From Wikipedia, the free encyclopedia - View original article
Silbo Gomero (Spanish: silbo gomero [ˈsilβo ɣoˈmeɾo], 'Gomeran whistle'), also known as el silbo ('the whistle'), is a whistled language spoken by inhabitants of La Gomera in the Canary Islands to communicate across the deep ravines and narrow valleys (gullies) that radiate through the island. A speaker of Silbo Gomero is sometimes referred to in Spanish as a silbador ('whistler'). It was declared as a Masterpiece of the Oral and Intangible Heritage of Humanity by UNESCO in 2009.
Little is known of the original language or languages of the Canaries, though it is assumed they must have had a simple enough phonological/phonetic system to allow an efficient whistled language. Invented before their arrival by the original inhabitants of the island, the Guanches, and "spoken" also on el Hierro, Tenerife, and Gran Canaria, Silbo was adapted to Spanish by the last Guanches and adopted by the Spanish settlers in the 16th century and thus survived. In 1976 Silbo barely remained on el Hierro, where it had flourished at the end of the 19th century. When this unique medium of communication was about to die out in the late 20th century, the local government required all Gomeran children to study it in school. The language's survival before that point was due to topography or terrain and the ease with which it is learned by native speakers. It now has official protection as an example of intangible cultural heritage.
As with other whistled forms of non-tonal languages, the Silbo works by retaining approximately the articulation of ordinary speech, so "the timbre variations of speech appear in the guise of pitch variations" (Busnel and Classe: v). The language is a whistled form of a dialect of Spanish.
Manuel Carreiras of the University of La Laguna and David Corina of the University of Washington published research on Silbo in 2004 and 2005 arguing that Silbo was understood by the brain in much the same way as a spoken language. Their study of speakers of Spanish (some of whom "spoke" Silbo and some of whom did not) showed (by monitoring brain activity with functional magnetic resonance imaging) that while non-speakers of Silbo merely processed Silbo as whistling, speakers of Silbo processed the whistling sounds in the same linguistic centers of the brain that processed Spanish sentences.
Ramón Trujillo of the University of La Laguna published his book "EL SILBO GOMERO análisis lingüístico" in 1978. This work containing almost a hundred spectrograms concludes in a theory that there are only two vowels and four consonants in the Silbo Gomero language. In Trujillo's work Silbo's vowels are given one quality, pitch, either high or low. However, the work of Julien Meyer (2005 - in French only (pg 100), 2008) gives a statistical analysis of the vowels of Silbo showing that there are 4 vowels statistically distinguished in production and that they are also perceived so. Also in 2005, Annie Rialland of the University of Paris III: Sorbonne Nouvelle published an acoustic and phonological analysis of Silbo based on new materials, showing that not only gliding tones but also intensity modulation plays a role in distinguishing different whistled sounds.
Trujillo's 2005 collaboration with Gomeran whistler Isidro Ortiz and others ("EL SILBO GOMERO Materiales didácticos" - q.v. pdf link below) revises his earlier assertions to state that 4 vowels are indeed perceived (q.v. pg 63 ref. cit.), and describes in detail the areas of divergence between his empirical data and Classe’s phonetic hypotheses. Despite Trujillo's 2005 work acknowledging the existence of 4 vowels, his 2006 bilingual work ("El Silbo Gomero. Nuevo estudio fonológico") inexplicably reiterates his 1978 two-vowel theory. Trujillo's 2006 work directly addresses many of Rialland's conclusions, but it seems that at the time of that writing he was unaware of Meyer's work.
Meyer suggests that there are 4 vowel classes of /i/, /e/, /a/, /u, o/. However Meyer goes on to say that there are 5 perceived vowels with significant overlap. Rialland (2005) and Trujillo (1978) both agree that the harmonic of the whistle matches the second formant of the spoken vowels. Spoken /a/'s F2 and whistled /a/'s H1 match in their frequency (1480 Hz). However there is a disconnect in harmonics and formants near the frequency basement. Spoken speech has a wide range of F2 frequencies (790 Hz to 2300 Hz), whistles are limited to 1200 Hz to 2400 Hz. Vowels are therefore shifted upwards at the lower end (maintaining 1480 Hz as /a/) increasing confusion between /o/ (spoken F2 freq 890 Hz, whistled <1300 Hz) and /u/ (spoken freq 790 Hz, Whistled <<1300 Hz). In whistling the frequency basement must be raised to the minimum whistle harmonic of 1000 Hz reducing frequency spacing in the vowels, which increases misidentification in the lower vowels.
Trujillo (1978) suggested that the consonants are either rises or dips in the "melody line" which can be broken or continuous. Further investigation by Meyer, and by Rialland suggest that vowels are stripped to their inherent class of sound which is communicated in the whistle in these ways: voice (/k/ vs /ɡ/) is transmitted by the whistled feature -continuity. A silent pause in the whistle communicates +voice (/ɡ/), while a +continuous consonant gives the quality -voice (/k/). Placement of the consonant (dental, palatal, fricative) is transmitted in whistle by the loci of the formant transitions between vowels. Consonant classes are simplified into four classes. Extra high loci (near vertical formant loci) denotes affricates and stridents, rising loci denotes alveolar, medial (loci just above the vowel formant) denotes palatal, and falling (low loci) denotes pharyngeal, labial, and fricative. This gives 8 whistled consonants, but including tone gradual decay (with intensity falling off) as a feature on continuous and interrupted sounds gives 10 consonants. In these situations gradual decay is given +voice, and continuous is given +liquid.