xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001680999
007 cr mnu|||uuuuu
008 051208s2005 flu sbm s000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001130
Effect of age on 11- to 18-year-olds discrimination of nuances in instrumental and speech phrase interpretations
h [electronic resource] /
by Andrew Sioberg.
[Tampa, Fla.] :
b University of South Florida,
Thesis (Ph.D.)--University of South Florida, 2005.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 83 pages.
ABSTRACT: This dissertation was a continuation of study on a theory of a learning window for the perception of expressive qualities in music and speech. The proposed theory suggested that a practice window must overlap a learning window before it closes around the age of 10. This dissertation sought to determine whether children older than the proposed learning window continued to improve in speech and musical discrimination skill, or leveled off in this ability. It also examined the impact of gender and private lesson experience on discrimination ability. Instrumental music students (n = 292) attending a public magnet school for visual and performing arts in North Carolina between the ages of 11 and 18 participated in the study. Each student was administered a forty-item listening test containing 20 speech items and 20 instrumental music items. Each test item consisted of three short speech or musical phrases.
Adviser: Dr Jack Heller.
x Music Education
t USF Electronic Theses and Dissertations.
Effect of Age on 11to 18-Y ear-Olds Discrimination of Nuances in Instrumental and Speech Phrase Interpretations by Andrew Sioberg A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy School of Music College of Visual and Performing Arts University of South Florida Major Professor: Jack Heller, Ph.D. C. Victor Fung, Ph.D. David Williams, Ph.D. Sheila Woodward, Ph.D. Date of Approval: April 13, 2005 Keywords: cognition, interpretation, perception, listening, nuance Copyright 2005 Andrew Sioberg
i Table of Contents List of Tables iii List of Figures iv Abstract v Chapter One Introduction 1 Background of the Problem 1 Purpose 5 Problem 5 Hypothesis 6 Chapter Two Review of Literature 7 Perception and Cognition Theories 7 Summation 16 Developmental Theories in Music Per ception and Cognition 19 Meter and Tempo Studies 25 Pitch Development 26 Auditory Perception 27 Rhythm 27 Aptitude 28 Interpretation 29 Music and Language 32 Chapter Three Methods 36 Problem 36 Development of the Items 36 Pilot One 40 Pilot Two 41 Study Population and Sample 45 Variables 46 Measures 46 Data Analysis 46 Chapter Four Results 47 Assumptions 47 Overall Test Results 49 Speech and Music Part Results 51 Chapter Five Discussion 58 Limitations 61 Future Study 62
ii References 64 Appendices 70 Appendix A: Percentage of Students who Answered Correctly Each Item of the Five Speech Categories 71 Appendix B: Percentage of Students w ho Answered Correctly Each Item of the Five Nonsense Categories 72 Appendix C: Percentage of Students who Answered Correctly Each Item of the Five Musical Except Categories 73 Appendix D: Percentage of Students w ho Answered Correctly Each Item of the Five Composed Melodies Categories 74 Appendix E: Pilot One: Descriptive Statistics 75 About the Author End Page
iii List of Tables Table 1 Pilot 1 Descriptive Statistics, Cronbachs alpha, a nd Pre/Post Pearson Product-Moment Correlation 43 Table 2 Pilot 2: 6 th 8 th 10-12 th Grade Descriptive Statistics 44 Table 3 One-Way ANOVA Summary Table of 6 th 8 th and 10-12 th Grade 45 Table 4 Age 11, 12, 13, 14, 15, 16, & 17-18 Overall Test Score ANOVA Descriptive Statistics 48 Table 5 7 x 2 x 2 Factorial ANOVA Su mmary Table Investigating the Relationship Between Age, Gender, and Prior Lesson Experience on Students Ability to Interpret Music and Speech Phrase Changes 50 Table 6 Age 11, 12, 13, 14, 15, 16, & 17-18 Speech Items ANOVA Descriptive Statistics 52 Table 7 Age 11, 12, 13, 14, 15, 16, & 17-18 Music Items ANOVA Descriptive Statistics 53 Table 8 7 x 2 x 2 Factorial ANOVA Su mmary Table Investigating the Relationship Between Age, Gender, and Prior Lesson Experience on Students Ability to Interpret Speech Phrase Changes 53 Table 9 7 x 2 x 2 Factorial ANOVA Su mmary Table Investigating the Relationship Between Age, Gender, and Prior Lesson Experience on Students Ability to Interpret Music Phrase Changes 55
iv List of Figures Figure 1 Overall Lesson Score Means by Age Group in Percentages 51 Figure 2 Speech Item Score Means for Lesson Experience Arranged by Age Group in Percentages 54 Figure 3 Music Item Score Means fo r Lesson Experience Arranged by Age Group in Percentages 56
v Effect of Age on 11to 18-Y ear-Olds Discrimination of Nuances in Instrumental and Speech Phrase Interpretations Andrew Sioberg ABSTRACT This dissertation was a con tinuation of study on a theory of a learning window for the perception of expressi ve qualities in music and speech. The proposed theory suggested that a practice window must overlap a learning win dow before it closes around the age of 10. This disserta tion sought to determine whet her children older than the proposed learning window continued to impr ove in speech and musical discrimination skill, or leveled off in this ability. It also examined the impact of gender and private lesson experience on discrimination ability. Instrumental music students ( n = 292) attending a public magnet school for visual and performing arts in North Carolina between the ages of 11 and 18 participated in the study. Each student was administered a forty-item listening test containing 20 speech items and 20 instrumental music items. Each test item consisted of three short speech or musical phrases. All three phrases in each item were the same written words or notated music, but one phrase was different in interpretation or expression from the other two phrases. Two of the phrases were intended by the performers to be the same in interpretation or expression and one was inte nded to be different in interpretation or
vi expression. Subjects were asked to determine which of the three phases in each item was different in interpretation or expression from the other two. Results of the study suggeste d that students with prior private lesson experience scored significantly higher than those students that had neve r taken private lessons. This study seemed to reinforce the proposed learning window for speech and music interpretation in that interpretation ability did appear to level off.
Chapter One Introduction Background of the Problem This study was designed to further test a theory of a learning window for the discrimination of expressive qualities in music and speech. Learning window theory is an established belief that specific time spans occur in a learners life when certain skills (in this case speech and music syntax) are easie st and most conducive to acquisition (Flavell, 1965). Campbell and Heller (1972) proposed a theory that a pract ice window must overlap this learning window before it closes around the age of 10. If students are given opportunities to practice music and speech befo re the closing of this window, children seem to learn the necessary implicit rules for music inte rpretation and language prosody. Validation of this theory reinforces the importance of music early in life for all children. If the music percep tion/learning window operates at a young age, exposure to music within that time period may be critic al for students to de velop related musical interpretation contingent with life-long mu sical learning and appr eciation. If learning implicit rules to music must occur at an early age for children to develop the ability to perform culturally appropriate music throughout lif e, it follows that other genres of music deemed of value must also be incorporated much earlier in music education. This may lead to more rigorous study earlier in stude nts development of jazz, world musics, and other genres deemed valuable by the profe ssion to assure the highest assimilation of implicit rules to those, and ot her, valued forms of music.
Research in the field of language studies has produced evidence supporting the theory of a learning window being in place for acquisition of language syntax. Nash (1997) stated that the window for acquiring syntax may close as early as five or six years of age, but the window for learning new words never closes. According to Nash, the ability to learn a second language was highest from birth to six years of age, and declined from there. While it is possible for people to learn language afte r the learning window closes, it is only at great di fficulty. Neuroscientists (see Nash) believed the brains greatest development closed by the age of 10. Nash recommended foreign languages be taught in elementary school or earlier As with different spoken languages, diffe rent musical language s require learners to know a specific set of imp licit rules to properly perform them. Campbell and Heller (1972) suggested this set of implicit rules for music was acquired at the same time syntax and the implicit rules for language were acquired. In 1982 Heller, Campbell, and Gibson demonstrated that this ability seemed to level off by age ten. According to Heller and Athanasulis ( 2002), whatever the musical genre or spoken language, the performer drew from specif ic, implicit stylistic or syntactic rules to deliver the performance in a manner deemed cu lturally appropriate. In both instances, how we perform the musical or spoken phrase was just as important as what was said or what was played. According to the resear chers, these implicit rules were not directly observable, and were therefore not easily de fined. However, their presence could be inferred from the resulting musical and sp eech nuances performers made, given the cornucopia of variation possibl e in the delivery of each musical or speech phrase. With regard to both music and speech delivery, va riation could be observed in the pitch,
timbre, rhythm, dynamics, meter, and timing of a phrase, among others. These possibilities could also occur in any combination and at any point of a musical or spoken phrase. In speech, these subtle variations we re called supra-segmentals. In music, the variations were often referred to as nuan ces. In both speech and music, Heller and Athanasulis (2002) believed the development of this discrimination skill took place at an early age. In addition to the complexity of the continuous variation in music and speech, Heller and Campbell (1981) argued that trad itional research methodology, isolating and observing a single variable, to investigate th ese implicit rules produced an artificial and inaccurate representation of the music or sp eech phrase. Their position was that music nuance and speech supra-segmentals were c ontext-dependant and could not be observed independent of the entire, unalte red, audible performance. The theory Heller and Campbell (1981) s upported moved away from the concept of music ability as a special talent, and move d towards the idea that music was an activity all humans were capable of achieving, given ac cess to a musically rich environment by a certain age. In addition, they proposed that perceptual development in music had similar characteristics to perceptual developm ent in spoken language. Music, like spoken language, involved a system by which multiple cues were implemented by the performer to evoke covert responses in the listener. Both were an active process involving the individual interacting with the environment, refining patterns and prominent features. Recognition of invariance to both music and la nguage was significant to development of perception; however, perceptual invariance di d not imply stimulus invariance (see Carl
Seashores normal illusions, 1938). These normal illusions also put significance on cultural conventions and their importa nce in perceptual development. In an effort to investigate this theo ry, Heller and Athanasu lis (2002) developed a 30-item speech test and a 30-item music test and administered it to 1st3rdand 5th graders. The speech test contained short phr ases spoken by a child, woman, and man. The music portion contained mel odic phrases performed by a clarinet, cello, and piano. In both tests, each item containe d three short phrases, with one phrase being different in interpretation from the other two. The task for the listener was to identify which of the three phrases was different in interpretation. The study found that stude nts were better at this task the higher the grade they were in, which does not sup port the learning window theory. The expectation was that fifth grader s ability (students about the age of 11) would not be higher than third graders (about the age of 9). In another related study, Heller (2003 ) investigated the ability of 66 undergraduate non-music majors to take simila r tests and compared their results with their musical aptitude scores based on the Advanced Measures of Music Audiation (Gordon, 1989). Results showed there was a si gnificant difference on the music items between the subjects who scored high and low in music aptitude. There was also a significant difference on the speech items between high and low music aptitude subjects. Those subjects scoring high in music aptitude scored high on both music and speech tests. The current study sought to continue to investigate the model of a learning/practice window from bi rth to about the age ten in order to further examine the viability of the theory. This dissertation exam ined the ability of students in the sixth to
twelfth grades (ages 11-18) to take a sp eech and music discrimination test. Continued improvement over time on this tests scores ma y suggest that the theory needs revision. Sustained scores over the grade levels w ould reinforce the theo ry. If there was no apparent growth in student discrimination ab ilities from grades six through twelve, a reasonable conclusion might be that the skills tested have already been developed by the sixth grade. Purpose The purpose of this study, then, was to ex amine the ability of children older than the proposed critical learning window. It sought to answer the question: Will students in grades 6-12 improve their ability to discriminate changes in speech and music phrase interpretations? Problem In order to address the stat ed purpose, the following questions must be addressed: 1. Can children older than the proposed lim its of the learning window demonstrate significantly higher scores in di scriminating differences in cult urally valid interpretations intended by performers competent in speech and music as age increases? 2. Is there a difference in spoken phrase a nd instrumental music phrase discrimination ability across age groups?
3. What effect does gender have on the ability of students to discri minate in music and speech phrases? 4. What effect does a students private music lesson experience have on his/her ability to discriminate changes in music and speech phrases? Hypothesis The following null hypotheses were esta blished for statis tical testing: 1. There were no significant differences between the mean scores on the spoken discrimination tasks across age. 2. There were no significant differences between the mean scores on the musical discrimination tasks across age. 3. There were no significant differences betw een male and female test scores on the discrimination test across age groups. 4. There were no significant differences between students that have taken private lessons with those students who have not taken privat e lessons in their te st scores across age groups. 5. There were no interactions.
Chapter Two Review of Literature There is abundant literature in the fi eld related to the proposed study. Scholars have been writing about the process of human perception and cognition since the beginning of the 1900s. Many theories and c oncepts have been proposed, revised, and developed in that time. A number of ar ticles have examined how perception and cognition are developed throughout the lifespa n of humans. Researchers have also examined how perception and cognition pertain to specific elements found in music such as pitch perception, rhythm, and aptitude, am ong others. Work has also been conducted on the relationship between music and speech. Perception and Cognition Theories Psychological research and discussion on the subject of music perception have appeared in the research literature for over a century. Throughout th e 1900s and into the 21 st century, a variety of theories and models have been proposed addressing the process of music perception. Notable theories and models were articulated by Seashore (1919) and his copy paradigm; Meyer (1956) a nd his expectancy theory; Moles (1958) and information theory; Serafines (1983) music as thought and her construction paradigm; Fiske (1996) and his metalanguage; and Gardners (1983) multiple intelligences. These theories have changed dramatically over th e years as continuing research has improved and refined the understanding of the process of music perception. It remains a core issue in music study since an operational definiti on and understanding of music perception and
cognition is deemed essential to music e ducation, pedagogy, and curriculum design. This chapter will examine significant contributions of the development of perception and cognition theory. Carl Seashore (1919) developed a theo ry stating that musical intelligence functions independently of general intel ligence. His work was grounded on what has been called copy paradigm, (Fiske, 1996) a beli ef that anything perceived, regardless of the senses involved, is directly related to meas urements of an actual, real, physical object. ...everything that is rendered as music or heard as music may be expressed in terms of the concepts of the sound wave (Seashore, 1938, p. 2). This meant that the emotional and aesthetic qualities found in music listening could be lo cated within the sound wave itself. In response to this logic, Seashore focused on carefully anal yzing and describing the sound wave and its acoustic properties believing he was also simultaneously discussing the structure of music perception. Out of this theory, Seashore developed a sophisticated series of tests, the Measures of Musical Talents to measure an individuals inherent musical abilities regarding the sense of tone qua lity, sense of consonance, sense of volume, and sense of rhythm to correspond with the four psychological attributes of sound: pitch, loudness, time, and timbre (Seashore, 1938, p. 2). Success in music was contingent to natural capacities of each ability in each individual, and this could be considered a level of musical aptitude. He believed musical tale nt was inborn to each individual (Seashore, 1919, p. 6). The theory was misleading, however when discrepancies occurred in the data collected between the listeners perception and actual acoustic measurement. They were described away as normal illusions because most people perceived the event in
the same manner (Seashore, 1938, p. 17). In addition, copy theory inferred that all listeners have the same internal representation of the music they are listening to. If listeners differed in internal representation of the music, one must be wrong because there could be only one correct representation. While Seashores work carefully described the acoustic signal, it did not accurately re present the perception of the listener. Meyer (1956) proposed an important theo ry regarding expectation in music. During the course of a musical work, listeners created on-going expectations about what particular tonal-rhythmic events were likel y to occur next in the composition. Those expectations were based on events that have already taken place in the music. Expectation then is a product of the hab it responses developed in connection with particular musical styles and of the mode s of human perception, cognition, and response the psychological laws of mental life (Meyer, 1956, p. 30). Based on past experience, if a stimulus in the present led the listener to expect a future consequent musical event, that stimulus then had meaning. According to Meyer, those stimuli that did not bring expectation had no meaning. Fiske (1996) believed there was a problem w ith testing expectancy theory because it did not offer a method for specifying differe nt degrees of expectancy strength either between musical events or for repeated listeni ngs. A partial solution to this problem was an expectancy profile offered by Carlsen (1982). A listeners expectancy profile hypothesized more than one pattern progressi on which may occur next in the piece. Fiske (1996) suggested measuring the specific probability of expected events and relative strength of their associated emotiona l arousals would produce better results.
Developing out of Meyers expectancy theory, information theory provided a statistical theory towards th e development of an understand ing of music perception. Fiske (1996) described information theory as the resu lt of a refined sensitivity to the relative probability of particular (musical) events and their future occurrence. Tonal-rhythmic messages sent by the performer to the liste ner resulted in meaning, but not because meaning existed in the message. Instead, mean ing was realized by the listener according to a set of conventions agreed upon by the performer and the listener (Fiske, 1996; Heller and Campbell, 1981). More specifically, Mo les (1958) attributed the weakness of studying musical acoustics as its lack of addressing the real problems raised by the creation of sonic structures: ...acoustics studied the r ubbing of the bow on the string, when the only thing that inte rested the musician was the tone this string produced (Moles 1958, p. 105). Moles (1958) proposed that information th eory defined the complexity of the message being sent from the performer to th e listener as the degree of originality. The more original the material in the message, the more complexity it contained. In the theory, original or unexpected events provided information, where as expected or previously heard events provided no information. However, there is no art without constrai nt. To say that music is an art is to say that it obeys rules. Pure chance represents total liberty, and the word construct means precisely to revolt agai nst chance. An art is exactly defined by the set of rules it follows. The role of esthetics... is to enumerate these rules and link them with universal laws of perception (Moles, 1958, p. 105).
Davies (1978) suggested listeners un derstand music, or music has meaning because it confirms or disconf irms certain expectations, which supported Fiskes (1996) discussion of information theory. Expectations and pred ictions depended on the organization of musical material and the know ledge a person had about the organization of the material. (Davies, 1978). Neisser (1976) suggested that the listener is constantly developing anticipations of what will come next, based on the information that has already been picked up. It was these anticipations formulated in temporal patterns that governed what would be picked up next, and were then modified by it. The anticipations were not highly specific, and were rarely definite, only guiding the liste ner in a general dir ection. Neisser defined cognition as the activity of knowing: the acquisition, organization, and use of knowledge (Neisser, 1976, p. 1). Serafines theory of music-as-t hought saught to s how that musical comprehension was a result of active cognitive construction and not a passive observation of musical structures. She fla tly rejected the idea that m eaning was in sound along with any theory suggesting that th e analysis of sound would produce musical meaning (Fiske, 1996). To say that the composer inverted the theme or the chord sequence modulated to V is to say that such themes and chords will be heard as inverted or modulated. The results of formal analysis should be consistent with cognitive reality, or they fail as explanations of art. (Serafine, 1983, p. 121)
According to Campbell (1991), Serafine argued that music comprehension required problem-solving strategies involving higher-order cognitive processes not unlike that of chess or mathematics. Serafi ne also thought mu sic study contained an emphasis on organization in a temporal context and not th e physical entities of sound. Field definition, which accounts for music characteristics, temporal organization (event to event processes), and nontemporal operations (formal characteristics of a work) are three cognitive processes she suggested were involved in music comprehension. Serafine also took a strong opposition to the assumption that musical elements were found in nature. She believed that scales, chords, and music theory in general had nothing to do with music cognition. Instead, they presented a system with which to order and describe music. Fiske observed that Serafine took the position of construction paradigm that musical understanding was: ...a product of pattern generating/recep tion processes rather than an aural copy of information contained in so und objects. Her theory emphasizes cognitive processes; perceived musical structures as a by-product of musical thinking; the shared cogni tive processes of the composer, performer, and listener; and the non-communicative function of music (Fiske, 1996, 51). Serafine believed that psychological research had been too narrowly focused on the artifacts of analysis than with music. Sh e believed music study mu st contain an emphasis on organization in a temporal context and not the physical entities of sound (Serafine, 1983).
Taking this constructivist theory even fu rther, Lerdahl and Jackendoff developed a theory that the listeners mind constructe d music entirely. Rhythm, beat, and melody were derived from a physical symbol (the sound). The listeners musical experience or intuition was the catalyst for the organization of the auditory signal (Lerdahl and Jackendoff, 1983). Studies by Krumhansl (1983) suggested that single tones within a tonal context were interpreted by the listener according to their function and or ganization. In addition, chords which were built with related ha rmonic and melodic information helped to develop the concept of tonality in the listener. Listeners a pparently made reference to knowledge of the regularities underlying music w ithin their experience to interpret the incoming sensory information. Fiske developed the metalanguage theo ry which addressed the issue of differentiating music sounds as different from other sonic events and also took into account the multifarious musical languages found around the world. According to Fiske (1996), music cognition was a three-stage pro cess. In the first stage, the auditory processing system received any and all incoming acoustic signals and did not yet distinguish music from any other sound. At th is point there was onl y the differentiation between speech and non-speech. The second stage involved a cogni tive decision-making activity where pitch-durational patterns were formed and compared to conclude what was being listened to (music, noise, speech). The third stage was the representation of patterns through encoded features for storage, recall, and recognition. Before a distinction could be made between music and speech, one must have been made between vocal and non-vocal sound through a micro-level temporal
psychoacoustic identification of timbre. Fiske pl aced this in stage one of his model. He postulated a dual-track processing system in stage two for pattern formation and identification of speech and non-speech s ounds using their tonal-rhythmic pattern processing hierarchies. Stage three made stipulations regarding speech-intended vocal sounds and nonspeech-intended vocal sounds and whether they were music or not using the individuals culturally in fluenced beliefs and experiences. In addition, it made a speech-music stipulation to account for unaccompanied song. For his theory to work, a number of assumptions were made. Fiske (1996) assumed music perception and cognition wa s a constructed process and not a copy process. With regard to the musical process, music was a part of all known cultures; was a uniquely human activity; occurred in a wide va riety of styles and genres that tended to change over time. Musical activity was a bout the generation a nd reception of tonalrhythmic patterns of sound, and music per ception and comprehension required time and effort. Finally, music comprehension was the result of an implicitly known set of cognitive processing rules which could have been either style-specific acquired by experience, or panstylistic inhe rent in the listening system. Elliotts (1995) writings supported the work of Fiske and Serafine regarding music cognition and perception, although it wa s discussed under his definition of listening. He suggested that the act of lis tening was both constructive and cognitive. A listeners internal organization and experience of music invol ved several kinds of musical thinking and knowing because the sonic mate rials being heard were always practicespecific and culture-specific constructions. In addition, music listeni ng was not restricting brain activity to just sonic stimuli, b ecause music was always context-dependent.
Elliott proposed the following: The procedural essence of music lis tening consists in such covert, nonverbal acts as constructing coherent musical patterns, chaining musical patterns together, making same-diffe rent comparisons among and between patterns, and parsing musical patterns into different types of textures. (Elliott, 1995, 85) It is important to note that Elliotts use of the word covert was hi s attempt at describing implicit or internal actions taken by the listener. Heller and Campbell (1981) rejected the idea that music contained meaning. Instead, they assumed a construction paradigm which argued that listeners construct their own perception (assign their own mean ing) of sound patterns in the brain. Heller and Campbells theory could be orga nized into a number of features. First, the performer produced a sound which provided th e listener with raw acoustic cues. Next, those cues were analyzed for micro-level patt erns which were linked together into longer patterns that eventually constituted musical elements. In the third, pattern identification was guided by an active, context-dependant, a nd implicit set of rules by which both the performer and the listener abide. The rules bo th parties follow were formulated through a social/cultural contract. Like the comprehension of different spoken languages, different musical languages required the participan ts to know different sets of rules to successfully interpret them. This musical percep tual ability must be acquired early in an individuals development. Heller and Campbell suggested that for discriminating interpretation in music nuance, this acquisition o ccurred by the 10 th year of life. Their theory stated that language perception for prosody was also developed by the 10 th year.
Critical periods cannot be rec overed; lack of appropriate mu sical experiences at the time during which the brain is most prepared to ac quire music social/cultural contracts results in a musically handicapped adult (Fiske, 1990, p. 7). Heller and Campbells theory sought to a void the concept of musical listening ability as a special talent, and instead pres ented a case to see music listening skill as an activity all humans are capable of achieving, provided they are exposed to a musically rich environment by a certain critical time period. If a listener missed this opportunity within the critical tim e period, music listening skills and implicit rules to musical interpretation could only be developed with great difficulty (or not at all) in a later period (Campbell and Heller, 1981). The theory proposed that perceptual development in music had similar characteristics to perceptual development in language. Both were an active process involving the individual interact ing with his or her environm ent, refining patterns and prominent features. Recognition of invarian ce in both music and language was significant to perception. However, per ceptual invariance did not imply stimulus invariance (Seashores normal illusions, 1937). Helle r and Campbell also put significance on cultural conventions and their importa nce in perceptual development. Summation Seashores work focused on the analysis of the sound wave and the components which comprised it. No matter how careful the analysis was or how specific the information was describing the sound wave, this method of study did not adequately address the emotional or aesthetic qual ities each person derives from a musical experience. An individuals response to mu sic is not measured by components of the
sound wave. Seashores work helped to contri bute to the study of music perception and cognition, but his theory was not di rectly applicable to this study. Meyers proposal that the listener created continuously evolving expectations as music was performed comes closer to a func tional address of per ception and cognition. The theory falls short, however, when Meye r suggested that stimuli which do not bring expectation (those musical events which ha ve not been experienced before) have no meaning. Artful music created over the cen turies has continuously challenged conventions of the time. Composers have ski llfully and intentionall y crafted their music in such a way as to delay or even defy e xpectation specifically to arouse meaning in the listener. At the same time, this position does not imply that music with little or no adherence to the listeners conventions would somehow have more meaning. Moles developed a similar position by suggesting that original or unexpected events in music provided information whereas previously hear d material provided no new information. If this were the case, there would be no need fo r previously experienced musical events in a composition. An important component of meaningful mu sic, then, is a careful balance of new material intertwined with the known, traditional conventions of the lis tener. This delicate balance of interest (new mate rial) versus clarity (known ma terial) continuously evolves as the piece progresses, for as new material is introduced, it loses its novelty with time. Neisser (1976) and Serafine (1982) both allude to this point. In music, original events and those events conforming to convention should be perceived as equally important in the listeners perception and cognition process. Music cannot be unde rstood without both.
Serafine (1983), and Lerdahl and Jacke ndoff (1983) provided another important component to music perception and cogniti on by proposing that meaning in music was constructed by each listener according to his/her individual listening experiences. The listeners construct, however, must be guided by some supervisory, decision-making process. Conventions, as discussed by of a number of scholars (M oles, 1958; Fiske, 1996; Heller and Campbell, 1981) were developed ove r the life experience of the listener and performer. Heller and Campbells (1981) pr oposal of culture was the most logical determinant for assembling this convention. Fiske (1996) enhanced this influence of culture on music perception and cognition by al so allowing for pro cessing rules to be inherent in the human auditory system. Hints of the auditory system having an impact on the listening process can be seen as fa r back as Seashores (1938) work and his identification of inhere nt musical abilities. Heller and Campbell (1981) took the music learning process one step further by seeking to match its cognition processes to th at of the speech learning process. Their theory proposed that musical interpretative ability was developed about the same time speech interpretative ability was obtained. Furthermore, both skills were acquired the same way; in an active process employing a recognition of invarian ce. The current study was developed to further examine the validity of this proposed theory.
Developmental Theories in Mu sic Perception and Cognition Scholars have long been examining the developmental processes of perception and cognition. Understanding when these skills are acquired is cri tically important to developing sound, productive educational pedagogy. The cognitive developmental psychology of Piaget is well respected and has had significant impact on music education (Pfled erer, 1967; Gardner, 1983; Campbell, 1991; Zimmerman, 1992 and 1986). In his theory, schemata continuously unfold in systematic stages as a result of environmental intera ction. Development is continuous and builds on that which has already been experienced. Age levels at which th e changes occur vary depending on each individuals cultural physical, and social environments. Piaget carefully described the progressive development of the learner as a series of stages. In the first eighteen months of life, intellectual pr ocesses are initially developed through predominantly sensorimotor behavior. In the second, or preoperational representation stage, a child undergoes a prec onceptual phase and an intuitive phase from about eighteen months to seven or eight years of age. Here the child reasons transducively, still relying heavily on perception for t hought. Failure of transducive reasoning causes intuit ive regulations and the formation of categories. Continuous and systematic variations in the childs focus of attention provide more impressions of reality in a move towards operational thought. By the age of seven, concepts acquired to this point are organized into coherent systems. Thought is no longer re liant on perception. In the formal operations stage from 12 to 15 the child develops reasoning which culminates in reversibility, the formati on of a thought which can then be unthought (Flavell, 1965).
A number of researchers have applied Piagets developmental theory to music and arts. Wolf and Gardner (1980) proposed four stages of artistic development corresponding to Piagets stages of cognitive development: (1) child as direct communicator, (2) child as sym bol user, (3) youth as craftsman (covers 5 or 7 to 11 or 13), and (4) youth as critic and full pa rticipant in the ar tistic process. Pflederer (1967) proposed th at processes involved in musical intelligence stem from perceiving, comprehending, and organizing the structure of mu sic. Music learning begins with perception and the sound structure, followed by musical concepts which permit one to think about what has been hear d. The essence of musical intelligence is found in the framework of rhythmic, melodic, harmonic, and formal relationships which have developed through a progressive organiza tion of musical experiences (Pfelderer, 1967, p. 221). Pflederer cited five conservation laws in the development of musical concepts. Identity applied to the maintaining of them atic materials over time. Metrical grouping represented an active organi zation of tonal stimuli around accents. Augmentation and diminution involved the presentation of a mu sical subject in doubled and halved values which requires a relational framework base d on Piagets reversible operation. The invariance of tonal organization was a cons ervation law employing transposition. Lastly, inversion of either melody or harmony used the reversible operation, a principle of conservation. Zimmerman (1982) suggested that proce sses involved in musical intelligence stem from perceiving, comprehending, and orga nizing the structure of music. Assessing comprehension was accomplished by observing th e cognitive behavior of the listener.
Her discussion was very closely related to th e work of Piaget. According to Zimmerman, learning processes begin with immediate and limited perception due to the limited framework for music in the child. Experience move s immediate perception to mediate perception, or a change from direct observation and stimulat ion to perception implied by or derived through something else. The initial precept move s to the emerging concept. Expansion of experience paralleled expansi on of cognitive framework. Concepts became interlocking and relational, allowing for advanced cogni tion such as transposition, modulation, inversion, augu mentation, and dimunition. Slobodas (1985) views also seem to parallel Piagets work. He proposed that human music skill was built out of a base of innate abilities and tendencies, and that the physical world and a persons experience in it develop this skill. He divided his discussion of musical learning and development into two main categories: enculturation and training. With regard to enculturation, there exists a set of inherent capac ities present in all humans at birth. There is also a set of shared experiences culture provides children as they grow. Finally, there is the impact of the rapidly changing general cognitive system as culture-supported skills are learned. These factors produced a relatively similar sequence of development for most children with in a culture, and a set of similar ages in which growth takes place. In addition, en culturation is achieved largely without a conscious effort to learn. Training, on the other hand, is a self -conscious effort to become more accomplished at a specific skill using explic it instruction. Individuals focus on specific experiences not shared by all members of a culture. While enculturation is the dominant
process of learning up to the age of 10, training takes an increasing importance following the age of 10. (Sloboda, 1985). Serafine (1988) proposed a theory of music cognition suggesting that musical thought is a result of the pos ing of artworks embodying fin ite and organized sets of temporal events described in sound. Temporal organization (succession and simultaneity), and nontemporal operations (c losure, transformati on, abstraction, and hierarchic levels) were defined as two categories of cognitive processes indicative of how music is understood. Sound is processed accordin g to a hierarchy of structural elements. She tested her theory using a series of te mporal and nontemporal studies incorporating musical tasks given to 15 subjects each at ages 5, 6, 8, 10, 11, and adult. Results indicated that tenand eleven-year-olds are in possession of most temporal and nontemporal processes. Intere stingly, eleven-year-olds showed a slight decrease in the pitch discrimination task, idiomatic construction task, and the motivic synthesis task. Serafine was expecting con tinued increases in performance throughout. Eight-year-olds could perceive hierarchic leve ls in simple melodies, easily identify simultaneous combinations of timbres, and di scriminate random melodies on a similar level to older students. Most st udents had difficulty with the rest of the tests in this age group. Because fiveand six-ye ar-olds demonstrated almost none of the processes, Serafine surmised that the years from eight to ten represent a period of rapid growth in music cognition. It was also noted that subject s with instrumental training fared no better than those subjects with no tr aining. Her studies suggested that nontemporal, formal, and abstract processes appeared to devel op earlier than temporal processes.
Gardners (1983) work suggested there exists several relatively autonomous human intellectual competences he referred to as human intelligences, one of which he identified as musical intelligence. He proposed that musical intelligence uses three systems of knowing to address symbols (m aking, feeling, and perc eiving). Symbols in music can be either denotati onal, expressive, or both. According to Gardner, from birth to 57 years of age the child develops from responding instinctively to relating actively to the environment through making, feeling, and perceiving representations. Instinctive refl exes evolve into symbol use and the three systems differentiate. Infants sing an d babble, produce undula ting patterns, emit individual sounds, and imitate prosodic patter ns and tones sung by others. Over time, the childs spontaneous vocaliza tions develop into echoing a nd reproducing song fragments. At the end of this stage, children can be heard inventing songs and singing songs from their culture. From age five to ten, elements of different symbol systems become linked to specific activities within domains. Child ren focus on learning rules, and have an interest in developing tec hnique. School music expands their song repertoire and knowledge base. After ten, children demonstrate continued skill development and integration of the three systems (making, f eeling, and perceiving). Critical acumen and critical reflection also develop, providing new insight towards style and music interpretation along with evaluation skills. Gardners work contributes much to the theoretical cognitive processes involved in human development and music. It serves as a guide for developing curricula and musica l activities for the developing student. (Campbell, 1991).
It seems there is consensus in the aca demic music community that cognitive development is continuous, compounding, and moves through stages. Many of the music scholars discussed have been clearly influenced by the work of Piaget. The theories offered here suggest how and when music concep ts are presented is very important to the success of students acquisiti on and retention of those con cepts. In a number of these theories, (Flavell, 1965; Sloboda, 1985; Sera fine, 1988; and Gardne r, 1983) scholars propose a shift in cognitive development does occur around the age of 10-11. Flavell (1965) in his discussion of Pi agets work proposed the de veloping mind transitions from perception-reliant cognition to reasoni ng and reversibility. Sloboda (1985) saw a transition around this time period from l earning unconsciously through enculturation to learning through formal traini ng. Serafines (1988) resear ch suggested most of the temporal and nontemporal processes were de veloped by this same age. Gardner (1983) suggested advanced integrati on of the three systems (making, feeling, and perceiving) along with critical acumen and critical reflecti on developed at this same stage. In the proposed theory to be tested in this study, language perception for prosody and music perception for nuance have been developed by this same time period. Following this time period it is possible that interpretative skill can be acquired, but only at great difficulty and perhaps never to the level of a person w ho acquired it at an earli er, more appropriate age. A number of studies have been cond ucted over the years examining different specific music skills and thei r development through different age groups of subjects. The following is a review of selected literature organized by music skill. While these studies are related to the present research in that they are trying to deve lop a better understanding
of music cognition and perception, the appro ach of isolating specific components of music for observation and study may not be the best method to observe these developments. The current study seeks to leave the entire music and speech phrases intact as they were originally reco rded and observe childrens ab ility to interpret changes in them. It is thought that keeping the musical and speech excerpts in their original, complex forms makes the task observed more realistic to musical listening a nd more generalizable to everyday musical activity. Meter and tempo studies To investigate the development of childrens ability to conceptualize and identify meter in music, Jones (1976) had 66 children aged 6-12 years perform 11 musical tasks in increasingly difficult order. His results suppor ted Piagets theory that childrens concept of time develops through three stages. The first stage was characterized by basic perceptual discriminations. The second stage, beginning after seven years of age, was characterized by a conservation of velocity, im mediate seriation, and some difficulty with double seriation. Jones defined seriation following the work of Piaget as the concrete operation that involves orderi ng stimuli along a quantitative dimension (such as length). Double seriation was an understa nding of symbols or objects that involve more than one meaning. By about age 9, there was immediat e double seriation, a grasp of simultaneity and succession, operation inclusion, and m easurement of physical time. The meter concept in Jones study developed between the ages of 9 and 10 with many children between 9 and 12 having some difficulty understanding meter c oncept and meter. As a follow up, Perney (1976) administer ed Tasks I-V from Jones study to second and third graders. Perney found th ere was no significant difference between
students who played instruments and those w ho did not, and females faired better than males in performing tasks. In addition, there was a relationship between performance on musical tasks and verbal ability for both sexes. Miller and Eargle (1990) examined th e effects of age and musical background independently on simple tempo discriminati on. The three age groups represented were middle childhood (7-11), early adolescence (1215), and adulthood (18-33). They found that subjects perception of changes in tempo during a piece was more sensitive to general developmental differences than mu sical training. However, acknowledging the maintenance of a standard tempo seemed mo re affected by musical training than age. Pitch development Davidson and Scripp (1985) asked children aged five to seven in a longitudinal study to write down a song they knew so that so meone else could sing it in an effort to capture evidence of those musical features they thought were most important in remembering how a tune goes. In those pict orial representations, they found development trends in childrens use of symbols, leve l of sophistications of their notations, and selective attention to musical features depe nding on the context. As the representations became more sophisticated, pitch emerges as the most robust component of cognitive musical development. The researchers specu lated that as pitch development occurred with age, it increasingly became independent of language, kinaesthetic, or number skills. They recommended more focus on the success of musical understandi ng children have of music through creative symbolic representation.
Auditory perception An investigation of a learning sequence fo r music listening skills associated with the detection of alte rations in timbre, rhythm, melodi c pitch patterns, and harmony was conducted by Hufstader (1977). The project ex amined almost 600 first, third, fifth, and seventh grade students from four school district s and their ability to take a test of aural perception skills designed by the researcher. Timbre mean scores reached an arbitrary criterion level for all four di stricts initially by fi rst grade. Rhythm was second at fifth grade for all four districts. Melodic pitch was achieved by two district s at fifth grade, and the other two at seventh grade. Harmony mean scores only reached the criterion level in two of the districts. This suggested an order to the development of specific listening skills used in auditory perception. Petzold (1963) investigated the deve lopment of auditory perception, phrase learning, and melodic reproduc tion using varying harmonies, timbres, and rhythmic ability. He found that age was a major factor in the development of auditory perception. For most tasks, a plateau was reached by the age of eight, and the most development occurs between the ages of six and seven Rhythm Cox (1977) examined almost 800 studen ts in grades 1-6 from 48 different classrooms in 8 schools to develop a descriptive analysis of their responses to beat, meter, and rhythm pattern. The results of her beat, me ter, and rhythm pattern tests indicated that music concepts develop gradually, rhythmic ab ility increases continuously with age, and the concept of meter does not matu re before the age of 9 to 9.
Atterburys study (1983) invol ved ten rhythm tests divi ded into three categories given to 20 seven-year-olds and 20 eight-year-o lds, half were normal achieving readers, and half were learning disabled readers. Results indicated that learning-disabled readers perceived simple same and different rhythm patterns simila rly to normal readers, but were less successful at difficult rhythm pattern s. Learning-disabled readers also could not reproduce rhythm patterns as well as norma l readers. In addition, tapped and spoken rhythms were significantly different than melodic or tapped for all subjects. Atterbury suggested including rhythm syllables when teaching rhythmic patterns to sevenand eight-year olds. Palmer (1990) engaged fourth-grade st udents from three elementary schools to examine the effectiveness of the methods of Mary Helen Richards and Edwin Gordon compared to a control group. The study s howed that both Richards and Gordons methods produced significant gains in rhyt hm reading over the control group, with Gordons method showing slightly better results than Richards. Shehan (1987) examined the effects of aural and visual approaches to rhythm reading and short-term retention with 2nd and 6th graders. She found that the simultaneous use of auditory and visual cha nnels facilitated learning of rhythms in both groups. More importantly, however, the older gr oup was able to learn rhythms twice as fast as the younger group re gardless of the presentati on method. In this study, age improved the musical skill of rhythm reading and short term retention. Aptitude Using Gordons Primary Measures of Music Audiation (PMMA), Flohr (1981) examined the influence of short-term mu sic instruction on five-year-old childrens
developmental music aptitude. Twenty-nin e children were pla ced in three groups consisting of improvisation instruction; si nging, playing percussion instruments, and movement; and no music instruction. Results su ggested students receiving either form of musical instruction scored significantly high er on the PMMA than the control group. De Yarman (1975) conducted a study administering the Musical Aptitude Profile for four years to fourth graders with varying training and found that th e type or amount of formal musical training had little effect upon ch ildrens musical aptitude prior to fourth grade. In addition, musical aptitude stabilized before age five or six and was resistant to further training. Schleuter and DeYarman (1977) did a follow up study which further supported the De Yarmans initial findings. Interpretation A few studies addressed sp ecifically the issue of mu sic interpretation. Johnson (2000) analyzed 24 piano performances of a portion of Beethovens Symphony Number 5 in c minor seeking to determine what rhythm ic and dynamic variations were used. Eight piano performing experts were asked to perform three times. The first time was played as written, the second time they added their inte rpretation to the performance to make it as musical as possible, and the third time they exaggerated their interpretation from the second effort to the point the music was unm usical. Results indicated that while dynamic changes were rather small for each take, changes in timing were much more dramatic. This suggested that in piano performan ce, the performer contributed more to interpretation using timing variations, and re lied more on the score for dynamic changes. Gibson (1986) examined the ability of 4to 6-year-old ch ildren to perceive changes in culturally identifiable interpre tative differences of speech and musical
performance. Visual stimuli were used first to determine if subjects could group categories of line drawings. If subjects were su ccessful at this first task, they proceeded to group categories of spoken, sung, and played interpretations. Gibson found an increase in mean score performance for each successive age group. Subjects seemed more adept at categorizing visual stimuli than auditory information. Rodriguez and Webster (1997) conducted re search to determine the nature of childrens verbal responses to repeated hearings of a brief musical excerpt. Thirty-three kindergarden through fifth grade children ga ve responses to questions which were recorded and submitted to three judges who categorized the responses into music listening or other developmental studies. Th ey found similarities between responses and the recognized developmental theory of Gardner (1983). Denardo and Kantorski (1995) conducted a study that investigated second and fifth graders musical cognition when listeni ng to four-phrase songs. The children were asked to determine if the second, third, and four th phrases were the same as, similar to, or different from the first phrase. There wa s no significant difference between the two groups in ability, accuracy declined the furthe r away in time the phrase in question was to the initial reference phrase, and subjects were more accurate in identifying phrase types that were different followed by similar (one parameter change), similar (two parameter changes), and same. This may s upport the learning window theory. Holahan, Saunders, and Goldberg (2000) examined college musicians, non musicians and 1st grade children to investigate development of tonal audiation skills and test the validity of a prev ious study conducted by Holahan and Saunders in 1993. Their work found the three groups differed from each other in regard to accuracy and speed of
response. College musicians demonstrated systematically higher scores than both nonmusicans and first graders. First graders te sted only slightly lower than nonmusicians. McBeth (1992) described music interp retation as the most difficult of all pedagogical efforts because it does not fall into the categor y of technique. Instead, it could only be achieved through a thorough understanding of the intent of the composer. Correct interpretation is the re-creation of th e composers intent (p. 17). This reinforced from a conductors perspective the learning window theory a bout the cultural relationship between performer and listener. Shaffer and Todd (1994) examined the perception of rhythm and timing by analyzing the nuances of rhythmic patterns in repeated piano perfor mances using actual performances of pieces. In this study, it was found that pianists slow down the tempo at places of stability in a very precise manner. At the cognitive level, this implied that the motor system was accessing a stable timekeeper and had a definite representation of the relevant timing parameters. At the musical leve l, it meant that tempo and rubato were not left to chance, but were the outcome of coded decisions, or implicit rules, of interpretation. Nettls work (1994) supported the th eory that music is culture bound. He suggested music expresses a subtext whose me ssage is determined by extraneous issues such as culture, class, gender, and personality. While he was unable to identify explicitly musical thought as different from other kinds of thinking, he did suggest the way musicians think about their musi c depends largely on the way they perceive their world at large. The way a society thought about the concept of music, shaped the way its musicians made music.
Music and Language Levman (1992) cited three major positi ons regarding the origins of music and language. The first was that they both are sepa rate and different faculties. This view supported the notion that languages roots were primarily gestur al in origin, the result of a primitive representation system early humans developed to navigate the environment. It began with humans mimicking each other alon g with other sounds in their surroundings. Initial word-sounds were emotive and expressive declarations of species identification, warning calls, cries for help, etc. The second position was that music developed out of language. Song was the result of emotional speech intensified and systematized. The idea was put forth by Spencer in 1857. Most musicologists of the ni neteenth and twentieth centuries subscribed to one of these two positions. John Blacking (1973) believed that music was a speciesspecific biological human impulse, separate from language, but also inseparable from the social context in which it develops. Nettls position proposed that bo th were inseparable as a form of communication and then grew apart (1983). In the third position, Levman (1992) sugge sted both music and language evolved out of a common proto-faculty which was primar ily musical in nature, out of impulse of the human to survive. The frequency com ponent contained the most sound perception. Frequency contained not onl y pitch, but timbral quality and vocal information. The concept of double articulation (L evman, 1992) separated music from language which until then was closer to mu sic than speech. Double articulation was the sonic division of a language into indivi dual phonemes. By themselves, phonemes had no
intrinsic meaning, but in combination they form ed morphemes. It was at this point where language ceased to be isomorphic and became symbolic. London (1996) proposed there was a cultural belief that music was a kind of language. According to London, music is treat ed as a linguistic phenomenon. There is a tendency to use the language framework to address all meaningful communicative behavior. Since the music is language framework is so well established in our musical training and the language we use to describe music, it becomes wholly transparent for many liste ners music becomes a subclass of linguistic phenomena. (p. 51) He contended that, if there was agreement that there was some sort of communication between composer and listener, in such cases the listener reli ed on his linguistic competence to interpret it. Fiske (1990) perceived the argument to ha ve two points. The first was whether or not music was processed cognitively in a similar fashion to language. The second was whether or not music had content and whether that content was specifically communicated as in language. Swain (1996) posed a paradox in musica l semantics and established a case for similarity between music and language. According to Swain, music seems full of meaning to listeners from beginning to adva nced, yet there is no consensus to specific meaning, a component essential to natura l language. Music and language both have varying degrees of semantic potential, but both are limited by the context in which they occur. The semantic range of a sentence is comparatively small because its component
parts define its context very well. Yet, when not in sentence context, individual words have much more range. While the semantic ra nge for an instrumental music passage is even greater still, Swain argues it is only in degree and not in kind. So while listeners often disagree on the symbolic im port of a piece of music for lack of context, the same can happen in language. Fiske agrees with Swain that the issue is based on semantics, and adds that one definitive char acteristic in language not found in music is that language is denotative. Heller (2000) continued refining his work in a paper designed to further examine the link between speechs supra-segmenta ls, or the inflection and nuances superimposed on words, with the subtle co mponents of expression or interpretation in music (stress, attack, vibrato, crescendo and others). Defining interpretation as the expressive characteristics of performance, He ller suggested similarities existed between prosody in speech and music interpretation (p. 77). To access this covert music listening pro cess, Heller rejected the idea of breaking the acoustic medium down into traditional measurements (frequency, amplitude, duration) because reduction of the full sound re moved the critically needed context and culture the listener relied on for understanding and compre hension. His work regarding music interpretation mirrored that of Ro mmeveit (1974) who proposed language was an evocative process involving the shared cultu ral experience of both the performer and the listener. The test for this study was based on functi onal and not acoustical similarity since the later has not proven a good indicator of lis tener response to inte rpretative gesture in music. Specific acoustic variables were not need ed to predict listener response to music.
The issue stressed was whether the listener c ould hear the musical interpretation of a phrase that was repeated ( but not identical), yet was intended to be the same interpretation. The main determinant of the listeners selection wa s whether the phrases had enough characteristics in the repetition to be considered different or not different in interpretation. If the performance was su ccessful, the listener should be able to categorize the interpretative intent of the performer.
Chapter Three Method Problem This study was designed to further examin e relationships between the perception of performed music and spoken language as reflected in a model of a learning/practice window that begins to close between the ages of 6 and 10 for the perception of music interpretation (subtle changes in th e expression qualities of performance). Development of the Items Questions for this study were developed from previous work conducted by Heller and Athanasulis (2002). In their process, three professional performers (a pianist, clarinetist, and cellist ) recorded six short melodies speci fically notated for interpretation. In each melody, three different interpretations were recorded. To accomplish this, each performer imitated the performed phrase as interpreted by the previous performer. Initial interpretation of the phrase was induced by the notation. For example, the cellist might start the sequence by playing melody one (with interpretation A), th e pianist and then the clarinetist would follow, intentionally trying to imitate as closely as possible the initial rendition of the perfor mer who preceded them. This sequence was repeated until all three performers agreed that at le ast three of these performances were good interpretative matches. The same procedure was followed for melody one with interpretations B, and C (each different from the other). This process was continued until
three different melodies, with three interpretations each, were recorded to the satisfaction of the three performers. This produced a total of 162 phrases (6 melodies x 3 interpretations x 3 repetitions x 3 instruments). All of these phrases were digitized and stored in separate computer files. The recordings were samp led at 18,500 bytes/second. The speech stimulus recordings were developed in a similar manner. The first stage required that three speaker s (a male, a female, and a child) recorded six specifically written short phrases (three were nonsense syllable phrases, the other three were meaningful short phrases) each with three different interpretations. Each phrase was about four seconds in length. Each speaker im itated the spoken phras es as interpreted by the previous speaker. Initial interpretation of the phrase was induced by the written words using italics for stress points For example, the male speaker might start the sequence by speaking phrase one (with interpretation A); the female speaker would follow, trying to imitate as closely as possible the rendition of the male speaker. This was then followed by the childs rendition trying to imitate the female utterance. The order of sequence was randomly changed for each written interpre tation. This process was repeated until all agreed that at least three of these utterances were good interpretive matches. The same procedure was followed for phrase one with inte rpretations B, and C (each different from the other). This process was continued until all six phrases, with three interpretations each, were recorded to the sati sfaction of the speakers and res earchers. As with the music phrases, the speech procedure produced a total of 162 short speech phrases (6 phrases x 3 interpretations x 3 repetitions x 3 speakers). These phrases were also digitized and stored in separate computer files. The record ings were sampled at 18,500 bytes/second.
The next portion of the test-making process consisted of organizing the music and speech recordings into usable items to be presented to the subjects. Three renditions of a melody or utterance, two with the same in tended interpretation, one with a different intended interpretation, were transferred from their separate computer files into a new file. Each of the three phrases in the file wa s a different playing, but two were intended to be within one interpreta tive category (e.g., Interpreta tion A, Interpretation B, Interpretation A). Each phrase was separa ted by one second of silence. Following the three initial phrases, 1.5 seconds of silence and the word a gain followed by another 1.5 seconds of silence was added. The three phras es were then repeated (A...B...A.again A...B...A). Each collection of th ree repeated phrases and sile nce was designated one test item. The length of each test item, three phras es repeated, was about 30 seconds. The task for the listener was to determine which inte rpretation did not belong with the other two and to indicate their choice on a bubble sheet. In Heller and Athanasulis (2002) study, a similar test was created that included thirty items constructed of 30 music items, and 30 speech items. No consideration was given to determine which combinations of phr ases would be most effective in testing listeners. For this dissertation, two pilot stud ies were designed specifically to examine the quality of a wide range of possible test it ems to help determine which items would be most effective in a final testing format. In each of the two speech categories (sp eech phrases and nonsense phrases), items were created five different wa ys. Items were created using on ly the childs voice for three interpretations; only the womans voice for thr ee interpretations; only the mans voice for three interpretations; a combination of child first, woman second, and man last; and a
combination of woman first, child second, and man last. In each of these five subcategories, 18 items were constructed re presenting a wide variety of the possible combinations. Ninety items were developed fo r each of the two speech categories. The speech portion contained a total of 180 possible test items. In each of the two music categories (m usical excerpt phrases, and composed music phrases), items were also created five different ways. Items were created using only the piano for three interpretations; only th e clarinet for three interpretations; only the cello for three interpretations; a combination of piano first, cello second, and clarinet last; and a combination of cello first, clarinet s econd, and piano last. In each of these methods, 18 items were constructed representing a wi de variety of the possible combinations. Ninety items were developed for each of the two music categories. The music portion contained a total of 180 possible items to choose from. The 18 individual items generated in each subcategory were then linked together into one listening presentation. Each item (with the three repe ated phrases and silence) was separated from the next by 1.5 seconds of silence; a vocalization of the number of the item in the order it was presented; and 1.5 more seconds of silence. Linked together, the 18 items varied in total length from 6 to 11 minutes, depending on the length of the individual phrases which comprised each item. A total of 20 listening assessments (ten in speech and ten in music) each containing 18 it ems was created out of the subcategories. The specific listener subject task was to decide which phrase in the three-phrase item was different in interpretation from the other two. In each item all three phrases were different performances (utterances) of the same phrase. That is, phrase one was followed by two more performances of that phrase, but one of the three phrases was
different in interpretation of the same mel ody or written sentence. The phrase that was the same interpretation was never a re-record ing of that interpretation. It was always another performance, but one that was inte nded by the performer or speaker to be the same interpretation. This makes the task more difficult, yet more generalizable to a realistic musical or speech situation. In a normal musical performance or in normal speaking the performer may intend to repeat spec ific interpretations of a phrase, but each repetition is not identical. Ther e is variability among phrases that are intended to be the same in interpretation. Musical or language decisions are made by the listener to interpretative category rather than to interp retive equivalence. So the listeners task in this study was to determine which of the thr ee phrases was intended to be different from the other two in interpretation. Pilot One The pilot assessments were administered to two sixth grade string classes in an urban, public middle school in the southeast Un ited States. Class A contained between 15 and 19 students depending on attendance. Class B contained between 20 and 25 students. Permission was grante d by the string teacher, who in corporated the assessments into her daily class routine as a listening task. Only one assessment was given in any class period. Both classes were given instructions and examples to practice on together as a group prior to their first effo rt to insure they understood the task. To keep students interested throughout the durati on of the assessments, the cl asses artificially competed against each other for the title of Best Listen ers, and a period of recreation with snacks was provided to the class with the highest average. Students marked their answers on a scantron and turned in their work without th eir names on the papers to assure anonymity.
Assessments were taken over a period of si x weeks depending on the school, class, and researchers schedules. Analysis was conducted to determine the percentage of students who successfully answered each item. Lists were then compile d for each item in each category (i.e. child speech, cello composed melody) w ith their percentage of success. Appendixes A, B, C, and D show the complete results for speec h, nonsense, melodic except, and composed melody categories respectively. Those two item s closest to 50% success in each category were selected for use in the study. In situations where more than two items were available, the two chosen were randomly selected from the available options. In categories where two items were not at the 50 % mark, the two closest (either high or low) were selected. This effort produced 2 items of relatively similar difficulty from each of the 20 categories for a total of 40 items to be placed in the final study assessment tool. Item selections to be incorporated into th e final assessment tool for each subcategory are also presented at the bottom of Appendixes A, B, C, and D. Appendix E lists each sub categorys descriptive statistics including N, mean, percentage correct, median, and standard deviation. Pilot Two The second pilot study examined the reliability of the final study assessment tool. For this pilot, the 40 selected items were placed into two randomly generated orders. Both forty-item assessments (A and B) were then administered to two new groups of 6 th graders in a different sch ool with similar socio-econo mic conditions whose music program contained students with similar mu sical experience. In addition, assessment A was also given to one eighth-grade class and one 10-12th grade class. As with the
previous study, permission was granted by the two music teachers, who incorporated the assessments into their daily class routine. All classes were given instructions and examples to practice on together as a group pr ior to their first effort to ensure they understood the task. Assessment A was conducte d on a different day than assessment B. The order was altered to prevent students from recalling the first test by simple repetition. Students marked their answers on a scan tron sheet in a similar manner of the previous pilot, but in this study, the stude nts were assigned a scantron sheet with a number written on it. This allowed the researcher to identify each students work on both assessments (A and B) while still maintaining anonymity. Descriptive statistics, Cronbachs alpha, and pre/post Pearson Product-Moment correlation of the two 6 th grade groups taking both assessments A and B are displayed in Table 1. The Cronbachs alpha score of both groups suggest a good internal reliability, reinforcing the item selection process of the first pilot. Scores of .83, and .89 from the two groups examined are well within the es tablished range of acceptability (Nunnally, 1978). Pre/post Pearson Product-Mome nt correlation coefficients of .42, and .58 suggest a small to moderate relationship betw een test A and test B. While the Pearson ProductMoment correlation coefficients are lower than expected, small sample sizes may be allowing for wide confidence intervals that may account for this small relationship. The high Cronbachs alpha scores suggest good internal re liability allowing the study to proceed.
Table 1 Pilot 1 Descriptive Statistics, Cronbachs alpha, and Pre/Post Pearson Product-Moment Correlation Group 1 Test N Mean SD Cronbachs alpha Assessment A & B r Assessment A & B A 23 .59 .16 .83 .42 B 23 .58 .14 Group 2 Test N Mean SD Cronbachs alpha Assessment A & B r Assessment A & B A 18 .54 .17 .89 .58 B 18 .51 .15 Further examination of an 11-12-year-o ld group, a 13-14-year-old group, and a 16-18-year old group was investigated us ing a one-way ANOVA, between groups design to determine if the higher grade improved test results. Descriptive statistics are displayed in Table 2. All groups showed within acceptable skewness and kurtosis with one exception. The 10-12 th grade group demonstrated a notab ly leptokurtic distribution. The 10-12 th grade group also had one extreme outlier.
Table 2 Pilot 2: 6 th 8 th 10-12 th Grade Descriptive Statistics 11-12-year-olds 13-14-year-olds 16-18-year-olds N 19 17 29 Mean in % 55.28 65.45 73.20 S.D. in % 17.13 13.80 15.38 Skewness -.61 -.42 -.73 Kurtosis -.53 -.15 1.44 Sample means of 55.28%, 65.45%, and 73.2% for the 6 th graders, 8 th graders, and 10-12th graders respectively, show enough variati on to be of interest. To provide a more standardized measure of effect size, Cohens f was calculated to be .486. This can be interpreted to mean that the group means typi cally deviate from the grand mean by about .4 standard deviations. An effect size, f of .4 is generally considered large. With N close to 20 for all groups, this ANOVA was relatively robust to violations of the assumption of normality. Students ha d no contact between grade level, and worked individually on the assessment, main taining the assumption of independence. Based on the descriptive statisti cs in Table 2, there does not seem to be any substantial violation of the normality assumption (ske wness<1.0 for all groups) or of the equal variance assumption (maximum sta ndard deviation ratio =1.1). An ANOVA was conducted examining the thr ee different groups taking the test, a summary of which is on Table 3. The obtained F (2,62) = 7.67, p = .0011, was judged to
be statistically significant at the .05 level suggesting that at least one pair of population group means differ. A Tukey test determined that the 6 th grade group differs significantly from the 10-12 th grade group. Table 3 One-Way ANOVA Summary Table of 6 th 8 th and 10-12 th Grade Source Df MS F p Between 2 295.76 7.67 .0011 Within 62 38.57 This pilot suggests there is evidence to contradict the theory that the learning/practice window closes at the age of ten. Study Population and Sample Assessment A, developed and refined from the previous two pilot studies, was administered to 292 students at a public, ma gnet school for visual and performing arts in North Carolina. Students participating in the study were enrolled in the instrumental music education program at the school. Th e test was administered during regular instrumental classes over a two-day period as a part of their daily class routine. Each class was given instructions and examples to practice on together as a group prior to their taking the test to insure th ey understood the task. Student s marked their answers on a scantron sheet. In addition to the answers, students included their grade level, age, gender, and whether they had any private lesson experience on their papers. Scantron sheets were turned in without names on th e paper to assure anonymity. Approval was granted by both the principal and the dist rict before the test was administered.
Variables The primary independent variable of inte rest in the study wa s the age of each subject. Gender of the subject and whether or not each student had any private music lesson experience was also considered. The dependent variables in the study in cluded the scores on th e speech and music items of the assessment. Measures The measures utilized in the study were sc ores obtained from the test prepared to assess the listening skill along with responses to self-report information obtained at the time the test was administered. Data Analysis To address the questions posed in th is study, three 7x2x2 factorial ANOVAs were performed examining the test scores of students with their age, gender, and prior private lesson experience. The first ANOVA examined th e three independent variables with the overall test score of each subject. The other two ANOVAs examined the same three independent variables with each subjects sc ore of the music and speech portions of the test by themselves. Because three tests we re being applied to the same data, a Bonferroni adjustment was used on the initial alpha level of .05 to account for the rising Type I error rate. The adjusted alpha level for the three tests was .017.
Chapter Four Results The purpose of this study was to investig ate the effect age had on the ability of students older than the proposed learning window (11-18 years of age) to discriminate subtle changes in phrase interpretation. For an alysis, students were divided initially into eight age groups: 11-year-olds ( n = 42); 12-year-olds ( n = 57); 13-year-olds ( n = 57); 14year-olds (n = 48); 15-year-olds ( n = 39); 16-year-olds ( n = 29); 17-year-olds ( n = 15); and 18-year-olds ( n = 5). The group of 18-year-olds with an n of 5 was too small to include in this investigati on as its own group. It was combined with the 17-year-old group (17-18-year-old group n = 20). Assumptions Before three 7x2x2 level factorial ANOV As, each with three between group factors, could be performed, three assumptions were determined to have been met in the study. The first was that the scores within groups could be cons idered independent because students did not interact during the treatment and were monitored by three teachers. The second assumption was that the populat ion distribution was normal. Each age group was evaluated in terms of mean, SD, skewness, and kurtosis (see Table 4). Generally accepted ranges for skewness and kurtosis are -1.0 to 1.0 and -0.5 to 1.0, respectively. Groups aged 11, 14, 15, and 16 disp layed kurtosis scores which gave cause for further examination. In all groups men tioned, the distributions were slightly
platykurtic. The first factorial ANOVA examining the overall test scores was robust to violations of the assumption of normality because the sample sizes were large. The last assumption was that of the homogeneity of population variances. Based on the standard deviation statistics in Table 4, there did not seem to be any violation of the equal variance assumption. Table 4 Age 11, 12, 13, 14, 15, 16, & 17-18 Overall Test Score ANOVA Descriptive Statistics Age Groups 11 12 13 14 15 16 17-18 N 42 57 57 48 39 29 20 Mean in % 56.55 58.43 64.25 62.40 61.73 69.48 63.13 S.D. in % 17.98 17.03 17.70 18.76 21.13 21.18 19.25 Skewness -.17 -.27 -.54 -.40 -.18 -.52 -.73 Kurtosis -1.24 -.22 -.42 -.53 -.93 -.90 -.14 Sample means of the groups found on Table 4 did not appear to have much variation. To provide a standardized measure of effect size, Cohens f was calculated to be .18. This could be interpreted to mean that the group means deviate about .18 standard deviation units from the grand mean. According to Cohen, this indicated a small effect size. Because none of the assumptions had been violated, an ANOVA could be conducted. In anticipation of running three factorial ANOVAs on the same data set to address the questions posed at the beginning of the study, the initial = .05 was adjusted down to
= .017 using a Bonferroni adjustment. This was necessary to account for an increasing Type I error rate as more than one te st was run on the same data set. A Pearson Product-Moment correlation coefficient of .75 was produced when examining students scores on the music items with those of the speech items. This score indicates a high positive correlation. An r 2 value of .56 indicates th at 56% of the total variance of music items is explained by the variance of the speech items. Cronbachs alpha on the overall sample was calculated at .87. According to Nunnally (1978) this score is within the established range of acceptability for internal reliability. Overall Test Results Once the assumptions were determined to have been met, the first 7x2x2 factorial ANOVA, three between group factors was performed. This ANOVA examined the independent variables of the seven diffe rent age groups, gender, and prior lesson experience with the dependant variable of ove rall test score to determine any possible levels of significance. A summary for the factorial ANOVA examining age, gender and lessons on the overall test is displayed in Table 5.
Table 5 7 x 2 x 2 Factorial ANOVA Summary Table Investiga ting the Relationship Between Age, Gender, and Prior Lesson Experience on Students Ability to Interpret Music and Speech Phrase Changes Source Df MS F r 2 Age 6 51.52 .98 Gender 1 206.19 3.93 Lesson 1 672.93 12.83* .04 Age x Gender Int. 6 16.11 .31 Age x Lesson Int. 6 38.96 .74 Gender x Lesson Int. 1 12.71 .24 Age x Gender x Lesson Int. 6 42.58 .81 Within 264 52.46 Significant at p < .017 after Bonferroni adjustment for a study-wide alpha level of .05. This overall analysis revealed a si gnificant main effect for lessons, F (1, 264) = 12.83; p < .0004. Students with private lesson experience, n = 66, Mean Score = 70.72%, scored significantly higher than those students with no private lesson experience, n = 226, Mean Score = 59.2%. Sample means for le sson experience arranged by age group are displayed in Figure 1.
Figure 1 Overall Lesson Score Means by Age Group in Percentages 020406080100AgeOverall Test Percent Correct No Lessons55.253.9364.4859.7358.9864.6359.63 Lessons69.387163.2571.3869.7580.2869.6511121314151617-18 The main effect for age was nonsignificant, F (6, 264) = .98; p = .44. The main effect for gender was nonsignificant, F (1, 264) = 3.93; p = .048. The interaction of age and gender was nonsignificant, F (6, 264) = .31; p = .93. The interaction of age and lessons was nonsignificant, F (6, 264) = .74; p =.62. The interaction of gender and lessons was nonsignificant, F (1, 264) = .24; p = .62. The interaction of age, gender, and lessons also was nonsignificant, F (1, 264) = .81; p = .56. Speech and Music Parts An examination of the test parts, music and speech, using two 7x2x2 factorial ANOVAs with three between group factors was performed to determine any possible
levels of significance. Before the tests were conducted, descriptive statistics were examined for any anomalies that would have violated normality assumptions. Descriptive statistics for speech items grouped by age are located in Table 6. All age groups fell within the accepted range of skewness and groups aged 11, 13, 14, 15, 16, and 17-18 demonstrated slightly platykurtic distributions. The large group sizes made the ANOVA robust to this normality violation. There did not seem to be any violation of the equal variances assumption. Table 6 Age 11, 12, 13, 14, 15, 16, & 17-18 Speech Items ANOVA Descriptive Statistics Age Groups 11 12 13 14 15 16 17-18 N 42 57 57 48 39 29 20 Mean in % 61.45 62.55 67.80 66.15 66.30 71.55 67.25 S.D. in % 21.9 19.2 21.35 22.1 23.15 23.6 19.95 Skewness -.25 -.53 -.47 -.62 -.33 -.71 -.42 Kurtosis -1.24 -.12 -.56 -.58 -1.04 -.84 -72 Descriptive statistics for music items grouped by age are located in Table 7. All age groups fell within the accepted range of skewness and groups aged 11, 12, 16, and 17-18 demonstrated slightly platykurtic distributions. The 17-18-year-old group demonstrated a notably leptokurtic distribution. The equal variance assumption had not been substantially violated.
Table 7 Age 11, 12, 13, 14, 15, 16, & 17-18 Music Items ANOVA Descriptive Statistics Age Groups 11 12 13 14 15 16 17-18 N 42 57 57 48 39 29 20 Mean in % 51.65 54.30 60.70 58.65 57.20 67.40 59.00 S.D. in % 17.40 17.60 16.55 17.85 21.25 20.8 20.15 Skewness .03 .22 -.36 -.01 -.1 -.14 -.63 Kurtosis -.7 -.57 -.2 -.34 -.42 -1.23 1.01 A summary for the factorial ANOVA examining age gender and lessons for speech items is displayed in Table 8. Table 8 7 x 2 x 2 Factorial ANOVA Summary Table Investigating the Relationship Between Age, Gender, and Prior Lesson Experience on Students Ability to Interpret Speech Phrase Changes Source Df MS F r 2 Age 6 11.34 .64 Gender 1 58.5 3.29 Lesson 1 185.47 10.42* .03 Age x Gender Int. 6 5.7 .32 Age x Lesson Int. 6 6.67 .37 Gender x Lesson Int. 1 .12 .01 Age x Gender x Lesson Int. 6 12.21 .69 Within 264 17.8 Significant at p < .017 after Bonferroni adjustment for a study-wide alpha level of .05.
The factorial ANOVA of the speech items revealed a significant main effect for lessons, F(1, 264) = 10.42; p < .001. Students with private lesson experience, N =66, Mean Score = 74.75%, scored significantly higher on speech test items than those students with no private lesson experience, N = 226, Mean Score = 63.1%. Sample means for lesson experience arranged by age group are displayed in Figure 2. Figure 2 Speech Item Score Means for Lesson Experience Arranged by Age Group in Percentages 020406080100AgeSpeech Item Percent Correct No Lessons60.458.6567.6563.2562.9566.563.1 Lessons71.2573.3568.575.97682.87511121314151617-18 The main effect for age was nonsignificant, F (6, 264) = .64; p = .7. The main effect for gender was nonsignificant, F (1, 264) = 3.29; p = .07. The interaction of age and gender was nonsignificant, F (6, 264) = .32; p = .93. The interaction of age and
lessons was nonsignificant, F (6, 264) = .37; p =.89. The interaction of gender and lessons was nonsignificant, F (1, 264) = .01; p = .94. The interaction of age, gender, and lessons also was nonsignificant, F (1, 264) = .69; p = .66. A summary for the factorial ANOVA exam ining age, gender and lessons for music items is displayed in Table 9. Table 9 7 x 2 x 2 Factorial ANOVA Summary Table Investiga ting the Relationship Between Age, Gender, and Prior Lesson Experience on Students Ability to Interpret Music Phrase Changes Source Df MS F r 2 Age 6 14.74 1.17 Gender 1 45.03 3.57 Lesson 1 151.84 12.04* .04 Age x Gender Int. 6 6.21 .49 Age x Lesson Int. 6 14.95 1.19 Gender x Lesson Int. 1 10.38 .82 Age x Gender x Lesson Int. 6 12.19 .97 Within 264 12.61 Significant at p < .017 after Bonferroni adjustment for a study-wide alpha level of .05. The factorial ANOVA of the music part of the test revealed a significant main effect for lessons, F (1, 264) = 12.04; p < .0006. Students with private lesson experience, N =66, Mean Score = 66.65%, scored significan tly higher on music test items than those students with no private lesson experience, N = 226, Mean Score = 55.35%. Sample means for lesson experience arranged by age group are displayed in Figure 3.
Figure 3 Music Item Score Means for Lesson Experience Arranged by Age Group in Percentages 020406080100AgeMusic Item Percent Correct No Lessons5049.1561.356.25562.7556.15 Lessons67.568.655866.863.577.864.311121314151617-18 The main effect for age was nonsignificant, F (6, 264) = 1.17; p = .32. The main effect for gender was nonsignificant, F (1, 264) = 3.57; p = .06. The interaction of age and gender was nonsignificant, F (6, 264) = .49; p = .81. The interaction of age and lessons was nonsignificant, F (6, 264) = 1.19; p = .31. The interaction of gender and lessons was nonsignificant, F (1, 264) = .82; p = .36. The interaction of age, gender, and lessons also was nonsignificant, F (1, 264) = .97; p = .44. In an examination of all three factorial ANOVAs conducted for this study, the r 2 accounted for no more than 3-4% of the variance. This indicates that while there was a significant difference observed between those students who had participated in lessons
and those who had not, it only explains a very small portion of the total variance. It is also noteworthy that in a ll three tests conducted, no si gnificance was found between age group and students ability to discriminate changes in instrumental and speech phrase interpretations.
Chapter Five Discussion The purpose of this study was to determine what effect age had on middle school and high school students ability to interpret subtle changes or nuances in music and speech phrases. It was designed to further test a proposed theory of a learning window for the perception of expressive qualities in music and speech. The theory suggests that a practice window must overlap a learning window before it closes around the age of 10 to properly develop music prosody and speech prosody skills. This study is a continuation and advancem ent of the work accomplished by Heller and Athanasulis (2002). In their study, Heller and Athanasulis developed a test comprised of thirty speech items and thirty music items, each consisting of three small phrases. It was admi nistered to forty 1 st -grade students (about six years old), forty 3 rd grade students (about eight years old), and forty 5 th -grade students (about 10 years old). Using an analysis of variance, Heller and Athanasulis (2002) found that in music items, growth was found to be statistically significant between th e 6and 10-year-olds (10.7%). A Scheffs post hoc comparison showed the mean difference between 6and 8year-olds was not significant (4.8%), but was significant in 8to 10-year-olds (5.8%). This small increase showed no leveling off in subjects as a function of age, which contradicted the proposed theory. In an analysis of the speech items, si gnificant growth was found in both groups (17.3% in 6to 8-year olds and 10.7% in 8to 10-year olds). The researchers concluded
that the reduction in growth between the pairwise comparisons in the speech items suggests a leveling off of growth which supports the theory. One limitation found in the Heller and Athanasulis study was in the development of the test items. An item analysis was not conducted, and it is possible that better items could have been selected given the large number of item possibilities to choose from. Another limitation was in the c hoice of students. While the goal of the study was to test a theory of a learning window that theoretically closes at the age of around ten, no students older than around ten were inve stigated. Closure of this learning window was not fully investigated. The current study attempted to follow up on the work of Heller and Athanasulis by first improving on the quality of the test Three hundred sixty items were developed and then tested to determine through item analysis which questions were most appropriate for the task. In addition, this st udy investigated the ab ility of students who were older than the proposed learning window theory to determine when and if the learning window closes at the proposed age of 10 to 11 for either music or speech phrase interpretation. In all three factorial ANOVA tests of the entire test scores and their parts, no significance was found between any groups aged 11 to 18. This contradi cts the results of the second pilot study which found that older stud ents improved in ability to take the test. One possible explanation for this difference in the pilot study is the smaller sample sizes and different schools being examined between the sixth and eighth graders and the 1012th grade group. The results of the current study reinforce the proposed theory that a learning/practice window seems to be in place for the acquisition of auditory
interpretative skill. What is of particular intere st in this finding is that the results seem to contradict the natural refinement of the quality of mu sic student over time. Younger groups (aged 11-12) are comprised of any st udent expressing in terest in learning instrumental music. As years pass, students who remain in instrumental music show an increasing interest and longer-term commitme nt to the development of their musical craft. The reduction of the reten tion in the program can be seen in the sizes of the groups examined in this study. It is noteworthy that students who remained in instrumental music for six years of public school music education did not fare statistical ly better than those just entering the program. At the time of the testing, beginni ng sixth graders would have only had one semester of instrumental music education. These results seem to indicate a leveling off of students ability to inte rpret speech and music phrase changes. In all three factorial AN OVAs, prior lesson experience demonstrated the only level of significance in the three independe nt variables or thei r interactions. Those students who had indicated they participated in private lessons in the past scored significantly higher than those students who had not taken pr ivate lessons previously. A possible explanation for this outcome may be the additional time devoted to music study implied by their decision to pursue additiona l music education outside the public school classroom. Those students al locating time and money on supplemental music education may have a more vested interest in the impr ovement of their musical abilities. Additional time practicing and studying music is inferred w ith the additional lessons It is interesting to note that those students who enrolled in private music lessons in the past also scored higher than those students who did not take lessons on the speech portion of the test. This
may be due to a higher overall commitment to academic excellence in general on the part of those students opting fo r additional music lessons. Limitations A number of factors could have affected the data obtained in this study. While the selection of a school containing six to twelfth grade students selected by lottery from the city population provided a good representation, a more useful sample may have been random subjects taken from a vari ety of different schools to better represent the general music education environment. While the studen ts obtained consistent education from the same music teachers throughout their tenure at the school, their experience and the results of this study may not be gene ralizable to the rest of the instrumental music education environments. Another limitation was in the wording of the private lesson question. Students were asked if they had ever participated in private lessons in the past. The question was not specific to privat e lessons prior to the closing of the proposed learning window. This means that a sixteen-year-old student coul d have answered yes to the private lesson question, but had lessons at the age of 14, well after the proposed learni ng window would have closed. While the private lesson informa tion is valuable, it could have been more useful with a more speci fically worded question. The length of the test may have been too long for consistent concentration and focus of each student on the task. Perhaps providing a break in the middle of the test would have improved students level of focu s. In addition, it is possible that smaller groups may have provided a quiet er environment to better hear the subtle changes in interpretation.
Future Study The results of this study reinforce th e idea that a practic e/learning window may exist for the development of the speech and music interpretation skill. If the theory is correct and a practice/learning window does seem to be in place for the acquisition of an interpretative skill in music and speech, mu sic educators must reexamine the order in which their pedagogy falls. Some forms of mu sic deemed valuable and important for public appreciation by the profession, with th e exception of sporadic good fortune in the occasional primary music education classroom, are not formally introduced until well after the proposed learning/ practice window closes. Jazz studies are routinely omitted from music education until late in secondary education, if at all. Masterworks throughout history are often not introduced until seconda ry education. Comprehensive world music study is sometimes neglected until college, even for music majors. One reason for this is that the genres called to attention are more technically advanced than others. More technically cha llenging music is put off until students have an ability to successfully accomp lish performing the music. In some cases this means late in their secondary education. If there is validity to this theory, exposure to these kinds of music is occurring too little and too late in students development. In fact, it could be argued that one reason students have a more difficult time negotiating these forms of music could be because they have had so little expe rience with it earlier in thei r education. This is not an advocacy for starting brass quintets and jazz ba nds in primary school. Instead, it may be a call for a more comprehensive, in depth a nd experiential learning program for children within the learning/pract ice window to the music of history and the world.
Future studies are needed to further e xplore the proposed learning/window theory. An examination of students who had privat e lesson experience within the learning window may demonstrate that more extensive music activity within the learning window may lead to improved ability to interpret phrase changes. A study using current popular or world mu sic phrases to exam ine peoples ability to choose which one of three phrases is most culturally appropriate should be considered. If young participants are able to easily de termine the appropriate phrase but older participants cannot, it may re inforce the suggested theory. It is important to continue to research and develop a better understanding as to how the rules of music listeni ng are learned. Understanding this process can lead to more efficient, age-appropriate pedagogy in music education.
References Atterbury, B. W. (1983) A comparison of rhyt hm pattern perception and performance in normal and learning-disabled re aders, age seven and eight. Journal of Research in Music Education, 31 (4), 259-270. Blacking, J. (1973). How musical is man? Seattle: University of Washington Press. Boyle, J. D. & Radocy, R. E. (1987). Measurement and evaluation of musical experiences. New York: Schirmer Books. Campbell, M. R. (1991). Musical learning and the development of psychological processes in perception and cognition. Bulletin of the Council for Research in Music Education, 107 35-48. Cox, S. M. O. (1977). A descriptive analysis of the res ponses to beat, meter, and rhythm pattern by children, grades one to six. The University of Wisconson-Madison, Ph.D. Davies, J. B. (1978). The psychology of music. California: Stanford University Press. Davidson, L. & Scripp, L. (1985). Young childre ns musical represen tations: windows on music cognition. In J. A. Sloboda (Ed.), Generative Processes in Music : The Psychology of Performance, Improvisation, and Composition (pp. 195-230). New York, NY: Oxford University Press. Denardo, G. F., & Kantorski, V. J. (1 995). A Continuous Response Assessment of Childrens Music Cognition. Bulletin of the Council for Research in Music Education, 126 42-52.
De Yarmen, R. (1975). An investigation of the stability of mu sical aptitude among primary-age children. Experimental Research in the Ps ychology of Music, Studies in the Psychology of Music, 10. Iowa City: University of Iowa Press. Elliott, D. J. (1995). Music matters: A new philos ophy of music education New York: Oxford University Press. Fiske, H. (1990). Music and mind: philosophical essays on the cognition and meaning of music. New York: Edwin Mellen Press. Fiske, H. (1996). Selected theories of music perception New York: Edwin Mellen Press. Flavell, J. H. (1965). The developmental psychology of Jean Piaget. New York: D. Van Nostrand Company. Flohr, J. W. (1981). Short-term music in struction and young childrens developmental music aptitude. Journal of Research in Music Education, 29 (3), 219-224. Gardner, H. (1983). Frames of mind: The theory of multiple intelligences. New York: BasicBooks. Gibson, B. L. (1986). Young childrens response to interpretation in music and speech. University of Connecticut, Ph.D. Gordon, E. E., Advanced Measures of Music Audiation, Chicago, IL: G.I.A. Publications, Inc., 1989. Heller, J. (2003). The ability to categorize interpre tations of music and speech phrases. a paper presented at the Research Alliance in Music E ducation Meeting, Corsham, England. Heller, J. and Athanasulis, M. (2002). Music and language: A learning window. Bulletin of the Council for Research in Music Education, 153, 18-22.
Heller, J. J. (2000). An implicit rule system for music listening. Bulletin of the Council for Research in Music Education, 147, 77-80. Heller, J. and Campbell, W. (1981). A theore tical model of music perception and talent. Council for Research in Music Education. 66-67, 20-24. Heller, J., Campbell, W., and Gibson, B. ( 1982). The development of music listening skills in children. Psychology of Music, 10, (3), 55-58. Heller, J. & Campbell, W. (1982). Music communication and cognition. Council for Research in Music Education, 72, 1-15. Holahan, J. M., Saunders, T. C., & Goldber g, M. D. (2000). Tonal Cognition in Pattern Discrimination: Evidence from Three Populations. Journal of Research in Music Education, 48 (2), 162-177. Hufstader, R. A. (1977). An investigation of a learning sequence of music listening skills. Journal of Research in Music Education, 25, 185-196. Johnson, C.M. (2000). Effect of adding interp retive elements to a musical performance on the rhythmic and dynamic variations. Bulletin of the Council for Research in Music Education, 147, 91-96. Jones, R. (1976). The development of the childs conception of meter in music. Journal of Research in Music Education, 24 (3), 142-154. Krumhansl, M. (1983). Perceptual structures for tonal music. Music Perception, 1 (1), 2859. Lerdahl, F., & Jackendoff, R. (1983). Theories of developmental psychology. San Francisco: W. H. Freeman. Levman, B.G. (1992). The gene sis of music and language. Ethnomusicology, 36, 147-70.
London, J. (1996). Musical and linguistic speech acts. The Journal of Aesthetics and Art Criticism, 54, 49-64. McBeth, W. F. (1992). Interpretation: unlocking the drama in music. The Instrumentalist, 47, 14-18. Meyer, L. B. (1956). Emotion and Meaning in Music Chicago: University of Chicago Press. Miller, L. K., and Eargle, A. (1990). The c ontributions of development versus music training to simple tempo discrimination. Journal of Research in Music Education, 38 (4), 294-301. Moles, A. (1958). Information theory and esthetic perception. Urbana: University of Illinois Press. Nash, M. (1997, February 3). Fertile minds. Time. 48-56. Nettl, B. (1983). The Study of Ethnomusicology Urbana: University of Illinois Press. Nettl, B. (1994). "Musical thinking" and "thi nking about music" in ethnomusicology: an essay of personal interpretation [Blackfoot, Persian, classical]. The Journal of Aesthetics and Art Criticism, 52 139-48. Nunnally, J. C. (1978). Psychometric theory. 2 nd ed. New York: McGraw Hill. Palmer, M. (1976). Relative effectiveness of two approaches to rhythm reading for fourth-grade students. Journal of Research in Music Education, 24 (3), 110-118. Perney, J. (1976). Musical tasks related to th e development of the conservation of metric time. Journal of Research in Music Education, 24 (4), 159-168.
Pflederer, M. (1967). Conservation laws applied to the development of musical intelligence. Journal of Research in Music Education, 15 (3), 215-223. Rodriguez, C.X. and Webster, P. R. (1997). Development of children's verbal interpretative respons es to music listening. Bulletin of the Council for Research in Music Education, 134, 9-30. Schleuter, S. and DeYarmen, R. (1977). Musi cal aptitude stability among primary school children. Council for Research in Music Education, 51, 14-22. Seashore, C. (1938). Psychology of music New York: Dover Publications, Seashore, C. (1919). The psychology of musical talent Boston: Silver, Burdett and Company. Serafine, M. L. (1983). Cognition in Music. Cognition, 14, 119-183. Serafine, M.L. (1988). Music as cognition: The development of thought in sound. New York: Columbia University Press. Shaffer, L. H. & Todd, N. (1994). The interpre tive component in musical performance. In R. Aiello (Ed.), Musical Perceptions (258-272). New York: Oxford University Press. Shehan, P. K. (1987). Effects of Rote Vers us Note Presentations on Rhythm Learning and Retention. Journal of Research in Music Education, 35 (2), 117-126. Sloboda, J. A. (1985). The musical mind: The cogn itive psychology of music. Oxford: Clarendon Press. Swain, J.P. (1996). The range of musical semantics. The Journal of Aesthetics and Art Criticism, 54 135-52.
Zimmerman, M. P. (1982). Developmental processes in music learning. In R. Colwell (Ed.), Symposium in Music Education: A Festschrift for Charles Leonard. Urbana: University of Illinois Press. Zimmerman, M. P. (1986). Music developm ent in middle childhood: A summary of selected studies. Bulletin of the Council for Research in Music Education, 86, 1835. Zimmerman, M. P. (1993). An Overview of Developmental Research in Music. Bulletin of the Council for Research in Music Education, 116, 1-21.
Appendix A Percentage of Students who Answered Correctly Each Item of the Five Speech Categories. Each categorys items are unrelated to the other categories listed. Speech Items Item Child Speech Female Speech Male Speech Child, Female Female, Child Only Only Only Male Speech Male Speech 1 95 76 100 83 68 2 95 100 100 83 84 3 68 53 100 78 53 4 95 88 89 89 42 5 82 59 83 83 84 6 95 94 94 22 63 7 0 65 61 39 21 8 55 88 72 83 89 9 41 41 83 78 32 10 14 71 72 28 89 11 82 100 89 83 26 12 41 47 89 44 53 13 05 41 67 67 63 14 64 0 72 72 79 15 68 71 67 61 32 16 23 53 17 39 58 17 95 100 94 89 32 18 59 88 50 67 42 Items Selected: 8, 9 3, 12 7, 18 12, 16 3, 12
Appendix B Percentage of Students who Answered Correctly Each Item of the Five Nonsense Categories. Each categorys items are unrelated to the other categories listed. Nonsense Speech Items Item Child Only Female Only Male Only Female, Child Child, Female Male Male 1 100 52 74 09 49 2 87 92 95 65 71 3 0 72 63 70 41 4 67 68 89 52 41 5 80 56 79 35 76 6 13 52 68 57 24 7 67 60 42 52 59 8 87 64 53 48 35 9 13 24 84 43 29 10 93 32 32 22 35 11 47 68 53 70 47 12 53 76 79 17 12 13 07 72 37 74 76 14 100 68 84 65 71 15 87 56 74 57 53 16 100 28 42 39 35 17 07 16 79 48 59 18 0 60 68 91 41 Items Selected: 11, 12 1, 6 8, 11 4, 17 11, 15
Appendix C Percentage of Students who Answered Correctly Each Item of the Five Musical Except Categories. Each categorys items are unrelated to the other categories listed. Musical Except Instrumental Items Item Clarinet Only Cello Only Piano Only Piano, Cello Cello, Clarinet Clarinet Piano 1 92 38 60 12 13 2 63 31 65 41 17 3 50 100 45 59 39 4 63 88 30 88 22 5 38 75 45 29 48 6 79 38 30 35 30 7 75 88 35 59 61 8 71 69 50 35 43 9 71 50 40 76 35 10 71 94 55 35 30 11 58 50 55 35 30 12 58 50 45 29 57 13 67 75 45 47 39 14 71 44 45 29 35 15 42 50 50 29 17 16 79 69 60 29 57 17 33 88 45 35 57 18 63 31 50 35 57 Items Selected: 3, 12 9, 11 8, 18 3, 13 5, 17
Appendix D Percentage of Students who Answered Correctly Each Item of the Five Composed Melodies Categories. Each categorys items are unrelated to the other categories listed. Composed Melody Instrumental Items Item Clarinet Only Cello Only Piano Only Piano, Cello Cello, Clarinet Clarinet Piano 1 56 83 62 26 38 2 72 61 29 65 44 3 50 48 43 61 38 4 78 96 14 09 50 5 67 61 76 74 50 6 78 65 67 35 50 7 67 61 62 52 56 8 67 17 67 52 63 9 50 87 67 43 75 10 83 96 62 43 63 11 72 22 67 30 56 12 72 78 14 48 56 13 56 78 81 35 38 14 67 70 38 26 88 15 83 74 38 26 69 16 78 83 57 35 88 17 67 65 33 52 31 18 50 0 19 78 56 Items Selected: 3, 9 2, 3 3, 16 8, 12 4, 5
Appendix E Pilot One: Descriptive Statistics Speech N Mean/Total % Correct Median SD Child 22 10.8/18 60 10.7 2.5 Female 18 11.7/18 65 11.6 3.3 Male 18 14/18 77.78 14.5 2.8 Child, Female, Male 20 10.7/18 59.44 11.6 4.2 Female, Child, Male 19 10.1/18 56.1 9.5 2.6 Nonsense Child 15 10.1/18 56.1 10.5 2.1 Female 25 10.2/18 56.67 9.6 3.7 Male 19 11.9/18 66.11 11.5 2 Child, Female, Male 17 8.4/18 46.67 7.7 2.5 Female, Child, Male 23 9.1/18 50.56 9.5 3.5 Melodic Excerpt Clarinet 20 8.5/18 47.22 7.7 2.6 Cello 24 11.4/18 63.33 11.6 3.3 Piano 16 11.3/18 62.78 10.6 2.9 Piano, Cello, Clarinet 17 7.4/18 41.11 7.7 1.8 Clarinet, Cello, Piano 23 6.9/18 38.33 7.6 2.5 Composed Melody Clarinet 21 9/18 50 8.7 2.5 Cello 18 12.1/18 67.22 11.6 3 Piano 23 11.6/18 64.44 11.6 3.3 Piano, Cello, Clarinet 23 7.9/18 43.89 7.7 2.6 Clarinet, Cello, Piano 16 10.1/18 56.1 10.6 3
About the Author Andrew Sioberg received a Bachelors Degree in Music from Radford University in 1994, a Masters Degree in Jazz Studies and Contemporary Media from the Florida State University in 1997, and a Masters Degree in Music Education from Syracuse University in 1999. Following his training at Syracuse University, Andrew taught instrumental music education at the Durh am School of the Arts in Durham, North Carolina for three years. He entered the Ph.D. program at the University of South Florida in 2002 and completed the degree in 2005. Th roughout his education and teaching, Dr. Sioberg maintained an active career as a prof essional performer, arra nger, and clinician.