<p>According to Lewis, counterfactuals should be evaluated with a weighted similarity relation between possible worlds, when used to analyse causal relations. Lewis is well-known for endorsing a specific desideratum on the similarity relation such that backtracking counterfactuals emerge as false (in standard contexts, under determinism). I argue against a similar desideratum under indeterminism and diagnose a flaw in what I call ‘NB position’, a broadly indeterministic extension of the original Lewisian constraint. A central part of my argument, presented as a formal testing case, draws on a notion of minimal patterns of sequential events. I demonstrate that backtracking counterfactuals are justifiable due to a characteristic type of minimal patterns of sequential events allowed by the broadly indeterministic framework accommodating the operations of both deterministic and probabilistic laws. Therefrom also emerges a new heuristic picture about causation.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Backtracking, Indeterminism, and Minimal Patterns

  • Yujian Zheng

摘要

According to Lewis, counterfactuals should be evaluated with a weighted similarity relation between possible worlds, when used to analyse causal relations. Lewis is well-known for endorsing a specific desideratum on the similarity relation such that backtracking counterfactuals emerge as false (in standard contexts, under determinism). I argue against a similar desideratum under indeterminism and diagnose a flaw in what I call ‘NB position’, a broadly indeterministic extension of the original Lewisian constraint. A central part of my argument, presented as a formal testing case, draws on a notion of minimal patterns of sequential events. I demonstrate that backtracking counterfactuals are justifiable due to a characteristic type of minimal patterns of sequential events allowed by the broadly indeterministic framework accommodating the operations of both deterministic and probabilistic laws. Therefrom also emerges a new heuristic picture about causation.