2000 Conference Proceedings

Go to previous article 
Go to next article 
Return to 2000 Table of Contents


Donald B. Egolf, Ph.D.
Department of Communication
1117 Cathedral of Learning
University of Pittsburgh
Pittsburgh, PA 15260
Voice Mail: 1-412-624-6763
FAX: 1-412-624-1878
Email: ratchet@vms.cis.pitt.edu

One factor characterizing many nonspeakers is the passive role they play in discourse. Because of their inability to speak, nonspeakers and their speaking partners manage their discourse by frequently resorting to a "20-questions" kind of interaction. The natural speaker becomes the interlocuter and the nonspeaker serves as the "yes-no" respondent. This mode of interaction has some functionality, particularly in determining the needs, wants, and preferences of the nonspeaker. At the same time, however, this mode of interacting is slow and it maintains the nonspeaker in a passive role.

The purpose of this study was to determine if at the time of the introduction of programmable, electronic, communication aids, the emergence of discourse initiation behaviors could be promoted and facilitated in nonspeakers.


Subjects in this study were seven nonspeaking, nonreading, school-age children. All subjects had experienced some form of "low-tech" communication augmentation but none had been exposed to a programmable, electronic, communication aid with synthesized speech output. All were cerebral palsied individuals whose symptoms were moderate to severe. The average age of the subjects was 12.8 years with a range from 9.9 to 18.8 years.

Pre-Study Observations
Subjects were observed interacting with their speech-language clinicians on two occasions, with the speech-language clinician from this project, with a teacher, and with a school-age peer. Each interactive period ranged from 15 to 20 minutes in length and each was videotaped for later review by the project staff. In the subsequent review the general nature of the subjects' interaction skills was evaluated. It was found that they had some knowledge of discourse skills, responding to questions, turn-taking, giving feedback, and initiation, for example. However, these discourse forms were manifested globally and primarily through nonverbal modalities. Moreover, pervading the interactions were "yes-no" responses from the nonspeakers. The nonspeakers were passive givers of information. Could these patterns be changed? This study sought the answer to this question. Equipment Equipment used in this study included videotape equipment and programmable, electronic, augmentative communication aids with synthesized voice output: Touch Talkers and Light Talkers. Each of the communication devices had a 128 space keyboard which could be rezoned in descending powers of 2 (64, 32, 16, 8). Recording Interaction Patterns During Discourse A recording system for analyzing the interactions in this study was developed. It was based on Bales (1950) Interaction Process Analysis and it included twelve categories into which subject utterances could be categorized.


Baseline-Type Session
Each subject was seen in a baseline-type interaction session prior to the introduction of the communication aids. In this session the project's speech-language clinician interacted with each subject with the only instruction being, "Have a chat." These instructions were analyzed using the Bales-based system, and the results of the analyses were later used for comparison data. Training Sessions Next, the communication devices were introduced to the subjects. Decisions about the assignment of a touch-activated device or a light-activated device were made in conjunction with the subjects' school-employed speech-language clinicians. While each decision was multi-faceted, of primary consideration was finding that part of the body that gave the most reliable voluntary motoric response.

Following the assignment of the devices, training sessions for subjects were held in the use of the devices. Further decisions were made about the zoning of keyboards in order to accommodate the respective motor disabilities of the subjects. Experimental Interaction Sessions Preparations for the experimental interaction sessions involved the coding of the communication devices. Since the subjects were nonreaders and nonspellers, programming required the use of icons on the keys instead of letters. We used some icons provided by the manufacturer of the Touch Talker and Light Talker but the majority of icons were produced by us. The communication devices were programmed so that the activation of a tactile or light sensitive "key" generated a synthesized speech message. Since we were interested in having our subjects become initiators in discourse and not passive recipients, we programmed into the device utterances that were requests: asking for opinions, information, or action. Complementing the requests were other utterances related to greetings, farewells, and likes and dislikes. Of prime importance in programming the devices was to devise tasks that promoted meaningful discourse and not just random and repetitive key activation.

To determine if the subject were responding in an appropriate and meaningful manner, communication tasks were designed so that the subject was requesting actions from a limited set. In Task 1, for example, subjects requested that the clinician give them a red, yellow, blue, or white piece of paper. There were five pieces of each color. Therefore, subjects could not continue asking for pieces of blue paper because after five such requests, the sixth request would be inappropriate since the supply of blue pieces of paper would have been exhausted.

In like manner in Task 2, subjects requested through their speaking devices that the clinician put specific coins from a limited coin set into a bank (an "Uncle Sam" bank where Uncle Sam would "drop" the coin from his hand into a pouch). If a subject requested that a dime be dropped into the bank and no dimes were available, the response was inappropriate. In Task 3, the subjects requested that specific Christmas-tree decorations be placed on a tree. Again, there was a limited set of decorations. To have the devices request colors, coins, and ornaments, "literal" icons were used -- color patches, coins, and small ornaments.

The requests for paper, coins, and ornaments were imbedded in the context of social discourse. For example, subjects had programmed into their devices greeting utterances, e.g., "Hello, how are you today?": Icon = smile face; farewell utterances, e.g., "Goodbye, have a nice day": Icon = waving hand; like and dislike utterances, e.g., "Do you like chocolate cake?": Icon = picture of a chocolate cake; utterances to correct the clinician, e.g., "You gave me the wrong one": Icon = circle with slash; courtesy utterances, e.g., "Thank you": Icon = praying hands; feeling utterances, e.g., "I don't feel well": Icon = medical thermometer; and pacing utterances, e.g., "Wait a minute": Icon = stop sign. Post-Intervention Sessions A post-intervention session was held for each subject after the three tasks were executed. This was similar to the baseline-type session except that subjects retained their augmentative communication devices. Paper, money, and tree ornaments were removed from the subjects' devices and icons related only to social discourse remained. Thus subjects could use their devices to talk about likes and dislikes of theirs and of the clinician, they could produce greeting and farewell utterances, express feelings, and could control the output of their speaking partners by asking them to repeat, slow down or tell more.


Data for the baseline, tasks, and post-intervention sessions were recorded on the Bales-based analysis system discussed above. Extracted from the Bales analysis forms were the frequencies with which subjects' utterances fell across the twelve categories. Visual inspection of the data showed that over 90% of the utterances fell into six categories, "Gives opinion, information, or action;" and "asks for opinion, information or action." These six categories were focused upon and the remaining were ignored since their incidence was marginal (<10%). Further, the six utilized categories were collapsed into two: (1) "Giving opinions, information, and action;" and (2) "Asking for opinions, information and action." This collapsing created two categories of interest in regard to nonspeakers: the first category reflects responding while the second reflects initiation, an area in which nonspeakers in general are, and our subjects in particular were, deficient.

The average frequencies of occurrence of the "asking" and "giving" responses during the task sessions were compared with the frequencies during the baseline-type session, and the baseline-type session frequencies were compared with the frequencies from the post-intervention session through the use of the Fisher Exact Probability Test (Siegel, 1956). Using the Fisher test (appropriate for an N as low as seven), it was found that during the communication tasks, a significant (p<.05) number (7 of 7) of subjects changed from being predominantly passive givers of opinions, information, and action to being predominantly "askers" of opinions, information, and action. An identical significant (p<.05) change occurred when the baseline-type session was compared with the data from the post-intervention session.


This study has shown that providing school-age nonspeakers with a properly programmed communication device with a voice synthesizer can lead to a significant change in their discourse styles. They changed from passive receivers to active initiators. We acknowledge that a criticism of our research is that we programmed the devices to get what we wanted from our subjects: namely, a significant increase in the frequency of their active initiation responses. If you give someone a hammer he will go about pounding the world so the aphorism goes and if you give a nonspeaker a device that can speak utterances that ask for opinions, information, or action, that is, indeed, what you will get. Our answer to this criticism is two-fold. First, we designed our tasks to guard against inappropriate responses, random and repetitious activations, for example. Inappropriate responses were not counted. Second, the purpose of any prosthetic device, communication devices included, is to correct some dysfunction. Eye glasses are designed to correct dysfunctional vision and communication devices are designed to correct dysfunctional communication. And so we did program the devices to enable the subjects to produce what we and they wanted to produce just as the optometrist prescribes lenses to obtain the desired result.

The results produced in this study constitute only a beginning in satisfying our subjects' communication needs. But a major initial step was taken. Our positive quantitative results were complemented by the subjects' expressions of joy and satisfaction when they produced spoken initiation utterances which gave them control over their conversational partners.


Bales, R. F.(1950). "Interaction Process Analysis: A Method for the Study of Small Groups." Reading, MA: Addison-Wesley.

Siegel, S. (1956). "Nonparametric Statistics for the Behavioral Sciences." New York: McGraw-Hill Book Co.

Go to previous article 
Go to next article 
Return to 2000 Table of Contents 
Return to Table of Proceedings

Reprinted with author(s) permission. Author(s) retain copyright.