Open Access

In this paper, we experiment with exploring linguistic networks as a new method in historical semantics. Our starting point is a long-term historical corpus (i.e., the Patrologia Patina) which we analyse regarding the conceptual stability of a key concept in medieval literature (i.e., virtus). Most analyses in historical semantics explore small data sets by focusing on narrow contexts of lexical usages, but we propose a more comprehensive method based on lexical networks that represent the underlying documents as a whole. We demonstrate both the topological stability of document-based lexical networks and their usefulness in providing empirical evidence in historical semantics.

The project Referenzkorpus Altdeutsch (‘Old German Reference Corpus’) aims to es- tablish a deeply-annotated text corpus of all extant Old German texts. In order to minimize manual work, an important target was to automate the retrieval of as much data as possible from existing sources. Whilst the texts themselves were already available in a digital form, the annotation data could to a large extent be found within a set of glossaries associated with each text. After digitizing these, the information contained in them could be automatically and semi-automatically linked to the texts. Subsequent manual editing focuses on any remaining gaps and misattributions, re- jecting inapplicable alternatives and adjusting details to the annotation standards of the project. Throughout the process, various problems have been encountered that require special attention to find particular solutions.

The Kassel Corpus of Clause Linking is part of a larger project on the grammar of New High German led by Vilmos Ägel. The project takes as its starting point the assump- tion that equal consideration of both oral and written language is essential in order to understand developments in grammar during the New High German period. The Kassel Corpus of Clause Linking includes four texts from the period 1650-1700 and four texts from 1850-1900, with three texts from each period exemplifying oracy (imme- diacy) and one text literacy (distance). The corpus was annotated for grammatical features which are relevant for clause linking, such as predicates, subjects and connec- tors. A major issue in the process of annotation was to identify correlations between single grammatical features and types of clause linking, such as coordination, subordination and ellipsis. In the paper we will explain the principles which formed the basis for compiling and annotating the corpus, and illustrate how correlations between single grammatical features and types of clause linking may be established. We will also provide an example (non-integrative ellipsis) of how the annotated corpus re- veals differences between immediacy and distance, as well as showing historical developments.

Although the alternation between was and were has been extensively recorded in modern varieties of British and overseas English, there is comparatively little information about the distribution of was and were in older varieties of speech. This has been largely due both to the scarcity of old regional material, and the consequent lack of dia- chronic dialect corpora. In light of this, this paper looks at some of the Lancashire texts included in the Salamanea Corpus. It examines the evidence provided by literary representations of the dialect with regard to past tense BE forms. Though largely neglected for linguistic investigation, literary samples of Lancashire English may go some way towards casting light on the forms of BE in the county between 1700 and 1900. Our aim is thus twofold: firstly, to contribute to previous research into past tense BE forms in Lancashire by adding historical data that have not been thus far considered, and secondly, to illustrate the linguistic possibilities of the corpus, arguing that it may serve as a complementary missing link to expand the database of English diachronic dialectology.

This paper draws attention to the newly available corpus resources for the study of Older Scots, and to the application of the lexical bundles method (Biber et al. 1999) in historical specialized discourse. The discussion concentrates on the method adopted from present-day corpus research, which illuminates historical questions which have so far proved unanswerable, e.g. which multi-word elements in text are stable and repetitive. I applied lexical bundles to legal and administrative texts written in Scots, to observe the degree of formulaicity in early specialized discourse. The results of the study show that the Scottish documents contain highly formulaic long lexical bundles (8-grams and 7-grams) when juxtaposed with other specialized discourse texts, such as the Bible. Similarly, shorter bundles helped to identify the impressive degree of formulaicity in comparison to speechbased legal genres, such as trials and depositions (Culpeper/Kytö 2010).