The concept of `diversity' has been one of the main open issues in the field of multiple classifier systems. In this paper we address a facet of diversity related to its effectiveness for ensemble construction, namely, explicitly using diversity measures for ensemble construction techniques based on the kind of overproduce and choose strategy known as ensemble pruning. Such a strategy consists of selecting the (hopefully) more accurate subset of classifiers out of an original, larger ensemble. Whereas several existing pruning methods use some combination of individual classifiers' accuracy and diversity, it is still unclear whether such an evaluation function is better than the bare estimate of ensemble accuracy. We empirically investigate this issue by comparing two evaluation functions in the context of ensemble pruning: the estimate of ensemble accuracy, and its linear combination with several well-known diversity measures. This can also be viewed as using diversity as a regularizer, as suggested by some authors. To this aim we use a pruning method based on forward selection, since it allows a direct comparison between different evaluation functions. Experiments on thirty-seven benchmark data sets, four diversity measures and three base classifiers provide evidence that using diversity measures for ensemble pruning can be advantageous over using only ensemble accuracy, and that diversity measures can act as regularizers in this context.

In this paper we propose cryptosystems based on subgroup distortion in hyperbolic groups. We also include concrete examples of hyperbolic groups as possible platforms.

A class of Xorshift Random Number Generators (RNGs) are introduced by Marsaglia. We have proposed an algorithm which constructs a primitive Xorshift RNG from a given prim- itive polynomial. We also have shown a weakness present in those RNGs and suggested its solution. A separate algorithm also proposed which returns a full periodic Xorshift generator with desired number of Xorshift operations.

The article investigates the usage of female occupation terms in English by discuss-ing a sample of terms denoting female physicians attested in the COCA corpus. Differences are examined between connotations of synonymous Sex+Profession compounds, such as woman doctor and lady doctor. The issue is considered whether such N+N combinations should be treated as appositional compounds or as attribu-tive compounds.

The Lotus Sutra (or Lotos Sutra) is a very important book for Buddhists because it claims to report the teachings of Buddha (Siddharta Gautama), the founder of Buddhism. It seems to go back to the 3rd century B.C., but English translations were only made from the late 19th century onwards, the two most recent ones by Watson (1993) and Reeves (2008). Judging from those two versions, the Lotus Sutra is not only a religious, but also a strongly rhetorical text, and binomials (word pairs) are one of the rhetorical figures that are frequently employed; a few examples are: births and deaths, clean and spotless, receive and retain. The binomials used by Watson and Reeves are in the focus of the present study. Among other things I give a brief definition of binomials (which can be extended into multinomials, such as birth, old age, sickness, and death) and provide a sketch of scholarship on binomials. I discuss their formal properties, e.g. their word-classes (mainly nouns, less frequently adjectives and verbs), the connection of their elements (mostly and, less frequently or), their basic structure as well as extended and reduced structures, and their morphological makeup. As far as their etymology is concerned, there are combinations of native words (births and deaths, body and mind), loan-words (causes and conditions, receive and retain), and combinations of loan-word plus native word (supreme and wonderful, soft and gentle). As far as meaning is concerned, there are three main groups, i.e. binomials that show synonymy (fine robes and superior garments, joy and delight) or antonymy (births and deaths, body and mind, good and bad); or various kinds of complementarity (leader and teacher, soft and gentle, etc.); I also discuss cultural aspects of binomials. Furthermore I look at the sequence of the elements and factors that determine or infl uence that sequence. The comparison of Watson and Reeves also shows that frequently one translator uses a binomial where the other does not, and even in passages where both have a binomial the wording is often different, but there are also some instances where both translators use the same binomial.

In this study we look closely at the set of dorsal fricatives in Ukrainian with the purpose to explain their distributional peculiarities and phonological behavior. The analysis focuses on certain phonetic characteristics and the phonological represen-tation of the segments in question. Some other issues addressed in this study include the segmental strength which is calculated by the complexity of a segment and the presence of headed elements, the consonant-vowel interactions in which the segments in question participate and the impact of vowels on the distribution of back fricatives in Ukrainian. The main aim of this study is to propose and justify a particular phonological representation of the back fricatives in Ukrainian.

The article explores the attitudes of Polish students of English towards their target language pronunciation. The data collected through a questionnaire are analyzed in correspondence to two major grouping variables: specialization (teacher vs. translator) and the level of studies (BA vs. MA level). In particular, the paper reports on five aspects of advanced learners’ opinions and beliefs about L2 pronunciation: concern for pronunciation accuracy, self-efficacy beliefs about pronunciation learning, attitudes to pronunciation instruction, pronunciation learning goals as well as strategies. Additionally, comments are made regarding the respondents’ attitudes to their target language accent and to the presence of native features in their English pronunciation.

This study is prompted by the recent economic situation in Europe, calling for special attention to the relationship between family economic problems and the well-being of adolescents. It seeks to shed more light on the experiencing of financial concerns in adolescents who face the challenge of the foreign language learning process in the context of formal education. The participants of the empirical research were 609 stu-dents from 23 randomly selected classes of the six secondary grammar schools in Opole, southwest Poland. The results prove that financial strain and worries are negatively related to achievement in the field of foreign language learning. However, the social nature of language learning, as well as the students’ personality profi les, attachment patterns or perception of social support in adolescence also play a role.

The article discusses the late Middle English replacement of the ordinal number other by the Romance loanword second. The major cause of the change was the ambiguity and polyfunctionality of the older native word. The study is based on the language material from the Dictionary of Old English Corpus, the Middle English Compendium and the Anglo-Norman Dictionary

The aim of this article is to examine the development and status of LEST – the most common subordinator introducing negative purpose clauses in Middle English. After presenting the relevant nomenclature of the subject and the etymology of the original structure, I analyse different meanings of LEST, i.e. avertive, in-case, apprehensive and apprehensional epistemic functions as well as its structural development throughout the Middle English period. The data for this study are drawn primarily from the Penn-Helsinki Parsed Corpus of Middle English because of the chronological order of the texts included which should enable tracing potential developments of the studied expression.

The paper deals with the rise of French sequences where two etymologically and functionally divergent markers, an erstwhile refl exive se and the l’un l’autre cluster, combine to encode reciprocity. Unlike in other old Romance languages, in medieval French (12th to 14th centuries) cumulative reciprocal constructions arise relatively late. In the meantime, they rely either on l’un ... l’autre (acting as substitutes of verbal arguments) or on se. The only exception involves pronominal verbs of movement that encode symmetric relations. Two independent mechanisms are claimed to have been responsible for broadening the domain of cumulative sequences: the weakening of anaphoric links of l’un ... l’autre to their antecedents and subjectifi cation. Since plural subject-verb agreement is obligatory in sentences with reciprocal se-verbs, the singular l’un is unlikely to fill a slot for subject. Therefore, it is coerced into drifting rightwards to form a single syntactic constituent with l’autre. Subjectification circumscribes cases where, notwithstanding the presence of l’un l’autre, se accompanying a verbal plural is suffi cient to convey reciprocity. As a consequence, cumulative sequences become widespread in 16th century. All in all, the shift is viewed as an instance of construction becoming increasingly schematic over time.

Weak verbs in Modern English are sometimes mistakenly identifi ed with regular verbs. Although most weak verbs are indeed regular, there remain some which belong to the irregular group, for example spend, put, make, burn. Apart from drawing a clearer distinction between strong and weak verbs in relation to the regular and irregular division, the aim of this paper is to explain where the irregularity of these irregular weak verbs comes from and to gather possible relicts still present in Modern English. The paper discusses 56 such irregular weak verbs without vowel alternations and 9 archaisms preserving traces of such infl ection. The 56 irregular weak verbs are divided into groups according to the patterns they display and they are additionally marked depend-ing on whether: (1) they have less common irregular preterite and past participle forms, which can be labelled as “literary”or“poetic,” (2) they are literary themselves, (3) they have irregular preterite and past participle forms chiefl y in North American English, (4) they have regular variants. The initial plan included all irregular verbs but the extent of the problem coupled with the editorial limitations as to the size of the paper led to the following decisions: first, to exclude the irregular strong verbs and save them for later analyses, and then, in the remaining irregular weak ones, to remove all those with vowel alternations (like keep, seek, lose, say) and to concentrate on the verbs without vowel alternations (investigated in the present paper). It is hoped that the aims of this paper as well as their realization can serve to make the content of the historical grammar course more meaningful to students by linking it to the problems present in Modern English and to the why-questions related to the subject of their studies.

In late medieval England learned medicine leapt the walls of universities and became available to people with no formal medical training (cf. also Jones 1999, Jones 2004). This widespread interest in medicine was partly triggered by the vernacularisation of medical writings. This process involved, among other things, (1) gradual evolution of conventions and norms for, e.g. recipe writing (cf. Carroll 2004) and/or (2) employment of various strategies to adapt the texts to the new audience.The study will attempt to explain what strategies were employed to adapt medical texts, in particular recipes, to the intended audience, i.e., “who speaks [writes] what language to whom and when” (Fishman 1979: 15). For instance, some recipes contain foreign (mostly French and Latin) or sophisticated terminology whereas other recipe collections make use of vernacular resources. This implies that the language of medieval recipes might be the indicator of a social distinction between the readers. The data for the paper come from the Middle English Medical Texts (MEMT), a computerised collection of medical treatises written between 1330 and 1500.

The present paper deals with a late medieval culinary collection, Liber Cure Cocorum. The collection differs from the other known culinary manuscripts of the time due to its being written in verse. Altogether the poem consists of 137 recipes and four other fragments which introduce four sections of the collection: pottages, sauces, roast foods and ‘small cookery’. Most of the instructions included in Liber Cure Cocorum are known from other medieval collections, written in prose (cf. Hieatt 2006). In the article the collection will be analysed from two perspectives. First, the struc-ture of culinary poems will be discussed in order to examine the degree of their compliance with the traditional model of the medieval recipe. Next, although the authorship of the collection is anonymous, we will try to reveal who its author was and whom he meant as the target audience. For this purpose, we will pay attention to fragments in which the author directly refers to himself and/or to the potential reader. Additionally, any details included in the particular recipe components which might expose the potential poet and/or the audience will be discussed. By looking closely at the structure of the recipes and the intended audience, we will try to an-swer the question why it was written in verse rather than in prose.

Most theoretical and empirical studies concerning intercultural communication seem to neglect the fact that numerous interlocutors’ speech and nonverbal behaviour is intentionally impolite and motivated by their racist and/or xenophobic views, which becomes visible in contact with representatives of other cultures, nationalities or countries. Such speakers’ behaviour may take the form of physical attacks, while their linguistic performance may include verbal/nonverbal signs of their prejudiced stance. Linguistic performance which is motivated by aggressive nationalism, intolerance or discrimination is termed hate speech. The available scientific literature on this topic is scant, except for some studies fuelled by the theoretical assumptions of critical discourse analysis. There are apparently no clear and precise criteria what constitutes an act of verbal and nonverbal violence. The authors of this paper aim therefore at portraying characteristic features of hate speech in Polish on the basis of articles collected throughout 2014 and 2015 from newspapers, magazines and Internet sites. The analysis took place as part of the RADAR project devoted to conducting research on hate speech directed towards migrants in Europe. The findings indicate there is a wide repertoire of words and phrases that are visible in Polish journalism and which promulgate racial, national or ethnic hate.

The height of the fundamental frequency (F0) is frequently cited as the major acoustic feature that distinguishes the female voice from the male voice. Women tend to have a signifi cantly higher mean F0 than men. Some studies have indicated that the variability of the fundamental frequency also differs between the two genders and women’s voices may involve higher prosodic explicitness than men’s voices. This study investigates the way in which these features are utilised in rendering the voice of male and female characters in the reading aloud of fiction. To achieve this aim, a representative sample of dialogues selected from audiobooks was analysed acoustically. The results reveal that the reader’s F0 tends to slightly increase in fragments with female characters, but other predictions have not been confirmed. There is no decrease of F0 in dialogues with male characters and, in general, the reader’s variability of F0 seems not to be influenced by the character’s gender.

This paper adopts a constructional approach to grammatical structure (Goldberg 1995; 2006) and a corpus-based method for investigating pairs of semantically similar constructions and the lexemes that occur in them. The method, referred to as distinctive-collexeme analysis (Gries and Stefanowitsch 2004a), is used to determine which of the lexemes occurring in two constructions are most distinctive for either of them. On the basis of the case study concerning the imminent-NOUNconstruction versus the impending-NOUN construction, the paper demonstrates that one construction attracts some nouns more strongly than the other does. Moreover, the results of the distinctive-collexeme analysis of this pair of constructions reveal that there are clearly distinctive collexemes for each of the two constructions, and that the frame-constructional semantics is a contributing factor in the selection between these two patterns.

The aim of the present study is to investigate the relationship between explicitation and directionality in simultaneous interpreting. Given that explicitation in this mode of interpretation is often triggered by the constraints inherent in the process of interpreting, it has been hypothesized that explicitating shifts might be more frequent in retour, which is considered to be more demanding. The study is both product- and process-oriented, relying on recordings and transcripts of interpreting outputs as well as retrospective protocols. The participants in the study were 36 advanced interpreting students. The analysed forms of explicitation range from cohesive explicitation (e.g. adding connectives, reiteration, etc.), through substituting nominalisations with verb phrases and disambiguating lexical metaphors, to inserting explanatory remarks. The present paper is aimed to be a pilot study for a larger project in progress.

In this article particular emphasis is laid on the compounds which constitute parts of the titles of the press articles which were published in the online version of the Süddeutsche Zeitung newspaper between 1 August 2016 and 31 October 2016. The examples selected from the author’s research corpus make it possible to state that title is the kind of text which is highly open to word games and linguistic puns. The research conducted demonstrates and explains the role that compounds – especially occasionalisms which are parts of the press article titles under analysis – play in terms of assigning functional tasks to the given press titles (headlines). The research corpus gathered for the purpose of the analysis in the present article, i.e. selected press article titles, serves as a means of describing not only general linguistic traits of the said titles from the semantic and syntactic point of view, but also, and to an even larger extent, their sociolinguistic, i.e. pragmatic functions. Particular atten-tion is drawn to the role of compounds in performing the communicative function assigned to press article titles.

The article deals with the role of discursive temporality, namely, the necessary time that is needed for the aspect-time conceptualisation of scenes conceived of as be-longing to the past and expressed by linguistic temporality, which in turn is closely associated with the choice of forms and indicators of time. This analysis of selected discourse of narration samples demonstrate an important connection between imagery (on the cognitive and conceptual level), register type (especially mediated register) that is realized in a discourse, and the coexistence of the verb forms that go beyond the so-called norm.

Metadiscourse has been extensively studied in various genres, e.g. newspaper discourse, casual conversation, textbooks, advertisements, and research articles. Studies focusing on metadiscourse in research articles often omit analysing abstracts and traditionally investigate research articles only according to the IMRAD structure. This paper explores metadiscoursive elements in Czech and English research article abstracts in philosophy and medicine at two levels of analysis, interlingual and interdisciplinary. The aim is to investigate whether scientific writers of research article abstracts identify more with their cultural identity or whether their identity is rather discipline-specific. The theoretical framework adopted in this study is a tax-onomy of metadiscourse markers proposed by Dafouz-Milne (2008) since it takes into account a functional differentiation of metadiscourse elements. The interlingual analysis reveals no major cultural distinctions, the interdisciplinary analysis proves that metadiscourse is more prevalent in humanities. Thus, we can conclude that academic writers of RA abstracts identify more with their disciplinary culture.

The law is not effective in forcing the advertisers to communicate complete and reliable information about advertised products: the mere inclusion of written disclosures does not guarantee their successful perception, let alone comprehension. TV commercials are especially troublesome in this respect, because not only are the disclaimers presented in microscopic font size, but their display time challenges the perceptive abilities of even the most effi cient readers, not to mention the distracting effect of the simultaneously presented visual images and voice-overs. Yet, their effective comprehension could very often significantly modify the overall message and in this way lead to altered (often in a way unfavourable for the advertiser) persuasive effect. The study presents the ways in which the small print components interact with the information presented by means of the more conspicuous linguistic and pictorial items in a collection of 20 Polish TV commercials for erection stimulants. In particular, attention is given to the potential changes in the communicated message in the conditions of full comprehension and intake of the small print items.