Friday, May 23, 2014

Larry Laudan on seriatim versus v. contemporaneous eyewitness identifications

Recent research has revealed that sequential lineup eyewitness identifications are less likely to falsely identify an innocent suspect as the culprit than are traditional simultaneous lineups. This has led numerous reformers to advocate (and many jurisdictions to accept) that the latter procedure should be replaced by the former. Clark has rightly pointed out that mis-identification data has another twist that almost everyone else has ignored; to wit, sequential lineups are much more likely to lead to false negatives than are simultaneous lineups. 

If, as we have every reason to believe, both types of lineups are more likely to include a guilty party than to include only innocent suspects, there are powerful reasons to say, as Clark does, that sequential lineups are apt to have higher aggregate error rates than simultaneous lineups do. That should give pause to the growing movement to replace the latter by the former.

The dynamic evidence page

Evidence marshaling software MarshalPlan

Thursday, May 22, 2014

How is fuzzy logic doing? Is it passé?

Professor Lotfi Zadeh sent the following message today to his discussion list (BISC):
How is fuzzy logic doing? A significant measure is the number of publications with "fuzzy" in title (annually). My administrative assistant, Ixel Chavez, has compiled the information which follows. Comments are welcome.



 Annual number of publications with "fuzzy" in title (
Google Scholar)1993: 5,030
1994: 5,700
1995: 6,340
1996: 6,620
1997: 6,810
1998: 7,130
1999: 7,650
2000: 7,620
2001: 8,260
2002: 8,650
2003: 9,240
2004: 10,900
2005: 12,300
2006: 13,900
2007: 14,800
2008: 16,000
2009: 17,900
2010: 18,700
2011: 18,900
2012: 18,700
2013: 17,000

Total: 238,150 (20 year total)

The dynamic evidence page

Evidence marshaling software MarshalPlan