Construction of a harmonic phrase

Alma Mater Studiorum of Bologna, August 22-26 2006 Construction of a harmonic phrase Ziv, N. Behavioral Sciences Max Stern Academic College Emek Yizre'el, Israel naomiziv@013.net Storino, M. Dept. of Music Bonfiglioli, L. Incasa, I. Caterina, R. Baroni, M. Dept. of Music ABSTRACT Background. Previous studies (Krumhansl, 2001, 2005; Tillmann, Bharucha, & Bigand, 2000) have established that the average listener possesses a certain representation of tonal grammar. This acquired knowledge, stored in long term memory, guides listening and helps create a schematic representation of musical pieces. Most studies (Povel & Jansen, 2002) either require participants to judge the appropriateness of short harmonic or melodic sequences, or ask participants to reconstruct, through the puzzle paradigm (Deliège, 1987, 1989, 2001), given musical pieces. Aim. The aim of the present study was to examine whether the grammatical knowledge of the tonal idiom is strong and stable enough to direct the active construction of a sequence of chords. The subjects were divided into three groups corresponding to different levels of presumable competence: from a minimum to a maximum. Their competence was valued according to different criteria, in order to avoid a too simple distinction between In: M. Baroni, A. R. Addessi, R. Caterina, M. Costa (2006) Proceedings of the 9th International Conference on Music Perception & Cognition (ICMPC9), Bologna/Italy, August 22-26 2006. 2006 The Society for Music Perception & Cognition (SMPC) and European Society for the Cognitive Sciences of Music (ESCOM). Copyright of the content of an individual paper is held by the primary (first-named) author of that paper. All rights reserved. No paper from this proceedings may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information retrieval systems, without permission in writing from the paper's primary author. No other part of this proceedings may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information retrieval system, without permission in writing from SMPC and ESCOM. and non-. Method. 66 participants took part in the study. Participants were presented with 8 piano chords (I I II III IV V V 7 VI) in random order. The participants were asked to listen to each chord, and then to combine them in order to create a satisfying musical sequence. Participants were allowed to try out as many combinations as they liked. Results. Two main criteria were chosen to evaluate the results: a positive criterion (a perfect cadence subdominant, dominant and tonic in the final position) and a negative one (no tonic in the first and final position). According to these criteria and to their possible mixtures (analyses focused on the positioning of the various chords in the sequence) the results were organized form a maximum to a minimum of apparent tonal competence. Preliminary results seem to indicate that, although a certain theoretical musical competence linked to musical studies seems to be necessary in order to get accuracy and precision in the construction task, different and not clear relationship can be found between the levels of musical competence and the correctness of the results. Conclusions. The research shows the necessity to establish more accurate distinctions among the participants of tonal test, beyond the two traditional categories of and non-. INTRODUCTION Previous studies have established that the average listener possesses a stable representation of tonal grammar. This knowledge is acquired even without formal musical training, and is stored in long-term memory (Krumhansl, 2001, 2005; Tillmann, Bharucha & Bigand, 2000). When listening to musical pieces, it guides the listener both in connecting between successive notes (Cuddy & Lunney, ISBN 88-7395-155-4 2006 ICMPC 1872

1995; Narmour, 1990) and in creating a schematic representation of the piece as a whole (Deliège & Mélen, 1997). Most studies require participants to judge given short harmonic or melodic sequences on a variety of parameters (Janata, Birk, Tillmann & Bharucha, 2003; Jansen & Povel, 2004). Participants in these studies are thus passive listeners to artificial musical stimuli. Other studies (e.g.: Deliège, 1996, 2001; Koniari, Predazzer & Mélen, 2001) require participants to actively reconstruct a given whole musical piece, through the puzzle paradigm. Studies comparing performance between and non- vary in the criterion for distinguishing from non-, defining as participants having 6 (Dowling, Tillmann & Ayers, 2002), 8 (Holleran et al., 1995) or 12 years of formal musical training (Bigand & Pineau, 1997). The present study was designed to answer two main questions: 1. Would the mental representation of tonal grammar guide participants in an active task of constructing a tonal harmonic phrase from a set of given chords? Presented with an unorganized set of diatonic chords, would participants juxtapose them to create a balanced harmonic phrase? 2. Would a finer distinction between levels of musical expertise, differentiating between and non- in more specific categories, explain differences in musical compositions? It was hypothesized that the higher the level of musical expertise, the closer the musical phrase composed would be to a balanced, hierarchical harmonic phrase, beginning on the tonic, and ending with a perfect cadence, according the rules of tonal grammar. METHOD Participants Sixty-six participants took part in the study (38 females, mean age= 25.71, s.d.=6.05, 28 males, mean age=31.36, s.d.=8.05). General mean age=28.11, s.d.=7.47. Participants were divided into three groups, according to musical expertise: participants having 8 years or more of formal musical training were considered professional (23 participants), participants having between 4-8 years of musical training were considered nonprofessional (15 participants), and participants having less than 4 years of formal musical training were considered non- (28 participants). Questionnaire A questionnaire regarding musical expertise was created for the study. Participants were asked whether they studied music, and if so, how many years. They were then asked to rate on a scale from 1 to 5 (1=not at all, 5=very much) their degree of liking of various musical styles: classical, rock, pop, jazz, techno, rap, folk, ethno, latin, new-age. They were also asked to indicate the number of hours they listen to music each day: 1= never, 2= less than one hour a day, 3= between two and three hours a day, 4= between four and seven hours a day, 5= all the time. Apparatus A special computer program was designed for the study. On the top part of the screen a row of squares was presented, numbered from 1 to 8. By placing the cursor on one of the squares and pressing the right button on the mouse, participants could hear each numbered segment, which was a chord. On the bottom part of the screen, eight empty squares appear, into which participants may copy the number of the chord in each position. By pressing a button marked "try the sequence" participants could listen to the series of chords they combined in succession. By pressing a button called "cancel the sequence" they could erase the combination they tried. Musical stimuli The chords used in the study were diatonic chords from the C major key. The degrees used were: I, I, II, III, IV, V, V 7, VI. The order of presentation in the initial series is presented in table 1: Table 1. order of presentation of chords Position 1 2 3 4 5 6 7 8 Degree VI II I I III V 7 V IV Procedure Participants were presented with the computer program, and received an explanation about how to use it. They were then asked to listen to the segments separately a first time, and then to try to combine them on the bottom line, in whichever order they liked. They were told they could try as many combinations as they liked, until they arrive at a combination which includes all segments, each segment only once (the program did not enable them to use a segment twice), and which sounds good to them. No further specifications were made, in order to allow participants as much freedom as possible in deciding on final combinations. After finishing the task, participants were asked to fill in the questionnaire described above, and to rate on a scale ranging from 1 to 5 (1=not at all, 5=very much) their degree of satisfaction with the combination they created. RESULTS Listening habits No significant differences between professional, non-professional, and non were found in hours spent listening to music each day. Participants in all groups listen on average between two and three hours a day to music. ISBN 88-7395-155-4 2006 ICMPC 1873

As for preferred musical styles, several significant differences were found between the three groups: Professional prefer classical music significantly more than non-, and less than non-professional and non- to techno. They listen significantly less than non- to rap and new-age music. Means and standard deviations of liking for the most liked styles are presented in table 2. Table 2. Means and standard deviations of preferences for musical styles Professional Non-prof. Non Classical* 4 (1.38) 3.33 (1.17) 2.71 (1.3) Rock 3.17 (2.28) 4.13 (1.12) 3.14 (1.43) Pop 2.96 (1.14) 2.47 (1.3) 3.07 (1.38) Jazz 3.09 (1.31) 3.47 (1.55) 2.82 (1.5) Techno* 1.09 (0.5) 1.73 (0.96) 1.68 (1.09) Rap* 1.3 (0.65) 2.07 (1.16) 2.29 (1.38) Folk 2.09 (1.31) 2.73 (1.43) 1.86 (1.2) Ethno 2 (1.24) 2.73 (1.33) 2.61 (1.39) Latin 2.26 (1.84) 1.8 (0.86) 2.54 (1.2) New Age 1.52 (0.59) 1.73 (1.03) 2.46 (1.2) *p<.05 As can be seen in table 2, the preferred styles of professional were classical music, rock and jazz in that order. The non-professional prefer the same styles, but their ordering is rock, jazz and classical music. The non- prefer rock and pop. A new variable was computed, in which the number of musical styles receiving a rating of 4 or 5 was counted for each participant. Although the differences between the groups were not significant on this variable, the nonprofessional received a higher mean on this variable (3.06, s.d.=1.43) than professional (2.3, s.d.=1.6) and non- (2.82, s.d.=1.46). This tends to suggest that non-professional have a wider range of musical styles which they listen to. Musical combinations In order to compare the final musical combinations achieved by participants, several dichotomic categorical variables were defined, focusing on the conclusion and the beginning of the sequence according to the rules of tonal musical grammar: 1. A sequence ending on V (7) -I or II(IV)-V (7) -I 2. A sequence ending on I 3. A sequence beginning on I 4. A sequence neither ending nor beginning on I The total number and percentage of participants fulfilling the new criteria variables are shown in table 3. Expected values were calculated as the probability of placing each chosen chord in a given position. For example, the probability of placing the I degree in the final position was 2/8, since the I degree appeared twice in the given segments. Table 3. Number and percentage of participants fulfilling combination criteria Yes No Chi square Ending on V (7) -I Or II (IV)-V (7) -I 25 (38%) 41 (62%) 94 Ending on I 44 (66%) 22 (33%) 61 Beginning on I 27 45 8.9 Neither Ending nor beginning on I (40%) 10 (15%) 56 (85%) p<.05 32 As can be seen in table 3, all results were significant, showing that in general, participants tended to place the I degree in the first and final position, and the V or V 7 in the position before last. Participants were relatively satisfied with their compositions (mean=3.77, s.d.=.76). No significant relations were found between fulfilling the different criteria and hours spent listening to music or preference for the various musical styles. Comparison between levels of musical expertise The number and percentage of participants within each group who filled these criteria are presented in table 4. Table 4. Number and percentage of participants fulfilling combination criteria Professional Non-prof. Non Ending V-I or II(IV)-V-I 13 (56.5%) 5 (33%) 7 (25%) Ending with I 20 (87%) 9 15 (53%) Beginning with I 12 (52%) 9 6 (21%) Neither ending nor beginning with I 1 (4%) 1 (6%) 8 (28%) No significant differences between professional, non-professional and non- were found in the first parameter, concluding the combination with a perfect cadence. However, significant differences were found between the groups in the three other parameters: ending the combination with the I degree (chi square= 6.722, p=.035), beginning the sequence with the I degree (chi square= 7.865, p=.020), and neither ending nor beginning on the first degree (chi square= 6.851, p=.033). In general, professional ' final combinations tended to conform more than non-professional ' and non-professional ISBN 88-7395-155-4 2006 ICMPC 1874

' to compositional norms of tonal music. It is interesting to note, however, that the percentage of nonprofessional beginning the sequence with the I degree was higher than that of professional. No significant differences between the groups was found in degree of satisfaction with final combinations. When looking at each group separately, several significant differences were found: in the professional ' group, a significant difference was found in ending the sequence on the I degree (chi square=12.565 ), and in neither ending nor beginning on the I degree (chi square=19.174, ). In the nonprofessional ' group, a significant difference was found in neither ending nor beginning on the I degree (chi square=11.267, p=.001). In non-, a significant difference was found in ending with a cadence (chi square=7, p=.008), in beginning with the I degree (chi square= 9.143, p=.002), and in neither ending nor beginning on the I degree (chi square=5.143, p=0.023). No significant differences were found within each group in the levels of satisfaction between participants who fulfilled each criterion and those who didn't. DISCUSSION The present study required participants to combine a set of diatonic chords to create a harmonic phrase. Results show that beyond levels of musical expertise, participants in general succeeded in recognizing the tonic, and tended to place it at the conclusion of the musical phrase. The use of the puzzle paradigm in the present study enhances results found in previous studies investigating the mental representation of the tonal system (e.g.; Holleran, Riess Jones & Butler, 1995; Jansen & Povel, 2004) by asking participants to actively manipulate the musical stimuli, instead of simply rating given musical segments on various scales. Results may thus give a stronger indication of the influence of previous knowledge on the processing of music. Moreover, the use of a simple series of chords, instead of a segmentation of a real repertoire piece (e.g.; Deliège, Mélen, Stammers & Cross, 1994), allows participants greater freedom in their manipulation, and does not imply strictly correct or incorrect answers. As was expected, level of musical expertise influenced the construction of the harmonic phrase. Participants having a high level of formal musical training tended to combine the given chords in a fashion conforming to tonal norms more than non-professional and non-. However, even within the two levels of lesser musical expertise, participants succeeded above chance in creating such a phrase. Another important remark is the influence of the musical styles preferred by the three groups of listeners. As we have seen, the professional listen to classical music more than non-professional and non. The different results in the construction of the harmonic phrase could depend not only on the musical competence but also on the characteristics of the usually listened musical styles. The rock and pop music are based on different sequences of harmonies respect of tonal music; they have not always the simple harmonic schema of tonal classical music. On the other side long many years the professional play tonal music constructed according the succession of chords used in the present research. This point suggests the need for further studies in order to investigate the real influence of the listened musical styles in task of evaluation and of reconstruction of tonal music. ACKNOWLEDGMENTS Thanks due to Fabio Regazzi for the creation of the computer program and for his precious suggestions. REFERENCES Bigand, E. & Pineau, M. (1997). Global context effects on musical expectancy. Perception & Psychophysics, 59(7), 1098-1107. Cuddy, L.L. & Lunney, C.A. (1995). Expectancies generated by melodic intervals: Perceptual judgments of melodic continuity. Perception & Psychophysics, 57(4), 451-462. Deliège, I. (1996). Cue Abstraction as a Component of Categorisation Processes in Music Listening. Psychology of Music, 24/2, 131-156. Deliège, I. (2001). Prototype Effects in Music Listening: An Empirical Approach to the Notion of Imprint. Music Perception, 18/3, 233-244, 371-407. Deliège, I., Mélen, M., Stammers, & Cross, I. (1994). Musical schemata in real time listening. In I. Deliège (Ed.), Proceedings of the 3rd ICMPC (pp. 271-272). Liège: ESCOM Publications. Deliège, I. & Mélen, (1997). Cue abstraction in the representation of musical form. In I. Deliège & J. Sloboda (Eds.), Perception and Cognition of Music (pp. 387-412). Hove: Psychology Press. Dowling, W.J., Tillmann, B. & Ayers, D.F. (2002). Memory and the experience of hearing music. Music Perception, 19(2), 249-276. Holleran, S., Riess Jones, M. & Butler, D. (1995). Perceiving implied harmony: The influence of melodic and harmonic context. Journal of Experimental Psychology: Learning, Memory and Cognition, 21(3), 737-753. ISBN 88-7395-155-4 2006 ICMPC 1875

Janata, P., Birk, J., Tillmann, B. & Bharucha, J.J. (2003). Online detection of tonal pop-out in modulating contexts. Music Perception, 20(3), 283-305. Jansen, E. & Povel, D.J. (2004). Perception of arpeggiated chord progressions. Musicae Scientiae, 8(1), 7-52. Krumhansl, C.L. (2001). Cognitive Foundations of Musical Pitch. New York: Oxford Press. Krumhansl, C.L. (2005). The cognition of tonality as we know it today. Journal of New Music Research, 33, 253-268. Narmour, E. (1990). The Analysis and Cognition of Basic Melodic Structures: The Implication-Realization Model. Chicago: of Chicago Press. Povel, D. & Jansen, E. (2002). Harmonic Factors in the Perception of Tonal Melodies. Music Perception, 20/1, 51-86 Tillmann, B., Bharucha, J.J. & Bigand, E. (2000). Implicit learning of tonality: A self-organizing approach. Psychological Review, 107, 885-913. ISBN 88-7395-155-4 2006 ICMPC 1876