“ Sort of ” in British Women ’ s and Men ’ s Speech

This paper (Note 1) examines the form sort of in British men and women’s speech, and investigates whether there is a gender difference in the use of this form. We do so through corpus analysis of the British National Corpus (BNC). We contend there is no quantitative difference in the use of sort of in men and women’s speech. Contrary to general belief, we claim women do not use hedges more than men. However, there is a functional difference. Men and women use the item sort of for slightly different functions: in particular, women use these forms as politeness devices more often than men. The results of this study do not present any evidence to support the widespread claim that women use the forms sort of more than men do. There is no significant gender difference in the frequency of sort of. In the light of these results, women’s allegedly greater usage of these forms could be regarded as persistent folklore. Furthermore, sort of has been treated mainly as a sign of powerlessness and uncertainty. Yet, this study indicates that it is a useful interpersonal resource, which both men and women seem to appreciate. When examining the affective function in greater detail, some marked gender differences emerge. The main finding is that positive affect is more common in women’s speech, whereas negative affect is clearly more frequent in men’s speech. In much of the research on gender differentiated conversational strategies, women’s language is presented as co-operative and men’s language as competitive. The results of this study show that, with respect to sort of, such a distinction cannot be drawn. Even if negative affect is more common in men’s than in women’s speech, by far the most instances of interpersonal function occur in essentially co-operative behaviour not only in women’s but also in men’s speech. This notion does not fit the description of men’s speech style being competitive and prone to conflict. Hence, the results of this study suggest that men use the marker sort of more co-operatively than many researchers allow for.


Introduction
There are widespread cultural stereotypes regarding differences between gendered speech: e.g., women engage in long, personal discussions; men only talk about impersonal matters, avoiding self-disclosure.Coates (2004: 86) challenges this "folkloristic knowledge".E.g many studies have proven men talk more than women in various situations (Tannen 1990;Coates 2004).Girls and women, and boys and men, have different interests in interaction, they are focused on different goals."...men pursue a style of interaction based on power, while women pursue a style based on solidarity and support" (Coates 2004: 126).Holmes (1995: 2) assumes that most women tend to use language to establish and develop relationships, whereas men are more inclined to see language as a tool for obtaining and conveying information.Cameron (1988) stresses that the hidden agenda in language differences between the genders is always female inferiority, in other words, continued male dominance.Tannen (1990) feels that the risks of ignoring the differences are greater than the danger of naming them, "there ARE gender differences in ways of speaking, and we need to identify and understand them.If we recognise and understand the differences between us, we can take them into account, adjust to and learn from each other's styles" (Tannen 1990: 17).
Generally speaking, politeness involves taking into consideration the feelings of others.Thus, a polite person makes others feel comfortable.A considerable number of studies concentrate on the issue of positive and negative politeness.Positive strategies might include recognition of the sensitivity of the addressee's wants and claiming common ground, for example by attending to the hearer's interests and needs, seeking agreement or using in-group markers.Negative strategies might include conventional indirectness, questions, apologies, and hedges.The notion of politeness is inevitably bound to the social context.Social distance is also crucial in determining the appropriate level of politeness.

Pragmatic Marker "Sort of"
Sort of has often been treated under the term hedge.Many have made the controversial claim that hedges are only related to women's language and to powerlessness.Hedges are linguistic forms that have the effect of damping down what we may say (Coates 1996: 152).The most common assumption that women's speech is tentative has been closely associated with the claim that women use more hedges than men (Coates 2004: 88).Earlier researchers assumed hedges equated with uncertainty about what one was saying.For instance, Lakoff (1975: 53) claimed that women use hedges in fear of not appearing too masculine.O' Bar and Atkins (1980) renamed women's language as "powerless language" and argued that anyone in a powerless situation would use such linguistic forms.They unquestionably accepted the claim that hedges are a feature of powerless language.
However, Coates (1996: 72) argues that hedges have been seriously misunderstood, precisely because of their association with women's ways of talking.Many researchers have failed to understand the multi-functionality of hedges.Although hedges may sometimes express doubt and uncertainty, their major functions are rather interpersonal in nature (Coates 1996;2004).They can be used to show sensitivity to other people's feelings.They are useful in protecting the face of the addressee, as they mitigate the force of what is being said.They encourage open discussion, as they prevent speakers from taking hard, inflexible positions.They are also used when avoiding "playing the expert", that is appearing superior to others, which is important when on a collaborative floor where social distance needs to be minimalised.
There has been criticism of earlier research conducted on hedges in men's and women's speech, for severe methodological problems.Many have just lumped all pragmatic markers (affective or referential) together and labeled them as hedges, loosely counting linguistic forms in either gender's speech, regardless of the particular function in a given context.This offers misleading results.It is important to identify the function of the pragmatic markers.Another error conducted in past research relates to the fact that the number of instances of usage was not compared to the total amount of speech collected, hence the results were statistically meaningless.
There is little consensus on terminology: the most common terms used are pragmatic particle, pragmatic marker, pragmatic expression, discourse particle, discourse marker, filler, turn holder, hedge, softener, hesitation marker, and connective.We have chosen Pragmatic Marker.We believe it is more appropriate than the term discourse, since the terms in question function on an interpersonal rather than on a textual level.In addition, we prefer marker to particle, as the two forms are not single-unit items, but phrases.Some researchers also argue that we should avoid assigning any specific function, such as hedging, to a specific linguistic form.Rather, we should prefer more neutral terms, such as pragmatic marker.
The pragmatic marker sort of is a particularly interesting case, due to its complex nature.Holmes (1988: 88-9) reports the variety of ways in which it had been described by different researchers.For instance, Bolinger (1972) has discussed the syntactic and semantic features of sort of, paying attention to its grammatical flexibility, a feature also noted by Crystal and Davy (1975).Quirk et al. (1972) concentrate on the position of sort of within the clause, and provide a detailed description of the syntactic restrictions on sort of in verb phrases.Aijmer (1984) and Coates (1987) focus on the function of sort of as a qualifier of individual words or phrases.Aijmer (2004) has also examined sort of, among other pragmatic markers, in advanced Swedish learners' spoken English.Finally, Holmes (1988;1990) has examined sort of in New Zealand women's and men's speech.
The pragmatic marker sort of is essentially a spoken form.E.g.Aijmer (1984: 121-2) states that in casual, relaxed conversation sort of helps the listener to reconstruct what the speaker has in mind, even if an exact linguistic expression is not available.She suggests that a sentence with sort of is usually not difficult to understand, even if it contains a word that is not precise.However, as Holmes (1984: 160) points out, the use of sort of requires shared background experience between the speaker and the hearer.This claim is supported by Aijmer's (2002) notion that sort of is likely to occur in informal conversation among friends or close colleagues 'who are able to make inferences or assumptions about each other's knowledge' (Aijmer 2002: 202).
The phrase sort of has been grammaticalised, thus it can be regarded as a pragmatic marker.Aijmer (2002: 180) explains that the hedging meaning of the phrase sort of is an extension of the more literal meaning, that is type of.Thus, the pragmatic marker sort of must be distinguished from the lexical item sort of which means type of.Both are identical in form, but different in function.In the following examples, sort of is not a pragmatic marker, thus such instances must be excluded from this study.
(1) What sort of people turn up to these displays?(Holmes 1988: 86) (2) What sort of coffee do you want?(Holmes 1988: 86) The syntactic position of sort of offers a clue towards the identification of the pragmatic marker usage.Holmes (1988: 87) explains that the type of variant always modifies a specific noun or noun phrase.The pragmatic marker sort of, on the other hand, has much more freedom of occurrence.Aijmer (1984: 120-1) states that it can occur before nouns (they're sort of pieces), verbs (well Sidney sort of lives upstairs), adjectives (but are the department themselves all very sort of happy and jolly), adverbs (I refused to answer these questions just sort ofsimply), numerals (I mean you had sort of six hundred trains per division I think) as well as phrases (we were in sort of absolute depths of despair financially).Furthermore, unlike the type of variant, the pragmatic marker sort of can occur even at the end of the sentence, as in: (3) But you know it's just one incident sort of.(Holmes 1988: 90) In addition, Holmes (1988: 94) argues that while the type of variant can only modify the following noun or noun phrase, the pragmatic marker sort of may refer to the whole utterance.She explains that this is frequently realised by its occurrence at the end of the utterance, as above, or in pre-verb position, as in: (4) The others sort of aren't considered sane.(Holmes 1988: 91) As a result of grammaticalisation, the pragmatic marker sort of has attained syntactic properties which differentiate it from other Noun + Preposition constructions.Aijmer (1984: 119) states that when sort of functions as a pragmatic marker, the noun sort can occur freely in relation to the indefinite article.For example, all of the following are syntactically possible: a sort of dinner, sort of a dinner, a dinner sort of.In addition, sort cannot itself be modified.Aijmer (1984: 119) explains that in a phrase such as a nice sort of dinner, the adjective nice modifies dinner rather than sort.She also points out that sort of does not necessarily agree with the preceding determiner, as in these sort of traditions.

Functional Properties of Sort of
Sort of serves a variety of functions in conversation.The hedging effect of sort of seems to be generally accepted.As Aijmer (1984: 118-19) states, sort of is always used in speech to make the reference vague and less well defined rather than clear and specific.According to Holmes (1988: 94), the range of functions served by sort of fall into the broad categories of epistemic modal and affective meaning.
Epistemic modal sort of is proposition oriented, that is to say, it focuses on the content of an utterance rather than on the addressee.According to Holmes (1988: 94), epistemic modal sort of functions as an imprecision sign.Holmes (1990: 198) provides the following example of epistemic modal sort of.
(5) He's got a sort of a skirty thing on I suppose you'd call it.(Holmes 1990: 198) Sort of can signal approximation.In such instances the speaker is not looking for a more precise expression, but the modified element is meant to be interpreted approximately.This function of sort of is frequent before numbers and measurements, as in: (6) I was sort of about six or seven.(Holmes 1988: 95) Sort of can also signal lexical imprecision, it indicates that the following word is not as precise as the speaker would want it to be.This conveys the speaker's lack of confidence in the following word or phrase.
(7) He's wearing sort of a floppy trouser thing.(Holmes 1988: 96) Sort of can function as a signal of syntactic self-repair, as in: (8) The majority of them just sort of they've always been that way.(Holmes 1988: 97) Sort of can be used to signal special style, such as the use of technical, rare, foreign, formal, vulgar or idiomatic words in informal context.Holmes (1988: 98) explains that the speaker wishes to signal that the word is in some way marked.She explains that as a special-style marker, sort of signals the speaker's 'awareness of the inappropriate nature of the term in the context' and reduces the 'effect of its precision' (Holmes 1988: 99).
(9) The other one was just a bit concussed you know sort of dazed and shocked.(Holmes 1988: 98) Holmes (1984: 160) argues that the primary function of sort of is to mark intimacy.She explains that affective sort of can be used to reduce social distance and express the speaker's wish for a relaxed relationship with the addressee.
(10) Do you think I could sort of come and watch a programme on your TV tonight.
It's only short?(Holmes 1990: 198) Affective sort of is a useful device for protecting the face of the speaker as well as the addressee.For example, recounting personal or embarrassing experiences can be very face threatening.Holmes (1988: 101) has discovered that such instances trigger the use of sort of.Coates (1987) explains that 'the speaker does not want to put herself in a position where she cannot retreat from the proposition if it turns out to be unacceptable to others' (Coates 1987: 121, quoted here from Holmes 1988: 101).This is a case of protecting the speaker's positive face.Furthermore, Holmes (1988: 102) suggests that sort of primarily serves an affective function, when it is used to attenuate the force of a critical comment or disagreement.In the following example the previous speaker has just been praising modern furniture.Hence, sort of appears to indicate concern for the addressee's positive face, rather than uncertainty about the proposition.
(11) I'm afraid I sort of like traditional things really.(Holmes 1988: 102) Besides critical comments and disagreements, strong opinions can be hedged due to concern for the addressee's positive face.In the following example, sort of seems primarily to express the speaker's wish to be reasonable rather than uncertainty about the opinion.
(12) She was sort of horrible really.(Holmes 1988: 100) Similarly, as Holmes (1988: 102-3) states, positive self-evaluation may be hedged for affective reasons, as in: (13) I used to be sort of top of the class.(Holmes 1988: 103) In addition, sort of can also serve an affective function by minimising any imposition on the addressee, that is, expressing negative politeness.For instance, requests fall into this category.( 14) Do you think I could sort of come and watch a programme on your TV tonight?(Holmes 1988: 103) Aijmer ( 2002) points out that collocations are the most important formal indication of what function sort of has in the particular context.She explains that some downtoning elements, such as pretty, slightly, rather or probably, regularly co-occur with the phrase sort of.She concludes that in such instances sort of is used 'with an illocutionary force modifying function' (Aijmer 2002: 189).By this, we assume, she means that the force of the utterance is reduced.
Sort of can also co-occur with strong expressions, such as absolutely, really and very, serving a polite down-toning function.In addition, Aijmer (2002: 203) explains that the frequently occurring collocation sort of you know can be considered to serve interpersonal, affective function, since it invites the hearer to treat the information as common ground.
Finally, sort of quite frequently co-occurs with the word thing, as in the expression that sort of thing.Holmes (1988: 87) argues that the phrase sort of thing is a fixed collocation, which is more restricted in meaning than the pragmatic marker sort of.Hence, she argues, it should be excluded from the analysis of the pragmatic marker sort of.However, the matter appears to be controversial, as Aijmer (1984) included the phrase sort of thing in her analysis.

Methodology
As this is a corpus driven study, and we are interested in investigating the speech of British speakers, we used the British National Corpus (BNC) as our source material.Refer to http://www.natcorp.ox.ac.uk/ for a full description of this corpus.We searched the spoken (demographic) component, which is approximately 4.2 million words in size.This is more than sufficient to draw relevant statistical assumptions.Any instances of non-pragmatic markers for sort of were excluded from the analysis.All remaining pragmatic markers (2 865 hits) were then examined for distribution of gender, age and social class, according to the divisions arrived at by the BNC.For reliable comparison between categories and across genders, all results were then normalized to no./10 000 words.All results were statistically tested.
All valid hits were then also classified according to the following function categorization: 1. Epistemic modal -the extent of a speaker's certainty e.g.She sounds like she's smoking sort of sixty a day.
2. Affect -positive affect (including positive and negative politeness) e.g.Is that the colour he chose?He looks sort of ghastly don't he? (Positive politeness -softened disagreement.)e.g.But if you find out about the insurance, see what that is and then perhaps sort of contact the gas people and see, I mean this was, how long ago?(Negative politeness -addresses face needs of hearer.)

Results
The following five figures below present the primary findings of this paper.Essentially, women do not use sort of significantly more than men.However, there are differences between the genders with respect to age and social class.There are also functional differences between the genders.Although neither gender is more epistemic than the other (that is there is no statistical significance), with affective function women do use more positive and more negative politeness then men.We can see from Fig. 1 above that although there is a numerical difference (5.46%) between the usage of sort of between males and females, this difference is in fact non-significant (x = 0.96, df = 1, p = 0.33) (Note 2).Hence, we can conclude, contrary to widespread belief, that women DO NOT use this pragmatic maker any more so than men do.
Figure 2. Overall distribution of sort of according to age and gender Fig. 2 shows that gender differences begin to emerge after the age of 15 for the use of sort of.Women use this pragmatic marker more so in the 15-24, 25-34 and over 59 age-groups, whereas the men use it more in the 35-44 and 45-59 age-groups.In fact, the difference is most marked for the 35-44 age-group, in which the men use this pragmatic marker more than twice as much as the women.As noted, the women use this marker noticeably more so then the men between the ages 15-34.Perhaps this could be attributed to the widespread assumption that sort of is a feature of women's speech.We could further speculate that in pursuit of socialization into appropriate gender roles the women regard the form as more desirable than men do.The opposite might be said of the men, they avoid the linguistic form that might be labeled as feminine.However, this "behaviour" is contradicted in the later speech patterns of both genders.We might possibly assume that at a younger age people are more sensitive to the stereotypical gender roles and expectations, but later in life other factors may become more relevant.The values presented here are all statistically very significant (x = 191.29,df=5, p = 0.01).To this point we have examined the results for gender, age and social class, but we have not yet investigated the functional use of sort of.Fig. 4 shows the results for such an examination.The usage has been divided between epistemic usage and affective usage.In both cases the women exhibit greater usage, but not to a statistically significant extent (x = 0.06, df = 2, p = 0.967).These results are interesting on several levels.Firstly, there appears to be no large differences between the amount of usage being epistemic or affective.There is only approximately a 2% difference between both categories for both genders.Secondly, although the women exhibit greater usage in both forms, in neither case are the results statistically significant.Once again, this tends to dispel the idea of women being more affective than men in conversation, at least for this particular pragmatic marker.
Figure 5. Distribution of sort of in the affect function according to positive and negative politeness and gender Fig. 5 shows a further division of the affect results found in Fig. 4. Here, the distinction is made between instances of positive and negative politeness.Again, the women are found to be both more positively, and negatively, polite, and again both genders seem to almost equally use both forms of affect.These results tend to agree with Holmes (1995: 6) and Brown (1980, cited in Mills 2003: 207), who suggest that women are more likely than men to express positive politeness.Yet, we must be careful not to generalize, as this is only one linguistic form, and men may use some other linguistic resources to convey positive politeness more than women do.In addition, as these figures are based on those found within Fig. 4, no statistical significance can be attributed to these observations.

Conclusions
Coates ( 2004) claims there is growing evidence that women use hedges more than men.But Holmes (1990) is of the opinion that research findings are contradictory, especially if they do not take into account the function in context.Her own distributional analysis of three linguistic forms generally regarded as hedges suggest women do not at least use any one pragmatic marker significantly more than men.Our findings concur with Holmes.We have shown that women do not use sort of any more then men, not to a statistically significant level.Women do seem to use sort of more than men as positive politeness devices.They use more positive affect than the men, but not significantly so.This concurs but does not corroborate with previous research, suggesting women are more co-operative and supportive in style, whereas the men are more competitive and challenging.Still, the marker sort of contributes more to positive than negative affect, even for the men, which tends to argue against the aforementioned folk-wisdom.We must also remember that this study only applies to British speech and to one pragmatic marker, so we should not draw absolute conclusions about gender speech styles.Still, the findings here do question previous conceptions.

Figure 1 .
Figure 1.Sum distribution of sort of according to gender

Figure 3 .
Figure 3. Distribution of sort of according to social class and gender

Figure 4 .
Figure 4. Distribution of sort of according to function, and gender