Effects of Prescriptivism on the Pronunciation of American English

Table of Contents

1. Introduction

2. Prescriptivism and descriptivism
2.1 Standardization of American English as a social and historical process
2.2 Standard American English and its use

3. The superiority of one pronunciation

4. Prescriptive rules on pronunciation
4.1. G-Dropping
4.2. Dissimilation of the (post)alveolar approximant /j/
4.3 Unreleased plosives in consonant clusters
4.4 Omitted /t/ in <-nt->-sequences

5. Methodology

6. Analysis
6.1 Examination: G-dropping
6.2 Examination: Dissimilation of the voiced postalveolar approximant /j/
6.3 Examination: Unreleased plosives in consonant clusters
6.4. Examination: Omitted /t/ in <-nt->-sequences

7. Outlook/Conclusion

8. References

9. Appendix
9.1 Summary: Phenomena of g-dropping
9.2 Summary: Phenomena of dissimilation of the (post)alveolar approximant /j/
9.3 Summary: Phenomena of unreleased plosives in consonant clusters
9.4 Summary: Phenomena of omitted /t/ in <-nt->-sequences

1. Introduction

Contemporary linguistics works according to descriptive rules. However, language ideologies on syntax, punctuation, and phonetics developed and supported by prescriptivists, are embedded in society's mind. Especially the pronunciation seems to be in the center of attention when, for example, radio broadcasters are being judged on the "correctness" and "standardness" of their pronunciation (Crystal 2019: 1-2). Nevertheless, speakers are considered to be passive followers of prescriptive rules, ergo, that those rules do not have any or much effect on the actual use (Ammon 2015: 65). This raises the question what the effects of prescriptivism on pronunciation are.

To analyze the effects, a similar version of the methodology presented by Kroch and Small (1978) will be applied for this analysis. They compared the syntax of two sociolinguistic groups, the hosts and the callers in an all-talk radio show. The hosts were thought to follow prescriptive rules more closely than the caller group, due to their social status and role in the radio show.

In this thesis, nationwide phonetic phenomena found in the United States of America will be analyzed. Namely, g-dropping, dissimilation of the voiced postalveolar approximant /j/, when /j/ occurs frequently in a word, unreleased plosives in consonant clusters and omitted /t/-sounds in <-nt-> sequences.

Instead of a radio talk show, interviews of the daytime talk show Megyn Kelly TODAY will serve as the corpus for this analysis. The television talk show host is Megyn Kelly and the show focuses on social topics. Megyn Kelly TODAY episodes are most easily found on YouTube on NBC's TODAY channel.

For the analysis, a transcription of the different interviews had to take place, as the corpus is in video format. Only the words and segments that contained the specific phonetic characteristics were phonetically transcribed according to the IPA system.

In the first chapter, the terms prescriptivism and descriptivism will be explained. The following section will give a brief introduction to the standardization of American English as a social and historical process. This will give a better idea of how prescriptive rules are conveyed and portrayed. Additionally, this section gives possible answers as to why prescriptivism is still embedded in society's mind. Section 2.2 goes into more detail on what standard American English is and when it is used. “The superiority of one pronunciation” explains when the “better” pronunciation can be expected, what social groups tend to speak in this accent, and which situations usually demand a more formal pronunciation. Chapter 4 describes the different phonetic phenomena that were chosen for this analysis. Finally, the phonetic occurrences are analyzed and evaluated.

2. Prescriptivism and descriptivism

Prescriptivism gives rules on how to use language “correctly”, on the basis of historical, logical or aesthetical criteria (Lexikon der Sprachwissenschaft 2002: 475). Or as Garner (2009: XXXVII) puts it: “prescriptivists give guidelines to language users on how to handle words as effectively as possible.” Typically, these rules are applied to the lexicon, syntax, punctuation, and phonetics and are the language rules that are taught in schools (Curzan 2014: 16).

In today's linguistic research descriptivism is practiced. So rather than prescribing rules on how language should be used, the focus is on how language is used by speakers (Garner 2009: XXXVII). In contemporary linguistics, prescribing language is seen as “an unrealistic goal” because language is not seen as something static, but as always changing (Curzan 2014: 2). Furthermore, Dubinsky and Holcomb (2011: 166-167) state that the field of linguistics does not include prescriptive approaches in linguistic analysis, as it is seen as an ideology by which authorities regulate the linguistic norms, even though other dialects or rather language varieties have working language systems of their own. With a descriptive approach to language, nobody is discriminated against who is not educated enough to know or follow the prescriptive rules (Dubinsky and Holcomb 2011: 166-167).

The approaches in the fields of linguistics have developed into descriptivism and distanced themselves from prescriptivism to the extent that prescriptive rules are not considered to have any effect on spoken language at all (Milroy and Milroy 1985: 4). However, those rules that identify a “better” language use, are still embedded in the minds of language communities. The general public accepts the prescriptive rules and adapts the prescriptive view of a “correct” and “incorrect” way of speaking (Milroy 1999a: 22). As prescriptive approaches to language are still established in today's society, it does not seem to be ideal to entirely ignore prescriptivism in linguistic research. The inclusion can give insight into how these prescriptive rules influence speech (Kroch and Small 1978: 45-55).

2.1 Standardization of American English as a social and historical process

It is necessary to investigate how prescriptive rules are conveyed and portrayed to understand why prescriptivism is still embedded in the minds of many people.

Standardization preserves the rules that underlie the standard language variety and makes sure that prescriptive rules exist throughout all levels of language (e.g. grammar, pronunciation, etc.). The varieties that are spoken by higher social classes and more educated people are thought to be “worthy” of becoming standardized (Dubinsky and Holcomb 2011: 168-169). Individuals and institutions that are in possession of power standardize language (Lippi- Green 2012: 22, 61).

An important part of standardization is codification. In this process vocabulary and grammar are recorded and analyzed in dictionaries and grammar books, a method first used for the English language in the eighteenth century (Thomas and Wareing 1999: 190). In this century, the aim of standardization was to show that even though language seems arbitrary, there is an underlying system, to give clear rules on language to avoid disputes and to make language users aware of common mistakes (Crystal 2000: 116).

Aside from prescriptivists, school teacher and social and political leaders preserve language norms to the extent that teachers especially will interrupt students to inform them about having used a “wrong” expression or pronunciation, etc. (Wolfram and Schilling 2016: 325-326). Those habits at school must be one of the reasons why the speech of the lower class is said to be less affected by standardization, as they typically have less access to education than speakers in the middle and upper class (McDavid and O'Cain 1977: 101).

The standardization process can become an issue for those speakers of a language, who do not master this variation and are therefore discriminated against. Nevertheless, standardization is efficient in terms of communication in areas like business, politics, education, etc. Specifically in education, it has several benefits, as a language can be taught with one example, without having to discuss any other language varieties (Dubinsky and Holcomb 2011: 168­169). These positive effects might be the reason why standardization is still in use, and therefore, the idea of a “standard” language is still ingrained in society.

2.2 Standard American English and its use

The following section will give an overview of the language variety that is considered the standard in the U.S. and the situations in which the standard is expected to be spoken.

In the eyes of prescriptivists and the public, the standard variety is seen as the most prestigious and most formal realization of a language. In contemporary linguistics, the “standard language” is seen as another variety, rather than the prescriptive belief of the standard being the better or correct language (Meyerhoff 2011: 18). To draw a complete picture of the standard variety, it is noteworthy that a distinction between a formal and an informal standard variety is made. The formal one is described as “codified, prescriptive, and relatively homogeneous, whereas the informal standard is more subjective, somewhat flexible, and tends to comprise a continuum.” (Wolfram and Schilling 2016: 323). In this bachelor thesis, the formal standard variety gives the ground for the analysis, as this variation underlies the prescriptive rules.

The Received Pronunciation (RP) is the known standard variety in the United Kingdom, a regionless accent, which is found in public settings, like television and radio (Trudgill and Hannah 2008: 15-16). The standard in the United States of America proves to be harder to pinpoint. Network Standard is usually seen as the standard American English variety. The variant does not consist of “any stigmatized features, as well as regionally conspicuous phonological and grammatical features” and is the most prestigious variant (Wolfram and Schilling 2016: 324). This standard is thought to be mostly found in the Northern Midwest U.S.A. and among middle-aged, educated, middle and upper class native American English speakers (Milroy 1999b.: 174-175).

Whether the standard variant is used, also depends on the situation in which an individual is expressing oneself through language. In formal settings, the tendencies for using the standard variety is much higher. For example, at work in a business meeting or at university when presenting a project. Generally, the more public a situation is. For example, on TV or radio, the language variety of the speaker becomes more that of the standard. Furthermore, the discussed topics influence the way one speaks. If the topic is emotional, the speech style tends to have fewer characteristics of the standard variety than when speaking about a less emotional subject. However, when individuals express themselves, the linguistic features available to them vary, depending on different social factors, for example, class, ethnicity, etc. Not everyone has the same ability to speak the standard variety (Wolfram and Schilling 2016: 282-283).

3. The superiority of one pronunciation

Garner's (2009: 665) prescriptive rules on pronunciation state that the “best” pronunciation is used by educated speakers, as they speak closest to the standard variety. In descriptive terms, pronunciation gives clues on social class and education. Furthermore, pronunciation is important in order to be understood by others and to show self- or group-identity (Crystal 2019: 8-9).

Some sociolinguists argue that the term “standard English” can only refer to features of grammar and vocabulary, as the spoken standard variety is always influenced by accents, so can therefore not be applied to the field of phonetics (Trudgill and Hannah 2008: 4). The term accent refers solely to pronunciation and varies from individual to individual depending on the geographical area and social factors, such as class (Lippi-Green 2012: 44-45). Considering that every person has an accent (Lippi-Green 2012: 44-45), it might be linguistically correct to exclude phonetics when speaking about the standard variety. Nevertheless, the fact that there is much evidence of speakers being judged on the “correctness” and “standardness” of their pronunciation, much more than on what they are saying, makes it clear that phonetics cannot be taken out of the discussion when talking about the standard English variety (Wolfram and Schilling 2016: 11).

Especially in the public eye, for example, in visual and audio media, the Network Standard is expected. However, even in those situations, the pronunciation varies due to the individual's accent (Crystal 2019 1-2). Crystal (2019 1-2) describes working at the BBC Radio 4 in the 1980 s , where many listeners sent in concerned letters about the standard of the language spoken by the radio broadcasters. While what they were saying was perfectly understandable, listeners judged the aesthetic as inferior, when the radio broadcaster did not pronounce certain terms “correctly”. Thus, the pronunciation was more important than what was said. It is noticeable that the listeners were influenced by prescriptive norms when judging the utterances of the broadcasters.

4. Prescriptive rules on pronunciation

According to prescriptive rules, only the Network Standard pronunciation is “correct” and should, therefore, be used in the United States (Wolfram and Schilling 2016: 324). The focus of this analysis will be on prescriptive rules that apply to phonetic occurrences that are represented throughout the United States and are not region-specific. This approach ensures that the phonetic phenomena could occur in the speaker's speech style, no matter the regional origin.

In the following sections, four national phonetic phenomena that are considered non-standard are described. Namely g-dropping, the dissimilation of the post(alveolar) approximant /j/, unreleased plosives in consonant clusters and omitted /t/ in <-nt>-sequences. To determine whether prescriptivists reject a certain pronunciation, often Garner's Modern American Usage: The Authority on Grammar, Usage, and Style (2009) and the online version of the Oxford Learner's Dictionary is consulted, to determine the prescriptive status of an utterance. The dictionaries are exemplary of other prescriptive approaches to language.

4.1. G-Dropping

G-dropping is a widespread non-standard phenomenon that occurs in word-final unstressed <-ing>-constructions, for example, when [kisiq] is pronounced as [kisin]. The term g-dropping suggests that the word-final <g> is “dropped” when pronouncing words ending in <-ing>, but this process is not really omitting a sound, rather the substitution of a sound for another: the voiced alveolar nasal /n/ for the velar voiced nasal /q/. According to Coupland (2007: 22-23), this phenomenon is stigmatized and associated with speakers that are less educated and part of the lower class.

The word-final phonetic occurrence can be expected to provide many examples, considering that this is a phenomenon that is found throughout the U.S. and with the knowledge that words ending in <-ing> are often used in the English language.

4.2. Dissimilation of the (post)alveolar approximant /j/

Another focus of this thesis will be on the voiced (post)alveolar approximant /j/, that occurs frequently in a word. Some speakers leave the sound out entirely, when /j/ follows a vowel (Wolfram and Schilling 2016: 53-54), for example when “surprise” is pronounced as [sepjaiz] instead of [sejpjaiz]). According to Wolfram and Schilling (2016: 54), this nationwide phenomenon is called dissimilation and is a process where two sounds with similar or the same features become more distinct. Garner (2009: 156, 666) underlines the negative connotation of the dissimilation of /j/, without specifically naming the phonetic phenomenon. Mentioned are words with a frequency of /j/, for example, “mirror” [mirer] and “deteriorate” [ditijiejeit] and the specification that those words shall not be pronounced as [mir] and [ditijieeit]. Obviously, the dissimilation of the voiced approximant /j/ is prescriptively stigmatized and can, therefore, give hints on the effect of the prescriptive rules on pronunciation on the speakers.

4.3 Unreleased plosives in consonant clusters

In spoken language, word-final plosives (/p,t,k,b,d,g/) are at times unreleased when the following word begins with a plosive as well (Sauer 2011: 30). Sauer (2011: 30) gives the example of “dead dog” which can be pronounced as [ded dog]. In this phonetic context, the word-final voiced alveolar plosive /d/ is unreleased (indicated by o), as there is a “stop closure but no release [of the airflow]” (Eddington 2007: 29).

However, in the online version of the Oxford Learner's Dictionary, the “correct” pronunciation of “dead” is transcribed as /ded/. When listening to the voice recording of the pronunciation of the adjective, the word-final plosive /d/ is pronounced with a released airflow, so [ded] and not [ded ]. Thus, the pronunciation of the plosive does not match the “correct” pronunciation prescribed by the dictionary.

In this analysis, the focus will be on word-final consonant clusters, whose last sound is a plosive and the following word starts with a plosive as well.

4.4 Omitted /t/ in <-nt->-sequences

Garner's Modern American Usage: The Authority on Grammar, Usage, and Style (2009: 156, 666) mentions another negatively connotated pronunciation.

The voiceless alveolar plosive /t/ in “intellectual” shall be pronounced as [intelektfuel] and not as [inelektfuel]. Actually, the omitted /t/ is an allophone of /t/, so a possible realization of the phoneme /t/ (Sauer 2011: 28). This realization can occur in specific phonetic contexts. But for this analysis, the focus will be on the phoneme /t/, when it follows the voiced alveolar nasal /n/ and precedes a vowel, as in “twenty” or the voiced alveolar liquid /l/, as in “differently.” In these sequences, the voiceless alveolar plosive /t/ can be pronounced as [twenti] and [difrentli], or the /t/-sound is omitted as in [tweni] and [difrenli]. While both pronunciations are found and used throughout the United States (Sauer 2011: 28), Garner (2009: 156, 666) proves that the omitted variation of the phoneme /t/ is stigmatized and will, therefore, be included in the analysis.

5. Methodology

The following chapter will explain the methodology, the choice of corpus and the basic idea and aim of the thesis. The necessary terminology for the analysis is mentioned in the preceding chapters.

The overall aim of the thesis is to find out what the effects of prescriptivism are on the pronunciation of American English. Kroch and Small (1978: 44-48) have researched the effect of standard norms, but with a focus on the syntax of spoken language. To do so, they have analyzed several hours of an all-talk radio show in Philadelphia. They compared the speech style of two sociolinguistic groups: the hosts/ host guests, and the callers. The hosts and guests were thought to speak more closely to the Network Standard than the caller group, due to the host's social status and “role in the radio talk-show [.] [and] generally greater loyalty to and concern for the norms of the standard” (Kroch and Small 1978: 48). By analyzing specific syntactic factors, they were able to prove that the caller group spoke, in fact, closer to the standard variety and that the “grammatical ideology is the cause of the intergroup difference” (Kroch and Small 1978: 52).

In this analysis, a similar version of the methodology presented by Kroch and Small (1978) will be applied, but with a slight change in corpus choice and a focus on the phonetic characteristics mentioned in the preceding chapter, instead of syntactic features.

Instead of a radio talk show, interviews of Megyn Kelly TODAY published by the National Broadcasting Company (NBC), a New York City-headquartered television network (Koblin 2017), will serve as the corpus for this analysis. Megyn Kelly TODAY episodes can be found on NBC's website, but since the daytime talk show was canceled in October 2018 (Nemetz 2018), the episodes are most easily found on YouTube on NBC's TODAY channel. The television talk show was hosted by Megyn Kelly and focused on social topics. Usually, individuals who have experienced something traumatizing or generally something emotional were interviewed live by Megyn Kelly on the incident (Koblin 2017).

Only episodes with interviewees with an American English accent were chosen for the analysis, to avoid phonetic phenomena that occur due to second language acquisition, and to avoid phenomena that might be considered standard in other English-speaking countries, but not in the United States. Furthermore, the interviewed individuals should not be considered “famous”, as it can be expected that those individuals are trained speakers and would therefore not be representative of a speech group that is allegedly less affected by prescriptive rules.

Moreover, only the introduction by the host, the prerecorded interview and the live interview will be analyzed. A short segment where the host reads a statement is also included, as it is part of the live interview. In the prerecorded interviews, scripted voiceovers of the host group occur, where additional information is given. As only spontaneous speech is relevant for this thesis, the voiceover is not included in the analysis.

The corpus choice by Kroch and Small (1978) is different from the corpus choice in this thesis. Therefore, the methodology by Kroch and Small (1978) cannot be applied exactly. Megyn Kelly falls into the “host and guest group” and the individuals who share their story are the “caller group”. The terminology to refer to the different groups for this analysis will be “host group” when speaking about Megyn Kelly, and “guest group” for the interviewed individuals.

While Megyn Kelly has been a controversial figure and reason for many social debates in the last year, it is important to point out that the daytime talk show was chosen for its appropriateness for the analysis and not for any opinions that the host has shared in the past.

Several reasons influenced the choice of Megyn Kelly TODAY as the corpus material. One factor was that the social characteristics of the individuals who are allegedly more affected by standardization also apply to the hostess of the daytime talk show. Megyn Kelly is educated, as she successfully studied political science and enrolled in law school. After having worked as a lawyer, she switched careers and became a correspondent for FOX news (Thomas 2015). Her occupations do not include any manual labor, rather she is working with her intellect and linguistic skill, thus one can categorize her as being part of the higher-middle class (Chambers 2003: 41-42).

A second factor is, that the episodes were broadcasted nationwide, and as mentioned in chapter 3, a “better” pronunciation is especially expected in the public eye. Therefore, the hostess and the guests will likely speak closer to the Network Standard. By using the Network Standard, it can be ensured that no regional stigmatized features are used, that might be frowned upon or not understood in other parts of the country.

Furthermore, the topics discussed on Megyn Kelly TODAY, are not expected to influence the hostess' speech style, as the topics are emotional and personal for the interviewee. But technically, they are not for the interviewer. The discussed issues demand a level of seriousness and formality from the hostess, due to their emotional value for the guest. Therefore, it can be expected that Megyn Kelly speaks the Network Standard variety.

To analyze the different phonetic characteristics of the speech style of the interviewer and interviewee, several episodes of the talk show had to be transcribed. Only the words and segments that contained the specific characteristics mentioned in chapter 4 were phonetically transcribed according to the IPA system and added to the appendix in chapter 9.

Finally, the transcription of the actual pronunciation in the interviews will be compared to the prescriptively “correct” pronunciation of the specific word or word combination. The online version of the Oxford Learner's Dictionary was consulted to determine the prescriptive pronunciation. By comparing the two pronunciations, a statistic can be developed, on how often the speakers pronounce the chosen phonetic occurrences according to prescriptive rules.

Then the two groups can be compared, and a conclusion can be drawn on whether one group shows more features of the prescriptive rules than the other and possible reasons for the occurrences. When presenting the findings, a descriptive approach will be applied to demonstrate the effects of prescriptive rules.

6. Analysis

In the following analysis, the transcribed interviews of the Megyn Kelly Today show will be analyzed, focusing on the phonetic factors mentioned in chapter 4. G-dropping will be the first element, following with dissimilation, unreleased plosives in consonant clusters and omitted /t/ in <-nt>-sequences. In the respective categories, the host group will be analyzed first, followed by the guest group, and lastly a comparison of the two.

6.1 Examination: G-dropping

In the analyzed interviews, a total of 193 <-ing>-sequences are found, out of which 83 sequences are uttered by the host. Four out of these 83 findings cannot be used for the analysis. Two of these four findings are not audible due to the applause of the audience (A.2.1.12 and A.3.1.54). Another reason to not include certain examples is that although a pronunciation does not conform with the prescriptively correct pronunciation, g-dropping is not apparent (A.1.1.4 and A.4.1.64).

Thus, 79 world-final <-ing>-sequences are relevant for this analysis and found in three different sections of the interview: the introduction, the interview, and when reading the statement. While in the introduction and when reading the statement all word-final <-ing>-syllables are pronounced “correctly” from a prescriptive point of view, all prescriptively “incorrect” pronunciations are found in the actual live interview.

A total of eight instances of g-dropping are found, so about 10.12% of the 79 relevant <-ing>-sequences are not pronounced according to prescriptive rules.

Three out of the eight prescriptively “incorrect” pronunciations could be categorized as a characteristic of a phonetic phenomenon known as assimilation, namely A.4.1.67, A.4.1.68, and A.4.1.80. Nevertheless, the word­final /q/-sound is replaced with a /n/-sound, and therefore the three examples are included in the overall analysis as examples for g-dropping, as it cannot be definitely determined whether assimilation or g-dropping takes place.

The host's utterance in the second interview “[...] you were talking about something [...]” shows another phenomenon. The host pronounces “talking” as [tokin] (A.2.1.27) and “something” as [sAmSiq] (A.2.1.28). Thus, in one sentence the speaker pronounces prescriptively “correctly” and “incorrectly.”

[...]