tag:blogger.com,1999:blog-1812431336777691886.post9121599004586273145..comments2024-02-29T03:57:00.088-05:00Comments on The Mermaid's Tale: Lucretius, and stories about the Nature of ThingsAnne Buchananhttp://www.blogger.com/profile/09212151396672651221noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-1812431336777691886.post-55873300199955849192014-03-04T11:10:51.690-05:002014-03-04T11:10:51.690-05:00It's true, but worse than that. The reason is...It's true, but worse than that. The reason is that falsifiability is not a good criterion for inference or establishing truth. One can 'falsify' an hypothesis because of bad study design or measurement. Evolution generates differences among individuals, populations, and species, so failure to replicate is not a necessarily a falsification. <br /><br />Or, many small effects may generate a result and our idea that any one of them we stipulate is involved by the kinds of statistical significance criteria we use, may simply be wrong--the factor may be involved but not enough to be detectable in a given study.<br /><br /> And then there is the Goedel principle that may apply to logical reasoning and empirical work--that there may be truths that our approaches just cannot confirm and we may not be able to know which ones they are.<br /><br />Sometimes the criterion works, and as you say sometimes more persuasive ideas come along--if we're lucky the successor has better empirical support or, if imperfect, seems at least to be more accurate.<br /><br />But sometimes it's just story-telling.Ken Weisshttps://www.blogger.com/profile/02049713123559138421noreply@blogger.comtag:blogger.com,1999:blog-1812431336777691886.post-69174736035841730852014-03-04T11:05:19.444-05:002014-03-04T11:05:19.444-05:00Too many hypotheses aren't falsifiable and are...Too many hypotheses aren't falsifiable and are only thrown out of the professional zeitgeist when better or preferred hypotheses are offered... Holly Dunsworthhttps://www.blogger.com/profile/05260104967932801186noreply@blogger.com