This paper describes the construction and annotation of the Late Latin Charter Treebank, a set of three dependency treebanks (llct1, llct2 and llct3) which together contain 1,261 Early Medieval Latin documentary texts (i.e., original charters) written in Italy between ad 714 and 1000 (about 594,000 tokens). The paper focusses on matters which a linguistically or philologically inclined user of llct needs to know: the criteria on which the charters were selected, the special characteristics of the annotation types utilised, and the geographical and chronological distribution of the data. In addition to normal queries on forms, lemmas, morphology and syntax, complex philological research settings are enabled by the textual annotation layer of llct, which indicates abbreviated and damaged words, as well as the formulaic and non-formulaic passages of each charter.

When exploring diachronic corpora, it is often beneficial for linguists to pinpoint not only the first or the last attestation dates of certain linguistic items, but also the moments in which they become more strongly established in the corpus or, conversely, the moments in which they, despite still being part of the language, become obsolete. In this paper, we propose an algorithm to assist the identification of such periods based on the frequency of items in a corpus. Our simple and generalisable algorithm can be used for the investigation of any linguistic item in any corpus which is divided into time-frames. We also demonstrate the applicability of our method using lexical data from the Corpus of Historical American English (coha), providing case studies on the statistics and characteristics of words that appear in or disappear from this corpus in different periods.

The use of corpora in translation studies has risen dramatically over recent years, contributing towards a growing body of empirical research focussing not only on what differentiates translated from non-translated language, but also on the changes or shifts that translators make from source to target texts. Most of these studies are centred on sub-sentential elements, such as the contrastive use of particular lexis or grammar. However, translation shifts can transcend the level of the sentence. For example, sentences can be split or joined in translation, or there can be complex shifts that combine the two. While there is some research on sentence splitting, there do not seem to be many studies about sentence joining, or indeed sentence splitting and joining together. This study seeks to address this gap. Using a bi-directional parallel corpus of Portuguese and English fiction, over 90,000 source-text sentences and their corresponding text in translation were analysed from a quantitative perspective, and a closer look was taken at a sample of over 1,000 parallel text segments involving sentence joining and splitting. The main findings were that in both translation directions (1) there was a strong tendency for sentence preservation, (2) the differences between sentence splitting and joining were not significant, and (3) changes in sentence boundaries were predominantly associated with the standardisation or normalisation of syntax and a tendency for explicitation.

This paper presents a corpus-based approach to Charles Dickens's use of direct thought presentation in his fifteen novels, an aspect that is traditionally under-explored, perhaps due to his characters’ lack of psychological depth. Thanks to the use of corpus tools such as WordSmith Tools 7 ( Scott, 2016 ) and Corpus Linguistics in Cheshire (CLiC), stretches of direct thought have been effectively retrieved and systematically scrutinised. The 244 stretches of direct thought presentation identified here have revealed meaningful patterns in form and function so far as Dickens's use of this mode of presentation is concerned. Not only will the analysis of these patterns give us a better understanding of Dickens's style, but it will also reinforce the findings discussed in Busse's (2010) study of direct thought presentation in nineteenth-century English fiction – the most comprehensive analysis to date.

The aim of this paper is two-fold: (1) to demonstrate the usefulness of corpus-driven methods in determining a research focus, and (2) to show that ‘people’-focussing is a key feature of UK and Ghanaian parliamentary debates. Identifying one's research focus in discourse analysis after collecting a large amount of data can be challenging. It is relatively easier to pinpoint the items and focus of the research when the researcher employs a corpus-based method, since he or she normally has a theory to validate, refute or refine. However, when the researcher has assembled his or her data without having in advance specific theories, features, themes and concepts relating to the discourse that he or she intends to study, it becomes difficult to identify the essential issues on which to concentrate and explore. Perhaps this is even more challenging when dealing with parliamentary debates data given the wide-ranging subjects and alternative courses of action deliberated by parliamentarians. In such a situation, a corpus-driven method can be highly useful. As part of a major study of parliamentary discourse, this paper demonstrates how a corpus-driven method can help in determining one's research focus in the study of parliamentary debates.

Although ‘stance’ and ‘evaluation’ are closely related theoretical constructs, they have typically been analysed in fundamentally different ways. Stance is normally described in terms of explicit lexico-grammatical features. As a result, stance has been studied through corpus-based methods that result in generalisable descriptions of different registers. In contrast, evaluative language has been regarded as implicit and context-dependent. As a result, analyses of evaluative language have focussed on the connotations of particular words and phrases, or on detailed descriptions of particular texts (rather than on generalisable descriptions of a register).Against this background, this study is motivated by an apparent contradiction that emerged from a recent study of register variation on the web (Biber and Egbert, 2017, forthcoming): end-users identified two categories of opinionated/persuasive documents on the web, one labelled ‘Opinion’ (OP) and the other labeled ‘Informational Persuasion’ (IP). However, a multi-dimensional analysis of register variation indicated that only the OP documents were marked for the use of grammatical stance features, while the IP documents were marked for the absence of those features.Our primary goals in this paper are two-fold. First, we undertake detailed linguistic analyses of lexico-grammatical stance features in OP and IP registers to confirm the general results of the earlier analysis. And, second, we use keyword analysis to explore the possibility that evaluation might be expressed lexically – rather than grammatically – in IP. In our conclusion, we discuss the broader theoretical implications of the study, including the possibility of using keyword analysis to explore a particular discourse function (i.e., evaluation); the possibility of analysing evaluative language in ways that can be generalised to an entire register; and the possibility of two complementary discourse systems (i.e., grammatical stance versus lexical evaluation) that are used as alternatives for the expression of personal attitudes and assessments.

This paper presents an approach to classifying large web corpora into genres by means of Functional Text Dimensions (FTDs). This offers a topological approach to text typology in which the texts are described in terms of their similarity to prototype genres. The suggested set of categories is designed to be applicable to any text on the web and to be reliable in annotation practice. Interannotator agreement results show that the suggested categories produce Krippendorff's α at above 0.76. In addition to the functional space of eighteen dimensions, similarity between annotated documents can be described visually within a space of reduced dimensions obtained through t-distributed Statistical Neighbour Embedding. Reliably annotated texts also provide the basis for automatic genre classification, which can be done in each FTD, as well as as within the space of reduced dimensions. An example comparing texts from the Brown Corpus, the BNC and ukWac, a large web corpus, is provided.

This paper uses corpus methods to support the analysis of data collected as part of a large-scale ethnographic project that focusses on inter-religious relations in south-west Nigeria. Our corpus consists of answers to the open questions asked in a survey. The paper explores how people in the Yoruba-speaking south-west region of Nigeria, particularly Muslims and Christians, manage their religious differences. Through this analysis of inter-religious relations, we demonstrate how corpus linguistics can assist analyses of text-based data gathered in anthropological research. Meanwhile, our study also highlights the necessity of using anthropological methods and knowledge to interpret corpus outputs adequately.We carry out three types of analyses: keyness analysis, collocation analysis and concordance analysis. These analyses allow us to determine the ‘aboutness’ of our corpus. Four themes emerge from our analyses: (1) religion; (2) co-operation, tolerance and shared communal values such as ‘Yoruba-ness’; (3) social identities and hierarchies; and (4) the expression of boundaries and personal dislike of other religious practices.

In this paper, I investigate online film subtitles from a quantitative perspective, treating them as a separate register of communication. Subtitles from films in English and other languages translated into English are compared with registers of spoken and written communication represented by large corpora of British and American English. A series of quantitative analyses based of n-gram frequencies demonstrate that subtitles are not fundamentally different from other registers of English and that they represent a close approximation of British and American informal conversations. However, I show that the subtitles are different from the conversations with regard to several functional characteristics, which are typical of the language of scripted dialogues in films and TV series in general. Namely, the language of subtitles is more emotional and dynamic, but less spontaneous, vague and narrative than that of normally occurring conversations. The paper also compares subtitles in original English and subtitles translated from other languages and detects variation that can be explained by differences in communicative styles.

Grammatical elements such as determiners, conjunctions or pronouns are very evenly dispersed across natural language data. By contrast, the uses of lexical elements have a stronger tendency to occur in bursts that are interspersed by long lulls. This paper considers two alternative explanations for this difference. First, it could be hypothesised that the more even distribution of grammatical elements is merely an effect of an element's high text frequency. Alternatively, it could be argued that a more even distribution is a symptom of greater generality in meaning. In order to assess the impact of both frequency and semantic generality, we conducted a corpus-based study that contrasts lexical and grammatical elements in Present-Day English. Our results suggest that evenness of dispersion is chiefly an effect of high frequency.