Nonsense Sign Repetition Task for BSL

The Nonsense Sign Repetition Task for British Sign Language (Mann, Marshall, Mason, & Morgan, 2010) has been developed within a research project investigating how phonetic complexity affects perception and articulation of handshape and movement in British Sign Language (BSL). A previous version of the task was piloted in Marshall, Denmark, and Morgan (2006).


Within the frame of the current study, two groups of children were tested: (1) Deaf children who acquired a BSL as a first language (3-11 years old) and (2) hearing children with no prior experience/exposure to signing (6-11 years old). The first group consisted of 91 deaf children, who were divided into three age groups: 3–5 years old (N = 26); 6–8 years old (N = 26), and (3) 9–11 years old (N = 38). Fourteen of these children had deaf parents, the remaining children had exposure to BSL from very early on through nursery school. The second group consisted of 46 hearing children, divided into two age groups: 6-8 years old (N = 23) and 9-11 years old (N = 23) (Mann et al., 2010).


The chosen methodology is based on the nonword repetition methodology which is used in spoken language acquisition research (e.g., Dollaghan & Campbell, 1998; Gallon, Harris, & van der Lely, 2007). Items of the Nonsense Sign Repetition Test for BSL consisted of items that were phonotactically possible, but did not carry any meaning in BSL. In order to ensure that the developed items did not exist in BSL, three native signers judged possible similarities to existing BSL signs. Signs that were detected by the three signers were removed from the item pool and were replaced by an alternative item. All items were recorded by a deaf signer and presented via a computer format (see Video 1).

Video 1: Examples of the Nonsense Sign Repetition Test for BSL (© Mann et al., 2008)

The authors of this test manipulated the phonetic complexity of nonsense signs along two phonological parameters: handshape and movement. For each parameter, items were either phonetically “simple” or phonetically “complex” (Table 1). With respect to handshape,  “simple” handshapes were the four unmarked handshapes, labeled here as “B”, “5”, “G”, and “A” (Sutton-Spence & Woll, 1999) and “complex” handshapes were all marked handshapes. With respect to movement, the authors defined just one movement, whether internal or path, as “simple”, and two movements (i.e., internal and path combined) as “complex”.


Table 1: Levels of Phonetic Complexity for the Nonsense Sign Repetition Task (from Mann et al., 2010)




Simple (unmarked)

Complex (marked)




Simple movement

(Path movement or hand-internal movement)

Level 0: 10 items

Level 1b: 10 items

Complex movement

(Path movement + hand-internal movement)

Level 1a: 10 items

Level 2: 10 items


Testing, scoring, and results
Testing of deaf children
: Each child was tested individually. The testing took about 10-20 minutes for each child. In addition to the Nonsense Sign Repetition Test, a motor skills task (bead-threading; White et al., 2006) was administered and for part of the sample (N = 65), scores from the British Sign Language Receptive Skills Test were available. Before the actual test, the children saw test instructions which were presented on video, explaining that children would see a sign which they should copy/repeat. Each item was presented only once. The 40 items were shown in blocks of 10 items, with a short break between each block during which the children were shown a short cartoon (Mann et al., 2010).


Scoring of deaf children: The results were scored by two hearing researchers, both of whom were fluent signers. The responses were coded/scored according to “overall response was correct and according to whether errors were made on the phonological parameters that we manipulated (handshape, path movement and internal movement)” (Mann et al., 2010, p. 70). It was alsocoded whether one of the movements in a movement cluster was deleted. The scores were then compared and any discrepancies were resolved. Fifteen randomly selected participants were coded also by a deaf native signer. The inter-rateragreement was high (85% for overall score, 88% for handshape, 87% for path movement, 83% for internal movement and96% for deleted movement in movement clusters; Mann et al., 2010).


Results of deaf children: As for the results of the motor skills test, the authors found only a relationship (by age groups) between the motor skills test and the Nonsense Sign Repetition Test in the youngest age group (Mann et al., 2010). As for the relationship of the scores of the BSL Receptive Skills Test and overall repetition accuracy of the Nonsense Sign Repetition Test, there was a significant correlation between two. (For a detailed discussion of the results see Mann et al., 2010).


Testing of hearing children: The 46 hearing children with no prior experience to BSL were also tested with the Nonsense Sign Repetition Test, but the test instructions were translated into English. They also completed the motor skills test. Overall, the results revealed that “the hearing children perform significantly less accurately than Deaf children of the same age” (Mann et al., 2010, p. 77)


Not much information about the usability is provided (e.g., time needed for scoring and analysis), but the length of the test is given (10-20 minutes). The authors of the Nonsense Sign Repetition Test suggest, based on the result of this study, that it can be used in the future as diagnostic tool for SLI in deaf children since “nonword repetition accuracy and language skillsare related whatever the modality of that language [is]” (Mann et al., 2010, p. 81). This is a new and promising approach. However, its diagnostic ability has been questioned in a study of 13 deaf signers with SLI aged 7-14 (Mason et al, 2010). In that study, only 4 of the 13 children with SLI scored significantly below the mean for their age group, a surprising result given that hearing children with SLI almost invariably perform extremely poorly on non-word repetition tests.


In a follow-up commentary paper, Marshall, Mann and Morgan (2011) noted that the nonsense sign repetition task was quite difficult when compared to non-word repetition tasks for hearing English-speaking children at equivalent ages.

They discuss the possibility that phonetic complexity in sign languages might differ from that of spoken languages and thus make it more difficult for signing children to process unknown familiar signs (Marshall, Mann, & Morgan, 2011). The authors (Marshall et al., 2011) argue that children have to deal with a larger repertoire of parameter values and parameter combinations compared to English. This “phonological heaviness” might be that there are “more “degrees of freedom” in the phonological composition of signs” (Marshall et al., 2011, p. 2). The authors state that this hypothesis needs to be fleshed out, but it might help “us to understand why modality differences in STM [short term memory] exist, and why STM deals particularly effectively with speech” (Marshall et al., 2011, p. 2).


Strengths: (1) Focuses on SLI in BSL, a new field which has not yet addressed very broadly in the literature, (2) child-friendly test interface, (3) information on the development of the test is provided.


Weaknesses: (1) Not much information about psychometric properties is available.




Summarized by Tobias Haug (2012, in cooperation with Chloe Marshall)

For more information regarding this test, please contact  Wolfgang Mann at the University of Roehampton.