pubmed.ncbi.nlm.nih.gov

Language processing in the natural world - PubMed

  • ️Tue Jan 01 2008

Review

Language processing in the natural world

Michael K Tanenhaus et al. Philos Trans R Soc Lond B Biol Sci. 2008.

Abstract

The authors argue that a more complete understanding of how people produce and comprehend language will require investigating real-time spoken-language processing in natural tasks, including those that require goal-oriented unscripted conversation. One promising methodology for such studies is monitoring eye movements as speakers and listeners perform natural tasks. Three lines of research that adopt this approach are reviewed: (i) spoken word recognition in continuous speech, (ii) reference resolution in real-world contexts, and (iii) real-time language processing in interactive conversation. In each domain, results emerge that provide insights which would otherwise be difficult to obtain. These results extend and, in some cases, challenge standard assumptions about language processing.

PubMed Disclaimer

Figures

Figure 1
Figure 1

Shown are a sample display, simulations and data from Allopenna et al. (1998). (a) Sample display, (b) simulations of fixation proportions using TRACE and the linking hypothesis, and (c) the behavioural data. All figures are adapted from Allopenna et al. (1998).

Figure 2
Figure 2

Sample display and proportion of looks to (a) the cohort competitor (picture of a cat) and (b) embedded carrier word competitor (picture of a captain) in utterance medial and utterance final positions.

Figure 3
Figure 3

(a) Sample stimuli. The small cube will fit into both cans but the large cube will only fit into the big can. (b) The mean latency to launch an eye movement to the goal with definite and indefinite instructions and one and more than one compatible goal referents.

Figure 4
Figure 4

(a) Sample stimuli for (i) one-referent (pencil) and (ii) two-referent (apple on napkin) conditions. (b) The proportion of looks to the competitor goal (the towel) for instructions with locally ambiguous and unambiguous prepositional phrases in one-referent and two-referent contexts.

Figure 5
Figure 5

Sample stimuli for trials with (a) action-compatible competitor (two liquid eggs) and (b) action-incompatible competitor (one solid egg). (c) The mean proportion of time spent looking at the competitor goal (the empty bowl) for instructions with locally ambiguous and unambiguous prepositional phrases with action-compatible and action-incompatible competitors.

Figure 6
Figure 6

(a) Sample stimuli. Both whistles can be moved by hand, but only the whistle with the string attached can be picked up with a hook. (b) The proportion of time spent looking at the competitor goal when the presence or absence of an instrument makes the competitor action-compatible or action-incompatible.

Figure 7
Figure 7

Schematic of the setup used in the referential communication task. Shaded squares and rectangles represent blocks and unshaded squares and rectangles represent stickers (which will eventually be replaced with blocks). The scene pictured is midway through the task, so some portions of the partners' boards match, while other regions are not completed yet.

Figure 8
Figure 8

(a) The proportion of fixations to targets, competitors, and other blocks by time (ms) for linguistically disambiguated definite NPs. The graph is centred by item with 0 ms=POD onset. (b) The proportion of fixations for the linguistically ambiguous definite NPs.

Figure 9
Figure 9

Schematic of part of the game board for the ‘questions’ experiment from the perspective of one of the participants. The animals in grey squares are in that participants privileged ground. The animals in the white squares are in common ground, that is, visible to both participants. The black squares contain animals that are only visible to the participant's partner.

Figure 10
Figure 10

The proportion of referents for responses to questions, and declarative questions, categorized by mutuality of the referent, from the viewpoint of (a) the speaker and (b) the addressee.

Figure 11
Figure 11

The average proportion of fixations to privileged ground cards (addressee's perspective) during interpretation of wh-questions and declaratives. Baseline region=1000 ms before the onset of each expression+200 ms; Critical region=1000 ms after expression onset+200 ms.

Figure 12
Figure 12

Hypothetical context for utterance After putting the pencil below the big apple, James put the apple on top of the towel to illustrate the implausibility of standard assumptions about context-independent comprehension. Note that the small (red) apple is intended to be a more prototypical apple than the large (green) apple.

Similar articles

Cited by

References

    1. Allen J.F, Byron D.K, Dzikovska M, Ferguson G, Galescu L, Stent A. Towards conversational human–computer interaction. AI Mag. 2001;22:27–35.
    1. Allopenna P.D, Magnuson J.S, Tanenhaus M.K. Tracking the time course of spoken word recognition: evidence for continuous mapping models. J. Mem. Lang. 1998;38:419–439. doi:10.1006/jmla.1997.2558 - DOI
    1. Altmann G.T.M, Kamide Y. Incremental interpretation at verbs: restricting the domain of subsequent reference. Cognition. 1999;73:247–264. doi:10.1016/S0010-0277(99)00059-1 - DOI - PubMed
    1. Altmann G.T.M, Steedman M.J. Interaction with context during human sentence processing. Cognition. 1988;30:191–238. doi:10.1016/0010-0277(88)90020-0 - DOI - PubMed
    1. Austin J.L. Harvard University Press; Cambridge, MA: 1962. How to do things with words.

Publication types

MeSH terms

LinkOut - more resources