Improvisation

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…

Read more

 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 …

Read more

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 fo…

Read more

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 langua…

Read more

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/

 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.

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?

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.

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/

 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.

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?

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.