The aim of this paper is to discuss different levels of segmentation considering linguistic analyses of sign languages. These levels of segmentation include the (1) word, sign by sign; (2) the utterance, based on each statement; (3) the syntactic segments focusing on the predicates; (4) the sentence, including subordinates, coordinates, complements, relative clauses; and, (5) translation. Each level of segmentation will be presented considering specific criteria. The segmentation of sign languages focused on in this presentation was proposed based on data from two sign language corpora: Swedish Sign Language Corpus and Brazilian Sign Language Corpus. We analyzed annotations from conversation settings of eight Deaf people, four from each country, each one an interactive setting in pairs. Conversation is a setting that involves more spontaneous production, without previous planning. This kind of setting needs additional criteria for segmentation to be analyzed at different levels of linguistic analysis.
An utterance is a full proposition, and is a segment including formal marks such as intonation and pauses in association with the context in which is produced. A syntactic segment is expressed through a predicate (verbal or nominal). Each predicate is separated in this specific segment. Following Börstell et al. (2016:19), we define a clause (here a syntactic segment) as a unit in which a predicate asserts something about one or more elements (the arguments). The base of the sentence is driven by syntax, while the utterance is driven by meaning. A full proposition can have more than one syntactic segment. In both cases, prosody is taken into account. Prosody includes non-manual markers, pauses, body- or gaze shifting, blinks and head nod (as analyzed for Finnish Sign Language and Swedish Sign Language, in Puupponen et al. 2016).
For syntactic analysis, we can consider multiple syntactic segments for studying different sentence levels of only one syntactic phrase or more, including different scopes of the sentence (such as a verbal phrase, or nominal phrase, an adverbial phrase, an adjectival phrase, a topic phrase, a focus phrase, a complement phrase).
The translation tier is created through utterances in another language (such as Swedish and Portuguese, and into English). We have seen that it might coincide with the utterance in sign language, but not always. This seems to happen because the proposition in each language may be slightly different.
The following examples illustrate the criteria established for both languages:
SSL (SSLC01_246 00:02:18.500-00:02:24.090)One utterance, four syntactic segmentsUtterance: TO DEAF YOUNG POINT.PL YOUNG PRO1 OLDER PU PRO1 MUST TELL POINT.PL KNOW-NOT WHO POINTSyntactic segments: TO DEAF YOUNG POINT.PL YOUNG / PRO1 OLDER PU / PRO1 MUST TELL / POINT.PL KNOW-NOT WHO POINTTranslation: When I, a little older, meet deaf young people, I usually tell them about him, they usually do not know who he is.SSL (SSLC01_246 00:01:12.936-00:01:15.756)One utterance, two syntactic segmentsUtterance: IMPORTANT GET SIGN-LANGUAGE GRAMMAR (facial expression) EFFECTSyntactic segments: IMPORTANT GET SIGN-LANGUAGE GRAMMAR / (facial expression) / EFFECTTranslation: It is important to acquire sign language grammar, it is a wow experience and a good start.Libras (FLN_G1_D1_CONVER_Escolasurdoouvinte 00:00:01:000-00:00:10:000)One utterance, three syntactic segmentsUtterance: SCHOOL INCLUSION HARD BECAUSE THERE-IS-NO THINKING KNOW DEAF CULTURE RIGHT?Syntactic segments: SCHOOL INCLUSION HARD / BECAUSE THERE-IS-NO THINKING KNOW DEAF CULTURE / RIGHT?Translation: The inclusive school finds some difficulty, because there is no knowledge of deaf culture, isn’t it?Libras (FLN_G3_D6_CONVER_EscolasurdoouvinteOne utterance, three syntactic segmentsUtterance: POINT.PL STUDENTS HEARING TALK PRO1 DEAF DV(stay-static) HELP NOTHINGSyntactic segment: POINT.PL STUDENTS HEARING TALK / PRO1 DEAF DV(stay-static) / HELP NOTHINGTranslation: The hearing students talk to each other, while I, a deaf child, stay still observing without help (to communicate) from the others.
The purpose of establishing the same criteria for segmentation is to make possible contrastive and comparative studies among sign languages.
2019.
TISLR13, the 13th conference of Theoretical Issues in Sign Language Research, Hamburg, Germany, 26-28 September, 2019