As a working musician I have a repertoire of several hundred songs and tunes, most of which live in my head. This cerebral cataloguing system is suitably eccentric and intensely personal. Reference marks include people I learnt tunes from, where I was when I learnt the piece, narrative details, tune shapes, finger patterns, memories of past performances and many other details. In this respect I don’t suppose that I am very different from many other musicians. One of the biggest problems I have encountered with the FARNE project has been the challenge of documenting thousands of melodies and variations. I am a musician, I have no background in libraries, museums or catalogues. It was time for some research.
As long as academics have been involved with folklore research there has been a desire to classify and document material. Initially researchers were only interested in song texts. The classification of melodies being only a relatively recent addition. Dialect and language provided a reasonable indication of geographical location, whilst subject matter gave some clue as to function. The classification system we now know of as the "Child Ballads" was arrived at using only textual references, song melodies being disregarded.
Cecil Sharp in his researches sought to give the Folk Song of England a dignity above and beyond the mere functional. To this end he utilised Mediaeval Church modes to describe and define his recordings. Of the twelve modes, The first eight were designated in the mediaeval era, as a way to classify pre-existing Gregorian chants. The next four were added by Glareanus in 1547. This gave way to our modern system of tonality, especially via the Ionian mode. Connecting the Greek names of these modes to Ancient Greek theory or any other tangible existence in Greece is dubious at best. The origin of the names is a sort of historical homage.
Sharp felt that this classification would give an added note of authenticity to his theories of folk song as a pure form of higher worth and merit than mere popular music. More recently this practice has been brought into question. The suggestion being that any attempt to describe a body of music using terms derived from the definition of music from another era and social status could hold no true worth. However, it still forms a useful starting point from which to begin other explorations. Bertrand H. Bronson was one of the last scholars to use a modal classification system , when he published the volumes of «The Traditional Tunes of the Child Ballads» between 1959 and 1972.
A most persuasive argument was put forward by Finnish scholar Ilmari Krohn (1867-1960) for a lexicographic method which was later taken over by Bela Bartók and to a greater extent by Zoltan Kodàly. It was Krohn’s idea to define the differences between tunes in terms of the similarities which they shared with others. In this way patterns could be used to define families of melodies. For easier comparison, the tunes must be reduced to a common final. Thus simple information like the start note and highest note (ambitus) can reveal a great deal of information about a given melody.
More recently scholars have focused on melodic contours. Suggesting that the patterns by which melodies rise and fall can be their unique defining feature. In the book, "The Directory Of Tunes", Denys Parsons has catalogued many classical themes using a simple contour system.
Each of these approaches has it’s own merits and pitfalls. All approaches are time consuming. Their success or failure rests on the relative use of the final results. Of these only time can tell.
Consider the following example:
|Kerrs Merry Melodies, pub. Glasgow, late 19thC.
Assessing these melodies in terms of modal characteristics we find that the melodies numbered 70. and 72. are definitely mixolydian; 77.,78.,79. and 80. Ionian; 74. Aeolian and 76. Dorian. The Melodies numbered 71., 73., and 75. are more difficult to describe, lacking the fourth, they have only a six note scale. They are all definitely in the minor key, but could be either Aeolian or Dorian. Interestingly these three melodies also show characteristics of the Italian Renaissance Passamezzo form.
The Passamezzo, a harmonic structure used by musicians as a basis for improvisation or accompaniment served very much the function that a twelve bar blues might do today. Indeed chordal or harmonic structure can be a very useful way of describing tunes.
Matt Seattle in his book "The Master Piper - Nine Notes That Shook The World" describes a simple harmonic formula underpinning the border pipe repertoire which is based on a duo-tonic structure. The two Mixolydian tunes in our selection could be said to operate from two separate scales one Mixolydian and the other Lydian in the formula of three bars of one followed by one bar of the other.
To investigate the approach used by Bartok and others we must first assume a common starting note for the scale of each melody. Thus simple expressions of starting note, range (ambitus) and cadence or final note become significant in defining both differences and similarities between melodies. Treated thus our selection might be expressed as follows:
74. 1,1, 1:10
80. 4,5, 1:9
The first figure showing the final note of the first part and the second figure the final note of the second part. The next two figures show the lowest note (a minus sign indicating that it is below the tonic) and the highest note. Thus we can see that most tunes have an ambitus of just over an octave, tune number 77. being unique in having a range of two octaves. Most tunes resolve on the tonic, but some are left unresolved. In classifying melodies this information may be useful in associating them with particular instruments or against common historical structures. Resolution and compass could also be a reasonable factor in defining regional or geographical stylistics. Certainly much of the Northumbrian pipe repertoire is left unresolved.
Finally let’s look at the shape of our melodies. Denys Parsons system uses a simple code defining individual notes as being the same as, lower or higher than, the note which it follows. An asterisk is used to represent the first note, the letters D,U, or R are used to indicate melodic movement. For ease of reading letters are grouped in blocks of five. Once catalogued the asterisk is ignored (this being common to all). Thus our selection might read as follows:
70. DDRUD DUDDR UUDUD
71. DUDDU UDUUU DUDDU
72. UUUDD DDUUD UDUDU
73. URDDD DDUDU DDDDR
74. UUDDD RUDDR UUDRU
75. DDUDU DUUDD UDDDU
76. DDRRR RRUUD UUUDD
77. UUUDD UUUUU UDDRU
78. URUDD DDDDD UDUDU
79. RRDDU UUURR DUUDR
80. UDUDU UDDUD UDUUU
On this small scale it is difficult to draw any meaningful conclusions from this process. However, I suppose that on a larger scale this could be used to make a melodic data base which was searchable by those with little or no musical knowledge. It may also reveal common musical patterns, common beginnings or tune families.
Indexing by title reveals little about tunes and certainly makes it difficult to identify similarities. Indeed it is common to find folk melodies identified in different ways from region to region, or within a region names can change with the passage of time. Historical texts may display variations in spelling. For example The tune Timour The Tarter as known in Scotland and the Borders, is known elsewhere as Blanchard Races. Tune Name Indexing would not reveal this, some of these methods might.
At the end of this investigation I don’t feel that I have any clear answers. My Question has to be, of what value are these systems to the FARNE Project, do they help me in ordering this body of information and will they be useful to users in searching the database?
By Mike Hirst