Abstract
The learning of jazz improvisation is often treated as the incorporation of stylistic vocabulary and development of technical dexterity. Although this methodology is effective, considering other aspects of jazz improvisation can make the learning process a more holistic and less technical endeavour. My experience teaching improvisation has led me to formulate a method based on the cognitive process of improvisation conceptualized as a multi-dimensional model consisting of five improvisational cognitive skills: performance of material; creation of material; continuation of ideas; structural awareness of the improvised material; and temporal awareness of the improvised events. This model indicates that, during improvised performance, the player shifts the focus from one cognitive skill to another; this ability to establish links between skills is what I define as improvisational intelligence. The proposed method develops this linking ability through exercises complementary to the more common methods of jazz improvisation. In this article I present the multi-dimensional model for improvisational cognition drawing from the existing literature. After breaking down the five cognitive aspects of the model and explaining how they work together, I provide exercises for each cognitive aspect in isolation and in conjunction. This method can be taught to high school and undergraduate jazz students and, with some modifications, to non-jazz musicians seeking to develop improvisation skills.
Introduction
Learning how to improvise is often treated as a process of acquiring stylistic vocabulary and procedural skills; in this process learners assimilate features of a style, applying them on their instruments thus focusing on the performance area of improvisation (see Aebersold’s play along books [1971]; Bailey, 1992; Berliner, 1994; Kratus, 1995). While this approach can provide high school and undergraduate students with fluency in the idiom of jazz, I have noticed the necessity to complement this methodology by addressing other areas of the improvisational ‘brain’. While improvisational cognition has been extensively researched by Pressing (1984, 1987) and Sarath (1996), their theoretical models have not been used to design a pedagogical method for improvisation that goes beyond the acquisition of stylistic features and technical ability. Drawing from both Pressing’s and Sarath’s models of improvisational cognition, and from my experience as a professional improviser and improvisation teacher, I propose a pedagogical approach that teaches improvisers not only ‘what to play’ but, more importantly, ‘how to think’ in improvisational terms. The underlying principle of this method is that, by understanding the mechanics of the improvisational ‘brain’, we can devise exercises, strategies, and activities that can develop cognitive faculties essential to improvising.
Through the traditional methods of learning stylistic features, the learner ends up developing a sense of improvisational thinking; in other words, after copying enough examples of successful improvisation the learner, unconsciously, starts to think as an improviser. The model proposed in this article intends to increase the focus on the thinking aspect of improvisation to match the level of acquisition of stylistic vocabulary. For example, most of the exercises in jazz pedagogy equip the player with scales and arpeggios to play during an improvisation over certain harmonic progressions, but not many exercises are specifically devised to consciously create contrasting, related, or recurrent material. This conceptual approach to improvisation is at best talked about and addressed in improvisation classes, but rarely is it explored through exercises. In this article, I introduce a multi-dimensional model for the cognitive process of improvisation. After first explaining each dimension, I then provide exercises to develop these dimensions in isolation and in combination as well. I have applied this model teaching high school and undergraduate students, and working with professional musicians. Jazz pedagogues and music educators who teach improvisation will find these examples useful; moreover, these may be adapted to accommodate for the demands and needs of different levels of instruction.
The model
Cognition in music improvisation differs considerably from cognition in other musical activities. For example, commonly ascribed to music composition, Wallace’s creative process, which consists of problem investigation, incubation, unconscious work, illumination, and verification (see Csikszentmihalyi, 1996; Runco, 1994), does not apply to improvisation in that revision is not possible. Furthermore, not only does improvisational intelligence activate areas of cognition exclusive to it, but also it connects these areas in a unique manner. I describe the cognitive model of improvisation as a multi-dimensional process; that is, several simultaneously active layers that can be accessed at any time and in any order by the improviser. This is analogous to a computer desktop with several programs running; the individual chooses which window to open while the other programs are still running and readily available at the click of the mouse. Similarly, the improviser establishes links between these ‘programs’ over the course of an improvisation. This linkage between improvisational skills is what I define as improvisational intelligence: it is a schema that connects specific musical skills over the time continuum without the possibility to revise (for schema theory, see Bartel, 2002; Marshall, 1995).
From the literature on improvisational cognition I include five aspects to my model of improvisational intelligence: (1) skilled performance of material; (2) creation of material; (3) continuation ideas; (4) structural awareness of the improvised work; and (5) temporal awareness of the improvised events. However, the most important feature of this model is not these five aspects themselves, but the constant linking between them. Admittedly, these five aspects are not exhaustive or definitive, but they have been the foundation for my teaching improvisation to beginner and intermediate learners. We can conceptualize these five aspects as different ‘brains’ or ways of thinking. For instance, let us compare improvisation with cooking. If I am making dinner for my family I may use my ‘planning brain’ to go over what we had last night and what’s available for cooking today so that I come up with a good option. I will also use my ‘timing brain’ to decide how long I have to cook and when dinner needs to be ready. My ‘schedule brain’ will prioritize what needs to be done first, as well as what can be done simultaneously. I will certainly use my ‘artistic brain’ to prepare the seasoning, dressing, and presentation of the meal. Similar to improvisation, I will switch back and forth between one brain and the others during the entire process of cooking. Let us now have a look at the different ‘brains’ used in music improvisation.
Performance ‘brain’
The first ‘brain’, which controls the physical aspects of improvisational performance, controls a three-stage process: (1) perceptual coding of incoming data; (2) evaluation of possible responses and choice of response; and (3) execution and timing of chosen actions (Pressing 1984, p. 353). This process operates as a feedback loop. First, the sensory system codifies the incoming data, which is processed in the second stage, evaluation. In the third stage, the result of the evaluation comes out through the motor system. In other words, while improvising, the player’s sensory system monitors the actions that the motor system performs. This first aspect, the sensory-motor brain, serves as the platform for the other four intellectual components of the cognitive process as we shall see in the next paragraphs. In a way, the performance ‘brain’ is the operating system of the improvisational ‘computer’.
Creative ‘brain’
The second aspect of my multi-dimensional model is the creative ‘brain’, in charge of generating ideas. While many theories have provided insights on the creative process, I favour the model presented by Johnson-Laird (2002) because of its applicability to improvisation. According to his theory, there are three algorithms for the creative process: neo-Darwinian, neo-Lamarckian, and a combination of the two. In the neo-Darwinian algorithm, a first stage generates random options and a second stage filters the viable ones. For instance, in an improvised solo over modal harmony, several motifs may be presented before choosing the best one to explore further; it is not uncommon for an improviser to ‘search’ for an idea at the beginning of a solo. During this search, the player generates ‘random options’ and, after a few bars, it becomes clear which motifs are the most promising.
The second algorithm for the creative process, the neo-Lamarckiam, proposes that
if an individual has mastered a set of criteria that suffice to guarantee the viability of the results, then the generative stage will yield a small number of possibilities, all of which will meet the constraints of the genre. The algorithm will be efficient, and it will never produce hopeless results. (Johnson-Laird, 2002, p. 421)
Using this algorithm, we can think of early swing as an example. Given the criterion that a solo in this style must sound diatonic, any combination of chord tones will render satisfactory solutions.
In the third algorithm, ideas are generated in a neo-Lamarckian model, and out of all the satisfactory possibilities created upon the set criteria, the neo-Darwinian model selects the strongest options. For example, in a passage of modal harmony the improviser plays several chord-tone motifs; the most interesting is then transposed diatonically and chromatically. The two ‘brains’ discussed so far relate to each other in that the improviser’s mind selects material (creative ‘brain’) from what is technically feasible for the player (performance ‘brain’).
Building upon this model of creative algorithms and its relevance to improvisation, it is important to point out that the three creative algorithms are used at different points over the course of an improvisation. Therefore the three-algorithm model in its application to improvisation is a multi-dimensional model in itself; this means that, along with mastering each algorithm in isolation, it is essential for the improviser to skilfully switch between one algorithm and the others. This linking quality renders the three-algorithm model, perhaps, the most suitable model of the creative process for my multi-dimensional model of improvisational cognition.
Continuation ‘brain’
In regards to the third aspect of our model, namely the continuation of ideas, Pressing proposes two alternatives: associative generation and interrupt generation (1987, p. 21). In associative generation, the improviser develops an idea by exploring its qualities. For instance, when relating to a motif, the improviser elaborates variations extracting all the potential content from it. In interrupt generation, the improviser has exhausted the motif and looks for new material or directions. The contrasting purposes implied in each of these procedural alternatives, association or interruption, trigger different cognitive skills. For instance, associative generation requires less evaluation over solutions given that they refer to an already existing satisfactory idea, whereas interrupt generation triggers a more alert sense of ‘quality control’ over the new possible directions to ensure satisfactory results from the possibilities considered.
It is important to notice that the creative algorithms mentioned in the creative ‘brain’ refer to the creative moment of constructing ideas, whereas the continuation ‘brain’ manages the succession of ideas. This means that the creative algorithms apply when the improviser thinks ‘how shall I make a new motif?’, whereas at continuation points the improviser thinks ‘shall I continue with this idea, or is it time for something new?’. Also, it should be stressed that creation in this case does not necessarily precede continuation as the creative algorithms may be accessed in the middle of a piece, for instance at a continuation point after the improviser decides it is time for new material. In this case, the improviser’s mind, relying on the ‘performance brain’, links the second and third ‘brains’.
Structural ‘brain’
The fourth ‘brain’ of the compound model that I propose facilitates the organization of musical ideas relating past and future events. According to Sarath (1996), this can be done in two different ways. A motif can either relate to its immediate predecessor, or it can result from the accumulation of all passed events. The former applies to most forms of jazz improvisation whereas the latter is more relevant to improvisation in classical forms such as theme and variations (p. 7).
The skill to organize musical past and future events is what I call ‘structural thinking’ (Monk, 2010); that is, the intended continuation, development, abandonment, or retake of an idea for the sake of conveying structural sense in the overall conceptualization of a piece. In relation to the other ‘brains’, the improviser decides, for instance, that it is time to present new material (continuation ‘brain’); this new material needs to be contrasting in order to convey the idea of a new section in the piece (structural thinking). This contrasting section will be in a key distant from the current one to guarantee differentiation (neo-Lamarckian algorithm). As we see, each decision made is supported by several layers of thought which operate simultaneously rather than sequentially; in other words, the conceptualization of a new, contrasting section in a different key links the different brains simultaneously producing a single construct. If the player worked out one aspect at a time, it would be impossible to play in streamlined time.
Temporal ‘brain’
Through the course of an improvisation, decisions are made at certain points called ‘key events’ (Sarath, 1996). It is at these key events when most of the linking between brains takes place. So far we have seen that at these key events the improviser’s mind engages in creative algorithms, decides how to move forward with the improvisation, and evaluates possibilities to convey a sense of structure, all this while performing, either at moments of musical rest, or during the production of sound.
The last aspect of the model presented in this article is the temporal ‘brain’ which adds a layer of retrospection and/or prospection to the decisions made (Mendonca & Wallace, 2004). This means that decisions are made in regards to past events while speculating about future ones. Although this ‘brain’ seems very similar to the structural ‘brain’, it differs in that the structural brain conveys organization whereas the temporal brain helps the improviser look back and forward. Consequently key events contain a layer of both structural and temporal thinking. For instance, when reaching a climax, the improviser conceptualizes this point as the peak of the solo and whatever motif is used at that moment will be related to the immediate past events (structural layer); as the peak becomes closer, the improviser applies prospective thinking anticipating or sensing the proximity of the peak (temporal layer). Over the structural layer, the improviser thinks ‘I am going to reach the highest note in my register which will mark the climatic point in the overall structure’. Over the temporal layer, the improviser thinks ‘OK, here comes the downbeat of the climax . . . Now!’. Thus structural thinking uses the ‘how-to-put-together’ brain, whereas the temporal dimension uses the ‘timing’ brain.
Pedagogical applications
The multi-dimensional model for improvisational cognition proposed in this article intends to shift the focus of teaching improvisation. Following pre-existing models, the teaching of improvisation has traditionally intended to develop procedural skills in learners that could be applied within the conventions of a style. Considering the proposed cognitive model, the pedagogy of improvisation can also develop cognitive improvisational skills that could be applied in a wide range of contexts including the well-established jazz language, the free-improvisation language, and any non-idiomatic improvisation that may apply musical material from various sources such as the vast language of 20th century harmony. Mastering the cognitive skills discussed here will facilitate improvised material that reflects the logic and individuality of thought of each improviser, a strength that constitutes one of the most important features of the model. The content of improvisations may come from three sources: previously learnt material; recently learnt material; or material discovered through risk-taking during improvisation itself. In summary, the pedagogical model is not geared towards producing one specific type of material, but to produce relationships between these materials that reflect the individuality of thought of each improviser.
In the following section I will first provide some examples of how to ‘practise’ each brain in isolation; I will then provide some examples of how to ‘practise’ the linking between the ‘brains’. The exercises are very simple and should be taken as basic points of departure for the adventurous teacher and learner to expand in complexity. In order to keep the explanation of the exercises clear, we will assume that all exercises are performed in a 4/4 metre, at a medium tempo, and over a modal harmonic background which is one chord only. The material can then be expanded to functional harmony, non-functional triadic harmony, quartal harmony, atonal harmony, and so on.
The ‘five-brain’ model
The pedagogical applications of the first brain discussed earlier, the performance ‘brain’, are covered by the motor skills developed through practising scales, arpeggios, patterns, and licks. Since the teaching resources for these skills are vast, I will move on to the less explored areas.
Under the auspices of the creative ‘brain’, the improvisation learner practises applying each of the three creative algorithms becoming aware of their implications in an improvisational context. The following exercise is an example of how to nurture this skill. Over a modal harmony groove, a soloist improvises three different motifs. These must be short and simple enough to be replicated by the player, and they need to be separated by one bar of rest. After the three motifs, the improviser selects which one ‘survives’ and repeats it twice with slight modifications. After this, assuming an ensemble setting, another improviser may take the next turn or the same improviser starts again. Although the exercise may seem very simple at first, it forces the player to improvise simple ideas. More importantly, this exercise develops a more selective type of playing as opposed to random playing. Eventually the player learns to critically listen to his/her own material discriminating average ideas from promising ones.
A common feature in the improvisations of novice improvisers is their moving on to new material too quickly leaving strong ideas undeveloped. Therefore, the outcomes of the creative ‘brain’ need another ‘brain’ to take creative ideas further. Using the continuation ‘brain’, the following exercise can prepare the player to say more with less, extracting more material from one promising idea, as well as judging when it is time to move on to new ideas.
Over a modal harmonic background, the player improvises one bar and rests one bar. Over 16 bars (eight odd bar numbers of sound and eight even bar numbers of rest) the improvisation develops the motif created on bar one; by developing the motif, I mean exploring different variations and ornamentations such as rhythmic variation and the addition of passing tones. On bar 16, which is a rest bar, the improviser must decide if bar 17 will be the beginning of a new 16-bar section or a continuation of the current one. If a new section starts, then, the next 16 bars will develop a new idea created in bar 17; then, at bar 32, the improviser will evaluate again how to continue, and so on and so forth. Alternatively, on bar 16, the improviser may decide to continue with the idea from bar one for another 16 bars, in which case at bar 32 there will be another point to evaluate how the improvisation will continue. Through this process of constant evaluation, not only does the learner start to listen to his/her own ideas more carefully, judging the potential of these ideas, but also starts to connect past events to upcoming ones. This connectivity takes us to the structural ‘brain’, the next layer of the model to be exercised.
Any musical component that relates to structure such as phrase-length and overall form can be trained through the structural ‘brain’. To illustrate how structural organization can be practised, I provide the following exercise to develop directionality; that is, building material towards a climatic peak. Over a modal harmony groove the learner is assigned 32 bars to improvise. The solo must consist of four eight-bar phrases. The material for each phrase may or may not be the same; that will be determined by the continuation ‘brain’. However, the density of each phrase must increase as the solo progresses. By density we understand how ‘busy’ the music is. For this exercise, we establish that the first eight bars will be quite sparse, with long rests and long-value notes such as half notes. Following the first phrase, in the second eight bars, the improviser will activate the rhythms using less rest and more eighth notes. Then, in the third eight bars, the improviser will incorporate 16th notes, after which, in the last eight bars, the improviser will be as busy as possible, perhaps repeating a busy riff on straight 16th notes. Reaching the peak on the last eight bars can also be supported by moving upwards in the register. The first eight bars can be played in the lower register, and on each subsequent eight-bar phrase the material moves up in the register, thus saving the highest notes for the peak. In this way, the structural ‘brain’ manages two musical aspects that support structure: density and register.
As the learner becomes fluent in this exercise, the length of phrases that use varying levels of density can be determined by layering the structural and the continuation ‘brains’. In this case, the structural ‘brain’ manages the density whereas the continuation ‘brain’ determines when it is appropriate to increase or decrease the level of density. For example, the player improvises five bars with low density on half-notes followed by one bar of rest to punctuate this first phrase. At that point, the continuation ‘brain’ questions ‘Shall I continue with low level density or is it time to start increasing the density?’. To solve this improvisational ‘problem’, the structural ‘brain’ decides to go for contrast; therefore, the creative ‘brain’ uses the neo-Lamarckian algorithm which indicates that any motifs constructed on 16th notes will propose a satisfactory solution.
The fifth component of the proposed multi-dimensional model is the temporal ‘brain’. To develop this ‘brain’ I propose an exercise that applies temporal awareness. Before starting the exercise, the player determines how long the improvisation is going to be; the length of the exercise must be established in minutes or seconds as opposed to amount of bars. For this exercise, it is essential that the improviser records his/her own work to check the accuracy of the timing and the progress of the improvisation throughout the predetermined time.
As mentioned earlier, the most distinctive feature of this multi-dimensional model is the linking between ‘brains’. This linking can also be reinforced through exercises, and I present two ways to do it. The more obvious approach is to practise the different ‘brains’ consecutively through a string of exercises that change every 32 bars. For example, over a modal background groove the improviser practises diatonic arpeggios (performance ‘brain’); without stopping, in the second 32 bars, the improviser practises the creative algorithm exercise; after that, over the third 32 bars, the improviser practises continuation points. This sequence continues until all five aspects of the model have been practised. To take this approach further, the next step is to reduce the number of bars for each exercise to 16, eight, four, two and one, thus increasing the speed of shifting from one ‘brain’ to another.
The other way to practise linking is by layering the ‘brains’. For example, the improviser can practise quartal arpeggios on the tonic chord (performance ‘brain’) with the intention to convey an arch form from quiet/sparse to loud/busy and back to quiet/sparse (structural ‘brain’). To take this exercise further, the improviser can incorporate arpeggios other than the tonic based on diatonic or chromatic roots. The creative ‘brain’ will then determine how far from the tonic the roots will be, and the structural and temporal ‘brains’ will indicate when it is time to deviate from the key centre.
Here is another example of layering, in this case, the temporal and the structural ‘brains’. The player records an improvisation for a predetermined amount of time (temporal ‘brain’) playing freely over a model harmony background. Once the improvisation is finished, the improviser listens to his/her work, evaluating if the management of some structural elements such as phrase length, dynamic contrast, and rhythmic density was effective (structural ‘brain’). The possible combinations of layering ‘brains’ and the strategies to develop these combinations are endless; teachers and learners should be proactive in creating their own exercises.
Conclusion
The purpose of the model presented in this article is to help music educators think of the learning of improvisation as a more holistic process that embraces cognitive aspects other than technical and stylistic vocabulary acquisition. The model featured in this article is not exhaustive, and many more dimensions could be added to the five presented here. It is also essential that the improvisation teacher as well as the improvisation learner devise their own exercises to address specific areas of concern.
From a pedagogical perspective, this model intends to develop individuality in the improviser. This individuality will become self-evident through the decisions made, the links established between ‘brains’, and the priorities given to the different aspects of an improvisation. Consequently this model can also be regarded as a template upon which improvisation learners can construct their own model.
I hope that exploring how the improvisational mind works motivates more music learners and teachers to experience jazz improvisation from a more holistic perspective. Analysing the process of improvisation from a cognitive perspective focuses on how our brain processes information as opposed to developing technical dexterity or musicianship. This model affirms that, if you have a brain, you can improvise.