The Ising Model for Changes in Word Ordering Rule in Natural Languages

The order of ‘noun and adposition’ is the important parameter of word ordering rules in the world’s languages. The seven parameters, ‘adverb and verb’ and others, have a strong dependence on the ‘noun and adposition’. Japanese as well as Korean, Tamil and several other languages seem to have a stable structure of word ordering rules, as well as Thai and other languages which have the opposite word ordering rules to Japanese. It seems that each language in the world fluctuates between these two structures like the Ising model for finite lattice. Model of changes of word ordering rules and Tsuonoda’s table Natural languages have been attractive objects of various interdisciplinary studies. The problem how the universals of languages evolve is drawing the attension of the physicists [1]. The importance of adpositions (prepositions and postpositions) as a parameter is recognized in word order typology [4, 5]. In previous studies [2, 6] using the word order data of Tsunoda (Table.1) [2], in which Japanese is taken as the standard of comparison, 130 languages are classified by using 19 word order parameters. The result is that the 130 languages are neatly divided into two groups with one exception or two: (a) prepositional languages and (b) other languages, i. e. postpositional languages and adpositionless languages. Adpositionless languages behave like postpositional languages in terms of other word order parameters. The second important parameter of word order is ‘numeral and noun’ [7]. Applying a statistical method for categorical data [8], the clustering of the 130 languages is well explained by the two word ordering parameters, ‘noun and adposition’ and ‘numeral and noun’. Greenberg [4] utilizes the position of adpositions in seven of the forty five universals he proposes (pp.110-12). For example in languages with prepositions, the genitive almost always follows the governing noun, while in languages with postpositions it almost always precedes it. Tsunoda’s data show, the parameters 1 (S, O and V), 3 (genitive and noun), 7 (relative clause and noun), 8 (proper noun and common noun), 9 (comparison of superiority), 10 (main verb and auxiliary verb) and 19 (purpose clause and main clause), have strong dependence on the parameter 2 (noun and adposition) [2, 6, 7]. For example our study of the data gives the statistical laws on cooccurrence with three parameters. (i) If a language has preposition, and if the language is the VO language, then the genitive tends to follow the noun.