Homepage » Acting (on) the Text: the Case of New Media (6. 2014) » Alexander Ronzhyn and Eugenia Kuznetsova

Issue 6

[.pdf]

Conveying the Message of Trust through
Written Texts in CouchSurfing.org

  

Alexander Ronzhyn and Eugenia Kuznetsova

      Abstract 

The present article describes the results of research on online identity construction during participation in hospitality social networks. One type of user submitted data, references, was analysed to uncover and describe the way trust is conveyed in Couchsurfing.org. Through corpus-based linguistic analysis, authors explore the relative importance of different types of user interactions in the network and describe how references contribute to the framework of trust built within the Couchsurfing.org social network. Among the findings are the increased use of adjectives in references and the concepts used by the Couchsurfing management. Trust is seldom used in the references directly, but rather expressed through euphemisms and metaphor. 

During the last decade, online social network services (social networks or SNS) have become an important part of everyday life for millions of people worldwide. Academic interest promptly followed. To date, most research is done on Facebook and Twitter, the SNSs with the highest number of users. Still, there are other internet services and networks that have a member base significant enough not to be ignored by social researchers. Couchsurfing.com, the most popular hospitality exchange network, has more than 4 million registered members (Couchsurfing) and enjoys worldwide popularity. The present research concerns one aspect of members’ experience on this network—the language people use to describe their interactions with other users on the network. Particularly, it focuses on the language related to the concept of trust, probably the most important concept within Couchsurfing ethos, which allows the hospitality exchange to function. Given the importance of trustworthiness to the network’s users it would seem safe to assume that trust is one of the central topics in user references. In reality, one of the findings of the present research is that trust is seldom expressed directly and often conveyed through other means: extensive use of adjectives, description of common activities and values, as well as trust-related metaphors. In order to understand why this is happening one should look at how the hospitality SNSs function and what exactly is the role of the references within Couchsurfing. 

Online hospitality networks appeared as a continuation of offline hospitality exchange services (like Friendship Force International or Servas Open Doors) and are primarily used to allow people to arrange accommodation without any monetary exchange. Basically a member of the Couchsurfing network, a couchsurfer, who wants to travel may contact any other member and ask to stay at his or her place, whilst the receiving end may decline or accept the request. After this hosting-visiting arrangement has taken place members may write a review of the common experience, each describing the counterpart. Such review is called a “reference” in the network and constitutes an important part of a user’s profile. Reading references left by others is often an important factor when deciding to write a request for staying at one’s place as well as for accepting or declining such a request. 

This paper attempts to build up on the previous work by Ronzhyn (“Online identity: constructing interpersonal trust”), which attempts to paint a picture of a typical Northern Spanish Couchsurfing user from the point of view of trust, discussing what exactly are the properties needed to appear trustworthy, and how the online user’s profile is constructed to convey the message of trust. Other researchers of Couchsurfing.org SNS e.g. Tan (The Leap of Faith from Online to Offline) and Vaicekauskas (Generalized trust in CouchSurfing.org), each have attempted to analyse the connection between members of the network by trying to investigate and explain the trust arising between complete strangers. Trust may be defined as an expectation that the other party will behave in a dependable, ethical, and socially appropriate manner (Zucker). In the case of hospitality SNS we are dealing with generalised trust as trust exercised effectively towards strangers, people outside one’s in-group. Several papers (see for example Adamic, Lauterbach and Te) concentrate on the connection between different user profiles, rather than the content of these connections. More sophisticated research by Paula Bialski (Becoming Intimately Mobile) explores the mechanism, origin and implications of trust in Couchsurfing by undertaking a qualitative approach and basing the research primarily on interviews with network members. Bialski’s research revealed the importance of similarity at the initial stage of profile evaluation—that is, that users tended to trust and contact more often profiles that are similar. [1] This article aims to fill the gap in understanding the message contained in the references, and more specifically to explore what language do Couchsurfing members use to portray an individual (or rather an individual’s profile) as good (trustworthy, interesting, engaging) through linguistic analysis of the references within the hospitality exchange network. 

Couchsurfing SNS (or CS for short) has several types of articulated connections between users: friend connections, references and vouches. References are connections that are established between users who know each other and include a 3-point scale describing the overall impression of a person (positive, neutral or negative) selected through a drop-down menu, and the option to provide a short text explaining the evaluation given. References are not necessarily reciprocal (though they are meant to be) and are usually created after a hosting-surfing arrangement through Couchsurfing has taken place. References are the most interesting links between member profiles and the most suitable for the present research for several reasons: first, they are (usually) established right after a meeting between members; second, they include textual explanation that contains important information regarding the hosting-surfing arrangements; and third, unlike text notes accompanying friendship connection, references might not be only positive or laudatory. 

Before describing the results, it is worth noting that the references are constrained by two factors. First, any reference may be ‘answered’ with another reference that would be placed next to the first one not unlike a comment on other SNSs. This makes writing negative references undesirable even when they are justified out of fear to get a negative reference in response, affecting the attractiveness of one’s profile for couchsurfers in future. Often users would abstain from leaving a reference altogether instead of leaving a negative one or still leave a positive reference but be brief and avoid complimentary language, which is characteristic of truly positive references. 

Second, the uneven positions of a host and a guest are an important factor. A user who was hosted essentially got a place to stay for free and thus has to compare all possible negative sides of the host against his or her assistance, which can be easily measured in economic terms (price of stay in hotel or hostel). Conversely, a Couchsurfing host is the one providing the service and thus has less incentive to be polite out of gratitude when something goes wrong. The hosting side is also the one who sets up rules for the interaction by providing particular guidelines for the guests. Inadequacy or perceived unfairness of the house rules might be another source of tension between the guests and the host. All this makes finding out the true ratio of good to bad experiences on Couchsurfing a very difficult task. A possible resolution to this problem may be to completely avoid the discussion of the underlying motives in the references and instead concentrate on the language used. Politeness, fear of reciprocity and adherence to established network’s norms may all contribute to a user’s motives but ultimately it is the author of the reference who decides how to describe one’s experience highlighting particular interactions or details and choosing particular linguistic tools to convey (or not to convey) the trustworthiness of the referenced party. 

For our research we have taken a sample of 45 couchsurfers: 15 random profiles from the 3 European cities with the highest number of registered Couchsurfing.org members according to the official CS statistics. These cities are Paris, London and Istanbul. The samples include members of different age groups and gender (including profiles that are described as belonging to “several people”) with the single condition that a selected member profile should have at least 50 references. In total 10 references from each of the chosen Couchsurfing members were randomly selected and processed for this research, resulting in 450 references for 45 experienced Couchsurfing members across the network’s three most popular European cities. Most of the selected members were from the 25-35 age group; nevertheless, several chosen CS profiles did not have any age indication. Among the 45 selected member profiles, 33 were of male users, 9 female profiles and 3 of “several people.”

 

Fig. 1: European cities with the highest number of registered Couchsurfing users (august 2013).

The system’s request for writing a reference itself helps the members to evaluate their experience. Normally we do not immediately write what we think about a person we just met–yet leaving a reference requires doing exactly that. This is a surprisingly interesting experience for reference writers, and interesting material for research on the relations that are being established between the two parties, the host and the guest. 

The analysed references were presented in a form of the Couchsurfing References Corpus (CRC, target corpus, word count 34,595 words). It is formed by 450 total references on average around 77 words each. The references vary greatly in relation to the amount of text they contain: some references are no more than 6-10 words long, while others are several paragraphs long. The corpus consists of three clusters: the Paris cluster, the Istanbul cluster and the London cluster according to the cities to which the references belong. The three clusters are analysed both individually and as a whole to see the possible differences between the cities. The corpus is annotated and lemmatised, meaning that a part-of-speech tag is assigned to each lexical unit of a text. For the part-of-speech tagging we used C5 and C7 tag sets of CLAWS4 tagger. CLAWS (the Constituent Likelihood Automatic Word-tagging System) is a part-of-speech tagging system developed by the UCREL research centre at Lancaster University and used for high accuracy automatic tagging of large amounts of textual data. For the comparison corpus we used a considerably larger and more general British National Corpus, in particular its spoken section (Davies). 

Part-of-speech analysis 

Adjectives

The analysis of part-of-speech annotations shows that references are highly descriptive: from 34,595 words of corpus, 2517 were adjectives (54 of them were superlative adjectives: the most often combination was “best host”). To compare the percentage of adjectives in references, a spoken section of British National Corpus (BNC) was measured. From 10 million words in spoken section of the BNC, 226,727 words appeared to be adjectives, presenting approx. 2.2% of the section. Meanwhile the rate of adjectives use in the CRC is 7.27%, which allows us to conclude that adjective-use in references is more than three times more intensive than in usual language. The Parisian cluster shows even higher results of 9.2% of adjectives. The extensive use of adjectives is normally associated with emotional and descriptive language. Most of the adjectives used are assumed to reflect empathy: very warm and open person, polite, gracious, thoughtful, humorous, intelligent, delightful, easy-going, friendly, positive etc. Except for the use of empathetic adjectives, there is a large group of adjectives reflecting personal and mostly positive emotions (interested, thankful, pleased, happy etc.) and a group of descriptive adjectives: knowledgeable, interesting, delicious, different, fabulous. The most used adjective was the word “good” (143 uses, 5.68% of used adjectives). “Good” is also the most used adjective in British spoken English. In the spoken section of the British National Corpus, the rate of usage of the adjective “good” is very close to the CRC, being 6.83% of all adjectives. In the Parisian cluster, the use of word “good” is almost identical to that of BNC (6.9%). Nevertheless, the rate of the adjective “great,” which holds second position in both corpora, differs considerably: 7.31 % in Couchsurfing references corpus and only 1.35 % in the BNC spoken section–that is more than five times less. 

Thus we can conclude that the extent, types and connotations of adjectives used in the CRC illustrate a high level of emotional esteem in the references and descriptions of positive experiences. There are almost no adjectives assumed to reflect personal distress or describe things in a negative light. This is natural if we consider the fact that negative references are extremely rare, and in case of bad experience most of the people would leave no reference than write a negative one. The adjective choice for the portrayal of a person (host or guest) functions as proof to the fact that some personal communication took place between CS members. If an individual chooses a variety of adjectives to describe his or her experience of spending a short period of time with another individual, this can be considered as a sign of good personal communication between the parties. There are even references consisting of adjectives only: “Bright Enthusiastic Adaptable Nice Determined. This is Bertrand!” In this example all adjectives describe the personality of the host in a way that demonstrates that the CS members had common experiences. Except for two general adjectives (“bright” and “nice”) Bertrand is being characterised as adaptable (i.e. their communication involved a certain situation where ability to adapt easily was needed), enthusiastic (i.e. they did some activities together) and determined (i.e. they shared life values). Put differently, the surfers and hosts not just tell about activities they did together, but vividly describe and evaluate each other’s personality. The following table presents the most common adjectives in the cities sections of the corpus.

Table 1: Top-10 of the most common adjectives in CRC according to the cities

The most interesting words from the top-10 table are those that are more specific and not so extensively used in everyday language, but still appeared to be among the most popular adjectives in all three of the analysed cities. This might be a sign of certain Couchsurfing discourse where there is a specific ideal of a couchsurfer against which all users are compared. This idealised image is thought to possess the qualities described with these most popular adjectives. One of the words present in all three columns is “open.” In the reference texts, this lexical item is used as shown in the table below. 

Table 2: Contextual representation of the adjective “open” according to the cities

 

As we can see from the context of usage of the adjective “open,” which is one of the most common adjectives used in all three cities, the lexical item is mostly being used for describing the person’s ability to share experiences and readiness to acquire new experiences: open-minded for doing new things, open to try new things, open for everything new, open to all sort of interesting experiences, open to learn etc. Emphasising the individual’s ability to try new things conveys that a short act of communication between the two parties took place (like an evening talk), as well as their willingness to share activities and maintain communication in the future. The use of the openness as a positive description is likely to be predetermined by the lexicon used by the network managers. Openness is declared as one of the core qualities of a couchsurfer, emphasised in mission statement and promotional materials (Couchsurfing.org). Such adaptation of a promoted attribute might serve as a further evidence of existence of a certain standard of an ideal couchsurfer as a part of Couchsurfing ethos.  

Most of the adjectives from the top-10 table are still referring to the emotional evaluation of the experience that the members had together. The words “wonderful,” “amazing,” “great,” “real,” “best,” describe individual features of the CS members (wonderful host, amazing couchsurfer, his generosity is amazing, great sense of humour, great person, etc.) or an experience they had together (wonderful days, wonderful experience in Istanbul, amazing stories, great conversations, etc.). Put differently, CS members prefer stressing the features connected to interest, knowledge or sharing conversations and experiences rather than concentrating on generosity, kindness or direct help/assistance received from the hosts or guests. 

Verbs 

The usage of verbs in the past tense that usually describe activities and deeds is almost equal in the CRC corpus and the spoken section of the BNC, even still in CRC the percentage of verbs in the past tense is slightly lower: 2.43% compared to 2.86% in the BNC. However, the use of the past form of the verb “do” (i.e. did) differs more. In the spoken section of the BNC the word “did” is used 34,597 times (0.34% of the general word count) and in the CRC “did” is used only 55 times (0.15%). Interestingly the lemma “be” (i.e. all the forms of the lexical verb “be”) is used extensively in both corpora;  moreover, the percentage of the lemma “be” in the past tense in the CRC is even higher than in the BNC: 1.2% compared to 1.1% in the BNC. The difference can be attributed to the fact that the past form of the verb “be” is often used to describe feelings, mental states or characteristics being accompanied by an adjective. 

He was very welcoming and the conversation was very good. He was very interested in every possible word we could squeeze in such a short time!” (Mari, Paris). 

Here, the verb “was” is just another way of characterising people at the moment of communication that took place in the past. 

The top-10 of the most common verbs excluding lemmas “have,” “do,” and “be” are represented in the following table organised according to the cities to which the references belong. 

Table 3: Top-10 of the most common verbs in the CRC according to the cities

 All the verbs present their lemmas–i.e. all forms of the verb in the reference texts. As we can see most of the verbs are common for all three columns. The most popular verb in all cities is “meet”: it holds the first position in the three columns. This is quite symbolic as Coushsurfing is considered to be a meeting place. Couchsurfing management explicitly underlines that meeting new people and communication are the main feats of CS travelling (Couchsurfing.org), while using the system as a way to minimise one’s expenses is decidedly against the network’s ethos. There is even a special member status of “coffee or drink” (when a person cannot or does not want to host) that supports the idea of meeting people for sharing and exchanging even without actually “surfing a couch.” 

“Burcu was our first CS experience in Istanbul, actually we were staying in a hotel and we wanted to share the local culture with someone who is experienced” (Claire, Istanbul)

Analysing direct quotations from references, it appears that among the most common verbs used many embrace the process of sharing values and exchanging experiences. For example, the word “know” that is in the top-10 for all three cities is usually used in general utterances like “glad to know him/her,” “nice to know you,” “he knows that he is always welcome” etc. However there are references involving the verb “know” that characterise the individual’s openness to other members:

“He is fan of talking and discussing different issues as well as listening to your ideas and opinions! Talking to Ozcan you do notice that he is completely involved in what you are trying to explain, he does not miss a chance to know you better even if you spent a little time together” (John, Istanbul) 

Among other verbs that are not in the top-10 but still serve as evidence of sharing values and experiences are “discuss,” “learn,” “teach”:

chat flows all night after dinner we discussed about many things, personal and general stuff” (Marta, Paris)
“she taught a lot of interesting things that I have never heard before” (John, Paris)
“there is something for everyone to learn from his guy” (Alex, Paris)
“They are an amazing friendly and welcoming couple who love learning other cultures” (Jerome, Paris)
“he taught me a bit about what Brazil is really like” (Patrick, London)
“we had good time with lots of discussions about traveling, CS and our life” (Tim, London)
“It was great meeting you, we had very interesting discussions after my visits in Paris. Very open and curious about other people and other cultures.” (Alex, Paris)
“She is an amazing girl who knows how to enjoy the life” (Michelle, Istanbul)
“We could talk about anything, surprised while discovering our similarities and had quite a lot of fun discussing our differences” (Bora, Istanbul)
“We could discuss with him in many different topics and also we learnt many things about turkish culture from him” (Mona, Istanbul) 

These quotations are only a few examples where the process of learning, teaching or discussing other cultures is directly described. The CS members in their references value the desire of learning or the ability to teach. The verb “talk” is also very popular in the three city clusters. Thus the value of conversation, discussion and communication in general is being emphasised in the analysed references. 

A verb that is not in the top-10 yet comes very close (22 uses in London, 14 in Istanbul and 12 in Paris) is the verb “share.” This has to be analysed separately. The act of sharing is a basic element in Couchsurfing philosophy and values: not only sharing place, food and shelter, but sharing something more. Trying to answer the question what do couchsurfers share, we have gathered the list of answers from the CRC.

Table 4: What do Couchsurfing members share?

The table illustrates that only a few of the things shared refer to some material values like a place, meal or drink. Most of the nouns following the verb “share” refer to life, values, thoughts, communication, advice and even secrets and “common ideologies of ways of making the world a better place to live.” Once again the users’ discourse mirrors that of the official Couchsurfing that emphasises sharing as one of the main activities on the network.

Nouns  

The percentage of nouns in the spoken section of the BNC is 5.1% while in the CRC is 13%. The difference can be primarily explained by the fact that references are normally very short texts and nouns form the basis for any text. Therefore if a reference is a unit of analysis it inevitably involves a number of nouns that are informative, while longer texts and particularly spoken language are normally less informative and involve less practical information that requires the use of nouns. The most used words both in the spoken section of the BNC and the CRC include some of the most common nouns, e.g. “people,” “time” and “thing.”  

Table 5: Top-10 of the most common nouns in CRC according to the cities

Similar to the use of verbs, the most common nouns of the references across all three cities are almost identical, differing slightly in their number of uses. The most illustrative words referring to the act of communication between the CS members are “friend” (see the analysis of the usage of the noun “friend” in the Trust section), “person,” “interest” and “experience.” Analysing the context in which these nouns are used supports the observation that these words stand as proof of close communication and the sharing of values and attitudes between the guest and the host. 

Table 6: Contextual examples of the most common nouns in CRC

Other nouns, while not being in the top-10 list, are quite illustrative of the sharing of ideas and opinions in CS. For instance, there are some direct messages about conveying knowledge on different subjects: 

“Alp is a funny and full of knowledge guy always willing to chat about everything since Turkish history and politics until philosophy or sociology” (John, Istanbul) 
“Theo has a wealth of knowledge, a contagiously spontaneous and practical approach to travelling and yet has an innate capacity to enjoy the moment and the people he's with” (Loraine, Istanbul) 

Equally, in a reference from the London cluster, the guest defined his host through a group of nouns, where one of the definitions is actually knowledge: 

“Grant=hospitality /Grant=generosity / Grant=source of knowledge / Grant=good food / Grant=best advice / Grant=London“ (Erika, London) 

This definition of Grant represents the entire spectrum of possible interactions between guest and host expressed through noun-noun utterance. Even so, nouns are not so informative about the attitudes and shared activities as adjectives and verbs are. 

Trust 

A word that has to be analysed separately is the word “trust.” “Trust” can be used as verb and noun; we also include the adjective “trustworthy” as a part of the lemma. Somewhat surprisingly only a few members directly write that they trust their host or guest. We would suggest that the reason is the possible negative connotation of the trust issue in general. That is to say, if I emphasise that I trust someone after meeting him or her in person, it can be interpreted that I started trusting directly after a meeting whilst before it was doubtful that that person could be trusted. Thus there is a certain tension between emphasising the trust and a policy of mutual trustworthiness in the community. Couchsurfing both accentuates the security of interactions within the system in general and provides justification for trustworthiness of individual members through the reference and vouching systems. A person who has a number of positive references, vouches and has hosted or surfed many times is deemed trustworthy. The usage of the lemma “trust” in the references can be classified into three groups: references where it is directly recommended to trust the CS member, references that use the adjective “trustworthy” listing the features of the host or guest, and references involving trust when describing one’s own feelings. 

Table 7: The usage of lemma “trust” in CRC

 

One of the Parisian users wrote in one of his references: "This about writing references feels a bit weird. Like putting a grade on someone, how good or bad people are... But I will do it again anyway as a habit it becomes” (Mari, Paris). This attitude has something in common with expressing a direct message of trust. Thus most users prefer conveying the message of trust through other linguistic means, other utterances and phraseological units. Describing shared activities or feelings, the message of trust is being conveyed to potential readers and an image of a trustworthy person is built with every positive reference without direct referring to the lemma “trust.” The most common form of conveying trust in a reference is to describe one’s own feelings about the host or guest in a way that would clearly illustrate that a user is trustworthy. Among the most popular ways of conveying the message of trust is the use of metaphors, such as the metaphor of a familiar person or the metaphor of home. 

Table 8: Metaphors in references

 

These examples illustrate that CS users, willing to express their trust, unintentionally avoid the lemma “trust” and prefer to construct more complex utterances in order to convey the message “you can trust him or her.” Interestingly, some examples of such euphemisms contain a clear indication that a single meeting was reason enough to start considering a person as a close friend or even a family member. Such expression of trust is evidence of deep and emotional communication between the host and the guest. This goes in line with the concept of “intimate mobility” introduced by Paula Bialski (Becoming Intimately Mobile), who found out that conversations between complete strangers when travelling (especially through Couchsurfing or hitchhiking) are often more open and intimate than those with close friends. Hosting is not only having a couch available in a living room but a communicative space where personal values and attitudes are shared. 

Another point worth mentioning is the honesty of the references. The reported ratio of negative to positive references in Couchsurfing is 1:2500 (Teng, Lauterbach and Adamic). Yet it is relatively safe to assume that the vast majority of users are honest when describing their positive experiences with other people. The low number of negative references is attributed to the fact that people decline to leave any reference in case of negative experience rather than them leaving a positive reference in place of a negative one. Even when a user decides to leave a positive reference, when neutral or negative would be appropriate, he or she may use more descriptive language (less adjectives, superlatives) or concentrate on facts rather than experiences, both avoiding to lie about the interaction and at the same time providing possible cues for other couchsurfers who consider this host or guest. Furthermore as references are not a requirement but rather an act of good will, one can easily refrain from leaving a reference without any consequences. With all these considerations in mind it seems doubtful that any significant number of couchsurfers would lie in the references left for other users. 

Future research 

This is the first research done on the content of members’ interactions in the Couchsurfing.org network. Even so, the present work does not attempt to offer a definitive assessment of on the language of Couchsurfing SNS. Indeed, there is lack of distinction regarding the source of the references. While the recipient of the reference is clearly defined and attributed, respondents who create the reference were not clearly systematised for this study. Thus future research may look into the reference content differences in relation to the different characteristics of the reference-givers: do references left by women are significantly different from those left by men? Is the age a defining attribute when it comes to the overall tone of the reference? and so on. Equally interesting are the possible applications of the research findings to other Internet reference and rating systems: whether the patterns (extensive use of adjectives and verbs, use of metaphors) found in the Couchsurfing references can also be observed in references and reviews left for services, products or establishments. Notwithstanding, the present paper is an important step towards measuring and describing how trust is conveyed and built through the interactions between members of an online community. Likewise, it offers initial ideas on how the question of the content of online interaction can be tackled. 

Conclusions 

The linguistic analysis of the corpus containing 450 sampled references left on the Couchsurfing.org website by hosts and surfers has uncovered several patterns that relate to the content and the implications of the references within the network. The analysis of part-of-speech annotations has shown that references are highly descriptive with very frequent use of adjectives (threefold compared to the number of adjectives contained in the British National Corpus). The use of such a variety of adjectives to describe the experience of spending a short period of time with another individual can be considered as a sign of good personal communication between people. Furthermore, the majority of the most commonly found adjectives refers to the emotional evaluation of the experience the members had together. Analysing direct quotations from references, a significant percentage of the most common verbs used are connected to the process of sharing values and exchanging experiences. Members tend to describe sharing values as being the cornerstone of the offline activity at Couchsurfing and entirely favourable. There is also evidence of a certain degree of normalisation of language, when users tend to describe others by employing language that is common on the network. The discourse on sharing, and the frequent use of the adjective “open” and its derivatives can be the evidence of users’ adopting the language used by the official Couchsurfing. 

Trust as a crucial issue at the Couchsurfing.org network is an important theme of the references. Still it is seldom expressed directly as that may be regarded as an indication of the opposite–the lack of trust. Among the most popular ways to address the question of trust in the references are euphemisms and metaphors. The latter are usually of a familiar person or the metaphors of home, when either the referenced person is described in a way that assumes a high level of trust, or the house of a referenced host is compared to or described as if similar to one’s own home. The direct use of “trust” can be found in references where it is directly recommended to trust the CS member; in references that use the adjective “trustworthy,” listing the features of CS member; or, in references involving trust in the description of one’s own feelings. 

Comparative analysis of the Couchsurfing Reference Corpus and more general British National Corpus showed both similarities (the use of verbs and some lemmas) and apparent differences, which stem from the purpose and the nature of the texts. The highly descriptive and emotional text of the Couchsurfing references highlights the personal and even intimate mode of communication between the users of the Couchsurfing.org social network. Lexicons of sharing ideas, experiences and activities as well as the question of trust occupy an important place in the texts, emphasising also the practical purpose of the references as an indicator of the trustworthiness of a referenced person. 

 

[1] For more explanation on the topic of members’  homophily and the use of similarity as a criterion for connection between network users/members see Bialski and Batorski.

  

Works Cited

Adamic, Lada A. , Debra Lauterbach and Chun-Yuen Te. “Rating Friends Without Making Enemies.” 5 July 2011. AAAI Publications, Fifth International AAAI Conference on Weblogs and Social Media. 27 March 2012. <http://www.aaai.org/ocs/index.php/ICWSM/ICWSM11/paper/view/2837/3280>. 

Bialski, Paula and Dominik Batorski. “From Online Familiarity to Offline Trust: How a Virtual Community Creates Familiarity and Trust between Strangers.” Zaphiris, Panayiotis and Chee Siang Ang. Social Computing and Virtual Communities. Boca Raton, USA: Taylor and Francis Group, 2010. 179-205. 

Bialski, Paula. Becoming Intimately Mobile. Vol. 2. Warsaw: Warsaw Studies in Culture and Society, 2012. 

CouchSurfing. CouchSurfing - Statistics. 2013. 4 May 2013. <http://www.couchsurfing.org/statistics.html>. 

Couchsurfing.org.  Couchsurfing Values. August 2013. <https://www.couchsurfing.org/n/values>. 

Davies, Mark. BYU-BNC (Based on the British National Corpus from Oxford University Press). 2004. 2 September 2013. <http://corpus.byu.edu/bnc/>. 

Ronzhyn, Alexander. “Online identity: constructing interpersonal trust and openness through partcipating in hospitality social networks.” Journal of Education Culture and Society 1_2013 (2013): 47-57. 

Tan, Jun-E. “The Leap of Faith from Online to Offline: An Exploratory Study of Couchsurfing.org” Lecture Notes in Computer Science (2010): 367-380. 

Teng, Chun-Yuen, Debra Lauterbach and Lada A. Adamic. “I rate you. You rate me. Should we do so publicly?” WOSN'10 Proceedings of the 3rd conference on Online social networks. Boston, USA: USENIX Association Berkeley (2010): 1-12. 

Vaicekauskas, Kristijonas Juozas. Generalized trust in CouchSurfing.org. Comparison of Nordic and Baltic countries. Uppsala: Uppsala Universitet, 2010. 

Zucker, Lynne. “Production of Trust. Institutional Sources of Economic Structure” (eds), B. Stow and L.Cummings. Research in Organizational behaviour. Vol. 8. London: JAI Press (1986): 53-111.