Phonological Similarity between Basque and Other World Languages Based on the Frequency of Occurrence of Certain Typological Consonantal Features

The aim of the present typological study of the Basque language based on the frequency of occurrence of consonants is to compare its sound chain to sound chains of various world languages. Then, one can construct typological distances between Basque and other world languages. The structure of the frequency of occurrence of consonants in the speech sound chain is a good clue of understanding the typological closeness of languages. Typological investigation of human languages involves the task of the analysis of their speech chains by a human being or computer. A human being can realise that this or that language sounds closer to his own native language without understanding the meaning. At the stage when it is hard to teach computer to understand a human language, it is quite possible to make it recognise the sound closeness of a language to this or that language on the basis of the analysis of its sound speech chain. We have computed the frequency of phonemic occurrence of 119 world languages as a teaching sample for the computer. Then we took Basque as a token language. Actually, Basque was chosen for only one reason. It was taken up mainly because Basque as well as Japanese, Korean, Ainu, Burushaski, Nivh (Gilyak), Yukaghir are considered to be isolated languages, i.e. languages that do not belong to the known language families. In fact, they may be relics of the former family of languages (Crystal, 1992: 425). Basque is a fair example of an isolate language. Efforts have been made to show a relationship with Caucasian languages, North African languages, and Iberian, but none has been convincing (Crystal, 1992: 40). Merritt Ruhlen considers John Bengtson to be correct to include Basque, Burushaski and Shumerian into the new Dene-Caucasian family (Ruhlen, 1994: 25). R. L. Trask believes Basque to be unquestionably the last surviving pre-Indo-European language in Europe. He links Basque with the dead Aquitanian language (Trask, 2001: 72). We tried to find out the distances between world languages with the help of computer in order to achieve the optimal and unbiased results. The computer had to measure the distances between a chosen language and the rest of the languages in the set. Then, the computer had to put it in a matrix: closer to some languages and far away from the others.We have chosen Basque and the other isolated languages (e.g. Basque, Yukaghir, Nivh, Korean and Japanese) because they are not categorically assigned to any language family. It is interesting to compare how the computer and how different linguists place them in different language families and super-families: Indo-European, Turkic, Finno-Ugric, Tungus-Manchurian, Uralic, Ural-Altaic, Paleo-Asiatic; Sino-Tibetan; Austro-Asiatic; etc. We measure the distances between Basque and the other world languages in the same way we did it with Japanese (Tambovtsev, 1988: 19). We believe that under the circumstances even some traits and hints from the typological point may help to find the languages genetically related to Basque (Tambovtsev, 2001: 83-85). In this study we have used the procedures that are usually used in pattern recognition First of all, one must choose the features which should be necessary and sufficient (Zagoruiko, 1972). We believe the chosen features to be the most informative from the phonetic point of view. Basque, as any other human language, has a specific structure of the speech sound chain. It can be distinguished by its structure from any other language. Every language has a unique structure of distributions of speech sounds in its phonemic chain. The distribution of Basque vowels will not be considered until the second stage of the investigation, that is, if the data on the frequency of occurrence of consonants will not allow us to distinguish between the languages under investigation. Let's point out that consonants bear the semantic load in the word, not vowels. Therefore, it is more possible to understand the meaning of the message by consonants, rather by vowels. Some writing systems (Hebrew, Arabic, etc.) are a fair example of that since they denote only consonants. However, if we fail to recognise and distinguish two languages by consonantal groups, then we resort to the second stage of investigation, in which the frequency structure of occurrence of vowels in the speech sound chain is taken into consideration. While comparing languages, it is necessary to keep to the principle of commensurability. Having it in mind, it is not possible to compare languages on the basis of the frequency of occurrence of separate phonemes, because the sets of phonemes in languages are usually different. The articulatory features may serve as the basic features in phono-typological reasoning. First of all, it is the classification of consonants according to the work of the active organ of speech or place of articulation (4 features: labial, front, palatal, velar).