Is event apprehension language-specific? A comparison of ......Cross-linguistic comparison Main...
1
Is event apprehension language-specific? A comparison of Spanish and German Monique Flecken & Johannes Gerwien University of Heidelberg Background: processes of event encoding Presentation of visual stimulus - initial phase of visual processing (apprehension) -> extraction of ‘gist’ - activation and structuring of concepts to be expressed -> ‘language plan’ - retrieval of word meanings and forms -> ‘formulation’ - articulation (model cf. Bock & Levelt, 1994) Approach: analysis of first fixations on stimulus Placement of first fixation is controlled by an action plan, which is drawn up based on the information extracted during apprehension. First fixations may indicate how detailed the information extracted during apprehension is. Research questions 1. How flexible is the process of event apprehension? 2. How specific is the information extracted during event apprehension? Test case - Manipulation of presentation duration of visual stimuli (cf. Dobel et al. 2007): Four conditions: 100, 300, 500, 700 ms - Cross-linguistic comparison: German and Spanish offer different grammatical means for event encoding. In Spanish, events may be encoded without explicit reference to a SPECIFIC AGENT of an ACTION (impersonal constructions, pro drop, etc.) Design Participants: N = 32 (16 NS of Spanish, 16 NS of German) Materials: 60 photos of everyday events, performed by a male or female actor (a man/woman drawing a house). Materials were pre-tested to ensure homogeneity of event descriptions, and to control for a potential bias of one over the other element (agent, action). Procedure “Describe ‘what is happening’ in the picture; at least try to mention the elements of the scene that you recognized” Pictures appeared randomly in 1 of 4 corners of the screen. Presentation duration manipulated between subjects. Sampling rate: 500Hz Analyses Linguistic data - Speech onset latencies - Specificity of event descriptions - Type of information expressed (related to event or only agent/action?) Eye tracking data - Total number of fixations in different conditions - First fixation latencies in different conditions - Region of first fixations in different conditions (agent/action/“in between”) Results Linguistic data General - Speech onset latencies sign. longer in 100 ms presentation duration - Increase in specificity of event descriptions with longer presentation durations (specific action verbs and objects) 100 ms > 300ms: SOT total ** 300 ms, 500 ms: SOT Ger > Spa * Results ctd. Cross-linguistic comparison Main differences between languages in short conditions (100 ms, 300 ms) - German speakers produce more complete event descriptions in 100 ms, compared to Spanish speakers - Spanish speakers leave out overt reference to agents more frequently than German speakers (overall) Eye-tracking data - First fixation region differs between languages, only in 300 ms condition Discussion 1) First fixation patterns Early overt attention allocation to stimuli is modulated by a) presentation duration of stimulus b) language of the speaker - Distribution of regions fixated first differs between time conditions, whereas the order of elements mentioned remains the same - First fixations do not generally correspond to order of mention - First fixation latencies are significantly shorter in 300 ms --> First fixations in the different conditions may reflect different phases of event apprehension and/or planning processes. 2) Cross-linguistic comparison German speakers fixate a specific scene element (agent or action) more frequently, whereas Spanish speakers fixate the “in between” region most frequently (300 ms and 500 ms). --> First fixation differences between languages indicate top-down effects: Different fulfillments of the task of event construal German speakers construct “who does what?” Spanish speakers draft a global picture of “what is happening?” Take-home message Speakers of different languages extract different types of information when under time pressure. Exposure time to the stimulus modulates the time of the execution of the first fixation. The processes that feed into the linguistic encoding system are specific and flexible. The top-down flow of information is time- and task-dependent. 300 ms: Span more “in between” First Fix * Ger more “action” First Fix * 500 ms: n.s. 700 ms: n.s. 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% German Spanish German Spanish German Spanish German Spanish 100ms 300ms 500ms 700ms Acon speci ficity speci fic acon verb unspeci fic acon verb no acon verb 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% German Spanish German Spanish 100ms 300ms Type of informaon expressed event descripon agent descripon object descripon no response 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% null agent indefinite noun definite noun presentaonal Reference to agents German Spanish * * 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% German Spanish 500 ms Agent Acon In between 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% German Spanish 700 ms Agent Acon In between 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% German Spanish 300 ms Agent Acon In between * * 230 235 240 245 250 255 260 265 270 300ms 500ms 700ms First fixaon latencies Total * - More fixations in longer presentation durations (no fixations in 100 ms) - First fixation latencies sign. longer in 500 ms (300 ms < 500 ms *) 2000 2200 2400 2600 2800 3000 3200 3400 3600 100ms 300ms 500ms 700ms Speech onset latencies Total German Spanish * * Contact information: [email protected] [email protected] *
Transcript of Is event apprehension language-specific? A comparison of ......Cross-linguistic comparison Main...
Is event apprehension language-specific?A comparison of Spanish and GermanMonique Flecken & Johannes Gerwien University of Heidelberg
Background: processes of event encodingPresentation of visual stimulus- initial phase of visual processing (apprehension) -> extraction of ‘gist’- activation and structuring of concepts to be expressed -> ‘language plan’- retrieval of word meanings and forms -> ‘formulation’- articulation (model cf. Bock & Levelt, 1994)
Approach: analysis of first fixations on stimulusPlacement of first fixation is controlled by an action plan, which is drawn up based on the information extracted during apprehension.First fixations may indicate how detailed the information extracted during apprehension is.
Research questions1. How flexible is the process of event apprehension?2. How specific is the information extracted during event apprehension?
Test case- Manipulation of presentation duration of visual stimuli (cf. Dobel et al. 2007): Four conditions: 100, 300, 500, 700 ms- Cross-linguistic comparison:German and Spanish offer different grammatical means for event encoding.In Spanish, events may be encoded without explicit reference to
a SPECIFIC AGENT of an ACTION (impersonal constructions, pro drop, etc.)
DesignParticipants: N = 32 (16 NS of Spanish, 16 NS of German)
Materials: 60 photos of everyday events, performed by a male or female actor (a man/woman drawing a house). Materials were pre-tested to ensure homogeneity of event descriptions, and to control for a potential bias of one over the other element (agent, action).
Procedure“Describe ‘what is happening’ in the picture; at least try to mention the elements of the scene that you recognized”
Pictures appeared randomly in 1 of 4 corners of the screen.Presentation duration manipulated between subjects.Sampling rate: 500Hz
AnalysesLinguistic data- Speech onset latencies- Specificity of event descriptions- Type of information expressed (related to event or only agent/action?)Eye tracking data- Total number of fixations in different conditions- First fixation latencies in different conditions- Region of first fixations in different conditions (agent/action/“in between”)
ResultsLinguistic dataGeneral- Speech onset latencies sign. longer in 100 ms presentation duration- Increase in specificity of event descriptions with longer presentation durations (specific action verbs and objects)
100 ms > 300ms: SOT total **300 ms, 500 ms: SOT Ger > Spa *
Results ctd.Cross-linguistic comparisonMain differences between languages in short conditions (100 ms, 300 ms)- German speakers produce more complete event descriptions in 100 ms, compared to Spanish speakers- Spanish speakers leave out overt reference to agents more frequently than German speakers (overall)
Eye-tracking data
- First fixation region differs between languages, only in 300 ms condition
Discussion1) First fixation patternsEarly overt attention allocation to stimuli is modulated bya) presentation duration of stimulusb) language of the speaker- Distribution of regions fixated first differs between time conditions, whereas the order of elements mentioned remains the same- First fixations do not generally correspond to order of mention - First fixation latencies are significantly shorter in 300 ms
--> First fixations in the different conditions may reflect different phases of event apprehension and/or planning processes.
2) Cross-linguistic comparison German speakers fixate a specific scene element (agent or action) more frequently, whereas Spanish speakers fixate the “in between” region most frequently (300 ms and 500 ms).
--> First fixation differences between languages indicate top-down effects: Different fulfillments of the task of event construalGerman speakers construct “who does what?”Spanish speakers draft a global picture of “what is happening?”
Take-home messageSpeakers of different languages extract different types of information when under time pressure. Exposure time to the stimulus modulates the time of the execution of the first fixation.
The processes that feed into the linguistic encoding system are specific and flexible. The top-down flow of information is time- and task-dependent.
300 ms: Span more “in between” First Fix * Ger more “action” First Fix *
500 ms: n.s.700 ms: n.s.
0%10%20%30%40%50%60%70%80%90%
100%
German Spanish German Spanish German Spanish German Spanish
100ms 300ms 500ms 700ms
Action specificity
specific action verb unspecific action verb no action verb
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
German Spanish German Spanish
100ms 300ms
Type of information expressed
event description agent description
object description no response
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
null agent indefinite noun definite noun presentational
Reference to agents
German Spanish
*
*
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
German Spanish
500 ms
Agent Action In between
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
German Spanish
700 ms
Agent Action In between
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
German Spanish
300 ms
Agent Action In between
*
*
230
235
240
245
250
255
260
265
270
300ms 500ms 700ms
First fixation latencies
Total
* - More fixations in longer presentation durations (no fixations in 100 ms)- First fixation latencies sign. longer in 500 ms (300 ms < 500 ms *)
2000
2200
2400
2600
2800
3000
3200
3400
3600
100ms 300ms 500ms 700ms
Speech onset latencies
Total German Spanish
*
*
Contact information: [email protected] [email protected]
*
![[MS-FSVCA]: File Set Version Comparison Algorithms](https://static.fdocuments.us/doc/165x107/6270b8b707e2f566200ba717/ms-fsvca-file-set-version-comparison-algorithms.jpg)
