Abstract
There has recently been a wave of attempts to make sense of the role of de se thoughts in linguistic communication. A majority of the attempts assume a Perryan or a Lewisian view of de se thought. Views with these assumptions, I suggest, come in four varieties: uncentering (Egan 2007, Kölbel 2013, Moss 2012), recentering (Heim 2004, Weber 2012), multicentering (Kindermann 2014, Ninan 2010, Torre 2009), and no centering (Kaplan 1989, Perry 1979). I argue first that all four varieties of centering are committed to what I call a shifting operation on the hearer's part. I argue second that, against common assumption, there is no real choice to make between the views. By showing that attempts to establish an advantage for some view over the others fail across the board, I make the case for neutralism regarding the varieties of centering – the claim that coverage of the empirical data is exactly the same for each view, and that the views are broadly equal in simplicity and elegance.