Presentation is loading. Please wait.

Presentation is loading. Please wait.

Inference suppression and moral dilemmas Barbara Kuhnert – University of Freiburg - October 2013 SPP1516 Project: Non-monotonicity, consistency and rationality.

Similar presentations


Presentation on theme: "Inference suppression and moral dilemmas Barbara Kuhnert – University of Freiburg - October 2013 SPP1516 Project: Non-monotonicity, consistency and rationality."— Presentation transcript:

1 Inference suppression and moral dilemmas Barbara Kuhnert – University of Freiburg - October 2013 SPP1516 Project: Non-monotonicity, consistency and rationality in human reasoning 1 Human non-monotonic reasoning and inference suppression under conditions of moral dilemmas

2 What if… 2SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 A runaway trolley is about to run over and kill five people. If a bystander throws a switch then the trolley will turn onto a side track, where it will kill only one person. The bystander throws the switch.

3 What if… 3SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 A runaway trolley is about to run over and kill five people. If a bystander throws a switch then the trolley will turn onto a side track, where it will kill only one person. The bystander throws the switch. 1.The trolley will turn onto a side track, where it will kill only one person. 2.The trolley will not turn onto a side track, where it will kill only one person. 3.Nothing can be concluded

4 What if… 4SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 A runaway trolley is about to run over and kill five people. If a bystander throws a switch then the trolley will turn onto a side track, where it will kill only one person. The bystander throws the switch. 1.The trolley will turn onto a side track, where it will kill only one person. 2.The trolley will not turn onto a side track, where it will kill only one person. 3.Nothing can be concluded

5 What if… 5SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 A boat is about to sink because of overweight. If the crew is told to throw the biggest person into the sea then the boat will not sink and the other three passengers will be saved. If the biggest person weights over 150 lbs then the boat will not sink and the other three passengers will be saved. The crew is told to throw the biggest person into the sea.

6 What if… 6SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 A boat is about to sink because of overweight. If the crew is told to throw the biggest person into the sea then the boat will not sink and the other three passengers will be saved. If the biggest person weights over 150 lbs then the boat will not sink and the other three passengers will be saved. The crew is told to throw the biggest person into the sea. 1.The boat will not sink and the other three passengers will be saved. 2.The boat will sink and the other three passengers will die. 3.Nothing can be concluded

7 What if… 7SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 A boat is about to sink because of overweight. If the crew is told to throw the biggest person into the sea then the boat will not sink and the other three passengers will be saved. If the biggest person weights over 150 lbs then the boat will not sink and the other three passengers will be saved. The crew is told to throw the biggest person into the sea. 1.The boat will not sink and the other three passengers will be saved. 2.The boat will sink and the other three passengers will die. 3.Nothing can be concluded

8 8SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Moral dilemma Suppression task Byrne (1989)

9 Leading thought 9SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Combination of two fields of research: Moral reasoning Non-monotonic reasoning and inference suppression  Inference suppression under morally difficult circumstances.

10 Central questions 10SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Do people make the same inferences and fallacies if they have to reason about moral content? Is it more difficult to make a valid form of inference especially if this inference is morally difficult? Do people hesitate while reasoning in morally difficult decision situations? Do people differentiate concerning the severity of moral decision situations?

11 Overview 11SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Theoretical approach –Non-monotonic reasoning –Suppression task (Byrne, 1989) Basic inferences Argument types –Moral dilemmas

12 Overview 12SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Pilot study –Design overview –Research questions –First results Future perspectives, critical issues and discussion

13 Non-monotonic reasoning 13SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Human reasoning obeys the rules of classical logics only partially. Formal logic  has a monotonic consequence relation, i.e. learning a new piece of knowledge cannot reduce the set of what is known.

14 Non-monotonic reasoning 14SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Non-monotonic logic  formal logic whose consequence relation is not monotonic. New knowledge may contradict old beliefs.  Extend classical logics in order to justify defeasible conclusions from knowledge or belief bases.

15 Non-monotonic reasoning - example 15SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Information 1:Tweety is a bird Information 2:Birds fly Conclusion  Tweety can fly

16 Non-monotonic reasoning - example 16SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Information 1:Tweety is a bird Information 2:Birds fly Conclusion  Tweety can fly New information:Tweety is a penguin New conclusion  Tweety can not fly Everyday reasoning is often non-monotonic (Johnson-Laird, 2010)

17 Suppression task (Byrne, 1989) 17SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013  to investigate non-monotonicity in human reasoning processes In the suppression task subjects were given two premises from which the correct conclusion resp. inference has to be drawn.

18 Suppression task - example 18SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Premise 1 (conditional): A  B Premise 2 (categorical): A Conclusion: B

19 Basic inferences: Modus ponens 19SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Premise 1: If Lisa has an essay to write then she will study late in the library. Premise 2: She has an essay to write Conclusion: She will study late in the library

20 Basic inferences: Modus tollens 20SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Premise 1: If Lisa has an essay to write then she will study late in the library. Premise 2: She will not study late in the library Conclusion: She does not have an essay to write

21 Basic inferences: Denial of the antecedent 21SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Premise 1: If Lisa has an essay to write then she will study late in the library. Premise 2: She does not have an essay to write Conclusion: She will not study late in the library

22 Basic inferences: Affirmation of the consequent 22SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Premise 1: If Lisa has an essay to write then she will study late in the library. Premise 2: She will study late in the library. Conclusion: She has an essay to write.

23 Argument types 23SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Simple Arguments Alternative Arguments Additional Arguments

24 Argument types: Simple argument 24SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Premise 1 If Lisa has an essay to write then she will study late in the library. Premise 2 She has an essay to write Conclusion She will study late in the library

25 Argument types: Alternative argument 25SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Premise 1: If Lisa has an essay to write then she will study late in the library. Alternative premise: If she has some textbooks to read then she will study late in the library. Premise 2: She does not have an essay to write. Conclusion: She will not study late in the library

26 Argument types: Additional argument 26SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Premise 1: If Lisa has an essay to write then she will study late in the library. Additional premise: If the library is open, she will study late in the library Premise 2: She has an essay to write Conclusion: She will study late in the library

27 Moral dilemmas 27SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 We select three current moral dilemmas known from the literature (e.g. Bucciarelli, Khemlani & Johnson-Laird, 2008).  in order to combine the suppression tasks with the central concepts of moral reasoning resp.  to investigate the inference suppression under morally difficult circumstances

28 Moral dilemma 1: Pregnant women 28SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013

29 Moral dilemma 2: Runaway trolley 29SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013

30 Moral dilemma 3: Overweight boat 30SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013

31 Combination: Moral dilemma - suppression task 31SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Example: runaway trolley scenario Introduction sentence A runaway trolley is about to run over and kill five people. Conditional sentence If a bystander throws a switch, then the trolley will turn onto a side track, where it will kill only one person. Categorical sentence (for modus ponens) The bystander throws a switch.

32 Combination: Moral reasoning - suppression task 32SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Example: runaway trolley scenario Introduction sentence A runaway trolley is about to run over and kill five people. Conditional sentence 1 If a bystander throws a switch, then the trolley will turn onto a side track, where it will kill only one person. Conditional sentence 2 If the switch has power then the trolley will turn onto a side track, where it will kill only one person. Categorical sentence (for modus ponens) The bystander throws a switch.

33 Pilot study: operationalization 33SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Questionnaire with 36 conclusion tasks Participants were recruited by self-selection on the online platform Amazon Mechanical Turk 42 participants (w=23 / m=19, mean age 41,3) 3 groups with 14 participants each

34 34 Runaway TrolleyOverweight Boat with 4 inferential figures each MPMT DAAC execute 3 scenarios each (12 tasks per scenario) Pregnant Women MP: Modus Ponens DA: Denial of the Antecedent MT: Modus Tollens AC: Affirmation of the Consequence Additional Argument Alternative Argument Simple Argument MPMT DAAC MPMT DAAC with 3 types of arguments each (4 tasks per argument type) 3x3x4  36 Items Moral DilemmaMoral Dilemma CounterpartNo Moral Dilemma 3 groups MDMDCNMD

35 Variables 35SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Independent variables: –Dilemma –Scenario –Argument types –Inferential figure Dependent variable –Correctness of the conclusion –Personal assessment of the conclusions` accuracy (Likeliness) –Response time

36 Research questions 36SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Implicated in the conclusion (of the groups with moral dilemmas) is e.g. that three persons die instead of one person. John Mikhail (2007) assumed that most people would prefer to let one person die instead of three or five persons, i.e., people base their decision to rescue the most number of people. What happens, if the logically correct conclusion interferes with this utilitarian principle? Can this have an impeding effect?

37 Research questions 37SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 The subjects in the groups with a moral dilemma should be affected by the inconsistency of the logical correct inference and the morally preferable answer.  The correctness of the group with no moral dilemma is higher than the correctness of the groups under conditions of moral dilemma.

38 Research questions 38SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 It is to assume that people need more time to answer and are less confident of the outcome in a morally difficult task than in a task without a moral dilemma.  The subjects of the groups with the moral dilemma will have a higher response time and a lower likeliness than the subjects of the group with no moral dilemma.

39 Research questions 39SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Byrne shows that the context of a situation influences the inferences people make (1989). The context of the tasks with morally difficult inferences and its counterpart inferences is similar.  It is to assume that the response time, correctness and likeliness of the two groups under conditions of moral dilemma do not significantly differ.

40 Research questions 40SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Do people differentiate concerning the severity of the three described moral desicion situations? “Imagine you had to make the decision to save lives, what kind of decision situation would be more difficult to make for you?” “How much more difficult is it for you to find a decision in the chosen decision situation?”

41 Results: Correctness 41SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 1: Average overall correctness per group

42 Results: Likeliness 42SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 2: Average likeliness of all tasks / likeliness of the correctly answered tasks

43 Results: Response Time 43SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 3: Average response time per group

44 Results: Weighting 44SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 4: Weighting  Pregnant Women vs. Runaway Trolley

45 Results: Weighting 45SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 5: Weighting  Pregnant Women vs. Runaway Trolley

46 Results: Weighting 46SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 6: Weighting  Runaway Trolley vs. Overweight Boat

47 Summary 47SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Research aim: Combination of two areas of research inference suppression and moral reasoning  Well known moral dilemmas were transformed into suppression tasks and arranged as questionnaire  Pilot study  Slight tendencies towards group differences  Subjects assess the moral situations differently

48 Future perspectives 48SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Future investigations and analyses are required to refine the questionnaire and its moral dilemmas. A larger sample size is necessary to ensure reliable results and to detect possible group differences Further analyses are required concerning within- group comparisons. (to analyze the effects of the different scenarios, argument types and inferential figures)

49 Critical reflection 49SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Is it even possible to cause a moral feeling with these kind of tasks? Do the subjects sense a moral dilemma, when they answer the questions? Is there any moral reasoning? How could it be possible to cause more intense moral feelings? (pictures / the first-person- perspective?

50 Thanks… 50SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 … for your attention and your interest! …Marco Ragni, Gregory Kuhnmünch and Stefan Wölfl for your tips and advices! … to the SPP for their support!

51 References 51SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Bucciarelli, M., Khemlani S. & Johnson-Laird, P. N. (2008). The psychology of moral reasoning. Judgment and Decision Making, 3(2), 121-139. Byrne, R. M. (1989). Suppressing valid inferences with conditionals. Cognition, 31(1), 61-83. Evans, Jonathan St B. T. (1982). The psychology of deductive reasoning. London ; Boston: Routledge & Kegan Paul. Mikhail, J. (2007). Universal moral grammar: theory, evidence and the future. Trends Cogn Sci, 11(4), 143-152. doi: 10.1016/j.tics.2006.12.007 Wason, P. C. & Johnson-Laird, P. N. (1972). Psychology of reasoning: structure and content. London: Batsford.

52 52

53 53 Examples and further results

54 Experiment: MD - example task 54SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 A pregnant woman is about to give birth to her triplets. If the doctors treat the woman then her triplets will live, but she will die. If the woman has lost too much blood then her triplets will live, but she will die. If a midwife helps the woman to deliver the babies then her triplets will live, but she will die. The doctors treat the woman.

55 Experiment: MDC – example task 55SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 A pregnant woman is about to give birth to her triplets. If the doctors treat the woman then she will live, but her triplets will die. If the woman has not lost too much blood then she will live, but her triplets will die. If a midwife helps the woman to deliver the babies then she will live, but her triplets will die. The doctors treat the woman.

56 Experiment: NMD – example task 56SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 A pregnant woman is about to give birth to her triplets. If the doctors treat the woman then she and her triplets will live. If the woman has not lost too much blood then she and her triplets will live. If a midwife helps the woman to deliver the babies then she and her triplets will live. The doctors treat the woman.

57 Results: demographics 57SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 MDMDCNMDOverall N women 79723 N men 75719 N overall 14 42 MSDMinMax MD36,612,22260 MDC42,812,62770 NMD44,413,42563 Overall41,312,92270 Table 1: sample size and gender distribution Table 2: age

58 Results: demographics 58SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 1: self-assessment - logical abilities Figure 2: self-assessment - grades in math

59 Results: Correctness - Modus Ponens 59SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 3: Average overall correctness per group

60 Results: Likeliness - Modus Ponens 60SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 4: Average likeliness of all tasks / likeliness of the correctly answered tasks

61 Results: Response Time – Modus Ponens 61SPP Workshop FRIAS: Barbara Kuhnert, „Inference suppression and moral dilemmas“ Freiburg 10/2013 Figure 4: Average response time per group

62 62

63 63

64 64

65 65

66 66


Download ppt "Inference suppression and moral dilemmas Barbara Kuhnert – University of Freiburg - October 2013 SPP1516 Project: Non-monotonicity, consistency and rationality."

Similar presentations


Ads by Google