Data annotation. We used ANVIL to code low-level behavior of part of the dataset according to a low-level Behavior repertoire. The repertoire data was incorporated in the first analysis tool of diver behavior – the DiverControlCenter. As a higher level of annotation proved more useful for behavior interpretation and segmentation, we developed a real-time annotation tool based on the diver tasks. We collected time-synchronized video material for all data collections for comparison with DiverNet data. The DiverNet motion recording is consistent with the motion displayed on the video material for all 36 cases mentioned above. DiverNet data and Video data was segmented according to these real-time annotations. The behavior segments were uploaded to the CADDY server. The DiverControlCenter can segment and analyze all DiverNet data. Task list: breathing with regulator breathing without regulator taking off the mask taking off the regulator controlling buoyancy moving up and down between 2 targets moving in the horizontal plane at a slow pace swimming quickly in a horizontal plane displacing an object free behaviour task T-posture The low-level behaviour repertoire for all diver limbs and breathing activity are shown in the next tables. Behaviour Definition hand signal hand performs defined diving signal manipulating own diving equipment hand touches, grabs, moves or picks up object related to the diver’s equipment manipulating other diver’s equipment hand touches, grabs, moves or picks up object related to another diver’s equipment touching object hand touches object (except equipment) picking up object hand grabs and picks up object (except equipment) carrying object upper limbs carry object touching person hand touches another person’s body touching own body hand touches part of the own body touching floor hand touches floor hanging on to something hand hangs on to object, floor or person still upper limbs still without contact with equipment, object, person or own body parts moving upper limbs moving without contact with equipment, object, person or own body parts none
Data annotation. It is a component for manual annotation and labelling of the data,
Data annotation. Each clause was viewed as a potential locus of agreement, e.g. (2) has two loci. Each clause was annotated for 10 parameters: Presence/absence of the plural morpheme on the predicate (dependent variable) Impersonal interpretation: YES / NO Word order: SV, SXV, VS, VXS Animacy of the subject: human / animate / inanimate Referentiality of the subject: definite / specific (referential) indefinite / non-referential Givenness of the subject: given (mentioned previously in discourse) / new Syntactic structure of the subject: plural NP, conjoined NPs, comitative NP, zero, etc. Presence of the following same-subject finite clause within the same sentence: YES / NO Semantics of the predicate: joint / distributive action (cf. [Xxxxxx 1993]) contrary to some previous claims [Xxxxxxxx 2008: 206; Xxxxx 1964: 91–93] this factor appeared to be irrelevant in our data; it is not discussed below. Syntactic structure of the predicate: various verb forms, non-verbal predicates. Why these parameters? occasional claims with respect to their relevance for agreement in Xxxxxxx, typically without much empirical evidence [Xxxxxxxx 2008; Xxxxxxx 1981; Xxxxx 1964]; studies on optional number agreement in other languages [Xxxxxxx 2000, 2006; Creissels 2010; Durie 1986; Iemmolo (in press); Пекелис 2013]; our hypotheses that emerged when collecting the data.
