1This article focuses on a phonological process which I have labelled Contemporary Palatalisation. Instances of Contemporary Palatalisation (henceforth, ICPs) are manifestations of the palato-alveolar fricatives and affricates /ʃ, ʒ, tʃ, dʒ/ in phonetic environments and lexical items in which they did not appear in the past. ICPs are variants mostly associated with younger speakers. Contemporary Palatalisation therefore seems to constitute a change in progress. As is frequently the case with linguistic changes, ICPs are often deemed ‘incorrect’ by conservative speakers. As far as linguists are concerned, it has been controversial whether such instances of palatalisation can be considered part of Standard accents.
2First, I will focus on the process of palatalisation that underlies ICPs and explain how it operates in four phonetic environments. After briefly surveying the different varieties of English in which ICPs may occur, I will explain how the phenomenon is the repetition of diachronic patterns that have led to palatalisation throughout the history of English.
3Then, I will particularly concentrate on the role of perception in sound change. I will propose explanations for the development of ICPs within the theories of sound change defined by Ohala (1981, 1989, 1993a, 1993b, 1994, 2003) and Blevins (2004). These theories do not account, however, for the diffusion of innovations to other members of the speech community. Smith’s cognitive model of change (2007) takes a wider perspective and considers that sound change normally goes hand in hand with social change, which can account for the diffusion of linguistic innovations. Smith’s theory will be used to explain the emergence of ICPs in the context of significant post-World War Two social and political changes, both within and beyond the British Isles.
4In this paper, palatalisation will be used as an umbrella term for several phonetic events. In historical linguistics, palatalisation is traditionally described as an articulatory shift that is both vertical and horizontal. While the tongue moves up towards the highest point of the hard palate, the horizontal shift consists either in a fronting of velars (e.g. /k, g/) or in a retraction of dentals or alveolars (e.g. /t, d, n/). The vertical shift in the place of articulation results in articulatory fortition, which is an increase of tension in the tongue muscles (Laborderie 1994: 85). In historical linguistics, it is traditionally considered that the production of a palatal consonant is not a genuine instance of palatalisation if it is not the result of those two simultaneous phonetic gestures (Straka 1979, Laborderie 1994). Indeed, genuine palatalisation is characterised by fortition and a horizontal shift in the place of articulation of the phoneme that has been palatalised (Straka 1979).
- 1 Today, [ç] can be an allophone of /j/ after /p, t, k, h/ and before /uː, ʊə/ (Cruttenden, 2008: 226 (...)
5In English phonetics, the term palatalisation usually has a broader meaning. According to Viel (2007: 20), palatalisation occurs when a phoneme becomes palatal because it is in contact with a palatal consonant. Strictly speaking, the palatal consonants of English are the palatal glide /j/ and the voiceless fricative [ç] (the status of which is no longer phonemic in contemporary English1). However, it is customary to give a rather wider meaning to the term palatal by including the palato-alveolars /ʃ, ʒ, tʃ, dʒ/ (Stévanovitch 2008: 21). This is the view adopted in this paper: the production of palato-alveolars can be the result of a process of palatalisation, even without articulatory fortition. The term palatalisation will therefore be based on the following definition given by Roach (2009b):
- 2 The bold characters correspond to the transcription chosen by Roach. It has been reproduced here.
It is difficult to give a precise definition of this term, since it is used in a number of different ways. It may, for example, be used to refer to a process whereby the place of an articulation is shifted nearer to (or actually on to) the centre of the hard palate: the s at the end of the word ‘this’ may become palatalised to ʃ when followed by j at the beginning of ‘year’, giving ðiʃ jiə.2
6In the example above, the process of palatalisation is synonymous with the retraction of the articulation of an alveolar fricative. It therefore consists in a horizontal shift that results in the production of a palato-alveolar (/s/ à /ʃ/). Throughout the history of English, both retraction of alveolars and fronting of velars have resulted in the production of palato-alveolars (cf. 2.5).
7Palatalisation is of course a frequently occurring process across word boundaries, as is illustrated by the following examples in Jobert and Mandon-Hunter (2009: 183):
(1)
|
|
|
/s/ + /j/ /s/ → /ʃ/
|
this year /ˈðɪʃˈjɪə/
|
/s/ + /ʃ/ /s/ → /ʃ/
|
nice shoes /ˈnaɪʃˈʃuːz/
|
/z/ + /j/ /z/ → /ʒ/
|
those euros /ˈðəʊʒˈjʊərəʊz/
|
/z/ + /ʃ/ /z/ → /ʒ/
|
She’s sure /ˈʃiʒ ˈʃɔː/
|
(2)
|
|
|
/t/ + /j/ → /tʃ/
|
Don’t you know? /ˈdəʊntʃəˈnəʊ/ (…)
|
/d/ + /j/ → /dʒ/
|
Would your parents mind? /wədʒəˈpeərəntsˈmaɪnd/
|
8In [1], the alveolars /s/ and /z/ become respectively /ʃ/ and /ʒ/ after the palatals /j/ or /ʃ/. In [2], the alveolars /t/ and /d/ coalesce with /j/ and become the affricates
/tʃ/ and /dʒ/.
9Palatalisation is a form of assimilation. All the examples given above are instances of regressive assimilation. The initial consonants of the words on the right, /j/ or /ʃ/, trigger the assimilation process. They are the segments that Pavlík (2009) calls assimilators:
The notion of assimilation presupposes the existence of at least two segments (phonemes/allophones), which, by influencing each other, change their phonetic properties. We may distinguish between the segment which is being assimilated – the assimilee, the segment which assimilates another segment (transfers some features to it) – the assimilator, and the segment resulting from the assimilation, i.e. the assimilee after the assimilation, which we will name, for want of a better term, the assimilant.
10As Roach (2009a: 111) points out, the assimilees do not disappear altogether, the duration of the assimilant being ‘more or less what one would expect for a two-consonant cluster’.
- 3 In the rest of this paper, this dictionary is referred to as LPD, followed by the date of the editi (...)
11In the Longman Pronunciation Dictionary3 (2008), Wells makes a difference between dealveolar assimilation and coalescent assimilation. The examples given in [1] are instances of dealveolar assimilation. In such cases, the alveolar acquires a feature from the following palatal assimilator, which remains unchanged. Wells labels assimilations such as those in [2] coalescent assimilation or yod coalescence. Two contiguous consonants fuse into one segment that retains features from both the assimilator and the assimilee. Yod coalescence involves no articulatory fortition.
12Connected speech processes across word boundaries have been described extensively in the literature and are considered neither incorrect nor substandard. In addition, palatalisation also occurs within words. While some instances of word palatalisation are considered to be standard in a great many varieties of English (e.g. issue, nature), the status of others, mostly associated with the latter part of the 20th century and the beginning of the 21st century, is more controversial from a prescriptive point of view. I have labelled such phenomena Instances of Contemporary Palatalisation (ICPs). As to the phonological process which leads to ICPs, I have called it Contemporary Palatalisation so as to distinguish it from previous processes in the history of the English language.
13This is a type of assimilation where the approximant /j/ (yod) fuses, or coalesces, with preceding /t, d/, resulting in affricates /tʃ, dʒ/, e.g. tune /ˈtjuːn/ à
/ˈtʃuːn/ ; dune /ˈdjuːn/ à /ˈdʒuːn/. The yod is the assimilator.
- 4 In Early Modern English, borrowings from French were gradually palatalised in items like nature, fo (...)
14In an article about Received Pronunciation (RP), Wells (1997) writes that ‘English has long had a tendency to convert /tj/ into /tʃ/ and /dj/ into /dʒ/’ (e.g. nature).4 He observes that the process spread to new words in the mid twentieth century to include words like actual, perpetual, gradual, graduate, whose everyday forms contain the affricate /tʃ/ or /dʒ/, their variants with /j/ being ‘mannered’ or ‘artificial’. Wells finally notes that a new change occurred in the late twentieth century, whereby coalescence continued to ‘widen its scope’, extending to stressed syllables in words like Tuesday, duke, dune.
15First, there was considerable resistance to accept coalescence in stressed syllables (e.g. Ramsaran 1990, Wells 1997). Linguists were reluctant to consider it as standard English and to include it in the description of RP. However, a study by Hannisdal (2006) turned the tables. As a direct consequence of her work, Wells decided to include coalescence in stressed syllables into descriptions of RP (LPD 2008). On his blog, he even explains that he had been wrong not to do it before:
In LPD I labelled these variants “non-RP”. Clearly I was wrong to do so (even if it’s true for people of my own advanced age).5
16The comment above seems to indicate that these variants constitute a change in progress, mostly associated with younger speakers, which is confirmed by various pronunciation preference polls in LPD 2008. Cruttenden (2008) also considers that yod coalescence in stressed syllables is a change that is “well-established” in RP.
- 6 Yod dropping (Wells 1982: 148) is the elision of /j/ in /Coronal+ j + u/ clusters.
17Of course, most accents of North America and other varieties in which yod dropping (Wells, 1982)6 is particularly prominent do not display this ICP.
- 7 The remarks made in this article concern all editions of EPD. This is the reason why no date is men (...)
18Words like presume and assume have traditional forms /prɪˈzjuːm, əˈsjuːm/ that can be palatalised into /prɪˈʒuːm, əˈʃuːm/. The assimilator is /j/, which retracts the articulation of both /s/ and /z/, making them palato-alveolar as opposed to alveolar. Therefore, /sj/ is palatalised into /ʃ/ and /zj/ is palatalised into /ʒ/. Again, the phenomenon is widely accepted in unstressed syllables but it is less so in stressed syllables. For instance, the palatalised variants are listed as non-standard variants in LPD 2008 and they are not even listed at all in the English Pronouncing Dictionary (EPD).7 The phenomenon is not as common as palatalisation in /tju, dju/ sequences, partly because yod dropping has the upper hand in these environments and partly because fewer items contain /sju, zju/ than /tju, dju/ sequences in the English lexicon. Nevertheless, these variants appear to be progressing:
Coalesced forms in the onset of accented syllables, e.g. in assume, presume are increasingly heard in RP. (Cruttenden, 2008: 227)
19As with the first ICP, most accents of North America and other varieties in which yod dropping is particularly prominent do not display this ICP.
20This third ICP concerns words like stop, start, stress, street, student, stew, which display the palatalised variants /ˈʃtɒp, ˈʃtɑːt, ˈʃtres, ˈʃtriːt, ˈʃt[j]uːdənt, ˈʃt[j]uː/. In items like stress, street and student, stew, it is the /r/ and the /j/ which respectively retract the articulation of the alveolar fricative and make it palato-alveolar. In the case of /st/ clusters, identification of a particular assimilator is much less obvious, as /s/ and /t/ are both alveolars. Such instances of palatalisation may well be the result of a paradigmatic type of assimilation.
21EPD does not list any palato-alveolars in these environments while the existence of palatalised variants of /str, stj/ is only mentioned in passing in LPD 2008 (52):
Some speakers of British English assimilate /s/ to /ʃ/ before /tr/ and /tʃ/, thus strong /strɒŋ/ à /ʃtrɒŋ/, student /stjuːdənt/ à /stʃuːd-/ à /ʃtʃuːd-/. This is not shown in this dictionary.
22There is a rather strange imbalance in the traditional distribution of initial /s/ - /ʃ/ when another consonant follows. If the following consonant is a sonorant, /s/ always occurs (e.g. slight, smile, snail, sue, swell), except before /r/ (e.g. shrimp, shrink) and recent borrowings (e.g. schmuck). Similarly, only /s/ occurs before obstruents (e.g. skate, spy, stake) except in a handful of recent borrowings (e.g. shtetl, shtick). In fact, it is remarkable how those borrowings do not pose any problems for the contemporary phonotactics of English, to the extent that it might be argued that /ʃtr/-type pronunciations for earlier /str/ are actually facilitated by these recent borrowings.
23In a study about the palatalisation of /str/ clusters, Rutter (2011) uses acoustic measurements to compare ten English speakers’ realisations of the onsets /ʃ/, /ʃɹ/, /ʃtɹ/, and /s/. He finds out that most of the occurrences of /str/ clusters produced by these speakers fall within their normal range for /ʃ/, as opposed to various intermediate phonetic realisations falling somewhere between a typical /ʃ/ and /s/. The results indicate that the change towards /ʃ/ is complete for those speakers.
24This fourth –and last– ICP can be found in items like anniversary, classroom, estuary, grocery, nursery, which have palatalised variants /ˌænɪˈvɜːʃri, ˈklɑːʃruːm, ˈeʃtjʊri, ˈgrəʊʃri, ˈnɜːʃri/ (LPD 2008). It is again the /r/ that triggers the assimilation process by retracting the articulation of /s/. These palatalised variants are listed in LPD, for both British and American English, but not in EPD.
25Overall, linguists have shown reservations as to the inclusion of ICPs within standard accents of English. It therefore seems perfectly reasonable to assume that they originate in non-standard pronunciations and that they might well be in the process of being standardised (at least to a certain extent). From a diachronic point of view, ICPs seem to be the continuation of a historic process whereby palato-alveolar fricatives and affricates have gradually diffused into English. Throughout the history of the English language, new palato-alveolars have repeatedly been created as allophones of pre-existing phonemes. The phonetic innovations thus produced were eventually fossilised (phonologised). All the examples below are taken from Stévanovitch (2008).
(3)
|
In Old English, [k] → [tʃ] in certain environments (e.g. Old English Cinn [k] → Contemporary English chin [tʃ]; Old English tæcan [k] → Contemporary English teach [tʃ]).
|
(4)
|
In Middle English, then Modern English (15th century: rare, 16th end 17th centuries: more common), [sj] → [ʃ] (e.g. nation [nasjõ] → [neɪʃən]; sure [syr] → [sjuːr] → [ʃʊə]).
|
(5)
|
In the 17th century [zj] → [ʒ] (e.g. measure [məzyr] → [mezjuːr] → [meʒə]).
|
(6)
|
In the 17th century [tj] → [tʃ] (e.g. nature [natyr] → [naːtjuːr] → [neɪtʃə]; fortune [fortyn] → [fortjuːn] → [fɔːtʃən]).
|
(7)
|
In the 17th century [dj] → [dʒ] (e.g. soldier [soldjər] → [səʊldʒə]).
|
- 8 Altendorf, 2003: 69; Altendorf and Watt, 2008: 213; Cruttenden, 2008: 87, Coggle, 1993: 51-52; LPD (...)
26On account of certain articles, ICPs have perhaps mistakenly been associated with the south-east of England and/or with Estuary English.8 However, as soon as one starts paying attention to the pronunciation of native speakers from the whole of the English-speaking world, it becomes obvious that contemporary palatalisation is a much more widespread phenomenon. This is apparent in the specialist literature. Indeed, the phenomena which I have labelled ICPs have been noted in many varieties of English in England (including RP), in Scotland, Canada, Australia, New Zealand, South Africa, as well as in several varieties of US English (cf. Glain 2011: 51-57 for detailed references).
- 9 There are of course many other factors that explain the sound change studied in this paper.
27Elsewhere (Glain 2011), I have shown that ICPs constitute a change in progress. ICPs are diffusing in the four environments described in this article by taking over from the traditional forms. It is a well-know fact that language change originates in language variation. Indeed, certain linguistic variants take over from others. To illustrate this point, I will now turn to one particular phenomenon that can partly account for the development of ICPs from the latter part of the 20th century: the role that perception plays in linguistic change.9
- 10 Grimm’s Law can partly be accounted for in this way. Aspirated allophones of /p, t, k/ are alleged (...)
28The hypothesis that the listener’s perception of phonetic events plays a part in language change is not new.10 As early as the beginning of the 20th century, Baudouin de Courtenay (1910, in Blevins 2004) highlighted the importance of listening errors as a factor of change, owing to the difficulty of the listener’s analysis of overlapping phonetic patterns. More recent studies have allowed linguists to formalise theories of change based on the role played by perception.
29The model of sound change proposed by Ohala (e.g. 1981, 1989, 1993a, 1993b, 1994, 2003) is based on the auditory perception of the acoustic signal. It is entirely non-teleological (Ohala 1981, 1989, 1993b) and has been studied in the laboratory. The principle of sound change relies on the phonetic variants –the number of which is seemingly unlimited – which are produced in actual speech (Ohala 1981: 179). The fact that both intra- and inter-speaker variation coexist is worth noting. In the mind of the speaker, the utterance that he/she produces ‘consists of a string of one or more discrete units’ (Ohala 1994: 239). However, things are not nearly as clear on the phonetic level and co-articulation of neighbouring phonemes frequently occurs. In connected speech, units actually overlap each other, which makes the signal imperfect. Thus, Studdert-Kennedy (1998: 169-171) explains:
Speech is not a simple left-to-right sequence of discrete and invariant alphabetic segments such as we see on a printed page. The reason for this is that we do not speak phoneme by phoneme, or even syllable by syllable. Typical rates of speech – 10 to 15 phonemes/second – are possible because we coproduce, or coarticulate, the units. At each instant, our articulators are executing overlapping patterns of movements that may correspond to several neighbouring phonemes, including phonemes in neighbouring syllables. (…) Thus, the units of the acoustic signal do not correspond, one for one, with the units of perception.
30Studies in instrumental phonetics show that units also overlap each other from the point of view of speech perception (Ohala 1994). What makes communication possible is that the listener parses the string of phonetic events in order to reconstruct the units intended by the speaker (Ohala 1994: 371–375). The reconstruction is made possible thanks to rules that the listener applies when exposed to the language that he/she normally uses (Wilhelm 2011a: 10). Ohala (1981) calls them reconstructive rules as they allow listeners to derive common underlying forms from different surface realisations. This is the process known as parsing. Most of the time, the parsing is correct and the speaker’s intended units are accurately reconstructed. In a minority of cases, the parsing is erroneous and the units that constitute the signal are not properly decoded. Ohala (1994: 376) explains that linguistic change is made possible when the listener fails to identify the ‘co-occurring phonetic events’:
Such mis-parsings, which can be found in controlled listening tests in the laboratory, apparently occur at modest rate in all exchanges between speaker and hearer. Most such errors are corrected but those that are not are potential sound changes.
- 11 I have chosen to use the terms hypo- and hyperadaptation proposed by Smith (2007) as opposed to hyp (...)
- 12 An example of hyperadaptation would be an item like chain realised as [ˈtjeɪn] through erroneous di (...)
- 13 In Ohala’s terminology, assimilation and coarticulation are even perfectly similar (Ohala 1993a: 16 (...)
31Ohala and Lindblom’s model of sound change is known as the H&H theory (H&H). It relies on the listener’s hypoadaptation and hyperadaption11 to the speech signal. According to Lindblom (1986, 1990), a speaker’s speech runs along a continuum that varies from hypoarticulation to hyperarticulation. The two ends of the continuum imply that sound change may operate in two different ways. When hypoadaptation occurs, the listener may underarticulate to reproduce sounds he/she has not properly perceived and thus ‘miss the target’ (Smith 2007: 18) when trying. This way, the realisation of a consonantal phoneme based on coarticulation may be perceived as an innovative variant of that unit. That innovation may well be generalised to other environments (Ohala, 1981). On the contrary, erroneous application of corrective/reconstructive rules on the part of listeners may lead them to hyperarticulate in order to correct ‘errors’ which have been perceived but not produced in actual fact. This is the phenomenon known as hyperadaptation. Hyperadaptation is not relevant to ICPs as it is related to the process of dissimilation12 (Ohala 1994). ICPs are the result of the opposite process, assimilation, which is closely related to coarticulation.13
32When a listener alters his speech through hypoadaptation, a ‘mini-sound change’ occurs. However, if that innovation is then copied by other speakers, this mini-sound change could become a regular sound change, i.e., characteristic of a well-defined speech community’ (Ohala 1981: 184). The model proposed by Ohala focuses on changes that occur at the level of individual speakers/listeners. Therefore, it does not explain how innovations spread throughout the speech community. Ohala considers that transmission and diffusion are beyond considerations linked to the causes of change. However, the temporal dimension implied by gradual changes is naturally compatible with the H&H theory as changes can take place over several generations of listeners/speakers.
33The H&H model of sound change allows us to make a parallel between synchrony and diachrony. Ohala (1994: 375) maintains that the coarticulation of phonemes in synchrony, and the errors of perception that it entails, have the exact same effect as other co-occurrences, which have been identified at the diachronic level. He illustrates his point with an example of palatalisation:
Synchronically, stops released before high vowels or glides show more intense noise in their burst and aspiration (if any); this is the same environment where diachronically stops develop into affricates (e.g., actual < [ækt + juəl] is [æktʃuəl]).
34In actual, we are dealing with an instance of fossilised coarticulation, which is a pattern of variation originally synchronic and phonetic that has become diachronic and phonological (Ohala 1993a: 155). A significant number of sound changes are the result of coarticulations that have been fossilised and, therefore, phonologised (Ohala 1993a).
35The parallel between synchrony and diachrony can be applied to yod coalescence after /t, d/ in stressed syllables and to yod palatalisation after /s, z/ in stressed syllables (cf. 2.1 & 2.2). In an item like tune, for example, the [t]+ [j] sequence can be misinterpreted as [tʃ] and reproduced as such, through hypoadaptation and failure to apply a rule of the following sort:
Confronted with a potential distortion, the listener can acquire sufficient experience to be able to factor it out. He knows that a slightly affricated release to a stop before a high front vowel or a glide is to be expected and that it is not part of the speaker’s intention (Ohala 1989: 185).
36Failure to apply reconstructive rules does not allow the listener to ‘factor out’ the phonetically-conditioned affrication, which can become a new norm for this particular listener.
- 14 We may again draw a parallel with diachrony. Historically, phonologisation of allophonic events has (...)
37To Ohala (1981: 185), a true sound change (as opposed to mere allophonic variation) occurs in the speech of a particular listener/speaker when they lose track of the environment that mechanically conditioned the phonetic change.14 This way, the listener may fail to identify a variant as a predictable one and store it as an innovation in his/her mental lexicon. His/her pronunciation will then rest upon the norm that has thus been created. In most cases, the result is an exaggeration of the phonetic variant originally produced, rather than a faithful reproduction of it (Ohala, 1993a).
38To illustrate the latest point, let us consider an item like tune. The traditional pronunciation [ˈtjuːn] may change into [ˈtʃuːn], and be stored as /ˈtʃuːn/, following several parsing errors.
-
The typical phonetic turbulence associated with the devoicing of [j] after the aspiration on the initial plosive may lead to the production of the palatal fricative [ç] (as in German ich), which is no longer phonemic in English today but which is often phonetically produced and which is phonetically very close to a devoiced [j] (Ashby 2005: 35).
-
The palatal [ç] can be perceived as the palato-alveolar [ʃ] as the two phones are phonetically and acoustically very close. What might induce an English-speaking listener to perceive a palato-alveolar is that /ʃ/ is part of his/her phonological system, while [ç] is not. The speaker would then map [ç] onto phonological /ʃ/ under the influence of the system of contemporary English.
-
When the listener becomes speaker, [ç] is exaggerated and produced as [ʃ]. The presence of the affricate /tʃ/ in his/her system further prompts the production of the affricate after [t].
39Similar processes of hypoadaptation can also account for palatalisation in items like dune, assume and presume. The [dj, sj, zj] sequences are misinterpreted as the palato-alveolars [dʒ, ʃ, ʒ] and hypoarticulated as such. Exaggeration of the fricative elements in [sj, zj] sequences may also play a part in the sound change. It follows that the process of assimilation that characterises ICPs is, at least partially, of an auditory nature. Pavlík (2009: 11) defines auditory assimilation as such:
Auditory assimilation is an assimilation that can be detected auditorily by means of auditory testing. Interestingly, there seem to be assimilations that occur only in perception, that is, some assimilations are articulatorily not present but they are nevertheless heard as assimilations.
- 15 Both consonants are voiceless fricatives. Only their place of articulation varies.
40Errors of perception due to the confusion of acoustically similar phones may lead to ‘mini changes’ in the course of interaction between two speakers/listeners. Let us consider two ICPs: palatalisation of /st, str, stj/ clusters and palatalisation of /s/ by /r/. The similarity that exists between [s] and [ʃ], both acoustically and articulatorily15, can certainly account for the process of palatalisation. Experiments of instrumental phonetics show that listeners divide [s] and [ʃ] along a continuum (Mann and Repp 1981, in Ohala 2003). In certain languages (e.g. Korean, cf. Yavas 2006: 37), the two phones are not contrastive, they are mere allophones. In /s, ʃ + C/ initial clusters, Germanic languages have evolved differently to yield different initial sibilants. For example, modern German (and Yiddish) only has /ʃ/ while Dutch only has /s/. In English, /s/ and /ʃ/ are fully phonemic. However, their particular phonotactic distribution in initial clusters (cf. 2.3) may participate in the listener’s possible confusion between the two phones. Auditory assimilation may therefore lead to the listener’s failure to discriminate between [s] and [ʃ]. Auditory errors of this type may trigger ‘mini changes’, leading in turn to ‘maxi changes’.
- 16 There exists a long tradition that aims at bringing together language change and Darwinian evolutio (...)
41Blevins’s model of Evolutionary Phonology (2004) shares a great deal with the ‘H&H theory’. Both models of change are non-teleological and are based upon the phonetic variation inherent to language and the listening errors that continuous speech implies. Furthermore, EP also links diachrony and synchrony, as it does phonetics and phonology, insofar as phonetic innovations are seen to mirror diachronic processes. EP argues that sounds often undergo similar changes across languages (Blevins 2004: 18) as a result of parallel evolution. The analogy with biological evolution and natural selection comes from the fact that certain genetic features develop similarly, yet independently, in several species (Darwin 1859, cf. Blevins 2004: 18). For instance, [ki] has repeatedly been palatalised into [tʃi] in a number of languages, as it was in Old English (cf. 2.5). This is due to the tendency for [ki] to be perceived as [tʃi]. Typically, consonantal assimilations are generalised and phonologised, which gives rise to a true sound change originating in phonetic variation. As with Darwinism16, not all innovations last. In short, ‘Evolutionary Phonology proposes historical, non-teleological, phonetic explanations for synchronic sound patterns. Cross-linguistic similarities which occur with greater than chance frequency are viewed as the result of direct inheritance or parallel evolution (…) the primary explanation for a synchronic sound pattern is historical’ (Blevins 2004: 81).
42ICPs seem to fit perfectly with this model of sound change. As we have seen in 2.5, they are the continuation of diachronic processes. In addition, the mechanisms that lead to palatalisation in English apply to other languages (e.g. [ki] à [tʃi]).
43Blevins’s model of sound change is composed of three types of evolutions called CHANGE, CHANCE and CHOICE (the ‘CCC model’). It can be summarised as follows (Blevins, 2004: 32-33):
GENERAL TYPOLOGY OF SOUND CHANGE IN EVOLUTIONARY PHONOLOGY
(S = SPEAKER, L = LISTENER)
CHANGE: The phonetic signal is misheard by the listener due to perceptual similarities of the actual utterance with the perceived utterance.
Example :
|
S says (anpa) L hears (ampa)
|
ii) CHANCE: The phonetic signal is accurately perceived by the listener but is intrinsically phonologically ambiguous, and the listener associates a phonological form with the utterance which differs from the phonological form in the speaker’s grammar.
Example :
|
S says (Ɂa̰Ɂ) for /aɁ/ L hears (Ɂa̰Ɂ) and assumes /Ɂa/
|
iii) CHOICE: Multiple phonetic signals representing variants of a single phonological form are accurately perceived by the listener, and due to this variation, the listener (a) acquires a prototype or best exemplar of a phonetic category which differs from that of the speaker; and/or (b) associates a phonological form with the set of variants which differs from the phonological form in the speaker’s grammar.
Example :
|
S says (kakáta), (kăkáta), (kkáta) for /kakata/ L hears (kkáta), (kăkáta), (kakáta) and assumes /kkata/
|
44It is hardly possible to see ICPs as an instance of CHANCE, where the listener correctly perceives the traditional variants [tj, dj, sj, zj, st, str, stj, s] in items like tune, dune, assume, presume, stop, strong, student or grocery but analyses them as /tʃ, dʒ, ʃ, ʒ, ʃt, ʃtr, ʃtj, ʃ/. CHANCE cannot really operate in the case of a change in progress because of the normative preference for traditional variants. However, CHANCE is much more likely to be valid in the case of changes already well-established in the speech community. A speaker pronouncing the item actually with a particularly devoiced [j] or with [ç] presents the listener with information that is phonologically ambiguous (as it is situated between /j/ and /ʃ/). It can be interpreted as /ˈæktʃuəli/ because it is the norm, rather than the exception, in English today.
45CHANCE may operate for the most frequent ICPs as high-frequency items are more likely to undergo reduction processes like assimilation (e.g. Bybee 2001, Shockey 2003).
46Certain ICPs can easily be accounted for through CHANGE, which implies the confusion of acoustically similar phones. Thus, CHANGE is particularly likely to occur in instances of palatalisation of /s/ into /ʃ/ in /st, stj, str/ clusters, as well as in the environment of /r/, particularly if it is retroflex (cf. auditory assimilation in 3.1). For example, the listener hears grocery and student as [ˈgrəʊʃri] or [ˈʃtjuːdənt] whereas the speaker’s pronunciation has actually been closer to [ˈgrəʊsri] or [ˈstjuːdənt], a typical case of auditory assimilation. In the long term, the underlying representations of grocery and student may well become /ˈgrəʊʃri/ and /ˈʃtjuːdənt/ for that individual, who will then produce [ˈgrəʊʃri] and [ˈʃtjuːdənt] when he turns speaker.
47Other ICPs can be accounted for through CHOICE, which implies a selection among several phonetic variants of a given lexical item on the part of the listener. According to Blevins (2004: 141), the choice is made on the basis of a prototypical form (an idealisation) and may influence the phonological representation (the underlying representation) of that item for that particular individual. Let us for example consider the item tune. The listener hears several variants of (tune) pronounced with a devoiced [j] or with [ç] (corresponding to various degrees of devoicing and of affrication) and analyses them as /ˈtʃuːn/.
48In CHOICE, the listener may also interpret /əˈʃuːm/ and /prɪˈʒuːm/ from variants of assume and presume that present various degrees of friction. Incidentally, Blevins (2004: 142) explains that change from /sj/ to /ʃ/ is typical of CHOICE. Theoretically, the same process may operate in items like grocery and lead to palatalisation of /s/ on account of the variation in the friction of the initial /s/ and of the different degrees of retraction of that /s/ in the environment of /r/. In CHOICE, the auditory input is characterised by the different variants produced by the same speaker (i.e. intra-speaker variation). The more frequent a lexical item, the more likely it is to trigger a change of a CHOICE type (Blevins 2004: 263), owing to the multiplicity of coexisting variants. Besides, the more frequently a particular variant is heard, the more likely it is to have an influence over the phonological representation of new learners (Blevins 2004: 41-42). This perfectly explains how variation leads to phonological changes from one generation of listeners/speakers to the next, through successive applications of CHOICE.
49To summarise,
CHANGE, CHANCE, and CHOICE are intrinsic features of the inexact language inheritance mechanism and constitute primary sources of sound change at the level of the individual (Blevins, 2004: 44).
50Blevins (2004: 19) acknowledges that a genuine sound change must spread throughout the community if it is to be recognised as such. However, as is the case with the ‘H&H theory’, the ‘CCC model’ is only concerned with changes at the level of the individual speaker. Such individual innovations constitute the necessary initial stage of a genuine linguistic evolution. Change at the level of the individual is therefore synonymous with potential linguistic change. It follows that social and sociolinguistic factors are necessary to account for the actuation of change, for the potential to become actual.
51Smith (2007) proposes a model that includes elements from the ‘H&H’ theory while considering the social dimension of change. The traditional dichotomy between internal and external change is not relevant as Smith (2007: 74) works within the framework of cognitive linguistics. Cognitivists ‘posit an intimate, dialectic relationship between the structure and function of language on the one hand, and non-linguistic skills on the other’ (Taylor 1996: ix). In other words, within a cognitive framework, there is no clear-cut boundary between the mental processes associated with human language and those related to the rest of human experience.
52Smith (2007: 19) agrees with the ‘H&H’ view of intra-speaker variation, whereby ‘speakers have the freedom to vary, that is to elaborate (overarticulate) or simplify (underarticulate) their speech under the control of communicative and situational constraints’ (Lindblom 1990). In other words, for any given phone, speakers have a repertoire of variants that they can choose from according to the situation. Among those variants, there exists a prototypical realisation of the phone that corresponds to its phonological representation. Smith reminds the reader that Jones’s definition of the notion of phoneme is ‘a family of related sounds’ (Jones 1956: 172). These related sounds are organised around the prototypical value, as is illustrated below.
A family of sounds (Smith, 2007: 20)
• = actual realizations
P = prototypical realization
53Jones’s definition corresponds to a mental, rather than a functional conception of the phoneme (Duchet 1998: 33). The prototypical value can vary from speaker to speaker. It follows that, owing to hypo- or hyperadaptation phenomena, the listener may change his/her pronunciation, through a process of identification with and adoption of the prototypical value of his/her interlocutor. Smith (2007: 11) maintains that a true phonological change occurs at the level of the individual if the adoption of the new value modifies the listener’s phonological system. In other words, ‘a sound change has taken place when a variant form, mechanically produced, is imitated by a second person and that process of imitation causes the system of the imitating individual to change’ (Smith 2007: 27). The more frequent the contact with speakers whose phonological system is different, the more significant the change in the listener’s system.
54Yod coalescence after /t, d/ can clearly illustrate the model proposed by Smith. When a speaker (A) who only has palatalised forms in unstressed syllables (e.g. actually, fortune, duality, durability) interacts with another speaker (B) who has palatalised forms in both unstressed and stressed syllables (e.g. actually, fortune, duality, durability, tune, tutor, dune, reduce), adoption of the prototypical values of B may lead to a modification of the consonantal system of A. Indeed, /tj, dj/ may merge with /tʃ, dʒ/, modifying A’s system from A1 to A2.
55A1 (initial system of A)
/tʃ/
|
actually, fortune
|
/dʒ/
|
duality, durability
|
/tj/
|
tune, tutor
|
/dj/
|
dune, reduce
|
56→
A2 (A’s modified system, following contact with B)
/tʃ/
|
actually, fortune, tune, tutor
|
/dʒ/
|
duality, durability, dune, reduce
|
57This way, A’s consonantal system is partly modified following a process of imitation.
58To Smith (2007) and Blevins (2004), sound change is the norm, rather than the exception. The reason why the ongoing evolution of sounds does not lead to the complete breakdown of communication is that the vast majority of the innovations that come from the interaction between two speakers are not diffused into the speech community (Smith 2007: 12-13):
It may well be that many, if not most, linguistic changes (…) are never diffused very far (…)
It therefore seems theoretically robust to suggest that many linguistic changes fail to diffuse beyond a very limited speech community.
59If change is always potential within variation, it is necessary to have an interaction between extra- and intralinguistic factors at a particular time for a particular change to occur and then to diffuse into the community (Smith 2007: 10). This raises the question of the actuation of change on a large scale. Why is a given change actuated at a particular time? Why not earlier or later? Smith (2007) argues that the reason why some innovations catch on in the community is often related to social considerations. The evolution may even originate in major historical events or ideological changes (Labov 2010: 44).
- 17 In the 1920s, the BBC had decided that all its presenters had to be RP speakers (hence the term BBC (...)
60Let us first consider Britain. The emergence of ICPs in the latter part of the 20th century coincides with the period when RP began to lose ground in Great Britain. At the same time, media interest in non-standard pronunciations arose. Regional accents appeared on the BBC17 and have been on the increase in all media ever since. That phenomenon had tremendous psychological repercussions. Prescriptivism in pronunciation had been particularly well-developed since the end of the eighteenth century and the conferences given by Thomas Sheridan, in which he lectured on the correct pronunciation of words. As a result of that long prescriptive tendency, non-standard speakers of British English felt linguistically insecure. Linguistic insecurity is the lack of confidence experienced by speakers when they believe that the way they speak does not conform to –and is inferior to– the standard variety (Calvet 2011: 47). Linguistic insecurity started to decline with the increase of exposure of non-standard varieties in the media.
61We have hypothesised that ICPs originate in non-standard varieties of English (cf. 2.5). This is how they have been treated in pronunciation dictionaries. If ICPs have been able to penetrate standard accents like RP, it follows that there must be some sort of dynamic relation between non-standard accents and RP. The influence of non-standard varieties on standard British English was noted by Sweet (1988: 54) as early as the end of the 19th century. More recently, the development of supra-local, highly influential, varieties of British English has been synonymous with the adoption of non-standard features into RP (Hannisdal 2006: 54). According to Wilhelm (2011b: 10), the presence of such features in both RP and local accents is the proof of a dynamic relation between RP and local, non-standard accents. RP is absorbing a number of non-standard innovations today (Kerswill 2007). As a result, variety and change are also present within RP (Wells 1982, Hannisdal 2006, Cruttenden 2008). It is thus possible to posit that the link between RP and non-standard accents may have played a significant part in the development of ICPs.
62Can the decline of the standard pronunciation in Britain (RP) at the end of the 20th century be explained within the model of language change proposed by Smith? To answer this question, we must try to find evidence for a link between the decline of RP and a specific social and/or historical context. If such a correlation can be found, it will be theoretically reasonable to posit that the development of ICPs was triggered by an interaction between intra- and extralinguistic factors.
63The decades following World War Two were characterised by significant social changes in Britain. Hannisdal (2006: 15) makes a link between this socio-historical context and the decline of RP:
Up until the middle of the 20th century RP reigned supreme as the unrivalled English pronunciation standard. But in the decades after the Second World War Britain underwent radical social changes which also left their marks on the linguistic development and on the attitudes towards accent. Along with the general social changes, the role of RP also changed considerably. Between 1944 and 1966 the number of universities in Britain doubled and higher education became available to people from diverse social backgrounds. The increased democratisation in the educational system extended into the occupational and public life. Professional and academic careers became open to people from the lower social strata, who of course were non-RP speakers. Regional accent features “massively invaded the realms of the social elite” (Wotschke 1996: 221) and the hegemony of RP was broken. An educated speaker was no longer synonymous with an RP speaker, and RP was no longer the exclusive property of a narrow social class.
64In 1970, Gimson (1970: 18-19) wrote:
The acceptance of the BBC accent, i.e. some form of RP, as a standard can no longer be said to be common amongst the younger people. The social structure of the country is much less rigid than it was forty years ago, and the young are particularly apt to reject authority of any kind. This general rejection includes the accent of the “Establishment”, i.e. RP.
65Thus, the decline of RP coincided with major social changes that went hand in hand with an increasingly equalitarian ideology in Britain. Kerswill (2007: 38) also notes that the loss of RP’s privileged status accompanied the strong social mobility in post-World War Two Britain. In addition, the emergence of non-standard accents in new contexts should be seen ‘in the context of the ideology, first emerging in the 1960s, of gender and racial equality and the legalisation of contraception, abortion and homosexuality – coupled with a generally greater access to education’ (Kerswill 2007: 51). All those social and historical facts are highly favourable to the actuation of linguistic change, within the framework of a cognitive model.
66If such a model is to be applied to the rise of ICPs, it must be tested beyond the British Isles. It so happens that the decline of RP as a sort of international standard after World War Two extended far beyond Europe at a time when Britain lost its Empire. In the United States, some sort of international English based on RP served as a model in high society. Thus, Labov, Ash and Boberg (2006: 46) explain that:
- 18 A passage from a speech by F.D. Roosevelt can be heard at the address below. This speech exhibits a (...)
r-less pronunciation, as a characteristic of British Received Pronunciation, was also taught as a model of correct, international English by schools of speech, acting, and elocution in the United States up to the end of World War II. It was the standard model for most radio announcers and used as a high prestige form by Franklin Roosevelt.18
67Of course, we must keep in mind that American English does not have yod coalescence after /t, d/ and yod palatalisation after /s, z/. However, the loss of the standard defined by Labov can certainly account for the development of the other two ICPs in US English. The post-World War Two decades also witnessed the decline of a certain form of formal speech in public contexts in the USA.
68The art of oratory has long been part and parcel of American culture. One easily recalls having heard important speeches from the 20th and the beginning of the 21st century (e.g. John Fitzgerald Kennedy, Martin Luther King, Barack Obama). Some other speeches have been passed down from one generation to the next even in the absence of recordings (e.g. Abraham Lincoln). McWhorter (2012: 109) explains that listening to speeches used to be a form of entertainment in the USA. For example, ‘before Abraham Lincoln delivered the Gettysburg Address, a professional orator named Edward Everett delivered a two-hour formal speech to entertain the crowd’. McWhorter (2012: 109-110) observes that the 20th century witnessed a gradual shift from formal and written-based types of public speaking to speeches that became more informal and much closer to real spoken English. This was part of a more general change. Indeed, ‘the difference between spoken and written language has been key in a general transformation in American language culture over the past several decades from one focused on written forms to one focused on spoken ones. This has been influenced in part by the spread of recording technology and in part by late 20th-century countercultural movements that rejected traditional forms of oratory’ (McWhorter 2012: 109).
69Let us leave aside the question of technology and focus on the countercultural movements. McWhorter (2012: 111) explains that the major changes came after the 1960s ‘and the general trend toward questioning the establishment’. One of the linguistic consequences of the ideology of the time was that a more natural, much less ceremonial form of language became the preferred style. This had longer-term repercussions as ‘by 1981, at which point countercultural America had settled back down into something more conservative again, American rhetoric, even in the most formal settings, had changed. Modern speechmaking was more like talking, and orators took pride in sounding more like the common man’ (McWhorter 2012: 111).
70It is entirely possible that, in shifting away from written-based, overarticulated speeches, modern speechmaking has participated in an overall change favourable to processes of phonetic reduction based on coarticulation and hypoarticulation, such as ICPs. In the USA, like in Britain, the 1945-1970 period witnessed a radical change in speaking standards. Those innovations were triggered by social changes.
- 19 Anthony Pym teaches translation studies and sociolinguistics at the University of Rovira i Virgili, (...)
71The loss of prestige of an RP-type model in the latter part of the 20th century also applies to varieties of English in the Southern Hemisphere. Let us consider Australia. Traditionally, linguists define three sociolinguistic varieties of Australian English that form a continuum: Broad Australian (informal register), General Australian (neutral register) and Cultivated Australian (formal register). Cultivated Australian has sometimes been called RP because of its similarity with British Received Pronunciation (Pym 2008). During the major part of the 20th century, it was common to observe speakers switch from one variety of Australian English to another in order to accommodate to their interlocutor(s) and to adapt to the context (Pym 2008). Things have radically changed over the past few decades. In the postcolonial world, the decline of a British model has coincided with a more typically Australian form of pronunciation. Pym (2008) has witnessed the evolution. He observes that the main features of Australian pronunciation have progressed and that General Australian has taken over from Cultivated Australian. When he goes back to Australia19 and talks to his former professors, he realises how Australian they now sound while they could pass for British speakers back in the 1970s.
72Trudgill and Hannah (2008: 22) note that ‘there has been a very marked decline’ in the prestige of RP in Australia, New Zealand and South Africa in ‘the past fifty years or so’. It follows that non-standard pronunciations have been on the rise in several varieties of English around the world since the 1960s, which cannot but have favoured the emergence of ICPs, originally non-standard variants par excellence. As is proposed by Smith’s cognitive model of sound change, a particular social, historical and ideological context helped trigger the actuation of the change inherent to ICPs, both within and beyond the British Isles.
73In this paper, I have introduced the phonological phenomenon which I have labelled Contemporary Palatalisation, as well as its lexical manifestations, Instances of Contemporary Palatalisation (ICPs). Various observations of palatalisation in different contexts, made by different linguists about different varieties of English, have been brought together under the common name of Contemporary Palatalisation, a process that operates in four different environments. What is common to all four ICPs is not the input but the output of the process, that is to say systematic palato-alveolars. The phonological view adopted here is therefore that Contemporary Palatalisation expresses product-oriented, rather than source-oriented generalisations (cf. Bybee, 2001: 126).
74While being phenomena mostly associated with the last fifty years, ICPs are the continuation of a long historical process which has systematically led to palatalisation in English. Contemporary Palatalisation is therefore an example of synchronic processes that are in fact manifestations of systematic, diachronic ones.
75I have suggested that ICPs can partly be explained by listening errors resulting from the coarticulatory nature of speech, together with extralinguistic social and ideological changes. I have therefore tried to bring an answer (at least in the specific case of ICPs) to what Weinreich, Labov and Herzog (1968) call the actuation problem and regard as ‘the heart of the matter’: why a change occurs at a particular time. Everything seems to indicate that the particular sound changes listed in this paper are part of a larger linguistic trend whereby spoken English underwent radical changes in the second part of the 20th century. Such evolutions operated within the overall context of a certain democratisation of society in English-speaking countries, which went hand in hand with a certain democratisation of language.