Cohesive Ties and Chains in Good and Poor Freshman Essays.

This research compared cohesive ties and chains (Halliday 8c Hasan, 1976) in the good and poor essays of college freshmen. None of 18 different kinds of cohesive ties was used more frequently by the good writers than the poor. In addition, the cohesive distances from individual precursors to coherers did not distinguish the good from poor writing. However, cohesive chains (a series of lexical collocations, reiterations, synonyms or superordinates and their pronouns all semantically related to one another) exhibit significant differences in their use of available cohesive distance, the variety of word types, and the maturity of word choices as measured by word frequency in the lexicon as a whole. The publication of Halliday and Hasan's Cohesion in English (1976) engendered a large body of research by educators, much of which attempted to estimate the potential role of the cohesion system in teaching and evaluating writing. To do this, researchers analyzed student writing with the cohesion system, searching for the characteristics that would distinguish good and poor writing, different modes of writing, or the writing of different age groups. Eiler (1979) took an exhaustive look at the writing of 15 ninth-grade honor students and discovered that various kinds of lexical cohesion seemed to be the best indicator of the students' response to literature and that reference cohesion was the primary evidence of ability to sustain a selfsufficient ("endophoric," in the terminology of Halliday and Hasan) text without appeal to the non-textual ("exophoric") environment. Eiler preferred the terms precursor and coherer for Halliday and Hasan's presupposed and presupposing. Hartnett (1980) tried to teach the cohesion system to basic writers at a Texas college and then used counts of different kinds of ties as a criterion for evaluation of the essays. She had mixed results, with no significant differences found for teacher, treatment, or mode of writing for the experimentals over the controls. The teacher x treatment interaction was significant, but in general the correlation of holistic score with number of types of cohesive ties was quite low, only .21 for all essays. Cherry and Cooper (1980) studied average and superior writers at grades The author wishes to thank James Collins of SUNY Buffalo for his assistance in this research. Requests for reprints should be addressed to the author at the Academic Affairs Office, Canisius College, Buffalo, New York 14208. Research in the Teaching of English, Vol. 21, No. 1, February 1987