While doing research on the Whistling Language of La Gomera, I came across a reference which I believe to be the earliest mention in literature of the existence of any whistling language. This was found in Bartolome de las Casas' 1527 "Historia de las Indias", written in the margin at the end of chapter XL, and is an excerpt from the monumental work of the Greek historian Diodoro of Sicily (fl. 1st century BC).

Translated into English, the relevant quote from Diodoro's work "Historia" (Book 3, chapter 13) reads "and thus they spoke, not only like men, but like birds singing".

I have not yet had the opportunity to see the Greek original, but I would imagine that Casas' translation is as faithful as any.

There are certain parallels in Diodoro's fanciful account which seem quite significant when related to Gomera's whistling language and other evidence of whistling for communicative purposes in the Canary Islands:

1. "una isla redonda"

The islands of La Gomera and Gran Canaria are both almost round.

2. "la lengua cortada, por medio a la lengua"

This is an obvious parallel to "Le Canarien"'s 1402 account of "& leur fit tailler leurs langues" ("and their tongues were cut"). Diodoro refers to a people whose tongues were cut lengthwise (medially) to give the appearance of being forked. This comes up again below in my comments on Nowak's article.

3. "Tienen de costumbre vivir hasta cierta edad y llegados allí ellos mismos se dan a la muerte"

This practice is significant in view of La Palma's tradition of "vacaguaré". When one felt death approaching, s/he would call together relatives and friends and announce "I want to die". S/he was taken to a cave, given a bowl of milk and the entrance was sealed. If they were given a special drug as well (as Diodoro mentions), it is not noted in my references to Palma's "death wish" custom.

4. "las mujeres tienen comunes"

All the earliest accounts of La Gomera (and Gran Canaria) mention that women were treated as common property.

5. If Diodoro's account were deemed to be an early reference to a Canary Island whistling language, it would prove that the original islanders' tongues were cut (and that they were exiled) long before the fall of the Roman Empire, ie. that this mutilation/deportation was not carried out by Hunericus and the Vandals in 477 AD as some researchers have suggested.

Because there are so many fantasy elements and other indications that do not jibe, it's tempting to dismiss Diodoro's entire quote out-of-hand. Navigating south (hacia el Mediodía) from Ethiopia one could run into Madagascar, the South Pole, or perhaps Australia. 5,000 stadia (an "estadio" is an ancient Greek measure of length = 606.75 feet) are approximately equal to 210 leagues, but a round island 575 miles in circumference and 183 miles in diameter does not exist in the neighborhood indicated (Gomera is 38 miles around, Gran Canaria is 96 miles around). Nonetheless, because some of the traits mentioned above hit so surprisingly close to home, I feel that a certain relevance and validity must be attributed to this ancient mariner's "fish story".

* * * * * * *

Everyone seems to agree that the whistling was developed by people whose tongues were cut (? out) and deported to the island, the implication being that the whistling language was developed because of the mutilation.

There is a reference to an annotation in the right-hand margin of pg 125 of "Le Canarien" (The 1402-5 / 1630 reference to La Gomera in "Le Canarien" is reproduced in the book "Whistled Languages" on pg 7.): "*Les relatiõs Angloises disent que ce furent les Romains. Voyés annot. Art. 28." (translated:) "The English say it was the Romans [who exiled them and cut (? out) their tongues]. See annotation Article 28."

I believe that I have located this "annotation 28" in Jean de Bethencourt's "Traicté de Navigations" on page 222. Since the Paris library refused to let me make a photocopy of this page (because of the age of the book) we'll have to content ourselves with the excerpt below:

(translated:)
"Origin: Some think that these people originate from Africa, and from there they were relegated to these islands by the Romans, who first cut (? out) their tongues for having blasphemed against their gods."

Many researchers have postulated that the Norman's earliest description of the people and language of La Gomera refers to its whistling language, but I'm by no means convinced that "le plus etrange langaige" (translated:) "the strangest language" is a reference to the Gomeran whistling language at all.

I believe that the Normans stayed close to the coast, and that since the whistling language was mainly used in the hills, they would not have become acquainted with it at all.

Even had the Normans gone into the interior, they would probably not have recognized a whistling language as such, more likely attributing the whistles to bird songs.

And had they encountered a whistling language, Bontier and Le Verrier would have written exactly that, ie. "un langaige sifflé". I'm quite certain that the verb "siffler" did exist in 15th century French.

* * * * * * *

With the assumption that the Normans' 1402 account is not a reference to the Whistling Language of La Gomera, I began the search for the first actual mention of the Gomeran whistling language in literature. As far as I've been able to determine, the very first specific mention of the whistling language on La Gomera is found in the account of Count Peraza's assassination either in Castillo/1686 or in Marín de Cubas/1687. Both of these authors later had revised editions of their work published (51 years later in Castillo's case, and only 8 years later in the case of Marín de Cubas). The question is not so much which of these two were the first to include whistled communication as part of the story of Peraza the Younger's death in 1488, but why it took two centuries for this to trickle into the history books!!!

Augustín Darias Ramos, the curator of the Casa Colón (The Columbus House) in San Sebastián de La Gomera, suggests that it may have been because many important documents were lost and burned in the 16th and 17th centuries during the various pirate raids.

Had the whistling language become extinct, its connection with the story of Peraza's demise would have seemed a mere romantic myth. In view of the whistling language's survival (probably thanks to Bishop Miguel de la Serna), the story of Count Peraza's assassination in 1488 becomes the crucial proof of the existence of Gomera's whistling language before the conquest!

There are earlier accounts of the 1488 Gomeran rebellion, but none of them include any mention of the whistled communications associated with the death of Count Peraza.

* * * * * * *

Scientific interest in Gomera's whistling language did not begin until the late 19th century. Quedenfeldt, in 1887, was the first to attempt a graphic representation of the Gomeran whistle. His main problem was that common music notation isn't (and will never be) flexible enough to notate a whistling language or any other form of "sliding" sonic phenomena (bird calls, whale songs, etc).

Also it's clear that Quedenfeldt should have used a violinist or trombonist (for example) rather than a pianist to conduct his research.

Even if he had used a violinist or trombonist, however, he still would have been stumped for a form of notation that others could "read" and consequently imitate with accuracy. He was on right track, but the tools he needed hadn't been invented yet.

* * * * * * *

Lajard's 1891 article, which is essentially correct regarding the mechanism of the silbo gomero, was summarily ignored; whereas Verneau's misunderstandings were repeated everywhere from encyclopedias to tourist guides.

* * * * * * *

Verneau's articles on the Gomeran whistle firstly deal with the anecdote of how his guide, answering whistled inquiries about the foreigner he was with, bragged about "the famous doctor from Paris" - Even after Verneau had specifically asked the guide not to let it be known that he was a medical doctor. This episode dramatically illustrates the effectiveness of the whistling language.

Secondly, Verneau's articles concern a demonstration of whistling techniques. The good doctor did not understand his guide's demonstration, and neither did anyone else who read his description of it.

Verneau must have asked his guide to show him "How you whistle". The guide obviously understood this to mean "Show me the various possible ways to whistle", at which point he promptly demonstrated several whistling techniques using different finger combinations.

There are at least 16 different ways to produce a whistle, and Verneau lists 8 of them. Verneau naïvely mistook his guide's demonstration to mean that changing finger positions is essential to the production of the whistling language.

The only whistler that Verneau would probably have seen in action at point-blank range would have been his guide. The other whistlers that the guide was communicating with would have been completely out of visual range or at least far enough away that it would be difficult if not impossible to make out exactly what a distant whistler was doing with his fingers.

The guide would have also been walking in front of the good doctor on the narrow trails, and while whistling, the guide would have turned away from Dr. Verneau in order to spare the his charge's ears. This means that Verneau would have had very little opportunity to ascertain exactly "how" his guide whistled.

At close range, however, Verneau would have been close enough to hear the grunts, etc that often accompany the physical effort of producing a loud whistle. This explains his reference to the larynx, as he would have assumed that these extraneous noises are somehow associated with the whistles and possibly aid in their comprehension.

Quite simply put, Verneau didn't stay on La Gomera long enough to become more thoroughly acquainted with the Gomerans' art of whistled communication.

What would Verneau have written if he had stayed long enough to really observe the phenomenon more closely? It's hard to say, but Verneau was trained as a medical doctor, not in the sciences of linguistics or phonetics, nor in the physics of sound.

* * * * * * *

In the book "Whistled Languages" (on page 3), there is an attack (probably by Classe) on Verneau's description of his encounter with the Whistling Language: "[Verneau's account is] not merely inaccurate but so wide of the mark that it appears to have been drawn from imagination rather than actual observation or even memory."

It seems strange that it never occurred to either of the authors of "Whistled Languages" that what Verneau saw was nothing more nor less than 8 different methods of producing a whistle. This condemnation is even stranger considering that on page 5 of "Whistled Languages" photos of five different techniques are shown. Clearly a case of not being able to see the forest for the trees.

After the Classe's initial attack on Verneau, the "relevant passage" (in French, with no translation) is "reproduced" (in fact it is badly garbled!!). If it is difficult to understand Verneau's quote in the original, it is absolutely impossible to comprehend the misprinted quote on page 3 of "Whistled Languages" (see Verneau's 1923 article).

On page 4 of "Whistled Languages" the unprofessional outburst continues with "This is pure fantasy and one could be forgiven for thinking that the writer had no first-hand knowledge of the [Gomeran whistle or the whistlers]." To err is human and to forgive is, at the very least, humane.

Line 11 of page 4 continues in the same vein stating: "As for the 'multitude de sons' [(translated:) multitude of sounds], this is palpable nonsense." Firstly, I do not feel that these strong words befit the scientific objectivity of a book of this type. And secondly, this is semantic hair-splitting. Considering that a multitude of words, phrases, functions and ideas can be expressed in the Gomeran whistling language, each one with its own distinct "sound", it is quite understandable that Verneau chose this term in his description.

Nowhere does Verneau imply changes in timbre, and it must also be taken into account that he was not trained as a musician either. Remember that there are many meanings that can be attributed to the word "sound".

Yes, it's true that Verneau misunderstood his guide's demonstration, and that his misunderstanding was misunderstood by everyone else (until now). Of course this was quite unfortunate, but we all make mistakes from time to time, and it's always best to rectify with diplomacy.

After having questioned Verneau's scientific integrity, Busnel/Classe inexcusably used Verneau's incomplete "Tu vas a misa hoy?" example from his 1923 article "Le langage sans paroles" (see "Whistled Languages" pg 9), instead of the authentic Quedenfeldt example which Verneau was referring to.

(qv. Quedenfeldt's "Pfeifsprache auf der Insel Gomera" <Zeitschrift für Ethnologie> XIX 1887 pg 739).

Verneau omitted 3 things from Quedenfeldt's work, two of which are absolutely essential:

a) Alltto. = Allegretto (not so essential)

This is a time indication. It means a little slower than Allegro (120-168), but faster than Andante (76-108).

b) portando = sliding, gliding, portamento (essential)

There is certainly no doubt that whistling the six consecutive tones (beep, beep, beep, etc) in the manner shown on pg 9 of "Whistled Languages" would result in zero comprehension.

While I'm not suggesting that one would have greater success by whistling Quedenfeldt's example in the manner it was written (sliding from G up to C, then sliding from G up to E, and finally sliding from G up to high G), at least one could not be accused of perpetuating Verneau's errors (especially considering that the authors of "Whistled Languages" dedicate the first several pages of their work to doing their utmost to destroy Verneau's credibility).

c) the slur/phrase markings (essential)

These show which note slides up to which other note, ie. they define the notes affected by portando.

* * * * * *

Page 109 of "Whistled Languages" finishes with "Fortunately many of them have been studied and recorded and will"...

There is no conclusion to the book's conclusion (Page 110 begins with "Acknowledgements"), and although I wrote Classe (care of the University of Glasgow) to ask about this, my letter was returned to me "addressee unknown".

* * * * * * *

At about the same time, (late 1970's) Herbert Nowak wrote an article in which he tries to quash the theory that Gomera's whistled speech was originated by men without tongues: (translated:) "the whistling language is not possible without a tongue, because it is precisely this [organ], by means of changing the size of the space in the oral cavity, that effects the changes in pitch".

Originally, my argument to Nowak's case was that it depends on how much of the tongue is cut off. But in view of Diodoro of Sicily's account (see above), the question is how the tongue is cut. Up till now no one has ever proposed that the original inhabitants' tongues were cut medially (lengthwise), and all researchers have tacitly assumed that the ends of the tongues were completely removed. It must be stressed that I do not see a medial cut as inconsistent with an interpretation of the French verb "tailler" or the Spanish verb "cortar".

(The English and the Germans have always chosen to translate the French verb "tailler" as "cut out" and "ausgeschnitten", quite possibly to the detriment of the original meaning.)

Not only would a "forked tongue" not impede whistling, it would almost inevitably produce whistling when attempting normal articulation. In fact it is entirely possible that the first exiled Gomerans' tongues were cut medially (as in Diodoro's account) as opposed to laterally, as everyone has always assumed.

* * * * * * *

Ramón Trujillo's 1978 book "EL SILBO GOMERO análisis lingüístico" is the most recent scientific work on the subject.

In 1990, this work was translated into English by J.Brent (entitled "The Gomeran Whistle: Linguistic Analysis"). In addition to translating Trujillo's monumental work into English, I also went ahead and took the time and effort to give it a much needed index (the Spanish original is not indexed). That book can be found in the Library of Congress, Washington DC.

Trujillo's 1978 work is mostly a refutation of Classe's 1956 "Phonetics of the Silbo Gomero" article. Here Trujillo claims that there are only 4 consonants and 2 vowels in the whistling language, and any and all other distinctions are figments of the imagination.

The whistlers themselves claim that they can clearly distinguish between all the different consonants and vowels, and do their best to pronounce (while whistling) to obtain the most accurate effect. But Trujillo's experiments with minimal pairs indicates that, even so, no matter how hard a whistler tries to pronounce the words distinctly, comprehension is limited or in many cases non-existent.

After reading Trujillo's book, one is left with the impression that communication in the whistling language is next to impossible, rather like the scientific proof that bumblebees are incapable of flight.

Trujillo's book deals almost exclusively with minimal pairs made deliberately confusing. In real life, humans talk in phrases, with questions and answers about situations and people which they are intimately connected with.

In face-to-face conversations that I had with my colleague Professor Trujillo, I voiced some of my opinions concerning the number of vowels, and confusion regarding consonants. He naturally sticks by his conclusions and also has the empirical data to back it up.

For example, I have mentioned to Trujillo that if the high vowels ([i],[e]) are in opposition in a diphthong, they can be perfectly recognizable. The vowel [i] is higher than [e], and the difference (to my ear) is undoubtedly noticeable when in direct opposition.

Recent research ("An Experimental Investigation of the Effectiveness of Training on Absolute Pitch in Adult Musicians" M.A. Rush 1989, Ohio State University) gives the definite impression that any human with normal hearing is capable of developing perfect pitch. It has long been known that relative pitch may be acquired by anyone, but only recently has it become public knowledge that perfect pitch might also be attainable by all.

In the case of the Gomerans, these innate abilities of perfect pitch and relative pitch (although the Gomerans are not aware of them as such), in combination, may (theoretically) be developed to such a degree that extremely precise nuances can be recognized and understood in the same way as any true living language, ie. NATURALLY, because the language is picked up from childhood.

At this point, this is all conjecture on my part, since I do not have a large body of empirical data consisting of phrases (as opposed to minimal pairs) to back up my hypotheses. However, I'm working on it and I hope one day to publish an exhaustive study to verify this once and for all.

Until I presented Trujillo with my (heavily annotated) copy of "Whistled Languages" in March of 1991, he was completely unaware of its existence. In light of the new material presented in said book, Trujillo has agreed to produce an updated version of his linguistic analysis which takes into account Classe's phonetic breakdowns presented therein.

[I have recently discovered that Trujillo has indeed authored another book on the subject of the silbo gomero. However this long awaited work is probably not by any means the one that Professor Trujillo envisioned back in 1991. This is the description of the course of study of the whistling language that has now become (since 1999) mandatory for all Gomeran children to learn in school. The book itself is available online in its entirety in pdf format El Silbo Gomero: Materiales Didácticos (© June 2005). While very academic, it will prove extremely interesting to any serious researcher.]

Classe's hypotheses in the book "Whistled Languages" regarding the phonetics of the Gomeran Whistle differ little from those put forth in his original paper published in 1956. For the benefit of future researchers, I have listed below the areas of phonetic divergence existing between the works "Whistled Languages" and "EL SILBO GOMERO análisis lingüístico".

Breakdown of the main areas of phonetic divergence between Classe and Trujillo:

(dc = dip continuous,     dd = dip discontinuous)
(rc = rise continuous,      rd = rise discontinuous)

Trujillo's work has almost a hundred spectrograms and other examples to support his conclusions. In contrast, the conspicuous absence of silbo gomero sonograms in "Whistled Languages" (only two - one on page 9, and the other on pg 33 fig. 16 #2) has a tendency to imply that Classe's conclusions were arrived at by subjective examination as opposed to objective observation of a large body of empirical data.

* * * * * * *

To wrap up this preface, I'd like to address two points which I feel need to be set straight. The first regards the question of the significance of relative volume changes with respect to comprehension. All other researchers categorically state that volume changes have zero influence on recognition of semantic features.

1. It has always proven inconvenient to devise a 2-dimensional graphic representation system which not only shows the pitch and duration features of whistled "curves", but which also shows the relative volume changes as well. I believe that it is for this reason that researchers working with sonograms/spectrograms, etc have chosen to ignore this facet of the characteristics of whistling languages.

2. A major problem with the sonograms/spectrograms presented in scientific works on the subject is that they all express frequency (Hz) as an arithmetic progression. In reality it is a logarithmic progression.

Example:

Whistling any given word or phrase in a certain range will produce some kind of curve. If you were to whistle said word or phrase a fifth lower against an arithmetic back-drop the curves would be flattened out. Similarly, whistling the same given word or phrase a fifth higher against an arithmetic scale would cause the curves to be exaggerated.

Using the proper logarithmic scale as the "backdrop" would mean that the curve's contours would look identical no matter what range said word or phrase is whistled at. In music, the resolution is 12 "steps" to the octave (the frequency doubles every octave). At this resolution the formula for each "step" works out at the 12th root of two. Doubling the resolution would make the formula for each consecutive step the 24th root of two. In order to make the read-out as accurate as possible a resolution a good deal higher than that is required.

* * * * * * *

I would hope that some of the above may prove useful and interesting to others investigating the phenomenon of whistled communication systems, and that various areas of confusion (both scientific and historical) have been resolved thereby.

* * * * * * *

Jeff Brent studied for his bachelors in mathematics and music at the University of Colorado. He is a professional musician. He is also multilingual and was formerly the director of the Santa Cruz de Tenerife (Canary Islands) branch office of The inlingua Schools of Languages (the world's largest organization of language schools).