Definitions
One of the methods by which a language can be analysed and studied is corpus linguistics, which “in its broadest sense encompasses corpus-based language research” (Utvić, 2013), i.e. use of corpus to analyze language. A corpus is a collection of texts, written or spoken, stored in a computer (O'Keefee, McCarthy, Carter, 2007), online in the cloud1, on web or in books, so it can be defined as a systematized collection of a natural language (Nesselhauf, 2005; Reppen, 2011). The Glossary of Language Technologies2 defines a corpus as a “written or spoken language resource collected and annotated with the purpose of: analyzing a language to determine its properties, analyzing human behavior (in the sphere of language use) in certain situations, training the system to adapt its behavior to specific language circumstances, empirical testing of a language theory, creating a test for a language-engineering technique or an application of the technique to see how it functions in practice.” The same glossary also gives a definition of computer corpora for which it claims that they are “encoded in a standard and consistent way with the intention of keeping them open for computer searches.” Information and communication technology plays a key role in development of corpora because “the creation of computer technology has made corpus research possible, as the computer is able to store, code, categorize, and retrieve massive amounts of information” (Durand, 2018, p. 132).
The term originates from the Latin word “corpus, corporis meaning a body, a whole, totality, a set, an almanac” (Utvić, 2013, p. 1). The term “corpus linguistics” was created in the 1980s (Leech, 1992) and it is was first mentioned by Aarts and Mejia in 1982 in the paper “Grammars and Intuitions in Corpus Linguistics”, and it appeared again in 1984 in the book “Corpus Linguistics I: Recent Developments in the Use of Computer Corpora”. In 1991, on an international symposium of British, Dutch, Swedish and Norwegian linguists, a new group of researchers under the name corpus linguists was formed. From 1996, they have also been publishing their own journal - The International Journal of Corpus Linguistics (Leech, 1992; Utvić, 2013). Utvić (2013) thinks that the authors of the definitions of corpus linguistics are stating their opinions about corpus linguistics and based on that treat it as a tool, method, methodology, methodological approach, discipline, theory, theoretical approach, theoretical or methodological paradigm or a combination of all of the above. He also claims that corpus linguistics is today considered to be primarily a methodology or a group of methodologies, not a separate theoretical discipline in the field of linguistics and also wonders whether or not corpus linguistics represents something more than mere methodology (Utvić, 2013). More detailed overview of definition and discussion on what corpus linguistics is can be found in Taylor(2008) “What is corpus linguistics? What the data says” where author states that “corpus linguistics is a tool, a method, a methodology, a methodological approach, a discipline, a theory, a theoretical approach, a paradigm (theoretical or methodological), or a combination of these” (Taylor, 2008, p. 180).
Corpus linguistics, apart from analyzing language with the use of corpora in its studies, wants to answer two important questions: “Which specific patterns are connected to lexical and grammatical features? and How do those patterns differ inside different corpora?” (Bennett, 2010)
As one of the leading figures during the rise of corpus linguistics, Leech (1992) states that computer corpora are not random collections of texts (p. 107). They are usually collected with a certain goal in mind and should be considered representative of certain types of text. According to Sinclair (2004), a corpus is “a corpus is a collection of pieces of language text in electronic form, selected according to external criteria to represent, as far as possible, a language or language variety as a source of data for linguistic research” (p. 16). Taking into consideration all of these definitions of a corpus, McEnery and Wilson (2001) conclude that a corpus should possess four main characteristics: sampling and representativeness, size, machine sensibility and standard reference.
In the book “Using corpora in the language classroom”, Reppen (2011) claims that a corpus is a large compilation of texts in a natural language stored in some type of electronic form, e.g. online or on computer. Author separately considers and explains the definition by claiming that the term “natural language” denotes a language in real everyday use, such as newspaper articles, letters, books, conversations etc. The part of the definition stating that a corpus is a coherent collection of texts refers to corpus design. For example, if the subject is a described language, then the corpus should contain different categories like fiction, academic prose, personal letters, memoirs, literature etc.