There are many studies on Chinese shortenings that cover almost each of their aspects. However, with regard to the principles that govern the formation of Chinese shortenings, most merely list and then explain them by using few examples. Do these principles really play their roles? Do other important principles exist? To what extent do these principles play their roles? How do these principles work together to produce various kinds of shortenings for different bases? Do these principles function universally across all shortenings in different languages, or do they merely work in the formation of Chinese shortenings?

To answer these questions, a big corpus comprised of 9036 Chinese shortenings is built. By surveying all Chinese shortenings in this corpus, especially the shortenings for tetrasyllabic and pentasyllabic bases, many useful findings about the principles governing the formation of Chinese shortenings are identified with the aid of statistics. Detailed statistics can explain whether these principles play their roles and to what extent they work. The main principles are as follows. No function words are chosen in shortenings. Shortenings take the same order as their bases. The length of shortenings is bisyllabic for tetrasyllabic bases and trisyllabic for pentasyllabic bases. Shortenings fully or partially represent their bases. First syllables, core morphemes, formal morphemes, or prominent morphemes are chosen for the shortenings. The morphemes in a shortening should be in accordance with one another in terms of style. Shortenings maintain the same syntactic structure as their bases. Shortenings should avoid ambiguity. The colloquial morpheme in a base in a colloquial context is chosen in the shortenings. The shared morpheme is preserved in the shortenings. Shortenings should be concise. The morpheme in a component that is often chose as the representative will be chosen again for another shortening. No morpheme in a redundant word in bases is preserved in the shortenings. The level, syntactic structure and prosodic structure that a base takes decide the lengths of its shortening to a great extent.

Next, the Correspondence Theory in the Optimality Theory is applied in this thesis because this theory provides the perfect explanation for how these principles (in OT, they are called constraints) work together. The constraints work together through ranking. After the evaluation of the rankings for all these constraints, based on previous conclusions, the only optimal shortening for a base is obtained. The overall ranking of these constraints for tetrasyllabic bases is NO FUN, LIN, COL-CON, 1st COR, NONAMB, PRO-MOR, ACC-STY » ECO, COR » FUL-REP, SHA-MOR, ACC-STR, FOR-MOR » FT-BIN, 1st + 1st, FIR-SYL » MAX-BS, CONTIG-BS. The ranking for most constraints for the pentasyllabic bases is NO RED, ANA, NONAMB, NO FUN » ECO » SHA MOR » SS-LEN » PS-LEN » FUL-REP » LEV-LEN, FT-TRI.

Lastly, by comparing cross-linguistically the shortenings in Chinese with the shortenings of three other languages (Korean, Japanese, and English), some universal principles governing shortening in different languages are obtained. the characteristics of Chinese shortenings are also clarified. Some reasons behind these differences are correspondingly offered.

(496 words)

.