Cormier, Kearsy. 1998. Grammatical and Anaphoric Agreement in American Sign Language. MA Thesis, University of Texas at Austin. Download thesis (PDF format)
Previously, researchers of ASL have proposed that structural agreement exists in all ASL sentences (e.g., Aarons et al. 1992), or that agreement exists in sentences containing one type of verb but not with other types (Lillo-Martin 1986). Following Bresnan & Mchombo (1987), I suggest that there are two types of locus agreement that occur in ASL, depending on whether the verb has overt or non-overt arguments. These two types of agreement are referred to as grammatical and anaphoric agreement.
The interaction of the locus agreement feature with these types of agreement will be presented using an analysis within the framework of Head-driven Phrase Structure Grammar (HPSG). Using as a basis the HPSG analysis of locus agreement presented in Cormier, Wechsler & Meier (1998), I will use argument structure to account for grammatical agreement and adjunction to account for anaphoric agreement. This analysis captures the fact that either grammatical or anaphoric agreement can occur without significantly changing the meaning of the sentence. Thus the real difference between grammatical and anaphoric agreement lies simply in the overt/non-overt nature of the verb’s arguments.
Cormier, Kearsy. 2000. How Does Modality Contribute to Linguistic Diversity? Presented at the Annual Meeting of the Linguistic Society of America, Chicago, IL, January 6-9, 2000. Download second qualifying paper (same title, PDF format)
Large-scale studies on linguistic diversity (e.g., Greenberg 1966) have attempted to "establish typologically, historically, and areally stable features in the world’s languages [and] to sort out genetic, geographical and universal determinants of linguistic patterning" (Nichols 1992). These studies generally sample a wide variety of the world’s languages from many different stocks and families, with the assumption that a more diverse sample leads to a more explanatory description/typology. Unfortunately, such studies have focused only on spoken languages. New insight into linguistic diversity can be gained by including languages in other modalities (such as the visual-gestural modality of signed languages), thus adding modality as yet another determinant of linguistic patterning.
In this paper I examine the distribution of one particular morphological feature,
inclusive/exclusive marking, among both spoken languages and signed languages (cf.
Forchheimer 1953; Ingram 1978 for spoken language data). Emphasis will be on American
Sign Language (ASL), although examples from pronominal systems of other signed languages
will also be discussed.
One reason that the pronominal systems of signed languages have not been typologically
compared to those of spoken languages may be that the pronominal systems of signed
languages are often taken to be fully indexic (Lillo-Martin & Klima 1990; Liddell
1995). Many signs in ASL and other signed languages are indeed indexic; indexic signs
"point to" their referents. However, I present evidence suggesting that
the pronominal system of ASL is not purely indexic (cf. also Meier 1990). I argue
that an inclusive/exclusive distinction can be found in both indexic and non-indexic
pronominal forms in ASL.
This study shows several interesting differences between spoken and signed languages
with regard to inclusive/exclusive marking. For example, markedness relationships
in ASL are different from those in spoken languages. Many spoken languages mark only
inclusive, but no spoken language marks only exclusive. In ASL, inclusive may or
may not be marked, but exclusive seems always to be marked. Specifically, indexic
forms can mark both inclusive and exclusive while non-indexic forms can mark only
exclusive. Thus even with pronominal forms that are not indexic, the signed modality
still appears flexible enough to allow very specific pronominal reference.
Another difference that arises from this study is that the inclusive/exclusive distinction
in spoken languages is an areal and genetic phenomenon, whereas the inclusive/exclusive
distinction may exist in all signed languages examined to date. The fact that the
signed modality practically entails an inclusive/exclusive distinction suggests that
modality itself may be a typological determinant that can be as robust as the other
determinants mentioned by Nichols (1992) - i.e., presence of other typological features,
area, and genetic affinity. This has major implications for Nichols’ framework and
design.
Accordingly, the fact that this morphological feature patterns so differently in spoken and signed languages highlights the importance of including signed languages in language typologies and studies of linguistic diversity. Any language typology that does not include signed languages may be making inaccurate generalizations about the world’s languages and about the possibilities of human languages in general.
Many researchers have noted that deixis in American Sign Language (ASL) is largely indexic; deictic signs such as pronouns and agreeing verbs "point to" locations associated with their referents (e.g., Liddell 2000). This property is referred to as directionality. Number has traditionally been considered to have little or no effect on directionality in signed languages - e.g. according to Liddell, number does not control the directionality of an agreeing verb. Thus, for purposes of simplicity, discussions of deixis and agreement in ASL have long focused on singulars. However, I argue that number may - at a miminum - influence the degree of directionality, affecting how precisely the verb points to an established location.
This study focuses on indexers (i.e. signs that serve to establish a referent or referents at a location in space) and on agreeing verbs; both indexers and agreeing verbs are considered highly indexic in their singular forms. Researchers often note that agreeing verbs index the same location as their coreferential pronouns, but no study has looked at how closely these two locations actually match in three-dimensional space. One goal for this study is to determine how closely a singular verb approaches the location established in space by the coreferential indexer (henceforth, the target indexer location). This study will also determine how precisely verbs indicating plural referents approach their target indexer location(s). If plural verbs are less prec ise than singulars, number may at least affect directionality (if not control it); this finding would be contrary to the traditional view of number in signed languages.
In order to measure this precision, I examined the distance between the locations of the indexer and the verb in three-dimensional space, using kinematic data collected from three Deaf native signers of ASL. Informants were presented with videotaped stimuli consisting of short skits, each one depicting different numbers of actors either performing or undergoing certain actions. Informants described what was happening in each skit and were recorded using 5 infrared-sensitive cameras to capture their hand movements in three-dimensional space. The locations of indexers were then compared with the locations of the verbs that agreed with them.
Preliminary results show that the mean distance between singular indexers and verbs is 124.9mm (SD=58.4). These results give us the first concrete estimate of the precision with which verbs agree with previously established locations. In contrast, the mean distance for plurals was 175.2mm (SD=79.9). Furthermore, the mean distance between singular verbs and their target indexers is significantly smaller than the mean distance between plural verbs and their target indexers (F(1,67) =9.31, p<.003). This suggests that singulars do indeed more precisely indicate their referents than plurals do.
These results have implications for the field of sign linguistics, in which the linguistic status of directionality has recently been a highly controversial issue (Lillo-Martin in press, Rathmann & Mathur in press). The fact that plurality affects directionality suggests that directionality must be at least partly linguistic because it is affected by the grammatical category number.
Cormier, Kearsy. 2002. Is the use of space in ASL linguistic or gestural? Evidence from plurals. Presented at Gesture: The Living Medium, Austin, TX, June 5-8, 2002.
Cormier, Kearsy. 2004. The categorical status of first person in signed languages. To be presented at the Annual Meeting of the Linguistic Society of America, Boston, MA, January 8-11, 2004.
Over the years there has been much debate among sign linguists about the status of person in signed languages. Some have claimed a three-person system (Friedman 1975), whereas others (Meier 1990) have proposed a two-person system which includes first and non-first persons. On still other analyses, sign languages do not exhibit person categories at all; instead, the use of signing space for reference may be fully gradient and therefore gestural rather than linguistic (Liddell 2000; McBurney 2002). Here I argue for the categorical status of first person in signed languages. I present data from two unrelated languages: American and British Sign Languages (ASL & BSL).
In both ASL and BSL, a singular pronoun referring to the signer is a point to the signer's chest. Both languages have several different first person plural pronouns, including (among others) WE and 3/4/5-OF-US. These pronouns are produced on -- or just in front of -- the signer's chest, often at the centre but possibly displaced to the right or left side.
For subject agreement with first person plurals (1PP) in both languages, the movement of the verb starts at the centre of the signer's chest. This is the case no matter where the 1PP pronoun is produced. Thus, even when a displaced pronoun is used (i.e. to the signer's left or right), the agreeing verb form starts at the centre of the signer's chest.
Object agreement is more robust in both languages and can occur with plural marking as well. For exhaustive (distributive) and multiple (non-distributive) plural marking with 1PP objects, the verb moves inward toward the signer and generally ends at -- or includes one verb path directed toward -- the centre of the signer's chest. Verbs may also be marked for collective plural, which for 1PP consists of a single inward path movement toward the signer's chest. If the verb is used with a 1PP pronoun that is displaced to the right or left side, the collective plural verb might also be displaced to that same side, but this displacement is not required in either language. If the displacement does not occur, the default location for 1PP agreement is the centre of the signer's chest.
These facts about subject and object agreement provide strong evidence for the claim that both ASL and BSL have a distinct first person category. Reference to the signer is unlike reference to the addressee(s) or non-addressed participant(s) for whom there is no single default location but "agreement" instead depends on the location of the referent(s) in space. For non-first person, the use of space for reference may indeed be fully gradient and best thought of as gestural rather than as a linguistic process of agreement. But for first person it is not; for both subject and object agreement, the default location for first person is at the centre of the signer's chest. The fact that this is true for two unrelated signed languages gives further support to the distinctness of the first person category in signed languages.
Cormier, Kearsy. 2004. Exclusive pronouns in ASL and BSL. To be presented at Sign Languages: A crosslinguistic perspective at the Annual Meeting of the German Linguistic Society (DGfS), Mainz, Germany, February 25-27, 2004.
Many researchers have noted that the pronominal systems of signed languages are largely indexic. That is, pronouns in signed languages “point to” locations associated with their referents. Furthermore, it has also been noted that signed languages are quite uniform with regard to the indexicality of pronouns. This does indeed appear to be the case for singular pronominal forms. However, the indexicality of plural pronouns is not so clear. In this paper I present evidence suggesting that the pronominal systems of signed languages, particularly first person plural forms, may not be purely indexic, suggesting that there may some variation in indexicality across signed languages.
American Sign Language (ASL) is one language that allows non-indexic forms of first person plural pronouns. One variant of the ASL pronoun WE is produced with a large arcing movement which begins and ends at the center of the signer’s chest. This form is indexic; some have claimed that this form is used to indicate referents who are present in the discourse.
However, another variant of the ASL pronoun WE (produced by two touches on the signer’s chest) is not indexic – this pronoun does not point to any of its referents other than the signer. This pronoun also has an exclusive variant (exclusive of any salient referent in the discourse), which is displaced to either the signer’s ipsilateral or contralateral side.
The exclusive variant (WE- displaced) may be indexic – that is, the location where the pronoun is produced (left or right side) may match the location associated with the included referents. For example, the pronoun WE- displaced can be located on the signer’s right side when the referents are also located on the signer’s right side. However, the exclusive variant is not always indexic. For example, WE- displaced may be located on the signer’s left side, even when the referents are located on the signer’s right.
The current study investigates whether British Sign Language (BSL), like ASL, also allows non-indexic first person plural pronouns. BSL has several different first person plural pronouns. One pronoun uses the index finger pointed downward with a small circular movement in the center of neutral space; this sign is not indexic. Like ASL, this pronoun has an exclusive variant (exclusive of any salient referent in the discourse), which is displaced to either the signer’s ipsilateral or contralateral side. If the referents are physically present, the displacement of the pronoun (i.e., to the signer’s left or right side) must match the location of the referents. Thus the use of non-indexic WE-displaced in BSL is ungrammatical. This is unlike ASL, where the displacement of the pronoun may or may not match the location of the referents.
We know that there is some variation across the pronominal systems of signed languages. For example, handshapes and locations can vary; there is also some variation in number distinctions. By and large, however, pronouns have been thought to be fairly uniform across signed languages. In one respect the data presented here support this assumption, since both ASL and BSL appear to have exclusive variants of first person plural pronouns. However, the data in this study also show some variation in pronominal systems, since BSL seems to be more restricted in its indexicality with exclusive pronouns than is ASL. These findings point to a need for more crosslinguistic research on signed languages, both where we might expect differences, but perhaps especially in those areas in which we expect strong similarities, such as the way that signed languages use space.
Tyrone, Martha E. & Kearsy Cormier. 2004. One hand or two: How ASL & BSL restructure fingerspelling in loans. Presented at the 12th Manchester Phonology Meeting, Manchester, May 20-22, 2004.
Lexical borrowing takes a variety of forms in signed languages, including mouthing, fingerspelling, and initialised signs (Sutton-Spence, 1994; Battison, 1978; Boyes Braem, 2001). This paper focuses on the last two of these, their structural relationship to each other, their influence on handshape inventories and their implications for theories of sign phonology. Signed language provides a unique perspective on lexical borrowing because the that is being borrowed from is often of a different modality (i.e. spoken rather than signed).
Although the United States and Great Britain share a common spoken language, the languages of the Deaf communities, American Sign Language and British Sign Language (henceforth ASL and BSL), are unrelated and have distinct lexicons and grammars. Another difference between ASL and BSL is the structure of the fingerspelling systems: ASL fingerspelling is one-handed, while BSL fingerspelling is two-handed. Fingerspelling is used to represent written words from the ambient spoken language. Each distinct configuration of the hand(s) corresponds to a written letter, and the letters are produced serially, but with movement overlap and coarticulation (Wilcox, 1992; Akamatsu, 1985). So even though ASL and BSL borrow mostly from the same language (i.e. English), they do so in remarkably different ways, because of the way in which words are borrowed. ASL fingerspelling relies primarily on contrasting handshapes, whereas BSL fingerspelling relies on both handshape and location on the non-dominant hand, to differentiate one fingerspelled letter from another.
Both fingerspelling systems use only a limited number of contrastive locations and movements compared to signs from the native lexicons: ASL and BSL fingerspelling are both relatively fixed in space and neither uses locations on the body or large movements to differentiate letters. The BSL fingerspelling system does not require more handshapes than its native lexicon because with two hands a larger number of formations with basic handshapes are possible. However, because ASL fingerspelling uses only one hand and very few locations and movements, it requires more possible values for handshape than its native lexicon. Consequently, a larger proportion of the handshapes in ASL fingerspelling (e.g. R and T) are physiologically difficult to produce (Mandel, 1979; Ann, 1996). Because multiple parameters are used to differentiate one sign from another in the native lexicon, these signs have proportionally more handshapes that are easy to produce (Ann, 1996, 1993). For example, handshapes with all the fingers flexed or extended, which are easily acquired and produced by humans (Schieber, 1996; Brooks, 1986), are extremely common in the native lexicons of ASL and BSL.
Initialised signs represent an intermediate form between fingerspelling and native signs; their structure is influenced by the structure of both sign and fingerspelling. These signs are produced by combining (usually the first) fingerspelled letter from a corresponding English word with one or more other phonological parameters. Because in BSL fingerspelling the two hands make contact, added locations on the body cannot be used to create new signs, due to phonological restrictions on the number of locations per sign (Brentari, 1998; Sandler, 1989). Consequently, because BSL fingerspelling uses both handshape and location, BSL initialised signs are often nearly identical in form to the fingerspelled letter. In contrast, initialised signs in ASL use handshapes from the fingerspelling system which can combine with many different movements and locations without masking the identity of the letter. This pattern of initialisation in the two languages indicates that signs which have developed from fingerspelled forms cannot violate phonological rules governing native signs, but can nonetheless recruit forms (i.e. handshapes) that are less common in the native lexicon. The larger handshape inventory of the ASL fingerspelling system thus leads to a much larger proportion of initialised signs in ASL than in BSL.
Cormier, Kearsy. 2004. Loss of indexicality with plural pronouns: Motoric and linguistic explanations. Presented at the 8th Theoretical Issues in Sign Language Research, Barcelona, Spain, September 30-October 2, 2004.
Many researchers have noted that the pronominal systems of signed languages are largely indexic. That is, pronouns in signed languages “point to” locations associated with their referents (Liddell, 2002). Furthermore, it has also been noted that signed languages are quite uniform with regard to the indexicality of pronouns (for example, see McBurney, 2002 for data on five different signed languages). Singular pronominal forms do indeed appear to be highly indexic. However, the indexicality of plural pronouns is not so clear. In this paper I present evidence from both American Sign Language (ASL) and British Sign Language (BSL) suggesting that the pronominal systems of these signed languages, particularly first person plural forms, may under some circumstances lose their indexicality. This loss of indexicality, I argue, is largely due to two types of tendencies: one motoric and one linguistic.
For the current study, data were collected from three ASL signers and three BSL signers. All were Deaf native signers from Deaf families. The participants were shown a questionnaire which was devised to elicit different forms of first person plural pronouns. The participants were asked to translate English sentences - each with a specific context - into ASL or BSL. In the context given with each sentence, the number of referents was manipulated so as to elicit dual, trial and plural first person forms. Possessive first person plural forms were also elicited. Furthermore, since the physical location of discourse participants may be crucial to how indices are set up in the signing space, a set of props was used to help the participants imagine real-world discourse situations.
The data were coded to determine how indexic the pronouns were with respect to the location of the referents (that is, the props that were placed in front of the signer during data collection). Pronouns were coded as located on the signer’s right side, at the centre of the chest, or on the signer’s left side. These values were then compared to the location of the props that the signer was referring to when producing each pronoun. If the referents were located on the signer’s right side and the signer produced the pronoun on his/her left side, then that pronoun token was coded as indexic (i.e. a match between referent and pronoun locations). However, in this situation, if the signer produced the pronoun either at the centre or of his/her chest or on his/her left side, then that pronoun token would be coded as non-indexic (i.e. a mismatch between referent and pronoun locations).
Results from both languages included first person plural pronouns such as WE, OUR, TWO-OF-US, and THREE-OF-US. Indexicality levels of these pronouns were less than what one might expect from singular pronouns. Of 134 tokens, 66% of BSL pronouns were indexic (i.e. the pronoun’s location matched the location of the referents in 66% of the pronouns produced), while 40% of ASL pronouns (N=115) were indexic.
Examining some examples of pronouns that were not indexic (i.e. the pronoun’s location did not match the location of its referents) reveals two major trends, which occurred to varying degrees with all of the pronouns noted above.
First, when the referents are located at the centre or on the signer’s contralateral side, the pronoun may be produced on the ipsilateral side causing a mismatch in location. I refer to this as ipsilateralisation. The reason ipsilateralisation occurs may be due to ease of articulation – that is, it may be physically easier to produce the pronoun on the ipsilateral side of the body than to produce it at the centre of the chest or to cross the midline and produce it on the opposite side.
Second, when the referents are located on the signer’s ipsilateral or contralateral side, the pronoun may be produced at the centre of the signer’s chest, again causing a mismatch in location. I refer to this as centralisation. There are two situations in which centralisation occurs in the data from this study. One is when the referents are located on the signer’s contralateral side, and the pronoun is produced at the centre of the chest. This is similar to ipsilateralisation as described above, i.e. the pronoun is produced more ipsilaterally than it perhaps would be if it were indexic. Thus the reason centralisation occurs may likely be due to motoric factors such as ease of articulation. With plurals, the locations of individual referents may not be as important as the fact that the number of referents is plural.
The other situation in which centralisation occurs is when the referents are located on the signer’s ipsilateral side and the pronoun is produced at the centre of the chest. In this case the pronoun is actually moving away from the ipsilateral side. The motorically easiest location for the pronoun would be the ipsilateral side, which would result in an indexic pronoun. How can we account for the loss of indexicality in these cases?
I suggest that the preferred location for first person reference in ASL and BSL is the centre of the signer’s chest, and that centralisation away from the ipsilateral side is due to first person marking. As with ipsilateralisation, the locations of individual referents may not be as important with plurals, and so the location of the pronoun can be centralised as a way of unambiguously marking that the pronoun is a first person pronoun.
The vast majority of the non-indexic pronouns from this study were ipsilateralised or centralised. The explanations offered here for ipsilateralisation are largely motoric (ease of articulation) while those for centralisation are either motoric (in cases of ipsilateralisation toward the centre) or linguistic (in cases of centralisation away from the ipsilateral side in favor of first person marking). These results support the notion that signed languages as grammatical systems are subject to both motoric and also linguistic constraints.
Cormier, Kearsy, Peter Carss & Clark Denmark. 2005. SignLab: Innovation and Technology in Sign Language Teaching. Presented at Moving Forward with Languages: Innovative Approaches to Language Teaching, University of Central Lancashire, Preston, UK, January 12, 2005.
Sign language teaching poses challenges not normally encountered with most language teaching situations. As visual languages, sign languages such as British Sign Language (BSL) cannot be easily transcribed or written. Instead, sign language teaching relies heavily on video technology. Indeed within the last 25 years, analog video has been very useful because it allows students and teachers to record sign language use and play it back again for study and review. However, analog video certainly has its disadvantages: videotapes wear out, get lost, are cumbersome to copy, and can be difficult to navigate with precision.
The arrival of digital video technologies offers potential for enormous improvement. The SignLab at the Centre for Deaf Studies (University of Bristol), takes advantage of such technologies. This lab allows students to watch digital video to work on reception skills. Students are also able to film themselves signing, for development of production skills. The software we use (Panda), developed exclusively at CDS, allows for live video capture and MPEG4 compression at the press of a button.
Our SignLab environment has drastically changed the way we teach our students. Digital video is higher quality and easy to navigate in a digital media player. Storage of teaching materials on a central server allows these materials to be shared by all tutors and students and also helps facilitate autonomous learning. Furthermore, our Panda software makes capture, compression and storage simple for even the most novice computer user.
Cormier, Kearsy. In press. Exclusive pronouns in American Sign Language. In Clusivity: Typology and case studies of inclusive-exclusive distinction, ed. by Elena Filimonova, 241-68. Amsterdam: John Benjamins.
This paper examines the distribution of inclusive/exclusive marking among spoken and signed languages, particularly American Sign Language (ASL). Pronominal systems of signed languages are often taken to be fully indexic – that is, signed pronouns ‘point to’ their referents. The results of this study show distinct exclusive marking in ASL. An exclusive pronoun need not be indexic as other ASL pronouns are. Furthermore, results also show the context-dependent nature of exclusion in ASL compared to spoken languages. While most spoken languages with this distinction can include/exclude only the addressee, ASL can use one exclusive marking to exclude any salient discourse participant.This paper highlights the importance of including signed languages in language typologies, to ensure accurate generalizations about the world’s languages.
Cormier, Kearsy & Adam Schembri. To appear. Sign lexicology: The structure of sign language vocabulary. In Sign Linguistics: An Introduction, ed. by A. Schembri, G. Turner & T. Johnston. Edinburgh: Edinburgh University Press.
This chapter provides an overview of the notions of native (i.e., sign) versus non-native (e.g., fingerspelled) vocabulary in signed languages, as well as the concept of the core/established lexicon (i.e., lexical signs) versus the non-core/productive lexicon (i.e., classifier signs). It also includes a discussion of lexical categories in signed languages, discussing some of the difficulties in applying these traditional notions to the vocabulary of signed languages.
Cormier, Kearsy, Schembri, Adam, and Tyrone, Martha. 2005. One hand or two: A cross-linguistic analysis of the non-native lexicon in signed languages. Paper to be presented at Sign Language Linguistics and the Application of Information Technology to Sign Languages, Milan, Italy, June 22-24, 2005.
Signed languages exist in a unique sociolinguistic situation compared to spoken languages–the constant contact between the signed language and spoken language used by the surrounding hearing community results in a range of language contact phenomena not found in spoken languages (Lucas & Valli, 1992). Such contact often has a major impact on the lexicon of a signed language, in particular resulting in a non-native part of the lexicon which includes a variety of different types of borrowings, including mouthing and also elements from fingerspelling systems. This paper examines the non-native lexicons of signed languages with one-handed versus two-handed fingerspelling systems, in particular American Sign Language (ASL) and the British Sign Language family (abbreviated here as BANZSL, this family includes British, Australian and New Zealand Sign Languages). The aim of this paper is to examine the model of the non-native lexicon as proposed by Brentari & Padden (2001) for ASL and to evaluate how well this model applies to signed languages with two-handed fingerspelling systems such as BANZSL.
In the ASL fingerspelling system (which is one-handed), the hand (usually the signer’s dominant hand) takes on different shapes to represent various different letters of the English alphabet. The letters are distinguished from each other for the most part by handshape, but in some instances by movement or orientation. The BANZSL fingerspelling system is two-handed and thus involves both the dominant and non-dominant hand. To represent almost all the letters of the English alphabet, the dominant hand contacts various locations on the non-dominant hand (-C- being the only one-handed exception). In fluid fingerspelling, the non-dominant hand is rarely completely passive; the non-dominant hand and fingers move to facilitate contact. Twenty five of the manual letters are distinguished from each other by handshape, location, and orientation (and in two cases by movement). Thus while the ASL fingerspelling system distinguishes letters largely by handshape, all of the major parameters of sign formation (handshape, location, movement, orientation and particularly the relationship between the two hands) are important for distinguishing almost all the manual letters in BANZSL.
Brentari and Padden’s (2001) model of the non-native lexicon is meant to account for various degrees of nativisation for fingerspelled elements. Thus they propose a set of four categories of non-native signs, each category more and more nativised–we will refer to these categories here as A through E. The two most nativised categories (A and B) consist of signs that violate no native rules. The first (A) includes signs (e.g. ASL #SAY-NO) which consist of handshapes, movements, and locations that are found in lexical signs. These signs may be best thought of as originating from the non-native lexicon but are now part of the native lexicon (in fact, in Brentari and Padden’s model this category actually overlaps with the native lexicon). The next category (B) also includes signs that violate no rules of the native lexicon. For one-handed systems, signs in this category have movement and locations found in lexical signs, but they use handshapes taken from the fingerspelling system rather than from the core lexicon. This category includes many initialised signs such as ASL WATER which uses the -W- handshape. The next category (C) includes signs the violate the constraint that each sign must have only one set of selected fingers–that is, these signs involve a non-native handshape (e.g. ASL WITHDRAW). The next category (D) includes signs that violate the selected fingers constraint in addition to the constraint that handshape change should maximise aperture. This category includes signs that have more than two handshapes (e.g. some fingerspelled loan signs such as #EASY). The final, least nativised category (E) includes signs which violate the same constraints as the other categories, but also violate the constraint that signs have at most two different movements. This category includes fully fingerspelled words in ASL (e.g. F-R-E-E).
Brentari & Padden’s (2001) model relies heavily (and almost exclusively) on handshape. Since fingerspelled letters in one-handed systems are distinguished almost entirely by handshape, a model based on handshape makes sense. But there are problems with this model even when only considering one-handed fingerspelling systems. One problem is that there is an assumption that all fingerspelled letters in ASL have marked handshapes (i.e. that handshapes in the fingerspelling system do not also occur in the native lexicon). This is largely but not entirely true. The handshapes used with the letters S and H are handshapes that also occur in the native lexicon (e.g. in the signs WORK and TRAIN, respectively). Thus for some signs (e.g. SOCIETY and HISTORY, initialised with S and H respectively), it is not clear whether they would belong in category A or category B.
More importantly, though, Brentari & Padden’s model is clearly limited to signed languages that have a one-handed fingerspelling system. Because signed languages with two-handed fingerspelling systems use all of the major parameters in distinguishing different letters, there are problems in trying to fit signed languages with two-handed fingerspelling systems into their model which relies heavily on distinctions in handshape.
Thus we propose a more inclusive model of the sign language lexicon that aims to account for varying degrees of nativisation based on the extent to which all native parameters (i.e. native handshapes, movements, locations and importantly native combinations of the three) exist within a given sign. According to our model, there are three main criteria for determining the extent of nativisation. Fingerspelled elements are more or less nativised along the continuum based on: 1) the extent to which native elements are added, 2) the extent to which non-native elements are reduced (e.g. letters present or lost), and 3) the extent to which native elements are integrated with non-native elements. On the non-native side of the continuum, few native elements are added, but on the native side of the continuum native elements are added (e.g. repeated movement in BSL GARAGE, arcing movement in ASL SOCIETY). Towards the non-native end of the continuum, not much is reduced (e.g. full fingerspellings), but moving towards the native end, more and more elements become reduced (e.g. fully nativised fingerspellings such as ASL #BREAD). Toward the non-native side of the continuum (e.g. sign + fingerspelled compounds like ASL S-T-O-C-K MARKET), elements are not tightly integrated together but are only sequentially combined, while toward the native side of the continuum (e.g. initialised signs like WATER) elements are more simultaneously integrated.
Our model of the non-native lexicon applies to ASL as well as BANZSL. The categories along the continuum are based on clear criteria (addition of native elements, reduction of non-native elements, and sequentiality vs. simultaneity). These specific criteria allow for a greater range of categories to be identified (eight categories here compared to Brentari and Padden’s five, with examples in each category identified for both ASL and BANZSL). Furthermore, these categories highlight some subtle distinctions that have not been recognised in the literature before, such as the difference between the two letter loan THURSDAY (for both ASL and BANZSL consisting of the sequence T-H), and the more nativised two letter loan THURSDAY (which in ASL deletes change in palm orientation and in BANZSL uses a middle instead of index finger for the letter -T-). By taking into account all native parameters, not just handshape, and by taking into consideration the markedness of those parameters independent of their relation to fingerspelled letters, our model more accurately represents the nativisation process and thus more completely represents the phonological structure of the two languages. Moreover, our model reveals that the non-native lexicons in ASL and BANZSL are much more similar than we might otherwise expect given how different their fingerspelling systems look on the surface.