English in Spoken Swedish Discourse: A Look at Swedish Youtube Celebrities

(1)

Look at Swedish Youtube Celebrities

Master’s Thesis / Language and Society / Universiteit van Amsterdam

(2)

1 Abstract

In this study, we take a look at how English structures emerge in spoken Swedish by examining the discourse used by a selection of Swedish Youtube celebrities in their video content. Four hours of video material revealed a robust general distribution of five types of multilingual structures corresponding to a range classical CS phenom-ena where Swedish functioned as the matrix language to English as a guest language. Youtubers appeared to utilize English structures in their discourse for the discursive benefits that CS offers and the stylistic dimension that English provides in the form of prestige and a certain cool factor. This work was intended to challenge traditional approaches to the study of inter-linguistic influence which typically has targeted loan-words and borrowings as they eventually enter dictionaries or popular texts. This study ultimately reveals that the breadth of English influence goes beyond the threatened domains wherein Swedish is already known to be losing ground.

keywords: spoken discourse, anglicism, code-switching

2 Introduction

The multilingual fabric of modern Europe is characterized by the coexistence of new and old immigrant languages with the continent’s many endemic languages and threaded together by varying degrees of English proﬁciency with the highest levels of proﬁ-ciency concentrated in and around Northern Europe. According to the EF English

(4)

Proficiency Index (EF EPI) - a global and adaptive test of English reading and listen-ing skills - Europe enjoys the highest levels of English proficiency of any otherwise non-anglophone region in the world. As shown in Image 1, there are five main tiers of English proficiency in Europe: the first and highest proficiency tier is concentrated in Northern Europe with the Netherlands followed by Sweden as the top two perform-ers. The second tier expands out and largely southeastward into the German- and parts of the Slavic-speaking world with Germany and Austria heading that tier. The third through fifth tiers expand out further east- and southward ultimately reaching Turkey and Azerbaijan.

Image 1. Ranking of EF standard English Test, scoring computed by average percentage of correct answers.

Northern European countries have a number of factors in common at the policy and cultural level which contribute to generally high English competence: starting in primary school, English is taught as a mandatory foreign language. In the classroom, teaching methods tend to emphasize communicative proficiency rather than grammat-ical accuracy, fostering effective communication skills early on. This high proficiency as a consequence of each nation’s education system combined with the heavy consump-tion of anglophone media in the form of television, film, music, and the like, frame the intersectional and intensive influence of modern varieties of English on the overall lin-guistic repertoires of Northern European populations. Europeans are not only bom-barded with English in their respective school systems and in their media consumption habits; Northern Europeans, as a consequence of aberrant wealth, can and do travel often, using English as a lingua franca. English is also used as a lingua franca among Europeans of disparate national origins and in general among people in the region who

(5)

do not share comparable proficiency in another non-anglophone tongue (“EF English Proficiency Index, Europe”, 2018). There is even a popular notion that if you speak English there is perhaps no reason to learn other European languages, especially lan-guages with smaller speaker populations and spheres of influences such as Dutch or Swedish. This is all to say that English is a very prominent skill in the Northern Euro-pean region and as such, very salient in the minds of its residents. Broadly speaking, I am interested in the ways in which English prominence might be disrupting the over-all linguistic ecosystem of the region. Historicover-ally, lingua francas have been powerful vehicles for both gradual and relatively quick linguistic change. How might English, in this way, be affecting the on-the-ground realities of other modern languages?

Having lived for a number of years in Sweden and as a native English speaker, I had to contend personally with the prominence of English as I endeavored to learn Swedish to fluency. Learning a new language as an adult is notoriously effortful, par-ticularly in the early stages; it is made still more difficult when the speaker population with whom you are dealing is also highly proficient in your mother tongue. When learning Swedish in Sweden it is often simpler to fall back on English once your lin-guistic resources have been drained yet communication must continue. Through insis-tence and a tremendous amount of patience, I persevered and managed to gain a great deal of competence in spoken Swedish, however several aspects of the process were striking to me: first, that tendency to rely on a more dominant language in the earlier phases of language learning, often referred to as crutching in linguistics, had a peculiar cadence when the language I was falling back on was English; I noticed that Swedes themselves would sometimes get tripped up when racking their mental lexicons for certain vocabulary and terminology, often resorting to English words or phrases when Swedish equivalents weren’t quickly accessible to them. English’s general salience in the minds of Swedish people seemed to lead to competition with what would be the appropriate Swedish correspondences, and English often won out in the end. For ex-ample, when mentally scouring for an applicable term a Swedish speaker might opt for the word research originating from English instead of the more established Swedish equivalent undersökning. In this way, where there seemed to exist lexical instability be-tween an anglicism and a codified Swedish lexeme, Swedes seemed to lean toward the English option.

I also noticed a tendency among Swedish speakers, especially younger Swedes, to inject a variety of anglicisms into their speech, sometimes even a multitude within a single conversation. The varieties of English structures ranged from single words to whole utterances. For example, one might frequently overhear teenagers tacking on tags or adding discourse markers such as basically, anyway, or ya know and using set expressions like I love you or makes sense. The sometimes overwhelming amount

(6)

of English present in otherwise Swedish discourse made my initial crutching on my native English less problematic and even naturalistic when used strategically and in accordance with the Swedish fashion; although the insertion of English is certainly a very productive phenomenon, there did seem to be structures and contexts where anglicisms seemed to violate a native Swedish speakers’ intuitions about their use. In other words, the use of anglicisms appeared to be systematic. A similar pattern of regularized anglicism usage also stood out to me when I arrived in the Netherlands and began listening to and engaging with people from the Dutch-speaking world. This hinted to me that a greater pattern was perhaps at work here.

These instances I describe are familiar to Swedes, the Dutch, and other North-ern Europeans; when questioned about them they usually laugh or, at the extreme, bemoan the decline of their language and culture as people are so want to do in the face of changes. They can rarely however provide a clear answer about why they use English in this way. Some will say that it’s just fun or that English just sounds better in certain contexts. A long-term curiosity about this phenomenon and these anecdotal experiences led me to this thesis, its research questions, and subsequent study. My central aim with this work is therefore to illuminate and explore how English struc-tures emerge in contemporary spoken Swedish. Whether the patterns that crop in this work are indicative of a greater trend across Northern Europe will remain, however, mere speculation.

3 Research Questions and Hypothesis

As mentioned in the introduction, most broadly speaking I am interested in how vari-eties of English might be disrupting the linguistic environment of Northern Europe. Research on the impact of one language on another is usually aimed at how features of the influencing language arise in the matrix language. These are typically sussed out via the analysis of what has been taken up and sanctioned in written texts, in other words what is eventually validated by arbitrary language authorities such as dictionaries and style guides; this is the field associated with the study of loaning and borrowing; how words and expressions gradually enter the language to the extent that non-linguists feel comfortable deeming these novelties real words. However, loanwords and borrow-ings are in essence artifacts of an earlier, usually tacit, sanctioning of usage within the spoken language. In this way, accounting for loanwords and borrowings only accounts for the final stages of language contact. Conversely, contemporary spoken speech is indicative of the latest realities of language use. It also captures more nuances of lan-guage use than simply a word or expression here or there ultimately garnering accep-tance by language authorities. For this reason, this study is positioned toward the

(7)

real-ities of spoken speech, speciﬁcally toward how English structures and features emerge within spoken Swedish discourse. The study’s formal research questions are as follows:

1.) How do English structures crop up in spoken Swedish, i.e. in what forms and in what contexts?

2.) What functions might these structures serve for the speaker and the discourse?

The goal of question 1 is to determine what forms English structures in spoken Swedish (henceforth referred to as ESSD or merely anglicism(s)) take by devising a scheme for their categorization. By doing so, it seeks to address the contexts in which instances of ESSD arise, i.e. as discourse markers, content words, post-sentential em-phases, etc, and serves to map out and deﬁne this phenomenon formally. Question 2 goes further to address what sociolinguistic and discursive functions might ESSD serve and therefore pertains to why this phenomenon occurs at all. In other words, these two research questions are intended to uncover what ESSD is and why it happens. A selection of Youtube video material from popular Youtube celebrities will function as the basis for data collection. These research questions are in correspondence with the following hypothesis:

Swedish speakers utilize English structures for their sociolinguistic and discursive beneﬁts. Furthermore, if language is a skill, subject to the gen-eral cognitive capacities of the individual, and relatively high English com-petence can be assumed on the part of the majority of the Swedish popula-tion, then English structures might be said to crop up in their L1 as a conse-quence of the regular activation of their L2.

4 Background

4.1 Overview of the Language Scenario in Sweden

Swedish is a Germanic language and the most widely spoken of the North Germanic languages. Varying degrees of mutual intelligibility exist among Swedish, Norwegian, and Danish, and the trio traditionally constitute the Scandinavian language contin-uum. The Language Act of 2009 established Swedish as the national language of

(8)

Swe-den for its population of 9,150,000 (2012 census). The language also enjoys the status of a national language in Finland where around 274,000 people speak Swedish as an L1 and 2,410,000 as an L2. Finnish is the most widely spoken ﬁrst language in Sweden after Swedish, however the Finnish speaking population declined by around 4 percent between 1975 and 2006 with around 201,000 L1 speakers in the latter year. During and after the fall of Yugoslavia in the 1990’s, Sweden saw a signiﬁcant increase in the population of Serbo-Croatian speakers whose numbers rose from around 34,000 in 1975 to 113,000 in 2006. Many other newcomer languages also coexist with Swedish in modern Sweden. The majority of newcomers have roots in Europe, the Mid-East, and North Africa. Today after Finnish and Serbo-Croatian, Arabic prevails as the most commonly spoken newcomer language in the country with a speaker population of around 93,000. The urban fabric of Sweden, like most other modern urban regions, is exceedingly multilingual in nature (Ethnologue, 2018, Parkvall, 2009).

The emergence of Swedish in its modern varieties began during the 16th century and is generally associated with the expansion of the Swedish empire; at no other time was Sweden more multilingual than during its reign as a great power or stormaktstiden in Swedish. After the decline of the empire and before World War II, the region that is now Sweden experienced a sweeping linguistic Swediﬁcation. This period was also characterized by an emphasis on nation-state building which favored cultural and linguistic homogeneity and therefore bolstered the regular use of Swedish and saw to its gradual standardization. (Parkvall, 2009).

From a linguistic standpoint, following the general trend among Indo-European languages, all mainland Scandinavian varieties experienced significant morphosyntac-tic simplification between the middle ages and the appearance of their early mod-ern forms. At present, Swedish is an SVO language with no system of case endings, no verb conjugation for person or number, and only two genders with masculine and feminine having fused into one gender referred to in English as common gender; how-ever the language maintains a rich system of noun morphology, famously its system of post-nominal endings indicating not only singularity, plurality, and gender but also definite- and indefiniteness (see Image 2).

(9)

Image 2, from Stensson, 2013.

4.2 The Historical Inﬂuence of Other Languages on Swedish

“Just as with immigrating peoples, words come along with other imports such as ideas, phenomena, and objects.” (Stålhammar, 2010)

Outside inﬂuence has come at the Swedish language from many sides and sources. Like most European languages, Greek and Latin words began to enter the Swedish lex-icon already in the Old Swedish period and with the christianization of Scandinavia. Some of the most common examples of loanwords from this period include kristen,

kors, präst, påve, altare, kalk, and kloster. The Renaissance and Enlightenment saw

re-newed lexical inﬂuence from Greek and Latin with respect to the cultural and techni-cal vocabulary of Swedish. German affected Swedish more extensively than any other language and at a particularly formative time during the Middle Ages when modern Swedish was in its infancy (Stålhammar, 2010).

“New organizations and functions were introduced, which mirrored both high-ranking and oﬃcial titles such as greve, hertig, hovman, riddare, borgmästare, fogde,

råd-man.” (Stålhammar, 2010)

German remained the most important foreign language up until the end of the Second World War when it gave out to English. A 2006 EU survey, however, revealed that German is the still the second most commonly learned foreign language in Sweden (Eurobarometer, 2006).

Although German has likely had the most lexical influence on Swedish quantita-tively, English has had a sustained barring on the language since the Middle Ages and the broadest across semantic fields with loanwords and borrowings entering Swedish via the Industrial, Scientific, and Tech Revolutions. Initially sparked by the advent

(10)

of trade and seafaring around 1200, imported words related to fashion, sports, poli-tics, and entertainment began to ﬂood into Swedish and continue to do so today. The British Empire and the Victorian Era brought words such as indiansommar, ranch,

veg-etarian, sandwich, slum, veranda, bumerang, curry, merkantilsystem, shop, and even terms

reﬂective of the more gruesome nature of the colonial period like slavdrivare and

lyn-cha. The word internet ﬁrst appeared in Swedish in 1992 and webben (the web) in 1995

(Stålhammar, 2010).

4.3 Attitudes toward English in Contemporary Sweden

English holds a powerful and sweeping position in modern Sweden. A large scale study of the stratiﬁcation of and attitudes toward English in Sweden by Catharina Nyström Höög (2005) revealed that even low-educated people indicate having a high degree of

command of English. Almost everyone involved in Nyström Höög’s study reported using

English at work and in their freetime during a typical working week; several general attitudes emerged, among them that English proﬁciency is necessary on the part of Swedish politicians in order to attend to the country’s interests on the world stage and in the European Union, that Swedish scientists need to know English for their work, and that English should function as the world’s lingua franca (Nyström Höög, 2005). Considering that Swedes exhibit the second highest level of L2 proﬁciency in English in the world, it is certainly in their best interest to maintain such attitudes.

On the subject of anglicisms within the Swedish language itself, most of the previ-ous research has focused on loanwords and Swedish society’s attitudes toward them. Although attitudes toward English as a language are generally positive, personal feel-ings about English imports in the language vary starkly (Stålhammar, 2010). The majority of informants in the Nyström Höög study held that Swedes use too many English-imported words in Swedish; a majority also believed that a translation loan should be used if said translation loan is more frequently occurring than the English loan. In practice, however, it seemed that informants in the study still used English words where valid Swedish equivalents also exist. Nyström Höög notes that infor-mants’ positions with regard to English imports seemed to be something which most informants had not previously considered, and when pressed to do so the most com-mon explanation for their use was pragmatic rather than ideological reasoning. For example, when taking a contra-position toward anglicism use informants argued that it makes language learning harder when one must parse out the English, while con-versely arguing that English use makes communication on the internet easier. Fur-thermore, informants opinions varied when choosing between certain high-frequency English loanwords with comparable Swedish equivalences or at least translation loans.

(11)

For example, 92% of informants agreed that livvakt was more favorable than

body-guard, while 79% held that design was a better alternative than the Swedish formgivning

(Nyström Höög, 2005).

As previously mentioned, the bulk of the previous work on the incursion of other languages into Swedish have focused on loanwords and borrowings, usually as they ap-pear in newspapers and articles, and attitudes toward them among different age groups and socioeconomic backgrounds. Foundational studies on the subject by Ljung in the 1980’s revealed that, despite concerns about the growing prominence of English, less than one percent of the words in the newspaper texts that were analyzed consisted of English loanwords; a later study by Josephson (2004) showed that English loans in newspapers had increased by some 10 percent since the 1980’s (Fälthammar-Schippers, 2014). This increase is reﬂective of the younger generation’s growing acceptance of an-glicisms. Ljung maintains that people under 30 are more likely than older people to accept English loaning with respect to words and phrases and even syntactic and con-struction loans. Ljung also establishes that age is the predominant factor with regard to whether an English- versus Swedish-based orthography will be selected whereby younger people are more likely to prefer an English-based orthography.

At the behest of a parliamentary committee in 2002, Kenneth Hyltenstam con-ducted a study which conﬁrmed Ljung’s ﬁndings with the following results:

Image 3, Hyltenstam, 2002

As Image 3 clearly exhibits, the majority of participants under 30, around 80 percent, either entirely or at least partially disagreed that too much English is used in Sweden, while those who agreed with the statement were from older generations (Svensson, Viberg, 2007).

(12)

Svensson and Viberg (2007) conducted a subsequent study along similar lines among high school age students and their teachers; both teachers and students had quite pos-itive attitudes toward English in general, however, notably only 53% of the students involved disagreed entirely or partially with the above statement while 100 % of the teachers agreed that too much English is used in Sweden. In response to the ques-tion - how many times have you spoken, read, or written in English this past week? - 64% of participants answered that they had used English many times each day, and everyone involved reported having used English at least one time that week, that is to say, fre-quency of English use was very high despite the variety of attitudes toward English; this is keeping with the trend among studies of English in Sweden wherein attitudes toward English are not mirrored by use, i.e. what is felt about English is not reﬂected in prac-tice. When study participants were presented with ideas that have both English- and Swedish-based possibilities, for example workshop versus arbetsseminarium, and were asked which option they prefered, 83% answered that they used both depending on the context; they couldn’t, however, demarcate precise contexts for the use of either possible term (Svensson, Viberg, 2007).

Anxiety around the general increase in loanwords is starkest with regard to the idea of domain-loss. Recent research by Fälthammar-Schippers (2014) revealed that, in accordance with aforementioned historical precedent, the domains of fashion, beauty, music, entertainment, ﬁlm, television, sports, leisure, food, health, technology, and engineering are among the most vulnerable (Falthammar-Schippers, 2014). The ever prescient Mall Stålhammar warns that in truth English’s eﬀect on Swedish is strongest where

it is noticed the least and maintains that while direct loans are immediately obvious to

us, translation loans determine how new Swedish words come about (Fälthammar-Schippers, 2014). Stålhammar (2010) also warns:

“English loanwords in everyday language, as a consequence of their prestige, can stir a nuanced usage of Swedish, but are hardly a threat to Swedish’s existence; however, a language’s continued use by an entire population and within every domain is threatening if one language begins to dominate within a particular area, a language domain. Contact is broken between the Swedish-speaking general public and the English-Swedish-speaking specialists within that domain, and eventually lay people lose access to special knowledge.” (Stålhammar, 2010).

4.4 What are loanwords?

So far, I have used loaning, loanwords, and borrowings mostly interchangeably, as well as the term anglicism when referring to English speciﬁc loaning, loanword, and borrowings;

(13)

however this will begin to shift in the theoretical framework as I approach the ESSD directly. Ljung deﬁnes three types of anglicisms in Swedish: direct loans, translation

loans, and construction loans (Ljung, 1988). Direct loans are characterized by the use of

individual words or phrases. Direct translations of words, phrases, and meanings are categorized by Ljung as translation loans, while construction loans relate to syntac-tic and morphological loans (Fälthammar-Schippers, 2014). In this study, I will focus only on direct loans used in spoken speech, however the nature of ESSD is somewhat outside the scope of traditional approaches to loanwords and borrowings.

5 Theoretical Framework

Several basic theoretical assumptions underpin my approach to the phenomenon of English structures cropping up in spoken Swedish discourse:

1.) Language is a skill and language learning is skill learning which emerges as a function of the input.

2.) Languages are not tightly bound systems but commingle in their cognitive representations.

3.) Linguistic forms point to social meanings.

5.1 Assumption 1: Language is a skill

Modern linguistics is a relatively young discipline. This is made obvious by the fact that even today linguists disagree about the fundamental nature of language as it is learned (or acquired as it is more often conceptualized) and how it is stratiﬁed in the brain. Although perspectives vary considerably, two schools of thought are tradition-ally borne out of the debate: the nativists and the non-nativists.

5.1.1 Nativism

The nativist perspective is founded on a number of key assumptions. Nativists tend to approach the theory of language based upon the ability to use language. In this way, the structure of language is viewed as something profoundly complex, i.e. natural grammar is seen as too difficult for a child to extrapolate merely from the available in-put. From this perspective, language is structure-dependent (Pinker,1994) allowing for constituents to be embedded and structured ad inﬁnitum with respect to complexity

(14)

and length (Pinker 1999, Hauser, Chomsky, and Fitch, 2002). This is famously known as the poverty of the stimulus argument and is largely associated with the Chomskyan school of linguistics. The poverty of the stimulus argument informs the nativist or

generativist idea that there must then exist an innate language mechanism in the brain

that has evolved over time as a consequence of random mutations in human evolution ultimately resulting in the human capacity to acquire and use language. The nativist position has developed considerably since its infancy in the 1950’s, but these core as-sumptions persist (Kruger, van Rooy, 2015). Furthermore, children who have not yet reached puberty are then said to make use of this innate system which allows them to acquire their native language without tremendous effort via a Universal Grammar (UG)(Chomsky, 1957); UG divides language into principles - abstract representations of grammar - and parameters - language speciﬁc properties found mostly at the surface structure level.

Two properties undergird the idea of language’s immense complexity: language is structure-dependent (Pinker,1994), and an innate language mechanism allows for con-stituents to be embedded and structured ad inﬁnitum with respect to complexity and length (Pinker 1999, Hauser, Chomsky, and Fitch, 2002). Structure-dependency en-tails the existence of categories of natural language structures. For example, rather than simply approaching lexemes directly as they occur in a linear order, linguistic operations are computed via natural categorical structures forming an invisible hier-archy of structures (Kruger, van Rooy, 2015). Kruger and van Rooy’s article The case for an emergentist approach (2015), cites the classical example of this property with Chomsky’s (1957) example of wh-extraction:

The interrogative pronoun who in 1a appears outside of the embedded clause. 1b illustrates the compounding complexity of structure-dependency as further clauses be-come embedded in the sentence structure. This is taken as evidence that rules exist which constrain the structural realities of language, typically captured by terms like

islands, subjacency, or bounding. Attitudes toward the nature of structure-dependency

have evolved over time as counter-arguments have been put forth, but the generativist school still insist that at least some property of recursion frames the structure of nat-ural language and in doing so allows for an inﬁnite set of rules generating an inﬁnite number of sentences (Kruger, van Rooy, 2015).

(15)

Whatever the particulars, this rule-governed innate system is believed to repre-sent the existence of a specialized mechanism within the brain for language which children can make use of when acquiring language before a certain age; what that age or age range might be has been the subject of much debate within the field of language acquisition. The hypothesis most closely associated with the nativist perspective con-ceptualizes this phenomenon as an ill-defined age range known as the critical period (CPH, Lenneberg, 1967). After this critical period, acquiring a language as if it were your mother tongue is considered very rare or impossible. Questions as to the ori-gins of this innate mechanism have been reduced to questions regarding human evo-lutionary biology. In general two views exist: the first view holds that gradual adapta-tion allowed an innate language mechanism to become entrenched in human anatomy (Bloom, Pinker, 1990, Jackendoff, Pinker, 2005); the second perspective, however, ar-gues that language emerged suddenly, perhaps from a single mutation (Chomsky, 2007, Chater, Christensen, 2008).

In sum, the nativist position on language is characterized by the assumptions that language is very complex and too little input exists for children to be able generalize the rules of a language’s grammar therefore implying some sort of innate mechanism. This innate mechanism might be conceptualized by a kind of universal grammar where universal principles govern all language and language-speciﬁc parameters dictate the grammatical eccentricities of speciﬁc languages. Finally, this would-be mechanism in the brain specialized for language is said to have evolved either gradually or suddenly during the course of human evolution as a consequence of human thought as opposed to a need for communication (Kruger, van Rooy, 2015).

5.1.2 Non-nativism

Non-nativist schools of thought are less united under a deﬁnite title and are more varied in the particular theoretical assumptions they address. Some common terms for non-nativist schools are functionalist, emergentist, usage-based, cognitive, among others. In general, they tend to address the use of language rather than the ability to use language. In principle, non-nativists view grammar in terms of constructions which is deﬁned by a pairing of linguistic form with either a meaning or a function:

“A construction-based view of grammar, like cognitive (Langacker, 1987) and functional (Croft, 1995) linguistics more generally, sees no strict diﬀerentiation between semantics and syn-tax, as the generativist view does, but rather embraces the relationship between linguistic form and meaning as central to an explanation for the shape language takes.” (Kruger, van Rooy,

(16)

Constructions then range from very abstract to so speciﬁc that they can be cate-gorized as ﬁxed, idiomatic expressions. Construction-based grammar does recognize very loose grammatical categories as in the following transitive construction:

Form: NP1 - VERB - NP2 Meaning: Agent - Process - Patient Example: The boy hits the ball.

The typical subcategories associated with generative grammar are reallocated to argu-ment construction schemas as seen above. These arguargu-ment constructions are very gen-eral but do not attempt to apply universally across one language or all languages. This is a fundamental characteristic of non-nativist approaches to language in that linguistic categories are not thought to be aprioristic but rather emerge from the use of language (Kruger, van Rooy, 2015). In other words, language is said to be rule-describable rather than rule-governed. Bybee (2010) maintains that much of the grammatical structure of language can be attributed to ﬁve domain-general cognitive processes shown to be at the fundament of cognitive processing in humans: categorization, chunking, rich memory

storage, analogy, and cross-modal association.

Categorization is derived through generalizations that can be made during linguis-tic exposure and mapped onto stored representations. Chunking is a mechanism for the analysis of sequences of linguistic units which can be combined to form even more complex units and higher levels of representation; chunks are eventually stored and accessible in whole units. Analogy is a device used for extrapolating to or from new ut-terances based on existing, stored constructions; in tandem with categorization, slots can be ﬁlled with novel yet similar constructions. Rich memory storage presupposes a wealth of mentally stored real-world examples of linguistic constructions which are annotated with - for instance - phonetic, morphosyntactic, pragmatic, and semantic information about each example. Finally, cross-modal associations involve the map-ping of meaning on the largest chunk possible rather than deconstructing them into their fundamental components and assigning smaller units of meaning to each compo-nent which would then be reanimated at the highest levels of representation (Bybee, 2010, Kruger, van Rooy, 2015). These cognitive processes are presented here in ser-vice to the idea that it is possible to develop a theory of language from the bottom up rather than the top down approach of the nativist school and to affirm that any theory of language architecture should incorporate a cognitive perspective.

Returning to the poverty of the stimulus argument, the non-nativist perspective is that children develop an increasingly more schematic and abstract network of con-structions through lived experience with linguistic input which ﬁlters through

(17)

domain-general cognitive processes (Kruger, van Rooy, 2015), i.e. linguistic skill emerges as a function of the input. Since children’s brains are still developing, their minds mold around the linguistic input and its use. As with most other skill-based tasks, people who learn and use a skill from a younger age can make use of the more implicit nature of child learning and become generally more proﬁcient or native to a particular skill. Contrastively, skill learning later in life is more effortful and, especially from a linguis-tic standpoint, will pattern differently with respect to the ultimate attainment of that skill as compared to those who begin their skill learning earlier in life. The issue of the poverty of the stimulus, therefore, can be addressed by changing the fundamental as-sumption: children are not learning an impossibly complex theory but rather building up a set of procedures for processing and producing language (Chater, Christiansen, 2018).

As Kruger and van Rooy (2015) point out, a central tenet of the theory of evolution is that there is no predesignated structure; instead, minute advantageous variations become the norm within a given species. Cultural evolution then reﬂects biological development in that it is not derived from preplanned design but rather develops from experience; therefore the non-nativist view argues that the communicative function of language is central to the development of languages in real-time and their subsequent relationships to one another (Kruger, van Rooy, 2015). The present study approaches language from a perspective based on these non-nativist assumptions about language wherein language is viewed as a skill emerging as a function of the input which has evolved in humans mostly for the purpose of communication.

5.2 Assumption 2: Languagesarenottightlybound structures

Lay speakers often view individual languages as largely separate systems subject to sets of language-speciﬁc rules. In linguistics, individual languages are traditionally broken into subdomains such as a phonetics, phonology, morphology, syntax, etc; every lan-guage then has a set of elements belonging to each of those categories. Linguists also know that languages do not have clearly deﬁned natural boundaries; instead they are distinguished from one another via human socio-political ideologies like culture or the nation-state. For the linguist, the degree to which linguistic varieties are mutually in-telligible is the most salient feature for distinguishing a language from a subset of a language - a dialect - but this axiom does not always hold in practice as some some varieties of a single language are not mutually intelligible with other varieties of the same language. Therefore, where one language ends and another begins is not always clear even to linguists and especially not to non-specialist every day speakers whose habits and attitudes play a leading role in the development of languages and linguistic

(18)

varieties.

The transition from one language to another in use is conceptualized as switching from one deﬁnite natural system to another, however is it really valid to view these as truly separate systems, i.e. are they housed in separate places in the brain?

“An inﬂuence of the active but not currently selected language(s) is omnipresent and appears to aﬀect all domains of language” (de Groot, 2011).

In a section entitled, The Interacting Languages of Bilinguals and Multilinguals: Eﬀects on

the First Language, de Groot (2011) discusses the competing linguistic inﬂuences within

the minds of multilinguals. De Groot provides evidence that all of the languages reg-ularly used by the individual interact even when not presently selected. This is what creates the phenomenon of an accent when a person’s first language exhibits influence over their phonetic and phonological use of a second or nth language. While accents are typically associated with phonetics and phonology, the term is often extended to other linguistic subdomains like syntax and semantics. Furthermore, de Groot (2011) maintains that this interaction is not limited to first language(s) influence on subse-quently learned languages; the reversed dynamic is also possible with later languages influencing the production and comprehension of the first language(s).

Already in the late 1970’s, Edith Mägiste (1979) assembled a considerable amount of evidence for multilingual cognitive interaction. These cross-linguistic effects appeared across a battery of studies which utilized a diverse set of comprehension and produc-tion tasks such as word reading and picture naming. In one of these studies - involving German and Swedish monolinguals, German-Swedish bilinguals, and trilinguals with another language in addition to German and Swedish - revealed that reaction times slowed down with each additional language. Another study by Flege (1987) showed the disruption in voice onset time (VOT) when comparing English and French mono-linguals with bimono-linguals of both languages; English speakers typically exhibit longer VOT’s than French speakers while bilingual speakers of both fell around the average for those two languages suggesting some sort of commingling cognitive representa-tions of each language’s system of phonetic categorization.

A study by Dussias and Sagarra (2007) displayed evidence that grammatical accent may be a side-effect of cross-linange inﬂuence. Their study involved Spanish-English bilinguals heavily immersed in their L2 English and examined the way they parsed a structurally ambiguous Spanish sentence:

(19)

actress who was on the balcony).” (Dussias, Segarra, 2007)

English monolinguals tend to assign the action to the closest agent, namely the actress, while Spanish monolinguals favor assignment to el hijo in ambiguous statements like this. The Spanish-English bilinguals, however, favored assignment to the actress sug-gesting cross-linguistic inﬂuence of participants’ L2 on their L1.

Clearly, there is precedence for multilingual interaction at the cognitive level which might suggest that languages are not neatly stratiﬁed in the brain as explicitly separate entities but rather blend together. From a cognitive perspective, languages are said to be kept separate during online production and comprehension via a system of language control. According to language-mode theory, in a given communicative context one language is chosen as the matrix language and this language is kept highly activated; the other languages present in the mind of the individual are referred to as guest languages and are activated in certain context such as with speciﬁc interlocutors or with respect to certain topics. The guest languages are activated to a degree congruent with these changing contexts yet still to a lesser degree than the matrix language. As contexts for guest languages arise and peak so too does the cognitive activation of these languages; it is here that the heaviest amount of language mixing occurs (de Groot, 2011).

Modern views on multilingualism are challenging this assumption that languages are static, strictly separate systems. Matras (2009) posits that the central function of language is efficient communication. All multilingual phenomena such as loaning and borrowing, code-switching, transfer, etc can be explained by way of a relatively small number of assumptions. One of these theoretical assumptions is that, instead of switching among sharply deﬁned systems, all language users make use of a linguistic

repertoire which houses all of their linguistic knowledge irrespective of language

(Ma-tras, 2009). In this way, all linguistic structural features coexist simultaneously in the minds of speakers just as the neuro-linguistic evidence suggests, however the need for efficient communication places constraints on what structural features are most suited for a given context and stratiﬁcation of interlocutors. Hypothetically, if every speaker had the same linguistic competencies - for example, if everyone held comparable com-petency in Arabic, Swahili, and Cantonese - there would be no practical impediment to keeping these three systems separate. We could utilise the sound systems, vocabu-laries, and even grammars of all three languages without fear that the person we were engaging with would not understand. It is merely for social reasons and in light of practical realities that we keep languages apart in our discourse. It is perhaps for this reason that multilingual structures such as code alternation ﬂourish in environments where people have diverse linguistic competencies.

(20)

linguistics and sociolinguistic theory might suggest, languages are not tightly bound systems but rather coexisting in the their neurological representations and evolve to suit the speaker’s need for efficient communication, the latter of which we will explore later in this work.

5.3 Assumption 3: Linguistic forms point to social meanings

The idea that linguistic forms are linked to social meanings is fundamental to the study of sociolinguistics. At some point during adolescence we become aware of this link-age. We can tell, for example, where someone comes from or what kind of socioeco-nomic background they might have by the way they speak. As children go through adolescence and begin to shape their identity and explore group membership, they become ever more acutely aware of how linguistic features index certain social mean-ings (Wagner, 2012b). Adolescents utilize stylistic variation to indicate their position among their peers and to distance themselves from their predecessors.

Gal and Irvine (1995) maintain that the boundaries of linguistic varieties are not natural and do not form outside of human will and political forces (Gal, Irvine, 1995). Speakers notice these linkages and place them within the greater framework of their socio-political ideologies. Linguistic structures are interpreted through social ideolo-gies which tend to reinforce and exaggerate differences. They explore three semiotic processes which engender the emergence of sociolinguistic indexes.

Iconicity or iconization is the semiotic process by which social images become so

inextricably linked to certain linguistic forms that it is as if a linguistic feature somehow

depicted or displayed a social group’s inherent nature or essence (Gal, Irvine, 1995). This creates

the illusion that because someone’s speech involves certain social markers it is iconic of them having the characteristics attributed to that group. For example, speaking in a certain regional dialect might be iconic of certain attributes such as maleness, tough-ness, low-education level, etc. Another semiotic process put forth by Gal and Irvine (1995) which they refer to a recursiveness, serves to proliferate sociolinguistic differences among speakers. This process involves attributing some difference in speech as being indicative of a difference at some other sociological level creating sub- and supercate-gories of oppositions at varying levels. In other words, observers of speech shape their sociolinguistic identities off of one another ad inﬁnitum. Finally, and most grimly, a process known as erasure involves the rendering of certain form-meaning linkages in-visible, either through an imagined linguistic heterogeneity or an active suppression of certain groups and their particular language habits (Gal, Irvine, 1995). It is processes such as these that bolster the idea of one-language, one culture, however they are merely mechanism through which human beings interpret sociological and linguistic

(21)

realities. They are, in any case, indicative of the power of these sociolinguistic linkages which are often viewed as implicit truths.

The assumption that every linguistic structure or feature carries an index to some social meaning informs this study’s approach to ESSD; the use of English structures by Swedes does not exist in a socio-political vacuum but rather must come with semiotic ramiﬁcations and sociolinguistic valuations.

5.4 ESSD as Code-switching

ESSD seems to arise in various forms ranging from single words to whole utterances but always embedded within the greater structure of predominantly Swedish discourse. I therefore have chosen to frame it as an instance of code-switching as it undeniably shares properties with this kind of language contact phenomenon.

Code-switching is typically associated with native or simultaneous bilinguals, how-ever all speakers can make use of the relevant elements of their linguistic repertoires to some extent. Even monolinguals (if such people truly exist) draw from a kind of repertoire when switching between linguistic registers or between dialects, a subset of code alternation referred to as style shifting. Code-switching (CS) encompasses a wide range of multilingual phenomena and is cumbersome to categorize deﬁnitively. CS might involve a single or even several words scattered throughout an otherwise differ-ent linguistic system (usually referred to as code-mixing); or a speaker might begin an utterance in one language and switch to another language at some point, usually at a clause boundary such as in the following classical example:

”Sometimes I’ll start a sentence in Spanish y termino en español.” (Poplack, 1980).

This type of switching is known as intra-sentential CS as compared to inter-sentential CS where the speaker alternates after the sentence boundary. However, intra-sentential CS, as in the example from Poplack, can tell us more about how the morphosyntax of both systems react in contact with one another.

Muysken (2000) offers a CS categorization schema consisting of three strategies: ﬁrst, alternation whereby two linguistic systems are kept separate in an A-B conﬁg-uration as in the example above. Second, congruent lexicalization emerges when two languages share a similar syntactic structure which can be populated with the lexicon of either contributing system. This is most common of course among languages who are closely related typologically. Finally, when words or phrases are installed into the syntactic structure of an otherwise foreign utterance, Muysken refers to this as

(22)

structural properties. These schematic properties are broad and do not attempt to capture all CS phenomena but rather serve as a jumping-off point.

Clearly distinguishing CS from other types of multilingual structures can be diffi-cult. For example, the term borrowing has also been used to describe a wide variety of structures ranging from the transfer of smaller linguistic features like phonemes and morphemes to the use of entire clauses, however borrowing is often typiﬁed by the inclusion of morphological and phonological features of the matrix language. For in-stance, the Japanese borrowing of the English word baseball adheres to the phonological morphological systems of Japanese in its established form as basubaru. In this way, bor-rowings acquire the feeling of being truly a part of the borrower language’s anatomy. This validating act of borrowing is often further strengthened when such loanwords are sanctioned by dictionaries and language authorities, which update their status to real

words by lay speakers. However, new borrowing can also be established in real-time by

the speaker as they draw from other linguistic systems in a hyper-contemporary way; these yet officially unattested borrowings are referred to in linguistics as nonce borrow-ing (Poplack et al, 1988). This range of borrowborrow-ings, from firmly established to quite novel, obscure the boundaries between what should qualify as a borrowing versus an instance of CS (Bullock, Toribio, 2004). It is my view that these are simply two nom-inal alternatives for structural objects ultimately existing on the same fluid spectrum of linguistic phenomena, i.e. the transfer of material from one language to another. CS might also be confused with other types of multilingual phenomena. For example, mixed languages such as the famous Media Lengua which incorporates the lexicology of Spanish within the morphosyntax of Quechua. Although systems like Media Lengua very well might have arisen via habitual CS among mixed Quechua and Spanish speak-ing populations, the language now has native speakers who do not necessarily perceive themselves as alternating between two disparate varieties. Furthermore, CS is usu-ally seen as a distinct phenomenon from diglossia wherein a speech community utilizes two or more languages or linguistic varieties within strictly defined domains (Bullock, Toribio, 2004). For example, in German-speaking Switzerland diverse Swiss German dialects are used colloquially while standard German is used in official context. To my mind, however, this might be considered merely an institutionalized form of CS.

Social environments in which CS ﬂourishes order from the level of multilingual families, as they negotiate each member’s competencies or indicate multicultural nu-ances, to the highest level of mass societal multilingualism. Societal multilingualism can appear as a consequence of many different socio-political forces; they have arisen very often in modern contexts as a byproduct of colonialism and nation-state build-ing wherein lbuild-inguistically and culturally heterogeneous peoples have been coerced into higher orders of either real or perceived homogeneity. Since culture is malleable so too

(23)

are the multilingual scenarios that contribute to large scale CS phenomena. Bentahila and Davies (1995) discuss the evolving roles of Moroccan Arabic and French among bilinguals of both across generational divides; for instance, younger bilinguals tended to use only French lexical insertions in their Arabic while the older generation uti-lized whole spans of French constructions in addition to Arabic. This displays how the changing status of each language in the society ultimately affects proﬁciency and therefore the shape that CS ultimately takes (Bentahila, Davies, 1995).

Despite lay speakers conception that CS involves seemingly random switching, CS is always systematic. Those who make use of CS have intuitions about its usage just as monolinguals do within their repertoire. CS structures are subject to the grammar of both languages always adhering to the syntactic properties of at least one of the component systems. My anecdotal experience with English structures in Swedish are therefore no exception and support my belief that ESSD does in fact constitute an instance of CS at the cultural level.

5.5 On a Deﬁnition of Discourse

Since the analysis of ESSD entails an approach to discourse itself, it is necessary to pro-vide an exploration of the definitions of discourse. Although definitions vary among formal disciplines, I include here some discussion of the relevant properties of dis-course as it pertains to spoken language in an effort to establish a firm definition suit-able for this study.

“‘When two people have been married for years they seem to become unconscious of each other’s bodily presence so that they move as if alone, speak aloud things which they do not expect to be answered, and in general experience all the comfort of solitude without loneliness. The joint lives of Ridley and Helen had arrived at this stage of community, and it was often necessary for one or the other to recall with an eﬀort whether a thing had been said or only thought, shared or dreamt in private.’ (Virginia Woolf)

In an article, entitled Sharing, shaping, showing: the deep uses of language, George W. Turner (1989) points out this observation by the formidable Virginia Woolf regarding the language used between long-married people. This aspect of relationships is famil-iar to us, and this peculfamil-iar kind of dialogue is routinely shaped by the context in which it continuously emerges. The comfortable, sometimes too comfortable, environment of marriage molds the way people communicate with one another; this is perhaps why we can quickly tell when two people have been in a close romantic relationship for some time simply by listening to the way they speak to each other.

(24)

According to Diane Macdonell in her book Theories of Discourse: An Introduction, dialogue, even in the most abstract sense, is the central condition for discourse as all speech is inherently social in nature (Macdonell, 1986). Every dialogue between a speaker and their intended audience occurs within certain contexts. Discursive con-texts are as diverse as the world we inhabit. The language we use in private rooms with our doctors differs from the language we use with our children or late at night in a pub with our friends. Time intersects with place to affect the form of our discourses when we consider that the language we use today with our romantic partners, or colleagues with whom we are only recently acquainted, is likely quite different than the language we would have used in those scenarios in previous centuries, eras, political climates, and so forth. The likely infinite number of times and places in which a discourse can play out is proportionate to our ability to construct an infinite amount of sentences with which to build a relevant discourse. By this definition, discourse is a dialogue in a context, and the possible amount of configurations of interlocutors and possible envi-ronments where they might be communicating are inumerable. This is the definition of discourse often associated with literary and discourse studies. To bring this idea home, Macdonell cites the Russian linguist, Volosinov:

[V]illage sewing circles, urban carouses, workers’ lunchtime chats, etc., will all have their own type. Each situation, ﬁxed and sustained by social custom, commands a particular kind of organization of audience’

The endless variety of formulations a discourse can take is undoubtedly symptomatic of the richness of the fabric of human societies and relationships. Ever-changing cir-cumstances shape and reshape how we speak to each other and dynamically construct our social realities ad inﬁnitum. Macdonell also points out that discourse is always about something or pertains to certain objects of discussion often to the exclusion, and perhaps detriment, of the things the discourse is then not about. She offers the exam-ple of literature departments whose discourse tends to not be about popular literature or women’s literature; in this way, the discourse by and about certain interlocutors is ignored or perhaps never occurs (Macdonell, 1986). This clearly speaks to the inherent power of discourse.

A competing but not incompatible deﬁnition of discourse, and one more closely associated with linguistic science, is that if syntax is the way words make up sentences then discourse is the way that sentences make up larger bodies of meaning (Buckholz, 2003). If all speech entails a certain context, one might then view the notion of context as something static rather than dynamic and separable from the discourse ultimately produced. This deﬁnition is clearly more simplistic and aims to address structural

(25)

dis-course without regard for the environment in which said disdis-course originated. For the purposes of this study, however, I subscribe to an integrated deﬁnition by establishing that discourse is: a dialogue made up of sentences, or strings of utterances, that form larger units

of meaning within a given contexts. In this way, we account for interlocutors, speech, and

circumstance as the basic conditions for discourse. I say speech here as the object of this study is to analyze the varieties of English-based structures that seem to emerge in spoken Swedish.”

6 Method

6.1 Introduction

In an ideal world, ESSD would be captured as it occurs in natural spontaneous speech in real-time among participants caught unaware of our greater aims so as not to con-taminate the data as a consequence of the Observer’s Paradox (Labov, 1972). In prac-tice, however, this is quite a laborious and time-consuming task. From anecdotal ex-perience with ESSD, I developed the notion that it arises particularly among younger people (as evidence from previously cited studies on attitudes toward English in Swe-den suggests) and also that it seemed to be very playful in nature. After all, the popular opinion appears to be that English is fun.

The constraints of fun and young brought me to the internet where I sought to iden-tify domains within popular media where these structures might crop up. Youtube is a website where users can create a proﬁle, share, like, comment on, and evaluate video content, usually for free. Since 2007, the platform has offered a monetization scheme for video creators whereby they receive revenue in accordance with the amount of views their content garners; this has resulted in the mass emergence of entrepreneurial

vloggers (Burgess, Green, 2009). These vloggers are popularly known as Youtubers or Youtube celebrities. Youtubers curate their own channels often creating video content

which meets the demands of their audience base. A enormous variety of stylistic approaches to this Youtube versus Youtube-user relationship have been established. One of the most common and widespread approaches among Youtubers is the simple monologue. These monologues consist of one speaker addressing their audience of subscribers (Youtube-users who opt for subscription to a Youtuber’s channel). These Youtuber monologues are not typically pre-designed before their presentation in an actual video. Therefore, a tremendous wealth of monologues, wherein Youtubers are speaking freely on an indeﬁnite number of topics, exists on the platform. Topics vary considerably from traumatic life experiences and funny stories to popularized chal-lenges in which Youtubers must perform some bizarre or difficult task. Meeting my

(26)

initial criteria, Youtubers are often quite young, in their teens and 20’s, and the register of their video content is usually very informal and playful. Furthermore, Youtube as a platform is very international-facing. Youtubers themselves can make no plausible estimation as to the ultimate reaches of their audience base since even Youtube-users who don’t subscribe to their channel can view their content

Furthermore, the platform houses Youtubers from all over the world, and therefore has produced Youtube celebrities of diverse linguistic backgrounds. A simple search will yield video content from practically any speech community, including videos where Youtube celebrities are producing freely spoken monologues in the desired language. All of these properties combined provide a kind of Goldilocks-zone, ripe for potential instances of ESSD on the part of Swedish-speaking Youtubers. For these reasons, I have chosen to use audiovisual material from Swedish Youtubers’ monologues as the basis for my data collection.

6.2 Data Collection

Summary of Collection Criteria:

–4 hours of audio-visual material - Swedish-speaking Youtubers - At least 50,000 subscribers - Disparate topics

- Freely spoken monologues

- Appear to be in their teens or 20’s

In order to collect data that might yield results which can elucidate answers to my re-search questions, I performed analysis on four hours of Youtube video material. The predominant language of each video had to be Swedish. Altogether 20 videos of vary-ing length were analyzed each with a new Youtuber. All videos involved one speaker speaking freely and addressing an audience who was not actually present. Each video dealt with a separate topic; topics varied somewhat in tone from very playful to quite serious. Selection was based on the network of related videos and suggested channels which the Youtube platform provides.

Youtubers were chosen based on their number of subscribers, with the lower limit being placed at 50,000 subscribers. The intention here was to involve Youtube celebri-ties who have established a very broad audience. In this way, the selected Youtubers can be said to be indicative of the language habits of younger Swedish people and also

(27)

hugely influential within the Swedish-speaking community; considering that the pop-ulation of Sweden today is around 10 million, 50,000 subscribers amounts to a rather massive sphere of influence. Unfortunately, demographic information on each Youtu-ber is limited as vloggers rarely provide credible information in their profiles such as their age, gender, place of residence, etc. Although most Youtubers appeared to be very young, either in their teens or 20’s, this was mostly taken at face value. Further-more, since self-declared information regarding each Youtuber’s gender identity and ethnic background is not available, the potential salience of these factors was disre-garded.

Summary of Collection Procedure: 1 - Review selected video 2 - Record each anglicism

3 - Record immediate preceding context 4 - Record immediate subsequent context

5 - Record time at which anglicism occurs in video

This ﬁve step procedure was devised for recording instances of English structures (also referred to as anglicisms for the sake of brevity) as they occurred in the greater Swedish discourse of each video. Anglicisms emerging merely as content words or short phrases were cross-referenced in the SAOL (2014), an online catalogue of Swedish words which have been recorded by Svenska Akademiens Ordlista (The Swedish Academy’s Dictio-nary). If a word or phrase was listed in the SAOL catalogue, it was not counted as an anglicism as these could be considered oﬃcial Swedish words by lay speakers.

As seen in Image 4, each anglicism was entered into an Excel spreadsheet where each step in the procedure was accounted for with its own column. The preceding context for each anglicism was recorded next. If the anglicism appeared within a sen-tence, the initial part of that sentence was taken to instantiate the preceding context; however if the anglicism occurred after or at the beginning of a sentence boundary, the previous sentence constituted the preceding context. Next, the subsequent or pro-ceeding context was recorded. Again, if the anglicism in question cropped up within a sentence, the subsequent context was taken to be the rest or end part of that sentence as long as the sentence itself did not go on for an unreasonable amount of time such that the subsequent context to that anglicism became irrelevant; in that case, only the relevant chunk of the proceeding context was recorded. At times, signiﬁcant pauses or disturbances occurred just before or after the instances of an anglicism. In those

(28)

cases, preceding and subsequent contexts were recorded as null. Furthermore, since all of these videos were signiﬁcantly edited by their creators, very often the preced-ing and subsequent contexts of anglicisms were cut out. Again in this case those slots were left blank. Finally, the time at which each anglicism became instantiated was also recorded. This was done with the hope that some greater pattern might arise whereby certain types of anglicism might cluster at speciﬁc terminal points within video.

Image 4, Excel Data Sample.

6.3 Notes on Mode, Register, and Stylization

It should be noted that the type of discourse present in the Youtube video material reviewed for this study does not reﬂect spoken speech in its most naturalistic context. Firstly, these videos are performed as freely spoken monologues directed toward an au-dience who is not present, and the recorded content can be edited later by the creator to ensure that their message is delivered in a way that suits their interests and tastes. Moreover, typical Youtube videos often adhere to a certain style both in the physi-ological presentation by the Youtuber and in the nature of their video-editing. Fast speech, emphatic exaggeration, quick-switch, and cut-to transition edits are among a constellation of stylistic choices with which self-styled Youtubers can align themselves with their fellow Youtuber peers. This creates a kind of Youtube culture wherein play-ful, to-the-point, and fashionable overall presentation are highly valued. For these rea-sons, the general video output is very informal in linguistic register which distinguishes Youtube content from traditional media such as television wherein institutional bod-ies maintain the authority with respect to editing, censorship, and judgments about what is appropriate content for a given viewership.

The discourse which crops up in these Youtube monologues does not really re-flect everyday naturalistic speech. These videos do, however, go largely unchecked before they are posted by vloggers in terms of needing prior approval by some sort of governing institution. Instead, it is up to Youtube-users to flag inappropriate or illegal content after it is produced and posted. In this way, even though the actual mode, i.e. monologue, is not really indicative of everyday speech, the unfiltered na-ture of Youtube discourse produces language use that is very informal and, to my ear,

(29)

closely resembles the linguistic register one might use among acquaintances or close friends. Furthermore, the kind of English-Swedish CS phenomena that ultimately emerge among Swedish Youtubers is starkly reminiscent of its use in everyday spoken discourse in Sweden according to my anecdotal experience, and I would venture to guess that the instances of ESSD that I have recorded will be very familiar in nature to Swedish speakers.

7 Findings

In the four hours of material reviewed, 246 anglicisms emerged. Theses anglicisms ranged from single content words to code-switches into whole English utterances. Sev-eral rounds of analysis were performed to determine how each ESSD instance could be categorized as compared to one another and according to their functional and struc-tural features. Ultimately, five coherent categories were pinned down and a sixth was used for structures who defied any discernible pattern. Each anglicism listed in the Excel catalogue was then assigned as one of the types listed below. Coincidently, this numbering of types is also reflective of the frequency of each type within the total sample with the exceptions of types 5 and 6 as seen in Table 1.

ESSD Categories:

Type 1: content word or phrase (may have discursive properties) Type 2: utterance with content (may have discursive properties) Type 3: interjection or pure discourse marker

Type 4: anglicism with Swedish morphological features Type 5: citation in English

Type 6: other, not easily categorized

Type Type 1 Type 2 Type 3 Type 4 Type 5 Type 6 Total

(30)

7.1 Explanation of Types

As mentioned in the section on code-switching in the theoretical framework, catego-rizing instances of CS deﬁnitively can be difficult. A considerable degree of gray area arises when one sets out to break multilingual phenomena into categories. Nonethe-less, the six types of ESSD I have identiﬁed were derived to be as clear-cut as possible though, at times, I had to simply use my best judgment in assigning each anglicism to a category. I will however note instances of hesitation or indecision in the discussion section of this thesis.

7.1.1 Type 1

The vast majority of type 1 anglicisms came in the form of nouns or noun phrases as well as adjectives or adjectival phrases. Nouns and nominal phrases usually referred to very speciﬁc terminology or physical objects to which the Youtuber needed to refer. Some typical examples included: beauty blender, comfort-zone, deconstructed burrito bowl,

inﬂuencers, and wannabe. Some of these type 1 ESSD instances seemed to stand in

where obvious Swedish equivalents were not available; however, this was certainly not always the case. Vloggers also used type 1 anglicisms where clear Swedish alternatives do exist, such as with terms like pressure, struggle, and bowl. The same was also true of the adjectives and adjectival expressions that Youtubers used, however instances where anglicisms were used despite the existence of very clear Swedish equivalents was even more prominent here: for example, vloggers used adjectives such as rough,

outstanding, and insane in their monologues, all of which correspond to obvious Swedish

alternatives. Some of these did occasionally seem to function partially as discourse markers.

7.1.2 Type 2

As the second most prevalent type, type 2 ESSD instances were the most striking as these involved code-switches into whole English utterances. Type 2 ranged from short phrases to long spans of utterances in the form of both intra- and inter-sentential code-switching. The structural features of this type will be discussed in the analysis section. Many instances of this type did carry discursive properties, however, as with type 1, they stood out from pure discourse markers or interjections.

7.1.3 Type 3

Type 3 anglicisms were the most difficult to set apart from the other categories. Deﬁni-tions vary somewhat as to what should count as a pure discourse marker. For example,