In the Japanese speech, clitics are almost always indespensable to make a sentence, and children frequently use them in their word-chain utterances from very early period (at least by the end of one-year-period, and normally about the middle of that period). The simplest type of the utterances which contains such clitics is that consisting of one independent word [A] (in early childhood, it is mostly a noun, a verb, or an adjective) and one clitic [B]. And this AB-type may be one of those most frequently observed wordchain utterances in one-year-old children. Furthermore, as these clitics are small in number, these characteristics are very useful in studying the early formation of word-compounding function.<br>The syntactical functions of the Japanese clitics can be classified roughly into the following: (1) Those which combine words logically [B′]; and, (2) those which appeal the speaker's sentiment to the hearer [B′′]. B′ expresses among others, grammatical “cases” in the sentence, while B′′ furnishes utterances with interpersonal-humanistic “tones” which are essential in everyday conversations. The clitics are one kind of particles (articles, prepositions, conjunctions, interjections, and the like in English), and B′ may correspond to preposition, and B′′ to interjection. But Bs (B′ and B′′) are essentially “postpositions” and are more important and more frequently used in speech than the particles in European languages are.<br>Procedures are as follows:<br>1. Longitudinal speech samples of five normal one-year-old Japanese children were analysed to clarify the development of their word-compounding processes. Original speech samples were magnetic-taped records of speeches recorded 30min a day per week through one year.<br>2. In the analysis, frequencies of the same utterances were not taken into account, and only varieties of AB-type were considered.<br>3. As indices of the word-compounding, we used the following: (i) Overall average number of the utterances connected to a word; (ii) frequencies of every type of the combinations (e.g. one B combined to 3As; 2Bs combined to one A, etc.); (iii) average number of the utterances combined to a word in every type of the combinations; (iv) difference of ratios between the number of each subword-class in A and B.<br>The results showed:<br>1. Children K and Y, who showed remarkable progress in their vocabularies also showed clear superiorities in the development of word-compounding. In the indices (i) to (iii), they make steady progress in increasing the scores. On the other hand, Children S and E, who were poor in vocabulary growth showed no signs of the development.<br>2. Data in the last period in Child K are the one to which index (iv) can be effectivelly applied. The result showed the sub-differentiations of linguistic functions of A (A′ and A′′, A′ means noun, A′′ not noun but mostly verb or adjective) and B (B′ and B′′).