Abstract
The growing interest in musical improvisation is exemplified by the body of literatures evidencing the positive impacts of improvisation learning on the musical apprentice’s aptitudes and the increasing presence of improvisation in Western classical concert halls and competitions. However, high-level Western classical music improvisers’ thinking processes are not yet thoroughly documented. As a result of this gap, our research addresses the following question: What strategies are implemented in the course of performance by Western classical music improvisers? To answer this question, semistructured interviews were conducted with five internationally recognized Western classical music expert improvisers. Each participant improvised, and immediately afterward, a retrospective verbal protocol with subjective aided recall data collection strategy was used to elicit the improvising musician’s strategies. After transcription, the interviews were coded and analyzed using NVivo 10 software, with a mixed (i.e., combining inductive and deductive coding) category approach. Our data revealed 46 improvisation strategies that were subsequently organized into five categories: preplanning, conceptual, structural, atmospheric and stylistic, and real time. Pedagogical implications arising from these findings are that (a) learners should be guided toward implementing various strategies and (b) the capacity to switch from one strategy to another according to the circumstances should be promoted.
In many musical cultures, including diverse non-Western cultures, the discrete specialization of performer, composer, and improviser does not exist; rather, every performing musician possesses a relative competence in assuming these three roles (Berkowitz, 2009). Such versatility was also frequent among Western classical musicians until the middle of the 19th century (Moore, 1992; Scott, 2007). Indeed, improvisation has been central to Western classical music for centuries; most Renaissance, Baroque, Classical, and Early Romantic musicians possessed a relative degree of ability in improvisation, composition, and performing from a musical score (Moore, 1992). However, from the second half of the 19th century, improvisation gradually vanished from Western classical music as musicians’ roles were increasingly separated and specialized (Berkowitz, 2009; Manderson, 2010; Moore, 1992). Various causes might explain this phenomenon, such as a greater emphasis placed on the precision of the execution of notated music and the increasing technical difficulty of the repertoire (Sloboda, 2004) together with the audience’s growing expectations in terms of “perfection,” expectations that are in turn fed by growing standards and technological improvement in the recording industry (Philip, 2004).
However, musical improvisation is increasingly evidenced in Western classical music performance (Després & Dubé, 2015; Dubé & Després, 2012). Simultaneously, throughout the past three decades, the body of theoretical and empirical research on musical improvisation has grown significantly (Solis & Nettl, 2009). A substantial body of research demonstrates the positive effects of improvisation learning on various musical and personal variables (e.g., creative thinking or motivation) and consequently advocates the (re)introduction of improvisation in Western classical music pedagogy (e.g., Azzara, 1992; Dos Santos & Del Ben, 2004; Kenny & Gellrich, 2002; Koutsoupidou & Hargreaves, 2009; McPherson, 1993; Montano, 1983; Whitman, 2001; Wilson, 1970; for an analysis of this literature, see Dubé & Després, 2012). Despite this support, improvisation is still rarely part of the Western classical curriculum in postsecondary music teaching institutions (McPherson, Bailey, & Sinclair, 1997). Why is this the case when empirical research has clearly shown that improvisation learning promotes the “completeness” or professionalism of the musician?
First, most Western classical music teachers (except perhaps those who learned through “active” approaches, such as Orff, Dalcroze, or Willems) have never been taught to improvise and therefore generally feel ill equipped to teach improvisation to their students (Ward-Steinman, 2007). As teachers tend to adopt pedagogical practices modeled on the way they learned (Dunn & Dunn, 1979), this cyclic process is hardly reversible. Moreover, the relative absence of improvisation in classical music training could be caused by a lack of empirical knowledge based on the actual practice of experts in the domain (Dubé & Després, 2012). Effectively, the cognitive processes involved in Western classical music improvisation are still relatively poorly documented (Berkowitz, 2010). While cognitive psychologists’ interest in researching musical improvisation has grown in the past few decades, most empirical studies are done in the context of jazz. By comparison, Western classical music receives significantly less attention (Solis & Nettl, 2009).
Rationale
Among the few research studies available in the domain of the generative process of expert improvisers in Western classical music are the case studies conducted by Bailey (1993), Berkowitz (2009), Johansson (2008), and Chamblee (2008). Bailey (1993) identified competencies such as the knowledge of harmony and stylistics norms to be central to improvise in various contexts and the ability to transform musical ideas into sounds in real time to be of uttermost importance for improvisers. In addition, within the constraints imposed by stylistic referents and with the help of information stored in his knowledge base, the musicians can use imitation and contrapuntal strategies during their improvisations or continuo realizations.
Aiming for a better understanding of the learning and production process of the improvisation of the classical period, Berkowitz (2009) analyzed nine pedagogical treatises of the mid-18th and early 19th centuries. He concluded that using a knowledge base of musical formulas is a key strategy for improvising in Western classical music. As the improvising musician cannot always use motives stored in long-term memory in their original form (for aesthetic, rhythmic, and tonal reasons), he or she has to develop permutation strategies such as transposition, variation, and recombination (Berkowitz, 2009). These strategies allow the improviser to adapt, on the spur of the moment, the motives drawn from his or her knowledge base to a specific musical context. Berkowitz also conducted interviews with Robert Levin and Malcolm Bilson, two expert musicians who specialize in the performance practice of the classical period who improvise in concert. Berkowitz identified two dialectic states of consciousness that can be experienced by the improvising musician: creator and witness. While the creator state is characterized by the intellect that “speaks,” the witness state is defined by the body that acts. The musician, while in a creator state, develops his or her improvisation consciously and deliberately, planning and evaluating his or her improvisation by using declarative knowledge. On the contrary, in a witness state, the improviser becomes a spectator of his or her own actions and develops the improvisation involuntarily through implicit procedural knowledge. This witness state could be analogous to the flow state documented by Csikszentmihalyi (1997).
In her doctoral dissertation, Chamblee (2008) compared organ and gospel improvisational processes. Her main finding was the identification of a circular improvisational process in which the declarative knowledge (“know-what”) and the procedural knowledge (“know-how”) of the musician interacted to generate “plan-based extemporizations.” While the improvisation is played, listeners evaluate it and accordingly provide feedback to the musician. This feedback in turn is integrated into the musician’s knowledge base, thus influencing future improvisations. The key point here is that the process of improvisation is described as something broader than the mere cognitive mechanisms of the musician: It encompasses the public’s reactions and their contribution to shaping the musician’s further improvisations.
Johansson (2008, 2011) studied the definition of improvisation in relation to the generative process of organists. Her findings suggested that organists conceive interpretation and improvisation not as diametrically opposed practices but as intertwined musical activities that form a continuum. Johansson (2008) also concluded that the context in which improvisation takes place has an important effect on the musician’s learning and discursive practices. For instance, in a liturgical situation, organists tend to conceive improvisation as a communication medium, while in a concert setting, improvisation acts mainly as an individual means of expression; therefore, both contexts call on different generative processes from the musician.
In sum, the body of research about Western art music improvisation’s generative processes revealed various strategies that could be implemented by the expert/professional musician in the course of a performance. A more comprehensive view of the strategies that expert improvising musicians implement can be gained by considering insights from recent research in other music genres, such as jazz and “free” improvisation.
Improvisational Thinking in Jazz and Free Improvisation Research
Mendonça and Wallace (2004) studied the cognitive processes of improvising jazz duets in four improvisational conditions: (a) familiar and simple, (b) unfamiliar and simple, (c) difficult and familiar, and (d) free. After improvising, musicians listened back to their performance and described their mental activity following the retrospective verbal protocols method. The analytical framework used for data encoding was closed. Only the data for conditions (a) and (d) were analyzed. The results of this research show that the conditions of performance (difficulty and familiarity) do not significantly influence the strategies used by duets of jazz musicians to develop their improvisation. This conclusion would probably not be corroborated by Csikszentmihalyi (1997), who, studying highly creative individuals in various domains, documented a flow state that can only be experienced when a person realizes a task of a difficulty level that matches the limits of his or her abilities. Also, from a methodological point of view, it is our belief that the analysis grid of this research would have benefited from being mixed (or even open) rather than closed because of the exploratory context of the study and the complexity of the phenomenon under study.
Reporting on the improvisational thinking of jazz expert improvisers, Norgaard (2008, 2011) collected individual musicians’ retrospective verbal protocols immediately after the execution of a standardized improvisation task. Norgaard’s main finding was the identification of two “ongoing processes” and four strategies used by jazz experts during improvisations. The two ongoing processes identified by Norgaard were (a) the sketch planning of what is going to be played and (b) the monitoring and evaluation of what has actually been played. The four strategies were (a) the idea bank, the use of pre-memorized musical material; (b) the harmonic priority, an approach based on chord tones; (c) the melodic priority, an approach based on melodic contour; and (d) the repetition of material previously played in the solo, either in an identical or a permuted fashion.
In the performance context of free improvisation, Menezes (2010) investigated musicians’ creative and communicative processes. As part of his research, three musicians produced a collective free improvisation before an audience. Menezes analyzed two main sources of data: first, through spectral analysis, the musical characteristics of the improvisations produced and, second, the retrospective verbal protocols of the musicians (collected after the improvisation). The researcher identified eight strategies that can be used to develop musical material during improvisation: (a) the act of listening and reacting, (b) the construction of a mental database of pre-composed musical objects, (c) the use of involuntary motor habits, (d) the necessity of the use of simple musical structures that may be easily called to working memory, (e) the use of reiteration as a process of creating tension and musical personality, (f) the conversational metaphor, (g) viewing errors as a motor of creativity, and (h) the importance of tone, tone color, and “timbral” aspects of improvisation (Menezes, 2010). Each of the three previous research studies used the verbal protocol data collection strategy to provide a better understanding of the improvisational thinking of high-level musicians.
Summary of Relevant Empirical Literatures
In summary, research studies on the generative process of improvisation in Western classical music are limited in number, have focused only on keyboard instruments, and have not benefited from the retrospective verbal protocol for privileged access to a musician’s mental activity, a protocol that has been shown to provide valuable results in jazz and free improvisation research. Therefore, we assert that the strategies used during performance by expert improvisers of Western classical music are not yet satisfactorily documented. Reviewing the research studies of Mendonça and Wallace (2004), Menezes (2010), and Norgaard (2008, 2011) demonstrated the retrospective verbal protocol to be a valuable tool for studying the strategies and cognitive processes involved in jazz and free improvisation. In effect, interviewing the musicians about a specific improvisation they had just played would significantly reduce biases such as memory flaws and thus enhance the validity of data collection. Using the retrospective verbal protocol method to study the strategies implemented by expert Western classical music improvisers could provide a better understanding of their thought processes, which in turn would carry significant pedagogical potential: The primary goal of improvisation instruction should be to structure learning activities in which students at all levels can experience ways of thinking that resemble those of artist-level improvisers. Accomplishing this goal requires that we more closely examine and more clearly describe not only what experts do but also how they think about what they do. (Norgaard, 2011, p. 124)
Based on the findings of our literature review, we formulated the following research question: What strategies are implemented by Western classical music expert improvisers during their performance?
Conceptual Framework
Two concepts were central to this research: expertise and improvisational strategies. We describe how these concepts have been defined in the context of this study, grounding our rationale in an analysis of the background literature.
Introducing experts and expertise
Studying experts’ behavior allows a better understanding of the limits of human capacities and the conditions of optimal learning (Ericsson & Charness, 1994). An expert can be defined as the distinguished or brilliant journeyman, highly regarded by peers, whose judgments are uncommonly accurate and reliable, whose performance shows consummate skill and economy of effort, and who can deal effectively with certain types of rare or “tough” cases. Also, an expert is one who has special skills or knowledge derived from extensive experience with subdomains. (Chi, 2006, p. 22)
Experts differentiate themselves from novices by the size and flexibility of their knowledge base, which allows them to anticipate problems and react with greater accuracy (Borko & Livingston, 1989) because they are able to implement strategies that are more adapted to the problems they face (Lemaire & Siegler, 1995). Similar observations are true in the specific domain of musical improvisation: For example, in jazz, a correlation can be established between successful improvisation and the size of the musician’s knowledge base, which consists mainly of overlearned patterns (Thompson & Lehmann, 2004). The criteria that have been used to acknowledge expertise in the specific context of Western classical music improvisation are described in the Participants section.
Introducing the role and definition of strategy
In his empirical work on high-level expert jazz improvisers, Norgaard (2011) distinguished two different types of improvisational thinking: ongoing processes and generative strategies. Ongoing processes “represent thinking that appears to be ongoing throughout the creation of each improvisation” (p. 118), but they do “not describe how musical material is initially conceived” (p. 118). As for generative strategies, they “describe ways of creating improvised material—making choices about which notes to play” (p. 118). Knowing how improvising musicians choose the actual notes to play is indeed a subject of great interest. However, many other musical decisions have to be made by the musician in situ, namely, in the formal, rhythmic, timbral, and textural dimensions. To include the strategic thinking related to these dimensions in our analysis, we more broadly defined strategy as a sequence of cognitive processes undertaken by the musician to ideate, evaluate, select or realize one or several musical aspects (form, harmony, notes, rhythm, timbre, etc.) of his or her improvisation.
While no single piece of previous research offered a thorough taxonomy of improvisational strategies, previous relevant literature was consulted to identify the generative strategies implemented by improvising musicians in the course of performance. Our initial list included 19 strategies drawn from empirical works on classical (Berkowitz, 2009, 2010; Johansson, 2008, 2011), jazz (Berliner, 1994; Hargreaves, Cork, & Setton, 1991; Norgaard, 2008, 2011), free (Menezes, 2010), and mixed genres (Bailey, 1993; Chamblee, 2008; Csikszentmihalyi & Rich, 1997) dealing with musical improvisation as well as theoretical studies on improvisation (Ashley, 2009; Clarke, 1988; Kenny & Gellrich, 2002; Pressing, 1988). The strategies we identified are listed in Table 1.
Initial Data Coding Scheme.
For the purpose of this research, the terms knowledge base and idea bank have been differentiated: The former refers to all the knowledge of the musician—including the idea bank—and the latter applies specifically to prelearned musical motives (which are also called formulas or licks in the literature).
This conceptual framework involving expertise and improvisational strategies guided the methodological design of our research. The list of improvisational strategies documented in the previous literature was used as the initial data coding scheme (see Data Analysis).
Method
We chose to structure our research around the retrospective verbal protocol data collection strategy. This decision was made based on the fact that previous studies in jazz and free improvisation implementing this method successfully identified expert musicians’ performance strategies. While this methodological choice has proved to be of great value in other genres, no research so far has used it to study Western classical music improvisation. The retrospective verbal protocol comprises two phases: (a) A task is completed by the participant, and (b) when the task is over, the participant verbalizes the thinking processes he or she implemented while performing the task (Lemaire & Fabre, 2005). Recent studies tend to show that the retrospective verbal protocol compared to the concurrent protocol is less reactive and provides more comprehensive verbalizations (van den Haak, De Jong, & Schellens, 2003). The validity of the retrospective verbal protocol can be enhanced with an aided (or cued) recall, namely, a stimulus used to “re-immerse” the participant in the original event (Omodei, McLennan, & Wearing, 2005), such as an audiovisual (A/V) recording of the participant while he or she was doing the task under study (Hoc & Leplat, 1983).
Building on Norgaard’s (2008) implementation of the retrospective verbal protocol with A/V aided recall and aiming to improve its validity, we reviewed recent studies on experts’ behavior in the domain of sports (Mackenzie & Kerr, 2012; Sparkes & Smith, 2012) and decision making (Omodei et al., 2005). We optimized Norgaard’s design, using a subjective perspective (Rix & Biache, 2004) for the A/V aided recall to increase the validity of our data and tackle more precisely the strategies implemented by expert Western classical music improvisers. A mini-camera was placed on the head of the musician to videotape, as closely as possible, his or her first-person point of view while improvising. Given the scarcity of empirical research on Western classical improvisation generation and pursuing our research question, we adopted an exploratory perspective and implemented a case study research design (Yin, 2009) to gain in-depth insight into our research topic.
Participants
We selected five expert musicians using the following criteria: The participant (a) is recognized by peers as an expert improviser, 1 (b) completed high-level Western classical music training (bachelor or conservatory equivalent) and/or had taught in a recognized Western classical music institution, (c) improvised on at least one professional album, and (d) improvised in nationally or internationally recognized concert halls. A presentation of the research was sent to 16 potential participants by email, with a positive response rate of 31%. Aiming for a unified sample, we contacted only musicians who were accustomed to improvising in a solo setting. A questionnaire was used to gather the participants’ demographic information and improvisational experience (summarized in Table 2). Pseudonyms have been used to ensure participants’ anonymity.
Participants’ Main Instruments, Gender, and Fulfilment of Expertise Criteria.
The purpose of the study was to acquire an exploratory perspective on Western classical improvisers’ strategies encompassing various instruments as most previous research in this field of study had focused on a single instrument. Given the exploratory nature of our study and the scarcity of Western classical improvisers that could fulfill our selection criteria, we omitted other factors that could have a negative influence on our sampling size, such as gender-related participant selection.
Procedure
An individual meeting lasting approximately two hours was scheduled with every participant. At the beginning of the meeting, the researcher adjusted a wide-angle mini-video camera
2
on the head of the participant to capture, as closely as possible, his or her point of view while performing (i.e., his or her in situ subjective experience). Then, the researcher invited the participant to improvise, using the following prompt: “Improvise in the language of your choice, as you would in concert, ending improvisation when you deem appropriate.” The improvisation was A/V recorded. The improvisations ranged in length between 3.10 minutes and 10.50 minutes; the average duration was 7.34 minutes. Right after the improvisation, the researcher gave the participant the following instructions: Listening back to your improvisation, try to narrate what was going on in your head while you played. I am very interested in hearing anything you can tell me. Try to answer questions such as “Where did that idea came from?” or “What was I thinking at this specific moment?” At any moment, you can pause playback, using the spacebar, so that you can verbalize your ideas. I might also pause the recording at certain moments.
These instructions were inspired by Norgaard’s (2011) prompt.
The researcher and the participant then watched the A/V recording of the improvisation together, and an open-ended, semi-directive interview took place. Both the researcher and the participant could pause playback at any time to discuss a specific section or event of the improvisation. During the interview, the researcher aimed to bring to light the strategies implemented by the improvising musician. This interview was both audio and video recorded. After completion, the interview was transcribed using Transcriptions 3 (Version 0.8.1; Haselberger, 2008) freeware.
Before data collection occurred, we conducted a pilot with one participant to ensure the appropriateness of our data collection method and the internal validity of the data. While the technological design was shown to be fully functional, we made minor adaptations to the interview method after our pilot study, changing the interview guide format from a table of textual questions to a lighter checklist (including the textual prompts for the improvisation and the beginning of the interview and questions such as “What is going on in your head at that moment?” and “How did you choose this idea?”), leaving more space and time for “full” silences 4 (Duchesne, 2000) and avoiding yes/no questions.
Data Analysis
All statements related to the musicians’ improvisational strategies were retained for in-depth analysis. We applied a thematic content analysis (Paillé & Mucchielli, 2012) to the retrospective verbal protocol responses based on the potential of this approach to provide “thick description of the data set” and “generate unanticipated insights” (Braun & Clarke, 2006, p. 97). The researcher coded the interview transcriptions with QRS International NVivo 10 software and retained a mixed model for coding (L’Écuyer, 1987). The mixed model combines the use of preexisting categories with the possibility of adding or removing categories during the analysis process. The preexisting categories retained to initiate the coding process were those outlined in Table 1. This preliminary coding scheme was used to begin interview coding, and new categories were created throughout the process in response to the themes addressed by the participants. The researcher formulated operational definitions for every new category created and refined the categorization throughout the coding process for succinctness and differentiation.
After completing the primary coding, interrater reliability (IRR) was verified by a senior researcher (PhD) on 10% of the data. Afterward, we discussed any divergent results and reworked operational definitions so that every category with an initial Kappa coefficient below 0.80 was adjusted to above 0.80, which is often perceived as the threshold point between “substantial” and “almost perfect agreement” (Landis & Koch, 1977). The contributions of this process are clearer category definitions and second-person feedback input to the analysis process. This allowed us to generate a final coding scheme (see Table 3), which the main researcher used to recode all the data. A coresearcher then validated the categorization for the entire data set.
Improvisation General and Specific Strategies.
Novel strategies (i.e., those that were not included in the preliminary coding scheme).
Results
The analysis of our corpus revealed a total of 46 strategies through an iterative coding-defining-synthesizing process. After analysis, these strategies were divided into two levels of categories: general and specific strategies, according to their function in the course of improvisation. General strategies—labeled with a noun—are broad categories that represent the point of focus of attention of the musician at any given moment of the improvisation. The strategies were labeled as general when the participants described what they were thinking about at any given moment of the improvisation without mentioning their specific aim or identifying the musical product linked to their mental process. General strategies were most widely used by the participants and were explicated using less extensive verbalization than specific strategies. Each general strategy could encompass various specific strategies. For their part, specific strategies—labeled with a present participle—were generally described by using more extensive verbalization; the participants described their mental process in relation to a specific goal or musical result. For example, all the segments in which the participants mentioned strategic mental processes related to chord or nonchord tones were coded under harmonic priority, which is a general strategy, while the peculiar case in which a participant mentioned using quartal harmony was coded separately as a specific manifestation of harmonic priority.
Then, attempting to organize the 46 strategies we had identified into a coherent taxonomy, we looked for their “highest common denominators” to “build a synthetic and structured representation of the analyzed content” (Paillé & Mucchielli, 2012, p. 257), following the thematic content analysis method. This interpretative process allowed us to regroup the strategies into five encompassing categories according to their respective functions:
preplanning strategies, implemented before the improvisation;
conceptual strategies, implemented to establish pitch, rhythm, or tonality;
structural strategies, implemented to organize the improvisation;
atmospheric and stylistic strategies, implemented to set a mood, a sonority, or a style;
real-time strategies, implemented to negotiate the extemporary nature of ideation and execution during improvisation.
Table 3 lists the categories and the general and specific strategies we documented. The participants mentioned using a total of 9 general strategies and 37 specific strategies; 28 (out of total of 46) general and specific strategies that were not part of the initial coding scheme were induced from our data. Daniel used the smallest number of strategies (19 in total), and Raùl used the largest number of strategies (31). The mean number of strategies implemented per musician was 24.
Preplanning Strategies
Preplanning refers to all the musical decisions made by the musician before the improvisation. The participants in the present research mentioned using three different pre-planning strategies: choosing a theme, fixing a frame, and setting the color palette. Choosing a theme is a strategy whereby the musician determines, before improvising, a unifying element of the improvisation. Similarly, by fixing a frame, the musician establishes the outlines of the improvisation while leaving space for undetermined musical decisions. For Suzanne, choosing a theme was a decision that was made in the very moments preceding the improvisation, while David’s fixing a frame process extended over a longer period. In fact, knowing that his participation in the research implied improvising, he prepared the frame of his performance beforehand, composing some musical sections and leaving other sections open to extemporization
5
: I tried to make a plan that stayed in related keys . . . it was built gradually. One day, I was improvising, and then I felt that the beginning worked well, and after that it just went anywhere. Then, the more I advanced, the least it was improvised . . . it is as if I was composing the piece more and more. (David)
Setting the color palette is a strategy used by the musicians to prepare the sonority of their improvisation without targeting specific notes, harmonies, or rhythms. When implementing this strategy, the participants interviewed either chose a peculiar scordatura (guitar) or prepared specific tone colors (guitar and organ). Planning a specific tuning or tone in advance allowed the musician to anticipate the “sound” of the improvisation or prepare an optional plan (e.g., a sonority for the final chord or something to be used in the case of a lack of inspiration).
Conceptual Strategies
Conceptual strategies center on pitches or rhythms. Pertaining to this category are four general strategies: harmonic priority, melodic priority, tonal approach, and rhythmical considerations. Harmonic priority regroups all improvisational thinking whereby notes are chosen according to their relationship to chord tones; for example, “I’m on a dominant and [sings] I play superimposed minor thirds. I have a diminished seventh chord with lower appoggiatura” (Daniel). A specific strategy related to this general strategy is using quartal harmonies (i.e., chords consisting of superimposed fourths). Similar to harmonic priority, melodic priority was used to determine the specific pitches of the improvisation. However, this choice was based on scales, modes, or intervals instead of chord tones. Specific strategies linked to melodic priority are thinking in intervals to select the pitches to play, using the chromatic scale, and using parallel motion. All the participants used a tonal approach at some point in their improvisation; namely, the tonality of the improvisation was at the foreground of their thinking process for a moment. As an example, David explained: I went in D minor and then passing tone G minor after that I quickly came to the relative major, F major. There I brought back the theme. After that . . . a kind of B flat, but really brief there . . . and then I pretended to return to F major but I went to F minor instead. I think it is the most distant tone that I used, which is, as you know, the homonymous tonality of the relative major.
Moreover, Jimmy mentioned using the specific strategy of “breaking” tonality, which falls into the tonal approach general strategy, by using various musical means to avoid the installation of a tonal center. Furthermore, all the participants mentioned having rhythmical considerations while they improvised; such strategies allow the musicians to set the tempo, pulse, or meter of their improvisation. For Raùl, rhythm and specifically groove, one of its components, is perhaps the most important element in improvisation: Once you get in the groove, you stay there until something else happens. So yes, this groove, I did not plan it much, but I felt there had to be something [i.e., a groove]. If not, it is not really serious and nothing happens and it bores. If there is no rhythm in a section, it bores.
Among other rhythmical considerations, Raùl also mentioned using a specific strategy that consists of choosing a comfortable (moderate) tempo for improvisation, one that allows him to use various rhythmical figures: It is always with a comfortable “groove.” I like to improvise with a “click” around 70-80. I feel that it is a very comfortable speed, I do not say I always do it like that, but it’s a speed that you can double, triple, triplets . . . you can do many things.
To summarize, conceptual strategies were used by the participants to determine the pitches, rhythms, or tonality of a segment of their improvisation.
Structural Strategies
Structural strategies were used to determine the organizational dimensions of the improvisation. General strategies in this category can be organized on a continuum ranging from the micro-structural (phrasing and articulations) to the macro-structural (formal construction of the improvisation). On this continuum, following the strategies linked to phrasing and articulations are meso-structural strategies related to the idea bank and the development of the improvisation. While three participants had thoughts linked to the phrasing and articulations of their melodic material, all of the participants drew prelearned musical motives from their knowledge base in the course of their improvisations. The specific strategies related to the idea bank are quoting a theme (or the head or the tail of a theme) drawn from long-term memory and expanding technical ideas, which consists of using overlearned technical elements with different pitches to create new musical ideas. As mentioned by Raùl, this strategy is an efficient way of maximizing the potential of technical elements: It’s an arpeggio of Leo Brouwer, from Brouwer’s Sonata. I said: “It sounds nice.” . . . It would be very stupid and very sad not to use this arpeggio again in my life and to use it in only 6/8 time. It will be very sad to use it only for this in my life. So, when I learn a musical work, I try to use the licks of that work to my benefit. So I remember exploring [musical example] and changing the rhythm. Rather than sixteenth notes in ternary rhythm, I do it in 4/4 [musical example]. It (sounds) pretty . . . and then I explore [this pattern with different] chords. (Raùl)
Development strategies allow the musician to create new musical ideas using pre-composed musical ideas or materials that were previously played during the improvisation. Development strategies comprise embellishing, varying or recombining using grace notes, variation techniques, or blending various musical ideas; repeating or sequencing 6 the ideas that have been played; creating a period, or an antecedent/consequent phrase; varying intensity, focusing on the creation of musical peaks or the alternation of tension and relaxation; and contrasting approach, by developing contrasting material in relation to what has just been played. These strategies are used to ensure the continuity of the improvisation while avoiding monotony.
The musicians interviewed used formal strategies to establish the improvisation architecture. Specific strategies related to the formal dimension of the improvisation include setting a harmonic structure underlying the improvisation (or one of its parts), building a bridge between distinct musical sections, and setting the duration (or thinking about the length of the improvisation). Suzanne used the latter strategy when she refrained from extending a part of her improvisation to maintain the audience’s attention: This is improvisation: one is not always doing what one expected. However, when you have an audience, you tell to yourself: “It’s going to be [too] long, it does not make sense to go in that direction . . . think about the others . . . don’t just have fun and let your fingers go.” . . . Rationally, I would not have continued to procrastinate for long like this, because it really sounds like musical procrastination. But it’s pretty, there are beautiful harmonies, but . . . you know as people used to say: “Get on with it” (laughs).
The previous quote brings to light how expert improvisers not only possess the capacity to generate their improvisation in real time but also have the ability to anticipate the audience’s reception of an anticipated musical segment and shape their improvisation according to this projection.
Atmospheric and Stylistic Strategies
Atmospheric and stylistic strategies determine the overall sonority of the improvisation without affecting directly its notes, rhythms, or structure. Six specific strategies pertain to this category: imitating another instrument, imitating a composer or a piece, imitating a style or a genre, setting the mood, setting an ostinato or a pedal, and focusing on the tone color. Imitation strategies were especially favored by the participants to emulate an instrument (a singer), a composer or a piece (e.g., Leo Brouwer or Star Wars Theme by John Williams) and a style or genre (recitative, “gesture music,” ouverture and scherzando). Setting the mood allowed the musician to establish a distinctive color or emotion for a given segment of her improvisation. For example, From here, I went into a mood . . . we are on a beach, and we completely change the landscape. And then I did not come back like in the beginning, as I wanted. However, harmonies came, and I thought it was a nice little diversion because it went fast, it went light, it went strong and all that, and then I said, “Oh yes, it feels good just these small waves.” (Suzanne)
Other atmospheric and stylistic strategies, such as setting an ostinato or a pedal and focusing on the tone color, are standard compositional techniques, the former being used when improvising on a repetitive pitch or motive and the latter being used when the attention of the musician is focused on elements such as the dynamics, timbre, or texture of the improvisation.
Real-Time Strategies
Real-time strategies are linked to the fact that improvisation is a live event. Because of the extemporary nature of improvisation, the musicians develop a peculiar relationship with their instrument, inspiration, and unexpected events. Since improvisation cannot be paused, the musicians learn to implement real-time strategies to keep on playing, even when they might temporarily run out of inspiration or play unexpected notes. The real-time strategies we inventoried include faking improvisation, gaining time, exploring the field, keeping a steady rhythm, using brief durations, avoiding the beaten paths, refusing the unexpected, and using the unexpected creatively. Faking improvisation consists of mimicking an improvisatory character while playing overlearned ideas, using techniques such as creating false hesitations in the course of performance. Gaining time was done by looping a musical segment for longer than the musician’s aesthetic sense suggested, waiting for inspiration to come. Exploring the field is a sequential approach whereby pitches are gradually integrated into the improvisation, thus installing and progressively creating a musical mode; when implementing this strategy, the musician was thinking in terms of “green” (good) and “red” (wrong) notes. For Raùl, keeping a steady rhythm can give coherence to any pitches played, even unexpected notes: [When] you have a steady beat, you can justify all the notes you want [example]. . . . Everything can be justified if you have the beat. Any chord: [musical example]. Whatever you want, if you get it into the rhythm.
Another way of justifying any pitch is by using brief durations. Again, for Raùl, any note can sound good when it’s played really fast and not sounded for a long period of time; therefore, playing rapid notes allows the musician to experiment. Avoiding the beaten paths is the strategy that consists of doing something deliberately new, avoiding one’s habits or clichés, for example by playing in an unaccustomed position or fingering, using less familiar voicing or tonalities, or avoiding staying too long in the “comfort zones.” Refusing the unexpected and using the unexpected creatively are two distinct reactions to unexpected occurrences: On one hand, the “error” is rejected and rectified in a compositional spirit, while on the other hand, the “error” is “accepted” and purposefully integrated into the improvisation, for example, by repeating or resolving the unexpected note.
Because improvising musicians need to realize their ideas in real time, they might at some point face the limits of their technique and instrument. Therefore, they develop various strategies linked to the morphology of their instrument or their technical ability, which we have grouped under the general strategy instrumentalities. For stringed instruments, an example of such a strategy is to change tonality at some place in the improvisation to have the possibility of using open strings. Under instrumentalities, we identified three specific strategies: using idiomatic techniques or instrument-specific techniques (e.g., harmonics, tapping, and microtones), keeping a visual contact with the instrument, 7 and simplifying an idea to be able to realize it technically in real time. In sum, the improvising musicians implement real-time strategies to reduce the cognitive load or technical demand of the improvisatory process or palliate unexpected events.
Discussion
From the methodological point of view, this study is, as far as we know, the first to investigate musical improvisation by using a retrospective verbal protocol with subjective aided recall, which has proven to be a fruitful data collection strategy. From the richness of the data gathered in this research, we conclude that subjective aided recall is a powerful tool that helped the participants “re-live” their in situ experience. Using this data collection method, we gathered the in-depth perspectives of five expert Western classical music improvisers about their performance strategies. A thematic content analysis of the participants’ interviews revealed 46 improvisational strategies that were organized in a three-level hierarchy: categories, general strategies, and specific strategies. All the strategies were classified into one of five top-level categories: (a) pre-improvisatory planning, (b) conceptual strategies, (c) structural strategies, (d) atmospheric and stylistic strategies, and (e) real-time strategies.
During the process of analyzing the data and organizing it into a coherent taxonomy, theoretical saturation was attained for the categories and general strategies before all the material was taken into consideration but not for the specific strategies. In other words, every participant used at least one specific strategy that was not implemented by the others, but all the categories and general strategies were shared by at least four participants. Therefore, it appears that the categories and general strategies we identified could be relatively representative of Western classical improvisation strategic thinking. However, we believe the list of specific strategies we inventoried is not definitive; a more substantial sampling would probably contribute a greater number of specific strategies. Furthermore, while no clear-cut generalization can be made, it appears that categories and general strategies could be of a more transversal nature—being used in various music genres or by various improvising musicians—while specific strategies tend to have a more genre-specific, or even idiosyncratic, character. As an example, because Western classical music is typically transmitted with the support of notation, improvisation in that genre might be more grounded in formal music theory than in other genres, which might cause an increase in the utilization of specific strategies pertaining to the conceptual or structural categories. Similarly, the quest for precision and perfection that is promoted in Western classical music performance might augment the implementation of specific strategies falling into the real-time category.
From another perspective, comparing our data with our initial coding scheme, we discovered 28 strategies in our data that were not mentioned in the background literature. Beyond providing a satisfactory answer to our research question, our results tell us that the mind of the improvising musician is not inert. The mean of 24 strategies used by the musicians in performance in a single improvisation (slightly above half the total of the strategies we identified, 46) shows that Western classical expert improvisers possess the ability to implement varied strategic foci and switch rapidly from one focus to another. Furthermore, all the musicians in the present study implemented strategies pertaining to at least four categories (conceptual strategies, structural strategies, atmospheric and stylistic strategies, and real-time strategies) during their improvisations, while four improvisers implemented preplanning strategies before improvising. Finally, the structure of the taxonomy, comprising three levels (category, general strategy, and specific strategy), and the grouping of the strategies under five categories is a contribution in itself because it provides a kind of organization that we observed to be too often lacking in the background literature investigating the generation of musical improvisation.
Limitations
The main limitations of this research project are our sample size and methodological approach. First, a more substantial sample would probably have revealed a greater number of strategies, and our small sample impedes the generalization of our results. Second, the retrospective verbal protocol does not allow access to unconscious cognitive mechanisms and leaves room for various biases and distortions (van Someren, Barnard, & Sandberg, 1994). Third, there may be strategies implemented by other types of instrumentalists (e.g., wind players) that were not utilized by the musicians in this study. Still, we consider that our methodological design constitutes an excellent exploratory approach to document Western classical music expert improvisers’ strategies.
Further Research
Building on the specific focus of this article, further investigations on the strategy use underpinning the creative process of musical improvisation would benefit from using a retrospective verbal protocol with a subjective aided recall strategy and/or comparing the results obtained using this strategy with those obtained using an objective (third-person) aided recall. Further research studies with a greater number of expert participants are necessary to generalize the exploratory taxonomy that emerged from this study. A larger number of participants would also help in identifying more precisely the strategies that are most frequently used, or those that are more qualitatively important for Western classical improvisers, as well as the strategies that are rather idiosyncratic. It is hoped that our taxonomy will be refined and developed through further empirical works on Western classical music improvisation strategic thinking. Furthermore, preplanning appears to be a potentially fertile area for investigation where more strategies might be evinced, for example, studying specifically what is going on in the head of the musician in the minutes before improvising. Hence, further research could use a similar method with a greater number of participants and/or focus on preplanning strategies.
Finally, further research should investigate the effect of a teaching and learning approach based on our results focused on the improvisational and other musical abilities of the learner. Both the capacity to implement various strategies (not only pitch-related ones, which are often at the foreground in improvisation methods) and the capacity to switch from one strategy to another should be developed in the learner.
Pedagogical Implications
Our taxonomy provides valuable insights to guide the development of a Western classical music improvisation curriculum. The pedagogy of Western classical music improvisation should consider the strategies we identified and instruct the learner to master progressively more and more complex strategies in all five categories of strategies we identified.
Various pedagogical ideas can be inspired by our results. In the first place, self-efficacy in improvisation should be developed in the learner. Creating solo, duo, or ensemble improvisation activities based on highly permissive atmospheric and stylistic strategies such as setting the mood (narrating a story or musically depicting a familiar significant place or event), creating an ostinato or a pedal, or focusing on the tone color can help the learner to develop self-confidence in improvisation. When presented in an open-ended, ludic context, these strategies allow for various musical responses and results within a frame, thus limiting what are perceived as “errors” or mistakes by the learner.
Second, to set the stage for conceptual and structural strategies, the teacher should aim to guide the learner in becoming a musical gold-digger; the learner should develop an acute interest for “valuable” musical ideas. The improviser has to develop awareness for musical ideas, themes, and elements of musical theory in the surrounding environment, recognizing their potential usefulness in the course of eventual improvisations. As an example, in developing such a consciousness, the learner might build autonomously his or her idea bank by hearing, inventing, or reproducing simple motives; learning popular themes; or exploring the creative potential of the technical exercises he or she is practicing. The key point here is that in doing so, the learner will enhance his or her awareness of the music surrounding him or her and thus develop critical judgment and aural abilities. Improvising from any element of music theory is a good way to turn the conceptual into the practical.
Third, the learner should become a craftsman, using micro-structural strategies, namely, learning to transform the musical gold he or she has gathered into crafts that are adapted to a specific need or context. As an example, by embellishing, varying, or recombining the musical ideas stored in his or her long-term memory, at first during improvisational exercises and eventually during a full improvisation, the learner will develop a practical understanding of the various dimensions of music theory and acquire polyvalent instrumental aptitudes. This will in turn widen the spectrum of musical contexts in which he or she will feel confident improvising.
Fourth, the learner should learn to become an expert decision maker. Beginners should learn to implement preplanning strategies. Making musical decisions before improvising—whether or not the actual improvisation follows the plan—will familiarize the learner with an embodied experience of enabling constraints. 8 More advanced learners could develop the aptitude to have a perfect Plan A and an at-least-almost-perfect Plan B before improvising while being able to change their minds midperformance if necessary.
Fifth, the improviser will become an architect by implementing macro-structural strategies by being aware of the details but also conscious of the whole while implementing development and form strategies to create a coherent musical structure in real time.
Finally, throughout his or her apprenticeship, the learner will have to deal with the extemporaneous nature of improvisation using real-time strategies. Experimenting with the strategies listed under this category and developing his or her own strategic toolbox, the learner will develop confidence in risk taking and flexibility.
Footnotes
Authors’ Note
Ethical approval for this project was given by the Comité d’éthique de la recherche avec des être humains de l’Université Laval (CÉRUL) [ref number 2013-110 / 10-07-2013].
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Social Sciences and Humanities Research Council of Canada (SSHRSC) and the Observatoire interdisciplinaire de création et de recherche en musique (OICRM).
Notes
Author Biographies
). Her field of specialization is research-creation in music with creators, particularly about modernization of Inuit throat-singing (with Tanya Tagaq), and the genetic analysis of the creative process in music (with Productions SuperMusique).