Word frequency and word order in freezes

The theoretical part of this analysis attempts to explain and predict sequencing phenomena by means of principles of cognitive economy: in order to avoid peaks of information that are difficult to process (Fenk and Fenk 1980), and also because of their easier 'lexical' and articulatory accessibility, the more frequent and therefore informationally poorer elements tend to occupy initial position (Fenk-Oczlon 1983a). In the context of 'freezes', this means 'more frequent (high-token frequency) before less frequent'. The arguments presented in this paper indicate that this new rule represents a principle thai is superordinate to such old rules as 'short before long', 'the first word has fewer initial consonants', 'me-first principle', etc. In order to discover whether these conclusions stand up to empirical examination, they were tested on the basis of 400 freezes using the corresponding statistical data (from Thorndike and Lorge, Josselson, Meier, Ruoff). The results include the following: - With 84% correct predictions, the new rule achieves by far the highest accuracy. - In paired comparisons (all possible combinations of five rules) no other rule achieves such a high degree of correspondenc e. - In order to explain those freezes which represent an exception to our rule, recourse must be had primarily to the iconic coding of spatial-tempo ral