GENDER DIFFERENCES IN COMPUTER-MEDIATED COMMUNICATION:
BRINGING FAMILIAR BAGGAGE TO THE NEW FRONTIER

Susan Herring

Program in Linguistics
University of Texas
Arlington, TX 76019
susan@utafll.uta.edu

(Keynote talk at panel entitled "Making the Net*Work*: Is
there a Z39.50 in gender communication?", American Library
Association annual convention, Miami, June 27, 1994.)

1. Introduction
Although research on computer-mediated communication
(CMC) dates back to the early days of the technology in the
1970's, researchers have only recently begun to take the
gender of users into account.[1] This is perhaps not
surprising considering that men have traditionally dominated
the technology and have comprised the majority of users of
computer networks since their inception, but the result is
that most of what has been written on CMC incorporates a very
one-sided perspective. However, recent research has been
uncovering some eye-opening differences in the ways men and
women interact "online", and it is these differences that I
will address in my talk today.
My basic claim has two parts: first, that women and
men have recognizably different styles in posting to the
Internet, contrary to the claim that CMC neutralizes
distinctions of gender; and second, that women and men have
different communicative ethics -- that is, they value
different kinds of online interactions as appropriate and
desirable. I illustrate these differences -- and some of the
problems that arise because of them -- with specific reference
to the phenomenon of "flaming".

2. Background
Since 1991 I've been lurking (or what I prefer to
call "carrying out ethnographic observation") on various
computer-mediated discussion lists, downloading electronic
conversations and analyzing the communicative behaviors of
participants. I became interested in gender shortly after
subscribing to my first discussion list, LINGUIST-L. Within
the first month after I began receiving messages, a conflict
arose (what I would later learn to call a "flame war") in
which the two major theoretical camps within the field became
polarized around an issue of central interest. My curiosity
was piqued by the fact that very few women were contributing
to this important professional event; they seemed to be
sitting on the sidelines while men were airing their opinions
and getting all the attention. In an attempt to understand
the women's silence, I made up an anonymous survey which I
sent to LINGUIST-L asking subscribers what they thought of
the discussion and if they hadn't contributed, why not.

3. Initial observations
The number one reason given by both men and women for
not contributing to the LINGUIST discussion was "intimidation"
-- as one respondent commented, participants were trying to
"rip each other's lungs out". Interestingly, however, men and
women reacted differently to feeling intimidated. Men seemed
to accept such behavior as a normal feature of academic life,
making comments to the effect that "Actually, the barbs and
arrows were entertaining, because of course they weren't aimed
at me." In contrast, many women responded with profound
aversion.

That is precisely the kind of human interaction I
committedly avoid. (...) I am dismayed that human
beings treat each other this way. It makes the world
a dangerous place to be. I dislike such people and I
want to give them WIDE berth.

When I analyzed the messages in the thread itself,
another gender difference emerged, this time relating to the
linguistic structure and rhetoric of the messages. A
daunting 68% of the messages posted by men made use of an
adversarial style in which the poster distanced himself from,
criticized, and/or ridiculed other participants, often while
promoting his own importance. The few women who participated
in the discussion, in contrast, displayed features of
attenuation -- hedging, apologizing, asking questions rather
than making assertions -- and a personal orientation,
revealing thoughts and feelings and interacting with and
supporting others.
It wasn't long before I was noticing a similar pattern
in other discussions and on other lists. Wherever I went on
mixed-sex lists, men seemed to be doing most of the talking
and attracting most of the attention to themselves, although
not all lists were as adversarial as LINGUIST. I started to
hear stories about and witness men taking over and dominating
discussions even of women-centered topics on women-centered
lists.[2] In contrast, on the few occasions when I observed
women attempting to gain an equal hearing on male-dominated
lists, they were ignored, trivialized, or criticized by men
for their tone or the inappropriateness of their topic.[3]
It wasn't until I started looking at lists devoted to women's
issues, and to traditionally "feminized" disciplines such as
women's studies, teaching English as a second language, and
librarianship, that I found women holding forth in an amount
consistent with their numerical presence on the list. I also
found different interactional norms: little or no flaming,
and cooperative, polite exchanges.

4. Different styles
As a result of these findings, I propose that women
and men have different characteristic online styles. By
characteristic styles, I do not mean that all or even the
majority of users of each sex exhibit the behaviors of each
style, but rather that the styles are recognizably -- even
steoretypically -- gendered. The male style is characterized
by adversariality: put-downs, strong, often contentions
assertions, lengthy and/or frequent postings, self-promotion,
and sarcasm. Below are two examples, one from an academic
list (LINGUIST) and the other from a non-academic list
(POLITICS).

1) [Jean Linguiste's] proposals towards a more transparent
morphology in French are exactly what he calls them:
a farce. Nobody could ever take them seriously -- unless
we want to look as well at pairs such as *pe`re - me`re*,
*coq - poule* and defigure the French language in the
process.

[strong assertions ('...exactly...', 'nobody...'), put-downs
('JL's proposals are a farce'; implied: 'JL wants to defigure
the French language')]

2) >yes, they did...This is why we must be allowed to
>remain armed...who is going to help us if our government
>becomes a tyranny? no one will.

oh yes we *must* remain armed. anyone see day one last
night abt charlestown where everyone/s so scared of
informing on murderers the cops have given up ? where
the reply to any offense is a public killing ? knowing
you/re not gonna be caught cause everyone/s to afraid to
be a witness ?

yeah, right, twerp.

> ----[Ron] "the Wise"----

what a joke.

[sarcasm, name calling, personal insults]

The second example would be characterized as a "flame" by most
readers because of its personally offensive nature.
Less exclusively male-gendered but still characteristic
of male postings is an authoritative, self-confident stance
whereby men are more likely than women to represent themselves
as experts, e.g. in answering queries for information. The
following example is from NOTIS-L.

3) The NUGM Planning meeting was cancelled before all of
this came up. It has nothing to do with it. The
plans were simply proceeding along so well that there
was no need to hold the meeting. That is my
understanding from talking to NOTIS staff just last
week.

[authoritative tone, strong assertions ('...nothing...',
'...simply...', '...just...')]

The female-gendered style, in contrast, has two
aspects which typically co-occur: supportiveness and
attentuation. 'Supportiveness' is characterized by
expressions of appreciation, thanking, and community-building
activities that make other participants feel accepted and
welcome. 'Attenuation' includes hedging and expressing doubt,
apologizing, asking questions, and contributing ideas in the
form of suggestions. The following examples, from a non-
academic list (WOMEN) and an academic list (TESL-L),
illustrate each type:

4) >Aileen,
>
>I just wanted to let you know that I have really
>enjoyed all your posts about Women's herstory. They
>have been extremely informative and I've learned alot
>about the women's movement. Thank you!
>
>-Erika

DITTO!!!! They are wonderful!

Did anyone else catch the first part of a Century of
Women? I really enjoyed it.
Of course, I didn't agree with everything they said....
but it was really informative.

Roberta~~~~~~~~~~~~~~~~~~~~~~~~~~

[appreciates, thanks, agrees, appeals to group]

5) [...] I hope this makes sense. This is kind of what I
had in mind when I realized I couldn't give a real
definitive answer. Of course, maybe I'm just getting
into the nuances of the language when it would be easier
to just give the simple answer. Any response?

[hedges, expresses doubt, appeals to group]

The female style takes into consideration what the sociologist
Erving Goffman called the "face" wants of the addressee --
specifically, the desire of the addressee to feel ratified and
liked (e.g. by expressions of appreciation) and her desire not
to be imposed upon (e.g. by absolute assertions that don't allow
for alternative views). The male style, in contrast, confronts
and threatens the addressee's "face" in the process of engaging
him in agonistic debate.
While these styles represent in some sense the extremes
of gendered behavior, there is evidence that they have symbolic
significance above and beyond their frequency of use. Thus
other users regularly infer the gender of message posters on
the basis of features of these styles. Cases where the self-
identified gender of the poster is open to question are
especially revealing in this regard. Consider the following
cases, the first involving a male posing as a female, the second
a suspected female posing as a male:

(i) A male subscriber on SWIP-L posted a message disagreeing
with the general consensus that discourse on SWIP should be
non-agonistic, commenting "there's nothing like a healthy
denunciation by one's colleagues every once in a while to get
one's blood flowing, and spur one to greater subtlety and
exactness of thought." He signed his message with a female
pseudonym, however, causing another (female) subscriber to
comment later, "I must confess to looking for the name of the
male who wrote the posting that [Suzi] sent originally and
was surprised to find a female name at the end of it." The
female subscriber had (accurately) inferred that anyone
actively advocating "denunciation by one's colleagues" was
probably male.

(ii) At a time when a male subscriber had been posting
frequent messages to the WOMEN list, another subscriber
professing to be a man posted a message inquiring what the
list's policy was towards men participating on the list,
admitting "I sometimes feel guilty for taking up bandwidth."
The message, in addition to showing consideration for the
concerns of others on the list, was very attenuated in style
and explicitly appreciative of the list: "I really enjoy
this list (actually, it's the best one I'm on)". This
prompted another (female) subscriber to respond, "now that
you've posed the question...how's one to know you're not a
woman posing this question as a man?" Her suspicion
indicates that on some level she recognized that anyone
posting a message expressing appreciation and consideration
for the desires of others was likely to be female.

The existence of gendered styles has important implications,
needless to say, for the claim that CMC is anonymous,
"gender-blind", and hence inherently democratic. If our
online communicative style reveals our gender, then gender
differences, along with their social consequences, are likely
to persist on computer-mediated networks.[4]
Entire lists can become gendered in their style as
well. It is tactily expected that members of the non-dominant
gender will adapt their posting style in the direction of the
style of the dominant gender. Thus men on women's special
interest lists attenuate their assertions and shorten their
messages, and women, especially on male-dominated lists such
as LINGUIST and PAGLIA-L, can be contentious and adversarial.
Arguably, they must adapt in order to participate
appropriately in keeping with the norms of the local list
culture. Most members of the non-dominant gender on any
given list however end up style-mixing, that is, taking on
some attributes of the dominant style while preserving
features of their native style, e.g. with men often
preserving a critical stance and women a supportive one at
the macro-message level. This suggests that gendered styles
are deeply rooted -- not surprising, since they are learned
early in life -- and that some features are more resistant to
conscious reflection and modification than others.

5. Different communication ethics
The second part of this talk concerns the value
systems that underlie and are used to rationalize communicative
behavior on the net. In particular, I focus on the phenomenon
of flaming, variously defined as "the expression of strong
negative emotion", use of "derogatory, obscene, or
inappropriate language", and "personal insults". A popular
explanation advanced by CMC researchers[5] is that flaming is
a by-product of the medium itself -- the decontextualized and
anonymous nature of CMC leads to "disinhibition" in users and
a tendency to forget that there is an actual human being at
the receiving end of one's emotional outbursts. However, CMC
research until recently has largely overlooked gender as a
possible influence on behavior, and the simple fact of the
matter is that it is virtually only men who flame. If the
medium makes men more likely to flame, it should have a
similar effect on women, yet if anything the opposite appears
to be the case. An adequate explanation of flaming must
therefore take gender into account.
Why do men flame? The explanation, I suggest, is that
women and men have a different communication ethic, and male
ethical ideals can be evoked to justify flaming. I stumbled
upon this realization recently as a result of a survey I
conducted on politeness on the Internet. I had originally
hypothesized that the differences in the extremes of male and
female behavior online -- in particular, the tendency for
women to be considerate of the "face" needs of others while
men threaten others' "face" -- could be explained if it turned
out that women and men have different notions of what
constitutes appropriate behavior. In other words, as a woman
I might think that adversarial behavior is rude, but men who
behave adversarially might think otherwise. Conversely, men
might not value the supportive and attenuated behaviors of
women.
In the survey, I asked subscribers from eight Internet
discussion lists to rank their like or dislike for 30
different net behaviors, including "flaming", "expressing
thanks and appreciation", and "overly tentative messages", on
a scale of 1 (like) to 5 (dislike). The survey also asked
several open-ended questions, including most notably: What
behaviors bother you most on the net?
My initial hypothesis turned out to be both correct
and incorrect. It was incorrect in that I found no support
whatsoever for the idea that men's and women's value systems
are somehow reversed. Both men and women said that they
liked expressions of appreciation (avg. score of 2), were
neutral about tentative messages (avg. about 3), and disliked
flaming (although women expressed a stronger dislike than
men, giving it a score of 4.3 as compared with 3.9 for men).
This makes male flaming behavior all the more puzzling;
should we conclude then that men who flame are deliberately
trying to be rude?
The answers to the open-ended questions suggest
another explanation. These answers reveal a gender contrast
in values that involves politeness but cannot be described in
terms of politeness alone. It seems women place a high value
on consideration for the wants and needs of others, as
expressed in the following comment:

If we take responsibility for developing our own
sensitivities to others and controlling our actions to
minimize damage -- we will each be doing [good deeds]
for the whole world constantly.

Men, in contrast, assign greater value to freedom from
censorship, forthright and open expression, and agonistic
debate as a means to advance the pursuit of knowledge.
Historically, the value on absolute freedom of speech
reflects the civil libertarian leanings of the computing
professionals who originally designed the net and have
contributed much of the utopian discourse surrounding it; the
value on agonistic debate is rooted in the western (male)
philosophical tradition.
These ideals are stirringly evoked in the following
quote from R. Hauben (1993) praising the virtues of the
Usenet system, on which 95% of the contributors are estimated
to be male:

The achievement of Usenet News demonstrates the
importance of facilitating the development of
uncensored speech and communication -- there is debate
and discussion -- one person influences another --
people build on each other's strengths and interests,
differences, etc.

One might think that uncensored speech if abused could cause
problems, but M. Hauben (1993) explains that there is a
democratic way of handling this eventuality:

When people feel someone is abusing the nature of
Usenet News, they let the offender know through e-mail.
In this manner...people fight to keep it a resource
that is helpful to society as a whole.

Consider, however, how the ideal of "people fight[ing] to
keep [the net] a resource that is helpful to society as a
whole", translates into violent action in the response of a
male survey respondent to the question: "What behaviors
bother you most on the net?"

As much as I am irritated by [incompetent posters],
I don't want imposed rules. I would prefer to "out"
such a person and let some public minded citizen fire
bomb his house to imposing rules on the net. Letter
bombing a annoying individual's feed is usually
preferable to building a formal heirarchy of net cops.

Another net vigilante responds more graphically as follows:

I'd have to say commercial shit. Whenever someone
advertises some damn get-rich-quick scheme and
plasters it all over the net by crossposting it to
every newsgroup, I reach for my "gatling gun mailer
crasher" and fire away at the source address.

These responses not only evoke an ideal of freedom from
external authority, they provide an explicit justification
for flaming: as a form of self-appointed regulation of the
social order, a rough and ready form of justice on the
virtual frontier. Thus a framework of values is constructed
within which flaming and other aggressive behaviors can be
interpreted in a favorable (even prosocial) light. This is
not to say that all or even most men who flame have the good
of net society at heart, but rather that the behavior is in
principle justifiable for men (and hence tolerable) in ways
that it is not for most women.

6. Netiquette
Further evidence that flaming is tolerated and
justified within a system of male values is the content of
written rules of network etiquette, or "netiquette", such as
are available on many public FTP sites and in introductory
messages to new members of some discussion lists. I analyzed
the content of netiquette rules from six lists, along with
those found in the guidelines for Usenet and in the print
publication _Towards an Ethic and Etiquette for Electronic
Mail_ by Shapiro and Anderson (1985). What do netiquette
rules say about flaming?
The answer is: remarkably little, given that it is
one of the most visible and frequently-complained about
"negatives" cited about the Internet. One might even say
there is a striking *lack* of proscription against flaming,
with the exception of a few women-owned and women-oriented
lists. And in the rare instances where flaming is mentioned,
it is implicitly authorized. Thus the guidelines for new
subscribers to the POLITICS list prohibit "flames of a
personal nature", and Shapiro and Anderson advise "Do not
insult or criticize third parties without giving them a
chance to respond". While on the surface appearing to oppose
flaming, these statements in fact implicitly authorize
"flames other than of a personal nature" (for example, of
someone's ideas or values) and "insulting or criticizing
third parties" (provided you give them a chance to respond).
Normative statements such as these are compatible with male
values and male adversarial style; the intimidating rhetoric
on LINGUIST and many other lists is not a violation of net
etiquette according to these rules.[6] Yet these are
behaviors that female survey respondents say intimidate them
and drive them off of lists and newsgroups. Can the Internet
community afford to tolerate behaviors that intimidate and
silence women? This is a question that urgently needs to be
raised and discussed net-wide.

7. Conclusions
To sum up, I have argued that women and men constitute
different discourse communities in cyberspace -- different
cultures, if you will -- with differing communicative norms
and practices. These cultures are not however "separate but
equal"; rather, the norms and practices of masculine net
culture, codified in netiquette rules, conflict with those of
the female culture in ways that render cyberspace -- or at
least many "neighborhoods" in cyberspace -- inhospitable to
women.
What can be done to address the imbalance? I'll
conclude by mentioning three ways in which I believe women
can promote their concerns and influence the discourse of the
net.[7] The first is to support and participate in women-
centered lists. Such lists provide comfortable fora for
women online, and are frequently models of cooperative
discourse whose norms then become available for wider
application if subscribers participate in other lists as
well. But separatism has its disadvantages, among them the
risk of ghettoization. Women must not let themselves be
driven by flame throwers away from mainstream, mixed-sex
fora, but rather should actively seek to gain influence
there, individually and collectively, especially in fora
where metadiscourse about the net itself takes place.
The second way to promote women's interests net-wide
is to educate online communities about the rhetorical
strategies used in intimidating others, and call people on
their behavior and its consequences when they use such
strategies. This is already happening on some women-centered
lists such as WMST-L and SWIP-L -- aware of the tendency for
a single man or group of men to dominate discussions even on
women-centered topics, female subscribers call attention to
the behavior as soon as they realize it is happening, and
interestingly, it is happening less and less often on these
lists. Group awareness is a powerful force for change, and
can be raised in mixed-sex fora as well.
Finally, women need to participate in any way they
can in the process that leads to the encoding of netiquette
rules. Instigate and participate persuasively in discussions
about what constitutes appropriate and inappropriate behavior
online -- seek to define in concrete terms what constitutes
"flaming", for instance, since women and men will probably
have different ideas about this. Take the initiative and
write down guidelines for suggested list protocol (or
modifications to list protocol if guidelines already exist)
and post them for discussion. No greater power exists than
the power to define values, and the structure of the Internet
-- especially now, while it is still evolving and seeking its
ultimate definition -- provides a unique opportunity for
individual users to participate in the normative process.
Indeed, it may be vital that we do so if women's
behavior is to be accorded value, and if we are to insure
women the right to settle on the virtual frontier on their
own -- rather than on male-defined -- terms.

NOTES

1. A notable exception to this generalization is the work
of Sherry Turkle in the 1980's on how women and men relate
to computers.
2. For an interesting example of this phenomenon on the
soc.feminism Usenet newsgroup, see Sutton (1994).
3. Herring, Johnson, and DiBenedetto (1992).
4. This problem is discussed in Herring (1993a).
5. For example, Kiesler et al. (1984), Kim and Raja (1990),
and Shapiro and Anderson (1985).
6. The discussion of politeness and communication ethics is
an abbreviated version of that presented in Herring (To
appear).
7. For other practical suggestions on how to promote gender
equality in networking, see especially Kramarae and Taylor
(1993).

REFERENCES

Hauben, Michael. 1993. "The social forces behind the
development of Usenet News." Electronic document.
(FTP weber.ucsd.edu, directory /pub/usenet.hist)

Hauben, Ronda. 1993. "The evolution of Usenet News:
The poor man's ARPANET". Electronic document. (FTP
weber.ucsd.edu, directory /pub/usenet.hist)

Herring, Susan. 1992. "Gender and participation in computer-
mediated linguistic discourse." ERIC document.

Herring, Susan. 1993a. "Gender and democracy in computer-
mediated communication". Electronic Journal of Communication
3(2), special issue on Computer-Mediated Communication, T.
Benson, ed.

Herring, Susan. 1993b. "Macrosegmentation in postings to
two electronic 'lists'". Paper presented at the Georgetown
University Round Table on Languages and Linguistics,
Presession on Discourse Analysis: Written Texts, March 1993.

Herring, Susan. 1993c. "Men's language: A study of the
discourse of the Linguist list." In Crochetiere, Boulanger,
and Ouillet, eds., Proceedings of the XVth International
Congress of Linguists. Quebec: Universite' Laval.

Herring, Susan. To appear. "Politeness in computer culture:
Why women thank and men flame." In Bucholtz and Sutton, eds.,
Communicating Across Cultures: Proceedings of the Third
Berkeley Women and Language Conference. Berkeley Women and
Language Group.

Herring, Susan; Deborah Johnson; and Tamra DiBenedetto. 1992.
"Participation in electronic discourse in a 'feminist' field."
In Bucholtz, Hall, and Moonwomon, eds., Locating Power:
Proceedings of the Second Berkeley Women and Language
Conference. Berkeley Women and Language Group.

Kiesler, Sara; Jane Seigel; and Timothy W. McGuire. 1984.
"Social psychological aspects of computer-mediated
communication". American Psychologist, 39, 1123-1134.

Kim, Min-Sun and Narayan S. Raja. 1990. "Verbal aggression
and self-disclosure on computer bulletin boards." ERIC
document (ED334620).

Kramarae, Cheris and H. Jeanie Taylor. 1993. "Women and
men on electronic networks: A conversation or a monologue?"
In Taylor, Kramarae and Ebben, eds., Women, Information
Technology and Scholarship, 52-61. Urbana, IL: Center for
Advanced Study.

Rheingold, Howard. 1993. The Virtual Community:
Homesteading on the Electronic Frontier. Reading, MA:
Addison-Wesley.

Seabrook, John. 1994. "My first flame." The New Yorker,
June 6, 1994, 70-79.

Shapiro, Norman Z. and Robert H. Anderson. 1985. Toward
an Ethics and Etiquette for Electronic Mail. The Rand
Corporation.

Sutton, Laurel. 1994. "Gender, power, and silencing in
electronic discourse on USENET." Proceedings of the 20th
Berkeley Linguistics Society. UC Berkeley.

Turkle, Sherry. 1984. The Second Self: Computers and the
Human Spirit. London: Granada.