The Impact of Brexit on Tourism as Released by the Press: A Corpus-Driven Lexical Analysis

The tourism and hospitality industry is one of the sectors that could be significantly affected by the exit of the United Kingdom from the European Union. The potential consequences of the separation started to be echoed by the mass media as soon as the Leave result had been released. In this study, a corpus of online news counting 201,108 tokens and covering the immediate reactions to the Brexit referendum in the tourism sector is analysed using WordSmith tools with the aim to explore the lexical choice of the press to report on the impact of Brexit on tourism. Although the results of the analysis present a temporary overview of the situation, corresponding to the time span covered (from the immediate aftermath of the referendum to the beginning of 2018), both negative wordforms and their extended context evidence the negative consequences forecast for the economy in general, whereas the opposite effect is found for the tourism sector. In addition, there are instances of newly coined words such as Brexit, Brexiteer and Brexodus, and the top keywords manage to provide a clear account of the topic that answers the 5 Ws characterising the news genre. All in all, a pervasive sense of uncertainty and concern about the future after Brexit is perceived all along the samples.


INTRODUCTION
A critical referendum for Europe was conducted on 23 June 2016 when the United Kingdom voted on its permanence or withdrawal from the European Union. The subsequent Leave result marked the onset of Brexit and a set of negotiations between the UK and the EU about the terms of the separation. Brexit may have farreaching consequences for both parties at all levels, but the benefits that the UK used to enjoy precisely as a member of the EU may all be jeopardised. The significant impact and direct consequences ensuing from the breakage are envisioned as a succession of potential scenarios and corresponding actions for the future (MacDonald, 2016).
Research on this phenomenon has been performed from multiple perspectives. Among them, early political studies attempted to identify the reasons leading to the Leave result and the factors that originated public support for Brexit (Andreouli & Nicholson, 2018;Goodwin & Heath, 2016); social studies shared a similar goal but sharpened their focus of analysis on the digital media and the public tweets posted on the event (Howard & Kollanyi, 2016;Maynard, Roberts, Greenwood, Rout, & Bontcheva, 2017). Subsequently, attention was bestowed upon the influence of the media on the Leave result and how such bias was verbalised by the press (Seaton, 2016). The event has also been analysed from a linguistic perspective approaching different means of communication on the subject such as social media platforms (Griebel & Heinrich, 2017), news articles (Ballmann, 2017) or both (Lalić-Krstin & Silaški, 2018). One of the earliest edited volumes in the field is presented as "the first comprehensive exploration of discourses surrounding the UK's departure from the EU and as such step towards understanding the reasons for, and processes of, Brexit" (Koller, Kopf, & Miglbauer, 2019: 1).
Although the definite repercussions of the separation cannot be fully anticipated, after the referendum a sense of uncertainty was felt in all areas of social, political and economic spheres. As far as tourism is concerned, there seemed to be a mood for pessimism as negative signs were expected in the sector, due to the fact that the different agreements between the UK and the EU would be annulled or, depending on the negotiations, undergo important modifications (Calder, 2016;Rhodes & Ward, 2016).
The present study aims to contribute with insight into the press discourse reporting about the potential consequences of the withdrawal of the United Kingdom from the European Union on one of the areas that would be significantly affected, namely, the tourism and hospitality industry (Rhodes & Ward, 2016). Adopting a corpus-driven approach, the study seeks to identify how the online press stories echoe the impact and the foreseeable consequences of Brexit on tourism. For this purpose, a linguistic corpus of online press news was designed and compiled ad hoc in order to point out the lexical preferences of the press when reporting on the issue.

THE IMPACT OF BREXIT ON TOURISM AS RELEASED BY THE PRESS: A CORPUS-DRIVEN LEXICAL ANALYSIS
Vol. 8(1)(2020):  In order to guide the development of the present study, two research questions were formulated: Q1: What lexical items are chosen by the press to convey the expected impact of Brexit on tourism?
Q2: What further lexical information is revealed by the syntagmatic relationships generated by the nodes of analysis, Brexit and tourism?
The results of the analysis will mirror the initial overview of the situation rendered by the language samples compiled in the corpus, covering the immediate aftermath of the referendum, from late June 2016 to the beginning of 2018.
The structure of the paper is as follows: after this introduction, the following section offers a review of literature pertaining to Brexit-related discourse analyses. Next, we describe the corpus gathered for the study and define the methodological approach adopted for the analysis of the samples. Then, the results are discussed with respect to the parameters of frequency and keyness, followed by an account of the findings obtained in terms of the syntagmatic relationships established by the particular nodes of analysis, Brexit and tourism, to finish with the main conclusions drawn from the study.

DISCOURSE ANALYSIS AND BREXIT: A LITERATURE REVIEW
A considerable amount of literature has been published on the discourse on Brexit. The call for a referendum marked a milestone in the history of both the UK and the EU, which aroused a surge of interest in how media discourse was shaping the perception of such reality. The Brexit discourse has been a productive topic of research in the field of critical discourse analysis where discursive articulations of visions of social reality and ideological foundation interwoven with the concept of Brexit have been explored (Buckledee, 2018;Lavery, 2019;Thommessen, 2017;Zappettini & Krzyżanowski, 2019). Political ideologies as well as sets of values and attitudes in the discourse of Leave and Remain campaigns have been examined by researchers, providing evidence of controversial topics associated to each party and the political repercussions leading to the final vote (Bennett, 2019;Buckledee, 2018;Miglbauer & Koller, 2019;Zappettini, 2019). Public opinion has also been under analysis through vox pops in the media illustrating the cultural distance between Leave and Remain voters and their position towards the recurring topics underlying the event such as migration, nationalism, economy, crisis, globalisation, among others (Miglbauer & Koller, 2019;Ruzza & Pejovic, 2019;Tolson, 2019). Similarly, critical discourse analyses of the press have pointed out how the European question is re-narrated within the Brexit context by the British press and the expression of English populism (Maccaferri, 2019), and how Europe acknowledges the event of Brexit as a transnational crisis (Ballmann, 2017;Krzyżanowski, 2019).
However, the use of corpora to research into the form and/or function of language as a communicative discourse introduces quantitative features as a crucial factor of the analysis (Partington, Duguid, & Taylor, 2013). Thus, corpusassisted discourse studies (CADS) combine the so-called quantitative approach with the more qualitative approach typical of discourse analysis (Partington, 2010). In this vein, the use of Brexit as a lexical item in the press discourse after the referendum has been analysed in terms of semantic prosody. The elevated number of collocates with a negative meaning revealed the negative connotations associated to the word (Broz, 2019). In turn, Alkhammash (2020) delved into the press discourse by comparing how the EU was viewed in the Leave and Remain campaigns. The author observed a common negative image of the EU and a populist ideology contributing to a eurosceptic attitude in the press. In addition, research has also focused on the discourse of Brexit in the press in order to trace the coinage and development of the Brexit term from its first use (recorded in May 2012) to the UK general election in 2015 (Fontaine, 2017). The myriad of neologisms generated by its linguistic impact has also been analysed (Lalić-Krstin & Silaški, 2018), as well as the use of the term and its different wordforms in the press before and after the referendum (Lutzky & Kehoe, 2019).
Nevertheless, to the best of our knowledge, little attention seems to have been paid in previous studies to the press discourse dealing with the practical consequences of Brexit, beyond the realms of the ideology underlying the report of events. Therefore, this research is intended to shift the focus of analysis onto the expression of the after-effects of Brexit on the tourism sector.
In the following section, we will describe the corpus gathered for the study and outline the methodological approach applied to analyse the samples.

DATA AND METHOD
A specific linguistic corpus was designed and compiled for the purpose of the study following the principles of Corpus Linguistics (Rea, 2010;Sinclair, 1991Sinclair, , 1996 and defined in terms of topic, mode, chronology, type of text or genre, variety, origin and size. As the language target is the press discourse about Brexit and tourism, the so-called News Corpus comprises digital press news stories dealing with the relationship of both topics in English, addressed to the general readership. The written pieces of news were released from 24th June 2016 to 23rd February 2018 by different authors writing for British and non-British newspapers published in English, under open access terms. 1 Only complete samples were collected reaching 1 The samples were collected in accordance with the BOAI definition of open access, that is, they were freely available on the public internet.

THE IMPACT OF BREXIT ON TOURISM AS RELEASED BY THE PRESS:
A CORPUS-DRIVEN LEXICAL ANALYSIS Vol. 8(1)(2020): 47-67 a median value of 620 words, ranging from 285 to 2,040 words. All texts were saved as .txt files for processing purposes. The final size of the corpus came to 201,108 tokens distributed throughout 309 texts.
Although the final number of texts that make up the corpus and the total of words gathered might seem small, the accessibility and availability of the samples in such a restricted area, that is, tourism and Brexit related news stories, would justify its size. On the one hand, the referendum and the subsequent Brexit process were recent events and had not finished yet at the time of the compilation process -a first deadline was set on 29th March 2019, later extended until 31st January 2020. On the other hand, it is worth noting the scarcity of open access to the different web platforms available. For example, in the UK, The Guardian is one of the few main British newspapers with unlimited access to content, unlike The Times. In contrast, other newspapers such as The Irish Times allow limited access online to non-paying users. Moreover, non-European newspapers report in their own language and offer limited sections in English, if any. All in all, the corpus covers three geographical sectors as depicted in Figure 1: (i) English from the United Kingdom; (ii) non-European English from the USA, Australia, Canada, China, Mexico, New Zealand and South Africa; and (iii) European English from Cyprus, Germany, France, Spain, Iceland, Ireland, Russia and Belgium. The methodological approach adopted in this study allows to examine the language on both a quantitative and a qualitative basis through the tools available for such aim. WordSmith Tools software (Scott, 2008) is the program selected for extracting corpus data and observing how words behave in texts. The immediate processing of the corpus yields two kinds of relevant data: basic statistical information (tokens, types, type/token ratio, etc.) and frequency list. Frequency analyses have revealed that the most frequent words are functional and tend to keep a stable distribution. Therefore, any noticeable change in the top ranking or the intrusion of a content word in this level may show a remarkable lexical behaviour, since around the top 100 most frequent words in a general corpus are functional (Schmitt, 2000;Sinclair, 1991). The unusual statistical behaviour of a word encourages its co-text analysis in order to describe the syntagmatic relations established, as shown by the word clusters it may generate and the collocates it may attract. In this paper, collocates are analysed under the concept of downward collocation (Sinclair, 1991) owing to its strength in statistical terms, that is, collocates are lower in frequency than the search word or node, which enables a semantic analysis, in contrast to upward collocation. Sinclair (1991: 116) maintains that there is a systematic difference between these types of collocation: "Upward collocation is the weaker pattern in statistical terms, and the words tend to be elements of grammatical frames, or superordinates. Downward collocation by contrast gives us a semantic analysis of a word". Therefore, only the collocates which are less frequent than the search word are considered for the purpose of the research.
In addition, any lexical behaviour statistically deviating from the general norm is significant both for the unusually higher or lower frequency of lexical items. Unusual patterns of frequency are identified by the Keyword tool in WordSmith. Scott (1997) argues that positive keywords (words whose frequency is unusually high) provide a good indication of the text's aboutness. In order to execute the Keyword tool to retrieve the most statistically significant lexical items in the news, we will draw on the general language corpus LACELL 2 as the reference corpus which establishes the norm and has been used in previous studies for research purposes (Marín, 2014;Marín & Rea, 2017, among others). The present study focuses on the positive keywords since they provide a good account of the subject content and are therefore assumed to reflect the particular lexical choice of the press to inform about the repercussions of Brexit on tourism.

THE LEXICAL CHOICE OF THE PRESS: RESULTS AND DISCUSSION
In the next subsection, the corpus is analysed in terms of frequency and keyness in order to identify the main lexical items deployed by the press to convey the expected impact of Brexit on tourism, focusing, therefore, on responding to our first research question.

Frequency and keyness
The general overview of the composition of both corpora is provided by the basic statistical information and frequency list after their first processing on WordSmith. Table 1 shows the statistical data yielded, whereas the figures indicating the frequency of occurrence can be observed together with the keyness values in the Appendix. In spite of the substantial difference in size, both type/token ratio and standardised type/token ratio indicate that the News Corpus attains a considerable lexical variety with respect to its size. In addition, lexical diversity is also determined by hapaxes, that is, the word forms that occur only once or twice in the corpus. Out of the 9,183 types in the News Corpus, there are 3,094 hapax legomena and 1,494 hapax dislegomena, which correspond to 33.6% and 16.2% of the corpus respectively. Regarding the expected word frequency ranking in relation to the steady distribution of functional words among the top 100, it is remarkable how the following content words appear within the 50 most frequent words in the News Corpus -frequency index is shown between brackets: UK (1770), Brexit (1197) (147), long (147), foreign (146), and take (146). Nevertheless, frequency index per se does not reveal the extent to which a lexical item is relevant in a particular domain. Indeed, word frequency is extremely important since it allows the quantification of the language and therefore statistical inferences such as the application of the log likelihood test to retrieve the keywords in a given corpus.
In the present study, positive keywords (Scott, 1997) are equated with the precise words selected by the press to cover the news in relation to tourism and Brexit, since such keywords are statistically singled out as to occur unusually more frequently than in the large general language corpus that establishes the norm. The top 100 keywords sorted by keyness index out of the overall 500 positive keywords generated by the program are shown in the Appendix. The frequency 53 CAMINO REA RIZZO Vol. 8(1)(2020): 47-67 count of the keywords in both corpora is given next to the keyness index. The results show that the top keyword is UK with a keyness (k) value of 11,840, leaving (k:104) is found in the middle, and the lowest keyword is significantly (k:42). It is noticeable how the word Brexit is ranking as the second most relevant keyword, whose index reaches 11,661.
A first observation of the 100-keyword list seems to provide the answers to the well-known questions of the 5 Ws formula -Who, What, Where, When and Why -the questions that a news story must cover in order to tell the readership in the most straightforward and shortest possible way what they need to know. Moreover, the top 5 keywords whose index is >5,000 succeed in summarising the news content to the utmost: UK, Brexit, EU, tourism, and travel. First of all, in order to answer the question "who are the subjects or doers of the event?", the UK (k:11,840) and the EU (k:10,033) are brought to the fore as the two main parties involved in the situation. Between them, the main news event answers what happened: Brexit (k:11,661), followed by two more central topics: tourism (k:9,535) and travel (k:5,098). Subsequently, a combination of topics and agents concerned with the issue alternate in relevance, such as pound (k:3,404), visitors (k:3,319), tourists (k:2,555), British (k:1,612), industry (k:1,558), and travellers (k:1,486). Next, the triggering event for the news occurs as one of the answers to the why question: referendum (k:1,328) as well as vote (k:1,227). Along with them, impact (k:1,276) emerges in this high relevance level informing about the powerful effect that the referendum may have.
As regards the rest of Ws, the keywords following the top 10 seem to disclose where the events occurred -leaving aside UK and EU mentioned previously. On the one hand, Brexit is related to toponyms like Britain (k:1,615), Europe (k:1,146), London (k:597), Ireland (k:350), Gibraltar (k:341), Scotland (k:319), Spain (k:314), and Schengen (k:234). On the other hand, there are keywords referring to places like overseas (k:1,140), destination (k:1,063), countries (k:664), abroad (k:648), border (k:321), and airports (k:260). Finally, the keyword which denotes a particular temporal reference to a key date is June (k:347), the month when the referendum was held.
Next, the lexicon singled out by the press is further examined by considering any outstanding numerical behaviour or meaningful lexical relationship in the results from the processing of the corpus which may help to describe such lexical choice. The fact that a keyword does not occur in the reference corpus may evidence the emergence of a new word. Not surprisingly, the third most frequent and the most relevant word (Brexit) in the news has 0 frequency in the reference corpus, since Brexit is a new blend coined to convey the exit of the UK from the EU; hence, it was not previously registered in the general corpus. Even in the early stage of the process where the language samples take place, there are records of a word form derived from the blend: Brexiteers (f:6), as well as the blend Brexodus (f:1) (see concordances 1-2 ). Nevertheless, Brexit is not attested in its plural form in our corpus unlike in Lutzky and Kehoe (2019) where the plural form is found to Another case of blending not found in the reference corpus is observed in the keyword staycation (k:418) and its derived form staycationers (k:48) (see concordances 3-4). This is a blend of stay and vacation, which refers to a holiday that someone spends near home or in their own country rather than travelling abroad. The term is associated to periods of financial crisis when overseas holidays become significantly more expensive. In this context, the case of VisitBritain (k:604) is also found. This is a tourist agency which tries to forecast the consequences of Brexit on the sector, among them on the journeys made for visiting friends and relatives also referred to as VFR (k:93) -with null frequency in the reference corpus. Even the big data company ForwardKeys (k:146) is registered in our corpus for the first time, performing a similar function to VisitBritain but on a large scale and on a different basis, as the company which analyses worldwide tourist bookings to predict trends and estimate their economic impact.
A second remarkable characteristic of the keywords obtained allows their taxonomy into three groups depending on the domain or field of activity they relate to, namely, politics, economy and tourism. A general negative sense is perceived in the economy group as evidenced by both negative word forms like uncertainty (k:599), weaker (k:429), fall (k:359), drop (k:143), volatility (k:140), decline (k:122), slump (k:104), affect (k:243), devaluation (k:212) and plummeted (k:100), and the negative meaning denoted by the context of use of other instances such as currency (k:661), fares (k:249), consumer (k:199), impacted (k:150), pound (k:3,404), among others, as evidenced in concordances 5-7. This lexical choice seems to confirm the initial pessimistic forecasts for the economy based on the outbreak of Brexit-induced uncertainty. 5 The drop suggests that the economic uncertainties around Brexit are not only impacting overseas businesses but domestic businesses also. 6 The sudden 13 per cent currency depreciation triggered by the Leave vote left sterling at its lowest level against the dollar for more than 30 years at one stage. 7 The pound fell because investors think leaving the EU's massive market will damage the U.K. economy.

55
CAMINO REA RIZZO Vol. 8(1)(2020):  In contrast to the economy domain, a more positive overview is detected in the tourism sector. In fact, lexical items like positive (k:171) and attractive (k:96) become prominent keywords in tourism contexts. Additionally, the salutary effect that Brexit has exerted on tourism contrary to the initial expectations is attested by the positive sense conferred to the contexts where keywords such as destination (k:1,063), bookings (k:590), tourism (k:9,535), inbound (k:993), visits (k:656), etc. are used and shown in concordances 8-11. 8 While London will continue to be a destination hotspot, a rise in domestic travel should also benefit hotels outside the Capital as UK residents explore other cities such as Manchester, Birmingham and Edinburgh. 9 After Brexit, we instantly noticed an increase in bookings and with many more people traveling from around the UK. 10 Tourism has been a winner from the 2016 Brexit vote which pushed down the value of the pound, making Britain a more attractive holiday destination for foreign tourists and encouraging British holidaymakers to stay at home. 11 It's now confirmed that Brexit had an immediate, positive impact on inbound tourism to the UK, which is converting into better-than-anticipated arrivals.
Within the tourism group, it is noticeable how some acronyms reach high keyness indexes. Their corresponding extended forms convey a sense of the sort of institutions and associations related to Brexit or affected by the process to a certain extent (see concordances 12-13). The most relevant acronym is ABTA (k:735), the Association of British Travel Agents, followed by ETIAS (k:331), the European Travel Information Authorisation System; ATTA (k:158), the Adventure Travel Trade Association; and ONS (k:142), the Office for National Statistics. Likewise, other acronyms in the field achieve keyword status characterised by their null occurrence in the reference corpus: ISHC (k:155), International Society of Hospitality Consultants; BHA (k:77), British Hospitality Association; APD (k:68), air passenger duty; and WTM (k:68), World Travel Market.
12 ABTA shared eight advantages of the EU for travel in the months leading up to the referendum, and now that Brexit has been announced, there may be some implications. 13 With the vote for Brexit now a reality, ISHC have put together a panel of experts to give their opinions on how Brexit will affect the hospitality and tourism industry.
Finally, it is worth mentioning the appearance of names and surnames as keywords, corresponding mainly to the politicians who played a substantial role in the process, that is, May (k:57) and Trump (k:53) (see concordances 14-15). At the time of the events recorded, Theresa May was the leader of the Conservative Party and the Prime Minister of the UK whereas Donald Trump, who had campaigned for Brexit, was the President of the USA. Similarly, the names of active journalists reporting the news about the phenomenon stand out particularly due to the nature of the corpus; that is the case with the US National Public Radio's correspondent in London Frank Langfitt (k:165), the most relevant surname. In the subsection to follow, the lexical items Brexit and tourism will be set as nodes for further analysis in terms of the syntagmatic relationships they establish, in order to examine closely the second research question formulated.

Syntagmatic relationship: collocates and clusters
The volume of data retrieved with respect to the collocates generated by the nodes Brexit (380 collocates) and tourism (503 collocates) is downsampled by resorting to the right and left immediate content-word collocates whose co-occurrence frequency is higher than 5 or, in other words, by resorting to the two-word clusters generated by the nodes whose frequency is higher than 5. In the second stage, the collocation span is widened to 5 words to the left and 5 words to the right of the node, aiming at spotting the top frequent collocates within the determined environment which may present a closer overview of the lexis associated to Brexit and tourism.
Despite the importance of downward collocation for revealing semantic traits, the results urge us to pay attention to some functional words, in particular to the remarkable co-occurrence of the prefix post (f:136), and the prepositions after (f:90) and following (f:14) preceding Brexit. They tend to be used in contexts confronting the different issues raised by the alteration of the agreements and rights of the UK as an EU member regarding tourism. Such contexts mostly display a pervasive sense of uncertainty and an overriding concern over inbound and outbound traveller's future as illustrated by concordances 16-18. This idea of the future after Brexit is reinforced by the addition of the prefix post to Brexit, which becomes an adjective in attributive position qualifying the following nouns : agreement,announcement,anxiety,appeal,arrivals,boost,border controls,bounce,Britain,budget,currency,England,Europe,fall,fund,government,immigration,landscape,Minister,pound,regime,sense,shock,situation,summer,tourism,travel,UK,vote,and world. 16 Tourism to suffer under Brexit border controls? POST-Brexit border controls which could potentially see EU nationals required to produce their passports rather than national ID cards when entering Jersey could damage tourism, an industry expert has said. 17 Britons travelling to continental Europe after Brexit could have to pay for permission to enter EU countries, Home Secretary Amber Rudd has conceded. 18 These rights are based on EU law. Whether they would be maintained following Brexit, and, if so, to what extent, is uncertain.

CAMINO REA RIZZO
Vol. 8(1)(2020):  As for content words, there is only one adjective qualifying Brexit: hard (f:14). In contrast, Brexit's right immediate collocates disclose a clear predominance of modal verbs showing different degrees of certainty or uncertainty (will f:74; could f:69; would f:33; may f:14; might f:13). Modal verbs are used to refer to situations, facts or events which may or may not be possible, which we expect or are not sure about, and here they serve the role of hinting at the critical conundrum raised by the event.
The co-occurrence of Brexit with a noun also needs to be highlighted. Brexit is immediately followed by 15 nouns whose frequency is higher than 5. In that position, Brexit is modifying its collocates as in Brexit + vote (f:139), negotiations When the collocation span is widened in order to pinpoint where collocates reach their top frequency so as to gain a better picture of the lexical content brought by the presence of the node Brexit, interesting findings are observed. The presence of impact is particularly notable for the clusters it generates with Brexit in combination with of, will, that and from. The most recurrent combination is the impact of Brexit on followed by what is impacted, that is, where Brexit turns out to be a matter of concern: the EU, arrivals, the tourism industry, Irish tourism, the northwest, the industry, medical tourism, the healthcare system, any future UK visa requirements, cruise tourism, travellers, and domestic tourists. With regard to the issues affected by Brexit, the R2 collocate mean reveals the conjectural answers to a series of questions posed about the potential results of the execution of Brexit in different areas of tourism, or the formulation of the indirect question itself where a modal verb is added in R1 position. The behaviour observed for the second top R2 collocate, affect, is closely analogous to the previous one, that is, affect and mean are found in fairly similar contexts as observed in concordances 19-20. 19 There is massive uncertainty about what Brexit will mean for the future of the country's economy and whether companies there will still enjoy easy access to other European markets. 20 Hopefully, this answers some questions about how Brexit will affect your winter holidays.

THE IMPACT OF BREXIT ON TOURISM AS RELEASED BY THE PRESS: A CORPUS-DRIVEN LEXICAL ANALYSIS
Vol. 8(1)(2020):  The top R3 post-collocate establishes a close connection with the second node under analysis. In fact, the same frequency is found between now the top L3 collocate, Brexit, and tourism, in Tables 2 and 3 because they refer to exactly the same concordances but from different angles.
Tourism Brexit (f:59) Tourism Impact (f:23) Tourism European (f:16) Tourism Table 3. Tourism's top collocates The concordance lines disclose a similar structure to the one generated by the collocate impact for Brexit as shown above. Here, Brexit and tourism are mostly found in the following combination: the + impact/effects + of Brexit on + adjective/null + tourism, like in concordance 21. There are also samples where Brexit is acting as the subject of a sentence whose object is tourism, for example in Brexit + gives/impacts/risks UK tourism. Such structure is sometimes lengthened by the introduction of a modal verb preceding the main verb, as found in Brexit + would harm/might help/may affect/could impact + tourism. Finally, Brexit is linked to a particular sector of tourism, such as Brexit and + Irish/international/medical/domestic/local + tourism.
21 So arguably the effects of Brexit on domestic tourism are much more important than what we do on our holidays. Table 3, all top collocates are located to the left of the node tourism. Among them, travel should be highlighted owing to its high frequency of co-occurrence in the L2 position. Such collocate is found adjoining the node by the coordinating conjunction and, to refer to travel and tourism industry or sector in general. Actually, this sequence of words is found preceded by British, consumer, strong, UK's, US, and world, and followed by businesses, council, flows, industry, and sector.

As depicted in
Surprisingly enough, L3 collocate European broaches the subject of aviation and low cost air companies facing Brexit. Its raw frequency (16) counts the repetition of three different combinations: the new European aviation and tourism agendas; the European Capital of Tourism project/initiative; and the European Parliament aligning tourism and aviation policy agendas (see concordance 22). The extended co-text in the corresponding news warns against the negative impact of Brexit on low cost airline travel and therefore, on tourism, employment and social mobility. Such issue was suggested to be included into Brexit negotiation agendas.
Following the procedure adopted to examine the previous node, the focus of analysis shifts into the right and left immediate content-word collocates of tourism whose co-occurrence frequency is higher than 5. Within such frequency level, there are 18 left co-collocates and 38 right co-collocates.
On A closer reading of the concordance lines retrieved for affect tourism provides several answers to a central question formulated as a direct interrogative sentence among the samples: How does Brexit affect tourism? Alternatively, the collocate and node are used to state an expected outcome. Their extended contexts reveal that major consequences for UK tourists seem to be related to the drop in the pound versus the euro, which would slowdown the UK's economic growth, and the regulations that would replace the European ones, which depends on negotiations.
The second verb, boost, is found in contexts dealing with the potential positive effect of Brexit on UK tourism with respect to staycationers and inbound tourists due to the devaluation of the pound.
Among the group of nouns to the left of tourism, leisure is worth highlighting due to its presence in the following combinations: (i) Hospitality, leisure & tourism, (ii) Leisure, tourism & retail businesses, and (iii) Leisure, Tourism & Retail. In relation to the Brexit issue their contexts discuss what is at stake. They manifest real worries about the potential restrictions on free movement of labour and goods that could lead to job losses and deter investment.
The high frequency yielded by the co-occurrence between Irish and tourism (f:116) introduces the critical issue concerning the status of the Republic of Ireland as an independent country that is a member of the EU and shares a border with the UK. Theoretically, the Republic of Ireland should not be affected by the exit of a European member. However, concordance 23 summarises the current situation: 23 "The impact of Brexit is already a reality for the Irish tourism industry and we estimate that the Brexit effect will cost Irish tourism at least €100m in 2017," Mr O'Mara Walsh said.
The rest of adjectives preceding tourism qualify different levels in the sector, ranging from a higher scope of British/UK/global tourism industry/sector in general, to more specific extents like inbound/outbound/domestic tourism and European/international tourism. Once again, the language samples inform about the positive results of Brexit for tourism within the UK as a consequence of the drop in their currency. Nevertheless, perhaps due to the early stage of the process where the news was released, there seems to be a more internal and local concern about the possible consequences of Brexit whereas the external repercussions are limited to the issue of Ireland and only four references to European tourism, as illustrated by concordances 24 and 25.
24 However, he cautioned that "while the industry was expecting a short-term boom in both domestic and inbound tourism over the next two years, the long-term prospects of the Tourism industry are dependent on the UK securing a deal with the EU that will protect the ability of UK and EU residents to easily travel." 25 The idea of a fund to protect areas hit by Brexit is supported by a number of officials from European tourism hubs popular with British tourists.
On the other hand, concerning the right co-collocates of tourism, nouns clearly outweigh other lexical categories. In fact, out of the 38 co-collocates encountered, there are 3 verbs (related f:6, driven f:5 and offering f:5) against 35 nouns. Among them, the top collocational frequency is attached to industry (f:232), followed by sector (f:82), Ireland (f:65), businesses (f:50), alliance (f:25), boom (f:14), and market (f:11); the frequency of the rest of collocates ranges from 11 to the minimum collocational frequency set for the study, that is f:5.
The terms tourism industry and tourism sector may be found in similar contexts so as to deal with the world of tourism in general. In the corpus, the first combination considerably exceeds the second accounting for the tourism industry in its broadest sense, which is defined as "the total of businesses that directly provide goods or services to facilitate business, pleasure and leisure activities away from the home environment" by the UNWTO. In turn, the tourism industry is constituted by a wide variety of sectors (accommodation, adventure tourism and recreation, attractions, events and conferences, food and beverage, tourism services, transportation, and travel trade) that provide diverse products and services to visitors. In consequence, the use of tourism sector would entail a more specific reference. In addition, tourism industry is also found in longer clusters like The Irish Tourism Industry Confederation (ITIC); Tourism Industry Council; UK's travel and tourism industry; UK's domestic tourism industry; and in combination with a place name in the genitive case: UK's/ Gibraltar's/ Ireland's/ Britain's/ Scotland's/ Suffolk's + tourism industry.
The next two collocates in ranking, Ireland (f:65) and businesses (f:50), are fairly related since 35% of the concordance lines retrieved for tourism businesses deal with tourism business in Ireland. Once again, the controversial situation of the island is felt as a source of concern and a critical issue facing Brexit as shown in concordances 26 and 27.
Another national organisation becomes visible thanks to the following cocollocate in ranking, Tourism Alliance (f:25), whose purpose is "to identify and develop policies and strategies to raise standards and promote quality within the industry and work with and lobby government on all key issues relevant to the growth and development of tourism and its contribution to the economy". 4 It comprises over 50 Tourism Industry Organisations representing around 200,000 businesses of different sizes in the UK. A closer observation of such combination's extended context reveals further potential problems. Among them, the risk of losing the hospitality workforce (near 25% of the workforce in tourism economy come from the EU) is overtly exposed on the grounds that workers might feel the lack of long-term career prospects in addition to visa issues. Likewise, the massive jump both in business rates and food costs as a result of the weak pound is repeatedly stated, together with the difficulties in setting up business ventures without European funding.

CONCLUSIONS
This research has presented an analysis of the press discourse reporting on the potential consequences of the UK leaving the EU for the tourism sector. The adopted corpus-driven approach has allowed a mainly quantitative analysis of the language samples enhanced by a qualitative perspective, thus providing a preliminary sketch of the relationship between Brexit and tourism as voiced by the press.
Concerning the first research question, which sought to identify the lexical choice of the press to convey the impact of Brexit on tourism, the quantitative results revealed those lexical items brought to the fore both by frequency and relevance parameters, which would identify the key factors at play. Among them, a straightforward answer was given to the well-known 5 Ws in journalism so that the reader can capture the essence of the event. In addition, figures evidenced the emergence of new words coined to convey new concepts related to the situation,

THE IMPACT OF BREXIT ON TOURISM AS RELEASED BY THE PRESS:
A CORPUS-DRIVEN LEXICAL ANALYSIS Vol. 8(1)(2020):  namely Brexit, Brexiteers and Brexodus. Another remarkable characteristic lies in the predominance of vocabulary on three major subjects governing the content of the news, that is, politics, economy and tourism. Precisely in such classification the initial expectation of a negative effect of Brexit was confirmed as evidenced in the economy group, firstly by the presence of negative word forms and later, by the negative meaning denoted in the extended contexts. However, against the pessimistic forecast for the tourism sector, the opposite effect was attested also by positive word forms and extended contexts. As for the second research question, further lexical information was revealed by the group of collocates and clusters generated by the two nodes under analysis: Brexit and tourism. In addition, specific instances of the areas that could be significantly affected were encountered when quantitative and qualitative methods were applied. It was not until syntagmatic relations were analysed that specific reference to the alteration of European agreements appeared. In that respect, the extended contexts where collocates and clusters are used generally try to make the reader notice hypothetical but likely scenarios. This is reinforced by the recurrent combination of the cluster the impact of Brexit on, followed by specific instances of areas or subjects where Brexit considerably hightens concern. Similarly, the cluster affect tourism is used within a series of direct questions which elicit the specific lexis to clearly display the potential consequences for tourism after Brexit. This mainly refers to a drop in the pound versus the euro, leading to the slowdown in the UK economy. Nevertheless, such consequence turns into a positive impact for tourism since the devaluation of the currency makes the country more attractive to foreign visitors and staycationers.
Special references are found to the restrictions on the free movement of workforce, to the explicit certainty that the Republic of Ireland will be unavoidably affected, and to the series of actions already initiated by different stakeholders in order to face any eventuality. Finally, the strong presence of modal verbs next to the nodes of analysis contributes to intensifying the widespread sense of uneasiness regarding the future of the UK tourism after Brexit.
To conclude, it is worth highlighting that the samples in the corpus pertain to a short timeframe after the Brexit referendum, compiled with the aim to capture the immediate reactions to the event in the tourism sector as delivered by the press. Therefore, a preliminary overview of the situation has been offered in this paper whereas the definite consequences are still to come and be linguistically analysed.