[BibTeX] [RIS]
Achieving reference via contrast in route instructions and spatial object identification
Type of publication: Inproceedings
Citation: TenbrinkKlippel05
Booktitle: Workshop on Reference, 21st Scandinavian Conference of Linguistics, June 1-4, 2005
Year: 2005
Abstract: Investigating the semantics of spatial projective terms such as left, in front of, or above requires accounting for interpretation processes that either focus on the semantic core of the expression or are due to contextual or pragmatic aspects. Herskovits (1986) proposes to capture the ideal meaning of spatial terms by describing the geometric relationships underlying their semantics, and to derive a range of different use types in which the ideal meaning is stretched and adapted according to context. In her account, the semantics of spatial terms involves a certain gradedness of interpretation. Basically, it means that projective terms can be applied best near a focal axis; for example, a good front position is one directly located on the straight line of sight of an observer. Positions further away from that axis are preferentially described using linguistic modifications or combinations of the terms. There are, however, good reasons for rejecting the notion that the gradedness of applicability is part of the terms' core semantics. In contrast, the effects of the spatial relation to the focal axis seem to depend on the discourse task. We differentiate three basic kinds of tasks in which the application of projective terms have been investigated: first, they are used to describe the location of one known object relative to another entity, for example, as an answer to the question Where is the object?. Second, they are used in route instructions where they indicate directions of movement (such as turn right) or indicate the position of landmarks along the way (such as there's a church on the right). Third, they are used to identify an object out of a number of competing objects, for example, as an answer to the question Which object is it? (answer: the one on the right). Generally, it seems that the gradedness of applicability most prominently comes into play in the first of these three discourse tasks (which is the one most often used for the investigation of projective terms), while in the other tasks different kinds of discourse processes seem to be more decisive. In route instructions, for example, people typically indicate the direction of movement in qualitative and simple terms, using an expression like turn left for a range of angles that a turn might take (e.g., Tversky & Lee 1999). Likewise, the positions of landmarks are typically not described in a very precise way; it is mostly sufficient to identify their existence near the route. Similarly, in spatial object identification tasks people may use an expression like the one on the left as long as the object referred to is positioned somewhere on the left half plane. Crucially, in both kinds of tasks the presence of competing objects or directions (i.e., other candidates for goal object or path direction) is decisive for the linguistic choices speakers make. For example, if there are two streets to the right, an unmodified expression like turn right will not be sufficient even if the intended change of direction corresponds to an exact 90° angle. In other words, speakers use the projective expressions in a way that is maximally contrastive with respect to the current context, enabling the listener to identify the intended object or direction, largely independent of the gradedness of applicability associated with other kinds of discourse tasks. We present evidence for this principle from empirical investigations in these two kinds of tasks. Since all linguistic expressions are intricately connected to the nonlinguistic concepts and real-world contexts associated with them, a term's semantics cannot be separated completely from its applicability in any case, so that it may be a matter of theoretic stance whether effects such as gradedness of interpretation are associated with the semantics or the pragmatics of a projective term. Our proposal is to assume a semantic core that is flexible enough to be compatible with a wide range of application contexts, which can then be specified systematically in order to gain more detailed information with regard to applicability and interpretation. References Herskovits, A. 1986. Language and spatial cognition. Cambridge University Press. Tversky, Barbara and Paul Lee. 1999. Pictorial and Verbal Tools for Conveying Routes. In C. Freksa & D.M. Mark (eds.), Spatial information theory- Cognitive and computational foundations of geographic information science (COSIT '99). Berlin: Springer.
Userfields: project={I1-OntoSpace}, status={Other},
Keywords:
Authors Tenbrink, Thora
Klippel, Alexander
Attachments
    Notes
      Topics