- Split View
-
Views
-
Cite
Cite
Jennifer S. Beal-Alvarez, Daileen M. Figueroa, Generation of Signs Within Semantic and Phonological Categories: Data from Deaf Adults and Children Who Use American Sign Language, The Journal of Deaf Studies and Deaf Education, Volume 22, Issue 2, April 2017, Pages 219–232, https://doi.org/10.1093/deafed/enw075
- Share Icon Share
Abstract
Two key areas of language development include semantic and phonological knowledge. Semantic knowledge relates to word and concept knowledge. Phonological knowledge relates to how language parameters combine to create meaning. We investigated signing deaf adults’ and children's semantic and phonological sign generation via one-minute tasks, including animals, foods, and specific handshapes. We investigated the effects of chronological age, age of sign language acquisition/years at school site, gender, presence of a disability, and geographical location (i.e., USA and Puerto Rico) on participants’ performance and relations among tasks. In general, the phonological task appeared more difficult than the semantic tasks, students generated more animals than foods, age, and semantic performance correlated for the larger sample of U.S. students, and geographical variation included use of fingerspelling and specific signs. Compared to their peers, deaf students with disabilities generated fewer semantic items. These results provide an initial snapshot of students’ semantic and phonological sign generation.
Many deaf children use sign language for communication and instruction. Across various-sized samples of the deaf student population, Gallaudet Research Institute (GRI) has reported between 28% (2013) and 46% (2009) use sign language for instruction, either alone or paired with spoken language. GRI's survey is estimated to represent about 65% of deaf/hard of hearing children in the USA (Knoors & Marschark, 2012), although this percentage likely varies by the total number sampled each year. Many factors affect children's sign language development, including their parents’ hearing status and whether they use sign language with their children, chronological age, age at language acquisition, amount of language exposure (measured by years in a signing environment), home language, and disability diagnoses (i.e., other than deafness) (Chamberlain & Mayberry, 2000; Corina, Hafer, & Welch, 2014; Marschark & Leigh, 2016; Mitchell & Karchmer, 2005; Moeller & Luetke-Stahlman, 1990; Moeller & Schick, 2006). In the current realms of bilingual (i.e., American Sign Language, ASL/printed English) and data-based instruction, educators, and researchers need to know how children currently perform, how they should be expected to perform, and how various factors may affect their performance to guide their language instruction (Common Core State Standards Initiative, 2010; Every Student Succeeds Act, 2015; Individuals with Disabilities Education Improvement Act, 2004). Documentation of children's language skills provides both individual data and developmental data to guide instruction, while deaf adults’ performance provides a model for fluent target production (Becker, 2009; Galvan, 1999).
Two key areas of language development include semantic and phonological knowledge. Semantic knowledge relates to word and concept knowledge. Phonological knowledge relates to how language parameters combine to create meaning. Children's semantic and phonological knowledge can be measured via verbal fluency tasks in which they rapidly recall and produce (via spoken or sign language) words or signs related to specific categories. Frequently used fluency generation tasks prompt production of animals and foods (semantic) or words that share specific language parameters (phonological), such as the same initial sound (spoken language) or handshape (sign language), within a one-minute time period (Marschark et al., 2004; Mitrushina, Boone, Razani, & D'Elia, 2005; Morere, Witkin, & Murphy, 2012; Tombaugh, Kozak, & Rees, 1999; Strauss, Sherman, & Spreen, 2006; Wechsler-Kashi, Schwartz, & Cleary, 2014). This study presents empirical data related to deaf children and adults’ semantic and phonological knowledge via verbal fluency tasks.
Expressive Semantic Fluency Tasks
The majority of previous research on the semantic fluency of deaf children involved receptive tasks, as opposed to generation tasks (Kates, Kates, & Michael, 1962; Kates, Kates, Michael, & Walsh, 1961; Kates, Yudin, & Tiffany, 1962; Koh, Vernon, & Bailey, 1971; Liben, 1979; Löfkvist, Almkvist, Lyxell, & Tallberg, 2012; McEvoy, Marschark, & Nelson, 1999; Tweney, Hoemann, & Andrews, 1975). Semantic and phonological generation tasks are accomplished via a cognitive process in which similarities among class members are recognized, stored, retrieved, and produced (Barr & Brandt, 1996; Boyd & George, 1973; Courtin, 1997; Ormel et al., 2010). To be successful in these generation tasks, participants must search their mental lexicon, generate words within a subcategory, be aware of when they struggle to produce more words, and transition to other subcategories, or clusters, when the previous category is exhausted (Kavé, Kigel, & Kochva, 2008; Lanting, Haugrad, & Crossley, 2009; Wechsler-Kashi et al., 2014; Witkin, Morere, & Geer, 2013). Food generation tasks have focused on food in general and the specific categories of fruits, vegetables, and items in a supermarket, all of which represent “a clear category of common items with which most people are familiar” (Morere et al., 2012, p. 146; see Mitrushina et al., 2005 for an overview). Broader categories, such as supermarket items, seem to elicit more category members than animal naming, while more restrictive categories, such as fruits or vegetables, produced slightly fewer items than the animal category, with fruit generation being the most commonly administered task for spoken language users (Mitrushina et al., 2005).
Two studies investigated the semantic generation of young deaf children who used cochlear implants and spoken language. Younger children, 6–7 years of age, performed similar to their hearing peers for animal generation, while older children 8–9 years of age (Löfkvist et al., 2012) and 7–10 years of age (Kenett et al., 2013) performed lower than their hearing peers. Kenett et al. (2013) concluded that the older deaf children had less-developed semantic networks than their hearing peers. To date, only one published study is available on animal and food generation for signing children (Marshall, Rowley, Mason, Herman, & Morgan, 2013). Nineteen deaf children, 6–15 years of age, four of whom were native signers, produced a mean of 15.6 (SD = 5.2) animals and a mean of 16.3 (SD = 5.2) foods, similar to their hearing peers (Marshall et al., 2013). Two younger deaf children, (i.e., 4–5 years old) were unable to complete the semantic generation task without prompting (Marshall et al., 2013). Deaf children within that study with a documented specific language impairment (SLI; N = 10, 7;5–14;0) generated slightly fewer animals and foods than their deaf peers without SLI but these differences were not significant. Marshall et al. (2013) reported a significant correlation between age and total animal responses (r = 0.601) and age and correct animal responses (r = 0.648), similar to hearing children (Halperin, Healey, Zeitchik, Lusman, & Weinstein, 1989); however, years of sign language exposure, as reported by teachers/parents, did not significantly correlate with animal and food scores. Deaf children generated similar responses within each category, which paralleled responses of hearing children (Crowe & Prescott, 2003; Nelson, 1974). For example, the most frequent responses for animals were giraffe, lion, elephant, tiger, horse, cat, and dog. There was no significant difference between the number of animal clusters (i.e., subcategories of related items; M = 3.9, SD = 1.6) and food clusters (M = 4.0, SD = 1.7) or the number of items within animal clusters (M = 3.2, SD = 0.82) and food clusters (M = 3.6, SD = 1.2) (Marshall et al., 2013). Marshall et al. (2013) did not present results for any fingerspelled responses by children.
Adults
Adult data are frequently used to indicate target performance (Becker, 2009; Galvan, 1999), in the present case as an indication of how children and adolescents should perform on these generation tasks when they reach adulthood. Available data for signing deaf adults suggest they perform similar to hearing spoken language users for the category of animals (see Marshall, Rowley, & Atkinson, 2014 for a review). Deaf adults who used British Sign Language (BSL), 21–60 years of age, 18 of whom had deaf parents, performed similarly between animal and food categories, with a mean of 23 animals (range 16–37) and 24 foods (range 15–39) (Marshall, Rowley, & Atkinson, 2014). Deaf college students (N = 49, M = 25 years) who used ASL (40% had at least one deaf parent) also performed similarly between animal and food categories, with a mean of 21 animals (range 10–38, SD = 5.0) and 21 foods (range 9–35, SD = 5.0) (Morere et al., 2012), although they performed slightly lower than Marshall et al.’s (2014) deaf adults. Based on Mitrushina et al.’s (2005) meta-analysis of 11 studies of hearing participants, those between 25 and 29 years of age are predicted to generate 24.4 animals (SD = 4.65). Morere et al. (2012) point out that their college student results were lower than but within one standard deviation of the hearing participants’ mean.
Although deaf adults produced more clusters for animals than foods, the average number of clusters and cluster sizes did not significantly differ between animals and foods, with a mean of 6 clusters and a mean of 3.8 items per cluster (Marshall et al., 2014). Higher scores resulted from more clusters, larger clusters, and more switches to activate new clusters within given categories. There were significantly more switches among clusters for food (7.8) compared to animals (6.5), suggesting more depth of knowledge for animals (Marshall et al., 2014). Deaf adults produced more animal and food clusters than deaf children (4 versus 6) but the mean cluster size was similar (3–4 items) (Marshall et al., 2013, 2014). Morere et al. (2012) suggested deeper knowledge of animals might stem from reading influences. One might speculate that food knowledge is more a result of daily experience, such as meals regularly served at home or school, while animal knowledge is derived from both lived experience, such as school-related zoo and aquarium field trips, and direct science instruction via school curricula, such as organization of animals into categories (e.g., mammals and reptiles).
Finally, fingerspelling was an uncommon phenomenon for deaf adults in the animal and food generation tasks, with only 1.6% of food responses and 2.1% of animal responses fingerspelled (Marshall et al., 2014). When fingerspelling did occur, it was used for items without conventionalized signs because of their low incidence (e.g., ENCHILADA), or use of lexicalized fingerspelling for high frequency items (e.g., H-A-M) (Marshall et al., 2014).
Expressive Phonological Fluency Tasks
Phonology in signed languages is represented by four parameters: handshape, location, movement, and non-manual markers (i.e., facial expression, eyebrows, etc.) (Marshall et al., 2014; Morere et al., 2012; Neidle, Kegl, MacLaughlin, Bahan, & Lee, 2000). Phonological generation tasks appear more difficult than semantic tasks for both adults (Harrison, Buxton, Husain, & Wise, 2000; Marshall et al., 2014; Morere et al., 2012) and children (Koren, Kofman, & Berger, 2005; Marshall et al., 2014; McQuarrie & Abbott, 2013; Sauzéon, Lestage, Raboutet, N'Kaoua & Claverie, 2004) with fewer responses and more errors than semantic generation tasks. In signed language phonological fluency tasks, participants are given parameters, such as handshapes, and prompted to generate as many related signs as they can in one minute (Marshall et al., 2014; Morere et al., 2012). Morere et al. (2012) prompted participants to generate signs for the handshapes of 5, 1, and U in this order (e.g., 5: PAPER, CHEESE; 1: RED, MEET; U: BUTTER, CUTE). Similar to Marshall et al. (2014) the selected handshapes escalate in level of difficulty to present a more challenging task (Morere et al., 2012; Morford & MacFarlane, 2003): the 5 handshape is deemed easy and frequent; the 1 handshape moderate but still frequent; and the U handshape “less frequent but not rare” in usage (Morere et al., 2012, p. 145). Forty-eight deaf college students generated 15–57 total signs for the combined 5, 1, and U stimuli (M = 35.0, SD = 9.9) (Morere et al., 2012).
In contrast to the 51U task, Marshall et al. (2014) limited BSL sign generation to six phonological categories. These included signs with three different handshapes (G or index finger; deemed most common in BSL), claw 5 (all digits apart and slightly bent), and I (extended pinky; deemed rare in BSL); signs produced within two specific locations (i.e., above the shoulders and palm of the non-dominant hand); and signs produced with one specific movement (two hands with both hands moving). Deaf adults generated a mean of 10.9 (range 3–19) correct items for the index handshape; 11.2 (range 4–22) for the claw 5 handshape; and 7.4 (range 4–14) for the I handshape (Marshall et al., 2014).
Marshall et al. (2014) noted that fluency was lower and apparently more difficult for phonological categories than semantic categories for their sample of deaf adults. While they performed similar to hearing adults on the semantic generation tasks, some deaf adults repeated given examples from instructions within their responses and generated non-signs to fit the given categories in the phonological tasks, something not observed in spoken language populations. Marshall et al. reported more errors, more repetitions of responses, and fewer total responses for the phonological tasks than the semantic tasks and attributed these differences to the complexity of the phonological tasks. Similar to Denmark & Atkinson (2015), Marshall et al. (2014) proposed that reduced experience with the explicit manipulation of phonological elements of signs inhibited responses, in contrast to typically hearing children, who engage in phonological play through nursery rhymes and English/Language Arts (ELA) instruction. Currently, most deaf children do not receive ASL Arts instruction equivalent to ELA instruction (Easterbrooks & Beal-Alvarez, 2013). Additionally, the simultaneous formation of signs via handshape, location, and movement arguably makes parameter extraction more difficult than segmentation of sequential spoken language units (Marshall et al., 2014).
Factors that Affect Fluency
Several factors affect children's verbal fluency, including chronological age, age of language acquisition, amount of language exposure, home language, and presence of a disability. Compared to same-aged hearing peers, younger and older deaf children tend to have similar knowledge of frequently occurring vocabulary and categorization strategies but they differ in the depth and breadth of their word knowledge and use of these strategies (Anderson & Reilly, 2002; Coppens, Tellings, Verhoeven, & Schreuder, 2013; Courtin, 1997; Koh, Vernon, & Bailey, 1971; Liben, 1979; Marschark & Everhart, 1999; Marshall et al., 2013; McEvoy et al., 1999; Ormel et al., 2010; Tweney et al., 1975; Wechsler-Kashi, Schwartz, & Cleary, 2014). Additionally, word knowledge tends to increase with chronological age, although variation exists within age groups (Beal-Alvarez, 2014, 2016; Mann, Roy, & Marshall, 2013). Later exposure to sign language results in later acquisition and less proficient language use and has been linked to reduced word knowledge (Chamberlain & Mayberry, 2000; Cuetos, Monsalve, Pinto, & Rodriguez-Ferreiro, 2004; Emmorey & Corina, 1990; Hildebrandt & Corina, 2002; Mann, Roy, & Marshall, 2013; Mayberry & Eichen, 1991; Mayberry & Lock, 2003; Mayberry & Witcher, 2005; Morford, Grieve-Smith, MacFarlane, Staley, & Waters, 2008). These language limitations may result in lower word generation due to a smaller lexicon or different lexical organization and access patterns (Marschark, Convertino, McEnvoy, & Masteller, 2004; Witkin, 2014).
Amount of sign language exposure is frequently measured by years of attendance at a school for the deaf. Henner, Hoffmeister, Fish, Rosenburg, & DiDonna (2015) reported that the longer deaf children with hearing parents attended schools for the deaf, the better they performed on ASL vocabulary and syntax assessments. Age of language acquisition predicted deaf children and adults’ picture naming performance (Cuetos et al., 2004) but not adults’ performance on an item exclusion task or their signed generation of items within given categories (Choubsaz & Gheitury, 2016). Similarly, Marshall et al. reported no significant item generation differences by adults’ age of BSL acquisition (2014) or deaf children's years of BSL exposure (2013).
In contrast, age of acquisition affects signers’ phonological processing. For instance, native signing adults (i.e., learned ASL from birth) did significantly better than early (i.e., exposed before 8 years of age) and late adult signers (i.e., after age 8 and largely in their teenage years) on a task that required them to organize given handshapes and movements into real signs (Corina et al., 2014). Similar results have been reported for other sign language lexical processing tasks (Carreiras, Gutierrez-Sigut, Baquero, & Corina, 2008; Emmorey & Corina, 1990; Hildebrandt & Corina, 2002; Morford & Carlson, 2011; Morford et al., 2008). Conflicting age of acquisition and exposure results may be tied more to the specific tasks at hand. In their review of the critical period of language learning, Choubsaz & Gheitury (2016) concluded that age of language acquisition may have differential effects according to the area of language in question, affecting mainly syntax and not semantics (Chomsky, 2000; Jackendoff, 2002).
Language modality (auditory versus visual) also may affect semantic and phonological generation (Brownfeld, 2008). ASL has a much smaller lexicon than English because single signs are simultaneously complemented by non-manual signals in lieu of a collection of synonyms, as in English (Marshall et al., 2014; Witkin, 2014). Additionally, fingerspelling of specific items requires additional time compared to sign generation, affecting the overall number of items generated within a given time period. Morere et al. (2012) provide the example of generation of the sign FISH, followed by fingerspelling different types of fish, which requires activation of English words and results in slower response time.
Home language may affect children's language skills as well. Around a quarter of deaf children are considered deaf/hard of hearing multilingual learners (DMLs), meaning a language other than English is used in the home (GRI, 2013). For the majority of these children, including those from Puerto Rico, the home language is Spanish (GRI, 2013; Ryan, 2013). Based on a survey of 336 deaf and hard of hearing school-aged children in Puerto Rico, 98% had Spanish as a home language, with 6% of families using English and 7% using ASL (Albertorio, Holden-Pitt, & Rawlings, 1999; Williams & Parks, 2012). Rodriquez (1993) noted that the majority of deaf children in Puerto Rico are exposed to some type of signing and use a combination of ASL, Puerto Rican Sign Language (PRSL), spoken Spanish, fingerspelling, gestures, and home signs with their peers and deaf adults. ASL is frequently used in Puerto Rico as a result of migration to the USA for educational services, contact with members of the Deaf community in the USA, use of ASL in schools for the deaf and university programs, and the recent increase of interpreting services between Puerto Rico and the USA (Rodriquez, 2001; Williams & Parks, 2012). Williams and Parks noted that PRSL is still used in less populated areas of Puerto Rico, such as the western and central areas of the country, which may result in some variation in sign production, including creation of unique signs; however, some locals indicate that ASL and PRSL do not differ greatly (Barish, 2009; Gonzalez, 2007; Rodriguez, 2001; Williams & Parks, 2012). Little is documented related to the signing skills of DMLs, specifically those from Puerto Rico, who may or may not exhibit variation in comparison to children from the USA on semantic and phonological generation tasks.
Gender may affect language skills. Previous researchers have noted advantages for males in visual–spatial tasks, such as mental rotation, and female advantages in language tasks, such as those related to vocabulary (see Goodman 2015 for a review). In samples of hearing adults, males and females produced similar overall scores but males produced fewer, larger clusters for both phonological and semantic tasks, while females switched among more clusters, suggesting use of different strategies during this task (Baron, 2004; Lanting et al., 2009). In the ASL 51U phonological generation task, Goodman (2015) reported no significant differences by gender for college-aged deaf students.
Finally, the presence of a disability, in addition to deafness, likely affects language skills, although limited data are available to support this conclusion (Henner, 2016; Marshall et al., 2013). The incidence of a disability co-occurring with deafness appears to be around 40% and types of disabilities have remained consistent across three years of GRI's Annual Survey data (Cupples et al., 2014; GRI, 2011, 2013). Beal-Alvarez (2016) reported that most deaf with disabilities in her sample (predominantly intellectual disabilities) scored within the average to low-average range on the ASL Receptive Skills Test (Enns, Zimmer, Boudreault, Rabu, & Broszeit, 2013) and showed less growth across time than their deaf peers without disabilities. Mann, Roy, & Marshall (2013) reported that deaf children with various disabilities performed similar to their deaf peers without disabilities on receptive and expressive measures of BSL vocabulary but with greater variation among scores. Expressively, deaf children with SLI performed similar to their peers without disabilities on a sign generation task but made more sign retrieval errors and accessed signs less fluently (Marshall et al., 2013).
Purpose
Little is known about signing deaf children, adolescents, and adults’ semantic and phonological sign generation ability and how they should be expected to perform across ages. Even less is known related to geographical effects, specifically deaf children in Puerto Rico. Previous researchers called for normative data for semantic and phonological tasks for the deaf population, including both adults and children (Denmark & Atkinson, 2015; Morere et al., 2012). The purpose of this study was to document the semantic and phonological sign generation performance of deaf children and adults who used ASL and investigate the effects of age, age of sign language acquisition, years of sign language exposure, gender, disability, and geographical location to provide insight on deaf children's development within these tasks.
Methods
Participants
This study included three groups of participants: deaf adults and deaf children from the USA, and deaf children from Puerto Rico. Within our study, we refer to the children as “students” due to the included age range. Similar to the USA, ASL is used in Puerto Rico, with some variation related to PRSL (Rodriquez, 2001; Williams & Parks, 2012). All 18 deaf adults (aged 22–50; M = 34.6, SD = 8.3) worked at one of two schools for the deaf in the same southeastern state within the USA, had hearing parents, varied in their language experiences and age of sign language acquisition, attended a variety of school settings in their educational experiences, used sign language as their primary mode of communication, and had some college experience or a master's degree. Adults completed only the animal and fruit generation tasks because we were unaware of the 51U task at the time of adult data collection (see Table 1).
Adult . | Age . | Gender . | Ethnicity . | Education . | Home language . | AoAa for ASLb . | Preferred language . | Schoolc . | Animals . | Animal clusters . | Fruits . |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | 29 | F | Black | BAd | SEe | 3 | ASL | L, R, R | 22 | 3 | 18 |
2 | 45 | F | White | MAf | Spoken English | 17 | ASL | D, L, L | 26 | 6 | 17 |
3 | 42 | F | White | BA | ASL, Spoken English | 4 | ASL/SE | L, D, D | 16 | 5 | 13 |
4 | 47 | F | White | MA | Spoken English | 12 | ASL | D, L, L | 22 | 5 | 15 |
5 | 37 | F | Hispanic | MA | ASL, Spoken English | 18 | ASL | L, L, L | 21 | 4 | 9 |
6 | 25 | F | White | BA | ASL, Spoken English, Pidgin | 2 | ASL/SE | R, L, L | 15 | 4 | 10 |
7 | 50 | F | Black | BA | Spoken English | 12 | ASL | L, L, L | 11 | 3 | 10 |
8 | 34 | M | White | MA | SE | 1 | Pidgin | L, L, L | 28 | 6 | 21 |
9 | 41 | M | Black | MA | Home Sign | 3 | ASL | L, L, L/R | 17 | 5 | 11 |
10 | 34 | F | White | MA | ASL | 1 | ASL | L , L, L | 20 | 4 | 17 |
11 | 30 | M | White | BA | SE | 2 | Pidgin | D, R, R/L | 21 | 6 | 11 |
12 | 36 | M | White | MA | Spoken English | 19 | ASL | D, D, L | 18 | 5 | 9 |
13 | 22 | M | White | BA | ASL | 2 | ASL | R, R, R | 27 | 6 | 11 |
14 | 23 | M | White | BA | SE, Spoken English | 2 | ASL | R, R, L | 21 | 5 | 8 |
15 | 31 | F | Hispanic | MA | SE, Spoken Spanish | 4 | ASL | L, L, L | 30 | 6 | 15 |
16 | 25 | M | White | BA | Home Sign | 3 | SE | L, L, L | 27 | 6 | 11 |
17 | 32 | M | White | BA | SE | 18 | ASL | L, L, R | 26 | 6 | 15 |
18 | 39 | F | Hispanic | HSg | SE, ASL, Spoken English | 4 | ASL | D, D, L | 14 | 4 | 26h |
Adult . | Age . | Gender . | Ethnicity . | Education . | Home language . | AoAa for ASLb . | Preferred language . | Schoolc . | Animals . | Animal clusters . | Fruits . |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | 29 | F | Black | BAd | SEe | 3 | ASL | L, R, R | 22 | 3 | 18 |
2 | 45 | F | White | MAf | Spoken English | 17 | ASL | D, L, L | 26 | 6 | 17 |
3 | 42 | F | White | BA | ASL, Spoken English | 4 | ASL/SE | L, D, D | 16 | 5 | 13 |
4 | 47 | F | White | MA | Spoken English | 12 | ASL | D, L, L | 22 | 5 | 15 |
5 | 37 | F | Hispanic | MA | ASL, Spoken English | 18 | ASL | L, L, L | 21 | 4 | 9 |
6 | 25 | F | White | BA | ASL, Spoken English, Pidgin | 2 | ASL/SE | R, L, L | 15 | 4 | 10 |
7 | 50 | F | Black | BA | Spoken English | 12 | ASL | L, L, L | 11 | 3 | 10 |
8 | 34 | M | White | MA | SE | 1 | Pidgin | L, L, L | 28 | 6 | 21 |
9 | 41 | M | Black | MA | Home Sign | 3 | ASL | L, L, L/R | 17 | 5 | 11 |
10 | 34 | F | White | MA | ASL | 1 | ASL | L , L, L | 20 | 4 | 17 |
11 | 30 | M | White | BA | SE | 2 | Pidgin | D, R, R/L | 21 | 6 | 11 |
12 | 36 | M | White | MA | Spoken English | 19 | ASL | D, D, L | 18 | 5 | 9 |
13 | 22 | M | White | BA | ASL | 2 | ASL | R, R, R | 27 | 6 | 11 |
14 | 23 | M | White | BA | SE, Spoken English | 2 | ASL | R, R, L | 21 | 5 | 8 |
15 | 31 | F | Hispanic | MA | SE, Spoken Spanish | 4 | ASL | L, L, L | 30 | 6 | 15 |
16 | 25 | M | White | BA | Home Sign | 3 | SE | L, L, L | 27 | 6 | 11 |
17 | 32 | M | White | BA | SE | 18 | ASL | L, L, R | 26 | 6 | 15 |
18 | 39 | F | Hispanic | HSg | SE, ASL, Spoken English | 4 | ASL | D, D, L | 14 | 4 | 26h |
aAge of acquisition.
bAmerican Sign Language.
CElementary, middle school, high school, D = day school for the deaf, L = local school, R = residential school for the deaf.
dBachelor's degree or some college experience.
eSigned English.
fMaster's degree.
gHigh school.
hFoods (not fruits).
Adult . | Age . | Gender . | Ethnicity . | Education . | Home language . | AoAa for ASLb . | Preferred language . | Schoolc . | Animals . | Animal clusters . | Fruits . |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | 29 | F | Black | BAd | SEe | 3 | ASL | L, R, R | 22 | 3 | 18 |
2 | 45 | F | White | MAf | Spoken English | 17 | ASL | D, L, L | 26 | 6 | 17 |
3 | 42 | F | White | BA | ASL, Spoken English | 4 | ASL/SE | L, D, D | 16 | 5 | 13 |
4 | 47 | F | White | MA | Spoken English | 12 | ASL | D, L, L | 22 | 5 | 15 |
5 | 37 | F | Hispanic | MA | ASL, Spoken English | 18 | ASL | L, L, L | 21 | 4 | 9 |
6 | 25 | F | White | BA | ASL, Spoken English, Pidgin | 2 | ASL/SE | R, L, L | 15 | 4 | 10 |
7 | 50 | F | Black | BA | Spoken English | 12 | ASL | L, L, L | 11 | 3 | 10 |
8 | 34 | M | White | MA | SE | 1 | Pidgin | L, L, L | 28 | 6 | 21 |
9 | 41 | M | Black | MA | Home Sign | 3 | ASL | L, L, L/R | 17 | 5 | 11 |
10 | 34 | F | White | MA | ASL | 1 | ASL | L , L, L | 20 | 4 | 17 |
11 | 30 | M | White | BA | SE | 2 | Pidgin | D, R, R/L | 21 | 6 | 11 |
12 | 36 | M | White | MA | Spoken English | 19 | ASL | D, D, L | 18 | 5 | 9 |
13 | 22 | M | White | BA | ASL | 2 | ASL | R, R, R | 27 | 6 | 11 |
14 | 23 | M | White | BA | SE, Spoken English | 2 | ASL | R, R, L | 21 | 5 | 8 |
15 | 31 | F | Hispanic | MA | SE, Spoken Spanish | 4 | ASL | L, L, L | 30 | 6 | 15 |
16 | 25 | M | White | BA | Home Sign | 3 | SE | L, L, L | 27 | 6 | 11 |
17 | 32 | M | White | BA | SE | 18 | ASL | L, L, R | 26 | 6 | 15 |
18 | 39 | F | Hispanic | HSg | SE, ASL, Spoken English | 4 | ASL | D, D, L | 14 | 4 | 26h |
Adult . | Age . | Gender . | Ethnicity . | Education . | Home language . | AoAa for ASLb . | Preferred language . | Schoolc . | Animals . | Animal clusters . | Fruits . |
---|---|---|---|---|---|---|---|---|---|---|---|
1 | 29 | F | Black | BAd | SEe | 3 | ASL | L, R, R | 22 | 3 | 18 |
2 | 45 | F | White | MAf | Spoken English | 17 | ASL | D, L, L | 26 | 6 | 17 |
3 | 42 | F | White | BA | ASL, Spoken English | 4 | ASL/SE | L, D, D | 16 | 5 | 13 |
4 | 47 | F | White | MA | Spoken English | 12 | ASL | D, L, L | 22 | 5 | 15 |
5 | 37 | F | Hispanic | MA | ASL, Spoken English | 18 | ASL | L, L, L | 21 | 4 | 9 |
6 | 25 | F | White | BA | ASL, Spoken English, Pidgin | 2 | ASL/SE | R, L, L | 15 | 4 | 10 |
7 | 50 | F | Black | BA | Spoken English | 12 | ASL | L, L, L | 11 | 3 | 10 |
8 | 34 | M | White | MA | SE | 1 | Pidgin | L, L, L | 28 | 6 | 21 |
9 | 41 | M | Black | MA | Home Sign | 3 | ASL | L, L, L/R | 17 | 5 | 11 |
10 | 34 | F | White | MA | ASL | 1 | ASL | L , L, L | 20 | 4 | 17 |
11 | 30 | M | White | BA | SE | 2 | Pidgin | D, R, R/L | 21 | 6 | 11 |
12 | 36 | M | White | MA | Spoken English | 19 | ASL | D, D, L | 18 | 5 | 9 |
13 | 22 | M | White | BA | ASL | 2 | ASL | R, R, R | 27 | 6 | 11 |
14 | 23 | M | White | BA | SE, Spoken English | 2 | ASL | R, R, L | 21 | 5 | 8 |
15 | 31 | F | Hispanic | MA | SE, Spoken Spanish | 4 | ASL | L, L, L | 30 | 6 | 15 |
16 | 25 | M | White | BA | Home Sign | 3 | SE | L, L, L | 27 | 6 | 11 |
17 | 32 | M | White | BA | SE | 18 | ASL | L, L, R | 26 | 6 | 15 |
18 | 39 | F | Hispanic | HSg | SE, ASL, Spoken English | 4 | ASL | D, D, L | 14 | 4 | 26h |
aAge of acquisition.
bAmerican Sign Language.
CElementary, middle school, high school, D = day school for the deaf, L = local school, R = residential school for the deaf.
dBachelor's degree or some college experience.
eSigned English.
fMaster's degree.
gHigh school.
hFoods (not fruits).
All students (5–21 years old) currently attended a school for the deaf and all but five had hearing parents. All U.S. students attended the same school; Puerto Rican students attended two different schools. Three U.S. students, one aged 5;3 and two aged 11;11, and two Puerto Rican students, aged 14;11 and 17;8, had deaf parents. Students from the USA attended a residential school for a range of 1–12 years, while students from Puerto Rico attended day schools for 1–16 years. The number of U.S. students included per task varied for two reasons. First, middle school age was the suggested age to begin administration of the 51U task (D. Morere, personal communication, September 6, 2014) and it was not administered to deaf students with disabilities, although these students completed the animal and food tasks. Second, video-recording errors resulted in an inability to code a few videos. Therefore, results are included for 68 students and 12 students with disabilities from the USA for the animal task; 63 students and 13 students with disabilities for the food task; and 48 students for the 51U task. All 17 students from Puerto Rico, aged 13;2–19;7, completed all five tasks. None of the Puerto Rican students had identified disabilities.
Procedures
Deaf adults volunteered to participate in this study. All completed a background data sheet. Data for students were obtained from school records. All sessions for all participants were completed one-on-one and video-recorded. The first author administered the animal and fruit tasks individually to the deaf adults at the schools in an unoccupied classroom. The first author and four trained graduate students in an interpreting preparation program administered the assessments to the students in the school media center, unoccupied classrooms, or dorm social area. Assessors prompted participants to generate as many animal signs as they could in one minute and followed the same procedures for foods (or fruits for adults). No examples of animal or food items were given. Assessors used their phones on the table top as stopwatches.
Unlike Marshall et al. (2014), who presented their phonological task instructions using a live deaf native signer, we used a video created by a deaf adult signer who taught ASL at an unrelated school for the deaf and who translated the task directions from English (Morere et al., 2012) into ASL. Similar to Marshall et al. (2014) examples (e.g., SKUNK) and non-examples (e.g., K-R-Y-S-T-L-E with name sign) were included in the 51U directions due to the “difficult and unintuitive” demands of the phonological task (Marshall et al., 2014, p. 594). Students were directed to generate as many signs with each handshape as they could within one minute in the following order: 5, 1, and U.
Data Analysis
Video recordings were transcribed into English glosses, entered into an Excel data base, and analyzed using SPSS. Similar to others, we calculated the number of responses per task by participant (Marshall et al., 2013, 2014; Morere et al., 2012) and coded each response as correct or incorrect (e.g., incorrect responses such as MONSTER for animals, DRINK for foods, pseudosigns or incorrect parameters for 51U) (Marshall et al., 2014). We calculated the number of fingerspelled responses across the semantic tasks (Marshall et al., 2013, 2014). Additionally, thematic and taxonomic categories emerged from the data, similar to Marshall et al.’s studies (2013, 2014). We calculated the total number of thematic (e.g., zoo or farm animals; breakfast) and taxonomic (e.g., birds or reptiles; meats or fruits) clusters for the animal and food tasks and the number of items within clusters to calculate cluster size (Marshall et al., 2013, 2014). In contrast to Witkin (2014) and similar to others (Marshall et al., 2013, 2014; Weschler-Kashi, Schwartz, & Cleary, 2014), we defined a cluster as two or more sequentially produced related items. Finally, in contrast to Marshall et al. (2014), who asked adult participants to suppress pointing during the index handshape generation task and speculated that this inhibition may have resulted in a cognitive expense in item generation, we counted signs that used the index finger for pointing, such as EYES, NOSE, and YOU (Morere et al., 2012, provided no data on coding pointing for the 1 handshape).
For adults and students, we calculated the effects of chronological age on overall scores and cluster size using correlations. For adults, we used one-way ANOVAs to investigate the effects of age of acquisition of ASL (i.e., early (0–4 years) and late (12–19 years)), gender, and education level (bachelor's degree/some college or graduate degree), on overall scores, number of animal clusters, and cluster size. For students, we used one-way ANOVAs to investigate the effects of years at the school site (1–12, as a proxy for age of exposure; Henner et al., 2015), gender, and geographical location (USA or Puerto Rico). We investigated students identified with a disability as a separate group. Similar to Marshall et al. (2013), we analyzed differences in animal and food performance across students who completed each task using paired t-tests. Finally, we analyzed item frequency and the frequency of fingerspelling in both semantic tasks.
Inter-observer agreement (IOA) for participant responses was independently conducted for 95 (20%) randomly selected videos across tasks by the first author, whose native language is spoken English, lives in the USA, has an Advanced Plus rating on the Sign Language Proficiency Interview (SLPI; Newell, Caccamise, Boardman, & Holcomb, 1983), and 8 years of experience teaching deaf students, and the second author, whose native language is spoken Spanish, lives in Puerto Rico, has 11 years of experience using sign language, and 2 years of experience teaching English to deaf students. Agreement was as follows: animals, 94%; foods, 84%; 5, 87%; 1, 81%; U, 77%. Additionally, a graduate student whose native language was spoken English and who had both deaf education and interpreting experience independently coded 5% of the videos, randomly selected, with agreement as follows: animals, 83%; food, 96%; 5, 86%; 1, 88%; and U, 68%. All three raters noted general difficulty in identification of specific signs produced by students in the 51U task due to the lack of context for generated signs, sloppy signing produced by some students, the intrusion of home signs, and variation or incorrect parameter production within signs. Additionally, the low number of responses for the 1 and U tasks provided greater weight to each disagreement, thereby inflating IOA.
Results
We asked: How does ASL semantic and phonological fluency, measured by the animal, food, and 51U sign generation tasks, develop across signing deaf adults and signing students from the USA and Puerto Rico? For deaf adults, we investigated overall performance and cluster performance by chronological age with correlations, and by age of ASL acquisition, gender, and education level with one-way ANOVAs. For students, we investigated overall and cluster performance by chronological age with correlations and by years at school sites, gender, and geographical location with one-way ANOVAs. We investigated the effects of disabilities as a separate group. For both adults and students, we analyzed differences in animal and food performance using paired t-tests. We present only significant findings across these factors, first for deaf adults followed by students.
Deaf Adults
Animals
Adults generated a mean of 20.4 animals (SD = 4.5; range 11–30) (see Table 1). No correlations or differences among adult factors listed above were significant. Dog and cat, tiger, and elephant were the most frequently generated first, second, and third items, respectively (24–29%). All other responses were diverse across animals. Adults generated a mean of 4.9 animal clusters (SD = 1.0) (i.e., 2 related animals sequentially). The mean number of cluster items was 3.0 (range 2.2–4.3; SD = 0.6). While pets was the most frequently generated first cluster, water animals (21%) and zoo (19%) were the most frequent clusters overall, followed by farm and pets (13% each), and reptile (11%). Seven adults (39%) repeated clusters, including water animals (4), zoo (3), and farm (1). Half of the participants used fingerspelling for a total of 18 responses (5.3%) and a mean of 1.0 animals (SD = 1.9; range 0–8).
Fruits
Adults generated a mean of 12.7 fruits (SD = 3.5; range 8–21). Again, no correlations or differences among factors listed above were significant. Apple (53%) and strawberry (18%) were the most frequently generated first items. Orange (29%) and banana (24%) were the most frequently generated second items. Banana and grapes (24% each) were the most frequently generated third items. Clusters were not analyzed for the fruit task. All but one adult used fingerspelling within their item responses, with a mean of 4.2 fingerspelled fruit items (SD = 2.0; range 1–9). Fingerspelled items represented 33% of adult responses. Commonly fingerspelled fruits included KIWI, MANGO, AVOCADO, APRICOT, CLEMENTINE, CANTALOUPE, GRAPEFRUIT, PEAR, PLUM, and various types of berries. Fingerspelling of one fruit seemed to prompt sequential recall of other fingerspelled fruits, such as KIWI, MANGO, PAPAYA, and PLATANOS; and CHERRY, BLUEBERRY, RASPBERRY, and PINEAPPLE in succession. The sole adult who completed the food task (DA 18) generated 26 foods and 3 clusters (breakfast, vegetables, and sweets). Adult performance between the two semantic categories significantly correlated (r = 0.484, p = .049) and was significantly different (animals: M = 21.7, SD = 5.2; fruits: M = 13.0, SD = 3.7) (t (17) = −7.56, p < .001).
Deaf Students
For students, we calculated overall animal and food scores, number of clusters and items within clusters, and differences between overall scores and cluster scores and the following factors: chronological age, years at the school site, gender, presence of a disability, and geographical location. We present only significant findings among these factors.
Animals
Sixty-eight U.S. students, aged 5;3–21;8, generated a range of 2–25 animals. Thirteen students with disabilities, aged 9;3–21;3, generated a range of 1–17 animals (see Table 2 for students without disabilities and Table 3 for students with disabilities). Animal score strongly and significantly correlated with age only for students without disabilities (r = 0.500, p ≥ .001). Animal score weakly but significantly correlated with years at school site only for students without disabilities (r = 0.26, p = .031). Scores for age-band and years at school site were not calculated for students with disabilities across categories due to the low number of participants. No other correlations or differences among factors were significant.
Age . | Animals . | Foods . | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Overall . | Clusters . | Items . | Overall . | Clusters . | Items . | ||||||||||
Na . | M . | SD . | M . | SD . | M . | SD . | N . | M . | SD . | N . | M . | SD . | M . | SD . | |
USA | |||||||||||||||
5–6 | 1 | 6.0 | 0.0 | 2.0 | 0.0 | 2.0 | 2.0 | 1 | 1.0 | 0.0 | 1 | 1.0 | 0.0 | 3.0 | 0.0 |
7–8 | 5 | 10.0 | 2.9 | 2.0 | 1.3 | 3.8 | 1.9 | 4 | 6.0 | 1.8 | 2 | 1.0 | 0.0 | 2.0 | 0.0 |
9–10 | 7 | 9.1 | 1.9 | 2.0 | 0.6 | 3.5 | 1.7 | 5 | 7.2 | 1.9 | 4 | 1.3 | 0.3 | 2.1 | 0.3 |
11–12 | 7 | 12.1 | 6.7 | 3.3 | 2.3 | 2.5 | 0.6 | 7 | 11.1 | 6.3 | 6 | 1.5 | 0.8 | 2.7 | 0.8 |
13–14 | 8 | 14.1 | 4.3 | 3.5 | 2.3 | 4.7 | 3.3 | 8 | 14.7 | 4.4 | 8 | 2.6 | 0.9 | 2.7 | 0.6 |
15–16 | 14 | 15.9 | 4.1 | 4.0 | 1.6 | 2.6 | 0.6 | 12 | 11.3 | 2.1 | 11 | 1.9 | 1.1 | 2.6 | 0.6 |
17–18 | 13 | 15.2 | 4.7 | 4.0 | 1.6 | 3.0 | 0.6 | 13 | 13.1 | 4.8 | 11 | 1.8 | 1.1 | 3.9 | 2.8 |
19–21 | 13 | 16.5 | 4.5 | 3.5 | 1.2 | 3.3 | 0.9 | 13 | 14.0 | 4.1 | 12 | 2.3 | 1.2 | 2.7 | 0.5 |
Puerto Rico | |||||||||||||||
13–14 | 2 | 12.5 | 2.1 | 2.0 | 0.0 | 4.5 | 3.5 | 2 | 10.0 | 2.8 | 2 | 1.0 | 0.0 | 5.5 | 2.1 |
15–16 | 3 | 20.7 | 6.5 | 3.7 | 2.1 | 3.2 | 1.4 | 3 | 14.0 | 2.0 | 3 | 2.3 | 0.6 | 3.1 | 0.8 |
17–18 | 9 | 16.3 | 3.1 | 3.8 | 1.2 | 2.9 | 0.7 | 9 | 14.0 | 2.8 | 9 | 1.9 | 1.1 | 2.9 | 1.2 |
19–21 | 3 | 18.7 | 4.9 | 4.7 | 2.3 | 3.0 | 0.4 | 3 | 14.3 | 3.2 | 3 | 2.3 | 1.5 | 2.6 | 0.6 |
Age . | Animals . | Foods . | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Overall . | Clusters . | Items . | Overall . | Clusters . | Items . | ||||||||||
Na . | M . | SD . | M . | SD . | M . | SD . | N . | M . | SD . | N . | M . | SD . | M . | SD . | |
USA | |||||||||||||||
5–6 | 1 | 6.0 | 0.0 | 2.0 | 0.0 | 2.0 | 2.0 | 1 | 1.0 | 0.0 | 1 | 1.0 | 0.0 | 3.0 | 0.0 |
7–8 | 5 | 10.0 | 2.9 | 2.0 | 1.3 | 3.8 | 1.9 | 4 | 6.0 | 1.8 | 2 | 1.0 | 0.0 | 2.0 | 0.0 |
9–10 | 7 | 9.1 | 1.9 | 2.0 | 0.6 | 3.5 | 1.7 | 5 | 7.2 | 1.9 | 4 | 1.3 | 0.3 | 2.1 | 0.3 |
11–12 | 7 | 12.1 | 6.7 | 3.3 | 2.3 | 2.5 | 0.6 | 7 | 11.1 | 6.3 | 6 | 1.5 | 0.8 | 2.7 | 0.8 |
13–14 | 8 | 14.1 | 4.3 | 3.5 | 2.3 | 4.7 | 3.3 | 8 | 14.7 | 4.4 | 8 | 2.6 | 0.9 | 2.7 | 0.6 |
15–16 | 14 | 15.9 | 4.1 | 4.0 | 1.6 | 2.6 | 0.6 | 12 | 11.3 | 2.1 | 11 | 1.9 | 1.1 | 2.6 | 0.6 |
17–18 | 13 | 15.2 | 4.7 | 4.0 | 1.6 | 3.0 | 0.6 | 13 | 13.1 | 4.8 | 11 | 1.8 | 1.1 | 3.9 | 2.8 |
19–21 | 13 | 16.5 | 4.5 | 3.5 | 1.2 | 3.3 | 0.9 | 13 | 14.0 | 4.1 | 12 | 2.3 | 1.2 | 2.7 | 0.5 |
Puerto Rico | |||||||||||||||
13–14 | 2 | 12.5 | 2.1 | 2.0 | 0.0 | 4.5 | 3.5 | 2 | 10.0 | 2.8 | 2 | 1.0 | 0.0 | 5.5 | 2.1 |
15–16 | 3 | 20.7 | 6.5 | 3.7 | 2.1 | 3.2 | 1.4 | 3 | 14.0 | 2.0 | 3 | 2.3 | 0.6 | 3.1 | 0.8 |
17–18 | 9 | 16.3 | 3.1 | 3.8 | 1.2 | 2.9 | 0.7 | 9 | 14.0 | 2.8 | 9 | 1.9 | 1.1 | 2.9 | 1.2 |
19–21 | 3 | 18.7 | 4.9 | 4.7 | 2.3 | 3.0 | 0.4 | 3 | 14.3 | 3.2 | 3 | 2.3 | 1.5 | 2.6 | 0.6 |
aIncludes only students who generated clusters.
Age . | Animals . | Foods . | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Overall . | Clusters . | Items . | Overall . | Clusters . | Items . | ||||||||||
Na . | M . | SD . | M . | SD . | M . | SD . | N . | M . | SD . | N . | M . | SD . | M . | SD . | |
USA | |||||||||||||||
5–6 | 1 | 6.0 | 0.0 | 2.0 | 0.0 | 2.0 | 2.0 | 1 | 1.0 | 0.0 | 1 | 1.0 | 0.0 | 3.0 | 0.0 |
7–8 | 5 | 10.0 | 2.9 | 2.0 | 1.3 | 3.8 | 1.9 | 4 | 6.0 | 1.8 | 2 | 1.0 | 0.0 | 2.0 | 0.0 |
9–10 | 7 | 9.1 | 1.9 | 2.0 | 0.6 | 3.5 | 1.7 | 5 | 7.2 | 1.9 | 4 | 1.3 | 0.3 | 2.1 | 0.3 |
11–12 | 7 | 12.1 | 6.7 | 3.3 | 2.3 | 2.5 | 0.6 | 7 | 11.1 | 6.3 | 6 | 1.5 | 0.8 | 2.7 | 0.8 |
13–14 | 8 | 14.1 | 4.3 | 3.5 | 2.3 | 4.7 | 3.3 | 8 | 14.7 | 4.4 | 8 | 2.6 | 0.9 | 2.7 | 0.6 |
15–16 | 14 | 15.9 | 4.1 | 4.0 | 1.6 | 2.6 | 0.6 | 12 | 11.3 | 2.1 | 11 | 1.9 | 1.1 | 2.6 | 0.6 |
17–18 | 13 | 15.2 | 4.7 | 4.0 | 1.6 | 3.0 | 0.6 | 13 | 13.1 | 4.8 | 11 | 1.8 | 1.1 | 3.9 | 2.8 |
19–21 | 13 | 16.5 | 4.5 | 3.5 | 1.2 | 3.3 | 0.9 | 13 | 14.0 | 4.1 | 12 | 2.3 | 1.2 | 2.7 | 0.5 |
Puerto Rico | |||||||||||||||
13–14 | 2 | 12.5 | 2.1 | 2.0 | 0.0 | 4.5 | 3.5 | 2 | 10.0 | 2.8 | 2 | 1.0 | 0.0 | 5.5 | 2.1 |
15–16 | 3 | 20.7 | 6.5 | 3.7 | 2.1 | 3.2 | 1.4 | 3 | 14.0 | 2.0 | 3 | 2.3 | 0.6 | 3.1 | 0.8 |
17–18 | 9 | 16.3 | 3.1 | 3.8 | 1.2 | 2.9 | 0.7 | 9 | 14.0 | 2.8 | 9 | 1.9 | 1.1 | 2.9 | 1.2 |
19–21 | 3 | 18.7 | 4.9 | 4.7 | 2.3 | 3.0 | 0.4 | 3 | 14.3 | 3.2 | 3 | 2.3 | 1.5 | 2.6 | 0.6 |
Age . | Animals . | Foods . | |||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Overall . | Clusters . | Items . | Overall . | Clusters . | Items . | ||||||||||
Na . | M . | SD . | M . | SD . | M . | SD . | N . | M . | SD . | N . | M . | SD . | M . | SD . | |
USA | |||||||||||||||
5–6 | 1 | 6.0 | 0.0 | 2.0 | 0.0 | 2.0 | 2.0 | 1 | 1.0 | 0.0 | 1 | 1.0 | 0.0 | 3.0 | 0.0 |
7–8 | 5 | 10.0 | 2.9 | 2.0 | 1.3 | 3.8 | 1.9 | 4 | 6.0 | 1.8 | 2 | 1.0 | 0.0 | 2.0 | 0.0 |
9–10 | 7 | 9.1 | 1.9 | 2.0 | 0.6 | 3.5 | 1.7 | 5 | 7.2 | 1.9 | 4 | 1.3 | 0.3 | 2.1 | 0.3 |
11–12 | 7 | 12.1 | 6.7 | 3.3 | 2.3 | 2.5 | 0.6 | 7 | 11.1 | 6.3 | 6 | 1.5 | 0.8 | 2.7 | 0.8 |
13–14 | 8 | 14.1 | 4.3 | 3.5 | 2.3 | 4.7 | 3.3 | 8 | 14.7 | 4.4 | 8 | 2.6 | 0.9 | 2.7 | 0.6 |
15–16 | 14 | 15.9 | 4.1 | 4.0 | 1.6 | 2.6 | 0.6 | 12 | 11.3 | 2.1 | 11 | 1.9 | 1.1 | 2.6 | 0.6 |
17–18 | 13 | 15.2 | 4.7 | 4.0 | 1.6 | 3.0 | 0.6 | 13 | 13.1 | 4.8 | 11 | 1.8 | 1.1 | 3.9 | 2.8 |
19–21 | 13 | 16.5 | 4.5 | 3.5 | 1.2 | 3.3 | 0.9 | 13 | 14.0 | 4.1 | 12 | 2.3 | 1.2 | 2.7 | 0.5 |
Puerto Rico | |||||||||||||||
13–14 | 2 | 12.5 | 2.1 | 2.0 | 0.0 | 4.5 | 3.5 | 2 | 10.0 | 2.8 | 2 | 1.0 | 0.0 | 5.5 | 2.1 |
15–16 | 3 | 20.7 | 6.5 | 3.7 | 2.1 | 3.2 | 1.4 | 3 | 14.0 | 2.0 | 3 | 2.3 | 0.6 | 3.1 | 0.8 |
17–18 | 9 | 16.3 | 3.1 | 3.8 | 1.2 | 2.9 | 0.7 | 9 | 14.0 | 2.8 | 9 | 1.9 | 1.1 | 2.9 | 1.2 |
19–21 | 3 | 18.7 | 4.9 | 4.7 | 2.3 | 3.0 | 0.4 | 3 | 14.3 | 3.2 | 3 | 2.3 | 1.5 | 2.6 | 0.6 |
aIncludes only students who generated clusters.
Age . | Gender . | Disability . | Animals . | Clusters . | Foods . | Clusters . |
---|---|---|---|---|---|---|
9;3 | M | MIDa | 9 | 2 | 4 | 0 |
9;4 | F | MID | 3 | 0 | 6 | 1 |
9;7 | M | MID | 1 | 0 | 3 | 0 |
9;7 | F | MID | 7 | 2 | 8 | 1 |
12;9 | F | OHIb | 12 | 2 | 8 | 2 |
13;7 | M | AUTc | 5 | 1 | – | – |
13;9 | F | MID | 13 | 2 | 8 | 1 |
14;4 | M | AUT | 12 | 2 | 12 | 2 |
16;5 | F | MODd | 13 | 4 | 6 | 1 |
16;7 | M | MID | 7 | 1 | 6 | 1 |
19;8 | F | MID | 17 | 3 | 15 | 2 |
20;10 | F | MID | 2 | 0 | 10 | 0 |
21;3 | F | MOD | 7 | 1 | 9 | 1 |
Age . | Gender . | Disability . | Animals . | Clusters . | Foods . | Clusters . |
---|---|---|---|---|---|---|
9;3 | M | MIDa | 9 | 2 | 4 | 0 |
9;4 | F | MID | 3 | 0 | 6 | 1 |
9;7 | M | MID | 1 | 0 | 3 | 0 |
9;7 | F | MID | 7 | 2 | 8 | 1 |
12;9 | F | OHIb | 12 | 2 | 8 | 2 |
13;7 | M | AUTc | 5 | 1 | – | – |
13;9 | F | MID | 13 | 2 | 8 | 1 |
14;4 | M | AUT | 12 | 2 | 12 | 2 |
16;5 | F | MODd | 13 | 4 | 6 | 1 |
16;7 | M | MID | 7 | 1 | 6 | 1 |
19;8 | F | MID | 17 | 3 | 15 | 2 |
20;10 | F | MID | 2 | 0 | 10 | 0 |
21;3 | F | MOD | 7 | 1 | 9 | 1 |
–, indicates no data.
aMild intellectual disability.
bOther health impairment (syndrome).
cAutism.
dModerate intellectual disability.
Age . | Gender . | Disability . | Animals . | Clusters . | Foods . | Clusters . |
---|---|---|---|---|---|---|
9;3 | M | MIDa | 9 | 2 | 4 | 0 |
9;4 | F | MID | 3 | 0 | 6 | 1 |
9;7 | M | MID | 1 | 0 | 3 | 0 |
9;7 | F | MID | 7 | 2 | 8 | 1 |
12;9 | F | OHIb | 12 | 2 | 8 | 2 |
13;7 | M | AUTc | 5 | 1 | – | – |
13;9 | F | MID | 13 | 2 | 8 | 1 |
14;4 | M | AUT | 12 | 2 | 12 | 2 |
16;5 | F | MODd | 13 | 4 | 6 | 1 |
16;7 | M | MID | 7 | 1 | 6 | 1 |
19;8 | F | MID | 17 | 3 | 15 | 2 |
20;10 | F | MID | 2 | 0 | 10 | 0 |
21;3 | F | MOD | 7 | 1 | 9 | 1 |
Age . | Gender . | Disability . | Animals . | Clusters . | Foods . | Clusters . |
---|---|---|---|---|---|---|
9;3 | M | MIDa | 9 | 2 | 4 | 0 |
9;4 | F | MID | 3 | 0 | 6 | 1 |
9;7 | M | MID | 1 | 0 | 3 | 0 |
9;7 | F | MID | 7 | 2 | 8 | 1 |
12;9 | F | OHIb | 12 | 2 | 8 | 2 |
13;7 | M | AUTc | 5 | 1 | – | – |
13;9 | F | MID | 13 | 2 | 8 | 1 |
14;4 | M | AUT | 12 | 2 | 12 | 2 |
16;5 | F | MODd | 13 | 4 | 6 | 1 |
16;7 | M | MID | 7 | 1 | 6 | 1 |
19;8 | F | MID | 17 | 3 | 15 | 2 |
20;10 | F | MID | 2 | 0 | 10 | 0 |
21;3 | F | MOD | 7 | 1 | 9 | 1 |
–, indicates no data.
aMild intellectual disability.
bOther health impairment (syndrome).
cAutism.
dModerate intellectual disability.
Students generated 33 different animals across the first three items. The most frequently generated first items were dog (28%), cat (21%), and elephant (13%). The most frequently generated second items were cat (23%), dog (18%), and lion (7%). Third items generated showed more variation, with bear, cat, and dog nearly equal at about 8% and dispersion among horse, elephant, monkey, and cow, among others. For students with disabilities, cat (40%, 29%) and dog (20%, 21%) were the most frequently generated first and second items; multiple items were generated for the third item and a few of those items were non-animal errors (e.g., COLD, FARM, and APPLE).
Students without disabilities generated a range of 1–8 clusters per student (see Table 2). The majority of clusters were thematic (71%), including pets (22%), farm (22%), zoo (21%), water (11%), woods (3%), and flying animals (2%). Taxonomic categories (29%) included birds (10%), reptiles (6%), dogs (3%), monkeys (2%), rodents (2%), and various other animals, all less than 1% of responses. Age significantly and moderately correlated with the number of animal clusters students without disabilities generated (r = 0.368, p = .002) but not the number of cluster items (r = −0.044, p = .726). Cluster size ranged from 2 to 8, with the exception of one student who generated only one cluster with 12 items. Several students across ages generated one cluster with seven or more items, including eight students and three students with disabilities for zoo and one student each for farm and water animals, showing a depth of animal knowledge within subcategories. Additionally, most of these students generated other clusters as well.
About a quarter of students without disabilities, 9–21 years of age, repeated the same cluster a second time with different items. The majority of cluster repeats (69%) was zoo, with two repeats each for farm and reptile and one for bird. Students with disabilities mirrored the same cluster generation as their peers (zoo 44%, farm 31%, and pets 25%); however, they generated only three different clusters and fewer cluster items (see Table 4). Fingerspelled animal responses were infrequent and diverse, representing 3% of total responses (26 out of 856) generated by 13 different students without disabilities (19%), aged 11;11–20;0. Twenty-three different animals were fingerspelled, including COYOTE, JELLYFISH, LEOPARD, SEA HORSE, and WILD HOG, among others. Students with disabilities fingerspelled no animals.
Age years . | 5 . | 1 . | U . | Total . | |||||
---|---|---|---|---|---|---|---|---|---|
N . | M . | SD . | M . | SD . | M . | SD . | M . | SD . | |
USA | |||||||||
11–12 | 4 | 7.5 | 5.9 | 7.5 | 3.1 | 3.0 | 1.6 | 18.0 | 10.1 |
13–14 | 6 | 9.2 | 2.1 | 10.3 | 2.7 | 5.0 | 2.5 | 24.5 | 5.4 |
15–16 | 14 | 11.2 | 4.8 | 8.9 | 4.1 | 4.9 | 2.7 | 25.0 | 9.2 |
17–18 | 13 | 10.5 | 3.4 | 9.9 | 4.4 | 5.0 | 2.5 | 25.5 | 7.7 |
19–21 | 12 | 9.2 | 4.4 | 9.4 | 3.6 | 5.8 | 3.1 | 24.4 | 9.2 |
Puerto Rico | |||||||||
13–14 | 2 | 8.0 | 0.0 | 9.5 | 2.1 | 5.0 | 2.8 | 22.5 | 5.0 |
15–16 | 3 | 12.3 | 5.7 | 9.0 | 2.0 | 5.7 | 4.7 | 27.0 | 9.5 |
17–18 | 9 | 12.1 | 2.3 | 10.6 | 3.6 | 6.4 | 2.0 | 29.1 | 6.3 |
19–21 | 3 | 17.0 | 2.7 | 11.7 | 1.5 | 9.0 | 1.0 | 37.7 | 3.2 |
Age years . | 5 . | 1 . | U . | Total . | |||||
---|---|---|---|---|---|---|---|---|---|
N . | M . | SD . | M . | SD . | M . | SD . | M . | SD . | |
USA | |||||||||
11–12 | 4 | 7.5 | 5.9 | 7.5 | 3.1 | 3.0 | 1.6 | 18.0 | 10.1 |
13–14 | 6 | 9.2 | 2.1 | 10.3 | 2.7 | 5.0 | 2.5 | 24.5 | 5.4 |
15–16 | 14 | 11.2 | 4.8 | 8.9 | 4.1 | 4.9 | 2.7 | 25.0 | 9.2 |
17–18 | 13 | 10.5 | 3.4 | 9.9 | 4.4 | 5.0 | 2.5 | 25.5 | 7.7 |
19–21 | 12 | 9.2 | 4.4 | 9.4 | 3.6 | 5.8 | 3.1 | 24.4 | 9.2 |
Puerto Rico | |||||||||
13–14 | 2 | 8.0 | 0.0 | 9.5 | 2.1 | 5.0 | 2.8 | 22.5 | 5.0 |
15–16 | 3 | 12.3 | 5.7 | 9.0 | 2.0 | 5.7 | 4.7 | 27.0 | 9.5 |
17–18 | 9 | 12.1 | 2.3 | 10.6 | 3.6 | 6.4 | 2.0 | 29.1 | 6.3 |
19–21 | 3 | 17.0 | 2.7 | 11.7 | 1.5 | 9.0 | 1.0 | 37.7 | 3.2 |
Age years . | 5 . | 1 . | U . | Total . | |||||
---|---|---|---|---|---|---|---|---|---|
N . | M . | SD . | M . | SD . | M . | SD . | M . | SD . | |
USA | |||||||||
11–12 | 4 | 7.5 | 5.9 | 7.5 | 3.1 | 3.0 | 1.6 | 18.0 | 10.1 |
13–14 | 6 | 9.2 | 2.1 | 10.3 | 2.7 | 5.0 | 2.5 | 24.5 | 5.4 |
15–16 | 14 | 11.2 | 4.8 | 8.9 | 4.1 | 4.9 | 2.7 | 25.0 | 9.2 |
17–18 | 13 | 10.5 | 3.4 | 9.9 | 4.4 | 5.0 | 2.5 | 25.5 | 7.7 |
19–21 | 12 | 9.2 | 4.4 | 9.4 | 3.6 | 5.8 | 3.1 | 24.4 | 9.2 |
Puerto Rico | |||||||||
13–14 | 2 | 8.0 | 0.0 | 9.5 | 2.1 | 5.0 | 2.8 | 22.5 | 5.0 |
15–16 | 3 | 12.3 | 5.7 | 9.0 | 2.0 | 5.7 | 4.7 | 27.0 | 9.5 |
17–18 | 9 | 12.1 | 2.3 | 10.6 | 3.6 | 6.4 | 2.0 | 29.1 | 6.3 |
19–21 | 3 | 17.0 | 2.7 | 11.7 | 1.5 | 9.0 | 1.0 | 37.7 | 3.2 |
Age years . | 5 . | 1 . | U . | Total . | |||||
---|---|---|---|---|---|---|---|---|---|
N . | M . | SD . | M . | SD . | M . | SD . | M . | SD . | |
USA | |||||||||
11–12 | 4 | 7.5 | 5.9 | 7.5 | 3.1 | 3.0 | 1.6 | 18.0 | 10.1 |
13–14 | 6 | 9.2 | 2.1 | 10.3 | 2.7 | 5.0 | 2.5 | 24.5 | 5.4 |
15–16 | 14 | 11.2 | 4.8 | 8.9 | 4.1 | 4.9 | 2.7 | 25.0 | 9.2 |
17–18 | 13 | 10.5 | 3.4 | 9.9 | 4.4 | 5.0 | 2.5 | 25.5 | 7.7 |
19–21 | 12 | 9.2 | 4.4 | 9.4 | 3.6 | 5.8 | 3.1 | 24.4 | 9.2 |
Puerto Rico | |||||||||
13–14 | 2 | 8.0 | 0.0 | 9.5 | 2.1 | 5.0 | 2.8 | 22.5 | 5.0 |
15–16 | 3 | 12.3 | 5.7 | 9.0 | 2.0 | 5.7 | 4.7 | 27.0 | 9.5 |
17–18 | 9 | 12.1 | 2.3 | 10.6 | 3.6 | 6.4 | 2.0 | 29.1 | 6.3 |
19–21 | 3 | 17.0 | 2.7 | 11.7 | 1.5 | 9.0 | 1.0 | 37.7 | 3.2 |
Seventeen Puerto Rican students (aged 13;2–19;7) generated a range of 11–27 animal responses. There were no significant correlations or differences by age, gender, or years at school site. Students generated 15 different animals across the first three items, with dog (41%) the most popular for the first item, followed by cat (35%) and lion (12%). Cat was the most frequent second response (47%), followed by dog (24%), and rabbit was the most frequently generated third item (18%). Other second and third items varied across animals (i.e., cat, dog, elephant, lion, tiger, snake, turtle, and zebra). Students generated 2–6 clusters and every student generated at least one animal cluster. Clusters were more prevalent for thematic than taxonomic groups, including zoo (23%), pets (21%), water (18%), farm (7%), and birds (11%). Less frequent taxonomic clusters included reptiles, rodents, horses, and cats. Only two students (12%) repeated a cluster with different items. The mean number of items per cluster ranged from 2 to 9. Four students generated more than seven items within the category of zoo while also generating other clusters. Of note, there was no significant difference in the number of clusters produced between same-aged U.S. (N = 35, M = 3.05, SD = 1.70) and Puerto Rican students (N = 17, M = 3.16, SD = 1.39) (F [1, 50] = 0.054, p = .817). Only two responses (of 290), both PANDA bear, were fingerspelled by two Puerto Rican students.
Food
Sixty-three U.S. students, aged 5;3–21;8, generated 4–21 foods. Twelve deaf with disabilities students, aged 9;3–21;3, generated 3–15 foods. Food score moderately and significantly correlated with age for students without disabilities (r = 0.467, p < .001) and strongly and significantly for students with disabilities (r = 0.616, p = .033). Similar to animal scores, food scores weakly but significantly correlated with years at the school site only for students without disabilities (r = 0.267, p = .033). There were no other significant differences. Foods showed more dispersion than animals across first-, second-, and third-generated items, with a total of 43 different items across the first three responses for students without disabilities. Pizza was the most frequently generated first item (23%), followed by hamburger (15%), ice cream (8%), and French fries (8%). Pizza and hotdog tied for the most frequently generated second item (14%), followed by hamburger and chicken (11%). Hamburger also was the most frequent third item (9%), with chicken, salad, and strawberry tied for the second most frequently generated third food item (6%). Students with disabilities varied in their responses. Only pizza, French fries, apple, and orange received multiple responses across these participants.
Overall, students without disabilities generated about half as many clusters for food as animals (N = 112), with a range of 1–4 clusters per student. In contrast to animals, the majority of clusters were taxonomic (91%), including meat (39%), fruit (28%), sweets and veggies (11% each), pasta (3%), and seafood (2%). Breakfast was the only thematic category (6%). Age moderately and significantly correlated with the number of food clusters students without disabilities generated (r = 0.422, p = .001). This was the only significant finding. The number of items generated within clusters ranged from 2 to 4, with the exception of one student who generated only one cluster with 12 items, showing a depth of fruit knowledge. Students with disabilities mirrored the most frequent clusters of students without disabilities for a total of 12 clusters, including meat and fruit (42% each) and sweets (16%). Two students generated only one food cluster (fruits) that included 12 and 9 items; the latter was a student with a disability who appeared unable to switch to another category and instead seemingly unknowingly repeated items within the fruits cluster. In contrast, another student generated 21 items but only one 2-item cluster. Fingerspelled foods were slightly more frequent than fingerspelled animals, representing 5% of responses (40 out of 744 responses), and fingerspelling was utilized by almost twice as many students for foods than for animals (31%; aged 11;11–21;8). Fingerspelled items were diverse, such as ROAST, KIWI, NOODLES, TACOS, ENCHILADAS, FUNNEL CAKE, PORK CHOPS, MAC, AND CHEESE. Similar to animal responses, students with disabilities produced no fingerspelled items. Finally, student performance between animals and foods significantly and strongly correlated for those without disabilities (aged 5;3–21;8) (r = 0.665, p < .001) and was significantly different (animals: M = 13.6, SD = 5.1; foods: M = 11.7, SD = 4.8) (t (64) = 3.86, p < .001). We divided U.S. students without disabilities into a younger subgroup (7;8–15;3 years) as a means of comparison with Marshall et al.’s (2013) results (6;4–15;3) and an older subgroup (13–21 years) as a means of comparison with the group of Puerto Rican students. Twelve students overlapped in these subgroups. Younger students’ correlation between categories was stronger (r = 0.800, p < .001) but their difference in performance by task was not significant (animals: M = 11.7, SD = 4.8; foods: M = 10.6, SD = 5.3) (t (25) = 1.73, p = .096). Older students’ correlation between categories was moderate (r = 0.423, p = 004) and their difference in performance was significant (animals: M = 13.1, SD = 4.0; foods: M = 15.3, SD = 4.2) (t (45) = −3.30, p = .002).
The Puerto Rican students (aged 13;2–19;7) generated 8–18 foods. There were no significant correlations or differences. Seventeen different foods appeared in the first three responses. The most frequent first item was rice (47%), followed by different fruits (24%) and bread (12%). The most frequent second item was beans (41%), again followed by fruits (24%). Various fruits (35%) and meat (18%) were the most popular third items, followed by varied responses (e.g., rice, hotdog, salad, and spaghetti). Additionally, 11 students generated rice and beans in sequence while only 1 student mentioned pizza. Students generated a range of 1–4 clusters and a range of 2.3–7.0 items per cluster. Similar to U.S. students, this was about half of the total number of clusters they generated for animals. Again, there was no significant difference in the number of clusters produced between same-aged U.S. (N = 45, M = 2.01, SD = 1.07) and Puerto Rican students (N = 17, M = 1.94, SD = 1.03) (F[1, 60] = 0.172, p = .680). The majority of food clusters was taxonomic, including meat (29%), fruit (23%), salad (19%), sweets (13%), seafood (13%), and vegetables (3%). Only three fingerspelled responses occurred (1.3%), including BBQ by one student and both CAMARONES (shrimp) and LANGOSTA (lobster) by another student. Finally, Puerto Rican students also exhibited a significant correlation between animal and food scores (r = 0.535, p = .027) and a significant difference in their performance between categories (animals: M = 17.1, SD = 2.8; foods: M = 13.6, SD = 4.4) (t (17) = −3.82, p = .002).
51U
Forty-eight U.S. students aged 11;11–21;8 completed the 51U task. No students with disabilities completed this task. Overall, students generated a combined 10–39 items for the 51U task. Student response totals decreased by subsequent handshape. As a means of comparison with Puerto Rican students, performance by handshape for U.S. students 13–21 years of age was as follows: 5: M = 10.2, SD = 4.0, range = 3–19; 1: M = 9.5, SD = 3.8, range = 2–16; and U: M = 5.2, SD = 2.7, range = 1–12. Besides the youngest students, performance was similar at about 25 total items. Finally, one of the older student with deaf parents (11;11) outperformed the other by 50% or more for each handshape on the 51U task: 16 v. 7 items on the 5 task; 12 v. 6 on the 1 task; and 5 v. 1 item on the U task, showing variation across the two students with deaf parents. Overall, Puerto Rican students generated a combined 17–42 items for the 51U task. Performance by handshape was as follows: 5: M = 12.5, SD = 3.8, range = 6–18; 1: M = 10.4, SD = 3.0, range = 7–18; and U: M = 6.6, SD = 2.6, range = 2–11. Again, response totals decreased by subsequent handshape. In this sample, the two students with deaf parents (aged 14;11 and 17;8) had the lowest scores (19 and 20). No correlations or differences were significant within either the U.S. or Puerto Rican samples. However, there was a significant difference for geographical location among same-aged peers (i.e., 13–21 years): Puerto Rican students (N = 17, M = 29.5, SD = 7.3) scored significantly higher than U.S. students on the 51U task (N = 44, M = 24.9, SD = 8.1) (F[1, 59] = 4.10, p = .047). Finally, we investigated correlations among semantic and phonological scores for both student groups. For U.S. students, both semantic tasks significantly correlated with the phonological task (animals and 51U: r = 0.334, p = .025; food and 51U: r = 0.553, p < .001). However, neither semantic task correlated with the phonological task for Puerto Rican students (animals and 51U: r = 0.117, p = .654; food and 51U: r = 0.190, p = .464).
Discussion
The purpose of this study was to document the semantic and phonological sign generation performance of deaf adults and students who used ASL. We investigated the effects of various factors related to performance within these tasks.
Semantic Tasks
Deaf adults in this study generated a mean of 20 animals, similar to deaf university students (Morere et al., 2012) and a few items lower than the predicted mean for hearing adults (Mitrushina et al., 2005) and deaf adults who used BSL (Marshall et al., 2014). They produced a mean of about one cluster lower than Marshall et al.’s (2014) adults, with a similar mean cluster size. Marshall et al. (2014) noted “no obvious differences in number or types of responses” (p. 592) for their sample composed mostly of native or early signers. In contrast, the present sample of adults all had hearing parents and six acquired ASL after the age of 4 years, yet their animal performance was similar compared to Marshall et al.’s adults. Additionally, no effects were found for chronological age or early or late acquisition of ASL for either semantic task for the present deaf adults. These findings are similar to those of Choubsaz & Gheitury (2016), who reported no differences for age of acquisition on adults’ signed generation of items within given categories.
The older deaf students, both from the USA and Puerto Rico, generated fewer animals than university students (Morere et al., 2012) and present adults with a similar number of animal clusters and cluster size. They generated 8–10 fewer food items than university students (Morere et al., 2012) and deaf adults who used BSL (Marshall et al., 2014). They generated fewer food clusters than deaf adults but generated similar cluster sizes (Marshall et al., 2014). About 40% of the university students had at least one deaf parent (Morere et al., 2012). The younger students (i.e., 7–15 years of age) generated fewer animals, fewer foods, slightly fewer clusters, and comparable cluster sizes across both tasks compared to Marshall et al.’s (2013) sample of same-aged BSL signers, five of whom were native signers. One might speculate that university students had significantly more opportunities for ASL interactions at Gallaudet University, similar to the deaf adults in this study, and in contrast to younger students at schools for the deaf. As noted by Marshall et al. (2013), it is possible that semantic fluency generation is more demanding in a signed language, perhaps due to deaf children having smaller vocabularies. However, it may be that any possible effects of smaller vocabularies or later acquisition of ASL disappear with increased language exposure as children transition into adulthood. Finally, a weak but significant correlation between semantic performance on both tasks and years at the school site, which was a proxy for years of ASL exposure, was found for U.S. students, in contrast to Marshall et al.’s (2013) finding of no effects for years of exposure based on parent or teacher report. Similar variation (i.e., standard deviation) in semantic results across students and adults in this study and other studies (Marshall et al., 2013, 2014; Morere et al., 2012), regardless of age of acquisition of and amount of exposure to sign language, supports Choubsaz & Gheitury's (2016) speculation that age of acquisition may affect syntactic knowledge more so than semantic knowledge. Additionally, performance may have been affected by a combination of reading and instructional influences, as suggested by Morere et al. (2012). These differences in semantic findings compared to previous studies may result from the greater prevalence of participants who had hearing parents, attended schools for the deaf for a variety of total years, had varied exposure to sign language, and who demonstrated geographical variation. Accurate data on age of ASL acquisition and the amount of ASL exposure outside school for larger samples would clarify differences found in this study.
No difference was found between the animal and food tasks for younger students, similar to Marshall et al. (2013) and results for hearing children 6–12 years of age (Halperin et al., 1989). However, adult and older student performance between the tasks in the present sample significantly differed, in contrast to previous findings for deaf adults (Marshall et al., 2014) and university students (Morere et al., 2012). Older students from both sites in this study generated about half as many clusters for food as animals. Clusters among U.S. and Puerto Rican students were similar. Zoo, a large thematic category, was the most frequent animal cluster generated by all participants, with nearly identical generation of the top three animals across groups (i.e., dog, cat, tiger, elephant, lion), similar to the children in Marshall et al. (2013) and results for spoken English samples (Crowe & Prescott, 2003; Nelson, 1974). These results support findings that more frequently switches among clusters, as opposed to larger cluster sizes, result in higher scores (Kavé, Kigel, & Kochva, 2008; Marshall et al., 2014). However, U.S. students, and adults to a lesser degree, identified farm animals with greater frequency than Puerto Rican students. Also, compared to U.S. students, Puerto Rico students used different signs for a number of animals, including pig, lizard, horse, and dog.
Within the food task, the top three generated fruits (i.e., apple, orange, and banana) by deaf adults and children in this study directly align with previous findings of children's most frequently generated fruits (Nelson, 1974). However, U.S. and Puerto Rican students most frequently generated dissimilar foods, including pizza, hamburger, ice cream, and French fries for the former and rice, beans, and fruit for the latter. Present food responses also differed from those of British children in Marshall et al. (2013). For example, “chips,” or the equivalent of American French fries, was the most frequently generated food by British children, while hamburger was the 10th and pizza the 13th. In contrast, hamburger and pizza were the two most frequently generated foods for the present U.S. students. Ice cream, which tied for third most frequent food, was not listed in British children's top 14 foods, even though it is cited as one of the foods acquired earliest in BSL (Vinson, Cormier, Denmark, Schembri, & Vigliocco, 2008). These differences in student responses for animals and foods may reflect cultural diversity in each geographical location, including the larger fast food market in the USA.
The small sample of students with disabilities mirrored the same cluster generation as their peers, although with fewer clusters and fewer items within clusters, and they used no fingerspelling for either task. Their semantic knowledge, as measured by these tasks, appears shallower than their peers and they made more sign retrieval errors, such as non-animal signs during the animal generation task, similar to Marshall et al.’s (2013) findings for those with SLI. These results add to the currently scant literature related to deaf students with various disabilities and provide a starting point for discussions of language development and language assessment for this sub-population (Beal-Alvarez, 2016; Mann, Roy, & Marshall, 2013; Marshall et al., 2013). These results must be interpreted with caution, given the small convenience sample and various effects of different disabilities on student performance. Larger samples of data will provide more insight into students with disabilities’ performance.
Deaf adults exhibited a much higher occurrence of fingerspelling within the fruit task compared to results for the food task for other deaf adults (Marshall et al., 2014) and present students, which may be reflective of more specialized knowledge (Courtin, 1997). One deaf adult explicitly noted that many fruits have fingerspelled names. Similarly, U.S. students used more fingerspelling in their responses than Puerto Rican students, and paired some signs with fingerspelling for clarification. In lieu of fingerspelled words, Puerto Rican students tended to elaborate in their descriptions, such as BIRD paired with a closed-beak classifier to demonstrate woodpecker. Less use of fingerspelling by Puerto Rican students was anecdotally noted by some of their teachers, independent of this task. They did use fingerspelling in other ways outside these tasks. When asked by the researcher what a sign meant, one student fingerspelled LAGARTO (LIZARD) in Spanish then looked to his English teacher for its spelling in English; another student fingerspelled HORMIGAS, looked up its English translation using an app on her phone, and fingerspelled ANTS. It is not clear if reduced fingerspelling within these semantic tasks is a cultural preference in how sign language is used or a lack of the label in their L2 for specific items (i.e., Spanish or English). The overall low incidence of fingerspelling in these semantic tasks supports Marshall et al.’s (2013) findings.
Phonological Task
Mean overall 51U scores were similar across U.S. students (13–21 years), with a mean of about 25 total items, while the smaller group of Puerto Rican students consistently increased in score with age and generated about 5 more total items. As a group, U.S. students varied by about one item more than Puerto Rican students, had lower minimum scores, and showed a larger difference in range. In comparison, deaf university students (M = 25 years) also showed a large range in performance (42 items), with a mean 10 items higher than U.S. students and about 5 items higher than Puerto Rican students. Again, perhaps increased ASL interactions increased university students’ ASL awareness (Morere et al., 2012). Looking at specific handshapes, Puerto Rican students had higher minimum items than U.S. students for the 5 and 1 handshapes, while performance was similar for the rarer U handshape. Similar to previous results, students decreased in their phonological responses with subsequent handshapes (Marshall et al., 2014; Morere et al., 2012) and no gender differences were found (Goodman, 2015). Neither chronological age nor years at the school site affected performance for either sample. It is unclear why the smaller sample of Puerto Rican students outperformed the U.S. students. This may be attributable to the greater variation within the larger U.S. sample, differences in instruction, or other factors. Of the four students with deaf parents who completed this phonological task, only one outperformed his peers with hearing parents. This lack of difference contrasts with receptive results for deaf children with hearing and deaf parents (Beal-Alvarez, 2014, 2016; Henner, 2016; Novogrodsky, Fish, & Hoffmeister, 2014) and supports the variability of language skills, even when children have deaf parents.
Generation Strategies
Overall, adults and students used similar recall and generation strategies for the semantic tasks. Some participants used the strategy of consciously repeating previously named items to stimulate generation of additional items. Several deaf adults and students were aware when they re-iterated items and made comments such as “I already said that” or “(item) FINISH”. Four deaf adults mentioned visualization of a supermarket when generating fruit. Other adults acknowledged how their background experience affected their performance. For example, one adult generated 27 animals, yet generated only 11 fruits and said “I don't eat many fruits” as he stalled in item generation. Similarly, another deaf adult said “I'm not an animal person.” Others stated that they were done prior to the end of the one-minute period, demonstrating the latency effect documented by Marshall et al. (2013, 2014). Adults’ and students’ background knowledge affected some of the unique items they generated, including boysenberry, bullfrog, killer whale, chameleon, lionfish, and wild hog.
During the phonological task, some students and adults looked at their hand in the designated handshape and paired it with different movements and locations to generate signs, which usually resulted in additional item generation, such as PAPER, MOVIE, and CHEESE (differ by movement) and SUMMER, UGLY, and DRY (differ by location). Two out of three parameters in these series of signs are the same; this suggests signers store similarly produced signs in proximity, demonstrated by their sequence in rapid recall (Marshall et al., 2013, 2014). A few students were inhibited in item generation by their recent instructional experience with handshape stories, namely for 5 and 1 handshapes, in which they rendered stories told with only one handshape that did not necessarily portray signs that were correct in the context of the handshape generation task.
Similar to previous results (Koren, Kofman, & Berger, 2005; Marshall et al., 2014; McQuarrie & Abbott, 2013; Morere et al., 2012; Sauzéon et al, 2004), phonological generation tasks appear more difficult than semantic tasks for the present students with fewer responses and more errors than semantic generation tasks. Marshall et al. (2014) surmised this difficulty may be due to the “more simultaneous and composite nature of handshape, location, and movement [that] makes them more difficult to explicitly extract from the sign in comparison to the onset of spoken words” (p. 604).
Need for Explicit Instruction
Finally, in the current realm of evidence-based instruction, we return to the use of students’ language data to guide instructional decisions for their individual needs. Ormel et al. (2010) concluded that deaf students made limited gains in semantic performance across grade levels and stressed the importance of “high-quality basic (sign) vocabulary knowledge, which is needed in order to develop semantic knowledge” (p. 357). Given the present results, it appears direct and explicit instruction in organization and retrieval of semantic and phonological knowledge is a requirement for some students compared to their same-aged peers, as shown by variation in overall animal, food, and handshape knowledge and breadth and depth of semantically related clusters. Additionally, explicit ASL phonology instruction may be needed for some students to fully understand their first language, how to manipulate it to change meaning (e.g., changing one parameter of a sign, such as location, changes meaning from MOM to DAD) and how to apply this knowledge to a second language, such as printed English or Spanish during reading (e.g., changing the initial phoneme in a word, such as “cat” to “fat” changes meaning). Some teachers embed phonological instruction within bilingual ASL–English educational programs (Crume, 2013), but limited evidence is available regarding how effective educators embed ASL instruction across the K-12 setting. For example, direct instruction increased students’ ASL classifier use within narrative productions (Beal-Alvarez & Easterbrooks, 2013) and their acquisition of key vocabulary in both sign language and text (Cannon, Fredrick, & Easterbrooks, 2010; Guardino, Cannon, & Eberst, 2014). Additional intervention studies are needed to investigate the effects of direct ASL instruction on students’ knowledge and use of ASL and how this knowledge relates to their knowledge of printed language for literacy (e.g., English or Spanish). Incorporating the strategies used by deaf adults when rendering printed English text into ASL for comprehension might be a starting point for bilingual literacy instruction (Banner & Wang, 2011).
Limitations
One limitation of this study is the lack of 51U data from the deaf adult sample, which prevents comparison of student performance to target performance of adults in their community. Another limitation is the small convenience samples of adults and students. Although adults varied in their education history they were from one geographical region. The inclusion of deaf adults from Puerto Rico and other geographical regions would provide more evidence related to target semantic and phonological performance and perhaps related to variation in performance. Moderate to strong age-related effects were found for the larger sample of U.S. students on overall performance and cluster performance for both semantic tasks, and for the smaller sample of students with disabilities on the food task, similar to Marshall et al. (2013). However, these correlations are restricted in utility based on small age subgroups in each student sample, which ranged from 5 to 8 for U.S. students aged 7–14 years and 2–3 students for all of the Puerto Rico groups except those 17–19 years. The lack of correlation between age and semantic performance for the smaller Puerto Rican sample and the lack of correlations between both age and years at the school site and phonological performance may be a result of these limited subgroup sizes. Educators might be cautious when comparing student scores for these semantic and phonological tasks by age. Larger samples from multiple geographic areas that include sufficient numbers of students across each age-based subgroup would provide more reliable and perhaps generalizable conclusions relative to semantic and phonological sign generation across ages.
Conclusion
In summary, deaf students’ performance on semantic and phonological sign generation tasks is similar in some ways and varied in others. While the present results present an initial snapshot of student performance across these semantic and phonological sign generation tasks, future research with larger samples from various geographical regions and instructional settings, such as local school systems in addition to schools for the deaf, would clarify current findings. Educators should use ASL data from individual students and deaf adults paired with evidence-based instructional strategies to tailor their instruction to students’ unique language and literacy needs.
Conflicts of Interest
The authors have no conflicts of interest to report.