Journal of Deaf Studies and Deaf Education, 2016, 83–93 doi:10.1093/deafed/env044 Advance Access publication September 23, 2015 Empirical Manuscript
empirical manuscript
Do Signers Understand Regional Varieties of a Sign Language? A Lexical Recognition Experiment Rose Stamp* *Correspondence should be sent to Sign Language Research Lab, Rabin Building, University of Haifa 31905 Haifa, Israel (e-mail: rose_stamp@hotmail. com).
Abstract The degree of mutual intelligibility of British Sign Language (BSL) regional varieties has been a subject of some debate. Recent research in which dyads of signers from contrasting regional backgrounds engaged in a conversational task showed no problems understanding one another. The present study investigated signers’ knowledge of different BSL regional varieties. Twenty-five participants from Belfast, Glasgow, Manchester, and Newcastle took part in a computerbased lexical recognition task in which they had to identify the meaning of 47 color signs from various regions in the United Kingdom. The results indicate that overall signers have a poor knowledge of regional signs for colors when signs are presented in isolation and without mouthing. Furthermore, signers with deaf parents performed better in the recognition task than signers with hearing parents and varieties from London and Birmingham were easiest to recognize. This article discusses how signers cope with regional differences and considers the features that facilitate in the recognition of regional varieties in BSL.
A characteristic of British Sign Language (BSL) is its considerable regional variation, particularly at the lexical level (Skinner, 2007; Stamp et al., 2014; Woll, Allsop, & Sutton-Spence, 1991). In a recent corpus-based study, Stamp et al. (2014) found considerable variation in the signs for colors, countries, and numbers. They also found regional variation was in decline with younger signers using less regionally marked variants compared to older signers. This reduction in the use of regional variants was described by Stamp et al. (2014) as leveling. Leveling is thought to be the outcome of regular face-to-face interactions between speakers of differing linguistic repertoires leading to linguistic convergence, also known as speech accommodation (Giles & Powesland, 1997; Trudgill, 1986). One of the main causes for language change in British English is the need for accommodation (Trudgill, 1983a). To consider the relationship between accommodation and lexical change in BSL, Stamp, Schembri, Evans, and Cormier (2015) investigated whether dyads of signers from contrasting regional backgrounds showed evidence of lexical accommodation during a collaborative task. Observation of conversations
revealed that signers had no problems understanding one another although few instances of lexical accommodation during the conversation suggested that lexical accommodation may not account for the lexical change found in BSL. Stamp et al. (2015) suggested that because BSL signers have a passive knowledge of other regional varieties there is little motivation to accommodate. However, no studies to date have empirically investigated signers’ knowledge of other BSL regional varieties. This study aimed to investigate to what degree signers are able to recognize the meaning of other BSL regional varieties. The following research questions will be addressed: (a) which groups of signers perform best at lexical recognition for color sign variants (e.g., younger, female signers)?; and (b) which regional varieties are easiest to recognize? This study is the first to explicitly investigate signers’ knowledge of lexical variation in BSL and the factors that affect it and therefore it makes tentative conclusions about the implications of this work for better understanding the mutual intelligibility of sign language regional varieties.
Received: March 31, 2015; Revisions received: July 31, 2015; Accepted: August 1, 2015 © The Author 2015. Published by Oxford University Press. All rights reserved. For Permissions, please email:
[email protected]
83
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
University of Haifa
84 | Journal of Deaf Studies and Deaf Education, 2016, Vol. 21, No. 1
Comprehension and Language Change
Comprehension of Spoken Language Regional Varieties Trudgill claimed that native speakers of different English varieties can “nearly always understand one another” (Trudgill, 1983b, 22). Despite this, there are plenty of examples of misunderstandings. For example, research has shown that native English speakers show difficulty comprehending other regional British accents in noise (Adank, Evans, Stuart-Smith, & Scott, 2009) or without context (Trudgill, 1983b). In a study conducted by Trudgill (1983b), 26 British speakers were shown a series of sentences containing grammatical structures known to vary in English regional varieties and were asked to select the intended sentence meaning from multiple options. Nearly a quarter of participants failed to understand the meaning of the regional form without context. Trudgill (1983b) concluded that despite speakers interacting on numerous occasions
Comprehension of Sign Languages There has been minimal research on the comprehension of regional varieties in sign languages; however, investigation of intelligibility across different sign languages has proved fruitful in identifying a number of features that aid comprehension. Hiddinga and Crasborn (2011) described what they call a “natural phenomenon” in which deaf people are able to comprehend other deaf people who use a different sign language with relative ease. This has been discussed by several researchers across a number of sign languages (Allsop, 1993; Allsop, Woll & Brauti, 1995; McKee & Napier, 2002; Monteillard, 2001; Rosenstock, 2004) and has been thought to relate to the shared use of iconicity across sign languages, as well as the use of mouthings and context to disambiguate meanings (Hiddinga & Crasborn, 2011; Safar et al. 2015). One of these shared features, mouthings, has been found to aid comprehension across signers in different sign languages that share a common spoken language. Mouthings are the full or partial articulation of a corresponding spoken word without voice at the same time as the manual sign is produced by a signer (Boyes Braem & Sutton-Spence, 2001). In a recent study by Safar et al. (2015), the comprehension of Flemish Sign Language (VGT) by signers of French Belgium Sign Language (LSFB) and Sign Language of the Netherlands (NGT) was investigated. Overall, LSFB signers performed best; however, NGT signers improved their comprehension score when mouthings were included. Given that the same mouthings are shared in NGT and VGT, the author suggested that mouthing plays an important role in comprehension. These studies, looking on comprehensibility across sign languages, take into account non-manual (e.g., mouthing, facial expressions, head movements) as well as manual features of a lexical sign. Despite the fact that signers may draw on these non-manual features to increase communicational efficiency, this study focuses only on lexical aspects of BSL regional varieties. However, the influences of non-manual features such as mouthing are considered in the Discussion section.
Comprehension of BSL Regional Varieties There have been contradictory claims as to signer’s comprehension abilities of BSL regional varieties in the last 30 years. Until the 1980s minimal interaction between varieties meant that mutual intelligibility was expected to be low. Jordan and Battison’s (1976), in a study interviewing deaf signers from the United States and other nations, stated the following comment was made: “A standard story, repeated by travelers and natives alike, holds that if you travel 50 miles in Britain you will encounter a different sign language that cannot be understood in the region you have just left.” (p. 59.) This comment was somewhat supported by a report conducted by Kyle and Allsop (1982) in which 227 deaf people living in the County of Avon in the southwest of England were interviewed. The report claimed that
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
One of the main theories on language change claims that when speakers from contrasting regional backgrounds interact, accommodation takes place and when frequently repeated this leads to long-term change (Trudgill, 1983a). The most widely described theory of accommodation is known as the Communication Accommodation Theory which claims that speakers converge to their interlocutor’s linguistic behavior with the intention of achieving their social approval, maintaining positive social identities, and attaining communicational efficiency (Giles, Taylor, & Bourhis, 1973). Other theories also draw the link between communicational efficiency and accommodation. Some claim that accommodation is a process to ensure speakers maintain consistency thereby reducing collaborative effort (Clark & Murphy, 1982; Clark & Carlson, 1981; Clark & Wilkes-Gibbs, 1986; Garrod & Anderson, 1987) and others claim that accommodation serves as a form of coordination to minimize communicative effort (Barr & Keysar, 2002; Bortfeld & Brennan, 1997; Niederhoffer & Pennebaker, 2002). For example, speakers may establish their lexical choice on the basis of their past references to the same objects—known as lexical entrainment—and this has been shown to serve a functional purpose in facilitating communication (Garrod & Anderson, 1987). According to Allwood and Ahlsén (1986), accommodation is a strategic process for making oneself understood. In their language-learning studies, they found that an increase in lexical convergence helped in memory retention of the lexical item and maintained conversational fluency. Lexical repetitions were higher when there were difficulties in understanding, suggesting that accommodative behavior in the form of lexical repetitions may serve to facilitate communication. Drawing on this, one might hypothesize that if signers have difficulties recognizing signs from other regional backgrounds, they may use accommodation as a coping mechanism. This presupposes that signers have a passive, if not an active, knowledge of other regional varieties in BSL; however, this has not been empirically investigated as yet. Therefore, determining signers’ knowledge of BSL regional varieties is vital for truly understanding processes of accommodation and language change in BSL. This article considers signers’ recognition of BSL regional varieties and what this tells us about accommodation with particular reference to the “attaining communicational efficiency” explanation.
with other regional varieties, in each interaction the speaker exploited the context rather than acquiring understanding of the feature usage. According to Trudgill (1983b), passive competence does not presuppose productive competence. Therefore, speakers may be aware of other regional varieties; however, they may not accommodate either because they choose not to, or because they do not possess the knowledge to do so. The next section will discuss the literature to date on comprehension of sign languages and regional varieties, with particular reference to studies of BSL.
Stamp | 85
Methods
Manchester. These specific sites were selected because the data collected there exhibited the most regional variation for the chosen signs under investigation (i.e., signs for colors) and because these sites also showed minimal change towards non-local variants, with younger signers continuing to use a high proportion of traditional variants for their region, ensuring that participants were likely to use traditional regionally distinct signs (Stamp et al., 2014). Twelve were recruited for the first time for this study. Thirteen were recruited as former participants in the BSL Corpus Project (2008–2011). The BSL Corpus is an online digital collection of video data that is openly accessible for researchers to analyze providing a large quota sample of language use from the British Deaf community (Schembri, Fenlon, Rentelis & Cormier, 2011; Schembri, Fenlon, Rentelis, Reynolds & Cormier, 2013). It consists of data collected from a total of 249 deaf participants living in eight sites across the United Kingdom: Belfast, Birmingham, Bristol, Cardiff, Glasgow, London, Manchester, and Newcastle. All participants in this study were deaf, British-born signers and had lived in their region for at least 10 years. Most had acquired BSL before the age of 7 years with three acquiring BSL between the ages of 8–12 years and one between the ages of 13–18 years. All participants were naïve as to the exact purpose of the study and were informed that the researchers aimed to investigate BSL variation. After data collection, participants were fully debriefed about the objectives of the tasks. Table 1 presents the participants’ demographics.
Lexical Recognition Task In this task, participants were shown a number of regional variants and asked to choose the color that best matched the meaning of the sign.
Stimuli Of the semantic categories known to show considerable regional variation (e.g., colors, countries, numbers), color signs were selected because, generally, the meaning of the sign is not as transparent in its form compared to number and country signs (e.g., the sign for three usually involves extension of three digits). The color sign variants were chosen from the data elicited as part of the lexical elicitation task in the BSL Corpus Project as representing either the most frequent or the traditional variant used in each of the eight original collection sites (Stamp et al., 2014). Specifically, for this lexical elicitation task, participants were shown PowerPoint slides or flashcards for 102 concepts (e.g., colors, days of the week and numbers). For each target concept, participants were asked to produce the sign variant they use most on a daily basis. Stamp et al. (2014) analyzed the responses for 41 of the 102 concepts. In the present study, participants were shown 47 unique regional variants for the five color terms (brown, green, grey, purple, yellow). A deaf native BSL signer signed the stimuli on video. The stimuli were produced without mouthings to avoid signers using mouthings as a way to disambiguate the meanings of signs. In a previous study (Stamp et al., 2015) looking at accommodation in signers from different regional backgrounds, signers sometimes accompanied signs with the mouthing of the equivalent English word and this may have facilitated comprehension.
Participants and Sites
Procedure
A total of 25 participants took part in a lexical recognition task: six from Belfast, Glasgow, and Newcastle and seven from
The experimental trials were presented using E-Prime, version 2.0.9.22 (Schneider, Eschman & Zuccolotto, 2002). Before each
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
around 40% of interviewees had never met a deaf person living further than 125 miles away and, in the period of a year, less than half had travelled to other parts of the country (more than 50 miles away). When asked about their knowledge of BSL varieties, those used in areas further north were reported as increasingly difficult to understand, with Scottish signers being the most difficult to understand. More recent studies suggest that most deaf people no longer experience difficulties understanding regional varieties to the same degree (Elton, 2010; Woll et al., 1991). Woll et al. (1991) used a collection of sign variants from a “See Hear”1 dataset as part of a production and comprehension task. This dataset consisted of BSL signs taken from eight years (1981–1989) of the BBC television program “See Hear”. First, signers were shown an English word corresponding to one of the signs in the “See Hear” collection and they were asked to produce their own sign variants for the concept. Secondly, signers were shown a random sample of sign variants from “See Hear” and they were asked the meanings of the signs and finally signers were involved in an interview. The results of the comprehension task revealed that older signers out-performed younger signers. Overall, most rated BSL regional varieties as relatively easy to understand with signers from London and Manchester performing significantly better in the comprehension task than signers from Scotland and Newcastle. The interview data revealed that signers rated Northern Irish to be the most difficult to understand (44%) and Scottish to be the second most difficult (12%). In Elton’s (2010) study analyzing the change of signing style used by English/BSL interpreters since the 1980s, deaf signers were asked to watch BSL translations of the Queen’s annual Christmas speech. The majority (i.e., 82%) reported to have no problems understanding regional signs used. It is unclear to what degree these translations included regional variants and where these regional signs originated, however. Anecdotal claims by the deaf community supported by recent investigations (Stamp et al., 2014) have shown that regional variation in BSL is in decline. This may have led to an improvement in the mutual intelligibility of BSL regional varieties. As discussed in the introduction, observation of 25 signers from Belfast, Glasgow, Manchester and Newcastle engaging in a collaborative “spot-the-difference” task with another signer from a contrasting regional background revealed no difficulties in communication between signers of contrasting regional varieties and few instances of lexical accommodation taking place (Stamp et al., 2015). Stamp et al. (2015) suggested that signers’ passive awareness of BSL regional varieties might be a discouraging factor for accommodation. However, no recent research has explicitly investigated mutual intelligibility of BSL regional varieties. In the present study, signers are shown a number of different BSL regional variants from across the United Kingdom. The aim of the study is to investigate to what degree signers recognize the meanings of other regional varieties, which sociolinguistic variables predict signers’ performance and which varieties are easier to recognize than others and for what reasons. An understanding of signers’ knowledge of other BSL regional varieties may point to factors that influence mutual intelligibility.
4 3 3 4 14
Non-mobile
2 3 3 1 (−1 excluded) 9 (−1)
2 3 4 2 11
Mobile Non-local
4 3 4 4 15
Local
2 3 5 2 12
Middle class
4 3 2 4 13
Working class
2 1 3 2 8 0 4 2 1 7
4 5 4 4 17
Hearing Deaf Other Older 60+ years
2 0 3 3 8 2 1 1 0 4
One of the aims of this study is to consider how lexical recognition ability correlates with participants’ social factors. The dependent variable was whether the participant correctly identified the meaning of the regional sign. The following variables were coded to reflect those also investigated by Stamp et al. (2015) using the same dataset: signers’ regional background (Belfast, Glasgow, Manchester, Newcastle), age (continuous variable), gender (female, male), social class (middle, working), language background2 (deaf, hearing), mobility3 (high, low), and school location (local, non-local). In addition, regional origin of the sign and block were included as variables. The regional origin of the sign variant was analyzed to determine which regions were better recognized than others (Belfast, Birmingham, Bristol, Cardiff, Glasgow, London, Manchester, and Newcastle). Block (1, 2, 3) was also included as a variable.
6 6 7 6 25
2 5 3 3 13
Results
Belfast Glasgow Manchester Newcastle Total
Total
Younger 16–39 years
Middle 40–59 years
4 4 3 3 14
Female
2 2 4 3 11
Male
6 2 5 5 18
White
Data Coding
Sites
Language background Ethnicity Gender Age
Table 1. Number of Participants in Each Social Category
trial, participants were first shown a fixation point to draw their attention to the screen in preparation for the forthcoming sign. Then, participants were shown a BSL regional variant from one of the following colors: brown, green, grey, purple, and yellow. When the sign was completed, participants were shown eight different colored square blocks with the equivalent English word displayed on them (i.e., black, brown, green, grey, purple, red, white, and yellow) (see Figure 1). Participants were asked to select from the closed set of options the response they felt best matched the meaning of the BSL sign viewed on the preceding screen. Participants completed three practice trials to become accustomed to the task. In the practice trials, participants were shown two signs with which they were likely to be familiar (i.e., a variant produced over a number of regions in the BSL Corpus Project data) and one sign that was less commonly known (i.e., a variant that is unique to one region). At the end of the three practice trials, participants were offered the choice of continuing to the real task or repeating the practice trials. Participants were informed that all signs were examples of BSL regional variants and that they would see signs that may not be understood. They were asked to respond as quickly and accurately as possible. Participants were given a break after each block and were able to re-commence the task at their will. The position of the color responses on the screen was kept constant for each participant to avoid the chance of errors. Four separate layouts were created with different color positions and randomly distributed to participants in each region to avoid any position effect. This was a self-paced task in which participants completed three blocks each containing 47 trials (141 trials in total). Multiple blocks were included to give participants an average score over the three blocks to accurately represent their abilities. The same 47 trials were repeated in each block and randomized for each participant.
One participant chose not to complete the lexical recognition task therefore only data from 24 participants could be analyzed. Overall accuracy scores on the lexical recognition task were low. On average, participants correctly identified the meaning for 38% (i.e., 1,029 tokens) of sign variants (range: 21–50%).
Regional Accuracy Differences The total number of tokens for the lexical recognition task was 3,384 (47 trials × 3 blocks × 24 participants). A repeated measures
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
Social class
School location
Mobility
86 | Journal of Deaf Studies and Deaf Education, 2016, Vol. 21, No. 1
Stamp | 87
Figure 1. Procedure for lexical recognition experiment starting with a fixation point, followed by the target sign and eight possible answers to choose from 254 × 190 mm. in black and white. Actual stimuli were presented in color, see online version.
Table 2. Multiple Logistic Regression Results for the Lexical Recognition Task
Region of sign*
Language background* Gender Mobility School location Block
Social class Age (+1) Region
Factor
Log odds
% of correct responses
Tokens
Centred weight
Birmingham London Newcastle Manchester Cardiff Belfast Bristol Glasgow Deaf Hearing Female Male Mobile Non-mobile Local Non-local Three Two One Middle Working (continuous) Belfast Manchester Newcastle Glasgow
1.363 0.523 −0.079 −0.095 −0.119 −0.353 −0.617 −0.624 0.176 −0.176 0.121 −0.121 0.177 −0.177 0.106 −0.106 0.074 −0.010 −0.063 0.067 −0.067 0.003 0.041 0.020 0.010 −0.070
68.4 51.6 38.4 36.5 34.2 30.9 24.3 26.3 42.8 35.6 39.3 36.5 40.2 36.2 37.5 39.1 39.6 37.8 36.7 40.5 36.5
288 432 216 255 360 324 432 399 906 1,800 1,461 1,245 1,236 1,470 1,806 900 902 902 902 1,017 1,689
0.796 0.628 0.48 0.476 0.47 0.413 0.35 0.349 0.544 0.456 0.53 0.47 0.544 0.456 0.527 0.473 0.518 0.497 0.484 0.517 0.483
38.9 39.7 35.6 37.6
666 798 702 540
0.51 0.505 0.502 0.482
Note. Degrees of freedom = 11, mean = 0.38, intercept = −0.386, deviance = 3335.673. Random effects (participant) standard deviation = 0.28. Random effects (lexical item) standard deviation = 0.232. Application value: correct response. 2,706 tokens. Factor groups are significant at *p < .05.
ANOVA with region as the between-subjects variable and the regional score as the within-subjects variable found that there was a significant sign region and region of residence interaction, F(28,140) = 11.909, p < .001. There was no main effect of region F(4,20) = 0.467, p = .759. All participants performed best identifying the meanings of signs from their own region (e.g., Belfast participants had higher scores for Belfast signs than other signers). In order to investigate which regional signs participants understood most, other than their own, 390 tokens where participants were viewing signs from their own region (e.g., signs associated with the Manchester region for Manchester signers) were removed. A further 288 tokens were removed as these signs were not associated with any particular region (e.g., single
manual letter signs such as -Y- for “yellow”). The data were organized with each token representing an individual trial. A total of 2,706 responses were analyzed. Table 2 presents the results for the lexical recognition task showing the log odds, number of tokens and the centred weight for each factor (with correct response as the application value). Gender, mobility, school location, block, social class, age, and participant’s region were not found to be significant. An Rbrul analysis found two significant factors predict performance in the lexical recognition task: region of the sign (p < .001) and language background (p = .0413). In all Rbrul analyses, participant and lexical item were included as random effects to account for individual variation and variation that is subject to the lexical items under investigation. The meaning
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
Factor Group
88 | Journal of Deaf Studies and Deaf Education, 2016, Vol. 21, No. 1
Figure 2. The mean identification scores for the sign variants by regional origin 254 × 190 mm.
To investigate why Birmingham color signs appear to be easier to understand compared to other BSL regional varieties, an Rbrul analysis was run to show how each variant correlated with accuracy in order of performance (from easiest to most difficult to recognize). Figure 3 shows the variants that participants found easier to recognize and Table 3 shows the Rbrul results for the analysis. One pattern that emerged is that 6 of the 10 easiest variants to recognize also incorporated an element of BSL, ASL, or ISL (Irish Sign Language)4 fingerspelling in the sign. For example, in BSL, there are some initialized sign variants5 for the color grey. There are also examples, which incorporate the BSL letters -G- and -Y- (i.e., the first and last letters of the equivalent English word representing the lexical concept).” Fingerspelling is a feature of many Western sign languages enabling signers to spell out words, such as names and places. In the example mentioned, a signer from any regional background can easily guess the meaning of the color sign as the fingerspelling of -G- narrows down the options to colors beginning with that letter (i.e., green or grey) and in some cases ending with -Y- (i.e., grey). Other examples include the sign for the color purple, which incorporates the first letter of the sign (e.g., “p”) with a modification to the movement (e.g., flicking of fingers on dominant hand). As the BSL fingerspelling alphabet is commonly known by all BSL users, the meaning of variants that incorporate fingerspelling properties is easier to guess than those that do not. It has not been investigated to what degree BSL signers know ISL or ASL fingerspelling systems. In order to investigate this further, another Rbrul analysis was undertaken excluding all variants that incorporate elements of fingerspelling (i.e., 6 variants which totaled 345 tokens). An analysis excluding these variants revealed that the same two social factors predict signers’ performance on the lexical recognition task: region of the sign and language background (see Table 4). The pattern for language background did not change; participants with deaf parents performed better at the lexical recognition task than participants with hearing parents. The order showing the BSL regional variants which participants guessed correctly did change. After removing the six fingerspelling variants, London variants appear to be the easiest variants to recognize other than the signer’s own regional variants. An independent-samples t-test was performed (using a Bonferonni correction of p = .007) to see if the differences across regions were significant. The results show that
GREENb Belfast
BROWN London
YELLOW3 Newcastle
PURPLEb Cardiff
GREY7 Birmingham
GREYb Birmingham
GREY London
GREEN Manchester
PURPLE Birmingham
YELLOW6 Glasgow
Figure 3. Ten variants that were found to be easiest to recognize by participants 254 × 190 mm.
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
of signs from Birmingham and London were found to be easier to recognize than signs from other regions (Factor weights, Birmingham = 0.796, London = 0.628, Newcastle = 0.48, Manchester = 0.476, Cardiff = 0.47, Belfast = 0.413, Bristol = 0.35, Glasgow = 0.349). Seven post hoc independent-samples t-tests were performed to see if the differences across these regions were significant. Using a Bonferroni correction, a critical p value of .007 was used. The results show that the difference in mean accuracy scores between Birmingham (M = 0.68, SD = 0.47) and London (M = 0.52, SD = 0.50) was significant (t(718) = 4.532, p < .001). The difference between London and Newcastle (M = 0.38, SD = 0.49) was also found to be significant (t(646) = 3.192, p = .001). The difference between Newcastle and Manchester (M = 0.36, SD = 0.48) (t(469) = 0.436, p = .663), Manchester and Cardiff (M = 0.34, SD = 0.47) (t(613) = 0.589, p = .556), Cardiff and Belfast (M = 0.24, SD = 0.44) (t(682) = 0.919, p = .358) and Bristol and Glasgow (M = 0.26, SD = 0.43) (t(829) = −0.666, p = .506) was not significant. After applying the Bonferroni correction, the comparison between Belfast and Bristol (M = 0.31, SD = 0.46) (t(754) = 2.010, p = .045) was not found to be significant. The second significant predictor was language background. Participants with deaf parents performed better at the lexical recognition task than participants with hearing parents (Factor weights, deaf = 0.544, hearing = 0.456). The results suggest that Birmingham color signs are the second easiest to recognize across all regions, after the person’s own region (see Figure 2).
Stamp | 89
Table 3. Multiple Logistic Regression Showing the Variants in Order of Accuracy in the Lexical Recognition Task (Most Difficult to Recognize to Easiest to Recognize)
Log odds
Tokens
% of correct responses
Centred weight
grey6 purple4 grey9 purple4 yellow5 grey3 brown2 green3 grey10 grey5 purple3 brown green8 yellow4 brown6 purple5 purple6 brown5 green7 brown grey4 grey2 grey8 green4 purple2 green5 purple2 green2 grey11 yellow2 brown4 green6 brown green brown yellow3 purple grey7 grey grey green purple yellow6
16.193 16.193 1.517 1.222 1.09 0.909 0.855 0.855 0.767 0.706 0.531 0.465 0.439 0.236 0.057 0.057 −0.121 −0.275 −0.344 −0.348 −0.486 −0.498 −0.596 −0.843 −0.999 −1.037 −1.177 −1.262 −1.54 −1.54 −1.597 −1.597 −1.748 −1.984 −2.005 −2.067 −2.13 −2.261 −2.399 −3.198 −3.212 −3.319 −3.513
54 72 57 72 51 54 72 72 57 72 54 51 57 54 57 57 72 72 57 72 72 51 54 54 54 72 51 72 72 72 72 72 54 54 72 54 72 72 72 72 51 72 57
100 100 94.7 93.1 92.2 90.7 90.3 90.3 89.5 88.9 87 86.3 86 83.3 80.7 80.7 77.8 75 73.7 73.6 70.8 70.6 68.5 63 59.3 58.3 54.9 52.8 45.8 45.8 44.4 44.4 40.7 35.2 34.7 33.3 31.9 29.2 26.4 13.9 13.7 12.5 10.5
>0.999 >0.999 0.82 0.772 0.748 0.713 0.702 0.702 0.683 0.67 0.63 0.614 0.608 0.559 0.514 0.514 0.47 0.432 0.415 0.414 0.381 0.378 0.355 0.301 0.269 0.262 0.236 0.221 0.176 0.176 0.168 0.168 0.148 0.121 0.119 0.112 0.106 0.094 0.083 0.039 0.039 0.035 0.029
the differences in mean accuracy scores between London and Birmingham (t(430) = 0.476, p = .635), Birmingham and Newcastle (t(286) = 0.486, p = .627), Newcastle and Manchester (t(469) = 0.436, p = .663), Manchester and Cardiff (t(613) = 0.589, p = .556), Cardiff and Belfast (t(682) = 0.919, p = .358) were not found to be significant. Finally, the difference in mean accuracy scores between Belfast and Bristol (t(754) = 2.010, p = .045) and Bristol and Glasgow was significant (t(772) = 2.925, p = .004). The means and standard deviations are presented in Table 5. This suggests that the meanings of color signs associated with London are easier to recognize by participants; however, there was no significant difference between London, Birmingham, Newcastle, Manchester, Cardiff, and Belfast. Participants performed significantly worse at recognizing Glasgow color signs compared to any other region (see Figure 4).6
Discussion This study aimed to investigate signer’s knowledge of BSL regional varieties from across the United Kingdom. The research questions were: (a) which groups of signers perform best at lexical recognition for color sign variants (e.g., younger signers, females)?; and, (2) which regional varieties are easiest to recognize? Despite the apparent ease of communication observed by Stamp et al. (2015) during a conversational task, results in the present study indicate that signers found identification of the meanings of regional color signs very difficult without presence of mouthing or context. This was reflected in the results with the average participant correctly guessing the meanings of 38% (i.e., 1,029 tokens) of color sign variants.
Variation According to Social Factors Two factors were found to influence participants’ performance on the recognition task: the region associated with the sign and the signer’s language background. As expected, participants performed best at identifying the meanings of signs associated with their own region. Participants with deaf parents performed significantly better than those with hearing parents in the lexical recognition task, suggesting that native signers are more aware of other regional varieties in BSL than non-native signers. This complements previous findings which show that native signers favor the use of regional varieties more than non-native signers (Stamp et al., 2014).
Regional Accuracy Differences In the present study, each sign variant was correlated with participants’ accuracy scores to investigate which sign variants were easiest to recognize. Knowledge of signs based on the sign origin was as follows starting with the highest recognition: Birmingham> London> Newcastle/Manchester/Cardiff/Belfast> Bristol/Glasgow. In spoken languages, the most intelligible dialect is often also the standard (Clopper & Bradlow, 2008). This may be because speakers are socially motivated to recognize and understand the standard or that the standard is often the most prestigious dialect (Casad, 1974). In this study, varieties originating in Birmingham and London, the second and first largest cities in England were found to be easier to recognize than any other variety. In BSL, there is no evidence to suggest that any single regional variety is considered to be more prestigious than any other. However, some varieties may be more intelligible because signers have frequent exposure to them via the media or widespread geographical use. This may be the case in BSL with signers being exposed to sign variants from Birmingham and London more than others. Upon further investigation of the Birmingham varieties, 6 of the 10 variants found to be easiest to recognize also incorporated an element of BSL, ASL, and ISL fingerspelling. As discussed, fingerspelling is a feature of signing enabling signers to spell out words, such as names and places. Therefore if fingerspelling is incorporated into the sign variant (e.g., yellow incorporating the letter -Y-) then signers, even those seeing the sign for the first time, can guess its meaning with relative ease.
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
Variant
In summary, overall performance on the task was poor. Signers from deaf families performed better than signers from hearing families. The results suggest that signers find it easier to recognize signs that originate from London and most difficult to recognize signs that originate from Glasgow.
90 | Journal of Deaf Studies and Deaf Education, 2016, Vol. 21, No. 1
Table 4. Follow‐up Multiple Logistic Regression Results for the Lexical Recognition Task Without Variants Incorporating Aspects of FingerSpelling Factor Group Region of sign*
Language background* Mobility Gender School location
Age (+1) Region
Log odds
% of correct responses
Tokens
Centred weight
London Birmingham Newcastle Manchester Cardiff Belfast Bristol Glasgow Deaf Hearing Mobile Non-mobile Female Male Local Non-local Middle Working (continuous) Belfast Newcastle Manchester Glasgow
0.531 0.403 0.255 0.216 0.075 −0.068 −0.417 −0.996 0.186 −0.186 0.202 −0.202 0.1 −0.1 0.115 −0.115 0.066 −0.066 0.005 0.084 0.062 0.021 −0.168
44.7 41.7 38.4 36.5 34.2 30.9 24.3 15.8 36.6 29.3 34.0 29.8 32.5 30.8 31.2 32.8 34.5 30.1
360 72 216 255 360 324 432 342 789 1,572 1,575 786 1,275 1,086 1,575 786 885 1,476
0.63 0.599 0.563 0.554 0.519 0.483 0.397 0.27 0.546 0.454 0.529 0.471 0.525 0.475 0.529 0.471 0.516 0.484
32.8 29.9 33.2 30.6
576 612 693 480
0.521 0.515 0.505 0.458
Note. Degrees of freedom = 18, mean = 0.317, intercept = −0.912, deviance = 2,816.964. Random effects (participant) standard deviation = 0.237. Application value: correct response. 2,361 tokens. Factor groups are significant at *p < .05.
The results here suggest that fingerspelling as a feature of some BSL regional varieties serves to improve intelligibility. After the variants incorporating fingerspelling were removed, London variants appear to be the easiest to recognize other than the signer’s own regional variants. Given that London is the capital city of England, it is perhaps unsurprising that signers are more familiar with these varieties. This also complements Stamp et al.’s (2014) finding in which London varieties were used more by younger signers above their own traditional regional varieties. Quite possibly the London varieties may be favored due to the high status of London, however, this requires further investigation. Another possible feature that might increase recognition of regional signs is iconicity. For example, a number of variants shown in Figure 4 may be considered iconic or partly iconic as folklore on the origin of BSL signs suggest that the variant green and yellow3 stem from their relationship to the signs for “grass” (or the concept of “green fingers”) and the sign for “sun” which may themselves be highly iconic, although not all signers will necessarily make the connection with their iconic origin. Despite this, many researchers consider color signs to be relatively non-iconic (Vinson, Cormier, Denmark, Schembri & Vigliocco, 2008). In Vinson et al.’s (2008) study, they collected normative data for 300 signs including an iconicity rating from 1 to 7 (with 7 rating indicating high iconicity). Only one variant, brown (traditional for London), from this present study was found in their database and was given a very low iconicity rating of 1.9. It is important to draw on iconicity ratings in future studies to better understand why signers recognize these particular varieties above others. In the present study, Bristol and Glasgow color signs were the most difficult variants to recognize. The Bristol finding is somewhat surprising given that previous studies have found
Table 5. Means and Standard Deviations for Each Regional Origin of the Sign
Region London Birmingham Newcastle Manchester Cardiff Belfast Bristol Glasgow
Mean accuracy scores
Standard deviations
0.45 0.42 0.38 0.36 0.34 0.31 0.24 0.16
0.50 0.50 0.49 0.48 0.48 0.46 0.43 0.37
that younger signers are moving towards the numbering system associated with Bristol (Stamp et al., 2014). The finding in Glasgow supports Woll et al.’s (1991) finding in which Glasgow signs were found to be difficult to recognize. One of the limitations of this study is that participants were recruited from only four regions. The sites of investigation were the four most northern of the eight original sites and were chosen for the reason that they differed most in their lexical variation from one another. However, it would be interesting to investigate southern regions (e.g., Cardiff and London) in order to see whether any differences exist and furthermore whether they are influenced by external factors such as social class, mobility and education.
Implications As discussed in the introduction, knowledge of signers’ understanding of regional sign varieties may contribute to an
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
Social class
Factor
Stamp | 91
Figure 4. The mean identification scores for the sign variants by regional origin 254 × 190 mm.
Conclusions The results from this study suggest that lexical recognition is relatively poor in BSL when participants are tested in an experimental setting without context or mouthing. However, previous studies have found that comprehension does not prove to be a problem within conversation between signers of contrasting regional backgrounds (Stamp et al., 2015). This disparity between experimental and conversational data is interesting. The results
Notes 1. “See Hear” is a magazine program aimed at the deaf community and presented or interpreted into BSL. 2. Language background represents the hearing status of the participants’ parents. If they have at least one deaf parent then they are considered to have a “deaf” language background as it is likely that they will have been exposed to sign language early in life compared to an individual with hearing parents. 3. Participants were considered to be “high” mobility individuals if they had lived in at least two different regions of the UK and “low” mobility individuals if they had remained in their home region for their entire lives. 4. A number of ISL signs have been found in the Glasgow data, which is thought to stem from their use by Irish nuns as a method of instruction in Catholic schools (e.g., St. Vincent’s School). 5. Initialized signs (Brentari & Padden, 2001) in BSL incorporate the handshape used to represent the first letter of the equivalent English gloss (e.g., grey uses a handshape for -Gfrom the manual alphabet in BSL). 6. As is conventional in the sign language literature, glosses for signs are displayed in small capitals (e.g., the color yellow will be shown as yellow), with the superscripted number to the right of the gloss depicting a different lexical variant of the sign as outlined by Cormier, Fenlon, Johnston, Rentelis, Schembri, Rowley, Adam, and Woll in Cormier et al., 2012. The glossing used in this paper reflects the glossing system used in the BSL Corpus Project and the BSL lexical database arising from it.
Conflicts of Interest No conflicts of interest were reported.
Funding British Sign Language Corpus Project, supported by the Economic and Social Research Council (ESRC) (Grant RES- 062-23-0825 and postgraduate award, Stamp, 2013). Economic and Social Research Council of Great Britain grants RES-620-28-6001 and RES-620-280002. Deafness Cognition and Language Research Centre.
Acknowledgments The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The research for this paper took place at the Deafness, Cognition and
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
understanding of the processes of language change. Leveling, the reduction in use of regionally marked variants, was identified to be taking place in BSL (Stamp et al., 2014). Given that spoken language research attributes leveling to be the outcome of linguistic accommodation (Giles & Powesland, 1997; Trudgill, 1986), Stamp et al. (2015) investigated lexical accommodation in BSL to further understand BSL variation and change. They found minimal examples of accommodation suggesting that the language change found by Stamp et al. (2014) could not be attributed to regular contact between regional varieties. Drawing on the “attaining communicational efficiency” explanation for accommodation, one could suggest that the absence of accommodation is an indicator that communicational efficiency exists in BSL or that it is attained through other coping strategies. Two coping strategies for attaining efficient communication have been discussed in this study: fingerspelling and mouthing. In cases of misunderstandings during Stamp et al.’s (2015) collaborative task, signers would often fingerspell to disambiguate the meaning of their regional sign. In addition, during the present study, varieties that incorporated fingerspelling were identified as the easier to identify in the recognition task because the meaning could be figured out through means of elimination. In Stamp et al.’s (2015) collaborative task, signers also used mouthing as a coping strategy to overcome misunderstandings. In the present study, mouthing seems to be important in accuracy given that scores were low in the recognition task without mouthings. Mouthing accompanies many signs in BSL and can serve to disambiguate homonyms such as the signs for yellow and young (both can be produced as Y-Y) or to disambiguate the meaning of polysemic signs such as a generic “color” sign (e.g., open palm facing outwards from the body) which can mean purple, brown, etc. (Fenlon, Cormier & Schembri, 2015). In the act of mouthing, a sign variant is produced simultaneously with a borrowed element of spoken English, providing two avenues for expressing the meaning of the sign. This example of bimodal simultaneity is a unique feature in sign languages. It is likely that this enables the coexistence of several variants in BSL without the necessity for accommodation for functional purposes.
of this study are consistent with previous research findings that mouthing and fingerspelling assist signers in recognizing BSL regional varieties. Mouthing, which may act as a way of disambiguating homonyms or polysemous signs, is unique to the signed modality. A study investigating mouthing and comprehension will help to fully understand its contribution. Should signers rely on mouthing and other disambiguating factors such as fingerspelling, context and iconicity to resolve misunderstandings across contrasting regions, one might predict the rate of lexical change may be slowed by eliminating the necessity to accommodate as a means of attaining communicational efficiency. This study thus contributes to the present literature by highlighting the differences in lexical variation and change processes between spoken languages and signed languages.
92 | Journal of Deaf Studies and Deaf Education, 2016, Vol. 21, No. 1
Language Research Centre (DCAL), University College London (UCL) as part of my doctoral thesis. I gratefully acknowledge the following people: colleagues at UCL’s Experimental Psychology department for their assistance with many of the statistical analyses; Daniel Ezra Johnson for his continued support with the data analysis using RBrul; our confederate for her help with the data collection for this study. We are also grateful to all the fieldworkers who assisted with the data collection as part of the British Sign Language Corpus Project and especially grateful to the 249 deaf community members who agreed to take part in the project; without their participation, none of this research would have been possible. Finally, I am also immensely thankful for the anonymous reviews and the useful feedback from Kearsy Cormier, Adam Schembri, Gerardo Ortega, Howard Giles, and Bronwen Evans, which greatly improved the quality of this paper.
References
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
Adank, P., Evans, B.G., Stuart-Smith, J., & Scott, S.K. (2009). Comprehension of familiar and unfamiliar native accents under adverse listening conditions. Journal Experimental Psychology: Human Perception and Performance, 35, 520–529. doi:10.1037/ a0013552 Allsop, L. (1993). The International Sign project. In R. Fischer & T. Vollhaber (Eds.), Collage: works on International deaf history (pp. 21–28). Hamburg, Germany: Signum. Allsop, L., Woll, B., & Brauti, J. M. (1995). International Sign: The creation of International deaf community and sign language. In H. Bos and T. Schermer (Eds.), Sign language research 1994 (pp. 171–188). Hamburg, Germany: Signum Press. Allwood, J., & Ahlsén, E. (1986). Lexical convergence and language acquisition. In O. Dahl (Ed.), Papers from the Ninth Scandinavian Conference of Linguistics, Stockholm, Denmark: University of Stockholm, Department of Linguistics. Barr, D. J., & Keysar, B. (2002). Anchoring comprehension in linguistic precedents Journal of Memory and Language, 46, 391– 418. doi:10.1006/jmla.2001.2815 Bortfeld, H., & Brennan, S. E. (1997). Use and acquisition of idiomatic expressions in referring by native and nonnative speakers. Discourse Processes, 23, 119–147. doi: 10.1080/01638537709544986 Boyes Braem, P., & Sutton-Spence, R. (2001). The hands are the head of the mouth: The mouth as articulator in sign language (Vol. 39). Hamburg, Germany: Signum Press. Brentari, D. & Padden, C. (2001). Native and foreign vocabulary in American Sign Language: A lexicon with multiple origins. In D. Brentari (Ed.), Foreign vocabulary in sign languages: A crosslinguistic investigation of word formation, 87–119. Mahwah, NJ: Lawrence Erlbaum Associates. Casad, E. H. (1974). Dialect intelligibility testing. Norman, OK: Summer Institute of Linguistics, University of Oklahoma, Oklahoma. Clark, H. H., & Carlson, T. B. (1981). Context for comprehension. In J. Long & A. Baddeley (Eds.), Attention and performance IX (pp. 313–331). Hillsdale, NJ: Lawrence Erlbaum Associates. Clark, H. H., & Murphy, G. L. (1982). Audience design in meaning and reference. In J. F. LeNy & W. Kintsch (Eds.), Language and Comprehension (pp. 287–299). Amsterdam, North-Holland. Clark, H. H., & Wilkes-Gibbs, D. (1986). Referring as a collaborative process. Cognition, 22, 1–39. doi:10.1016/0010-0277(86)90010–7 Clopper, C. G., & Bradlow, A. R. (2008). Perception of dialect variation in noise: Intelligibility and classification. Language and Speech, 51, 175–198. doi:10.1177/0023830908098539
Cormier, K., Fenlon, J., Johnston, T., Rentelis, R., Schembri, A., Rowley, K., Adam, R., & Woll, B. (2012). From corpus to lexical database to online dictionary: Issues in annotation of the BSL Corpus and the development of BSL SignBank. Poster presented at the 5th Workshop on the representation and processing of sign languages: Interactions between corpus and lexicon, Istanbul, Turkey, 27 May 2012. Elton, F. (2010). Changing the way we sign: An analysis of the signing style used by translators in the Queen’s Christmas speech since the 1980s. Birkbeck College, University of London, UK. Fenlon, J., Cormier, K., & Schembri, A. (2015). Building BSL SignBank: The lemma dilemma revisited. International Journal of Lexicography. 28, 169–206. doi:10.1093/ijl/ecv008 Garrod, S., & Anderson, A. (1987). Saying what you mean in dialogue: A study in conceptual and semantic co-ordination. Cognition, 27, 181–218. doi:10.1016/0010-0277(87)90018-7 Giles, H., Taylor, D. M., & Bourhis, R. (1973). Towards a theory of interpersonal accommodation through language: Some Canadian data. Language in Society, 2, 177–192. doi:10.1017/ S0047404500000701 Giles, H., & Powesland, P. F. (1997). Accommodation theory. In N. Coupland & A. Jaworski (Eds.), A sociolinguistics reader (pp. 232–239). Basingstoke, UK: Macmillian. Hiddinga, A., & Crasborn, O. (2011). Signed languages and globalization. Language in Society, 40, 483–505. doi:10.1017/ S0047404511000480 Jordan, I. K., & Battison, R. (1976). A referential communication experiment with foreign sign languages. Sign Language Studies, 5, 69–80. Kyle, J., & Allsop, L. (1982). Deaf people and the community. Bristol, UK: University of Bristol. McKee, R., & Napier, J. (2002). Interpreting into International Sign pidgin: An analysis. Sign Language & Linguistics, 5, 27–54. doi:10.1075/sll.5.1.04mck Monteillard, N. (2001). La Langue des Signes Internationale. Acquisition et interactionen langue etrangere, 15, 97–115. Retrieved from http://aile.revues.org/1347 Niederhoffer, K. G., & Pennebaker, J. W. (2002). Linguistic style matching in social interaction. Journal of Language and Social Psychology, 21, 337–360. doi:10.1177/026192702237953 Rosenstock, R. (2004). An investigation of International Sign: Analysing structure and comprehension. Washington, DC: Gallaudet University dissertation. Safar, A., Meurant, L., Haesenne, T., Nauta, Y. E., De Weerdt, D., & Ormel, E. (2015). Mutual intelligibility among the sign languages of Belgium and the Netherlands. Linguistics, 53, 353–374. doi:10.1515/ling-2015-0004 Schembri, A., Fenlon, J., Rentelis, R., & Cormier, K. (2011). British Sign Language Corpus Project: A corpus of digital video data of British Sign Language 2008–2010 (1st edn). Retrieved from http://www.bslcorpusproject.org Schembri, A., Fenlon, J., Rentelis, R., Reynolds, S., & Cormier, K. (2013). Building the British Sign Language Corpus. Language Documentation and Conservation, 7, 136–154. Retrieved fromhttp://hdl.handle.net/10125/4592 Schneider, W., Eschman, A., & Zuccolotto, A. (2002). E-Prime reference guide. Pittsburgh, PA: Psychology Software Tools. Skinner, R. A. (2007). What counts? A typological and descriptive analysis of British Sign Language number variations. University of London, UK. Stamp, R., Schembri, A., Fenlon, J., Rentelis, R., Woll, B., & Cormier, K. (2014). Lexical variation and change in british sign language. Plos One, 9, e94053. doi:10.1371/journal.pone.0094053
Stamp | 93
Stamp, R., Schembri, A., Evans, B., & Cormier, K. (2015). Regional sign language varieties in contact: Investigating patterns of accommodation. Journal of Deaf Studies & Deaf Education. Stamp, R. J. (2013). Sociolinguistic variation, language change and contact in the British Sign Language (BSL) lexicon (Unpublished doctoral thesis). UCL (University College London). Retrieved from http://discovery.ucl.ac.uk/1393284/ Trudgill, P. (1983a). Linguistic change and diffusion. On dialect. Oxford, UK: Blackwell, 52–87. Trudgill, P. (1983b). On dialect. New York, NY: New York University Press.
Trudgill, P. (1986). Dialects in contact. Oxford, UK: Basil Blackwell. Vinson, D. P., Cormier, K., Denmark, T., Schembri, A., & Vigliocco, G. (2008). The British Sign Language (BSL) norms for age of acquisition, familiarity, and iconicity. Behavior Research Methods, 40, 1079–1087. doi:10.3758/ BRM.40.4.1079 Woll, B., Allsop, L., & Sutton-Spence, R. (1991). Variation and recent change in British Sign Language: Final report to the ESRC. Bristol, UK: University of Bristol.
Downloaded from http://jdsde.oxfordjournals.org/ by guest on February 5, 2016
