Section Six

Data Analysis – The Results

In accordance with the aim of the research, the data was analysed with respect to tokens associated with the (ju) variable.  Once the relevant tokens had been identified and transcribed, they were classified according to the preceding phonological environment.  Finally, the individual speaker scores were totalled in order to represent both age groups as well as both genders.

The following table illustrates the percentage of yod-dropping and yod-retention across all preceding consonants except /t/ and /d/ with respect to age:

(6.1)

This shows that whereas the older generation of islanders retain /j/ in 39.04% of the possible 356 tokens, younger speakers retain /j/ in 91.38% of the relevant 290 tokens.  Thus, the percentages show that neither group of speakers had a categorical usage of one form, although the younger islanders did exhibit complete yod-retention in some environments.  Consider the graphs below which illustrate the age-group percentages according to the preceding consonants:

(6.2)

(6.3)

Except for the 100% score of /u/ before [T] (achieved through a single token enthusiast by a 74 year old male), there are no preceding environments where the older speakers demonstrate categorical use of either /ju/ or /u/.  Conversely, the younger informants exhibit 100% yod-retention following /f v k g h l/ which (except in the case of /l/) is in keeping with previous analyses which claim that coronals are more likely to induce /j/-loss.  Therefore, the lingering, albeit slight persistence of /u/ after /n/ and /s/ (17.55% and 12.5% respectively) for this sample of younger speakers reflects the coronals’ power over yod-deletion.  However, the persistence of /u/ following labials /p/ 9.62%, /b/ 8% and /m/ 15.38% in this age group is reflected in the robustness of /u/ in these environments for the older informants who exhibit 62.5%, 75.61% and 48.15% respectively.  This observation contrasts with both Trudgill (1974) and Spurling (2004) who rank /p/, in particular, higher only with respect to the dorsals /k g/ regarding the promotion of /j/-loss.

The implication of the data at this stage is that yod-dropping, as a feature of the Mersea dialect, is on the decline as the new generation islanders emerge.  Therefore, further decomposition of the data to illustrate gender variation may provide an insight into which gender is more likely to be advancing the shift towards yod-retention.

(6.4)

(6.5)

Figures (6.4) and (6.5) demonstrate how older female speakers were recorded as using more non-standard /u/ than their male counterparts, while this pattern was reversed for the younger age group.  Although the older female percentages can be said to have been influenced by one informant’s categorical use of /u/ throughout the interview (a performance which was not emulated by any of the older males), this still contrasts with the robust sociolinguistic findings that women use proportionately more standard forms than men (Coates 2004:68).  However, the reversal of this pattern in the younger informants suggests that the females, as greater users of the standard form, are conforming to the notion that “women seem to initiate changes above the level of consciousness” (Coates 2004:185), which tend to be in the direction of prestige norms. 

The rankings for preceding consonantal environments for older and younger speakers are therefore given below as well as the ranking once both sets of data have been amalgamated (- nb for a phonological perspective regarding these rankings see section seven below).

(6.6)

Older generation -         /T/> /b/> /f/> /s/> /p/> /st/> /k/> /h/> /n/> /v S/> /m/> /l/> /g/

Younger generation -     /n/> /m/> /s/> /p/> /b/> /f v k g h l/

Total -             /T/>/st/>/s/>/b  S/>/f/>/n/>/v/>/h/>/p/>/m/>/k/>/l/>/g/

(6.7)

Yod-Coalescence – Preceding /t/ and /d/

The combined totals for all informants for the relevant variants [ju], [u], [§] and [½] are given below:

(6.8)

Percentages for

all informants –

(6.9)

Overall, yod-coalescence is significantly more robust than either [ju] or [u] variants and is worth examining in further detail.  Once the data has been categorised by age, it is evident that younger speakers coalesce [dj] at every opportunity.  This is also predominantly true of [tj] (since less than a tenth of [tju] tokens remained as so), meaning that the variant [u] did not feature at all in the speech of the younger age group.  It is also worth noting that these speakers were exhibiting yod-coalescence in both word-internal and word-initial positions. 

(6.10)

Younger

Generation –

(6.11)

Younger

Generation –

The position of the preceding consonant is crucial when interpreting the percentages of older informants:

(6.12)

Older

Generation -

(6.13)

Older

Generation -

Although the data indicates that this age group produces 71.88% [§] and 12% [½] throughout the tokens, such coalescence does not occur in word-initial positions.  By way of illustration, items such as Tu!esday and du!ring featured alternations between only [ju] and [u] variants, whereas items such as situa!tion and gra!dually allowed the option of coalescence.  Although these examples correspond with Bauer’s (1994:108) assertion that coalescence is more common in unstressed syllables, there were no tokens in the older generation data which had unstressed word-initial /tju/ or /dju/ from which to draw a comparison of [§ ½] usage.  However, there were examples within the younger generation data of coalescence in items with unstressed initial syllables such as tui!tion

[ `§u."I.S«n].

Therefore, with respect to coalescence of /tj dj/, the following rankings can be extracted:

(6.14)

Older generation -         t à  /§/> /ju/> /u/            d à  /u/> /ju/> /½/

Younger generation -     t à  /§/> /ju/> /u/            d à  /½/> /ju/> /u/

Total -             t à  /§/> /ju/> /u/         d à  /½/> /u/> /ju/

Dialectual Development of Palatalisation

As mentioned above, Bauer (1994) proposes that coalescence of /tj dj/ are examples of fortition due to the retention of the palatality associated with /j/.  Therefore, it follows that this feature must be licensed, and thus present in the surface form in order for the merger of elements to occur.  With this in mind, we propose that dialects with consistent yod-dropping must feature /tju dju/ forms before coalescence can take place, and that syllables which are not subjected to main stress that lead the change (see the stress data below).  This results in a transitional period during which time an increase in /ju/ realisations replace /u/ before they themselves decrease in favour of the coalesced forms.  Thus, regarding the older generation data in (6.12), the low frequency of [u] following [t] (9.38%) is complemented by double the amount of [ju] (18.75%) which provides greater opportunities for coalescence (71.88%).  Conversely, [u] following [d] is still the most prolific variant and, although there is 16% yod-retention, the change is not yet advanced enough (unlike the state of preceding /t/) to excel the spread of /½/ at the expense of /dju/ and /du/. 

This observation contrasts with the younger generation since [dj] experiences 100% coalescence and is therefore more advanced than /§/.  With respect to preceding /t/, the elimination of the [u] variant has reacted to create an increase in standard [ju] which we suggest has also decreased, and will continue to decrease in favour of /§/.

Effects of Stress – All preceding consonants except /t d/

Wells (1982a) claims that yod-deletion in Cju syllables is less likely when the syllable is unstressed.  This is demonstrated by the Mersea data below:

   nb- ‘non-main stress non-adjacent’ indicates that there is an unstressed syllable between the potential /ju/ and the main stress

(6.15)

Older Generation-

(6.16)

(6.17)

Younger

Generation –

(6.18)

These graphs illustrate how the [u] variant is more prolific when the syllable is under main stress.  This observation contrasts with Tent’s (2001) analysis which demonstrates that yod-dropping is more common in unstressed syllables.  However, Bauer (1994:108) asserts that the patterning of (ju) under main stress is the same when separated from main stress by an unstressed syllable.  As the above data illustrates, this is not the case with respect to the data of Mersea informants which instead shows a corresponding pattern between both classifications of non-main stressed syllables.

Effects of Stress – Preceding /t d/

The effect of stress on (ju) is significant when considering the process of coalescence.

(6.19)

Older

Generation –

(6.20)

The older generation islanders correlate with the expectation that coalescence is much more likely in unstressed syllables.  In addition, it can be seen that even elements of secondary or tertiary stress inhibit the coalescence process.  Even though the former evaluation is true of the younger generation, these informants retain a minimal amount of [ju] even when the syllable was unstressed, reducing the amount of coalescence produced. 

(6.21)

Younger

Generation –

(6.22)

Since it has been noted above (see section three) that coalescence is an example of fortition, its occurrence in unstressed syllables may be a device to add phonetic strength to an otherwise phonetically weak syllable.

 Home        Next