Electronic Letters to a City Council
factors influencing the composition of email messages
This study is an analysis of
electronically transmitted letters (i.e. email messages) and traditional paper letters
(i.e. handwritten or typewritten letters on paper) from citizens to the authorities of the
city council in Göteborg city. The norms for email are still in the process of being
established. People are uncertain how to formulate themselves in this rather new medium.
Most studies of email have been made on email messages of one-to-many interaction in
public mailing lists. This is a study of public one way, one-to-one email
the receiver is an unknown authority.
The overall purpose of this study is to try to establish which factors
influence how people formulate themselves in a textbased electronic medium. Do email
messages to authorities conform to the business template of traditional formal letters, or
is it the ease and rapidity of the electronic medium that pose the greater influence on
the way the senders formulate their messages? Or are there combinations of other
Results from this study confirm results from previous studies of email (Herring, 1996, Du
Bartell, 1995, Danet, forthcoming), suggesting that email messages to authorities are less
formal and shorter than the formal business template. Email often seems to serve other
communicative purposes replacing phone calls (Severinson-Eklundh, 1994).
4. Material and method
About the Author
We frequently use computer-mediated communication (CMC) today through
the written medium via networked computers, both for private and for professional
purposes. One of the reasons for using the fast transmitting computers is the convenience
it brings. Email has a number of advantages compared to traditional written letters: it is
an easy, fast way of getting in touch with people, and it is also low in cost.
Transmission time is very much shorter than that of the traditional postal service, and a
reply could possibly be received within minutes. It is also possible to attach files of
various kinds (e.g. sound files, word processor documents, etc.) to email messages.
As email messages are written, certain demands and constraints are put
upon both sender and user. Spoken interaction is multimodal, making use of several
channels simultaneously for sending information. Written interaction has to rely on the
single and linear channel of vision for communicating textual messages. Strategies, such
as the use of smileys or abbreviations (see "This study" below) have been
developed to overcome the difficulties of the written medium in order to avoid
misunderstandings and ambiguities, and still be able to make use of the speed of
transmission that CMC technologies allow for.
It is imaginable that the evolving possibilities of electronic
communication change the way people approach each other in writing. What psychological and
contextual factors influence the way writers compose their electronically transmitted
messages? Studies have shown that the purposes for communication, as well as topic and
medium for communication play a part in the way messages are formulated (Baym, 1996; Du
Bartell, 1995; Hård af Segerstad, forthcoming1). Other factors, such as the relation between sender and addressee
(Danet, forthcoming), grounding and closure on the actions (Clark, 1996) also have their
share. Whichever medium one selects to communicate through, it will have an impact on the
potential forms of language which may be manifested (Du Bartell, 1995: 232). The ease of
access to sending messages and the users relative anonymity might also influence the
way in which electronic communication is formulated.
Other studies have shown that the faster the medium, the more like
spoken language the written messages get (cf. Horowitz and Berkowitz, 1964). Other studies
have shown that email messages often get more informal both in terms of composition
(salutation and closing conventions) and form (spelling, syntax) (cf. Herring, 1996).
Email messages are rapidly composed and transmitted, and low in cost
and effort of production and transmission. They are written rather than spoken and allow
the sender to remain relatively anonymous.
(To the top)
This study aims at analysing what could
be the underlying factors behind how people compose their electronic messages (email) that
they send to an unknown authority at the city council of the city of Göteborg, Sweden. A
comparison with traditional paper letters of the same type will be made. Intuitively, one
expects that the relation between sender and addressee will have an influence; letters
approaching "authorities" will conform to the formal business template (cf.
The ease and rapidity of production and transmission is hypothesised in
the present study to make email messages more "speech-like" and less formal than
traditional letters, which are conforming to the business template (see "The Swedish
business letter template and norms for informal letters" below). At the same time,
email messages are still written and need to rely more on the typed words than on
contextual information, and in this sense email messages tend to be more
"written-like". The written mode may also make people feel that they may remain
relatively anonymous and stay "hidden" behind the text, as it were. Most email
software automatically include the senders name and email address, which one would
expect would lead the sender to omit his or her name in closing the message.
The analysis was made by examining whether the traditional paper
letters and the email messages conform to the formal business template with respect to
epistolary conventions such as salutation and closing conventions (cf. Danet,
forthcoming), in combination with an analysis of contextual factors such as the status of
the sender, the medium, the purpose and topic of communication.
(To the top)
Analysis of messages in an electronic Swedish
Extensive work on electronic mail as a medium for dialogue has
been made. Severinson-Eklundh (1994) argues that the computer medium may serve
communicative purposes previously reserved for direct, spoken conversation. Her analysis
of a body of messages from the Swedish conference system COM (1986), showed that messages
were structured into coherent dialogues in a way distinct from other forms of written
communication. Dialogue sequences appear in email simply as a result of linking between
messages and their replies. The exchange of email messages may contain a range of
speech-like or "conversational" features. This applies both to the character of
individual messages and to the structural features of the entire dialogue.
Severinson-Eklundh states that the email messages in her study typically were short
compared to regular letters. Furthermore, almost all messages used a direct, informal
style of address.
Analyses of messages in electronic mailing lists
Du Bartells (1995) study of the features of the messages
in a mailing list suggests that the spoken and written-like characteristics in a written
medium result from the constraints imposed by the computer medium the machine
architecture. The computer medium permits texts which seem both written-like and
spoken-like. CMC messages display linguistic characteristics typically associated with
spoken language and other forms of written language in addition to linguistic features
specific to the medium. Du Bartell argues that we expect written language to be edited,
planned, articulated without recourse to non-standard constructions, slang and vulgar
expressions. From speech we expect more or less the opposite: we expect slang, the
non-standard grammatical constructions, the sudden topic shifts and spontaneity. "CMC
gives us these in writing. CMC discourse exhibits the type of grammatical constructions
that appear in non-edited non-standard spoken language of face-to-face interaction"
(DuBartell, 1995: 233).
Herring (1996) analysed the schematic organisation of electronic
messages posted to two academic mailing lists, one mostly male and the other mostly
female, in order to evaluate the popularly held view that men and women use email for
different purposes (information exchange vs. social interaction). Her results did not
support the stereotype, but showed that womens and mens messages are
structured differently, with female users exhibiting alignment, and male users opposition,
towards their addressees.
The basic electronic message schema was analysed into epistolary
convention of salutation, introduction, body, and close. She concludes from analysing 136
messages that "Surprisingly few messages are preceded by a salutation (only 13 % on
average), and fewer yet are followed by a complimentary close or a postscript."
(Herring, 1996: 87).
Both Herring and Du Bartell explain the relative lack of epistolary
conventions to be partly because of the fact that a header is added automatically to each
message by the electronic mailer, including a separate line for whom the message is
"from", whom it is addressed "to", and the date and time of posting.
Partly as a result of having a subject displayed, email messages frequently omit even the
typical salutations and farewells associated with other media, regardless of whether or
not the speakers2 know each other. Email messages do display rather informal register
characteristics, even between people unknown to each other (Du Bartell, 1995).
Baym (1996) argues that although CMC is written it is marked by many
features associated with face-to-face interaction. Her study of Usenet messages3 showed that Usenet
interaction is a hybrid between oral, written, interpersonal, and mass communication. Baym
concludes that the message features of her study stem from five interrelated factors
the Usenet medium, the institutional context of work, the topic, the
participants gender, and the social context which the participants strive to create.
Analysis of private email
Because of the problems of ethic considerations, not many
studies analysing private email, rather than public mailing list messages, have been
published. For this reason, and for the reason of analysing private, one-to-one email
messages, Brenda Danet analyses portions of her own email correspondence (in Keybo@rd
K@perz [sic] artful communication on the Internet, forthcoming). Her
email study focus was on letters sent to her by people who did not know her but knew her
name and status. She remarks that
] a writer of a first letter is likely to take special care
in its formulation. Traditional norms for letter form are likely to be salient, and
writers are likely to be especially conscious of the impression their message may make on
Danet analysed her email messages holistically, using the criteria of
the business letter template. In the case of openings and closings, letters were coded
"yes" only if they contained both an appropriate opening and an appropriate
closing. Abbreviations, spelling, typography, punctuation, and use of exclamation points
conforming to the norm were coded "yes". She found that none of the letters of
her corpus conformed to all her criteria, and that the variability was extreme. Most
letters conformed to expectations regarding syntax and vocabulary, as well as those for
spelling, typography and layout, but almost none followed paper letter practice regarding
openings and closings.
Danet argues that the new medium invites informality even in business
or official contexts. This is not just due to the technology per se, but converging with a
general trend which she takes to have been in place already. She remarks that the novel
medium can facilitate changes of style and substance in much shorter time, than would
paper letters have done. Style, or register, may apply to substantive domains of human
communication and action not to all communication in a medium (cf. Allwoods
activity based communication analysis [1976, 1995]). Danet concludes that the language of
email is in a state of transition. She predicts, among several things, that an informal,
partially speech-like email style will increasingly characterise public as well as
personal communication. Our normative expectations will change to provide increased
legitimation of a more informal style.
Closure on actions
Herbert Clark (1996: 222) argues that a fundamental principle
of intentional action is that people look for evidence that they have done what they
intended to do. People need closure on their actions. He argues that to get closure on an
action, one looks for evidence that one has succeeded. This principle applies to
intentional actions of all kinds. Evidence of success must be valid to be useful; it must
be reliable and interpretable. Evidence must also be easy to get, economical in effort.
Evidence must be timely. Without such evidence, one may try the action again, or try to
repair what went wrong. In conversation people ordinarily go to some extent to reach joint
closure on their actions. An answer to a question gives evidence that the question is
perceived and understood. This applies not only to spoken conversation, but to written
communication as well.
The analyses made by Severinson-Eklundh, Du Bartell, Herring
and Baym above all concern email dialogue. Email dialogue consists of an ongoing
discussion comprising a series of messages which are interconnected: a person sends a
message which is met with one or more replies; sends off another message or replies in his
or her turn. The mailing list messages are contributions in a many-to-many interaction.
However, this study deals only with single one-way messages4 from individuals to a remote and unknown
"authority". Like Danets study, the material consists mostly of
"first letters" from people who do not know the recipient personally. Unlike
Danets material, the senders did not know the recipient by name, but possibly only
had a notion of the higher status of the authorities at the city council.
Written language in general lacks some of the information cues that are
conveyed in spoken language. An utterance conveying words such as, e.g. "Its
your fault" may not lead the risk of being taken as an accusation or an insult. The
listener takes into consideration more than just what is being said. The speakers
and listeners shared background knowledge, the context in which the words are
uttered, non-verbal information picked up from the tone and intensity of voice, facial
expressions, gaze, gestures etc. can all add up to an interpretation of the utterance as a
jestful remark or sarcasm or the like. The same contribution in nothing but plain text
leaves the receiver in so much more doubt as how to interpret the message. Text-only
without clarifying comments can be very ambiguous and difficult to decipher.
In a previous study (Hård af Segerstad, forthcoming) dealing with
strategies in written language in a chat room, there emerged some distinct strategies for
the purpose of overcoming some of the constraints on written language. Strategies such as
the use of
- Emoticons, or smileys, in
resemblance of facial expressions
- Abbreviations and acronyms
- Words or phrases written in capitals only
- No mixed cases
- Extensive use of punctuation marks
- Fonts and colours
- Asterisk framed words or phrases
are used to enhance the written language,
prevent misunderstanding, and reduce the time and effort of production. Some of these are
innovations of the written language, and specific for CMC. Others have been used for the
same purpose in traditional written language (e.g. abbreviations, punctuation marks), but
perhaps to a lesser extent.
According to Horowitz & Berkowitz study from 1964, part of
the many differences found between spoken and written expression are due to the greater
ease of speaking ones mind than writing it. Any mode of writing5 that increases the ease of
production of this mode should result in the production of cognitive and linguistic
material closer to that produced in spoken expression (Horowitz & Berkowitz, 1964:
The Swedish business letter template and norms for informal letters
The Swedish business letter template differs slightly from the
British-American one as described by Brenda Danet (forthcoming): "Most generally, the
standard paper business letter is supposed to be cast in a formal styleto use
language appropriate to formal situations". The British-American formal letter opens
with a salutation or greeting ("Dear Sirs" or the like), has blank space between
the salutation and the body of the message and between the body of the message and the
closing. Furthermore, it has to conform to the norms and conventions of spelling,
punctuation and orthography (e.g. sentence initial capitalisation, no contractions and no
The Swedish formal business template6 opens with the topic or concern of the
letter, often underlined or in bold face. It is not the practice to open with a
salutation. The senders address, as well as the receivers and the date are
often placed at the head of the letter. Like the formal British-American business template
it has a blank space between the salutation and the body of the message and between the
body of the message and the closing. It also has to conform to the norms and conventions
of spelling, punctuation and orthography.
On the other hand, informal letters in Sweden, like letters between pen
pals, mostly open with a salutation or greeting. The informal "hej"
("hi"), or variants of it ("hejsan", "hallå") are often
used. Danet also notes that "In the Anglo-American tradition, personal letters have
always been more conversational and informal than business or official ones", which
is true also in the Swedish case.
(To the top)
The data of this study consists of electronic mail from citizens to
the city council of Göteborg, Sweden. The messages are public material, open to anyone
who wishes to take part of the city council material. The material consists of 183
messages sent between April 1, 1998 and August 31, 1998. The web page of the city council
provides a "Questions & Answers" service, at which anyone has the
possibility to send messages of any sort concerning issues that they hope the people at
the city council might be able to help them with. Originally, the service was designed for
questions and remarks concerning the web page itself, but people sent messages of all
kinds, and it was eventually decided that this service should pose no restrictions on what
kind of messages to handle.
People were asked to fill in a form with the following fields:
"Sender" (email return address in order to get an answer back), "Subject/
topic" (subject of the message), and lastly the field "Text" in which the
message was composed. After filling in the form and writing the message, the message was
sent by clicking a "send" button at the bottom of the page7. The form looks basically
like most email software, but the return address is not automatically added.
Traditionally written paper letters
Data from 25 traditionally written paper letters was also analysed:
4 were handwritten, 17 used a computer, and 4 were typewritten. The letters were of the
same kind as the email messages: open8 letters from citizens asking questions, requesting help, or the like.
The comparatively low number of traditional letters is due to the problems that the staff
at the city council archive had with abstracting material out of their filing system. A
further study will have to include an analysis of a larger number of traditionally written
letters, which would be more suitably comparable to the email corpus.
Both the traditional letters and the
email material were collected from the city council with the help of the staff at the
information and archives departments9. It was stored, and analysed digitally using an automated tool
TRASA which was developed by Leif Grönqvist at the Department of Linguistics,
Göteborg University, Sweden. For this study we used this software for quantitative
analyses of the occurrence of abbreviations and punctuation marks, closing and
introductory words in messages, mean length of utterance, etc.
The main focus of analysis was qualitative. Qualitative analyses of
both email messages and traditional letters aimed at rating them for whether or not they
conformed to the business letter template (cf. Danet, forthcoming). Focus was on
salutations, pre-closings, closings, and signatures. In order to try to establish factors
influencing how people compose their messages, cross analyses was made of the gender of
the sender, the status of the message (sent for private or for professional purposes),
type of communication (e.g. question, complaint, etc.), and topic of the messages.
(To the top)
Overview of the corpora
Figure 1: The distribution of messages sent by females, males,
unknown, and multiple senders in the email and paper letter corpora (percentage of total).
An overview of who sent the messages
shows that most of the messages in both corpora were sent by men. When a message contained
no clues as to gender, it was classified as being sent by an "Unknown sender".
The senders email address often contains clues, but many addresses were of the
cryptic type, made up by a combination of letters and numbers. Such cryptic senders who
did not sign their messages with their name had to be classified as "Unknown
senders". Some messages may have been sent by someone using someone elses email
account, a fact which one could bear in mind but which will have to be ignored as it could
not possibly be discovered. A number of messages were sent by more than one sender
a family or two people working together, for instance and thus classified as
"Multiple senders". Note that none of the paper letters was categorised as being
sent by an unknown sender.
Type of message
Figure 2: Distribution of type of message in email and paper
letter corpora (percentage of total).
The main type of the email messages consisted of questions (78%). The
main type of the traditional paper letters consisted of complaints (44%), and requests or
appeals for help (28%). Most of the messages were questions or requests for information of
various sorts. Other messages were remarks or complaints; the largest category of the
paper letters was complaints. In email cases complaints were mostly about the failing
information at the web site. Several messages were complaints about matters in the
opening hours of the museums or parking facilities, for instance. A number of messages
contained offers of services or suggestions: people offered to work or to send information
that could be useful at the web page or for the people at the city council, some suggested
co-operation. Many of the messages contained both a question and a remark of some kind,
but were classified as that which seemed to be the primary cause for communication.
Reasons for communication
Figure 3: Email messages and paper letters sent for private
or professional reasons
(percentage of total).
Most of the messages were clearly sent
for private reasons: citizens of Göteborg or someone seeking information of various
kinds. 74% of the email data, and 84% of the paper letter material were sent for private
reasons. 25 % of the email messages, and 16% of the paper letters were sent for
professional, or business related reasons. This was concluded from topic, signature and
senders address. The better part of the messages, though, was sent for private
Analysis of the letters
My analysis of the letters is
influenced by Brenda Danets. She categorised her email into groups conforming in
various degrees to the business template with regards to salutation and closing
conventions. Her smaller corpus allowed for a more scrutinising analysis of each message,
but I was forced, for limitations of time and space, to exclude a closer analysis of each
letter in this study with respect to spelling, punctuation, and occurrence of informal
The email messages as well as the paper letters of this study were
analysed for salutation conventions and whether or not the messages were signed with the
senders name. The Swedish business letter template does not require salutation, but
a topic or the letters concern, of which the subject line of the email format is
taken to be the equivalent. Four categories emerged:
- Messages introduced with a salutation
and signed with the senders name (+salutation/+signature).
- Messages introduced with a salutation, but not signed with the
senders name (+salutation/-signature)
- Messages left without salutation, but signed with the senders
- Messages neither introduced with a salutation nor signed with the
senders name (-salutation/-signature).
Salutations and signatures
Figure 4: Diagram visualising the distribution of email and paper letters
into categories of salutation and signature (percentage of email and paper letter corpora
As seen in fig. 4, the category
+salutation/+signature occupied 48% of the total number of email messages, and 8% of the
total number of paper letters. This category of the email messages
(+salutation/+signature) covered 48% of all questions, 50% of all suggestions, and 35% of
all complaints. The only two instances of smileys occurred in this category which conforms
to the norms with respect to salutations and signature. Both messages were informal in
spelling and syntax. Smileys did not occur at all in the paper letters.
The second category,
+salutation/-signature, consisted of 5% of the email material and had no representations
in the paper letter data. Only 5% of all email messages were opened with a salutation and
not signed at all. The two types of the email messages (+salutation/-signature) were
questions and complaints. This group featured 6% of all questions, 7% of all complaints.
The third category, -salutation/+signature, occupied 25% of all email
messages and 92% of the total number of paper letters. 25% of the email messages had no
salutation, but were closed with a signature. This category conforms the most to the
traditional Swedish business letter template, if we allow for the subject line to serve as
the subject opener of paper letters.
In this category, in the email material we find 22% of all questions,
27% of all suggestions, 50% of all requests, and 42% of all complaints.
The last group, -salutation/-signature,
consisted of 22% of the total number of email messages and was not represented at all in
the paper letter data. These messages conformed the least to the business letter template,
in having no salutation and no signature. Only four of them had a pre-closing of some
sort. This group is the most "email like", and it is also in this category that
we find the most unknown senders (11% of all unknown senders). This fact is probably due
to the email format which normally give name and address of the sender automatically, but
wich the web form does not provide (cf. Herring, 1996). In this category we find 23% of
all questions, 22% of all suggestions, 14% of all complaints, and the only clear example
of a nonsense message (see Conclusions).
Conclusions from the categories
As we saw from the diagram above, the
email messages are spread out over the four categories, whereas the paper letters only
figure in two categories. In this sample, the email style is more varied. Moreover, it
seems that the email messages and the paper letters almost show opposite features.
None of the traditional letters were categorised into -salutation/-
signature or + salutation/- signature. All traditional letters but two, were the most
consistent with the Swedish business template: not preceded by a salutation but ended with
a signature. The same category of email messages consisted of 25% of the email messages.
The largest category of the email messages, 48%, were opened with a salutation and closed
with a signature, but only 8% of the traditional letters in this small sample were in this
These results do confirm the results of previous studies of email in a
way which may seem contradictory at first glance: Herrings findings from the mailing
lists suggested that surprisingly few messages were preceded by a salutation (Herring,
1996) - only 13% on average, where we find 48%. This is a somewhat contradictory fact if
we take the business template to be the form to refer to, as the British-American formal
letter requires a salutation and the Swedish one does not. This suggests that the email
messages of this study are less formal than the business template, which is quite in line
with the results from previous studies.
5% of the email messages were preceded by a salutation but not signed
with the senders name this category scored 0 in the traditional paper letter
The most "email like"
messages, the ones with neither
salutation nor signature, would perhaps be expected to be more frequent had the messages
been posted in an ordinary email programme for private use. The messages in this study
were composed and posted in an electronic form at a web page and the senders name
was not automatically added. Even so, 22% of all email messages were of this type. This
category had no equivalent in the paper letter corpus.
The last category described above only appears in the email corpus and
does not appear at all in the paper letter data. These facts confirm the argument that the
normal email architecture and the information automatically given ("From", date
and time of posting) make this unnecessary to type. In paper letters, on the other hand,
it is necessary to include name and address manually if the receiver is to know who the
Previous studies have suggested that email messages are shorter than
paper letters (c.f. Severinson-Eklundh, 1994). This study clearly confirms this fact: the
mean length of the email messages was 52,78 words, and that of the paper letters was
Salutation/signature and gender
Below follows a cross analysis of
salutation/occurrence of signature and gender of the senders. This combination was chosen
mainly to investigate whether or not there are differences between the email and letter
writing conventions of men and women. Other combinations of parameters are of course
possible, but will have to be left for further studies.
Figure 5: Percentage
of the total number of messages distributed into four categories of email
Fig. 5 shows the distribution of the
gender of the senders of email messages over the four categories. Men seem to keep to the
informal paper letter conventions of category 1 slightly more than women do, but this is
at the same time also the category which was most represented by both male and female
senders. We can draw no clear conclusions of differences of gender in email style with
respect to salutation and signature conventions. Category 4, the most "email
like" group of messages, had a high representation of unknown senders. This is also
the case with category 2, which was the least represented of the four. It could be argued
that this is due to the anonymising effect of the email medium: it is possible to hide
behind the text, as it were, or behind a cryptic email address that gives nothing
The messages were analysed for
different ways of salutation. Of the email messages, 98 opened with some sort of
salutation, and 86 were sent without salutation. All salutations were of the informal kind
(variations of e.g. "Hej" ["Hello" or "Hi"]) followed by
variants of conforming to the norms for punctuation: exclamation mark, comma, full stop or
no punctuation at all. I agree with Herring and Du Bartell (see "Background"
above) that this is a common characteristic most probably stemming from the fact that both
topic and sender is announced in the header of each message.
Of the traditional paper letters, only two opened with a salutation:
one with the informal "Hej!" ("Hi!"), while the second one, written in
English for professional reasons, was introduced with "Dear Sirs", and conformed
to the business letter template in all aspects.
44% of the email messages had closing
phrases of some sort. 3% were closed with the signature only, and 17% were sent with
neither closing nor signature.
The use of abbreviations in closing was common. 19% of the email
messages were closed with some variant of abbreviating the normative, formal closing
phrase (cf. Stenson) "Med vänliga hälsningar" (literally "With friendly
regards"). The abbreviated phrase showed a wide range of variations: examples are
m.v.h., MVH, mvh, M.V.H., and so on. None of them is the correct way of signing a formal
letter. 23% of the email messages were closed with variants of the formal "Med
vänliga hälsningar". Examples are "vänligen" ("in a friendly
manner"), "Vänlig hälsning" ("Friendly regards"), and so on.
40% of the traditional paper letters were closed with the full formal
"Med vänliga hälsningar". None of them was closed with the abbreviated form.
24% were only signed with the senders name. 8% closed with the date and
senders name. 28% closed with some pre-closing ("Tack på förhand"
["Thanks in advance"] or the like) in combination with the senders name.
(To the top)
The aim in this paper was to analyse
electronically transmitted written messages from citizens to the city council of
Göteborg, Sweden, and thus try to establish what psychological and contextual factors
influence peoples way of composing messages in electronic communication to unknown
authorities. It was hypothesised that factors such as the purposes for communication,
topic and medium for communication play a part in the way messages are formulated. Other
factors such as the relation between sender and addressee, grounding and closure on the
actions, the ease of access for sending messages, and the users relative anonymity
might also influence the way in which electronic communication is formulated.
The results seem to confirm suggestions from previous studies that the
norms for email are still in the process of being established. People are feeling
uncertain as to what conventions to use and what the effects of their messages will be.
People are juggling with both written and spoken conventions when formulating their
electronic messages. Are email messages to be like speech formulated after telephone
behaviour (without visual and vocal cues) or after the written norms of traditional
letters? The style of the messages in this study cover the range from formal, letter-like
messages, to informal messages with all the features to be expected of email messages.
They are clearly shorter than traditional paper letters, supporting evidence from
Just about 48 % of all the messages in this analysis kept to the
traditional way of writing informal letters and were introduced with some sort of
salutation and concluded with some kind of closing convention. Only 5 % of the messages
were introduced with a salutation of some sort and concluded with no closing convention at
all. 25 % of the messages were not introduced with any kind of salutation convention but
concluded with closing conventions of some sort. 22 % of the messages were neither
introduced with a salutation nor concluded with any kind of closing convention.
Herring (1996) and Du Bartell (1995) explain the relative lack of
epistolary conventions found in the messages of their analyses, to be in part accountable
to the architecture of email programs. Email technology in itself is easily accessible,
and the typing of messages and the transmission time is rapid. Sending an email is also
low in both cost and effort. E-mail and the increased accessibility seems to make people
less scared of having their say in writing, and the quick and easy way of sending
email messages may influence the way people write, and what they write about. Email is
still in the written mode, and allows the sender to remain relatively anonymous, while at
the same time having to struggle with the monomodality of written language.
Danets suggestion of a more
socio-/psycholinguistic kind, is that
the relation between the sender and the receiver is a factor that might influence how
people compose their messages. First letters to an authority should conform to the formal
norm of letter writing. This is perhaps not so well confirmed in this study, and might in
part be due to cultural differences in letter writing. The senders age was not
possible to find out in this study, and gender was in some cases difficult to establish
too. Still, I believe that the senders age and gender most probably affect the
The closure time (Clark, 1996) of an email is uncertain. The sender
does not in advance know whether s/he will get an answer to his or her mail, nor is it
clear whether the message was sent or not.
The choice of style may not always be apparent to the writer. One of
the reasons for choosing a particular style seems to be the purpose of the communication.
If the purpose is to obtain help or information, there is a tendency to mind ones
language and send "correct" messages. If the purpose is to complain or remark,
there seems to be a tendency towards lesser politeness and lesser formality displayed by
the lack of salutations or closing conventions.
Anecdotal evidence from the staff at the city council suggests that
email often replaces telephone calls. They also suggested that the types of messages they
get through email are of a different sort than traditional paper letters. It seems that
people feel they can hide behind the text and also behind screens, and remain relatively
anonymous. The ease and rapidity of sending an email make people send messages which they
probably would not have bothered to send had they been forced to find pen, paper,
envelope, stamps, letter box, and so on and so forth. Besides, when people sent the email
messages of this particular study, they were already on-line and surfing the Net. Below is
an example which I believe illustrates this beautifully, and which I cannot bear to
withhold from the reader. I dare guess that it would never have been sent in a traditional
Ämne: Hata lantisar
Ni götborgare är jävla lantisar. Ni kan sluta va så jävla kaxiga. Ni är en liten
skitstad mot Stockholm.
From: the Stocckholmer
Subject: Hate country bumpkins
You gothenburgers are real country
Stop being so damn cocky. You are a hole compared to Stockholm.
This could be seen as evidence that the accessibility and
least-possible-effort it takes to contact someone in writing produces messages of a
different type than that of which traditional letters normally are.
Time and space limit the scope of this analysis, and leave many
questions unanswered or not analysed in as much detail as a fuller study would provide. It
is my intention that further studies will address some of these issues in a more
(To the top)
Ylva Hård af Segerstad is a PhD student at the Department of
Linguistics at Göteborg University in Sweden. She is currently working on her thesis
which deals with the adaptation of Swedish written language to different kinds of
electronic communication, including studies of email, chat, avatar chats, and web
The research for this article lies within
the project Anpassningen av svenskans skriftspråk till elektronisk kommunikation
adaptation of Swedish written language to electronic communication) which is funded by
HSFR, Sweden.Funding from Birgit and Gad Rausing Stiftelse för Humanistisk
Forskning provided the computers with
which all data was collected and analysed - and on which this article was written.
(To the top)
1. Hård af Segerstad, Ylva. Forthcoming. "Swedish Teenagers
Written Conversation in Electronic Chat Environments". In: Penrod, Diane (ed.). WebTalk
- writing as conversation. Mahwah, NJ: Lawrence Erlbaum Associates.
(Back to the text)
2. Du Bartell employs the terms speaker and
listener regardless of the medium of linguistic communication.
(Back to the text)
3. Usenet is an enormous collection of topically organised discussion
groups distributed through the Internet. (Back to the
4. On inquiring, I was informed that all messages were in fact
eventually answered, but that material was sadly enough not given for analysis and thus
not included in this study. (Back to the text)
5. Horowitz & Berkowitz compared handwriting, typewriting and
stenotyping in 1964. (Back to the text)
Skriva i Tjänsten ("Professional Writing")
(1997) was consulted for the norms of the traditional business template.
(Back to the text)
7. The web page form can be found at
<http://www.goteborg.se/wwwdb/gbgwww.nsf/fragorochsvar> (5 August, 2000).
(Back to the text)
8. "Open" in this sense means that the letters were addressed
to anyone at the city council, and not to someone in particular.
(Back to the text)
9. My warmest thanks to the friendly and helpful staff at the city
council, Göteborg, Sweden: Udo Metz, Lotta Sundström, and Annette Johannesson.
(Back to the text)
Allwood, Jens. (1995): "An
Activity Based Approach to Pragmatics". In Gothenburg Papers in Theoretical
Linguistics 76, Dept. of linguistics, University of Göteborg. Forthcoming in Bunt
& Black (eds.) Approaches to Pragmatics.
Baym, N. K. (1996): "Agreement and Disagreement in a
Computer-Mediated Discussion". Research on Language and Social Interaction,
29(4), 315-345. Lawrence Erlbaum Associates, Inc.
Clark, H. H. (1996): Using Language. Ch. 8. Cambridge: Cambridge
Danet, B: Forthcoming. Keybo@rd K@perz - artful communication on the
Du Bartell, D. (1995): "Discourse Features of Computer Mediated
Communication: spoken-like and written-like". In Organization
in Discourse. Wårvik, Tanskanen, Hiltunen (red.) Turku: Turun Yliopisto.
Herring, S. C. (1996): "Two Variants of an Electronic Message
Schema". In Herring, S. C. (ed.) Computer-Mediated Communication: Linguistic,
Social and Cross-Cultural Perspectives. Pragmatics & Beyond: New Series 39.
Amsterdam/Philadelphia: John Benjamins.
Horowitz, M. W. & Berkowitz, A. (1964): "Structural Advantage
of the Mechanism of Spoken Expression as a Factor in Differences in Spoken and Written
Expression". Percept. Motor Skills. 19: 619-25.
Hård af Segerstad, Ylva: Forthcoming. "Swedish Teenagers
Written Conversation in Electronic Chat Environments". In: Penrod, Diane (ed.). WebTalk
- writing as conversation. Mahwah, NJ: Lawrence Erlbaum Associates.
Stenson, Per (1997): Skriva i tjänsten. Stockholm: Timbro.
Severinson-Eklundh, K. (1994): "Electronic Mail as a Medium for
Dialogue". In van Waes, L., Woudstra, E. and van den Hoven, P. (eds.), Functional
Communication Quality. Amsterdam/Atlanta: Rodopi Publishers.
(To the top)
© Ylva Hård af Segerstad 2000
to Human IT 2-3/2000