• No results found

2.   Literature Review


2.6.2   Corpus Analyses

conform to a particular issue.123 In doing so, one is better placed to develop a theory, in Temperley’s sense of the word, of a particular style, artist or body of songs. Such an approach also refines analytical methods because one can see when, and why, a particular tool may not be appropriate — for example, is Schenkerian analysis applicable or necessary for Britney Spears as well as The Beatles?

Stax recordings generally followed traditional structures (i.e. AABA,

verse/chorus, chorus/verse or blues); a third of the songs used only chords I, IV and V, while almost half involved harmonic movement in thirds; and nearly all the songs fit within two bands of tempi, either a slow ballad or between 102 and 132 bpm. Because Bowman’s sample spans a decade and features multiple songs by single artists, he uncovers historical trends, such as changes in chord voicings, and common aspects of an artist’s style.

Significantly, Bowman had earlier conducted interviews with the musicians, who gave insight into why the “Stax sound” came about. For example, one may relate the infrequency of minor chords to some cultural aesthetic; perhaps, the band wished to present an image of ‘happiness’ as they attempted to woo the white audience. Rather, Jim Stewart, studio producer and owner, just “didn’t like minor chords,” according to keyboardist Isaac Hayes.126 Similarly, guitarist Steve Cropper believed that the crash and ride cymbals were avoided because

“high end” percussion “offended the female buyer.”127

In this sense, Bowman links his analysis with musical experiences. His work is important in two complementary respects: first, he identifies the core musical features that distinguish the “Stax” style; and second, his ethnographic

investigation explains part of the studio’s aesthetic and provides greater depth to the analysis.

Jon Fitzgerald’s study is structurally similar. He examines the most successful black “crossover” songwriters from 1963-1966, those who had eight or more hits on the U.S. Top 40 chart, leading to 91 songs by five

songwriters/songwriting teams.128 Fitzgerald analyses the songs in terms of six

126 Quoted in ibid., 298.

127 Quoted in ibid., 308.

128 The songwriters are Sam Cooke, Curtis Mayfield, Smokey Robinson, William Stevenson and the Holland-Dozier-Holland team. See Jon Fitzgerald, “Black Pop Songwriting 1963-1966: An Analysis of U.S. Top Forty Hits by Cooke, Mayfield, Stevenson, Robinson, and Holland-Dozier-Holland,” Black Music Research Journal 27, no. 2 (2007), pp. 97-140. This article derives from Fitzgerald’s doctoral dissertation that involved analysis of close to 400 hit songs from the same time period to ascertain songwriting trends in a wider context. I am grateful to

categories: lyrical content, melody, rhythm, harmony, form/structure and production. He seeks the common musical details of these songs, which were historically important because they represented ‘black’ music shifting into the mainstream. Fitzgerald also finds similarities within the sample’s subsets, such as a songwriter’s fingerprint or traits associated with the Motown studio.

His analysis reveals that much of the crossover pop utilized elements from traditional black music, an expected, yet valid conclusion. Fitzgerald links gospel and the new mainstream style; that is, a preference for verse/chorus structures, the use of short, repeated harmonic progressions and

call-and-response phrasing. But it is also notable that some musical features shifted away from their historical roots, namely melodic structures, which, while still based on pentatonic scales, lacked the arch contour of gospel music.

Fitzgerald’s article can also be read in conjunction with his study of Motown music, in which he returns to the artists and songwriters for insight into the creative process.129 In doing so, one can locate the musical features within a particular context and establish a relationship between the music, the musicians and their aesthetic beliefs. One could further complete the picture by placing this context within a wider context (i.e. 1960s America, cultural change, ‘British invasion’130 etc.), but Fitzgerald’s work assists understanding of a significant period in popular music history.

Bowman’s and Fitzgerald’s studies are relevant in two ways. First, their

methods provide a model for studying the Nature’s Best songs. Their analytical tools, both at individual and corpus levels, are useful; likewise, both authors vindicate the need to conduct analysis with the musicians’ creativity in mind.

Jon Fitzgerald who, following email contact in early 2011, sent me his dissertation in unpublished form.

129 Jon Fitzgerald, “Motown Crossover Hits 1963-66 and the Creative Process,” Popular Music 14, no. 1 (1995), pp. 1-11. As the title suggests, Fitzgerald uses the same data as his doctoral dissertation and the previously cited article on black pop music of the same period.

130 Fitzgerald does make a brief comparison between Motown, “teen” hits and British pop songs with regards to the their form. There is a stark contrast between the Motown verse/chorus structure and the British tendency towards AABA forms, from which one could possibly draw further conclusions. Ibid., 4.

Second, the authors’ modes of presentation are exemplary. They aim to find trends and commonalities in quantitative terms — for example, 27% of the Stax songs followed a verse/chorus/bridge structure.131 Such an approach runs the risk of turning music into statistics, which is far-removed from creating, performing and listening to music. The authors circumvent this problem by constantly referring to the songs in question, demonstrating how the various musical parameters actually play out.

For example, when outlining the structures of The Astors’ ‘Candy’ and Eddie Floyd’s ‘Knock on Wood,’ Bowman highlights the common feature — the verse/chorus/bridge form — but also the subtle differences — phrase lengths and use of instrumentation in the bridge section.132 This is a basic point to make, but it is still important. From these studies, one understands how certain songs are musically connected, without losing sight of individual attributes, the idiosyncratic features differentiating one song from the next.

One could compare these studies to Tsai et al., who use computer and statistical techniques to blindly cluster singers. They envisage a scenario in which

someone has multiple unknown records and wants to group them according to the singer, thus separating original and cover versions.133 This is an interesting and valid research proposition.

Part of the problem is the authors’ statistical jargon, which will deter most musicians, but is appropriate for their discipline. That said, they do not refer to actual examples and consequently, one struggles to understand this work outside its academic context. Whereas one can engage with Fitzgerald’s analysis and hear it in the 1960s pop songs, it is not clear that one could read about blind clustering of singers and then use that information practically. The authors note their work is only introductory and call for further research134; however, the

131 Bowman, “The Stax Sound,” 294.

132 Ibid., 294.

133 Wei-Ho Tsai, Dwight Rodgers and Hsin-Min Wang, “Blind Clustering of Popular Music Recordings Based on Singer Voice Characteristics,” Computer Music Journal 28, no. 3 (2004), pp. 68-78.

134 Ibid., 76.

necessary path for an empirical or quantitative study is one grounded in the

“aesthetic experience” of music as demonstrated by Bowman and Fitzgerald.

Trevor de Clercq and David Temperley have recently analysed songs from the Rolling Stone “500 Greatest Songs” list.”135 The authors select the top twenty songs from each decade, 1950-2000, and subject them to a Roman numeral harmonic analysis.136 Although, the study is narrowly focused, compared to Bowman’s or Fitzgerald’s, it is relevant in this context because it engages similar methodological issues to the present study.

Their corpus is comparable to Nature’s Best because both lists arose when a group of people, at a certain historical juncture, voted for these songs. As de Clercq and Temperley point out, “there is disagreement as to what rock is and what it is not.” Therefore, there is no guarantee that the “500 Greatest Songs”

are “rock & roll” (whatever that may be) or that they are the “greatest” (however that may be judged)137, a situation analogous to Nature’s Best. That said,

neutrality may be sought but will never be achieved. Thus, for corpus analysis, the Rolling Stone list, like Nature’s Best, is a good starting point.

Furthermore, de Clercq and Temperley create a set of data against which the current results can be compared. The study reveals the frequency of particular chords (in terms of Roman numerals), the distribution of two-chord progressions and common three-chord progressions. The authors also present the distribution of harmonies according to each decade. Although acknowledging the limitations of a small sample, they contend that rock’s harmonic language “matured” in the 1960s and then changed very little in the following decades. Given Nature’s Best covers similar years, this makes for a potential point of comparison. The authors suggest, at the end of their article, that new corpuses could substantiate

135 Trevor de Clercq and David Temperley, “A Corpus Analysis of Rock Harmony,” Popular Music 30, no. 1 (2011), pp. 47-70.

136 Since their initial research, de Clercq and Temperley have expanded the list to include the next 100 highest placed songs not already analysed. The songs used in this study can be found at http://www.theory.esm.rochester.edu/rock_corpus/ (accessed 17 May 2011).

137 De Clercq and Temperley, “A Corpus Analysis of Rock Harmony,” pp. 50-51.

their findings or offer different perspectives on popular music harmony.138 This research will, therefore, complement their forays in this area.

Finally, Walter Everett’s mammoth study, published as The Foundations of Rock, references approximately 6500 songs, including the 2459 that appeared in the top twenty of the Billboard “Hot 100” singles chart between 1955 and 1969.

As Everett states, this era represents the “cauldron out of which rock was born.”139 The book “explores every domain of rock and pop recordings in greater depth than experienced anywhere else.”140 Everett presents chapters on structural, production and melodic materials as well as discussing the range of instruments appearing in popular music of this period. His main focus, however, is harmony with four chapters moving from basic chord construction through diatonic harmonies to more complex chromatic progressions.

Unlike the aforementioned studies, Everett presents his analytical results in quasi-list form, similar to Moore’s “Appendix” of harmonic progressions in rock music.141 Thus, in the chapter on “Chromatic Harmony,” he compiles a table in which are outlined twelve voice-leading patterns; under each heading, he lists the songs featuring these patterns and the particular harmonic

progression used.142 Everett aims not to discover the most common occurrences of particular musical features143, but to provide requisite information so that “a listener should be able to identify the sources of any and all sounds in a


The results in this thesis will not be directly comparable to Everett’s. But, his research is useful as it provides a resource of harmonic progressions, melodic lines, rhythmic characteristics and so forth. In the case of unconventional musical details appearing in Nature’s Best songs, they may be derived from an earlier era. If Everett is correct about the “embryonic nature” of 1960s rock and

138 Ibid., 69.

139 Everett, The Foundations of Rock, vi.

140 Ibid., v.

141 See Allan Moore, “Patterns of Harmony,” Popular Music 11, no. 1 (1992), pp. 82-106.

142 Everett, The Foundations of Rock, pp. 275-276.

143 Although one can often infer this from the number of examples presented.

144 Everett, The Foundations of Rock, x.

pop, then it is likely such a precedent could be found in The Foundations of Rock.145

This concludes the literature directly informing this research. A position has been established and justified of studying the musical text with regards to its musical parameters. Although formalistic, this standpoint complements, and is complemented by cultural, sociological, historical and receptive readings of the same songs. Ignoring these other perspectives does not diminish their

importance; rather, it is impractical to cover every angle in one project. The primary focus, thus far, has been theoretical and methodological issues without examining, in depth, analytical findings. These, and other, sources will be further considered in Chapters 4 and 5.