Return to index.

The Effects of 'Powerful' and 'Powerless' Speech Styles on Speech Effectiveness in a Non-Courtroom Setting

Keith D. Green

Volume 1, 1987

Communication research is helpful in investigating the relationship between subtle language phenomena and the perceived effectiveness of the communication event. In this complex relationship, the effect of subtle nonverbal and verbal factors is of special and, according to Myers and Myers, relatively new interest (1985, p. 224). To further study these areas, the emphasis of this inquiry is on the effect of 'powerful' and 'powerless' speech styles on speech effectiveness. Specifically, this study seeks to determine if there is a relationship between the grades given student speakers by teaching assistants and the number of the power indicators present in their speeches.

This concept of powerful or powerless speech has been derived from studies in socio-linguistics. According to Erickson, Eind, Johnson, and O'Barr in "Speech Style and Impression Formation in a Court Setting: The Effects of 'Powerful' and 'Powerless' Speech," these socio-linguistic studies found that certain speech style variations were associated with social status and authority. In a given situation, individuals with higher social status tended to use a more powerful speech style; whereas, those with lower status tended to use a more powerless speech style (1978, p. 267). Power in speech is indicated by a multitude of verbal and nonverbal variables. Erickson and her associates identified seven primary language characteristics:

Individuals with low social power and low status vis-à-vis the court tended to make frequent use of itensifiers ("so," "very," "surely," as in "I surely did."), hedges ("kinda," "I think," "I guess," etc.), especially formal grammar (the use of bookish grammatical forms), hesitation forms ("uh," "well," "you know," etc.), gestures (e.g. the use of hands and expressions such as "over there" while speaking),questioning forms (e.g., the use of rising, question intonation in declarative contexts), and polite forms ("please," "thank you," etc.) These features tended to occur in the speech of low-power witnesses, and their frequent use constitutes what we term here the "powerless, style of speaking (1978, p. 267).

As the number of these power indicators rises, the speech style becomes more "powerless" and vice versa.

The ramifications of the attribution of powerfulness or powerlessness are multifold. If a speaker uses a powerless speech style, this may indicate less confidence. In turn, the credibility and attractiveness of that speaker may fall. The speaker may be perceived as less knowledgeable and less assured.

Furthermore, to listen effectively to a speaker using a large number of vocal segregates and other powerless features is especially taxing (Erickson, et al. 1978, p. 268). All of these weaken the speaker's effectiveness, and since they do so in a variety of ways, powerless speech styles may be especially damaging to the speaker. Since the ramifications of power style are significant, it is important to consider these language variables in greater depth.

The use of these subtle language variables constitute what James J. Bradac and Charles R. Berger label the expressive function of speech." They state, "By the expressive function of speech, we mean to suggest that performed language conveys information to the hearers, quite apart from the intentions of speakers" (1982, p. 53). The language variables, the receiver's perception will support or deny the intended message. In other words, these powerless variables may transmit information contradictory to the message intended by the source.

In the basic speech course, such powerless styles are usually encountered. As speech teachers, we are well aware that beginning speech students use a large number of these variables. In a fundamentals class, however, attention is often given to content over delivery. As a result, delivery concerns are either neglected or minimized in order to avoid placing more pressure in an already anxious situation. Accordingly, speech effectiveness should be graded by content, not by delivery. However, can instructors avoid being influenced by these language factors? Thus, the null hypothesis may be stated: The number of power indicators present in a speech does not affect its perceived effectiveness.

By testing this hypothesis, a better understanding of the effects of power styles may be gained. Moreover, by investigating these relationships in the classroom, a better understanding of the pedagogical impact of these variables may also be reached.

The primary monograph studying speech styles and power was the previously mentioned 1978 Erickson et al. study. Several language variables were isolated then manipulated to determine altered perceptions. The investigators used intensifiers, hedges, especially formal grammar, hesitation forms, gestures, questioning forms, and polite forms reproduced in varying levels in actual courtroom testimony.

They concluded:

The powerful-powerless speech style manipulation affected not only subject's perception of the speaker's credibility and attractiveness but also their acceptance of the information contained in the speakers testimony (p. 276).

Thus, in a courtroom setting, a more powerless speech style reduced the credibility and attractiveness of the speaker thereby reducing the effectiveness of the speaker's message.

John W. Wright and Lawrence A. Hosman, in their 1983 study, sought to test Erickson's findings. They accepted, with some reservations, that these language variables do reduce the force of the statement, hence the power is lowered; however, they raised questions regarding the interaction of the various power indicators in singular and combined appearance. They asserted:

Contrary to Erickson's contentions, it is conceivable that intensifiers are perceived as increasing the expressor's certainty...a certainty which Erickson, et al., associates with low power"p. 143).

While Erickson et al. found that intensifiers lower power, Wright and Hosman hypothesize that they add certainty. By testing these variables individually, they sought to define and specify the interaction of power factors with perceived credibility.

The results of their study supported Erickson's findings. Their results suggested that someone who expressed certainty in their assertion by not qualifying them was liked more than someone who expressed uncertainty. The present study confirmed the Erickson, et al. implications....(p. 149).

They concurred with Erickson and associates in that, in a courtroom setting, the variations in the powerful-powerless styles did affect the receivers' perceptions of credibility. The higher the speaker's uncertainty of his own message, the lower the perceived power, hence credibility, of the speaker.

These findings are supported by James J. Bradac, Michael R. Hemphill and Charles H. Tardy (1981), and Allan E. Lind and William M. O'Barr (1979) in their respective works. All have concluded:

Hedges, intensifiers, polite forms, hesitation forms, and deitic phrases produced judgments of low power, and the absence of these forms produced judgments of high power (Bradac, Hemphill and Tardy 1981, p. 334).

Clearly, these studies demonstrate that the level of powerless speech variables does affect effectiveness; however, the above findings are all based in courtroom settings, not on other communication events. If these language variables do affect speaker effectiveness, an investigation of these variables in a non-courtroom setting is certainly warranted. By discovering how they may function in other settings, specifically in the basic speech class, we will better understand the process of communication, as well as important factors affecting the relationship between student as speaker and teacher as evaluator.

An important point in these studies was the use of a situation in which credibility was of primary concern, such as a courtroom. Dale Leathers asserted that in a persuasive situation the perceptions of the speaker's credibility are assigned added importance to the effectiveness of the communication event (1976, p. 136). For this reason, persuasive speeches given in the Fundamental of Public Speaking course were observed. Disregarding any intrapersonal variation of intent, by definition the student speakers must have been concerned with personal credibility in order to increase their effectiveness. The students had been taught the Aristotelian modes of persuasion and were to apply them in the persuasive speech. By observing these persuasive speeches and correlating the data with the grade for the speech, the relationship of power style to speech effectiveness was investigated.

For the present study, five power variables were chosen: hesitation forms, intensifiers, hedges, gesture forms, and questioning intonation. These hesitation forms include such phonemes as "uh" or "ah" and such morphemes as "okay" or "you know" when they are used as a nonfluency, imparting meaning superfluous to the intended message. Of the three major studies using language variables to analyze perceived power, Erickson, et al.; Bradac, et al., and Wright and Hosman, only the latter did not make hesitation forms a primary nonfluency. The preceding two used hesitation forms as one of the primary forms analyzed. These hesitation forms may indicate a lack of message confidence in the speaker. Accordingly, the greater the number of hesitations, the lower the perceived speaker credibility. Wright and Hosman focused primarily on intensifiers and hedges. Intensifiers are those words, such as "very," "certainly," and "definitely" which seem to increase the force of the statement. Although they appear to add certainty, they most accurately indicate a lack of certainty in the message (Erickson, et al., 1978, p. 205; Wright and Hosman 1983, p. 149). If the source must use an intensifier to enhance the message, the speakers own lack of confidence as to the accuracy of the statement may be revealed. In this situation, the meaning is derived from the vocal emphasis placed on the intensifier.

Similar to the intensifier, although superficially appearing to be opposite, is the hedge. While the intensifier infers uncertainty by virtue of its over-reinforcement, the hedge more directly displays that uncertainty. Such words as "sort of," "kind of," "maybe," and "I think" add a sense of uncertainty over the assertion. By qualifying the message in such a form, the speaker transmits his/her doubts over the validity of his/her intended meaning (Erickson, et al. 1978, p. 271). While the hedge forms may be used to acknowledge the presence of admitted exceptions, they become hedges when they are used to indicate the speaker's reticence toward the message. The hedges, like the intensifiers, suggest uncertainty over the message being transmitted.

The final two variables are more subtle and generally appear less often than the previous three. Gesture forms are those phrases, such as "over there" or "like this," usually accompanied by a gesture which suggests the speaker's desire to resort to physical placement in order to overcome the lack of lexical diversity to adequately explain an idea. If the speaker is hesitant in a verbal description of an event, idea, concept, or thing, he/she will often attempt to use hand gestures or physical placement to replace or support the verbal message. message.

The final power variable is the use of rising inflection, Usually indicative of questioning forms. By ending with rising inflection, the information is put into an interrogative form which casts doubt upon the message. While the structure of the message may be in statement form, the delivery will place it into a questioning form indicating the speaker's uncertainty over asserting himself/herself which then lowers speaker credibility.

The choice of subjects was not random in the pure sense. In selecting speakers, five sections held at times accessible to the investigator were selected and whichever students had been assigned to speak that day by the classroom instructor were observed. Although each classroom instructor was aware of the observations, none adjusted their speaker order as a result. Each individual teaching assistant had determined the speaking order prior to being aware of the study As a result, randomness cannot be claimed, but specific selection was also avoided. furthermore, none of the teaching assistants were informed as to the purpose or method of the study prior to the investigation. As a result, the natural relationship between the teacher's grade and the student's performance was maintained.

While observing each speech, the student's use of the five variables was counted by the investigator. The speaking rate was ranked slow, moderate, or fast. This rating determined the words-per-minute figure used to calculate the language variable ratios. The word-per-minute figures used were 120, 147.5, and 175 respectively, based on the figures proposed by Ehninger, Gronbeck and Monroe (1981), and Reid (1972).

After determining the number of appearances of each variable, the language variable ratio was computed by dividing the total number of variables by the approximate words in the speech, as found by using the wpm figure assigned to the speaker. By doing this, differences in the lengths of the speeches and speaker rates were made negligible.

After these ratios had been determined, each teaching assistant provided the grade for the students observed. The letter grades were given a numerical value, A+ = 13, A = 12, . . . F = 1. Using the Pearson's R, these grades were then compared with the nonfluency ratio in order to determine significance.

The only additional limitation placed upon the use of data in the study was that only native English speaking students were used. By doing this, nonfluencies and other variables based on lexical ignorance or uncertainty over the language were not used. These are based on separate factors and, as a result, are not applicable to this study.

Since the standard significance factor needed to reject a null hypothesis is between 0.01 and O.OS (Spiegel, p. 212), a figure fitting these parameters is required here.

After the data had been subjected to a Pearson's Correlation test, an r of -.261 was found. From this factor, two deductions may be made. First, a negative relationship, as indicated by the r being a negative number, reveals that as one variable increased the other variable decreased. This demonstrates that there is an inverse relationship between the number of power variables in a speech and the effectiveness of that speech as indicated by the grade. Moreover, the r is equal to a significance of approximately O.OS. This indicates that the inverse relationship is probably a sound one. As the number of variables rise, the rated effectiveness is lowered. The r of -.261 indicates that there probably is a causal relationship between the language variables and the grades.

For these two reasons, the null hypothesis,"The number of power indicators present in a speech does not affect its perceived effectiveness," is rejected in favor of the contention, "As the number of power indicators in a speech rise, the perceived effectiveness of that speaking event declines." The findings of Erickson, et al., Bradac et al., Lind and O'Barr, and Wright and Hosman have been aaffirmed as applicable in a non-courtroom setting.

An aspect of the results must be noted, however. From the findings of the study as listed in Table 1, one can see that the language variable ratios for a given grade varied, some quite widely. For instance, the spread in the ratios for a B varied by .038. The ratios of .056 and .018 both appear in the group. This would indicate that other factors as well as the tested language variables affects speaker credibility.

Table 1
Results of Language Variable Observation by Grade

X=variable ratio: total variables/total words in speech.
Y=mean of variable ratios by grade.


























In considering the implications of these results, several points arise. Clearly, this study has shown that these language variables do affect a listener's perception of the speaker and, logically following, the message itself. When these variables appear in number, the overall effectiveness of the speaking event declines. While this study did not seek to determine why these variables caused the effect they did, such a relationship may be hypothesized.

The variables may serve to indicate a gap in the message. The source is searching for words or otherwise considering the message prior to transmitting it. The receiver, it appears, picks up these nonverbal cues and makes an inference about speaker confidence. If the speaker uses a more powerful speech style, the speaker is perceived as more credible and, accordingly, the speaking event is more effective. On the other hand, if the confidence is low, the power is lessened, and the speaking event is less effective.

Since the building of speaker credibility is most often based on honesty, sincerity, and enthusiasm, these language variables play a large role. The level of the speaker's credibility based on these intangible qualities may be most clearly indicated by these variables. If a speaker hesitates, pauses, overly enforces his/her message or otherwise indicates an uncertainty over the message, his/her credibility will drop.

In the classroom, we as teachers of speech must be aware of the impact of these variables. First, we must instruct our students in the ramifications of powerful and powerless speech styles. We have all intuitively known that the manner in which one speaks significantly affects the listener's perception and image of the speaker. This study adds further emphasis on this need to be cognizant of the impact of our language choices and delivery techniques.

Second, as teachers we must avoid believing we are immune to the effects of these speech styles. Although we may have a better academic understanding of their existence and effects than does the layman, this knowledge does not free us from their subtle influences. Upon being informed of the nature of this study, after the observations were complete, each of the teaching assistants commented that they did not grade based on variables such as these. Rather, they asserted that they graded primarily on content. From this it is clear that we must be aware of the impact of these variables not only n on outside audiences, but on ourselves as well. This study clearly suggests that the grades we give may be strongly influenced by these factors, even if we do not consciously use them as grading criteria.

There are several questions to be raised from this study. Since the language variable ratios had such a wide deviation in some grades, what other factors affect effectiveness? It would be of interest to do two variations on this study to help determine other factors.

First, the individual variables could be isolated more precisely than has been done in the past. This could give a more specific comparison of paravocal phenomena and speaker effectiveness. Also, it would be of interest to examine how content and language variables are related. Can a good speech override a powerless speech style? To study this, a well structured speech could be delivered with different levels and types of power variables, much like Erickson, et al. did in their study.

By continuing to establish relationships between variables in the communication event and the success of that event, awareness of how effective communication is attained as well as maintained can move ahead. We as communication teachers will be better prepared to instruct our students in specific skills, to enhance their communication competence, and to use their communication skills effectively in business or pleasure.


Bradac, James J. and Charles R. Berger. Language and Social Knowledge. London: Edward Arnold Publishers, Ltd., 1982.

Bradac, James J., Michael R. Hemphill and Charles H. Tardy. "Language Style on Trial: Effects of 'Powerful' and 'Powerless' Speech Upon Judgments of Victims and Villains." Western Journal of Speech Communication 45 (1981): 327-41.

Camden, Carl and Carole Kennedy. "Interruptions as an Index of Communication Dominance. " ERIC, 1981. ED 198 598.

Ehninger, Douglas, Bruce E. Gronbeck and Alan H. Monroe. Principles of Speech Communication. Glenview, Illinois: Scott, Foresman and Company, 1980.

Erickson, Bonnie et al. "Speech Style and Impression Formation in a Court Setting: The Effects of 'Powerful and 'Powerless' Speech." Journal of Experimental Social Psychology 14 (1978): 266-79.

Kramer, Cheris Gamble. Women's and Mens' Perception of Female and Male Speech. ERIC, 1975. ED 120 886.

Leathers, Dale G. Nonverbal Communication Systems. Boston: Allyn and Bacon, Inc., 1976.

Lind, Allan E. and William M. O'Barr. "The Social Significance of Speech in the Courtroom. "Language and Social Psychology. Ed. Howard Giles and Robert W. St. Clair. Baltimore: University Park, 1979l

Miller, Gerald R. and Murray A. Hewgill. "The Effects of Variations in Nonfluency on Audience Ratings of Source Credibility." Quarterly Journal of Speech 50 (1964), 36-44.

Myers, Gail E. and Michele Tolela Myers. The Dynamics of Human Communication: A Laboratory Approach. New York: McGraw-Hill Book Company, 1985.

Putnam, Linda L. and Linda Skerchock. "Verbal and Nonverbal Determinants of Dominance in Sex-Type Mail, Sex-Type Female, and Androgynous Dyads." ERIC, 1978. ED 159 745.

Spiegel, Murray R. Probability and Statistics. New York: McGraw-Hill Book Company, 1972.

Wright, Jon W. II and Lawrence A Hosman. "Language Style and Sex Bias in the Courtroom: The Effects of Male and Female Us of Hedges and Intensifiers on Impression Formation." The Southern Speech Communication Journal 48 (Winter 1983): 137-52.