On Reduction in English: What the English Don’t Say

. The aim of this article is to examine which sounds are most often omitted in official oral pub - lic performances by native speakers of British English. Such terms as reduction , elision (and its types), and connected speech are explained; the literature cited is concerned with elision of vowels (triphthong smoothing being treated separately), of consonants, and of whole syllables. The study presents the results of an analysis conducted on selected material available on the Internet – this comprises three British English oral performances of the total length of approximately 20 minutes. With regard to the nature of the data, they were divided into scripted, semi-scripted, and unscripted samples. The research has shown that earlier preparation significantly lowers the number of phonetic reduction phenomena. Furthermore, if the speech is given in front of an audience rather than being pre-recorded, the number of reduction phenomena is higher. Besides, the analysis has shown a few important tendencies in mod - ern British pronunciation, such as omission of /t/, realisation of final /t/ as [ ʔ ], omission of final /t/ in contractions, which therefore are realised as [n] instead of [nt].


Reduction, elision, and clipping
In order to specify the topic of the article, 1 let us consider two meanings of reduction: reduce (v.) (1) A term used in the phonological classification of vowel sounds, referring to a vowel which can be analysed as a centralized variant of a vowel in a related form.For The article is based on the author's BA thesis in English Philology at the Pedagogical University of Cracow, written under the supervision of Anita Buczek-Zawiła, Ph.D.
(2) A further phonological use of the term is found in the context of phonological rules, where it refers to a process of simplification which affects certain types of sound sequence.The most important category is consonant cluster reduction (e.g.clock becoming /gɒk/), which is common in early child language (Crystal 2008, 406).
In this paper, the focus is on the second meaning, whereby reduction is understood as elision.In phonology, elision, as defined by Crystal (2008, 166), is "the omission of sound or sounds in connected speech". 2 It may affect vowels, consonants and even whole syllables.Elision may be further divided according to the position of the affected element (Skandera and Burleigh 2005, 96-97): • aphaeresis: elision of an initial element or elements (would have [wəd ˈaev]); if the initial omitted element is a vowel, the process can be named aphesis; • syncope: elision of a medial element or elements (answer is [ˈɑːnsɹ ɪz]); • apocope: elision of a final element or elements (kept complaining [ˈkep kəmˈpleɪnɪŋ]).However, if a part of a word or a whole syllable is omitted, we speak of clipping.Skandera and Burleigh (2005, 97), for instance, distinguish three types of clipping: • fore-clipping: affects the initial part of a word, e.g.telephone -phone; • back-clipping: affects the final part of a word, e.g.laboratory -lab; • fore-and-aft clipping: fore-and back-clipping operate simultaneously, e.g.influenza -flu or Elizabeth -Liz.In fore-and back-clipping, we understand the beginning and the end of a word in terms of its morphological borders.
Since vowel reduction (i.e.reduction in the first meaning) is not the subject of this paper, this change in grammatical words 3 is not going to be considered. 4However, every non-standard pronunciation of these words will be noted and counted if affected by elision.What will not be counted are contracted forms, which are also treated as reduction, but counting them would be pointless because they are normative.

2
Connected speech is understood here as a continuous sequence of words typical of natural utterances and conversations.Units (words or phrases) sound different when they are pronounced in connected speech or in isolation.Another important process operating in connected speech, apart from elision, is assimilation (cf.Crystal 2008, 101).

3
That is, strong and weak forms of auxiliary verbs, conjunctions, prepositions, pronouns, articles etc.

Triphthong smoothing
Although triphthong smoothing is also an instance of vowel elision, it has been treated separately, because it affects a discrete category of sounds in English phonetics and phonology, namely triphthongs.These also undergo elision, or to be precise -syncope, which is an omission of the medial element of a triphthong (/ɪ/ or /ʊ/), especially when the following /ə/ is not a separate morpheme.Thus, triphthongs undergo following changes, which are known as smoothing (Cruttenden 2008, 145-146) (Cruttenden 2008, 147).

Elision of consonants and syllables
The main factor triggering consonant elision is complex consonant clusters.The sounds that are most frequently omitted are plosives and fricatives.Skandera and Burleigh (2005, 95)  As we read in Roach (1998, 127), the middle plosive is usually dropped in a cluster of two plosives and a fricative or of three plosives, which can be illustrated with such examples as acts /aekts/ → [aeks], looked back /lʊkt ˈbaek/ → [lʊk ˈbaek].Cruttenden (2008, 303) lists also even more complex clusters: "Thus elision is common in the sequence voiceless continuant + /t/ or voiced continuant + /d/ (…) followed by a word with an initial consonant," which is exemplified by such phrases as next day  2008,304).
In informal speech, the elision of only two-consonant clusters can be encountered, too, as in I want to [aɪ ˈwɒnə], He went away [hɪ ˈwen ə'weɪ], Give me [gɪmɪ] or Let me [lemɪ].Interesting forms are also the informal gonna, gotta, and wanna, which stand for going to, have to (or 've got to), and want to (or want a), respectively.The elision also affects the /v/ in of in informal speech when it precedes a consonant, e.g. in lots of them [ˈlɒts ə ˈðem] or waste of money [ˈweɪst ə ˈmʌni] (Roach 1998, 127).
In regional informal speech, /h/ can be elided at the beginning of a word and the word then behaves as if it started with a vowel, namely it takes the an indefinite article or [ði] definite article, for instance an hill [ənˈɪɫ] or the house [ðiˈaʊs].Even among some RP speakers, one can find such words as historical, hotel, or hysterical realised without the initial /h/ and the grammatical result is the same, as in an historical novel [ənɪ'stɒɹɪkɫ̩ ˈnɒvɫ̩ ].Such realisation, however, is not common (Cruttenden 2008, 205).
Furthermore, whole unaccented syllables can also be omitted.If the same consonant appears again in the subsequent syllable, the first syllable of the pair is likely to be dropped.This can be observed in numerous words, such as library [ˈlaɪbɹɪ] or particularly [pəˈtɪkjəli] (Skandera and Burleigh 2005, 96).

Research
In this section, we are going to examine some fragments of genuine official English speech in terms of the reduction phenomena discussed above.The aim of the research is to look into the frequency of the reduction phenomena.Although In Touch is of longer duration and features more speakers than the other two samples, it conforms with the remaining ones, since each speaker talks individually and undisturbed by the others, and only rare interaction between the speakers is observed.
The audio material was downloaded together with transcripts from the source websites.Then the recordings were slowed down and listened to, with the instances of relevant phonetic phenomena being marked on the transcripts.

Hypotheses
The occurrence of the reduction phenomena is dependent not only on phonetic and phonological features, but also on individual preferences of a speaker, as well as his or her speech tempo or regional origin.Nonetheless, we managed to find samples produced by speakers who use the standard variety of British English.Thus, two hypotheses may be forged.
Firstly, if the samples are not scripted, the number of the reduction phenomena is noticeably higher.Thus, Studio School is expected to show a significant number of the phenomena, while Pottery Cattle barely any.In Touch was labelled as semi-scripted, since three of the speakers are journalists but the remaining two are not.Therefore, the number of reduction phenomena in In Touch should be somewhere between Studio School and Pottery Cattle.
Secondly, the sounds that are most frequently reduced are stops.Here, apart from /p/, /b/, /t/, /d/, /k/, /g/, affricates /tʃ/ and /dʒ/ are considered as well, since they also contain a stop element.In order to check this, we kept thorough records of the numbers of reduction phenomena and calculated the percentages.
The results obtained are collated in tables, each dedicated to one sample.

Studio School
Studio School is a monologue, similar to a very short lecture, delivered in front of an audience.Because of that, it was labelled as unscripted.The reduction phenomena which occurred in the text are listed in Table 2 below. 9What follows, however, is a vowel. 10In this case, aspiration of /p/ in an unstressed syllable and appearance of /r/ occur as compensatory phenomena for the elision. 11Notice that two seconds later (6'4"), in the phrase for an answer for some children, the speaker uses both linking /r/ in for an and /ə/ in answer, in contrast to the preceding phrase with a similar wording.The form [ɑːns] also appears earlier, in 1'43", with linking /r/ following. 12At least is again reduced to [əʔ ˈliːs] in 6'9" in the phrase at least can help us. 13In 1'38", however, the speaker uses the full form [ɪmˈpɔːtᵊnt]. 14 Interestingly, in 4'58", in the phrase right at the heart, we observe no elision: [ˈɹɑɪt ət ðə ˈhɑːt].
With regard to this, the percentage result of reduction is 3.81% (1,050 words in total, 40 reduced 16 ).The recording gave us 4 examples of syncope, 27 of apocope, 4 of aphaeresis, 9 of weakening.The total number of reduction processes is 44.Stops were reduced in 33 cases (75%) and 26 of them were /t/, vowels in 5 cases and other consonants in 6.The results are presented in Table 3.

In Touch
In In Touch, there are five speakers altogether: two men, who are professional radio journalists, and three women, one of whom is also a professional radio broadcaster; the two remaining speakers are not trained for professional recording -they are people to whom the topic of the programme (the cancellation of the DLA mobility component) applies.They are not guests in the studio so the sample cannot be considered a discussion.Each answer to the journalist's questions is given in an elaborate way (at least a few sentences), which is why this fragment has been chosen for the study.The recording has been classified as semi-scripted.Table 4 presents the reduction phenomena found in the recording.Moreover, all of the speakers showed a strong tendency to substitute the final /t/ with /ʔ/, not only in pronouns or prepositions, but also in lexical words.Furthermore, they also tended to drop the final /t/ from negative contracted forms.Hence, these instances have been considered as a general inclination and have not been counted as reduction. 25aving said this, we may state that 65 words out of 1,775 were reduced, which amounts to 3.66%.There were 68 reduction processes, 47 (69.12%) of which were examples of apocope, 12 (17.64%) of syncope and 3 (4.41%) of aphaeresis, 5 (7.35%) of weakening and 1 (1.47%) of clipping.In total, stops were elided in 61 cases (89.7%), vowels in just 2 cases and other consonants in 6.26Among stops, /t/ was reduced 25 times.These results are presented in Table 5.

Pottery Cattle
Pottery Cattle comes from one speaker, with the exception of two very short sentences, which are quotes from two professors who contributed to the story, one sentence each.This material was professionally recorded in a studio by a man who, as we assume, had familiarised himself with the text of the speech and prepared for the recording -therefore, it was labelled as scripted.The speaker's pronunciation is much more accurate and clear than in the previous samples.Hence, very few reduction instances have been found.They are presented in Table 6.
New Horizons in English Studies 2/2017 in front of an audience and a prepared radio programme with journalists and guest speakers).It seems that our findings favour the second hypothesis as well.In the material, there were 119 elements reduced altogether, 96 of which were stops, which gives the result of 80.67%.The remaining elements were: vowels -11 (9.24%), fricatives -8 (6.72%), and other consonants -4 (3.36%).
The frequency of occurrence of reduction in the whole material is 3.39% (that is, 111 words reduced out of the total of 3,276).Table 9 presents the detailed data.

Conclusion
The research shows that prior preparation and speaking with the aid of a script influence the occurrence of elision.The audio material has provided us with a straightforward answer, although one must bear in mind the differences between the samples analysed.Furthermore, stops proved to be the most frequently elided elements in the audio material: they constituted more than 80% of the elements reduced.Amongst all the stops elided, the most significant number falls on /t/, which constituted 54% of them.This means that /t/ alone constituted more than 40% of all the elements affected by reduction.Furthermore, final /t/ was reduced in 43 cases, which is more than a third of all the sounds reduced.The second most frequently elided elements were vowels, but they amounted to less than a tenth of the total.
Another conclusion which may be drawn from the analysis of the audio material is that in present-day colloquial English, there is a strong tendency to use [ʔ] in the place of final /t/ before a vowel.Another easily observed tendency is an omission of the final /t/ of the negative contracted forms, so that -n't, normally pronounced as [nt], is realised as [n].
This study has shown that phonetic reduction is a topic worth studying, since it affects native speakers' pronunciation to a noticeable extent.The analysis has confirmed that earlier preparation for speaking influences the pronunciation, namely there [neks ˈdeɪ], raced back [ɹeɪs ˈbaek], last chance [lɑːs ˈtʃɑːns], left turn [lef ˈtɜːn], send round [sen ˈɹaʊnd] or caused losses [kɔːz 'lɒsɪz].Further on, we read: "Similarly, word-final clusters of voiceless plosive or affricate + /t/ or voiced plosive or affricate + /d/ (…) may lose the final alveolar stop when the following word has an initial consonant," which is again illustrated with such examples as kept quiet [kep ˈkwaɪət], helped me [ˈheɫp mi], stopped speaking [ˈstɒp ˈspiːkɪŋ], changed colour [ˈtʃeɪndʒ ˈkʌlə], urged them [ˈɜːdʒ ðəm] or judged fairly [ˈdʒʌdʒ ˈfeəli] (Cruttenden

3. 1
Selection of the material and procedure I have decided to analyse approximately 20 minutes of data, consisting of three samples.The first sample, A short introduction to the Studio School (henceforth Studio School), is a speech by Geoff Mulgan on TED.com from 2011 about new, more practical, and less boring schooling.The second sample is a fragment of a BBC 4 podcast series In Touch about the cancellation of the DLA mobility component for disabled people living in residential houses.The last piece of the material, A History of the World in 100 Objects.Episode 8. Egyptian Painted Pottery Cattle from BBC Radio 4 (henceforth Pottery Cattle), is a fragment of a podcast series about clay statues of cattle in ancient Egypt.

Table 1 .
The data analysed

Table 2 .
Reduction in Studio School 5678Weakening used in the table stands for no audible release of a stop.However, the records below the table, used for statistics, include both unreleased plosion and reduction to /ʔ/.Some of the examples in Table2require a comment.First of all, we have observed the speaker's general tendency to substitute a /t/ preceding a consonant with a /ʔ/.We have found a number of instances in such words as:9101112 6 7The same occurs in 2'52" in the repeated phrase.However, in 2'50", in the phrase learn best in teams, the speaker uses the full form: [ˈlɜːn ˈbest ɪn ˈtiːmz].• important -important priority [ɪmˈpɔːʔn̩ pɹaɪˈɒɹəti] (56") and most important of all (3'23"); 13 • importance -the importance of [ði ɪmˈpɔːʔn̩ s əv] (

Table 3 .
Reduction in Studio School (unscripted)

Table 4 .
On Reduction in English: What the English Don't Say Reduction in In Touch1 718192021222324 LANGUAGE

Table 5 .
Reduction in In Touch (semi-scripted)

Table 9 .
The whole empirical material