What makes a Lament a Lament?

What follows are personal observations offered as  stimuli for research and debate; hypotheses, not facts. If anyone wanted to test them in an empirical, statistical study, I’d be delighted to collaborate. I am curious to know the degree to which each hypothesis holds true, but also conscious of serious hazards facing the researcher – such how the encoding of pibroch would handle cadence E.

With this in mind, my doctoral research is focusing on cadences and other questions raised by the diversity of notations found in pibroch’s source material. It is not like analysing Palestrina’s polyphony, or Haydn’s string quartets: these are relatively straightforward to encode because they use a consistent form of notion and complete editions exist to scan. Pibroch, by contrast, presents an encoder’s nightmare. The notation is far from consistent, the editions are far from complete, and it is hardly worth encoding a smaller sample than the 312 tunes transcribed before 1841 because you need to randomly set aside half the data for the results to be taken seriously. Limiting the sample to what has been published would skew findings to reflect the preferences of 20th-century editors, and I am more interested in the thinking of pibroch’s early transcribers.

For experimental results to be valid, the design has to be robust and I’d like to thank David Huron for helping me think how to approach this. Without actively seeking failure, the results will reflect personal bias and prejudice rather than scientific truth. David delivered an inspiring lecture series at the University of Cambridge, Faculty of Music, and generously helped me to articulate these hypotheses in February 2013. I’m publishing them here, untested, in order to help my co-editor, David Hester, to develop the ‘Explore by Genre’ pages of this site:

Hypothesis 1. In light of the higher pitch exhibited by vocal crying, we predict that laments will tend to exhibit a higher tessitura, particularly in the second halves of grounds which, for psychoacoustic reasons, leave a more powerful impression on the listener. [N.B. a single variation doubling from each work suffices for the sample – perhaps only the second half.]

Hypothesis 2. In light of the social functions of laments (assuaging grief and leaving lasting memorials), we predict that the grounds of laments will be longer than those of non-laments.

Hypothesis 3. In light of hypothesis 1 and the leading role that crathaidhnean (double beats) play in sustaining pitch height, we predict that lament grounds will exhibit more crathaidhnean.

Hypothesis 4. In light of the resemblance of the sound of finger strikes to the ‘cracking’ or ‘breaking’ of the human voice in crying, we predict that crathaidhnean in lament grounds will exhibit a larger average pitch excursion (i.e. that double beats on D and E carry greater weight as lament markers than other double beats).

Hypothesis 5. In light of the emotional state associated with grief, we predict that lament grounds will exhibit more dramatic surprises, less internal repetition, and less geometric order in their construction.

Hypothesis 6. In light of psychological research investigating examples of non-pibroch music that move people to tears,  we predict that laments are more likely to contain falling sequences than non-laments.

I predict that these six hypothetical characteristics of a pibroch lament carry different weight and are present in different proportions, with only a couple needing to be present for a pibroch to be perceived by a cultural insider as a lament. Knowledge of other laments is what defines the insider: someone who has subconsciously formed multiple connections  within the repertoire. Allan MacDonald has suggested that there were two types of lament: a more anguished keening lament and a more consoling, monumental cumha.

Psychoacoustic considerations add complexity. For example, how do we measure higher tessitura? Lets consider the laments for Mary MacLeod, MacSwann of Roaig, and Cameron of Dungallon. The brain does weird and wonderful things interpreting physical acoustic input, and what is required here is not simplistic note-counting but something that approximates the complexity of human perception. Why does my brain – perhaps yours too – perceive a high tessitura in these laments? I suggest that sustained high stretches in the second half can be just as powerful as a high tessitura across the whole ground.

I’d look for reinforcement between different lament markers and put everything that doesn’t fit the pattern centre-stage. For example, the high tessitura of Salute on the Birth of Rory Mor is recognised as a evoking joy, not grief. Note its melodic liveliness and the absence of repeated double beats, wriggling about rather than lingering on one pitch for any length of time. Looking for things that don’t fit, rather than things that do, is the best way to sharpen our understanding.

Two insoluble questions arise – aspects of interpretation that I don’t think can be tested empirically because our notations do not provide sufficient resolution. They are:

  • In the internal timing of double beats, were performers at an earlier period more likely to lengthen the second finger strike, and to increase the separation of the two strikes, in laments?
  • Were they more likely to play longer appoggiaturas (or ‘held streams’) in laments?

Recent psychological research explains why, of all the pibroch genres, laments are valued the most highly. This Wall Street Journal article provides a good introduction. This in turn helps us to understand why non-laments might gradually come to be played like laments. In an environment in which each piper only presents one tune, lament treatment turns a non-lament into a superior piece of ‘listening music’.


1 thought on “What makes a Lament a Lament?

  1. So, what we need is a methodology to help us test these hypotheses. Short of direct psycho-acoustical research, are there empirical methods we can use that will help a) prevent us from predetermining outcomes and b) be replicable and repeatable, such that others can assess and take on themselves?

    What I have begun doing on the AltPibroch.com website is a very tentative exploration into mildly-rigorous efforts to develop a method of empirical musicology to start a conversation about genre definitions and boundaries. (A more rigorous method would be to scan and statistically analyze the material in depth. Mine is more a manual count through observation.)

    But what I think is defensible is that a) I present a method, b) I use the method, c) I lay out the results. The method includes: a) assemble all tunes whose titles make reference to a given genre category (“gathering” or “lament”, for example); b) look for proposed indicators (for “gathering” tunes, Urlars built upon rapid-fire fosgailthe taorluath rhythms and woven tune structures); c) once done within the assembled “genre-titled” tunes, look at the broader corpus of tunes for possible inclusion despite lack of designation in the title; d) publish results.

    For laments, prior to the publication of these six hypotheses, I considered three proposed indicators:

    high tessitura – a majority of theme notes are above D (that the tessitura is expanded to high notes does not necessarily mean the tessitura is predominantly high pitched)

    double-beats – a significant number (admittedly impressionistic at this time), including hiharins (which are double beats); where a significant number is not found, the presence of dari movements may be considered (I think of dari movements not as keening, but as mimicking ululation – admittedly, this may be a cultural mis-match)

    complex tune structure – lyrical (interlaced, free, rounded) and progressive structures

    Interestingly, the resultant initial list (I have only begun to look at the whole corpus; there are 77 genre-specific titles) show 39 tunes that fit these criteria (including 4 unnamed tunes, and the Massacre at Glencoe). An additional 17 tunes had combinations of two of the three indicators.

    That left 26 tunes exhibiting only one proposed indicator (or none at all).

    Now, this represents a bit of a quandary: It could very well signal that these indicators are too exclusive. It could also mean that I did not weight them properly. But in looking at the list of “excluded tunes”: woven structure, absence of double-beats (except for the occasional presence of a hiharin or two), and a generally low tassitura would suggest these tunes are either NOT “laments”, or that a significant change in audience expectation, performance and setting may be represented by these tunes. Without having looked at Allan MacDonald’s proposal, it may possible that these “excluded” tunes represent “consoling” laments. (This needs to be tested.)

    I suggest that the result of such an empirical approach to genre is an awakening to difference long since elided by the deliberate performance of pibroch as a single genre, taking place over hundreds of years; in particular, the deliberate performance of all pibroch as “laments”. We have been living in a world where all cats are gray.

    Through genre analysis we are beginning to become sensitive to sometimes subtle, sometimes profound (and often obvious but overlooked) differences among the tunes when they are no longer viewed homogeneously. Not only do we see “rowing” tunes and “gathering” tunes (and “preludes”), each of which bring different demands to the performer with respect to tempo and expression, but we may now begin to see different types of “laments”.

    I have published my results here and here. They are incomplete, but should keep the conversation going.

Leave a Reply