Methodological issues in studying variation1

Ceil Lucas

1. Introduction

This paper will discuss methods for the sociolinguistic studies in sign language communities and will touch on four main topics: data collection (Section 2), defining variables and constraints (Section 3), data reduction (Section 4), and dissemination of the findings (Section 5).

2. Collecting data

2.1. Selection of participants

A major component of data collection, of course, is the selection of the participants. Sociolinguistic studies want to be able to determine the correla- tion between variation and speaker – in this case, signer – characteristics, including age, gender, ethnicity, region and socioeconomic status. Although some characteristics like gender, age and ethnicity might be common to all studies of linguistic variation, many of these characteristics need to be artic- ulated more fully when they are put into research practice in a particular community. This is particularly true for studies of linguistic variation in Deaf2 communities. Notions like socioeconomic status or even age cannot be simply wholly borrowed from studies of variation in spoken language communities. The differences in social characteristics when applied to Deaf communities are of two types. The first type includes characteristics, like age and region, which may have a different meaning when the history of Deaf communities is taken into account. The second type includes characteristics, like language background, that are unique to Deaf communities. For deaf people, regional background or where they were born, may be less important than where they attended school (especially if it was a resi- dential school) or where their language models acquired American Sign 286 Ceil Lucas

Language (ASL). Age as a characteristic may have different effects on linguistic variation because of the differences in language policies in schools and programs for deaf children. Some differences in language use may be the result of changes in educational policies, like the shift from oralism3 to Total Communication or from Total Communication to a bilingual-bicultural approach. These language policies have affected not only what language is used in the classroom but also teacher hiring practices that support hiring deaf teachers who know the sign language in question or hearing teachers who can’t sign. These language policies have affected deaf children’s access to appropriate language models and this access may have varied across time to such an extent that it has affected the kind of variation that we see in sign languages today. One strong example of this concerns the Black Deaf community in the United States. Following the Civil War, seventeen states and Kendall School in the District of Columbia established separate schools for Black deaf children or opened “departments” on the campus of the school for white children, i.e., separate buildings. Even though deaf education started propi- tiously in the United States in 1817 at the American School for the Deaf in Hartford, Connecticut with (ASL) as the medium of instruction, by 1880, oralism was firmly established in the schools for white deaf children, with many Deaf teachers being fired. However, the policy of oralism was not extended evenly to the schools for Black deaf children and the use of sign language as the medium of instruction was widely allowed. In addition, some schools for Black deaf children had white Deaf ASL-signing teachers, providing the children with ASL input. Then, following the Brown vs. Board of Education decision in 1954, Black and white deaf children slowly began to attend school together (but some states like Louisiana managed to delay integration until 1978!) and the practice of mainstreaming began to take over education at residential schools, so deaf children had increasingly more contact with their hearing peers. All of this helps explain the variation in Black ASL that McCaskill et al. (2011) have found, e.g. noticeably less mouthing in older signers, since they had less direct exposure to oralism and hearing peers who spoke English. This illustrates clearly that the selection of participants for sociolinguistic studies of sign languages must take into account the meaning of age, ethnicity and region in Deaf communities, in order for the resulting analyses to be meaningful. Furthermore, large studies of sociolinguistic variation in ASL (Lucas, Bayley, and Valli 2001) and other sign languages such as Auslan (Schembri et al. 2009) have clearly shown the importance of whether a subject comes from a Deaf family in which the sign language is used or from Methodological issues in studying sign language variation 287 a non-signing family, be it hearing or deaf. For example, Lucas, Bayley, and Valli (2001) demonstrated that participants from Deaf families were more likely to use the standard “citation” forms of signs, such as signs like KNOW produced at the forehead as opposed to lower locations.

2.2. Contact people

Central to the selection of participants are contact people. The approach to participants in Lucas, Bayley, and Valli (2001) and McCaskill et al. (2011), for example, was guided by the work of Labov (1972a, 1972b, 1984) and Milroy (1987). Groups were assembled in each area by a contact person, a Deaf individual living in the area with a good knowledge of the commu- nity. These contact people were similar to the “brokers” described by Milroy, individuals who “have contacts with large numbers of individuals” in the community (1987: 70). They are compensated and participate of their own free will. It was the responsibility of the contact people to identify persons suitable for the study – in the case of the 2001 and 2011 studies, fluent life- long users of ASL who had lived in the community for at least ten year. Community members can be decidedly reluctant about participating in a study and can outright refuse. This is not at all unique to Deaf communities. As Wolfram (2012) explains, “community members may have underlying questions and concerns about sociolinguists’ motivations in working in their community. What are they really doing in their community? Why are they so obsessed with the minutia of language? Do they have an underlying socio- political agenda in terms of language?” He goes on to say that “We need to enter the community fully understanding and appreciating the legitimacy of the community’s practical cautions and concerns about the motives of socio- linguistic researchers”. Feagin (2002: 26) observes that “skin color, class affiliation, speech, or education may all set the investigator apart”. There are particular concerns in Deaf communities, directly tied to the history of deaf education and to how research on sign languages has taken place. Oralism, the belief that spoken language is inherently superior to sign language, played an important role in deaf education. Even though deaf education in the United States began in 1817 with sign language as the medium of instruction, by 1880, the oral method of instruction was well established in the White schools (Lane, Hoffmeister, and Bahan 1996) and also in many countries around the world. As Burch and Joyner (2007: 21) note, “the rise of oralism motivated schools across the country to replace deaf teachers with hearing instructors who would speak to students rather 288 Ceil Lucas than sign with them”. In the mid-1970s, in light of low reading levels in deaf students, the transition was made in the United States to the simultaneous use of speaking and signing, on the theory that deaf students could thus see English being produced on the mouth and hands, to help them in their learning of English. Specific manual codes for English (MCEs) were devised such as Signing Exact English (Gustason, Pfetzing, and Zawolkow 1972), which purported to represent the syntax, morphology and lexicon of spoken English. As Ramsey (1989: 123) states, “The developers built the requisite MCE lexicon by borrowing ASL signs, modifying ASL signs with hand- shape features from the manual alphabet, and inventing signs specifically to represent English derivational and inflectional morphemes”. And she goes on to observe that “[t]he materials used to construct SEE 2 are highly valued linguistic resources in the Deaf community: ASL lexical items and the medium of signing itself. These resources are being used to promote the linguistic values of another community” (Ramsey 1989: 143), i.e., the teachers and parents and educational administrators who see MCEs as an answer for teaching deaf children. MCEs were devised in many countries around the world. At the same time that MCEs were being devised, research on the structure and use of sign languages was getting underway in many places, with many members of the Deaf community serving as informants and sign models for hearing researchers. It was not infrequent that this research was published with only a brief mention or no mention of these informants and models, naturally leading to resentment. Singleton, Jones, and Hanumantha (2012) conducted a focus group study with members of the Deaf community and researchers and report that two main issues emerge: lack of trust and confidentiality. The lack of trust has to do in part with feelings of tokenism on the part of Deaf researchers, “feelings of being exploited and that they had not received adequate credit for their contributions to the work”. Resentment can also arise concerning the ownership of the research findings and “Some resented the academic superiority of English over ASL in the publication world and the fact that published materials are predominantly in English”. Issues that have arisen in doing research in Deaf communities have also been explored by Baker-Shenk and Kyle (1990) and Harris, Holmes, and Mertens (2009). Harris, Holmes, and Mertens (2009: 114) state: “It is critical that researchers attempt to determine the ways in which Sign Language community members feel and think about the world and give these the recognition they deserve”. Based on Osborne and McPhee’s (2000) indigenous terms of reference, they propose six principles to be considered when sign language communities are studied, principles which center on openly involving and empowering the Methodological issues in studying sign language variation 289 members of the community. The fact that community members have most often not been involved and empowered has led to caution and often, reluc- tance by community members to cooperate with researchers, a reluctance which the contact people have to mediate.

2.3. The Observer’s Paradox

Another factor that shapes data collection is the sensitivity of signers to the audiological status and ethnicity of the audience. The amount of attention language users pay to their language production has been addressed by soci- olinguists starting with Labov (1972a), who discussed what he referred to as “Observer's Paradox”. That is, our interest is in the language signers and speakers use when they are not being observed. However, to obtain the kind of data that we need for analysis we need to record signers' (or speakers') production in situations that often lead to self-consciousness. The problem is particularly acute in sign linguistics because, on video, there is no way that signers can be completely anonymous. In addition, as scholars who have discussed the influence of audience design on language use have shown (see e.g. Bell 1984, 2001; Giles 1973, 2001; Giles and Powesland 1975), speakers and signers often adjust their language use to accommodate to what they perceive as the preference of their interlocutors. In sign linguistics, Lucas and Valli (1992) demonstrated that ASL users are very sensitive to an interviewer’s audiological status and ethnicity, i.e., hearing or deaf, Black or White. This sensitivity may be manifested by rapid switching from ASL to Signed English (a manual code for English) or Contact Signing (an outcome of the contact between ASL and English char- acterized by core features from both languages and continuous mouthing). As explained by Giles’ Accommodation Theory (1973), many Deaf people will adjust their signing to bring it closer to what they perceive to be the preference of their interlocutor. Issues of the lack of anonymity need to be clearly and carefully handled during the consent process, so that participants consent (or don’t) to having their images shown as part of conference presentations or in publications. As Singleton, Jones, and Hanumantha (2012) state, there may be a “need to translate informed consent documents into the native language of the host community (i.e., videos using American Sign Language) to ensure that Deaf participants, who may have limited English proficiency, are offered acces- sible information regarding their rights as research participants”. Since inter- views often take place at sites to which the researcher cannot easily return to 290 Ceil Lucas gain continuing consent, the consent form is worded to clearly explain that the participants’ images will be part of conference presentations and publica- tions from the time of consent forward. Potential participants are of course entirely free to decide against participating in the project. To avoid code- switching caused by the interviewer, researchers need to be very mindful of the audiological status and ethnicity of the interviewer. One instance from the Black ASL project illustrates this quite starkly: the Black ASL team inter- viewed an elderly signer. The team consisted of four Black Deaf researchers and one White hearing researcher. At the end of the interview, the elderly signer wanted to know who everyone was. She knew that the interviewer was Black and deaf and confirmed, one by one, that the other three Black team members were also deaf. The hearing White researcher – who had stayed out of sight during the interview – then introduced herself and because she is hearing, the elderly signer who had been signing for ninety minutes with no voice turned on her voice and spoke English for the remaining thirty minutes that the team was in her house. As mentioned, the hearing White researcher had stayed out of sight during the interview. Had she been present, the chances are good that the data would be quite different. Of course, the actual tools used can include interviews, structured elici- tation, and questionnaires, as well as free conversation sessions. The first three must be designed with the issues discussed here in mind: Who is doing the interviewing and the elicitation, are questionnaires written in English entirely accessible to all deaf ASL users or is the questionnaire signed and does the researcher need to go over the questionnaire with the subject, and so forth (see also Van Herreweghe and Vermeerbergen 2012). Recent advances in communication technology have directly affected the collection of sign language data and the kind of data that can be collected. For example, researchers can now administer survey questions virtually and participants can respond to them virtually. That is, participants can see the researchers asking them the questions in sign language on a computer and can respond to the researchers in sign language on the same computer. Researchers no longer have to recruit participants to come to a physical loca- tion for data collection. This has naturally led to adjustments in the tradi- tional signed consent form and to considerations of the effects of providing virtual responses on the nature of the resulting data. In addition, there now exists a whole new source of publicly-available data, in the form of YouTube and vlogs (videologs), for example, also raising the issues of permission and the wide variety of uncontrolled data collection settings. Finally, the video relay service (VRS) is now being used widely for interpreting, whereby a deaf person and a sign language interpreter can actually see each other. Methodological issues in studying sign language variation 291

An interpreter may be interpreting for a deaf person from anywhere, not just in their immediate area, and issues of how to handle lexical variation have frequently arisen and are being researched (Palmer, Reynolds, and Minor 2012).

3. Defining variables and linguistic constraints

3.1. Variables

There are a number of phonological and morpho-syntactic features in sign languages that behave variably. We see phonological variation in the hand- shape, location, palm orientation and movement of a sign. Some signs which are two-handed may be produced with one hand. In terms of syntactic varia- tion, most sign languages are pro-drop languages and hence show variability in their use of subject pronouns. Variability is also seen in WH and negative elements, in the use of constructed dialogue and constructed action and in the size of the signing space. There is of course also widespread lexical variation. The challenge is to identify which can be identified and analyzed as sociolin- guistic variables. As seen with Hoopes (1998) and his study of pinky exten- sion, one part of the challenge can be confirming a variable. The pinky can easily be extended as a part of signing and the issue was to determine if this extension constituted a sociolinguistic variable. He focused on signs such as TOLERATE and WONDER and demonstrated that pinky extension tended to co-occur with prosodic features of emphatic stress and that it tended to occur with lexemes used repeatedly within a discourse topic, before pauses and with lexemes lengthened to almost twice their usual duration. Another example of confirming a variable can be found in the Black ASL project. During the process of writing the grant proposal to obtain funding for the project, the researchers viewed tapes of interviews done with Black signers and noticed that the signers used a lot of repetition, of lexical items, clauses or whole sentences. This use of repetition was added to the list of features that might distinguish Black ASL as a separate variety of ASL and analysis showed that it was indeed one such feature. Sometimes the challenge can be in identifying a variable. An example comes from studies of Tactile ASL (TASL), the signing used by deaf-blind people in the United States,4 specifically those with the genetic condition, Usher Syndrome type I. Individuals with this syndrome are born deaf and later, usually in their teen years, start losing vision in varying degrees due to retinitis pigmentosa. Crucially, most American deaf-blind people in 292 Ceil Lucas this category grow up using ASL and are fluent signers by the time they begin to lose their sight. A variety of ASL has emerged in this community that accommodates the loss of sight at all linguistic levels: phonological, morphological, syntactic, and discourse. One of the consequences of the loss of sight is that deaf-blind people no longer have access to the numerous ASL grammatical and discourse markers produced on a signer’s face. Remark- ably, these non-manual (facial) markers are produced on the hands in Tactile ASL. For example, the raised eyebrows required for yes/no questions or the nodding required for back-channeling are produced manually (see Collins and Petronio 1998 and Collins 2004 for fuller accounts). As mentioned, features of Tactile ASL are manifested at every level of the language and there is a vigorous community of deaf-blind signers who use Tactile ASL. Tactile ASL qualifies as a clear example of a variety of ASL but arriving at this conclusion required identifying the features that would make this so and clearly distinguishing them from the features of “visual” ASL. Confirming and identifying variables can happen with reference to earlier studies. For example, studies done in the 1970s had shown that two-handed signs could become one-handed (Woodward and DeSantis 1977), that signs could be lowered and moved to a more central signing space (Woodward, Erting, and Oliver 1976; Frishberg 1975) and that signs such as DEAF could experience metathesis (Liddell and Johnson 1989). Confirming and identifying can also happen through observation, as in the case of repetition described above. And it can happen via informed intuition. This was the case in the Black ASL project (McCaskill et al. 2011) with the incorporation of features of spoken African American English into ASL signing. It simply stood to reason that younger Black signers who had attended integrated educational programs would incorporate such features into their signing and in fact, they do. Expressions such as “Girl, please” and “My bad” find their way into ASL, as does slapping five in agreement during the course of a conversation.

3.2. The envelope of variation

A related issue has to do with defining the so-called “envelope of variation”, that is, deciding which instances of a particular phenomenon will be included or excluded from an analysis and why. A sociolinguistic variable may be Methodological issues in studying sign language variation 293 defined as two ways of saying or signing the same thing, with the distribution of different forms influenced by linguistic and social factors (Labov 1972b). For example, an ASL user may sign DEAF from ear to chin or from chin to ear. However, the meaning of the concept DEAF does not change. And, as Bayley, Lucas, and Rose (2000) have shown, signers' choices among the different forms of DEAF are affected by the grammatical class to which the sign belongs as well as by a number of social factors including the region where the signer lives and age. Turning to an example from variation between one-handed and two-handed versions of the same sign, it is clear that DEER signed with two hands and DEER signed with one hand refer to the same creature. Before we begin coding examples, then, we first need to decide which signs to include in the envelope of variation and which signs to exclude. We can consider the case of two-handed signs that can be signed one-handed. The task is complicated to a certain extent by the fact that ASL contains different types of two-handed signs. Signs of Type 1 can be produced with only one articulator. Examples include: DEER, DON'T KNOW, FINISH, HORSE, NOW, PONDER, SICK, TIRED, and WANT. Signs of Type 2 may be divided into two subtypes. Type 2a signs cannot usually be produced with only one hand. Examples include CAN'T, CHEESE, CHURCH, SOCKS, STAR, and WORK. Signs that can be designated as Type 2b can be produced with a substitute base or produced one-handed in very particular discourse situations. Examples include RIGHT, PAPER, SCHOOL, and SHOES. In an earlier study of handedness (Lucas et al. 2007) and in the Black ASL study (McCaskill et al. 2011), we coded only signs that can be produced with two hands or one hand unremarkably. Signs in Type 2a and 2b were excluded as being outside the “envelope of variation” or subject to constraints that have nothing to do with linguistic variation such as the presence or absence of a convenient substitute base. Similarly, in Lucas, Bayley, and Valli (2001), signs with a 1 handshape that are depicting verbs (also widely referred to as classifier predicates) were excluded, since the handshape in those signs is semantically motivated. Table 1 provides a summary of variables looked at in studies of ASL, Black ASL, Australian and New Zealand Sign Languages (Auslan, NZSL) and Italian Sign Language (LIS), along with the number of tokens of each variable. 294 Ceil Lucas

Table 1. Summary of variables in different sign languages

Language Variable Number of tokens American Sign Language 1 handshape 5195 (ASL) location 2862 DEAF 1618 pro-drop 429 2 hand → 1 hand 2258 Black ASL 2 hand → 1 hand 818 (+2258 earlier study) location 877 (+2862 earlier study) size of signing space 2247 clausal or phrasal 172 repetition mouthing 221 constructed dialogue/ 1021 constructed action Auslan/New Zealand Sign location 2667 Auslan, 2096 NZSL Language (NZSL) Italian Sign Language (LIS) WH elements 884 annotated, 413 coded

3.3. Linguistic constraints

Recall that internal constraints on variation are linguistic features that may play some role in the occurrence of variation. Wolfram (personal commu- nication, 1994) has stated that the internal constraints on variables may be compositional, sequential, functional, or having to do with structural incorpo- ration. Compositional constraints are those that have to do with the linguistic nature of the variable itself. Battison, Markowicz and Woodward (1975), for example, stated that thumb extension in such signs as FUNNY and CUTE had to do with how many other fingers were extended and whether there was internal movement in the sign. In the signs FUNNY and CUTE, the fingers oscillate. Sequential constraints are features in the immediate linguistic envi- ronment surrounding the variable, such as the handshape or palm orientation of the sign immediately preceding or following the variable sign. Functional Methodological issues in studying sign language variation 295 constraints relate to the role that the sign’s grammatical category plays in the variation, while the constraint of structural incorporation concerns the syntactic environment in which a variable finds itself. Finally, there may be pragmatic constraints on variation, as Hoopes (1998) found for pinky ­extension. Earlier studies of sign language phonology (Liddell and Johnson 1989) claimed that variation such as the lowering of signs such as KNOW or the metathesis observed in the sign DEAF had constraints of a sequential nature, as in spoken languages, such that DEAF followed by a sign on the forehead might be signed chin to ear while DEAF followed by a sign on the torso might be signed ear to chin. However, the results of Lucas, Bayley, and Valli (2001) and Hoopes (1998) demonstrate that the analysis of internal constraints on variation in sign languages needs to be done carefully because the identi- fication of such constraints is not always completely straightforward. For example, further examination of DEAF, signs like KNOW and 1-handshape signs has shown that the grammatical category of the target sign may have a significant role in explaining the variation. For all three variables – DEAF, signs like KNOW, 1-handshape signs – grammatical category is more impor- tant than or as important in explaining the variation as sequential features such as the preceding or following sign. For example, when DEAF is a predi- cate (“I am deaf”), it tends to be signed ear to chin. Adjectives (“deaf cat”) and nouns (“Deaf [people] know that”) use both ear to chin and chin to ear forms, while compound signs (DEAF^WORLD, DEAF^CULTURE) tend to favor the contact-cheek form of DEAF. The main question is why do grammatical and prosodic constraints seem to have a more important role than the features of the preceding and following signs in conditioning phonological variation in ASL? The first answer is simply that, as in spoken languages, phonological variation in ASL is not constrained exclusively by phonological factors. The focus heretofore may have been on features of the preceding and following signs, but large data-based quantitative studies such as the one undertaken by Lucas, Bayley, and Valli (2001) show that grammatical factors must also be considered. A second answer concerns differences between spoken and sign languages. Having established that sign languages are indeed “real” languages, research on all aspects of sign language structure has begun to show some funda- mental and most likely modality-related differences between spoken and sign languages. Of most relevance to the present discussion are the basic differences in how morphology functions and how the differences manifest themselves in variation. In many of the spoken languages in which phono- logical variation has been extensively explored, morphology is a “boundary 296 Ceil Lucas phenomenon”. That is, meaningful segments are added to the beginning or end of other units in the language in the form of plural markers, person and tense markers, derivational affixes, and so forth. These units are essentially added to an existing phonological environment. It stands to reason that when variation occurs, a good place to look for the cause of this variation is the immediate environment to which units have been added (i.e., the preceding and following segments). In fact, many studies of spoken language variation have demonstrated the key role of the immediate phonological environment in governing variation. However, morphology in sign languages is by and large not a boundary phenomenon, at least not to a great extent. There exist very few sequential affixes. Morphological distinctions are accomplished by altering one or more features in the articulatory bundle that makes up a segment or by altering the movement path of the sign. For example, segments are not usually added to other segments to provide information about person or aspect. Rather, the location feature of a segment (e.g. near or away from the signer) indicates person, and movement between locations indicates the subject and object of the verb in question. Similarly, a particular movement path indicates contin- uative or inceptive aspect. As Emmorey (1999: 173) states with specific regard to aspect marking in ASL: “In many spoken languages, morphologically complex words are formed by adding prefixes or suffixes to a word stem. In ASL and other signedlan- guages, complex forms are most often created by nesting a sign stem within dynamic movement contours and planes in space…, ASL has many verbal inflections that convey temporal information about the action denoted by the verb, for example, whether the action was habitual, iterative, continual. Gen- erally, these distinctions are marked by different movement patterns overlaid onto a sign stem. This type of morphological encoding contrasts with the primarily linear affixation found in spoken languages. For spoken languag- es, simultaneous affixation processes such as template morphology (e.g. in Semitic languages), infixation, or reduplication are relatively rare. Signed languages, by contrast, prefer nonconcatenative processes such as redupli- cation; and prefixation and suffixation are rare. Sign languages’ preference for simultaneously producing affixes and stems may have its origins in the visual-manual modality.” The results in Lucas, Bayley, and Valli (2001) indicate that these fundamental differences manifest themselves in the variable components of the language. That is, the immediate phonological environment turns out not to play the major role in governing phonological variables, in part because the variables themselves are not affixes. The grammatical category to which the variable Methodological issues in studying sign language variation 297 in question belongs, is consistently the first-order linguistic constraint (in this regard, see also Brentari 2002). This finding has important implications for our understanding of- vari ation in spoken and sign languages. As the modality differences between spoken and sign languages manifest themselves in the basic phonological, morphological, and syntactic components of the language, so they also seem to appear in the patterns of linguistic variation. As the phonological and morphological processes go, so apparently goes variation.

4. Data reduction

Once the data has been collected, decisions need to be made about how and to what extent it will be reduced. The central issue is how to record in a useful way manual-visual languages that do not have written forms.5 As seen in some of the examples in this paper, one very common approach is to “gloss” signs using the written language with which the sign language is in contact so written English for ASL and Auslan, written Italian for Italian Sign Language (LIS) and so forth. So a gloss for the English word “table” would be TABLE; a gloss for the Italian word “casa” (“house”) would be CASA. Glosses are not limited to single lexical items; sentences and whole texts can be glossed. Glosses can be very basic, with just the signs indicated, or they can be very complex, with indications of non-manual signals, indexing, repetition and other features. Issues arise, of course, concerning the general- izability of glosses, that is, can the glosses developed by one researcher be understood and of use to another researcher ? Given lexical variation, which sign for the gloss WATCH (verb) will a signer envision – a V handshape with the palm down or a C handshape at the mouth ? Seeing the gloss for WALK, will a signer envision a V handshape iconically like “legs” or two B hand- shapes alternating in movement, palm down? And how much of this detail should be included with the gloss? Sign language researchers naturally want to refer to the actual signing event and this is now possible with software such as the EUDICO Linguistic Annotator (ELAN) annotation system developed at the Max Planck Insti- tute. This allows for the video clip being glossed to appear on the screen and it also allows for numerous levels – “tiers” – of information to be coded. One tier can be a basic gloss and other tiers can have non-manual informa- tion, separate coding for each hand, occurrences of gestures, instances of constructed action and constructed dialogue and so forth. 298 Ceil Lucas

Other analyses require detailed phonetic notation. The first such system was introduced in 1965 by William Stokoe in his Dictionary of American Sign Language, in which he presented signs according to their handshape (tabula, “tab”), location (designator, “dez”) and movement (signation, “sig”). This was the first sign dictionary to present signs not by the alphabetical order of the English word to which they correspond but by their handshapes, locations and movements. For example, the first sign in the dictionary is the sign for “boss” or “person in charge”, produced with an A handshape. One significant shortcoming of the Stokoe system was that he did not recognize sequentiality in the production of the parts of signs, sequentiality of handshapes, locations, movements, palm orientations and non-manual signals that is easily demon- strable. Liddell and Johnson (1989; Johnson and Liddell 2011) introduced their Movement-Hold model which represented signs as sequences of move- ment segments and hold segments, parallel to sequences of consonants and vowels in spoken languages, with the features of handshape, location, palm orientation and non-manual signals as an articulatory bundle contained in each segment. Their system has evolved over the years and of course is used for very close analysis of sign phonetics and phonology. Finally, there are some situations in which a translation from a sign language to a written language might be needed. This would be in cases where a broad comparison of meaning might be required and would be of use in research on sign interpretation or in representations of the data in which the point is to convey a general meaning. For example, the interviews in the Black ASL project (McCaskill et al. 2011) revealed some striking percep- tions about the differences between Black signing and White signing. Partic- ipants were simply asked if they noticed differences between Black signing and White signing and as part of their responses, a number of them expressed the view that “White signing is better than Black signing” because “White signing is more advanced, more complex.” White signing is of course not more advanced or complex than Black signing – and in fact, one of the major findings of the project is that because of social and geographic isolation, Black signing uses more traditional, standard forms – and these perceptions are clearly a reflection of the oppression and inferior education that Black deaf students had to endure. And the most powerful way to present this infor- mation particularly to hearing, non-signing audiences is with the stark trans- lation of the signing, i.e., “White signing is better than Black signing.” Methodological issues in studying sign language variation 299

5. Dissemination of the findings

We finally come to the issue of disseminating the research findings. There are, of course, the usual avenues in scholarly journals and books. For confer- ence presentations and workshops, a distinction must be made between dissemination to hearing, non-signing audiences and signing audiences that may include both deaf and hearing people.

5.1. Hearing, non-signing audiences

One issue has to do with the relevance of sign language studies for spoken language linguists. In the fifty years since research on sign languages has been undertaken, research findings have informed spoken language studies in significant ways in the areas of gesture (Taub, Galvan, and Piñar 2008), the structure and use of metaphor (Taub 2001), bilingualism and language contact (Lucas and Valli 1992), first and second language acquisition (Chen Pichler 2010, 2012; Petitto 2000), variation and the emergence of varieties of language (Lucas, Bayley, and Valli 2001; McCaskill et al. 2011). Making this relevance and the sign language-spoken language connections clear is a chal- lenge faced by sign language researchers. This is accomplished by a steadily increasing presence of sign language researchers at scholarly conferences and in scholarly journals that traditionally focus on spoken languages and by increasing collaboration of spoken language and sign language researchers working on the same topic with the shared goal of understanding how human language works, be it spoken or signed. However, since presentations about sign languages usually involve either live demonstrations of signs or filmed examples and also deaf presenters with interpreters, it is quite common for hearing non-signers to conclude that, since a presentation is about sign language, “I can skip this one, because I don’t sign.” The challenge for sign language researchers is to think about how captions, voice-over, slow motion and careful explanation can be used effectively to make the information fully accessible and of interest to a hearing non-signing audience. Researchers also need to think about the starting point: How much does the audience already know about sign languages and research on sign languages and how much of a foundation needs to be provided. 300 Ceil Lucas

5.2. Signing audiences

A signing audience is one whose members may be both deaf and hearing, as most hearing researchers who work on sign languages sign and can follow lectures in sign language. While some deaf members of a signing audience may be trained linguists, many others may be community members from many different walks of life who attend a conference because they are inter- ested in what is being said about their language. As mentioned earlier, sign languages are highly valued cultural resources and tension can arise when the scientific terms and academic style being used to describe themare incomprehensible to community members, “What are you saying about my language?!” Researchers are challenged to handle this tension. This relates directly to Wolfram’s (2012) observation that the right nontechnical level for presenting the findings needs to be found. He goes on to say that “there is often a tension between the specialized expertise of the linguist and the community perspective on language.” It may happen that the findings being presented are emotionally diffi- cult for an audience to receive. For example, for reasons clearly having to do with oppression and segregation, a number of participants in the Black ASL study, when asked if there are differences between Black signing and White signing, immediately responded that “White signing is better, more advanced.” One might expect this response from older signers but it was the response of younger signers as well. Many in the audience at the 2009 meeting of the National Black Deaf Advocates (NBDA), where these find- ings were presented, were shocked and saddened by seeing signers giving these responses, surprised that the younger signers still had these percep- tions. And it was the challenge of the researchers presenting these findings to handle these reactions very carefully. Related challenges have to do with whether hearing presenters choose to speak or sign. They may choose to speak if their signing skills are not good enough for a formal presentation; they may choose to sign to clearly establish their connection to the commu- nity. If they choose to speak, as Singleton, Jones, and Hanumantha (2012) point out, this may also “raise the possibility of misanalysis of linguistic data and [ question] the researcher’s ability to run the study effectively and objec- tively when they possess limited sign proficiency”.

5.3. Giving back to the community

One final issue related to dissemination has to do with giving back to the community. Wolfram (2012) describes Labov’s Principle of Error Correction Methodological issues in studying sign language variation 301

(1982: 172) and Principle of Debt Incurred (1982: 173), along with his own Principle of Linguistic Gratuity (1993: 227), stating that Labov’s two prin- ciples “focus on the obligation of linguists to expose misunderstandings and misinterpretations about language” (Wolfram, forthcoming) and to use the knowledge for the benefit of the community being studied, while Wolfram urges researchers to “pursue positive ways in which they can return linguistic favors to the community” (Wolfram 1993: 227). Giving back to the community is of course an issue in sign language research. Some of this giving back can be said to be of a “global” nature while some of it is of a more “local” nature. Stokoe’s pioneering work in the early 1960s clearly demonstrated that sign languages are full-fledged linguistic systems structurally independent from the spoken languages with which they come in contact and all of the work that flowed from that on the structure and use of sign languages has constituted giving back of a “global” nature: this legitimization of sign languages as “real” languages has made possible the use of sign languages as the media of instruction in the education of deaf children. This has meant that many deaf children all over the world now have unfettered access to academic content through the visual sign languages that they understand, as opposed to through oral systems or systems which combine speaking and signing (Johnson, Liddell, and Erting 1989). It has led to the recognition of sign languages as official languages in many coun- tries, which in turn has meant that deaf people can participate in the social and legal systems of their countries by means of interpreters. Sociolinguistic research on bilingualism and language contact, variation, language planning and language policy, and language attitudes pertaining to sign languages has reinforced the status of sign languages as “real” languages and has revealed both parallels and differences with spoken languages. Giving back of a “local” nature has involved sharing the research findings in an accessible way, as in the book and DVD entitled What’s Your Sign for PIZZA?, (Lucas, Bayley, and Valli 2003) which provides an introduction to variation in ASL for a popular audience and of which copies were provided to the communities that participated in the original study. This giving back had Wolfram and Schilling-Estes’ (1997) model for the Ocracoke commu- nity specifically in mind, with their book produced for a popular audience, the royalties for which went to the historical society of the island. The royalties for PIZZA are donated to a scholarship fund for deaf students. A similar book-DVD package has been published about the Black ASL project (McCaskill et al. 2011). And local giving back also includes numerous free workshops about projects for community groups and interpreter training programs and the incorporation of project findings into course curricula. 302 Ceil Lucas

Finally, deaf undergraduate and graduate students serve as research assis- tants on sociolinguistic research projects and thus receive a very direct intro- duction to the world of sign language research. Many of these students go on to advanced degrees and become teachers themselves.

Notes

1. Portions of this paper are adapted from Lucas (2012), Hoopes et al. (2001), Lucas, Bayley, and Valli (2001), and Lucas and Bayley (2010). 2. The use of upper-case Deaf here indicates cultural deafness, as opposed to the strictly audiological condition indicated by lower-case deaf. Both uses are conventional in the literature on deafness. 3. Oralism is a method that involves teaching deaf children solely through speech, to the total exclusion of any signing; Total Communication (Denton 1976) was a philosophy supporting the use of both speech and signing, most often simultaneously 4. Tactile Signing is not limited to the United States; see, for example, Mesch (2000) on tactile signing in Sweden. 5. There have been efforts to devise writing systems for sign languages, most notably SignWriting devised by Valerie Sutton; see www.signwriting.org

References

Baker-Shenk, Charlotte, and Jim Kyle 1990 Research with deaf people: Issues and conflicts. Disability, Handicap and Society 5: 65–75. Battison, Robbin, Harry Markowicz, and James Woodward 1975 A good rule of thumb: Variable phonology in American Sign Language. In Analyzing Variation in Language, Ralph Fasold and Roger W. Shuy (eds.), 291–302. Washington, DC: Press. Bayley, Robert, Ceil Lucas, and Mary Rose 2000 Variation in American Sign Language: The case of DEAF. Journal of Sociolinguistics 4 (1): 81–107. Methodological issues in studying sign language variation 303

Bell, Allan 1984 Language style as audience design. Language in Society 13: 145– 204. 2001 Back in style: Reworking audience design. In Style and Sociolinguistic Variation, Penelope Eckert and John Rickford (eds.), 139–169. Cambridge: Cambridge University Press. Brentari, Diane 2002 Modality differences in sign language phonology and morphophonemics. In Modality in Language and Linguistic Theory, Richard Meier, David Quinto-Pozos, and Kearsy Cormier (eds.), 35–64. Cambridge: Cambridge University Press. Burch, Susan, and Hannah Joyner 2007 Unspeakable: The Story of Junius Wilson. Chapel Hill: University of North Carolina Press. Chen Pichler, Deborah 2010 Sources of handshape error in first-time signers of ASL. In Deaf Around the World, Donna J. Napoli and Gaurav Mathur (eds.), 96–121. Oxford: Oxford University Press. 2012 Acquisition. In Sign Language: An International Handbook, Bencie Woll, Markus Steinbach, and Roland Pfau (eds.), 647–686 (Handbooks of Linguistics and Communication Science (HSK) series). Berlin: Mouton de Gruyter Collins, Steven C. 2004 Adverbial morphemes in tactile American Sign Language. Ph.D. dissertation, The Union Institute. Collins, Steven C., and Karen Petronio 1998 What happens in tactile ASL? In Pinky Extension and Eye Gaze: Language Use in Deaf Communities, Ceil Lucas (ed.), 18–37. Washington, DC: Press. Denton, David 1976 The philosophy of Total Communication. Supplement to the British Deaf News. Carlisle, UK: The British Deaf Association. Emmorey, Karen 1999 The confluence of space and language in signed languages. In Language and Space, Paul Bloom, Mary A. Peterson, Lynn Nodel, and Merrill F. Garrett (eds.), 171–209. Cambridge, MA: MIT Press. 304 Ceil Lucas

Feagin, Crawford 2002 Entering the community: Fieldwork. In The Handbook of Language Variation and Change. Jack Chambers, Peter Trudgill, and Natalie Schilling-Estes (eds.), 20–39. Oxford: Blackwell. Frishberg, Nancy 1975 Arbitrariness and iconicity: Historical change in American Sign Language. Language 51: 696–719. Giles, Howard 1973 Accent mobility: A model and some data. Anthropological Linguistics 15: 87–105. 2001 Couplandia and beyond. In Style and Sociolinguistic Variation, Penelope Eckert and John R. Rickford (eds.), 211–219. Cambridge: Cambridge University Press. Giles, Howard, and Peter F. Powesland 1975 Speech Style and Social Evaluation. London: Academic Press. Gustason, Gerilee, Donna Pfetzing, and Esther Zawolkow (eds.) 1972 Signing Exact English. Los Alamitos, CA: Modern Signs Press. Harris, Raychelle, Heidi Holmes, and Donna Mertens 2009 Research ethics in sign language communities. Sign Language Studies, 9: 104–131. Hoopes, Rob 1998 A preliminary examination of pinky extension: Suggestions regarding its occurrence, constraints and function. In Pinky Extension and Eye Gaze: Language Use in Deaf Communities, Ceil Lucas (ed.), 3–7. Washington, DC: Gallaudet University Press. Hoopes, Rob, Mary Rose, Robert Bayley, Ceil Lucas, Alyssa Wulf, Karen Petronio, and Steven Collins 2001 Analyzing variation in sign languages: Theoretical and methodological issues. In Signed Languages: Discoveries from International Research, Valerie Dively, Melanie Metzger, Sarah Taub, and Anne Marie Baer (eds.), 135–162. Washington, DC: Gallaudet University Press. Johnson, Robert, Scott Liddell, and Carol Erting 1989 Unlocking the Curriculum: Principles for Achieving Access in Deaf Education. Gallaudet Research Institute Working Paper 89 (3). Washington, DC: Gallaudet University. Johnson, Robert, and Scott Liddell 1 201 Toward a phonetic representation of signs, 1: Sequentiality and contrast. Sign Language Studies 11: 241–273. Methodological issues in studying sign language variation 305

Labov, William 1972a Sociolinguistic Patterns. Philadelphia: University of Pennsylvania Press. 1972b Language in the Inner City. Philadelphia: University of Pennsylvania Press. 1982 Objectivity and commitment in linguistic science. Language in Society 11: 165–201. 1984 Field methods of the project on language change and variation. In Language in Use: Readings in Sociolinguistics, and Joel Sherzer (eds.), 28–53. Englewood Cliffs, NJ: Prentice-Hall. Lane, Harlan, Robert Hoffmeister, and Ben Bahan 1996 A Journey into the DEAF-WORLD. San Diego: DawnSign Press. Liddell, Scott, and Robert Johnson 1989 American Sign Language: The phonological base. Sign Language Studies 64: 195–278. Lucas, Ceil 2013 Methods for sign languages. In The Oxford Handbook of Sociolinguistics. Robert Bayley, Richard Cameron, and Ceil Lucas (eds.), Oxford: Oxford University Press, 280–298. Lucas, Ceil, and Robert Bayley 2010 Variation in American Sign Language. In Sign Languages, Diane Brentari (ed.), 451–475. Cambridge: Cambridge University Press. Lucas, Ceil, Robert Bayley, and 2001 Sociolinguistic Variation in American Sign Language. Washington, DC: Gallaudet University Press. 2003 What’s Your Sign for PIZZA? An Introduction to Variation in ASL. Washington, DC: Gallaudet University Press. Lucas, Ceil, Amber Goeke, Rebecca Briesacher, and Robert Bayley 2007 Phonological variation in ASL: 2 hands or 1? Paper presented at NWAV 36, University of Pennsylvania, Philadelphia, October. Lucas Ceil, and ClaytonValli 1992 Language Contact in the American Deaf Community. San Diego: Academic Press. McCaskill, Carolyn, Ceil Lucas, Robert Bayley, and Joseph Hill 1 201 The Hidden Treasure of Black ASL: Its History and Structure. Washington, DC: Gallaudet University Press. 306 Ceil Lucas

Mesch, Johanna 2000 Tactile Swedish Sign Language: Turn-taking in signed conversations of people who are deaf and blind. In Bilingualism and Identity in Deaf Communities, Melanie Metzger (ed.), 187–203. Washington, DC: Gallaudet University Press. Milroy, Lesley 1987 Observing and Analyzing Natural Language. Oxford: Blackwell. Osborne, Ray, and McPhee, Rob 2000 Indigenous terms of reference (ITR). Presented at the 6th UNESCO-ACEID International Conference on Education, Bangkok, December 12–15. Palmer, Jeffrey, Wanette Reynolds, and Rebecca Minor. 2012 “You want WHAT on your pizza ?!”: Videophone and video relay service as potential forces on lexical standardization of American Sign Language. Sign Language Studies 12 (3): 371–397. Petitto, Laura A. 2000 On the biological foundations of human language. In The Signs of Language Revisited: An Anthology in Honor of Ursula Bellugi and Edward Klima, Karen Emmorey and Harlan Lane (eds.), 447–471. Mahwah, NJ: Lawrence Erlbaum. Ramsey, Claire 1989 Language panning in Deaf education. In The Sociolinguistics of the Deaf Community, Ceil Lucas (ed.), 123–146. San Diego: Academic Press. Schembri, Adam, David McKee, Rachel McKee, Sara Pivac, and Trevor Johnston 2009 Phonological variation and change in Australian and New Zealand Sign Languages: The location variable. Language Variation and Change 21: 193–231. Singleton, Jenny, Gabrielle Jones, and Shilpa Hanumantha 2012 Deaf friendly research? Toward ethical practice in research involving Deaf participants. Deaf Studies Digital Journal 3. Stokoe, William 1960 Sign Language Structure: An Outline of Visual Communication Systems of the American Deaf. Studies in Linguistics: Occasional Paper 8. Buffalo, NY: University of Buffalo, Linguistics Department. Stokoe, William, Dorothy Casterline, and Carl Croneberg 1965 A Dictionary of American Sign Language. Silver Spring, MD: Linstok Press. Methodological issues in studying sign language variation 307

Taub, Sarah, Dennis Galvan, and Pilar Piñar 2008 The role of gesture in crossmodal typological studies. Cognitive Linguistics 20: 71–92. Taub, Sarah 2001 Language from the Body. Cambridge: Cambridge University Press. Van Herreweghe, Mieke and Myriam Vermeerbergen 2012 Data collection. In Sign Language: An International Handbook, Bencie Woll, Markus Steinbach, and Roland Pfau (eds.), 1023– 1044 (Handbooks of Linguistics and Communication Science (HSK) series). Berlin: Mouton de Gruyter Wolfram, Walt 2013 Language awareness in community perspective: obligation and opportunity. In The Oxford Handbook of Sociolinguistics Robert Bayley, Richard Cameron, and Ceil Lucas (eds.), Oxford: Oxford University Press, 754–772. Wolfram, Walt 1993 Ethical considerations in language awareness programs. Issues in Applied Linguistics 4: 225–255. Wolfram, Walt and Natalie Schilling-Estes. 1997 Hoi Toide on the Outer Banks. Chapel Hill: University of North Carolina Press. Woodward, James, Carol Erting, and Susanna Oliver 1976 Facing and hand(l)ing variation in American Sign Language. Sign Language Studies 10: 43–52. Woodward, James, and Susan DeSantis 1977 Two to one it happens: Dynamic phonology in two sign languages. Sign Language Studies 17: 329–346.