The Routledge Handbook of Language and Digital Communication

Alexandra Georgakopoulou, Tereza Spilioti, Alexandra Georgakopoulou, Tereza Spilioti

• Publisher: Routledge
• Year: 2015
• ISBN:
  9781317439295
• Edition: 1
• Topic: Langues et linguistique
• Subtopic: Études sur la communication

The Routledge Handbook of Language and Digital Communication provides a comprehensive, state of the art overview of language-focused research on digital communication, taking stock and registering the latest trends that set the agenda for future developments in this thriving and fast moving field. The contributors are all leading figures or established authorities in their areas, covering a wide range of topics and concerns in the following seven sections:

• Methods and Perspectives;

• Language Resources, Genres, and Discourses;

• Digital Literacies;

• Digital Communication in Public;

• Digital Selves and Online-Offline Lives;

• Communities, Networks, Relationships;

• New debates and Further directions.

This volume showcases critical syntheses of the established literature on key topics and issues and, at the same time, reflects upon and engages with cutting edge research and new directions for study (as emerging within social media). A wide range of languages are represented, from Japanese, Greek, German and Scandinavian languages, to computer-mediated Arabic, Chinese and African languages.

The Routledge Handbook of Language and Digital Communication will be an essential resource for advanced undergraduates, postgraduates and researchers within English language and linguistics, applied linguistics and media and communication studies.

Part I Methods and perspectives

1 Approaches to language variation

Lars Hinrichs

DOI: 10.4324/9781315694344-1

Overview

The study of linguistic variation

In linguistics, the term variation refers to the phenomenon of heterogeneity of linguistic form. If the language faculty of the human brain were a simply programmed implementation of rules, one would expect all users of a language to use the same linguistic forms each time they express the same concept or grammatical relation. But they do not: there are many sites in the linguistic repertoires of speakers and writers that force them to choose between alternative ways of expressing something; most of the time, they make these choices automatically, without giving them much thought. For example, in English the possessive relation between two noun phrases can be expressed using either the of-genitive or the ’s-genitive. Speakers and writers are forced to make a choice between those alternatives every time they express the possessive relation between two noun phrases; linguistic study has shown that this choice is conditioned systematically by multiple competing factors (Hinrichs & Szmrecsanyi 2007). Similarly, writers of digital discourse can choose between standard spelling such as <you, aren’t> and alternative spellings such as <u, arent>. They can also choose whether or not to insert features specific to Computer-Mediated Discourse (CMD), such as emoticons, into their writings.

The study of variation is one branch of what Susan C. Herring circumscribes in her framework for the study of online language known as computer-mediated discourse analysis (CMDA, Herring 2004). Variation occurs in digital communication at the micro and the macro levels of linguistic analysis, and at all intermediate levels. Likewise, there are different types of cause of variation. In variationist research, the causes of variation are referred to as factors, and they are grouped, at the most fundamental level, into internal and external types. Internal factors are planes of variation related to the linguistic system. They include grammatical categories, as well as – in multilingual or multidialectal digital communication – the choice of linguistic variety itself. External, or social, factors include properties of the interactional context such as age, gender, or ethnicity of the participants, etc.

This chapter first introduces the two subfields of linguistics that most directly inform the study of variation in digital discourse, namely variationist sociolinguistics and corpus linguistics, and then provides an overview of research on variation in online language. The following discussion makes practical suggestions for researchers planning their own studies on issues of variation in digital discourse, and the chapter concludes with some reflections on future developments in the field.

Variationist sociolinguistics

Broadly, language variation is an umbrella concept that refers to difference of linguistic form in any of a number of dimensions, for example: differences between one individual’s ways of speaking across different situational parameters; differences between multiple individuals’ ways of speaking; differences between the language of speech communities located in different areas, i.e. geographical variation; or differences between social groups (e.g., ethnicities, genders, generations).

The fact that language is not always the same everywhere, or that it is fundamentally heterogeneous, is a common observation in daily life. Accents, ways of speaking, and linguistic idiosyncrasies draw attention to themselves in most people’s daily experience. In the 1960s, William Labov initiated the field of sociolinguistics when he observed that language variation is in many cases systematic rather than random. In the formulation of the foundational article he co-authored with Uriel Weinreich and Marvin Herzog, language displays ‘orderly heterogeneity’ (Weinreich, Labov, & Herzog 1968: 100). The field of language variation studies is dedicated to discovering and describing the regularity and order of that heterogeneity in language. For example, Labov’s well-known study of variation in New York City English (1966a) demonstrated the correlation between the social status of speakers and certain forms of speech, proving the existence of regular, systematic correspondences between higher socioeconomic status and lower frequencies of usage of linguistic features that are socially valued as informal.

The central theoretical construct of language variation studies is the linguistic variable (Cheshire 1987; Labov 1966b; Labov 1978; Wolfram 1991). The term refers to any unit of the linguistic system that can be realized in more than one possible shape in linguistic production. To reintroduce an example mentioned above, Standard English has more than one way of expressing the concept of possession linking two noun phrases, i.e. two alternative ways of constructing genitives: the of-genitive (the plan of the administration) and the ’s-genitive (the administration’s plan). In many cases, both genitives are equally possible choices, but it can be shown that variation between the two occurs according to several independent constraints that are statistically significant: if the possessor noun phrase denotes an animate entity, the ’s-genitive is the more likely choice (e.g. the man’s car); if the possessor noun phrase ends in a sibilant, the of-genitive is the more likely choice (the angle of the pass), and so on (Hinrichs & Szmrecsanyi 2007). At the level of phonetics and phonology, linguists may study variation between different pronunciations of the same phonemic speech sound that are all possible according to the rules of the language variety being spoken. For example, in all English-speaking parts of the world, words ending in –ing may be pronounced with the final sound realized either as [ŋ] (reading) or as [n] (readin’). Generally, the variant ending in [ŋ] has been shown to correlate with more formal social features (of the speaker’s identity, the situation, the topic, etc.) and the [n]-variant with informal features.

In the early account by Weinreich et al. (1968), language variation studies were conceived of as a necessary addition to Chomskyan linguistics, which had by then come to dominate much of linguistic discussion. Transformational grammar presupposes that the human language faculty is a regular system of rules that translates the meanings that speakers wish to express from abstract representations to linguistic forms that are grammatical in the language being spoken. The study of this translational apparatus is the object of Chomskyan linguistics, and for a long time during the initial phase of the field’s establishment, its data relied heavily on made-up linguistic examples that were grammatically correct and possible according to the linguist’s own judgments. In other words, Chomskyan (or formal) linguistics emerged as a reflection on language that took an abstract, ideal state of language as its object of study. Spoken language as it was actually used in the speech community was dismissed as unimportant, and variation as noise in the data. But early protagonists of language variation studies pointed to a key fact: over time, language variation leads to permanent changes in the grammar of a language (this is the fundamental claim of Weinreich et al. 1968; for elaboration see Labov 1972). Because the responsible study of syntax must therefore incorporate an understanding of language change, as variationists argued, and because language change results from variation, variation and change have come to be viewed as two sides of the same coin.

The quantitative paradigm of language variation studies, whose backbone is sketched above, is also known as the ‘Labovian’ type of sociolinguistics. In recent years it has been complemented by several influential studies that add ethnographic and qualitative insights to an overall Labovian research design. In an influential review, Eckert (2012) describes three waves in the methodological development of linguistic variation studies. The first wave departs from Labov’s work on variation in New York City (Labov 1966a), which inspired similar studies of variation throughout the Anglophone world (e.g., Trudgill 1974; Wolfram 1969), but also in non-English-speaking locales such as Tehran (Modaressi-Tehrani 1978). It marked a fairly radical departure in linguistics because (i) it established beyond doubt that linguistic variation is structured and systematic; and (ii) it ‘introduced a new quantitative empiricism into linguistics, with supportive theoretical underpinnings’ (Eckert 2012: 88).

Another of Labov’s studies initiated what Eckert considers the second wave of variation studies. His work on Martha’s Vineyard investigated the link between variation as part of an ongoing sound change and individual identity within the community (Labov 1963). It became an influential model for studies that combined the quantitative angle with ethnographic methods in an effort to gain better insights into the links between local practice, identity, and linguistic variation. Studies in the second wave include Milroy (1980), Cheshire (1982), and Rickford (1986).

The third wave of variation studies takes the ethnographic approach even further, placing ethnography fully at the centre of its methodological paradigm. Unlike the first two waves, which were interested in describing social structure and in correlating dynamics of variation with social categories, the third wave aims to explain variation by making the ‘social meaning of variation’ its overall descriptive goal. In the third wave, the ethnographic method is used to discover linguistic practices and the ways in which they contribute to the construction of locally meaningful social styles. Therefore, the social unit of analysis is no longer the speech community, as it was in the first two waves. Rather, third wave ethnographies are located in communities of practice (Eckert & Wenger 2005; Holmes & Meyerhoff 1999), i.e. social units which exist and constantly re- construct themselves over time around a common and concrete purpose. For example, Eckert (2000) studied the ways in which high school students participated in social styles that stood in locally meaningful contrast to each other, i.e. the ‘jocks’ and the ‘burnouts’. Variation was shown to systematically contribute to an individual’s stylistic identity within, near, or outside of one (or both) of these two macro-identity categories. The community of practice, in this case, was the school where the research was conducted. Other studies in the third wave include Bucholtz (2011), Campbell-Kibler (2007), Mendoza-Denton (2008), and Podesva (2007).

Corpus linguistics and language variation

The sociolinguistic approach is not the only way of looking at language variation. Another tradition that studies variation – and is of particular relevance to written digital language – is corpus linguistics (Biber 2012).

A milestone achievement in the founding days of corpus linguistics was the creation of the Brown corpus and the subsequent publication of Kucˇera & Francis (1967). Brown is an early example of a digitized collection, consisting of 500 samples of texts written in American English and published in 1961, i.e. representing a specific time slice. The methodological innovation championed by corpus linguistics lay in its strongly empirical approach to language studies, grounded in the study of vast amounts of natural language data, which points to an obvious intellectual kinship with sociolinguistics. Even early corpora were often grammatically annotated in some way, for example through part-of-speech tagging (see Hinrichs, Smith, & Waibel 2010 for a discussion of Brown and three more recent corpora, as well as their levels of annotation).

The major difference between the sociolinguistic and the corpus linguistic approach to variation lies in what each takes as its primary unit of quantitative analysis. Variationist studies focus on individual instances of linguistic forms, while corpus linguistic approaches quantify the number of occurrences for a certain form throughout a text, or across an entire collection of texts (Biber 2012: 12) – usually in order to provide a basis for comparison with other texts, or corpora of texts. A sociolinguistic approach to variation between variant A and variant B of variable X would compile data in long form, which implies keeping a record of each instance of variable X that was encountered in the data, and recording not only its realization as A or B (i.e. the form in which the dependent variable occurred) but also a number of contextual factors (i.e. the independent variables). By contrast, a study in corpus linguistics would typically count all occurrences of A and of B in a given section of their data and then compare these frequencies with those in another subset of data; this way, only one data point would be recorded for the first sub-corpus and another for the second. This form of quantitative data recording is called aggregate-level, or short-form, data collection. For example, one might compare the relative frequencies of two variants of one grammatical form, A and B, in two corpora representing texts from 2000 and 1950, normalizing frequencies in case the sizes of both samples should differ. The differences between the two sets of frequencies would provide a first approximation of language change over time.

The next section discusses how the linguistic interest in variation has been applied to, and developed in, research on digital discourse.

Foci of research on variation in digital discourse

In contrast to the focused research programme of foundational variation studies, which emerged out of an explicit engagement with formal linguistics, research on variation in digital language incorporates a remarkable range of applied research interests and goals. The following thematic overview traces the conventional hierarchy of levels of linguistic analysis. It proceeds from the micro-levels of typography and orthography, through morphology and lexis at the word level, to syntax at the unit of the sentence/utterance. Further, variation at the level of the linguistic system is considered by discussing code-switching in online contexts.

Phonetic/phonological variation in digitally transmitted speech has not been widely studied to date, though steady improvements in audio-visual transmission as well as text-to-speech interface technology suggest that such research is imminent (Jenks & Firth 2013).

Typographic variation can be broadly conceived of as the alternation between typographic uses that are accepted in conventional printed language, and those that are not. In contrast to orthography, typography refers to the use of non-alphabetic symbols such as numbers, punctuation, and special symbols. Some of the non-standard, or innovative, uses that occur frequently in CMD have become iconic of online writing in popular perception. A representative list of these features (see Crystal 2006 for an introduction) includes:

• Innovative uses of lower and upper case letters, aS iN tHiS eXaMpLe.
• The use of emoticons, such as:-),;-), or:-@ to represent emotive expressions – here, a smile, a smirk, and an angry face, respectively. Dresner and Herring (2010) show that the pragmatic force of emoticons is not simply to convey information about emotional states, but also to signal the illocutionary force of utterances.
• The optional use of traditionally required punctuation marks. For example, contracted forms that in standard practice require an apostrophe to mark elisions, such as can’t or isn’...