly-wise. They have some idea what language is. They know it is a system with rules and principles that have somehow to be acquired. They worry about their failure to work that new system. They want to understand more about the nature of the task in the hope that this understanding will help them more. At the same time, the increase in awareness about the language being learned often seems not to be accompanied by a corresponding increase in proficiency, hence the question: do rules help?

From this account, it might be reasonably assumed that adults then somehow obstruct natural learning capacities? Perhaps, because of embarrassment or self-consciousness, adults are unwilling to experiment or play with language; after all if you don't try then you can't learn. On the other hand, Reber (1993) warns against the assumption of mysterious or innate powers of language competence in the child, reminding us how, during the first seven years of the child's life '... just how much learning takes place' (p.151). If, then, a child consciously practices language for 3 hours a day, '... we still find a very large number, 8,214 hours of practice in one's native language. Cut the estimate to 1 hour a day and we still end up with 2, 738' (p.152). This number he argues is consistent with research which asserts the amount of time needed to '... learn the structure of any rich, complex, abstract domain such as those that underlie natural sciences like physic, social sciences like psychology, or performing arts like music or dance. Numbers between 1000 and 5000 hours keep cropping up, no matter what the discipline' (p.152).

To return to Sharwood Smith's question: 'Do rules help?' it seems sensible to explore the different meanings given to rules by individuals in different learning contexts. 

For example, when Sanz & Leow (2011) conclude: 'Whatever the domain, learning the patterns, regularities or underlying concepts of a complex problem space or stimulus environment with explicit instruction, direction, and advances clues, heuristics, or organisers is always better than learning without any cues at all' (p.38); they refer to rules as cues - a substantially more friendly term than rules. Sharwood Smith (1994) also widens the terms of reference: 'The idea of a rule - whether it is called a rule, a routine, a plan or a process is immaterial...' (p.37), to illustrate the plain existence of rules as a means of structuring or organising experience. Perhaps the main problem with rules then is our usage of them; the danger being that in some learning environments, theory takes precedence over performance and rules '... may be couched in very abstract and technical terms such that they do not appear to offer much of immediate benefit to teacher or learner' (Sharwood Smith, 1994, p.34). Taken to further extremes, rules can be couched in coercive terms - a characteristic of much musical teaching of the 19th century conservatoire approach. In this sense, rules can be more than cues for structuring perception and understanding; instead they are conditions under which music-making is accepted or rejected. Such rules or 'Rules', if taken literally by the student, dominate the learning process in particular through their influence on attentional processes and feedback, as the student seeks naturally to align the ouput of improvisation with the idealised (imaginary) product of the Rules. 

One interesting aspect of Sharwood Smith's article (1994) is his elaboration of the different kind of knowledge needed for speaking, rather than listening or thinking about language: '...we may assume that the on-line production and reception processes during actual language performance are governed by principles requiring theoretical explanations of quite a different character to those used to explain competence' (p.37). This interested me because, in my experience, my competence (in the Chomskian sense of implicit or idealised knowledge) in improvisation greatly exceeds my improvised performance. For example, while driving to church I can plan and imagine the most marvellous improvisations, which, when I approach the organ seem to fade, and when I start actually improvising evaporate entirely, leaving little if any trace behind. Clearly, when I start to play, I am under the authority of a different set of constraints altogether, and these constraints or principles, as Sharwood Smith suggests, require a different theoretical explanation. (This, being a private or inner experience takes the investigation of improvisation in a new direction: if we are to discover or describe the consciousness of improvisation then a new framework is needed, a phenomenological account which could (i) describe this experience in embodied, first-person terms, (ii) explore research into proprioceptive memory and experience, and (iii) analyse ways in which musical material is inwardly represented.) 

For Sharwood Smith (1994), what follows from this distinction between the knowledge we bring to language use (or improvisation) and the knowledge we actually use while  speaking (or improvising), is that rules of construction once more assert themselves into the process. Improvising and speaking take place in real time, and thus, events have to be organised into a serial stream. Ideally, at expert level, this stream is fluent and is enacted without breaks and pauses, meaning that the individual has to have the material-to-be-improvised sequentially organised. While we may theoretically adopt a relaxed attitude towards construction (for example, it is often said in improvisation circles that mistakes don't matter and can be turned to creative advantage), the fact is that the improviser must '... determine what takes place first, what second, and what takes place simultaneously' (p.37); if they cannot determine this, for what ever reason, then faults emerge in their performance. This necessity for organisation can, as Sharwood Smith suggests, best de described as rules, and, whether applied to language or musical improvisation, such 'performance rules determine what happens millisecond by millisecond, second by second, etc.' (p.37).

Thus, performance itself is governed mercilessly by its own rules - rules of production. The acquisition of these rules, through experience and learning is the subject of other posts.

 

 

 

 

 

  • rules and creativity
  • attitude to rules
  • acquisition of rules - contrast between theory and performance.

References:

Sharwood Smith, M. A. (1994). The Unruly World of Language. In Ellis (Ed.), Implicit and explicit learning of languages (pp. 33-44). Bodmin: Academic press.

 

New lessons!

I recently started a new course of study with the composer, organist and improviser Jurgen Essl in Stuttgart.

This is a wonderful and salutary experience for me, having worked so much on my own, to bring my development as an improviser to the attention of someone with this level of experience and deep musical knowledge.

My first lesson was, from the point of view of my own proficiency (i.e., as a demonstration of skill), a disaster! Clearly I was unable to piece together even the simplest material in the Baroque style which I had been studying for so long.. From a psychological point of view, perhaps it was more successful: rather than feeling cast down by my miserablt show of abailities, I was able to focus on Jurgen's instructions which clearly formed a basis for further development. I think if this kind of experience had happened to me as a young man, I would have been very upset and humiliated, so clearly I have learnt something from my psychological investigations!

In general, I knew it was not going to be easy, but still I was surprised at the extent to which I was unable to produce a simple, fluent improvisation. In spite of my recent work improvising in baroque style at home, it seemed as if a vast gulf exists between what Chomsky (in a linguistic context) would refer to as my competence (mental awareness, implicit knowledge and ideals) and performance (what I actually produce). Let's consider then what I had been doing:

  • I had been piecing together lots of ideas in a baroque style, practising figurations, making sequences, and trying to understand the various relationships of consonance and dissonance. 
  • I had been studying works by Bach and Buxtehude, analysing some of the structural plans, writing out the rhythmic motifs.
  • I had been practising fugue constructions, using a variety of different fugal themes.

This practice had left me with an unfocused and generalised knowledge base.

While playing at home, I could sometimes achieve nice results - but these results were momentary, interspersed amongst a great deal of trial and error. Probably my biggest achievement was emotional, in the sense that I was not afraid to improvise: I could enjoy the process of trial and error, and could play around with all different types of musical construction. My biggest problem was that, when it came to a 'product' moment - when I actually had to come up with something coherent, then I didn't know what to do. Playing in church for example was an ordeal, because I had no resources or techniques with which to organise my material. I was left with a very complex mental state, of incoherent goals, and (corresponding) emotional reactions. Note, that the incoherence of my knowledge (by this I mean the procedural knowledge of 'what to do') left me with other problems to confront: I questioned my relationship with the listeners, the context in which I found myself, my attitude towards music, performance, etc.

Working with Jürgen taught me the following:

  • To focus unrelentingly on one aspect of the texture - in this case the harmonisation of a chorale theme.
  • Jürgen told me to decide on a harmonisation of the chorale melody, and to keep this harmonisation for further treatment (variations). 
  • From this harmonisation, came tasks such as putting the melody in the alto, the tenor and the bass.
  • To harmonise with three voices in the right hand, just the bass in the left; to harmonise then with two voices in the left (alto and tenor), with an ornamented melody in the right, and the bass in the pedals; also to harmonise with two voices in each hand.

Deciding on a harmony - i.e., to have a memory of that harmony was something I found I couldn't do as my approach to harmonising was too random. My goal (to this point) when harmonising was to find something that fitted and then to move on. This meant that the harmony I selected had no real function or relation to something which I could define. It was ad hoc, - to hand.. For me, this ad hoc-ness was quite an achievement, as it meant I had progressed from searching for the 'correct', theoretical version, to something improvised, something which allowed for margins of error. I had stopped searching for the version, and was now searching for version; I had emerged from a product state of mind, to a process approach.

Jürgen's advice seemed at first to be a return to the product view: I had to define my harmonisation, and then to write it down. Why do this? To define more clearly what the harmony is? Actually, I think (I write now from the perspective of a week's fairly intensive experience) the process of writing down is to understand more clearly what the role of a harmonisation is - perhaps in terms of its possibilities, or functions. Remember that before I was using a bricolage or ad hoc technique. Anthing could be used so long as it sounded good in that moment. Once I'd played my choice of harmony, it was gone, because that moment had gone - it was not connected to anything else in the improvisation.

Writing down is an iterative technique, as one returns from writing to playing, and repeats this process until a version is arrived at. It is also an elaborative technique, as to write is to see one's harmonisation more objectively, perhaps too to see it with the same mechanisms as one would view a printed score. No longer is the harmonisation a thing of air and fantasy, the possibilities of a moment; it is corporeal, and can be repeated verbatim. After writing, I also see visually some connections within the harmony which I didn't perceive when playing; from this I find that some problems can be viewed graphically and perhaps solved analytically. Writing down the harmony does not produce a solution, as often new 'solutions' arrive through playing, in which case writing takes a form of dictation: writing usually takes its cue from the playing, but it helps to refine and instruct the process of playing.

Learning can be a difficult process, because one's faulty or incomplete knowledge leads one's instincts and capacities for doing astray. It's hard to arrive at a solution and be told that it is not a good solution! But, in this disciplined form of improvising there are clear and practical reasons for doing certain things (i.e., ways of harmonising) in certain ways. But this is the subjetc of another post..

View older posts »

Comments

There are currently no blog comments.