Abstract: The present study examined the possibility that users will model or can be shaped to the vocabulary and phrase structure of a program's output in creating their own inputs. Novice and occasional computer users interacted with four versions of an inventory program ostensibly capable of understanding natural-language inputs. The four versions differed with respect to the vocabulary and/or the phrase length presented on the subjects' CRT. One-half of the subjects were unknowingly restricted to input phrases identical to those used by their respective program, the other half were not. Additionally, one-half of the subjects communicated with the program by speaking, the other half by typing. The results indicate that recognition rates of natural-language processors will increase if users are provided with a consistently worded program output to model and then are shaped with nonthreatening error messages that reiterate those vocabulary and/or phrases that the processor can understand.

Keywords: Empirical studies; Intelligent/Expert systems; Keyboard input; Natural language; Natural language; Screen output; Voice/Speech input; Voice/Speech output; Wizard of Oz

Originally published: Proceedings of the Human Factors Society 28th Annual Meeting, 1984, pp. 768-772

Republished: G. Perlman, G. K. Green, & M. S. Wogalter (Eds.) Human Factors Perspectives on Human-Computer Interaction: Selections from Proceedings of Human Factors and Ergonomics Society Annual Meetings, 1983-1994, Santa Monica, CA: HFES, 1995, pp. 24-28.