Download Sign Language: Cross-Lingual View - Pronouns, Verb Agreement, and Spatial Representation and more Study Guides, Projects, Research Sign Language in PDF only on Docsity!
P 108 Sign language: a cross-linguistic perspective AG 7
Kearsy Cormier Exclusive pronouns in ASL and BSL Donnerstag/Thursday: 10:
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.
AG 7 Sign language: a cross-linguistic perspective P 108
Barbara Hänel The acquisition of verb agreement in signed languages (DGS/ASL) Alternate
I will base my talk on two ideas: First, according to linguistic theory, child grammar should be constrained by the same syntactic rules and principles as has been suggested for adult grammar (Pinker 1984). Second, parameters specify the possible choices which languages can make. They usually relate to functional heads (Chomsky 1991, Ouhalla 1991). While Universal Grammar provides the possible sets of functional categories and features, children have to determine which one might be used for actual language. In this paper I will argue that the development of morphological agreement markers is related to the realisation of the functional category INFL or an equivalent functional element. Following Meir (1998) and Aronoff et al. (2000) I assume that verb agreement in signed languages is expressed by copying the locative index of the structurally relevant noun phrase onto the agreement morpheme. Agreement verbs are underspecified with respect to locus features (beginning and / or end point of the movement), thus making empty slots available for the copied locus features. Copying is available if the two elements are in an specific structural relationship, usually a specifier-head relation. The prediction is that at the same time when children use inflection markings productive the feature sharing process should have been acquired, namely the [R-locus] feature being copied to the open slots of the verb. The study will be based on an analysis of two deaf children growing up with DGS as a first language in their deaf families. The results will be compared to earlier studies of children acquiring ASL as its first language (Meier 1982, Loew 1984, Lillo-Martin 1991). It will be shown that the DGS children as well as the ASL children acquire the underlying agreement mechanism during the same time span. Despite this crucial developmental parallel, one difference shows up. Whereas DGS learning children start using verb agreement with non- present referents, children learning ASL cease to produce this from during the period in question. I will tentatively put forward the idea that this is a performance rather than a competence effect.
Daisuke Hara Maximum syllable complexity constraint in ASL and JSL Mittwoch/Wednesday: 18:
A sign language syllable is composed of elements in at least three basic components, i.e., the handshape, movement, and location components. Each of the components has a finite number of elements. The previous models impose restrictions on each component, and succeed in licensing only the elements that are actually used and in excluding the ones that do not occur in syllables. Remarkably, many of the restrictions so far proposed have been argued in terms of individual components only, i.e., intracomponentially. There is no doubt that intracomponential restrictions are indispensable to argue well-formedness of syllables. But, it is also true that they alone are not sufficient to discuss well-formedness of sign language syllables. We also need a mechanism that controls relationships among the components, i.e., intercomponential relationships, since not all the combinations of the legitimate
AG 7 Sign language: a cross-linguistic perspective P 108
relationship between culture-specific gestures, and the way these may or may not be grammaticalized in sign languages. Jordanian Sign Language (or Lughat il-Ishaarah il-Urduniah, LIU) appears to be one dialect of a larger Eastern Arabic Sign Language, that also comprises at least Lebanese, Syrian, Iraqi, and Palestinian Sign Language. Although Arab cultures tend to use the backward tilt of the head, often accompanied by a tongue-click and a raise of the eyebrows, the grammatical status of this gesture in LIU is questionable. Although its form seems to be reduced compared to the gesture used by the hearing population, the distribution of the gesture as used by hearing and Deaf Arabs seems to be very similar. That is, it is not generally used as a clause negator, but as a “negative interjection”, occurring on its own as a one-word utterance, although it may also accompany manual negator signs. Unlike the headshake in many sign languages, however, it does not appear to be able to occur as a clause negator by itself. LIU has several manual negator signs, the most common of which are an existential negator MA-FI (which in Arabic means ‘there is no’, but has a much broader meaning in sign language) and a side-to-side “handshake”, using a 1-hand. These two signs can often be used interchangeably, but there is a subtle difference, in that MA-FI can apparently not be used for advice or warning, whereas the “handshake” can. This difference in distribution may be tense- related. Tense-related negation has been described for Israeli Sign Language (ISL), a language which appears to be unrelated to LIU, but is geographically very close. Apart from these two signs there are negative signs that have a specific defensive, warning, apologetic or emphatic function. LIU also appears to have a negative post-clitic, which is very similar to a negative derivational suffix in ISL, and can be added to some verbs and adjectives. This sign can also occur as a separate word and may occur with nouns as well as adjectives and verbs. When it occurs as a separate word, however, it does not affect the place of articulation or movement of the preceding sign in the way it does when it is a clitic. The clitic occurs with a strong negative facial expression, which has been observed to negate a sentence without an accompanying manual sign or head-movement, although this seems rare. Although more research needs to be done into the precise use of different negators in LIU, and into the use and domain of negative facial expression, it appears that LIU is typologically interesting because it does not use headshake as the primary means of negating clauses, although a very subtle headshake or head turn may accompany a manual sign. It has a variety of manual negators that all seem to have subtly different meanings, which may in one case interact with tense. It also has a negative clitic which is very similar to the negative suffix described for ISL, which appears to be an unrelated sign language.
Annette Hohenberger Cross-linguistic variation between sign languages: UG, modality and typological aspects Donnerstag/Thursday: 10:
This paper is concerned with delineating the possible range of cross-linguistic variation between sign languages. The range of variation is determined by three factors: UG, modality, and typology. As for UG, there are general constraints on the possible format of natural languages which hold for any language, despite modality and typology. Such constraints include, among others, universal syntactic principles (e.g., Full Interpretation) as well as design properties of the language system (having, e.g., hierarchical, recursive, asymmetric representations). They are a part of the general conception of how the relation between grammar and its interfaces (Articulatory/Perceptional (AP) and Conceptual/Intentional (CI)) is to be construed. The
P 108 Sign language: a cross-linguistic perspective AG 7
version of the conceptualization of the human language faculty adopted here is the Minimalist Program (Chomsky, 1995). As for modality, I will investigate in how far general modality effects pertaining to all sign languages restrict possible cross-linguistic variation. This part of the analysis includes the discussion of modality differences between signed and spoken languages. Generally, no modality differences are to be expected between signed languages and spoken languages on the level of Logical Form (LF), given that the principles of forming concepts or drawing on them is universally invariant. The means for expressing universal concepts may be modality- dependent, though. In particular, I will focus on differences with respect to language processing. I will rely on Brentari's (2002) distinction of "horizontal processing" (for spoken languages) and "vertical processing" (for sign languages). Brentari claims that "sign languages constitute a typological class unto themselves" (p 57) which she characterizes with respect to canonical wordshape, namely as "polymorphemic/monosyllabic". This view corresponds to the "traditional view" of emphasizing the similarities between signed languages as opposed to spoken languages. I will adduce comparative evidence from language production, namely slips of the hand in German sign language (Deutsche Gebärdensprache, DGS) and slips of the tongue in spoken German in order to point out the effect of modality (Hohenberger, Happ, & Leuninger, 2002). Cross-linguistic variation is to be expected to reside within the field of phonological, morphological, and syntactic typology. Within grammar, sign languages do in fact have at their disposal the whole range of variation of establishing a formal and arbitrary connection between meaning (the CI-system) and form (the AP-system) from which they can choose. Morphology "as the characteristic domain for idiosyncratic aspects in linguistic structure" (Bierwisch, 2001) is the most promising domain for cross-linguistic variation. It comes as no surprise that many cross-linguistic differences between sign languages investigated so far are morpho-phonological and morpho-syntactic in nature. Insofar comparative production data will be accessible to me by the time of the conference, I will adduce evidence for cross- linguistic differences in slips of the hand in different sign languages, most notably between ASL and DGS. I expect differences to exist in phonology (e.g., in the restriction of the use of the non-dominant hand), morphology (e.g., classifiers, inflection), and position of functional categories, inevitably showing up in language production errors in different sign languages.
Wiebke Iversen & Jill Morford Counting on your hands: The language-dependent nature of number processing Donnerstag/Thursday: 11:
Cross-linguistic differences can influence not only language processing, but also other cognitive processes that are mediated by language. We report the results of a study investigating the effects of cross-linguistic differences on number processing. Specifically, we investigated the effect of form differences in the lexicons of German Sign Language (DGS) and American Sign Language (ASL) on parity (odd-even) judgements. Cross-linguistic differences in parity judgements have already been documented in speakers of German and signers of German Sign Language. Iversen and colleagues (2003) found that speakers of German identify odd numbers more quickly when responding with the left hand and even numbers more quickly when responding with the right hand for both Arabic numerals and number words (in German, e.g., “eins”, “zwei”, etc.). In contrast, signers of German Sign Language identified odd numbers more quickly with the right hand, and even numbers more quickly with the left hand for Arabic numerals. These divergent results have
P 108 Sign language: a cross-linguistic perspective AG 7
extended from a fist and pointing upward. It signifies a significant discourse theme. There does not appear to be a THEME buoy in Swedish Sign Language (SSL). Signers create fragment buoys by pointing at the final configuration of a just produced sign. This gives the sign fragment the ability to remain in place with the significance of that just produced sign. The POINTER buoy is produced with the same handshape as the THEME buoy, but oriented so that the finger points toward a significant entity being discussed while the strong hand continues signing. What all these examples have in common is the maintenance of a meaningful buoy on one hand while the other hand produces connected discourse. We will focus our discussions on the use of buoys in American Sign Language, Norwegian Sign Language, and Swedish Sign Language. We will, in turn, describe the use of buoys in the sign language each of us studies. We observe pointer, fragment, and list buoys in all three sign languages. While Norwegian Sign Language and Swedish Sign Language have static list buoys like those commonly used in American Sign Language, they also make frequent use of sequentially built lists. In constructing these sequentially built lists, the focus appears to shift to the extent of the list rather than on associating specific entities with fingers. Our general approach will be to contrast the apparent uniformity of the categories of spatial representations used across sign languages with language specific differences in how the use of space is manifested within the category buoys.
Gaurav Mathur & Christian Rathmann Cross-sign-linguistic variation in the frequency of verb agreement forms Mittwoch/Wednesday: 15:
Introduction. All signed languages documented to date clearly exhibit verb agreement. What is not clear, however, is the extent of cross-linguistic variation across signed languages with respect to the form of verb agreement. In re-examining the distribution of agreeing verbs previously elicited from four signed languages (American Sign Language - ASL, German Sign Language - DGS, Australian Sign Language - Auslan, and Japanese Sign Language - Nihon Shuwa), we ask the following question: are certain forms of modulation and types of handedness more frequent than others, and if so, how much is this subject to cross-linguistic variation? Form of modulation. With regard to form of modulation, we find three patterns. First, for all signed languages, a change in both orientation and direction of movement (= Form OrDir) is more frequent than a change in just one of them. Whether a change in orientation (= Form Or) only is more frequent than a change in direction of movement (= Form Dir) only varies cross- linguistically. The second finding is that Form OrDir is more frequent than a similar change which involves an additional change in the relative position of the hands with respect to the body (= Form OrDirHnd). Whether this change is more frequent than another change in which involves just the orientation and the relative position of the hands (= Form OrHnd) depends on the signed language. The last pattern is Form Or and Form Dir are more frequent than Forms OrDirHnd and OrHnd, with the exception of ASL, where Form Dir is the least frequent. The high frequency of Form OrDir can be explained by the fact that it requires signs that are easy to manipulate from an articulatory point of view. Forms Or and Dir appear less frequently, because signs are not as manipulable so that they change only in orientation or direction of movement. Forms OrDirHnd and OrHnd are even less frequent, since they are more complex in that they also change the relative position of the hands with respect to the body.
AG 7 Sign language: a cross-linguistic perspective P 108
Type of handedness. Within Battison’s (1978) typology of handedness, one finding is that two-handed, non-symmetric signs with the hands having different handshapes (= Type 3) appear most frequently in all signed languages (in Auslan, one-handed signs (= Type 0) are equally frequent as well). The other finding is that two-handed, non-symmetric signs with the hands having the same handshape (= Type 2) appear least frequently in all signed languages. It is the relative frequency of one-handed signs and two-handed, symmetric signs (= Type 1) that is subject to cross-linguistic variation. It is not surprising that Types 3 and 0 are the most frequent, since they are the most prototypical types of sign, i.e. two-handed asymmetrical and one-handed signs, respectively. Type 1 may be more frequent than Type 0 in Nihon Shuwa, because Nihon Shuwa has a high number of agreeing verbs that are two-handed. Finally, Type 2 seems to be relatively rare owing to the fact that when the two handshapes are the same, there is a strong preference for the sign to be symmetrical (i.e. Type 1), as predicted by Battison’s (1978) Symmetry Condition. Conclusion. The findings above suggest that motor economy can predict which of the forms are most frequent and therefore appear universally across all signed languages, whereas it is in the less frequent forms that we see cross-linguistic variation (see Ann 1996 for a similar conclusion regarding handshape in lexical signs in Taiwanese SL and ASL). Frequency in forms of modulation, as well as frequency in types of handedness, in an apparently universal phenomenon like verb agreement is thus a tool for investigating the range of cross-linguistic variation across signed languages at the phonological-phonetic level.
Marie Nadolske & Rachel Rosenstock A preliminary cross-linguistic study of word pictures in ASL Mittwoch/Wednesday: 17:
Non-manual aspects such as eye-gaze, body shift, eyebrow movement, etc. have been widely recognized as a part of the grammar of natural sign languages (Baker-Shenk 1983, Valli & Lucas 1992). One component that is still highly controversial is the use of the mouth to form what resembles spoken words (Word Pictures). While the general assumption by both Deaf and hearing researchers seems to be that the phenomenon of Word Pictures (WP) does not exist in ASL (see Hamilton 1983 for discussion), the phenomenon has been observed and described by various European researchers for sign languages in e.g. Norway (Vogt-Svendsen 1983), Germany (Ebbinghaus & Hessmann 1994) or the Netherlands (Schermer 1990). In the debates of the function of WP and their possible role as a part of sign language grammar, the obligatory nature of WP remains a point of discussion and speculation. The present study attempts to demonstrate the occurrence of WP in American Sign Language. A quantitative method will be applied to measure the frequency with which WP occur simultaneously to a specific word class. Additionally, various factors of variation (situation, onset of learning ASL, age) were taken into consideration. The relative frequency of occurrence can indicate the status of WP as either code-switching, nonce-borrowing, or borrowing. The data for this study consists of 80 minutes of various videotaped ASL signing. All data was glossed and the single signs categorized into word classes. Demographic characteristics and background information were noted for each individual signer. Each item produced by the signer was then classified into three categories of mouthing: (1) WP; (2) other mouthing; (3) no mouthing. More than 6000 items were included in the study.
AG 7 Sign language: a cross-linguistic perspective P 108
In addition to the temporal domain of the mouth marker, we examine the functional domain of its meaning, distinguishing whether it adds to the meaning of the manual sign itself or to the meaning of a whole phrase or sentence it is associated with. Per language we categorize these meaning patterns, for instance whether the mouth marker adds adjectival, aspectual or adverbial (or idiosyncratic) meaning. Results are presented in terms of quantitative similarities and differences between the behaviours of the mouth in the three languages. Furthermore, the time precise alignment of the transcriptions that are embedded in ELAN enables us to very precisely investigate the relation between the rhythm of the manual movement and that of the mouth movement. The idea presented in the mouthing literature is that the mouth and the hands move in tandem, and that the mouth follows the rhythmic pattern and time span of the hands (Woll 2001). This rhythmic relation proposed for lexical signs is investigated in the broader context of sign sentences in the three languages.
Asli Özyürek & Deniz Ilkbasaran The use of space in transfer verbs in Turkish Sign Language (TID): Implications for typological variation Mittwoch/Wednesday: 15:
Turkish Sign Language (Türk Isaret Dili (TID)) is still an undocumented language with 2, million users within Turkey. Here we report on one of the first investigations on TID grammar, specifically on grammatical constructions involving transfer verbs (e.g., GIVE and TAKE). The data was collected from 4 native signers of TID from Istanbul. The analysis in this talk is restricted to the grammatical constructions of GIVE and TAKE verbs used in the following contexts: a) spontaneous stories, b) picture book narrations (Balloon Story used in Lillo-Martin, 1991), c) single sentences elicited in response to simple GIVE and TAKE events. All the sentences in which GIVE and TAKE verbs used for non-present referents were coded for the number of arguments expressed, the word order, the direction of the verb signs and the type of person marking in expressing arguments (i.e., lexical versus location in space). The analysis shows that, first of all, Turkish signers most often prefer to use SOV order in sentences where the subjects and the objects are lexicalized. Furthermore, the nouns or the pronominal forms are rarely located in space before the use of the directional verbs. Subject nouns are either lexicalized or they are mapped onto the signer without explicit pointing toward the signer. Similarly the object nouns are not localized before the use of the directional GIVE and TAKE signs. Furthermore, we find the verb directions to be half of the time directed to the sides (right or left) and half of the time directed in front of the signer on a sagittal axis. Finally, for many of the cases the GIVE or TAKE events were found be depicted with two verbs per event as in:
(1) MAN GIVE-BALLOON (directed from signer to his right) BOY TAKE-BALLOON (directed from right towards the signer) (depicting an event where the balloon man gives a balloon to a boy).
These findings indicate that Turkish signers use space in different ways for transfer verbs than has been described in the literature for ASL, NGL, ISL etc. (e.g., Lillo-Martin & Klima, 1990; Meir, 2002 etc.). They suggest that Turkish signers do not use spatial agreement morphology in a consistent way according to the definition of verb agreement proposed for sign languages as in “ a verb agrees with its arguments if its initial and final locations are determined with the R-loci of its arguments” (Meir, 2002). Rather, TID signers might be relying on regular word
P 108 Sign language: a cross-linguistic perspective AG 7
order for the specification of argument roles for these event types. These results are suggestive of typological variation in the use of these verbs.
Pamela Perniss & Asli Özyürek “How many apples in the bowl?” – number and quantification in German (DGS) and Turkish Sign Language (TID) Donnerstag/Thursday: 12:
This paper explores the domain of quantification in two sign languages – German Sign Language (DGS) and Turkish Sign Language (TID). The domain of quantification encompasses all expressions that give an indication of quantity. Thus, we are concerned with morphological categories of number, the use of numerals and quantifiers, and other lexical or morphological means of expressing the size, extent, or configuration of a certain quantity of entities. The visual-gestural modality makes the investigation of the expression of number and quantification in signed languages especially interesting for three reasons, in particular.
(1) It is difficult to find morphological criteria that straightforwardly distinguish word classes in signed languages. Many signs appear in different grammatical functions without undergoing any derivational changes of form. Furthermore, articulatory constraints may block morphological processes in the domain of quantification, e.g. for indicating plural (cf. Zeshan 1999). (2) The shape or configuration of entities constituting a plurality can be iconically represented in sign space. The topographic use of space makes it possible to create representations in space that correspond schematically or isomorphically to real-world quantity configurations. (3) Communication in signed languages is necessarily face-to-face, and is thus always highly context-dependent. This context-dependency has considerable influence on the interpretation of signs with respect to number.
The effects that these features of the modality have on the means of quantifying entities and events are discussed based on the presentation of similarities and differences between German (DGS) and Turkish Sign Language (TID). The conditions under which singularity and plurality are expressed and the means of expression are contrasted in these two signed languages. For both languages, an underlying transnumerality of signs is assumed (cf. Perniss (2001) for DGS). This means that the difference between one and many denoted objects is not obligatorily marked as a difference in the morphological form of lexical signs. In both DGS and TID, the marking of plurality is an optional morphological process that is realized under certain conditions. The morphosyntactic means of marking plurality include the use of numerals, the use of quantifiers, the repetition of a sign, the use of indexes, the use of size and shape denoting handshapes or tracing movements, as well as movement patterns like a horizontal sweep of the hand through sign space. These means of marking plurality are used in both DGS and in TID, but important differences exist in word order, spatial modification of signs, frequency of occurrence, and the realization of specific forms. There are differences, for example, in the ways in which adjectives and nouns can be spatially modified according to the form of an accompanying quantifier in DGS and TID. With respect to word order, numerals and
P 108 Sign language: a cross-linguistic perspective AG 7
An analysis with further movement of the negative modal to Cº will be defended. Interaction with aspectual adverbs will be shown to provide extra support for the claim. This extra step is arguably triggered by a [+focus] feature residing in the complementizer position of these structures.
Ronice M. de Quadros, Diane Lillo-Martin & Deborah Chen Pichler Clause structure in LSB and ASL Freitag/Friday: 11:
In this paper we discuss syntactic similarities and differences between Brazilian Sign language (LSB) and American Sign Language (ASL). We aim to account for both the similarities and differences with a theory employing the minimum technical machinery possible (Chomsky, 1995). LSB and ASL have both verbs which can be inflected to indicate the subject and object (‘agreeing’ verbs) and verbs which cannot (‘plain’ verbs). They also have ‘spatial’ verbs, which like agreeing verbs are inflected by being signed in a referential location, but the inflection indicates a physical location rather than a referent. In addition, both allow some plain verbs to be signed in a particular location. We examine the ways in which plain verbs signed in a location behave like spatial verbs – they convey information about physical location, and they are distributed syntactically like spatial verbs. We follow Quadros’(1999) proposal that LSB agreeing verbs are taken from lexicon fully inflected, while plain verbs are inserted into the syntactic structure ‘bare’ and must merge with an adjacent virtual affix. This morphological distinction leads to an important syntactic effect: the two verb types project different syntactic structures. On this analysis, plain verbs project only IP, while agreeing verbs project TP and AgrP. This difference accounts, among other things, for the different patterns with negation. Spatial verbs, and plain verbs signed in a particular spatial location, behave like agreeing verbs in LSB with respect to negation, and also with respect to word order and null arguments. This indicates that they project the same structure in this language. Despite the morphological similarity of LSB and ASL, the verb types behave syntactically differently across the two languages. In particular, the differences in LSB negation placement are not found between agreeing and plain verbs in ASL. We conclude from this that the verb types project the same structure in ASL. Theory-internal arguments lead to the postulation that this is the structure which includes both TP and AgrP. While LSB and ASL behave differently with respect to basic clause structure, they are very similar in their use of doubling constructions, described in detail by Petronio (1993) and Quadros (1999). We find that both LSB and ASL have strict restrictions on focus doubling in contexts where more than one potentially focused element is available. For example, in WH- question contexts, only the WH-element may be doubled, not a modal or verb (as originally observed by Petronio, 1993). We propose an account for this distribution which reduces it to a type of Relativized Minimality (Rizzi, 1990). In particular, on our account only the structurally highest eligible element can be involved in focus doubling. These results are expected under the hypothesis that languages have deep similarities with systematic lexical differences affecting a range of structures.
AG 7 Sign language: a cross-linguistic perspective P 108
Ninoslava Šarac, Katharina Schalber, Tamara Alibašic & Ronnie B. Wilbur Crosslinguistic comparison of sign language interrogatives Freitag/Friday: 13:
Our research compares interrogatives in Croatian (HZJ), Austrian (ÖGS) and American (ASL) sign languages. Addressed are (1) word order and the position of interrogative signs, and (2) nonmanual markers and their scope. (1) Word order and position of interrogative signs : HZJ appears to be SVO, whereas ÖGS is SOV (Šarac 2003; Schalber et al. 1999; Wilbur 2002, in press). Although unrelated, HZJ looks like ASL with head-final CP, and other phrases head-initial. In contrast, ÖGS has head- initial CP, and head-final IP and VP. We are intrigued by this result because there was substantial contact between ÖGS and HZJ, which is not reflected in the syntax so far. For y/n questions : (a) We are not aware of any reports of SLs that require inversion of a verb form (auxiliary/modal/verb) and a subject. However, under certain limited circumstances (doubled modals), such inversion is tolerated in ÖGS. (b) HZJ has an interrogative sign JE - LI ‘is-it’; we argue that it is an optional left adjunct (from signed Croatian) and not part of the interrogative structure itself (cf. QMwg; Neidle et al 2000). For wh-questions : We are not aware of any SLs that do not permit wh-fronting nor any that require it (i.e. prohibit in situ wh-signs). In HZJ the wh-signs can be initial, final or doubled, with optional copying on the right either to C (no pause) or in Tag (following a pause) (Alibašiü 2003; Šarac 2003); that is, it behaves like ASL. In contrast, ÖGS has C on the left, so there is no place for standard copying; a doubled wh-sign can only occur in TagQP (after a required pause). (2) Nonmanuals : There appears to be a typological split with respect to the main nonmanual used to differentiate yes/no from wh-questions. From the literature and our own work, three SLs - ASL, BSL, and Swedish SL - use brow position: raise for yes/no, furrow for wh-q. Three SLs - HZJ, ÖGS, and LSQ - use head/chin position/movement instead: down(ward) for y/n, up(ward) for wh-q. (There may be other nonmanuals that serve this function in other SLs
- we hope to discuss this with conference attendees.) Note that unlike the syntactic descriptions above, where ASL and HZJ seem to pattern together, with respect to the nonmanuals, HZJ patterns with ÖGS. Consistency of marking : The markings on y/n seem to be much more regular than those for wh-q, which could be related to the presence of wh-signs that might make the nonmanuals less crucial to wh-q interpretation. ÖGS and HZJ appear to differ in the degree of obligatoriness of nonmanual marking in wh-q. In HZJ, the nonmanual marking is crucial and consistent, whereas in ÖGS, wh-q can be found marked with upward or downward tilt, as well as forward thrust of the head. This result could be due to an interaction of the nonmanual similarities (use of head) and the syntactic differences (SVO, SOV). Intensity : Like ASL, in HZJ y/n questions, the intensity of the nonmanuals is highest at the end of the clause. Unlike ASL, in HZJ wh-q the highest intensity is initial (that is, with the first occurrence of [+wh]). In HZJ y/n questions, the sign (^) JE - LI can be initial (or final if in Tag). JE - LI is not connected to the question operator or nonmanual intensity; hence we accord it left IP-adjunct status. This analysis has implications for the treatment of ASL IF / SUPPOSE (as a potential counterexample to the analysis of head-final CP). These data suggest: (1) typological differences in syntactic phrase structure, but that e.g. all SVOs are not alike; (2) typological differences in choice of nonmanual articulator for marking interrogatives (and other operators); and (3) a hypothesis that nonmanuals may be more easily
AG 7 Sign language: a cross-linguistic perspective P 108
Myriam Vermeerbergen, Lorraine Leeson, Adam Schembri & Trevor Johnston Considering cross-linguistic analysis of constituent ordering in sign languages: A presentation in three parts Mittwoch/Wednesday: 14:
This presentation explores some of the salient issues relating to the cross-linguistic analyses of constituent ordering in sign languages. There are three main themes which we address: Part 1 considers the existing literature on constituent order in sign languages; Part 2 focuses on problems encountered in compiling the results of this body of data for cross-linguistic comparison, and Part 3 outlines the process of developing a small comparative analysis of declarative utterances in three sign languages using a widely applied elicitation task and a shared set of guidelines regarding interpretation of the data. Part 1: Part 1 deals with consideration of the literature regarding constituent order in sign languages, focusing specifically on data collated using the Volterra et al. (1984) picture elicitation task, which discusses three types of declarative sentences (locative, reversible, and non-reversible). This material has been applied to a wide range of unrelated sign languages which leads one to expect that a broad range of data could be relatively easily compared cross-linguistically. However, there is a range of difficulties involved in comparing and contrasting the data. Part 2: Here we problematise the issue of cross-linguistic comparisons which arise for a range of reasons. These include for example, the fact that the linguistic framework for analyses may differ significantly from study to study; the fact that the definition of basic notions, such as subject and object, are either not defined at all or are interpreted in different ways; certain structures are treated as special cases or marginal cases and excluded from any core discussion of results (e.g. polymorphemic verbs, simultaneous constructions); the assignment of semantic roles/ macro-roles is applied inconsistently across the literature, and more basic issues such as defining the scope of the sentence/ clause often remain undiscussed. All these issues conspire to make valid comparison of the available body of data on constituent order difficult. Given these issues, we designed a small three-way cross-linguistic “test-case” where the criteria for data collection, and most specifically, for the analysis and interpretation of the data were clearly defined at the outset. Part 3: In this section we outline the results of our collaborative “test-case” which focused on three sign languages, namely Flemish Sign Language (VGT), Irish Sign Language (ISL) and Australian Sign Language (Auslan). The research was driven by the Volterra et al. picture elicitation task data. Four informants contributed to the data in each of the three sign languages (i.e., a total of 12 informants). To minimize sociolinguistic complexities associated with generational variation, we controlled for age, including only signers in the age group 30- 40 years. We also considered gender as a potential influencing factor and thus this test-case includes only male signers. Native competence was also controlled. The presenters suggest that this analysis is unique insofar as it represents a collaborative attempt to apply the same notions to constituents and to the interpretation of data, allowing for a valid cross-linguistic comparison.
