Chapter 2 - Method

Jump to

Table of contents — Introduction — Chapter 1 — Chapter 3 — Chapter 4 — Conclusion — Bibliography — The Spreadsheet — Summaries of analysed works — Thesis in PDF

Contents

2.1 Process overview

The purpose of this experiment was to determine the existence of a link between harmony and emotion in French music written in the early 18th century. This was accomplished by compiling a dataset of harmonic sequences and their associated text, in which to look for patterns in how the two are related. The process was straightforward: select the works to be analyzed, analyze them as precisely and exactly as possible, log the results, and then search for patterns within the compiled results.

2.2 Works analyzed

Due to the time constraints involved in conducting this research, strict criteria were applied to the selection process when seeking works to analyze. Those selected were dramatic works with instrumental accompaniment composed in France between 1700 and 1730 (with the exception of Castor et Pollux for reasons noted below). They were composed by prominent musicians, who were judged as such due to records of their works having been performed at court or significant religious institutions. From these works, only secco recitatives with detailed continuo figures were analyzed. The complete list of works analyzed is as follows:

Summaries of the above works can be found in the Appendix II.

The analysis process began with Rameau’s Castor et Pollux, despite the composition date of this work falling just outside the chosen period. Rameau was very thorough in his basso continuo notation, which would allow for faster analysis and more confident results. It was also supposed that Rameau, as a forerunner of the very concept of harmony, would make particularly thoughtful and precise figuring choices. Once the Rameau analysis was finished, the analysis proceeded with more speed and confidence. The next work analyzed was Jacquet de la Guerre, followed by Campra, and finally Clérambault.

2.3 Harmonic analysis method

There were two main options available to perform a harmonic analysis of recitative passages, namely: Roman numeral analysis, where each scale function is symbolized by a Roman numeral, or basso continuo analysis, where basso continuo figures describe chord structure. The present research required as precise and exact a harmonic description as possible, so both methods were considered.

Basso continuo analysis is more appropriate for the music of the time, as a vertical conception of continuo realization was still a very novel concept in the years that this study focused on. By its nature, basso continuo analysis describes very accurately the structure of each chord individually. It does not, however, give information about the context in which a given chord exists. Looking for patterns within a basso continuo analysis would result in grouping harmonic passages by their chord compositions—seventh chords together, first inversion chords together, etc.—but the chords in each group would potentially have very different roles within the harmonic phrase.

Roman numeral analysis, on the other hand, identifies each chord by its function within the scale. It describes each chord’s role in the context of the musical phrase. The analysis of a chord using this system, therefore, depends entirely on the chords around it. Unlike basso continuo analysis, however, simple Roman numerals give little to no information about the structure of individual chords. Looking for patterns within a roman numeral analysis would yield chords grouped by role—dominants together, tonics together, etc.—but the chords within each group would differ vastly in terms of structure and note composition.

Considering the above, the ideal choice was deemed to be a combination of both harmonic analysis methods. After all, the goal during this stage was to gather as much data as possible. Using both methods maximized the efficiency of the information obtained from each analyzed passage. It was not clear, before starting this research, whether patterns would emerge in the functions of chords, or in their structure, or both. It seemed unfortunate to risk missing potential conclusions by using only one method, when both are so complementary that using them simultaneously did not make the analysis considerably longer. It was also predicted that both analysis methods would yield hints as to where patterns might be found in the other.

Eventually, it was concluded that the importance of the analysis method lay first and foremost with its consistency. This study was not looking to describe how each composer conceived of their writing, or how their choices could be understood and interpreted within a historical context. On the contrary, the goal was to look only at the music itself and find patterns within it. To do so, all that was needed was for similar chords and chord progressions to be identified in a similar way. What exactly that way was did not matter, if it was always the same, and could easily be catalogued and identified at a later time. Both a basso continuo and a Roman numeral analysis were therefore performed on the studied musical passages. Before starting, a clear system of analysis was established for both, to ensure that the results would be consistent and allow for an efficient examination.

2.4 Notation system

Works where the composer was very deliberate and generous in their markings of figures were prioritized. This ensured that the analysis would simply follow the composer’s choices without the risk of ascribing them decisions they did not make. A figure was only added when the vocal line contained a note that was not in the written figure but was significant within the chord. In these cases, the system of basso continuo used by French composers at the time was followed, to avoid any later confusion.

The choice of Roman numeral analysis system required a little more reflection, as there were several different systems of categorization and notation to choose from. For purposes of efficiency and eventual pattern recognition, a system was chosen where all functions are indicated by capital letters, regardless of the mode of the chord. Most analysis systems taught in English-speaking Canada use lowercase letters for minor chords, but this felt redundant within this research, for two main reasons. The first is that the key of each passage was recorded separately, and once the key of a musical phrase is established, the mode of each chord within it is implied by its harmonic function. Cases where the mode of a chord was different from expected within the given key were dutifully recorded. The other reason is that, once the choice was made to use both Roman numeral and basso continuo analysis, it was clear that the basso continuo figures would fill the gaps of whatever the Roman numerals did not cover. For the same reason, chord inversions were not added to the Roman numerals.

It is important to note that the harmonic role was prioritized over the chord structure when the function of a chord could be described in different ways. A prime example is that of a diminished-seventh chord preceding a tonic chord in a cadential moment. Though such chords could be notated as VII, they were always notated as V—a dominant function with a missing fundamental. Indeed, within the combined analysis system, the context that Roman numeral analysis provides is particularly important, as it complements the chord structure description of basso continuo analysis. Contextually, notating a diminished seventh chord as a dominant when it is filling that role makes the most sense, especially since the basso continuo figures will clearly indicate that the fundamental is missing.

Finally, ambiguous notations in the facsimiles resulting from the random use of figuring in music of this period were generally left out, to avoid confusing the data by making choices that the composer may not have made.

2.5 Sample selection

At the beginning of the analysis process, it became apparent that the task of selecting which phrases and harmonic progressions to catalogue was extremely murky. In theory, it was predicted that the passages of note would leap out from the page, ready to be recorded. In practice, this was not the case; without an idea of harmonic rhythm or cadence structure in this music, it was initially difficult to determine which chords were significant. It was also unclear at first how passages should be delineated. Therefore, a large percentage of secco recitatives from Rameau’s Castor et Pollux were analyzed before beginning the selection process. This was very useful in bringing out the harmonic rhythm of the recitatives, as well as which chords and progressions were common. As it turns out, French recitative was a particularly judicious choice for this research; the rhythm and meaning of the text dictate the music it is set to, so the harmonic phrases are quite clearly delimitated. Most phrases are short and end with an obvious cadence. Once this became clear, it was easier to select the passages worthy of being catalogued.

When working on Rameau’s Castor et Pollux, more passages were selected than were probably needed. A frame of reference as to what was recurring, and therefore relevant, was not yet established. Almost all cadences were selected, as well as any passage including secondary chords or unusual continuo figures. Once the analysis had moved on to the selected works by Jacquet de la Guerre and Campra, it was possible to make more discerning choices as to what to keep. Harmonic progressions that were extremely common, like II-V-I, appeared too often to carry any specific meaning, so they were no longer selected. Particular attention was paid to any intense expression of emotion in the text, particularly a negative one, as those appeared most likely to be given special harmonic treatment.

2.6 Data cataloguing

Once the selection process was complete, the data was entered into a detailed spreadsheet. In this spreadsheet, each harmonic passage is logged in a single line containing the following information:

Back to top - To chapter 3

Footnotes

  1. Jean-Philippe Rameau, Traité de l’harmonie (New York: Broude Brothers, 1965), 64-65. ↩