One of the first own projects that we launched in Cuerdas Pulsadas was the tablature search engine , that we introduced about five years ago and that basically has the tablature collection of Sarge Gerbode , although I always thought of expanding it by adding other similar collections to this search engine. The truth is that I always had in mind to carry out an analysis of the data of that collection, but it is only now that I have found the opportunity to do it as it deserves. I would like, before going on to share the analysis, to clarify some questions about it :  
Is this an exhaustive or rigorous study on the heritage of the plucked string?
At all, my intention is only to study a particular collection of tablatures, so the conclusions are partial and the sample is totally biased. I am also not a specialist in the field, just an amateur, so I do not have a deep field knowledge to answer some of the questions that can be deduced from the data. My only intention with this analysis is to display the data from that collection in a different way.  
Can any degree of reliability be given to the conclusions?
No, since the origin of the data is not either. We can not trust that the dates, authors, titles or modes are correct and neither can particular data, such as the difficulty of a tablature.   Having clarified that this is nothing more than a fun, curious and certainly interesting vision of a specific collection of tablatures, I share the conclusions.

Details on the origin of the data

The Sarge Gerbode collection has a spreadsheet that details all the elements, with information such as:
  • Author
  • Year
  • Mode
  • Difficulty
  • Title
  • Type of score
  • Document and page
  • People who have worked on the transcript (intabulando, revising, editing)
  • Instruments or type of ensemble for which the piece is intended
Therefore, it is a very interesting data source and practically ready to be analyzed.  

Data processing

Given that the source is well structured, it has not been necessary to apply many changes, although I would like to highlight three operations in particular, which will allow us to understand some of the final aspects:
  • The “year” field required processing, to correctly treat it as a numeric data and to eliminate incorrect values.
  • I added a calculated “decade” field, to group the data into time ranges.
  • The fields “title” and “subtitle” also required treatment, to extract the value words.
In this last point, it is worth noting the elimination of what is known as “ stop words “, that is, those words in different languages that do not have a specific meaning (prepositions, articles, conjunctions). This means eliminating “el” in Spanish, “the” in English, “le” in French … a relatively easy task at present, since there are dictionaries in several specific languages on “stop words”. Specifically, I have eliminated the “stop words” in Spanish, English, French, German and Italian.

Here is the fun and interesting part, the data analysis:  

Tablatures timeline

This graph shows the evolution on the composition date of the tablature collection. Obviously, we cannot draw general conclusions about the distribution of the plucked string corpus over time, but it may be correct to highlight which periods there is the greatest interest in intabulating today (since, remember, this does not it ceases to be a collection of intabulations) or, for example, the existence of modern compositions for the lute. The sample covers from 1250 to 2011, with a marked volume in the period between 1600 and 1670.      

Composer analysis

Here you can see the distribution of works by composer, that is, a top of composers in this data set. Again, no general conclusions can be drawn about the plucked string, but rather about which authors are the most practiced or known, or which authors have had the greatest impact on having a complete catalog of their work. Beyond the anonymous that stands out in first position, we have some surprises in this graph. It is not surprising to find John Dowland, Miguel de Fuenllana or Francesco da Milano in the top ten, but at least for me it has seemed unusual to find Elias Mertel, Emanuel Adriaenssen since they are not habitual in concert programs or in the recordings.    

Composers Timeline

I consider this graph to be one of the most interesting views of the data set, mainly because it allows us to appreciate the evolution over time of the main composers, in a very illustrative overview.    

Timeline of modes

Another particular view of the evolution of the plucked string corpus that, in this case, allows us to analyze the evolution in the use of modes over time. In this view, I find it interesting to highlight what may be specific or transitional styles, such as the use of the B flat major and A minor modes around the decades of 1590 to 1610. Surely someone has a well-grounded answer to some of the questions that this graph may pose.    

Global use of modes

If we condense the previous view into a model where time is not a variable, we obtain this view in which, in general, we can see which modes are the most used. Quite sharply, lute music, regardless of epic, is articulated around sol, do, re and fa.    

Global use of musical forms

I also found it interesting to know which musical forms are the most practiced, so that this view allows you to see it in a single graph. The most representative ones should be highlighted, such as the song, the madrigal, the fantasy, the pavana, the gallarda or the intabulation of vocal music. Perhaps it would be interesting to add to this information the time variable, which would allow us to analyze the beginning, emergence and decline of musical forms as styles and fashions advance in time.    

Difficulty analysis

As this is clearly subjective, it is also relevant to analyze what degree of difficulty we perceive in the corpus of the plucked string. Clearly, there is a marked distribution in medium difficulties (values two, three and four), although there is also a small portion around the highest difficulty. For the most curious, 50% of the highest difficulty pieces are from our beloved John Dowland, followed by Daniel Bacheler and Alessandro Piccinini, in that order.    

Analysis of the difficulty over time

If we add the time variable to the difficulty analysis, we can see that difficulty has gradually been incorporated into the interpretation. Clearly an ascending line is perceived, a marked tendency to increase the complexity in the music as the styles became more refined and the language of the lute became richer.    

Analysis of the most used words in titles

Perhaps one of the most entertaining views of the analysis, as it allows us to see which words are used the most to title the works. Musical forms appear predominantly, since many compositions do not have a specific title. Thus, we can see gallardas, duets, preludes, fantasies, ricercares, pavanas, courantes, alemandas … But beyond the musical forms, it is fascinating to find words like “love” (love, amour), “lady”, “lord”, “queen”, “away”, “dear”, “dance”, “sweet”, “delight”, “fortune”, “jour”, “eyes”, “night” … which I think very well reflect the intention and purpose of plucked string music.    

Leave a Reply

Your email address will not be published.