Syntax and parsing

UNCONTRACTED NEGATIVES AND NEGATIVE CONTRACTIONS IN COMTEMPORARY ENGLISH: CORPUS-BASED STUDY.

Author: CASTILLO GONZÁLEZ MARÍA DEL PILAR.
Year: 2006.
University: SANTIAGO DE COMPOSTELA [www.usc.es].
Place of defense: FACULTAD DE FILOLOGÍA.
Place of preparation: FACULTAD DE FILOLOGÍA.

Summary: The denial has been and remains one of the topics studied in the English language. However, the variation between contractions negative and their corresponding forms without entering into contemporary English has not caught the attention of many linguists, with the exception of Biber (1987), Hiller (1987), Kjellmer (1998), Biber et al. (1999), Tagliamonte & Smith (2002), Yaeger-Dror, Hall-Lew & Deckert (2002) and Castillo-González (2001, 2003). In this context, the primary objective of this dissertation is to offer, from a perspective variacionista, a more complete and thorough than is provided by these studies on the distribution of forms negative alternatives, not only in different dialects of English today, but also in different kinds of texts, both oral and written, and responding to various structural factors. This paper is divided into two main blocs. The first, cutting more theoretical, is devoted to the review of the literature on the use of contractions and negative forms without contracting. For his part, the second section focuses on the empirical study of the data extracted from nine córpora computer contemporary English: three from the British, two of whom are written, The Lancaster-Oslo-Bergen corpus of British English (LOB ) and The Freiburg-LOB corpus of British English (FLOB), and an oral The London-Lund Corpus (LLC); other three American English, two letters, The Brown University-Corpus of American English (Brown) and The Freiburg- Brown Corpus of American English (FROWN), and an oral The Corpus of American English Spoken Professional (PEGSCO), a corpus of written English Australian, The Australian Corpus of English (ACE) and two English córpora representative of New Zealand, a written, The Wellington Written Corpus (WWC), and the other oral The Wellington Spoken Corpus (WSC). The starting point for the first part of the work has been the distinction between denial clausal, namely that of the scope of the clause is any denial, and denial subclausal when it denies is just one part of clause (cf. Klima (1964), Quirk et al. (1985) and Huddleston (1995), among others). The establishment of such a distinction has proved essential, since my study focuses on an exclusive basis, in the first of these two types of denial. In contemporary English denial clausal with negative particle can not materialize in three patterns alternative: (a) First, one that shows how without contracting operator followed by not (he is not), which I have referred to in this study Uncontracted negative (UncN). (B) Secondly, the contraction operator with the subject (he's not), the so-called contraction Operator (OpeC). (C) Finally, the merger with the operator of negative particle not (he is not), which I mean as Not-contraction (NotC). A total of sixteen operators can deny the particle not in a position post-verbal: be, have, do, will, would, shall, should, can, could, must, may, might, dare, need, and ought to used to. Most of these operators admit both UncN as NotC, while the OpeC has a narrower scope and is possible only in the following cases: this be the operator, the operator will have, both present and past, and operators will, and shall would. With the aim of studying the actual change patterns among the three alternatives mentioned above, this research has focused on precisely these operators admit that the three possibilities for denial. Operators under review have had negative contractions from older periods in the history of the language until today. The most significant changes in the system of denial and contractions negative from a diachronic perspective are summarized in Section II.2. The review of the literature to which I have devoted Section II.3. This thesis has allowed me to establish a series of variables which, to a greater or lesser extent, depends distribution of forms not contractas and their 8 corresp 1ff8 ondientes contractions negative in contemporary English. Other factors include the following: (a) Type of text. As assert Fries (1940: 8), Forsheden (1983: 36), Quirk et al. (1985: 123ff) or Biber (1988: 243), among others, contractions tend to be associated with the informal language or oral texts, and often considered inappropriate in written texts, particularly in those classes of more formal, as texts scientific or official documents. So Biber (1987), in his study on the use of contractions in British and American texts from different records, detected a higher proportion of contractions in the "Fiction", relatively close to the spoken language, which in more formal texts , as may be the case of "Official Documents". Other authors, as Kjellmer (1998), Biber et al. (1999), Huddleston & Pullum (2002: 91, 800) or Yaeger-Dror, Hall-Lew & Deckert (2002: 81ff), confirm the existence of the correlation between informal texts and a higher proportion of forms contractas. (B) Dialects. The dialect differences also play a key role in the selection of the alternatives under consideration. According to this variable, the contractions seem to be more prevalent in American English than in British English, both in formal texts into categories as closest to the oral language, as a result of studies of Biber (1987), Hiller (1987) and Castillo-González (2001). The selection of two types of contractions, OpeC and NotC also seems to be tied to some extent by criteria dialect. For example, in the south of England there is a preference for the NotC on OpeC, while in northern England and Scotland tends to be used more frequently OpeC that NotC. (C) Social class, sex and age. The use of certain contractions can be indicative of social class. The contractual form that serves as a marker of social class in a more obvious is ain't. Authors like Freeborn (1986) or Jørgensen (1979), among others, regarded as a non-standard used mainly by speakers of lower classes. In regard to the gender variable, women tend to use contractions less often than men, however, as stated by Coates (1986), women using more contractions in the so-called "question tags". Finally, with respect to the variable age, we can say that the highest percentage in the use of contractions can be seen among the youngest speakers. (D) Structural factors such as type of clause or type of subject. With regard to the first of these factors, Kjellmer (1998), Westergren (1998), Biber et al. (1999) or Yaeger-Dror, Hall-Lew & Deckert (2002) argue that the contractions are more often in terms interrogativas and imperative that affirmative. As regards the use of forms contractas with those verbs that can operate as auxiliaries and lexical verbs (be and have) there is a unanimous agreement among linguists. Some authors claim that the contractions are more common when these operate as auxiliary verbs are verbs that when lexicons (cf. Quirk et al. (1985), Sinclair (1990) or Biber et al. (1999)), while others, such as Philips & Reynolds (1987), argue that with contractions be dominated when it is a copulation. Also, the type of subject also appears to be determinant in the distribution of the variables under consideration, as it forms no contractas benefits are subject to more complex (nominal phrases, clauses, etc.). While contractions are more frequent with pronominales subject. Here, the frequency with which two or more elements appear together in the speech, the so-called "string frequency" with Krug (1998), seems to have a decisive influence, and that greater frequency of a given sequence tends to be associated with a greater proportion of contractions. With regard to the distribution of the two forms contractas, the literature mentions the prevalence of OpeC on NotC with all operators except be (cf. Quirk et al. (1985), Biber et al. (1999) or Kortmann ( 2003), among others). However, the distribution of the two alternatives contractas seems to depend to a large extent on factors mentioned by Hiller (1987), Kjellmer (1998), Biber et al. (1999), Tagliamonte & Smith (2002) and Yaeger-Dror, Hall-Lew & Deckert (2002), such as the type of text, geographic variation, sex, structural factors or criteria fonológicos (cf. Section II.3.2.3. ). As commented at the beginning of this summary, the second part of this dissertation focuses on an empirical study of the data extracted from the nine córpora computerized mentioned above (LOB, FLOB, BROWN, FROWN, ACE, WWC, LLC, PEGSCO and WSC). The first sections of this part of the thesis describes, in short, each of these córpora, as well as the problems which I have found in the stages of extraction, classification and analysis of data. Most of these córpora containing a total of about 1,000,000 words and a similar structure, with different kinds of texts that illustrate different levels of formality. The only exceptions are the ACE, which were selected only those categories that are common to other writings corpora (SR), around 926,000 words; LLC, which contains only 500,000 words and the PEGSCO, which has been selected texts representing all categories included in the body by a total area of 1,000,000 words. Therefore, the total corpus used for the elaboration of this thesis round of 8,426,000 words. To obtain the data I used the program WordSmith Tools version 3.0, focusing (as previously justified) operators who allowed both UncN, as OpeC and NotC. Therefore, the forms have been analyzed as follows: am not, are not, is not, 'm not,' re not, 's not (for both is to you), ain't, are not, isn 't have not, has not, had not,' is not, 'd not (had both would like to), have not, has not, had not, will not,' ll not (both will for shall), will not, shall not, shan't, would not and would not. The total score of examples for these 24 ways has risen to 24,708. However, not all cases these examples can be considered free of variation between the three alternatives denial (cf. Section III.2.). We have excluded contexts such as: (1) examples of denial subclausal; (2) forms neutralized between two operators; (3) forms of negative operator shall, as from the historical point of view, the contraction 'll up to and will not shall; (4) forms of negative had better and would rather, as there Quirk et al. (1985) and Denison (1998), the use of NotC these expressions is restricted to terms interrogativas negative; (5) other cases that do not fall into any of the previous groups. The total excluded examples in the section III.2. Totals 956. Once discarded cases mentioned above, have been analyzed a total of 23,752 examples, of which 9,151 (38.53%) are UncNs, 5,423 (22.83%) OpeCs and 9,178 (38.64%) NotCs, which means that, globally, the material used for this dissertation shows a preference for contractions versus negative ways not contractas. However, only 19,067 of these examples freely allow the election of at least one of two ways merged under review, while the remaining 4,685 cases have been classified as contexts KO (cf. Section III.3.2.). For example, the OpeC is not possible in a wide variety of structures, including the following: "yes-no questions", "wh-questions" investment subject and the verb, "question tags", subordinate clauses conditional investment subject and the verb, cláus 8 ulas cook 1ff8 rdinadas in which the subject of the second clause is omitted, cases where the subject is not present, those examples involving the third person singular present indicative of be or have on the subject ends in-s, examples in which the subject and operator are separated by some other element clausal, where the speaker emphasizes both the operator and in the particle or those negative examples where there is a break between the subject and operator. For its part, NotC is not possible in the so-called "split VPs" or in cases where the speaker emphasizes the subject and operator or pauses between it and the particle negative. After analyzing individual examples of those groups that do not allow free choice of the three alternatives under consideration, the remaining cases (19,067 in total), have been considered taking into account various factors, such as dialect (British English, American, Australian and New Zealand), medium (written language in front of oral language), the date of composition of the texts (the decade of 1960 compared to that of 1990 in British and American English writing), registration (formal versus informal), Class operator (be, have, will, and would), type of subject (pronoun, nominal phrase, clause, there existential and others) or the potential influence of the so-called "string frequency." The main findings that he obtained from the analysis of each of the nine córpora are as follows: (a) In the written texts, without contracting forms predominate in general on ways contractas. However, those texts written closer to the oral language, ie the categories of fiction, show a preference for negative contractions versus their corresponding forms without contracting. The only exception is the category M ( "Science fiction"), whose behavior is closer to that of text types more formal and J ( "Learned and scientific writings"). In the case of oral texts, the most formal and the informal, favoring the use of contractions. (B) detects a preference for UncNs with all operators in all written texts with the exception of FROWN, which favors this option only with the verb be. In oral córpora again, the contractions are more common than their counterparts not contractas regardless of the operator. (C) In terms of the variation between the two types of accounts, OpeC and NotC, the former tend to be used more frequently to be the operator, in both oral and written texts, thus confirming the above by authors such as Dillard (1980), Freeborn (1986), Hughes & Trudgill (1996) or Anderwald (2002). However, in American English the years 1960, the trend is going in the opposite direction, as there is a higher proportion of NotCs with this operator. These data show that the authors mentioned above, can, at times, fall into generalizations do not always realize the great linguistic diversity that exists. (D) With respect to the use of contractions with those verbs that can operate as auxiliaries and lexical verbs (be and have), the data from my study revealed heterogeneous performance of the various córpora. Thus, the verb lexicon be conducive contractions, rather than auxiliary be written at all except LOB córpora and BROWN, ie for the years 60, and American English Oral (PEGSCO). On the contrary, the operator have nine córpora agree show a higher frequency of use of the forms contractas when operating lexicon as a verb. In regard to the distinction between assistant be as progressive and as assistant passive, the trend is to find a higher percentage of forms contractas with the first of these uses. However, the proportion of NotCs is to be greater than the liabilities be progressive. (E) With regard to the data provided by each body based on the type of subject, the overall trend is that the contractions, particularly NotCs, are preferred subjects with pronominales and with the existential there, while the forms not contractas are normally associated with more complex types of subject, as subjects nominal or clausales. Also, the importance of frequency in the range of negative variants has been confirmed in this view, since very common sequences as he is not, favoring contractions more than others is not as a man who rarely appear repeated the material used. Once analyzed data for each body of an individual (see Sections III.3.4. And III.3.5.), Have established a series of comparisons from different points of view (cf. Section III.3.6.), Which it has enabled me to come to the following conclusions: 1. From a diachronic perspective, a comparison between, on one hand, the LOB and FLOB, both British English, and, secondly, the Brown and FROWN of American English suggests that, despite the fact that UncNs are option predominant in the written language contractions have experienced a remarkable progression from the years 1960 to the decade of 1990. In fact, the forms contractas have increased their use of considerably not only in informal texts, as the categories of fiction, but also more formal categories such as "Learned and scientific writings" (Cat J). As for the variation between forms contractas, British English says a predominance of NotCs face OpeCs with all operators except be, both in the years 1960 and in 1990. The dominance of the NotC with all operators except be too visible in American English in the 1990s (FROWN) compared with data from BROWN, with material from the mid-1960's, where the verb be also favors NotC . 2. Taking into account factors dialect, we can say that in English written in the four dialects studied, without contracting forms predominate on ways contractas. However, the frequency of contractions is highest in American English than in British English, Australian or New Zealand. By contrast, in the three dialects oral texts surveyed favor the use of contractions in more than forms without contracting. However, the proportion of contractions in American English is lower than in British English or New Zealand, due to the greater degree of formality of the texts present in the PEGSCO. In relation to the variation between the two types of contractions in the oral texts, the role of dialect is also critical, since in English and American English in New Zealand the OpeCs predominate over NotCs, while the latter is the preferred option in English British. 3. In regard to the medium (written language in front of oral language), as has already been demonstrated in the preceding paragraphs, córpora writings promote the use of non-standard forms contractas, while the latter prefer the use of contractions, even in the more formal texts. Finally, the analysis of the nine córpora selected for the elaboration of this thesis has allowed me to establish further comparisons between the data obtained from my study and the authors who have paid attention to the same subject, although to a lesser extent ( cf. Section III.3.7.). For example, my data confirm the results obtained by Biber (1987) and Kjellmer (1998) in relation to the distribution of the two variants of denial in written texts, in which the dichotomy formal texts versus informal texts plays a role decisive. Moreover, data from my research only partially overlap with those of Hiller (1987), as the tendency for the OpeC prevail on the NotC with the operator not be confirmed in American English writing of the years 1960. Finally, my data not ratify all afirmacione 8 s made 5df by Biber et al. (1999) for British English, since the texts included in the category "News" in my favor research forms without contracting. In short, this dissertation contributes to offer a broader view that the existing so far in the patterns of distribution of forms negative contractas and without entering into contemporary English operators who allowed the three alternatives of denial, UncN, OpeC and NotC taking as a starting point the analysis of texts from different decades, dialects, records and means. However, this work is still far from providing a general description of the subject matter. For example, the analysis of other variables such as age, sex or social stratification, among others, should expect to be tackled in future research.

SYNTAX AND PARSING