The perceptions of women's speech wasinvestigated. Thirty six undergraduate studentsread one of three passages characteristic of male,female and neutral language. then rated thespeakers on the following nine characteristics:impoliteness, confidence, status. intelligence,aggressiveness. unattractiveness, extroversion.power and gender. All students were uninformed ofthe true nature of the experiment for fear that theywould give the socially acceptable answers until theend of the experiment. The results were found to beinsignificant for a majority of the characteristics.except for the fact that women's speech was ratedsignificantly more polite and more feminine thanmale's speech. However. a trend was foundindicating that women are rated more negatively;they were rated less confident. less aggressive. andless powerful. Thus. it can be concluded that anegatively stereotyped attitude exists towardswomen today which may be the cause of genderinequality in our society. Since these inequalitiesare ingrained in language, they are extremelydifficult to combat. The only solution is to keeppeople aware and committed to change in hope thatthings will get better. Sex Stereotypes as a Functionof Genderlect. When we meet a person for the firsttime. our perceptions of that individual tend to beformed by what we see and hear. Since languagereflects one's social standing, self confidence,education etc .., it is constantly used in impressionformation. From the way a person speaks, we thusmake inferences about his or her character.

Although not consciously formed, many sexdifferences in language do exist. Numerous studies(LakofT.1990 and 1976; Haas, 1979; Rasmussen &Moely. 1986) have clearly demonstrated thedifferences between masculine and femininegenderlect. These differences could be what accountfor the different perceptions we hold of men andwomen. Lakoff. (1976) notes that "linguisticimbalances are worthy of study because they bringinto sharper focus real world imbalances andinequities". To study these inbalances LakofT(1976and 1990). noted eight linguistic categories thatwomen and men use to a different degree, which

show up in all levels of the grammar of English.She found differences in the choice and frequencyof lexical items in the and in the situations inwhich certain syntactic rules are performed.

1. Certain colour words. Women make farmore precise discriminations in namingcolours than men; words such as beige. ecru.mauve. aquamarine. and lavender arepresent in woman's active vocabulary. butabsent from the vocabulary of most men."Men find such deSCription trivial. andirrelevant to the real world" (LakofT.1976).2. "Empty"adjectives. There are a number ofadjectives that are mainly confined towomen's speech. LakofTstates that womenuse words such as 'adorable'. 'charming'.'sweet'. 'lovely'. and 'divine' more often thanmen do. She also states that there areneutral adjectives which are available toboth sexes. such as 'great'. 'terrific' and 'neat'.3. Expletives. Lakoff claims that thestronger swear words are used by men andweaker ones are reserved for women. Anexample of a strong expletive would be 'shit'as compared to 'oh dear' used by women.4. Hedges. Women's speech, in general.seems to contain more instances of 'well','y'know', and 'kinda'. Terms such as 'I guess'or 'I wonder' are also used. These terms areused to convey uncertainty or used ~hen onecannot vouch for the accuracy of theirstatement (LakofT.1976).5. The intensive 'so'. Although men can usethe intensive 'so', its use is morecharacteristic of women's speech. Men use'so' in sentences that do not refer tothemselves or that are unemotional innature. Women. on the other hand. use 'so'instead of 'very' in sentences such as 'I likehim so much' instead of 'I like him verymuch'. "This weasels on intensity andconveys that women do not want to showstrong emotions or assertions" (Lakoff.1976).6. Tag questions. "Atag is midway betweenan outright statement and a yes no

question... It is used when the speaker isstating a claim. but lacks full confidence inthe truth of the claim" (Lakoff. 1976). Anexample is 'It's a nice day. isn't it?' By addingthe question part at the end of a declarative,it gives the impression that the speaker isnot sure of themselves and "looks to theaddressee for confirmation and even havingno viewof their own" (Lakoff. 1976).7. Hyper correct speech. Women use moregrammatically correct speech than males.Studies have been done which show thatwomen will pronounce the full form of theending "ing" for example 'singing' vs. themale form 'singin' (Fisher. as cited inFasold. 1990). Another study found thatmen delete or reduce final consonantclusters that serve grammatical functionsmore often than women (Wolfram. as citedin Smith. 1985).8. Polite forms. Women's speech is morepolite than men's. They use please andthank you more often and make requestsrather than making demands using theimperative form. For example. men will usethe direct order 'Close the door' whereaswomen will use the compound request Won'tyou please close the door?" Lakoff (1976)states that by using he imperative, malesmake the assumption that they are superiorto the addressee.

Lakoff goes on to state that women's language(language descriptive ofwomen alone) "submerges awoman's personal identity. by denying her themeans of expressing. herself strongly, on the onehand. and encourages expressions that suggesttriviality in subject matter and uncertainty aboutit". The ultimate effect of these discrepancies, isthat they contribute to various stereotypes that wehave about the sexes and their roles in society.

Broverman. Vogel, Broverman, Clarkson,and Rosenkrantz (1972). have noted the existence ofsex-role stereotypes 1.e.. "consensual beliefs aboutthe differing characteristics of men and women". Itis known that groups invent stereotypes about othergroups. The dominant group is the one thatestablishes the stereotypes of the other groups anddecides which groups. on the basis of thesestereotypes are 'good' and 'bad' (Lakoff, 1976).Broverman et. al. (1972) indicate that men andmasculine characteristics are more highly valuedin our society than women and femininecharacteristics, "of the 41 items defined asstereotypic. 29 had the masculine pole chosen asmore desirable by a majority of each sample.Moreover. the men and women showed a highagreement about which poles were sociallydesirable". They found that these malecharacteristics form a group of related behaviours

which reflect a 'competency' cluster. Included inthis cluster are attributes such as competence.assertion. logical. confident. and leadershipability. A relative absence of these traitscharacterizes the stereotypic perception ofwomen.

Lao. Upchurch. Worwin. and Grossnicle.1975. also found that negatively stereotyped viewstoward females are shared by both males andfemales. This supports past research (Mckee &Sherriffs. 1956; Sherriffs & Jarret. 1953) whichfound that males are seen as the high prestigiousgroup. and this preference for males is seen on thepart of both men and women. In this study by Laoet. al. (1975) it was found that "unfavourableattitudes towards females do exist". They looked atintelligence and likeability and concluded thatsince there is no concrete evidence that males aremore intelligent than females (Maccoby. 1963;Wechsler 1958. as cited in Lao et. al.) the fact thatsubjects evaluated males as more intelligent isindicative of a biased attitude held towards females.The same reasoning was applied to the likeabilityvariable.

A study by Seigler and Seigler (1976) alsofound that syntactic forms associated with malesare rated more intelligent and those associated withfemales less so. They also comment on Goldberg's(1968) findings that students consistently ratedpublished articles credited to women lower than theidentical articles credited to men.

What do the differences between men andwomen's speech mean? Since the manner in whichan individual speaks conveys information aboutthe speaker. we would like to know how thepredominant forms in women's speech areperceived. Thus. the purpose in the presentexperiment is to provide information about whatthese forms imply about women today. In doing sowe attempt to determine the extent to whichnegative stereotypes still exist about women.Broverman et. al (1972). have noted that "thesestereotypes are widely held, persistent. and highlytraditional". We too believe that negativestereotypes. formed by language perception. are stillprevalent in our society.

Subjects read passages typical of male.female and neutral speech and rated· the speakerson a list of nine characteristics developed from thesex role literature and Lakoffs work. Thecharacteristics are impolite. confident. status.intelligent. aggressive. unattractive. extrovert.power, and gender. These adjectives were chosen totest the hypothesis that women. speakers ofwomen's language. should receive more negativeratings than men. speakers of men's language onthese characteristics.

University of Western Ontario participated in ourstudy. Their ages ranged from 18 - 22 with a meanage of 19.7 years. These students were randomlyassigned to one of three groups. Group one waspresented a passage typical to a woman's speech.group two received a paragraph typical of male'sspeech and the third group received a neutrallanguage paragraph. Thus, each group consisted oftwelve subjects. six females and six males with eachsubject receiving one passage and answering onequestionnaire.

Three passages were used: one wascharacteristic of women's speech. one male'sspeech. and one neutral speech (see Appendix A).The passages were very short containing from sevento nine lines each. They were controlled for topicand length and hence only differed on thevocabulary Lakoff indicated as differentially usedby males and females. For example. a sentence inthe female passage read 'I drove so qUickly to gethere on time'. Whereas the male correspondingsentence is 'Shit. I drove like hell to get here' and theneutral sentence read 'Boy. I drove faster than usualto make it here'.

The passages were created by inserting asmany of Lakoffs characteristics into each passage.Thus. the female passage contained color words'turquoise'. tag questions 'haven't I?'. the intensive'so' and hedges such as 'y'know'. it was also moregrammatically correct and more polite. To be surethe passages were typical to each gender. 15 subjectswere asked to read the passages and judge if theythought it was a male or female. All 15 judged thegender of the speaker correctly.

Questionnaires were prepared on which thesubjects were to rate the speakers in the previouspassages (See Appendix B). A 7-point LIkert scalewas used where the subject was to put an 'X on theline closest to the characteristic they believed thespeaker to hold.

The subjects were approached on campus (inthe library. Center spot or the Social science:bullding) and asked if they had a few minutes to

/ answer a questionnaire. Each subject waspresented randomly with one of the types ofI passages. They were asked read the presented

I passage and to then rate the speakers on theI following characteristics. The subjects signed) consent forms (See Appendix C) however. they were, not told the true nature of the experiment until theywere debriefed after the experiment (See AppendixD). This was done for fear that they would answerthe questionnaire with the socially acceptable

answers instead of with their true feelings, had theybeen told the study was about stereotypes. Thequestionnaire took approximately 5 minutes tocomplete.

We can see that on Status. Intelligence. andextroversion. there were' no differences between thegroups. However. with impolite. confidence.aggressiveness. unattractiveness, power. and genderwe notice a trend where the female speaker wasrated lower on all these items as compared to themale speaker. Note that low scores indicate less ofthe characteristic. hence a low score on impoliteindicates more polite. and a low score on confidencemeans low confidence.

Table 1: Mean and standard deviation scores as afunction of charachteristics and the gender of thepassage.

Passage GenderCharacteristic Male Female Neutral

Nonpolite 4.385 3.333 2.583(1.805) (1.731) (0.996)

Confidence 4.154 3.833 5.333(1.573) (1.697) (1.371)

Status 4.538 4.5 4.333(1.506) (1.446) (1.371)

Aggressiveness 4.923 '4.167 4.833(1.553) (1.801) (0.937)

Nonattractiveness 3.769 3.250

3.333(1.481) (1.055) (1.155)

Extroversion 5.538 5.50 5.50(1.45) (1.0) (l.0)

Power 3.923 3.75 5.167(1.553) (1.215) (L03)

Gender 5.462 3.5 4.583(1.33) (1.732) (1.782)

Intelligence. 4.0 4.0 4.167(1.291) (1.758) (0.835)

Note: Standard deviations appear in parentheses.The group size for male and female equals 12.

An analysis of variance was performed onthe data. A insignificant main effect for thelanguage of the passage was found. however asignificant main .effect for the characteristics wasfound (E (8. 272) = 8.10. II < .001) and a significantinteraction effect was found between language of

passage and the characteristics (E (8. 272) = 2.13. n<.01). Thus. the subjects' ratings depended on thepassage they received.

In order to determine which characteristicsdiffered significantly a post hoc analysis wasperformed using the Newman Keuls procedure andan alpha level of .05. The nature of the interactionwas examined by holding the characteristic of thepassage constant and comparing the passages at thefollowing characteristics: Impolite. confidence.aggression. power. and gender. For a graphicalrepresentation of the means.

Using a pooled error term interesting resultswere found. The only characteristic that showed nosignificant results was aggression. For gender. themale passage differed significantly from the femalepassage (g (3. 272) = 4.917. n < .05). and the femalepassage almost differed from the neutral passage.whereas the male and neutral passages did notdiffer significantly from each other. On thepoliteness variable the male and neutral passagediffered significantly (g (3.272) = 4.516. n < .05). themale and female passage almost did as well (g (2.272) = 2.637. n < .05). On the confidence and powervariables. both the male and female passagesdiffered significantly from the neutral passage.

The present experiment investigated theperceptions of women's speech characteristics. andattempted to determine the extent to which negativestereotypes exist about women today. Previousstudies have noted that "unfavourable attitudestowards females do exist" (Laoet al .• 1975), howeverthe present study fajIed to replicate these studieswith significant results. The only characteristicsthat were found to be statistically significant weretwo: First that women's speech was rated morefeminine and second. it was rated more polite thanmale's speech. However. a trend was noticed. it wasfound that women were rated more negatively: theywen~ rated as less confident, ~ess aggressive.unattractive and less powerful. However. I couldnot replicate Seigler and Seigler (1976) and Lao etal. (1975) who found that women were rated as lessintelligent than males.

There are many explanations as to whythese results could not be replicated. Firstly thesample size: due to time constraints only 36subjects were used compared to previous studiesthat used use as many 643 subjects (Laoet al.• 1975).As well. the male and female subjects' responseswere not separated. It was assumed that sinceprevious studies found both men and women to holdthese negative stereotypes toward women(Broverman et al.. 1972; Lao et -al. 1975) it was notnecessary to separate their responses. This shouldhave. however. been done to see if past research onthis aspect could have been replicated and thus that

indeed women do have a prejudice against their ownsex. As well. although we tried to control for thegood subject effect. by not telling the subjects aboutthe true nature of the experiment until the end. theycould stll1have been able to figure it out and answerwith the socially acceptable answer. By askingsubjects directly how they feel about lingUisticforms, this caused them to be self conscious aboutthem. Thus. the subjects wanted to say the rightthing, especially today when the awareness ofnegative gender stereotypes towards women isheightened.

Seigler and Seigler (1976). comment thatrecently reported "changes in attitudes" towardwomen (Baruch. as cited in Seigler & Seigler, 1976)may have been more a function of changes in thesocial desirability of expressing anti-feministprejudices than of real changes in attitudes. Thus.they used a matched groups design, where one groupJudged the intelligence of statements and the otherJudged the sex of the speaker, in effect reducingsocial pressure. If this design had been used thesubjects attitudes probably would have been morevalidly reflected. Further evidence of a responsebias lies in the fact that the majority of our scoreswere found to fall around the mean. The subjectsprobably did not want to give a true score, butpreferred to have no extreme opinion. Furtherresearch should be done with matched groupshaving one group evaluate the passage and the otherindicate the gender. "'.•.,... -'.. ..i: _.

Another methodological problem could liein the fact that written speech samples were used.The subjects received part of a conversation ofwhich they were only able to 'hear' one speaker.Other research has been done using written stories(Rasmussen &Moley 1986) and sentences (Seigler&Seigler. 1976), however these have been criticized inthat language differences between the sexes occur inspoken informal style. Lakoff (1976) mentionedthat speech materials need reference to real worldcontext "they are a complex and subtle combinationof judgments that would be virtually impossible to-reproduce--in a natu-ral-way L."1an-experimentalsituation". These materials then, since they are notpart of connected dialogue and are unrelated tocontext. could too be the cause of differentialfindings. I could not decide upon another method oftesting for female stereotypes. without usingwritten samples. for in my opinion auditorysamples (video or cassette recordings) would createa bias since the voices are heard as definitely maleor female. Further research needs to be done todevelop a better methodology to test for the presenceof attitudes. .

Another problem could be in the fact thatthese passages were not actual speech samples.They were fabricated to be typical of male, femaleand neutral speech for the study. Perhaps had Itranscribed actual conversations more influential

results could have been found.The trend that was found has some serious

implications. The fact that women are ratednegatively on confidence, aggressiveness andpower, can explain the differences in theinequalities in the work force. Since thesecharac~eristics are the ones that are valued forsome occupations, opportunities that require suchtraits may be seriously limited for females. As well.when we consider that these results were found withliberal, young educated university students theseresults are more disturbing for two reasons. First.these are. the people who are usually more openminded and more sympathetic toward the goal ofgender equality. If these people rate women as morenegati~ely than men, then similar strongerattitudes must be held by the majority of society.Second! lh,.ese are the people who will fill the keypositions in the future. The fact that they holdthese ~ed attitudes toward women certainly doesnot shpw. that women are well on their way ofach1~ gender equality, these attitudes show nohope of tieing erased. Consider the fact that thesestereotypes appear to be unconscious. Ask anyoneof the~e ~d1vtduals and they all claim to be forwomen's equality. Because they do not realize theyhold sqch attitudes these stereotypes are even moredifficult to combat.

However there is no quick, clear cutsolution. ~n though women are seen in a negativelight, theY may be seen in an even worse light whenthey take on traditionally male sex roleappropriate behavtours. They are in what someauthors call a double bind, the dammed if you do.dammed. it you don't phenomenon (Lakoff, 1976;Lao et al. 1975; Rasmussen & Moely 1986). Theseresearchers have found that women, acting in aninappropriate sex-role manner, such as beinghighly assertive, were evaluated as even lessintelligent and even less likable (Lao et al., 1975).Rasmussen and Moley (1986) even found thatwomen ~ere rated as uppity when they acted inmale sex roles.

As long as genci~r inequalities exist, womenwill be in a double bind communicatively: any waythey differ from men they will be rated negativelyand anyway they act to approximate men they willbe rated,as unfeminine and uppity. The ideal would.be for both sexes to move their ways of

I,cJmmunicating closer to some middle ground.~owever, this is only an ideal and would never

I ~ork. Partly because men will have little to bei ained by compromise and mostly because it isi tebatable whether a compromise even exists. Is

Jere really a compromise between aggressive andubmissive? Society's only hope is to make

individuals aware of the differences between men'sand women's language and alert people that theyneed to remedy them. 'We may not all be ready orable to undertake the deep and sincere changes ofattitude that are prerequisite to change inbehaviour; that will take a long time" (Lakoff.1990). However. if people are kept aware andcommitted. things are likely to get better.

References CitedBroverman, I., Vogel,S., Broverman, D., Clarkson.

F., & Rosenkrantz, P. 1972. Sex role stereotypes:A current appraisal. In The Joumal of SocialIssues, 28, 58-77.

Haas, A. 1979. Male and female spoken languagedifferences: Stereotypes and evidence.PsyclwlDgical.Bulletin, 86, 616-626.

Fasold. R. 1984. 11le sociolinguistics of society.Massachusetts: Basil Blackwell Inc.

Lakoff, R. 1976. Language and Woman's Place.NewYork: Octagon Books.

Lakoff, R. T. 1990. Talking power. United States:Basic Books.

Lao, R. C., Upchurch, W. H., Corwin, B. J., andGrossnickle, W. F. 1975. Biased attitudes towardfemales as indicated by rating of intelligenceand likeability. Psychological Reports, 37,1315-1320.

McKee, J.P., and Sherriffs, A. 1957. The differentialevaluation of males and females. Journal ojPersonality, 25, 356-363.

Rasmussen, J. L., and Moely, B. E. 1986. Impressionformation as a function of the sex roleappropriateness of linguistic behaviour. SexP~les, 14, 149-161. -_ ...- ._.

Sherrtffs, A. C., and Jarrett, R. F. 1953. Sexdifferences in the attitudes about sexdifferences. Journal ofPsyclwlDgy, 35, 161-168.

Siegler, D. M., and Siegler R. S. 1976. Stereotypes ofmales' and females' speech. PsychologicalReports. 39, 167-170.

Smith. P. M. 1985. Language, the sexes and. society.New York: Basil Blackwell Inc.