Anda di halaman 1dari 5

Alma Mater Studiorum University of Bologna, August 22-26 2006

Harmonizing a tonal melody at the age of 615 years


Pirkko Paananen
Department of Music University of Jyvskyl 40013 Jyvskyl, Finland ppaanane@campus.jyu.fi

ABSTRACT
In this study, 615-year-old participants (N=44) harmonizations of a tonal melody were analyzed. The context consisted of two verses of a simple tonal melody in C major. Four marked keys (C, D, F, G) of the synthesizer, each of which producing a whole major triad, were used in improvising an accompaniment as a real-time performance, recorded with Micro Logic 3.5 / Logic Education 5.5. Age-related development of harmony and metre was examined by analyzing the products statistically (correlation, ANOVA). The cases were clustered with hierarchical cluster analysis. The frequency of chord C (I) as the final chord, the distribution of chords to strong metrical positions, and the number of periodically harmonized measures increased with age. The number of events and chord transitions decreased with age, as well as the frequency of the out-ofkey chord D. In addition, seven clusters of varying developmental levels were found. At the earliest level of development children focussed on rhythmic impressions of chords or metre, and coordinated melody and chords poorly, producing a high number of events. At next level,

coordination of melody and harmony improved, and tonic closure became common. The chords emerged either at strong beats, or formed a harmonic riff. In the most developed products the melody and chords were wellcoordinated, the most frequent chord being C(I), the tonal closure being common, and the majority of chords emerging at the strongest beat.

Keywords
Harmony, chords, harmonization, tonality, metre, rhythm, development, children, school age.

INTRODUCTION
From the very beginning infants show sensitivity to melodic contour, familiar scale contexts and sensory consonance in perception (Trainor, Tsang & Cheung, 2002). However, the development of tonality and tonal harmony takes several years. Around 5.5 years of age, a moderate intervallic accuracy, key stability between phrases is achieved in singing (Davidson, McKernon & Gardner 1981; Dowling 1988; Kreutzer 2001). Harmonic perception in childhood has been studied from several viewpoints: consonance, tonal closure, local chord changes and implicit harmony. Zenatti (1969) found that 6-year-old children performed above the level of chance in identifying a chord that was different between pairs of short progressions. The ability to identify the chord was depending on the number of chords in the progression. Hair (1973) found that first graders were able to perceive differences between I and V7 chords. OHearn (1984) found clear improvement between first and third grade in perceiving a change in harmony in ongoing music. CostaGiomi (1994) found that 5-year-old children identified over 50% of the chord changes of a two-chord accompaniment

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.

ISBN 88-7395-155-4 2006 ICMPC

484

ICMPC9 Proceedings

played without melody. In a series of studies Costa-Giomi (2003) found remarkable improvement in the perception of implied harmony around 9 years of age under which performance was poorer in spite of harmonic instruction. Schellenberg, Bigand, Poulin-Charronat, Garnier & Stevens (2005) found that 611-year-old childrens judgements of the final, target chord in a chord sequence was faster and more accurate when the target chord was the tonic chord versus a less stable chord. While all these studies point to the development of harmonic perception during the school years, measured mostly in relatively local contexts, very little is known about childrens or adolescents abilities to harmonize tonal melodies. Some case studies, such as Wiggins (1994, 1995), have shown that 1012-year-old children learn to harmonize and even transpose basic chord transitions to different keys quite easily after a training period of few weeks.

Grade 0 (=preschool) (range 6.256.50 yrs), five in Grade 1 (range 7.177.83 yrs), six in Grade 2 (range 8.258.92 yrs), six in Grade 3 (range 9.259.83 yrs), six in Grade 4 (range 10.8010.92 yrs), six in Grade 5 (range 11.1711.83 yrs) and eight in Grades 89 (range 14.4215.75 yrs). The melody (fig.1) was a tonal children-song like tune in C major (oboe sound). However, the tonic is not present until measure 6, just before the cadence. The most frequently occurring tones in the melody are G(V), D(II), E(III) and C (I). The chords including melodic tones are above the melody in figure 1.

AIM OF THE STUDY


In this study, 615-year-old participants (N=44) harmonizations of a simple tonal melody were analyzed. Age-related development of harmonic features and chord rhythm was investigated, as well as the types of harmonization. The aim was to examine, whether children/adolescents rely on 1) random or rhythm-based production of chords, 2) a global sense of key or, 3) melodic intervals or implicit harmony, as a result of segmentation and chord recognition (see Jansen & Povel 2004). Furthermore, the role of metre in the development of harmony was investigated. On the basis of earlier studies of tonal hierarchy (Lamont & Cross 1994; Paananen 2003) it was assumed that the youngest (67 yrs) would prefer random, rhythm based or global key-based strategies. The ability to segment the melody and select the chord on the basis of the melodic interval (see Bartlett & Dowling 1980) or implicit harmonic relations (Costa-Giomi 2003) was not expected to develop earlier than the age of 89 years. The studies of tonal closure (Imberty 1969/1981), tonal hierarchy in perception (Lamont & Cross 1994) and in melodic improvisation (Paananen 2003), as well as studies on harmonization (Wiggins 1995, 1996) and harmonic perception (Costa-Giomi 2003) point to the possibility that children from the age of 10 onwards would be able to focus on both melody and harmony in a more differentiated manner in perception, and to harmonize a simple tonal melody. Figure 1. The melody and the chords including melodic tones. Chord production was reduced to the use of four marked keys (C, D, F and G) of a synthesizer, each of which produced a whole major triad (piano sound). With this method it was possible to harmonize a tonal melody as a real-time performance and, provide an opportunity to use more improvisatory or global strategies of production, with reduced motor demands. The participants explored the keys and listened to the two verses of the melody before harmonization. All participants created two harmonized versions (= two tracks with Micro Logic 3.5 / Logic Education 5.5.) of the two-versed melody, both of which were analyzed statistically.

RESULTS Age-related development of harmonic features


Age-related development of harmonic and harmonicrhythmic features was analyzed by ANOVA (67, 89, 10 11 and 14-15 yrs) and correlation and, the different types of harmonization with hierarchical cluster analysis. These features were: the frequency of each chord, the frequency of out-of-key chord D in the end of measure 5 (-> strong dissonance), the frequency of the I chord as the ending chord, coordination of melody and harmony in each phrase at -note level, the number and frequency of chord transitions and V-I-transitions, number of events, number of periodically harmonized measures and the frequency of the eight metrical positions (=the distribution of chords to metrical positions).

METHOD
Forty-four children/adolescents took part in the test (N=44; 18 girls and 26 boys). The participants studied at Norssi School in Jyvskyl, Finland. All, except one 10-year-old participant, had some experience of playing the synthesizer or some other keyboard instrument. Five participants had received at least one year of formal music training, while the rest had received general music education at school for 12 hours per week in each grade. Six children were in

ISBN 88-7395-155-4 2006 ICMPC

485

ICMPC9 Proceedings
0,33 0,31 0,29 0,27 0,25 0,23 0,21 0,19 0,17 0,15 C D chord F G 6-7yrs 8-9yrs 10-11yrs 14-15yrs

coordinated the chords and melody best in the end of the piece. Number of chord transitions correlated negatively with age results being non-significant; however, differences between the four age groups were significant (F (3,40)= 3.78; p=0.02). In similar, number of events correlated negatively with age results being non-significant, but differences between the four age groups were significant (F (3,40)= 4.28; p=0.01). No age-related results were obtained for the frequency of V-I-transitions. All metrical variables obtained significant age-related results: Number of periodically harmonized measures correlated with age (r=0.392**), as well as the frequency of metrical position 1 (r=0.324*) indicating the 1st beat of the measure, the frequency of metrical position 1+5 (r=0.353*) indicating the 1st and 3rd beats and, the frequency of metrical position 1+3+5+7 (r= 0.386**) indicating the note level. Differences between the four age groups were significant (F (3,40)= 2.87; p=0.05) in metrical position 1+5. The 67-year-olds differed from the 10-11-year-olds.

Figure 2. The harmonic profiles of the four age groups. Figure 2 presents the harmonic profiles (the frequencies of each chord) for the four age groups, 67, 89, 1011 and 1415 yrs. A slight preference for chord C is observable already at 6-7 years of age, however, the use of out-of-key chord D is almost as frequent as the tonic chord C and, as frequent as the subdominant chord F. The most developed profile is the 1415-year-old participants one. The 10-11year-olds profile reflects a preference for chord F as the tonal centre. Frequency of chord C (I) as the ending chord (Fig 3.) obtained significant results, correlation with age being moderate (r= 0.302*; p<0.01**), and differences between the age groups being significant (F (3,40)= 4.25; p=0.01), Scheff post hoc test showing differences between the youngest the three older age groups. Frequency of chord D (V/V) in the end of measure 5 correlated negatively with age, results being significant (r= -0.372*; p<0.05*).
Frequency of C (I) as the ending chord 100 %

Harmonizing types
The cases were clustered with hierarchical cluster analysis (average linkage between groups, measured by squared euclidean distance, standardized Z-scores). Age was not included in analysis. Seven clusters were found. The clusters and their salient features are as follows: Rhythmic (mean 98 months): No clear tonal centre, 25% ending with I chord, coordination of melody and harmony average, number of events average, number of periodically harmonized measures low, 44% of events distributed to metrical position 1+5 (1/2-note level), and only 24% to metrical position 1 (strong beat). Harmonic-metric early (mean 131 months): No clear tonal centre but 46% ending with I chord, coordination of melody and harmony average, number of events average, 76% of events distributed to metrical positions 1, 3, 5 and 7 (1/4-level) and 42% of events to metrical position 1 (strong beat). Harmonic-rhythmic (mean 147 months): Clear preference for chord G(V), 50% ending with chord I, coordination of melody and harmony high, number of events high, 47% of events distributed to metrical positions 1, 3, 5 and 7 (1/4level) and only 27% of events to metrical position 1 (strong beat).

80 % 69 % 60 % 43 % 40 % 29 % 20 % 33 %

0% 6-7yrs 8-9yrs 10-11yrs 14-15yrs

Figure 3. Frequency of C (I) as the ending chord for the four age groups. Coordination of melody and harmony was measured on the basis of common tones of melody and the chord. In the first two measures (12) this variable obtained nearly significant results (F (3,40)=2.77; p=0.05): the 14-15-year-old participants performed better than the younger age groups already in the beginning of the melody. All age groups

Metric (mean 119 months): Clear preference for chord F(IV), 19 % ending with chord I, coordination of melody and harmony low, number of events low, 84% of events distributed to metrical position 1 (strong beat). Puzzled (mean 121 months): Clear preference for chord F(IV), only 12,5% ending with chord I, coordination of melody and harmony low, number of events very low, number of periodically harmonized measures low, but 78% of events distributed to metrical position 1+3+5+7 (1/4level).

ISBN 88-7395-155-4 2006 ICMPC

486

ICMPC9 Proceedings

Harmonic-metric, developed (mean 146 mos): Tonal centre C(I), frequency of chord C(I) being 40%, 54% ending with chord I, coordination of melody and harmony high, uses more V-I-transitions than any other cluster, number of events relatively low, 85% of events distributed to metrical position 1+3+5+7 (1/4-note level) and 68% to metrical position 1 (strong beat). Harmonic, early (mean 92 months): Tonal centre C(I), frequency of chord C(I) being 40%, 58% ending with chord I, coordination of melody and harmony high, however, frequency of chord D in the end of measure 5 very high, number of events high, number of periodically harmonized measures very low, 39% of events distributed to metrical position 1+3+5+7 (1/4-note level) but only 19% to metrical position 1 (strong beat).

Costa-Giomi, E. (1994). Recognition of chord changes by 4- and 5-year-old American and Argentine children. Journal of Research in Music Education, 42, 68-85. Costa-Giomi, E. (2003). Young childrens harmonic perception. Annals of New York Academy of Sciences, 999, 477-484. Davidson, L. & McKernon, P. & Gardner, H. (1981). The Acquisition of Song: A Developmental Approach. In R. E. Taylor (Ed.) Documentary Report of Ann Arbor Symposium. National Symposium on the Applications of Psychology to the Teaching and Learning of Music. Reston, Virginia: MENC Reston. Dowling, W. J. (1988). Tonal Structure And Childrens Early Learning of Music. In J. A. Sloboda (Ed.), Generative processes in music. The psychology of performance, improvisation and composition (pp. 113128). Oxford: Clarendon Press. Jansen, E. & Povel, D-J. (2004). Perception of arpeggiated chord progressions. Musicae Scientiae, VIII (1), 7-49. Hair, H. I. (1973). The effect of training on the harmonic discrimination of first grade children. Journal of Research in Music Education, 73, 85-90. Imberty, M. (1969/1981). Tonal Acculturation And Perceptual Structuring of Musical Time in Children. In Basic musical functions and musical ability (pp. 107-130). Stockholm: Publications issued by the Royal Swedish Academy of Music 32. Kreutzer, N. J. (2001). Song Acquisition among Rural Shona-Speaking Zimbabwean Children from Birth to 7 Years. Journal of Research in Music Education, 49 (3), 198-211. Lamont, A. & Cross, I. (1994). Childrens Cognitive Representations of Musical Pitch. Music Perception, 12 (1), 27-55. OHearn, R. N. (1984). An investigation of the response to change in music events by children in grades one, three, and five. Dissertation Abstracts Int. 46: 371A. Paananen P. (2003). Monta polkua musiikkiin. Tonaalisen musiikin perusrakenteiden kehittyminen musiikin tuottamisja improvisaatiotehtviss ikvuosina 6-11. (Many Paths to Music. The development of basic structures of tonal music in music production and improvisation at the age of 611 years). Dissertation. Jyvskyl Studies in Humanities 10. University of Jyvskyl, Finland.

DISCUSSION
The youngest, 67-year-old children, seemed to benefit of the given method in the sense that it allowed them to focus on more global level of the piece. In general, they focused rather on the key than local chords and produced more events than the older participants. As expected, they produced a high number of events grouped randomly or into rhythmic patterns, lacking the typical chord rhythm of tonal music. However, the harmonic early type of production (mean 92 months) was able to coordinate melody and chords well, using a strategy of favouring the tonic chord C and producing a high number of events. Finding an appropriate chord was easiest in the cadence already for the 67-year-old children. The oldest age group (1415 yrs) generally ended the piece with the tonic chord C. A tendency of placing the chord at metrically strong positions developed clearly with age, and was common from age 10 onwards. These older children could also add a chord in the end of the phrase, or used riff-like patterns. The frequency of V-I transitions did not increase with age, which might be due to the given method. It is quite possible that children needed more time for focusing on local chord transitions. The 10-11-year-old children seemed to be puzzled of the task more than the younger or the older. This age group produced clearly less events than the younger. They seemed to have the general idea of what harmony should sound like, but they did not find the tonal centre C. It is possible that these children were attending to the melodic intervals more locally than the younger ones, and were in the middle of the shift from global to analytic way of representing tonal harmony.

REFERENCES
Bartlett, J. C. & Dowling, W. J. (1980). The Recognition of Transposed Melodies: A Key-distance Effect in Developmental Perspective. Journal of Experimental Psychology, 6, 505-510.

ISBN 88-7395-155-4 2006 ICMPC

487

ICMPC9 Proceedings

Schellenberg, E.G., Bigand, E., Poulin-Charronnat, B., Garnier, C. & Stevens, C. (2005). Childrens implicit knowledge of harmony in Western music. Developmental Sciente, 8 (6), 551-566. Trainor, L.J., Tsang, C.D. & Cheung, V.H.W. (2002). Preference for sensory consonance in 2- and 4-month-old infants. Music Perception 20 (2), 187-194. Wiggins, J.H. (1994). Childrens Strategies for Solving Compositional Problems with Peers, Journal of Research in Music Education, 42 (3), 232-252.

Wiggins, J.H. (1995). Building Structural Understanding: Sams Story. The Quarterly Journal of Music Teaching and Learning, VI (3), 57-73. Zenatti A. (1974). Perception et appreciation de la consonance musical par lenfant entre 4 et 10 ans. Sciences de lArt, 9 (1 & 2), 1-15. Zenatti A. (1969). Le developpement gntique de la perception musicale. Monographies Francaises de Psychologie No 17.

ISBN 88-7395-155-4 2006 ICMPC

488

Anda mungkin juga menyukai